Citation
Human information processing for decisions to investigate cost variances

Material Information

Title:
Human information processing for decisions to investigate cost variances
Creator:
Brown, Clifton E., 1947-
Publication Date:
Language:
English
Physical Description:
x, 193 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
Conservatism ( jstor )
Cost accounting ( jstor )
Cost allocation ( jstor )
Cost efficiency ( jstor )
Efficiency decisions ( jstor )
Employee efficiency ( jstor )
Motivation ( jstor )
Simulations ( jstor )
Statistical discrepancies ( jstor )
Variable costs ( jstor )
Accounting thesis Ph. D ( lcsh )
Dissertations, Academic -- Accounting -- UF ( lcsh )
Industrial management -- Decision making ( lcsh )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Includes bibliographical references (leaves 188-192).
Additional Physical Form:
Also available online.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Clifton E. Brown.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
022909022 ( ALEPH )
05267304 ( OCLC )

Downloads

This item has the following downloads:


Full Text










HUMAN INFORMATION PROCESSING FOR DECISIONS
TO INVESTIGATE COST VARIANCES











By

CLIFTON E. BRO14N















A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
,IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY





UNIVERSITY OF FLORIDA 1978




HUMAN INFORMATION PROCESSING FOR DECISIONS
TO INVESTIGATE COST VARIANCES
By
CLIFTON E. BROWN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1978


Copyright 1978
By
Clifton E.
Brown


ACKNOWLEDGEMENTS
I am grateful to my supervisory committee; Rashad Abdel-khalik,
Douglas Snowball, Gary Hoi strum, and Richard Griggs. Without their
valuable contributions this dissertation would have had far more
obscurities.
I am indebted to my experimental assistants; Doug Snowball,
Gary Hoi strum, Ron Teichman, Ed Bailey, John Wragge, Robert Thompson,
Nancy Hetsko, Andy Judd, and Michael Gift (even though he was not on
time). Without their help the experimental procedures used within
this study would not have been possible.
Finally, my greatest debt is to my wife, Sandy. Without her
encouragement and support this dissertation never would have been done.


TABLE OF CONTENTS
ACKNOWLEDGEMENTS iii
ABSTRACT viii
CHAPTER I. INTRODUCTION AND OBJECTIVES 1
Variance Investigation Decision Processes 1
The Research Objectives A
Dissertation Organization 6
CHAPTER II. AREA OF INVESTIGATION 8
Managerial Accounting Concepts 8
Managerial Decisions 8
Management Task Planning and Control 9
Standard Cost Variance Investigation 11
Psychological Concepts 16
Psychophysics 16
Human Information Processing 20
CHAPTER III. RESEARCH METHODOLOGY AND DESIGN 23
General Conceptual Development 23
Decision Situation Structure 23
Available Information Set 25
Individual Information Processing Efficiency 26
Individual Ability to Expand the Information Set 27
General Research Design 28
Selection of Independent Variables 29


General Research Methodology 33
General Experimental Environment and Task 34
Operationalization of Variables 35
Hypothesis Formation 46
Simulation of Optimal Model Performances 46
Simulation of Investigation Decision Performances 49
CHAPTER IV. THE EXPERIMENT 76
Experimental Environment 76
Subjects 79
Experimental Materials 80
Background Information 80
Variance Investigation Decisions 80
Elicitation of Heuristics 83
%
Elicitation of Subject Motivations . . 83
Experimental Procedures 83
Assignment of Subjects to Treatment Conditions 84
Training Phase 85
Experimental Phase 86
Final Debriefing 87
CHAPTER V. ANALYSIS AND RESULTS 88
Summary of Results 88
General Method of Analysis 88
Training Phase Analysis and Results 98
Initial Decision Anchors 98
Decision Anchor Adjustment 102
Individual Decision Model Sensitivity 106
The dj Dependent Variable 106
v


Relative Decision Model Sensitivity 107
Intrinsic Motivation and Decision Model Sensitivity 114
Individual Decision Criteria 116
Relative Decision Criteria 116
Relative Decision Criteria Conservatism 121
Individual Attributes and Decision Criteria 124
Individual Long-Run Decision Efficiency 126
Relative Decision Costs 126
Individual Attributes and Long-Run Decision Efficiency ... 138
Long-Run Decision Efficiency and Training Phase Performance 140
Individual Attributes 141
Subject Grade Point Average 141
Subject Motivations 142
CHAPTER VI. DISCUSSION OF RESULTS 146
Conceptual Development Revi si ted 146
Overall Results 148
Discussion of Results 150
Decision Anchor Selection and Adjustment 151
Individual Decision Model Sensitivity 153
Individual Decision Criteria 156
Individual Long-Run Decision Efficiency 159
Limitations 163
Implications for Accounting 166
Value of Additional Information 166
General Standard Setting Process 167
Future Research 168
APPENDIX A. THEORY OF SIGNAL DETECTION DERIVATIONS 172
vi


APPENDIX B. ORAL PRESENTATION TO ELICITE VOLUNTEER SUBJECTS .... 174
APPENDIX C. BACKGROUND INFORMATION BOOKLET 176
APPENDIX D. PRIOR INFORMATION SHEETS 182
APPENDIX E. SUBJECT HEURISTIC ELICITATION QUESTIONNAIRE 184
APPENDIX F. SUBJECT MOTIVATION QUESTIONNAIRE 186
BIBLIOGRAPHY 188
BIOGRAPHICAL SKETCH 193


Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
HUMAN INFORMATION PROCESSING FOR DECISIONS
TO INVESTIGATE COST VARIANCES
By
Clifton E. Brown
December 1978
Chairman: A. Rashad Abdel-khalik
Major Department: Accounting
Two major functions of management are those of planning and con
trol. Broadly defined, control is the management process which assures
that selected alternatives are implemented and executed in accordance
with plans. Important aspects of the control process concern analysis
and investigation of standard cost variances provided within accounting
reports. A substantial portion of the accounting variance investigation
literature has employed the normative model approach-- researchers have
created variance investigation models which a manager should use. Rarely
has attention been given to how the manager would interprete and integrate
information required by the various normative models.
The principal focus of this dissertation was on the effects of
specific situational variables upon a manager's information processing
for purposes of making variance investigation decisions. The specific
objectives were 1) tc develop a conceptual framework which will predict
effects cf specific situational variables on a manager's relative effi
ciency in information processing and in variance investigation decision
making, and 2) to empirically test some implications of this conceptual
framework.
viii


Variables that can affect the manager's variance investigation
decision process include 1) structure of the decision situation, 2)
contents of the available information set, 3) manager's information
processing efficiency, and 4) manager's learning efficiency. The
structure of the decision situation depends upon number possible states
of control, relative frequencies of the states, various statistical
relationships among the states, and various relationships between the
decision/state outcomes. The contents of the available information set
include information known by the manager prior to his decision. The
manager's information processing efficiency relates to the particular
strategies employed in combining and weighting various items of infor
mation. The manager's learning efficiency refers to the manager's
ability to learn from his experiences with the controlled process. Both
information processing and learning efficiency are the results of the
manager's particular decision strategies. These decision strategies
were expected to adapt a general form, referred to as anchoring and ad
justment. A natural starting point is used as a first approximation for
the investigation decision rule and is then adjusted as the manager
learns from his experiences.
Two experimental methods were used to derive and test implications
of the conceptual framework. Simulation techniques were used to opera
tionalize hypotheses and a laboratory experiment was used to test these
hypotheses.
The experimental environment was that of an assembly department
within a simulated manufacturing company. The assembly department as
sembled a single product and performance of this department was deter
mined completely by the assembly -workers' labor efficiency. Student
IX


subjects, who assumed the role of assembly department operational manager,
made labor efficiency variance investigation decisions based upon a series
of independent standard cost variance reports. The manipulated indepen
dent variables included contents of the available information set,
distributional properties of the states of control, and cost effects of
the investigation decision errors. Parameters of the psychological theory
of signal detection permitted measurement of sensitivity and criteria of
a subject's decision model. The subject's investigation decision costs
compared to normative investigation decision costs derived under similar
situations were employed as a relative measure of decision efficiency.
Overall, the implications of the conceptual framework were supported
by the obtained results. To explain the few major deviations from
expectations, an ex post hypothesis was introduced. This hypothesis
posits that the standard introduces a subjective adjustment bias. Depend
ing upon the direction of adjustment the standard subjectively inhibits
complete adjustment either to optimal decision values close to the
standard or to optimal decision values distant from the standard.
x


CHAPTER I
INTRODUCTION AND OBJECTIVES
Business organizations generate internal accounting reports for
use by managers for purposes of evaluation and decision making.1 One
important evaluation process and decision task of management is that of
standard cost variance analysis and investigation. Based in part upon
standard cost variance reports generated by the internal accounting
system, managers estimate the likelihoods that various production pro
cesses remain under control and decide whether to investigate particu
lar variances. The present study concerns the effects of selected
variables on standard cost variance investigation decision making by
operational control managers. This chapter discusses general aspects
of the standard cost variance investigation decision process and the
variables which may affect the process and presents the research
objectives of the study.
Variance Investigation Decision Processes
An important factor in the evaluation and use of standard cost
variance reports is the manager's perception of the validity of the
xThe American Accounting Association 1966 Statement of Basic Accounting
Theory states that "the objective of accounting for internal use is to
provide information to persons within an organization that enables them
to make informed judgements and effective decisions which further the
organization's goals" (p. 38).
1


2
information provided. In particular, the manager's perception of the
standard setting process will have considerable influence upon his
variance investigation decision process. If the manager believes that
the standards are unrealistic (i.e., that the standards have been placed
too far from the in-control distribution) he may rescale either the
standards or the variances to correspond with his own perception of the
in-control distribution.2
If it is assumed that a manager accepts the standard setting pro
cess as realistic and does not rescale the standards or variances, his
variance investigation decisions may be viewed as the culmination of a
two-stage process. The first stage concerns the detection of the
particular distribution (e.g., in-control or out-of-control) that gen
erated the variance. The manager's performance of this task is a
function of the sensitivity of his decision process (model). This
sensitivity is affected by the structure of the particular situation
and by the manager's knowledge of this structure. The extent of the
manager's knowledge of the situation, in turn, is affected by the
available information (both contained within the variance report and
provided from other sources), and by the manager's ability to learn
from his experiences with the processes being controlled. In situations
where the controlled process distributions have some area of overlap
the results of a manager's detection process are probabilistic (i.e.,
prior to completing an investigation the existence of any given state
is uncertain). Furthermore, the greater the area of overlap, the more
difficult the discrimination task becomes.
2Within this context, an in-control distribution concerns statistical
congruence of production output and planned output in terms of
controllable resource utilization.


3
The second stage concerns the manager's investigation decision
criteria. Having arrived at a conclusion (albeit probabilistic) about
the distribution that generated the variance, the manager must integrate
and process various objective function parameters in order to arrive at
his variance investigation decision (these parameters can belong to
either the manager's objective function, the organization's objective
function, or both if the same). The potential investigation criteria
can be divided into two major categories: structural criteria and
behavioral criteria. The first category can be divided further into
structural cost criteria and structural probability criteria. The
structural cost criteria include such variables as the additional cost
of operating an out-of-control process (given that the process can be
returned to the original in-control state), and the costs of variance
investigation and correction (which can differ depending upon the
actual state that generated the variance). The structural probability
criteria include such variables as the probability that the source of
the variance is controllable (i.e., that it can be returned to the
original in-control state through managerial action), the probability
that the process will return to the in-control state without managerial
action, and the prior probabilities of each state distribution. The
behavioral criteria include such variables as the manager's perception
of the effect of the investigation upon his performance evaluation and
reward structure, and the manager's perception of the effect of the
decision upon employee performances and attitudes.
Although the variance investigation decision process has been
described as two stages, the manager may not actually utilize such a
sequential stage process. The manager's actual decision process is


4
labeled a heuristic. Within this context, heuristic refers to the
learned set of rules or principles that are utilized by the individual
in making the particular decisions required of him. The manager's
specific heuristic can be affected substantially by his individual
characteristics (e.g., intelligence, cognitive complexity and style,
decision process sensitivity, motivations, etc.). However, a general
heuristic, labeled anchoring and adjustment (Tversky and Kahnemar,
1974), is expected to describe the general form of the manager's vari
ance investigation decision process.3
The above examination of the manager's variance investigation
decision process indicates that the variables which can affect both
the process and the results of the process include 1) the structure
of the decision situation, 2) the contents of the available information
set, 3) the manager's information processing efficiency, and 4) the
manager's learning efficiency (from his experiences with the controlled
process). The manager's knowledge concerning the structure of the
decision situation may be limited by the contents of the available
information set and by his information processing and learning effici
encies. The accountant, to a large extent, has control over the con
tents of the available information set.
The Research Objectives
Many problems in accounting reduce to one of choosing among
alternative information sets that could be provided to a decision maker
(American Accounting Association, 1972). Two basic approaches to
3The anchoring and adjustment heuristic is described in greater detail
within Chapter II.


5
deciding on the information set to be presented within the cost variance
report have been advocated. First, the accountant can determine those
models which managers use in making variance investigation decisions and
provide an information set which would permit the implementation of
these models. Problems with this individual model approach are 1) the
possibility exists that different individual models use a wide range of
different information sets (which are not costless), and 2) the indi
vidual models may not be optimal (i.e., they could be inefficient with
respect to other decision models). The individual model approach cen
ters on the psychological question of what the manager is doing (or
would do) with the available (or additional) information. Second, the
accountant could create a normative model of the variance investigation
decision and provide the information required by that model. Problems
with the normative model approach are 1) the information provided may
not optimize the individual's investigation decisions (which may con
tinue to be made using the individual's model), and 2) the costs
associated with operationalization of the normative model may exceed
the benefits (cost savings as a result of more optimal investigation
decisions and greater congruence of manager goals with overall organi
zation goals) offered by the model. The normative model approach
centers on the analytical question of what the manager should be doing
(or should do) with the available (or additional) information.
A substantial portion of the accounting variance investigation
literature has focused upon the second approach-- the normative model
approach.4 However, rarely has attention been given to how the manager
4This literature is reviewed within Chapter II.


6
would interprete and integrate the information sets of the various
normative (optimal) variance investigation models. In many instances
the implicit assumption has been that the manager would process the
information with the same efficiency as the normative model.
The principal focus of this research is the effects of specific
situational variables on a manager's variance investigation decisions
(relative to the investigation decisions of an optimal model) and on
a manager's processing of available information (relative to the infor
mation processing of an optimal model). Specific objectives are:
1) To develop a conceptual framework which will predict the
effects of situational variables on a manager's relative
information processing efficiency and to empirically test
the implications of this conceptual framework.
2) To develop a conceptual framework which will predict the
effects of situational variables on a manager's relative
variance investigation decision efficiency and to
empirically test the implications of this conceptual framework.
Dissertation Organization
Chapter II presents certain general concepts from accounting and
psychology. These concepts are necessary for the development of the
specific environment studied in this research, and for the development
of a conceptual information processing and decision making framework
within this environment. The accounting concepts are concerned with the
nature of managerial decisions, managerial task planning and control,
and standard cost variance investigation. The psychological concepts
are concerned with the theory of signal detection and with cognitive


7
aspects of human information processing such as the general heuristic
of anchoring and adjustment.
Chapter III synthesizes the general concepts of the previous
chapter and develops a conceptual framework of standard cost variance
investigation employing the methodology of the psychological concepts.
This conceptual framework was operationalized using simulation tech
niques, and experimental hypotheses were derived from the simulations.
The details of a laboratory experiment designed to test the
hypotheses derived from the conceptual variance investigation framework
are presented in Chapter IV. The experiment employed a between-subjects
design and was manipulated factorially using specific variance investi
gation situation variables. Modified parameters from the psychological
theory of signal detection were employed to measure subject's information
processing efficiency and decision model sensitivity.
The results obtained from the laboratory experiment are presented
in Chapter V. The model comparison procedure of non-orthogonal analysis
of variance was the primary method of analysis employed. Chapter VI
discusses the results, develops a modification (ex post) of the
original conceptual framework, and discusses some implications for
accounting and for future accounting research.


CHAPTER II
AREA OF INVESTIGATION
Managerial Accounting Concepts
This study relies on the synthesis of certain psychological con
cepts with certain accounting concepts. This section discusses the
accounting concepts: the nature of managerial decisions, managerial
task planning and control, and standard cost variance investigation.
Managerial Decisions
Since the information set (accounting report) exists to support
the manager's tasks, a conceptual framework of managerial decisions would
facilitate the accountant's information set selection. A conceptual
framework of managerial decisions should differentiate managerial de
cisions along dimensions that allow insights into the informational
needs of those decisions.
Anthony (1965) provides one such dimension along which he iden
tifies the purposes or orientation of managerial activities: strategic
planning, management control, and operational control. Another dimen
sion is provided by Simon (1966) who distinguishes between programmed
and nonprogrammed decisions. The underlying dimension is concerned
with the manner in which managers deal with their problems. The cri
teria for classifying a decision consist of the extent of structure
associated with the problem solving phases of the decision.
8


9
A conceptual framework of managerial decisions that synthesizes
these dimensions is presented by Gorry and Morton (1971). Simon's
dimension is modified by replacing the terms "programmed" and "non-
programmed" with the terms "structured" and "unstructured" and adding
a third category labeled semi-structured. Gorry and Morton's conceptual
framework is used in this research as a model that will permit the
identification of a standard cost system within the overall managerial
decision framework.
Management Task Planning and Control
Standard cost systems present information sets to managers to
support various decisions concerning task planning and control.
Demski (1967) provides an excellent conceptual discussion of the manage
ment task planning and control process, and his approach (with some
modification) is adopted in this research.
Figure 1 presents a model of the management task planning and
control process. Environmental information, largely external to the
standard cost system, facilitates the planning of overall goals and
policies within the constraints imposed by the environment. Overall
objective control feedback 1) provides evidence for the ccntinuted
validity of overall assumptions made in forming the organization ob
jectives, and 2) facilitates strategic planning evaluation of organi
zation performance.
Both the environmental variables and the overall task control
feedback of the management control activity 1) provide evidence for
the continued validity of the overall task assumptions made in forming


Overall
Objective
Control
Strategic
Planning
Overall
Objective
Planning
Overal1
Task
Control
Management
Control
Overall and
Specific Task
Planning
Specific
Task
Control
V
Operational
J-
Control
Specific Task
Structuring
Source: Modified from Demski, 1967.
Environment
Output
FIGURE 1
MANAGEMENT TASK PLANNING AND CONTROL PROCESS


11
the task plans, and 2) facilitate management control evaluation of
operational control performance.
The operational control manager uses the specific task control
feedback to 1) decide whether the physical system performance is in
agreement with the specific task plan, and 2) to decide whether the
task can be restructured to bring performance back into agreement with
the plan (if performance and the plan differ). A task that can not be
restructured could have significance for the overall task control feed
back to the management control activity and possibly may lead to
modification of the original task plans. Such modification could in
turn have significance for the overall objective control feedback to
the strategic planning activity, and a modification of the original
overall objectives may follow.
Standard Cost Variance Investigation
Within the operational control activity the specific task plan
provides standards of performance in terms of expected component costs
and usages, and the desired physical outputs. The operational control
manager structures the tasks within the physical system and periodically
receives a standard cost variance report describing the system output
in terms of the task plan and the actual results. Upon receiving a vari
ance report the manager must decide the nature of the given variances.
Dopuch et al. (1967) present a classification of standard cost
variances that is based on the expected source of the variances. A
variance resulting from a random fluctuation of the physical system,
labeled a Type 1 variance, requires no operational control response if


12
not statistically significant. Whether a variance resulting from a
change in the physical system, labeled a Type 2 variance, requires an
operational control response depends on whether the underlying cause of
the change is a temporary rather than a permanent phenomena. A tem
porary or controllable variance, labeled a Type 2a variance, is one
that operational control can correct in the future (i.e., the physical
system can be returned to the previous in-control condition). A per
manent or noncontrollable variance, labeled a Type 2b variance, is one
that operational control can not correct in the future.
If the manager decides that the variance is of Type 1 no action
is required. If, however, the manager decides that the variance is of
Type 2 he must then decide if the variance should be investigated for
its underlying causes. Should the variance be of Type 2b it would
have little operational control significance; should the variance be
of Type 2a the cause of the variance could be eliminated by restructur
ing the physical system. The present study confines itself to the
assumption that all variances which result from a change in the physical
system are controllable by the operational control activity. In other
words, it assumes that the standard cost variances have no significant
affect on either the strategic planning or the management control
activities.
The information set contained in a standard cost variance report
can be viewed as relating to one of two categories of information:
distributional properties of system states and investigation decision
criteria. The distributional properties category can include such
variables as 1) the mean and standard deviation of the output when the
system is known to be in-control, 2) the mean and standard deviation of


13
the output when the system is known to be out-of-control, 3) the actual
output for the period, and 4) the deviation between the standard output
and the actual output for the period (the standard output variance).
The investigation decision criteria category can include such variables
as 1) the prior probabilities of the system's states, and 2) the
costs and benefits associated with each possible investigation decision
in combination with each possible system state.
Information relating to investigation decision criteria that is
not typically presented within the variance report includes 1) the
probability that the cause of the variance is controllable (assumed by
this research to equal one), 2) the probability that the system will
return to the in-control state without managerial action (assumed by
this research to equal zero), 3) the perceived effect of the investi
gation decision on manager performance evaluation, and 4) the perceived
effect of the investigation on employee performances and attitudes.
The variance investigation literature is concerned mainly with
the investigation significance of Type 1 and Type 2a variances. A
general approach described within the literature is that of the Shewhart
X chart procedure (Probst, 1971; Koehler, 1968; Luh, 1968; Jeurs, 1967;
Zannetos, 1964). This approach, based upon classical statistics,
involves sampling the system to construct an in-control mean and stan
dard deviation. Arbitrary control limits (generally plus or minus three
standard deviations) are used as the criteria for making a variance
investigation decision. Only one of the above studies considered the
effects of the information provided on the manager's investigation
decision. Probst (1971), in an industrial field study, found foremen
unwilling to accept the procedure as their investigation decision rule


14
when the procedure was constructed objectively (the foremen indicated
the reason for not accepting the procedure was that they felt it
ignored their experience). However, Probst did not analyze his results
in terms of the performance efficiency of the foremen.
Other research has noted what are considered to be significant
drawbacks to the Shewhart X chart procedure: the failure to consider
the costs and benefits associated with the various investigation de
cision outcomes and the failure to consider information from prior
periods. Bierman et al. (1961) proposed the incorporation of investi
gation decision costs and benefits within a classical statistics frame
work. Thus managers may use statistical information to calculate the
probability that the variance reflects the in-control state, and com
bine this probability with the investigation decision costs and benefits.
Most of this information is assumed to be provided by the accountant
and accurately processed by the manager.
Kaplan (1975) and Jacobs (1978) have considered another deficiency
of the classical statistics approach-- the failure to use prior period
information. They analyze the use of the cummulative sum procedure,
whereby the cumulative sum of the variance is charted for each period.
Theoretically, under a stable state these sums should follow a random
walk, and any drift would indicate the system is out-of-control. Eval
uation of such a drift would be accomplished on the basis of information
provided by the accountant or derived from the manager's experience.
The economic cumulative sum procedure incorporates the effects of
estimated investigation decision costs within the drift evaluation.
Several studies in the variance investigation literature use
decision theory to construct a variance investigation model (Kaplan,


15
1969; Dyckman, 1969; Kaplan, 1975; Dittman and Prakash, 1978). If
all the parameters required by these models were available these
models could replace the manager as the variance investigation decision
maker. However, to the extent that not all the parameters are opera-
tionalizable (given seme cost constraint), the manager must continue
to make the variance investigation decisions using the information
provided by the accountant and by his own experiences.
A conclusion that can be drawn from the variance investigation
literature is that the additional information proposed for the use of
managers is becoming both diverse and complex. Furthermore, there is
a paucity of research relating to the manager's ability and efficiency
to interpret and integrate this additional information. Demski (1970,
1971) was one of the first to note the problem associated with the
decision-implementation interface (i.e., the effect of the control
system information on individual behavior and performance). Some
analytic research on this problem has followed. Using simulation
techniques, Magee (1976) compared average total cost for seven decision
rules under various operational conditions. Although the simpler
decision models (rules) tended to have larger average costs than the
more sophisticated decision models, the difference in average cost was
relatively small. Indeed, if model implementation and information costs
are considered the simpler models may be more efficient than the
sophisticated models. Magee also noted that because of the use of
different manager performance measures (average operating costs, average
number of months below standard, etc.) simpler decision models may re
sult in rational choices (i.e., choices that maximize the manager's
expected utility). Closely related to Magee's research is an empirical


16
study by Magee and Dickhaut (1977). Using human subjects, Magee and
Dickhaut found support for the proposition that different manager per
formance (payoff) measures will affect the specific decision rules
(heuristics) employed by the decision maker (as a result of the decision
maker attempting to maximize his subjective utility).
Psychological Concepts
This section discusses the psychological concepts employed in
developing a conceptual framework of manager standard cost variance
investigation decisions. These concepts include the psychophysical
theory of signal detection and the human information processing con
cepts of decision heuristics.
Psychophysics
Psychophysics studies the relationships between physical and
psychological scales of measurement. Modern psychophysics adopts the
view that subjects can make meaningful evaluations of the magnitudes of
their sensory experiences, and therefore sensory magnitudes, as well
as physical magnitudes, can be quantified. One approach of modern
psychophysics is based upon the theory of signal detection (TSD). TSD
permits the separation of the decision maker's ability to discriminate
between classes of stimuli (sensitivity) from his motivational response
biases (decision criteria). Two comprehensive theoretical descriptions
of TSD are presented by Green and Swets (1974) and Egan (1975); general
surveys of the TSD theory are presented by Coombs et al. (1970), Watson
(1973), and Pastore and Scheirer (1974).


17
The basic TSD experiment utilizes the single-interval procedure.
This procedure consists of a series of trials, each trial consisting of
an observation interval and a response interval. The possible stimulus
events during the observation interval are 1) the observation contains
a meaningful signal added to a background of noise (sn trial), and 2)
the observation contains only a background of noise (n trial). It is
assumed that each trial is independent of all other trials and that
the prior probabilities of n and sn are given and remain constant. The
background noise fluctuates at random from trial to trial; the stimulus
(usually a fixed level) is added to the noise. Therefore, the obser
vation fluctuates randomly from trial to trial. The task of the subject
is to detect whether the observation was generated by the signal plus
noise (sn) or by the noise alone (n) distribution. That is, in the
response interval the subject will respond with either "Yes, the signal
was present" (Y response), or "No, the signal was not present" (N re
sponse).
On any trial there exist four possible outcomes of the subject's
decision in conjunction with the actual distribution: 1) sn was pre
sented and the subject said "Yes" (a hit), 2) sn was presented and the
subject said "No" (a miss), 3) n was presented and the subject said
"Yes" (a false alarm), and 4) n was presented and the subject said
"No" (a correct rejection). A conditional probability matrix for a
series of these events is given by the following (Green and Swets, 1974):
RESPONSE
Yes No
:3 sn P(Y|sn) P(N|sn)
rs
~ n P(Y|n) P(N|n)
UO


18
Since the cells of this matrix are both exhaustive and mutually exclu
sive, the row-wise conditional probabilities must sum to one (this does
not necessarily hold for the sum of the column-wise conditional
probabilities). All parameters of the TSD model are derived from this
conditional probability matrix (see Appendix A for a more detailed
discussion of the conditional probability matrix including its rela
tionship with Bayes1 theorem).
In the single-interval task the subject analyzes the evidence
and classifies the stimulus into one of two categories according to
his criteria. The criteria are determined by his objective function.
The objective funcition of interest within this research is the
maximization of expected value. Assume that the subject has some value
(utility) fqr each of the four event outcomes. A payoff matrix of these
values related to the four outcomes is. given by the following (Egan,
1975):
DECISION
No
Yes
sn ^sn,Y ^sn,N
n vn,Y vn,N
The decision rule for maximizing the expected value (see Appendix A
for a derivation of this decision rule) is:
P(x|sn) P(n) V,N Vn.Y
P(xjn) P(sn) V$n,Y ^sn,N
At the point at which the above expression is an equality the subject
should be indifferent between saying "Yes" or saying "No." This point
can be considered the critical value of the likelihood ratio of the
observations, L(x0). The decision rule for saying "Yes" is expressed
by the relation L(x) > L(x0).


19
The critical value L(x0) for this decision rule has two possible
values: a theoretical value which is a measure of the criteria of an
optimal (or ideal) subject and a subjective value which is a measure
of the criteria of an actual subject.
One set of TSD models assumes that both conditional probability
distributions are Gaussian; i.e., $(x|n) and $(x|sn) are normally
distributed. With the additional assumption of equal variance for both
distributions, the parameters which measure individual discrimination
sensitivity and individual decision criteria are labeled d1 and g,
respectively. The discriminability measure, d', has the following
theoretical definition:
where, ysn= the mean of the signal plus noise distribution;
vn = the mean of the noise alone distribution;
a = the standard deviation of both distributions;
zn = the value of the normal distribution function associated
with the noise alone distribution and any decision axis
cutoff value common to both distributions; and
z = the value of the normal distribution function associated
sn
with the signal plus noise distribution and any decision
axis cutoff value common to both distributions.
The d' measure theoretically is independent of the decision criteria
measure.
The decision criteria measure, 3> has the following theoretical
definition:
3 = i>(zsn) / 4>(zn)


20
where <¡>( ) denotes the normal density function for the point in
parentheses and the z parameters are the same as defined for the d'
measure.
Although the assumption of equal variance normal distributions
is employed within this research, such an assumption is not necessary
to employ TSD. Egan (1975) demonstrates the use of TSD with exponential
distributions, chi-square distributions, Bernoulli distributions, and
Poisson distributions. Grier (1971) develops nonparametric measures
of discriminability and decision criteria.
Traditionally, psychophysics has employed TSD to study perceptual
processes: i.e., sensory processes such as audition and vision. Over
the last decade, however, TSD has been applied to conceptual processes.
These extensions to conceptual processes have included numerical
processing (Lieblich and Lieblich, 1969; Hammerton, 1970; Weissmar. et al.,
1975), medical diagnosis (Lusted, 1969; Lusted, 1971; Swets, 1972),
conceptual judgement (Ulehla et al., 1967a; Ulehla et al., 1967b), and
memory (Bernbach, 1967; Banks, 1970).
Human Information Processing
Human information processing (HIP), a subset of cognitive psy
chology, studies human judgement and decision making with particular
emphasis on the processing of information that determines these
activities. An area within HIP is the construction of models of human
decision making. The work within this area is typically classified
into schools of research which employ different paradigms, the two
major paradigms being the Bayesian and regression approaches (Slovic
and Lichtenstein, 1971). The TSD model is related to the Bayesian approach.


21
The Bayesian approach is a normative model specifying how a
decision should be made given certain internally consistent relation
ships among probabilistic beliefs. The basic beliefs of this approach
are that decisions should be based on subjective probabilities and
that these probabilities should be revised upon the receipt of addi
tional information in accordance with Bayes's theorem.
The major findings of Bayesian research are labeled conservatism.
The subjects, after receiving additional information, revise their
posterior probabilities in the same direction as the optimal model but
the revision is insufficient. Much of the research has focused on an
explanation of the cause of conservatism, the major explanations being
misperception (Peterson et al., 1968), misaggregation (DuCharme and
Peterson, 1968), and response bias (DuCharme, 1970).
Another area within HIP is the study of subjective information
processing principles and decision rules, labeled heuristics. A
heuristic, within this context, refers to a learned set of rules or
principles which are utilized by individuals in making the particular
decisions required of them. Research in this area has been concerned
with identifying systematic biases (relative to some definition of
optimal) of subjective heuristics within certain types of decision
tasks. Those information processing and decision rule biases iden
tified thus far have been labeled as an anchoring and adjustment
heuristic (Tversky and Kahneman, 1973), a representative heuristic
(Kahneman and Tversky, 1972; Swieringa et al., 1976), an avail-
ablility heuristic (Tversky and Kahneman, 1973), and the law of small
numbers (Tversky and Kahneman, 1971).


22
This study will make particular use of the anchoring and
adjustment heuristic. In many situations, individuals first make
decisions by starting with an initial anchor (decision point) and then
adjust this initial anchor as they learn from their experiences. The
initial anchor can be suggested by the structure of the decision
situation, or can be the result of a partial computation or estimate.
Empirical tests involving the anchoring and adjustment heuristic in
dicate individuals do not sufficiently adjust their initial decision
point. That is, their adjustment is less than that which would allow
optimal processing of the available information (Slovic and Lichtenstein,
1971; Slovic, 1972; Alpert and Raiffa, 1968; Tversky and Kahneman,
1974).


CHAPTER III
RESEARCH METHODOLOGY AND DESIGN
General Conceptual Development
A general problem confronting decision makers is that the indi
vidual must decide subjectively which state of nature is most probable
based upon some incomplete set of information. When an individual
deals repetitively with a similar situation his long-run decision
efficiency or "decision correctness" can be affected by 1) the struc
ture of the particular decision situation, 2) the contents of the set
of available information, 3) his efficiency in processing the available
information, and 4) his ability to expand the available information
through experience with, and observations of, the various states of
nature. Each of these four elements is discussed below with specific
reference to the objectives of this research.
Decision Situation Structure
The structure of the particular decision situation primarily de
pends upon several key variables. These variables include the number of
possible states of nature, the relative frequencies of the states, the
various statistical relationships among the states, and the relation
ships between the various decision outcomes (the costs incurred given a
specific decision and the existence of a specific state). Depending on
23


24
the specific values which these variables may assume, the structure of
the particular decision situation can affect both the difficulty and
the importance of the individual's discrimination among the states. In
general, as the number of possible states increases and as the area of
distributional overlap of these states increases the discrimination task
becomes more difficult. Furthermore, the discrimination task becomes
more important (in terms of incurred costs) as the relative frequencies
of the states become equal and as the costs associated with the possible
decision outcomes which involve decision errors (an incorrect decision
given the existence of a specific state) become unequal.
The general situation employed in this research, selected for
its relevance to the accounting discipline, is the cost variance
investigation decision. In this situation two states of nature are
possible: 1) in-control (the underlying physical process described by
the standard cost variance report is functioning as planned), and 2)
out-of-control (the underlying physical process described by the stan
dard cost variance report is not functioning as planned). In reality,
the possible states of nature may be located on a continuum whose end
points are the states of in-control and out-of-control. Between these
two end points are any number of states that take the general form of
partially out-of-control or moving out-of-control. In the present
research the possible states of nature are confined to the two end
points. The relative frequencies of the two states are controlled as
constants with their values being close to equal. The statistical
relationships among the two states and the relationships between the
various decision outcomes are manipulated as independent variables.


25
Available Information Set
The contents of the available information set refers to the infor
mation known by the individual prior to his decision. Such information
can be of two types-- singular or distributional (Tversky and Kahneman,
1977). Singular information consists of information that specifically
relates to the current decision. Distributional information consists
of information that relates to the relative frequencies and to the
statistical relationships among the states of nature.1 The difficulty
of the discrimination task may be affected by the presence or absence
of certain items of information. In general, the less information con
tained in the available set the more difficult the discrimination task,
for missing information required by a decision model must be estimated
by the individual. Compared to statistically derived estimates, these
subjective estimates are likely to have greater uncertainty and in
efficiency associated with them.
Within the variance investigation situation, the information con
tained on each variance report constitutes the singular information.
Two types of singular information are employed in this research-- the
actual results of the physical process and its variance from a standard,
and the marginal costs associated with each of the two possible decisions
in combination with each of the two possible states of nature. The
presence of both these singular information types is controlled as a
constant within this research. The presence or absence of specific
xThis definition of distributional information implicitly assumes that
the relative frequencies and statistical relationships are stable over
the relevant time frame. If these variables were non-stable, revised
estimates of their specific values would be required prior to each
decision, thus classifying them as singular information. This research
assumes that both of these variables are stable across all decisions.


26
distributional information items (the statistical means, variances, and
distributional shapes of the two states) is manipulated as an indepen
dent variable. The presence of other distributional information items
(the relative frequencies and the allowed standard) is controlled as a
constant.
Individual Information Processing Efficiency
The individual's efficiency in processing the available infor
mation relates to the particular heuristics or strategies employed in
combining and weighting the various items of information. The term
efficiency implies a relationship between the individual's process
output (his decision) and a normatively correct or optimal decision.
The individual's decision and information processing performance can be
evaluated by comparing his performance against an optimal model.
Optimality refers to the best possible performance under given condi
tions. Since the optimal decision model relies on an incomplete infor
mation set rather than certain knowledge,even its performance can be
affected by both the structure of the situation and the contents of the
available information set.
The optimal models used in this research are a function of the
experimental environment: the single-interval procedure found within
the psychological theory of signal detection and the basic decision
situation of standard variance investigation. Within this environment
the optimal decision rule for minimizing (maximizing) the expected cost
(value) of a set of decisions is based upon an extension of Bayes'
theorem that takes into account the relative costs (values) of various


27
possible decision outcomes. The parameters required to fit the optimal
model include 1) the relative frequencies of the two states of nature,
2) the mean of each state, 3) the statistical variance of each state,
and 4) the costs associated with each of the two possible decisions in
combination with each of the two possible states of nature. Since the
individual's decisions are to be evaluated by comparisons with the out
puts of the optimal model, it would seem reasonable that the information
available to the optimal model be the same as the information available
to the individual. Consequently, for decision situations in which some
of the parameters required by the optimal model are not contained in the
available information set the optimal model must make estimates of the
missing parameters.
Individual Ability to Expand the Information Set
The individual's ability to expand the available information set
over time refers to his ability to learn from his experiences with the
states of nature. Such learning can occur through improved estimates
of unknown items of distributional information and through modifications
of information processing strategies to incorporate state relationships
which were unknown or undetected previously.
The general form of the expected individual information processing
strategy is described by the heuristic of anchoring and adjustment
(Tversky and Kahneman, 1974).2 A natural starting point, or anchor, is
used as the first approximation for the decision. This anchor is then
2A heuristic, within this context, refers to a learned set of rules or
principles which are utilized by an individual in making the particular
decisions required of him.


28
adjusted as the individual learns from his experiences with the states
of nature.
Within the present study the initial decision points or anchors
were expected to fall near the geometric intersection (either actual or
estimated, depending upon the available information) of the distribu
tional curves of the two states. Adjustments from these initial deci
sion points by the individual were expected to occur during the train
ing phase of the experiment as the individual gained experience with
the states and received feedback as to his performance (relative to an
optimal model). The extent of an individual's adjustments (his learn
ing efficiency) is measured using several variables. Each variable
measures the relative extent of adjustment (or lack of adjustment) from
the original decision anchor tov/ard the optimal value of that variable.
General Research Design
As stated previously, the focus of this research is on human
decision making and information processing within a particular decision
context-- that of standard cost variance investigation. Much of the
variance investigation literature within accounting has focused on how
the decision maker should integrate the available information and make
the investigation decision (see Kaplan, 1975 for a review of this
literature). Very little research has been concerned with how the
decision maker does accomplish these processes. The major objective
of this research is to study the processes used by the decision maker
in reaching variance investigation decisions. In particular, this re
search will examine within the conceptual framework discussed previously:


29
1) the effects of the decision situation structure, the available infor
mation set contents, the individual's information processing efficiency,
and the individual's learning efficiency on the individual's long-run
decision efficiency; and 2) the effects of the decision situation
structure and the available information set contents on the individual's
information processing and learning efficiency.
Selection of Independent Variables
The effects of two types of variables are of primary interest
within this study: these are the situation variables and the process
variables. The following discussion describes the selection of each
of the variables employed in this study.
The situation variables
The situation variables are the quantity of information available
to the individual prior to his decision, the statistical structure of
the two states of nature, and the cost structure of the possible decision
outcomes.
The available information set. The effects of the contents of
the available information set are studied by manipulating the presence
and absence of certain distributional information. This involves the
specification of two levels of available information set content. Since
the difficultly of the discrimination task may be increased by the
absence of certain information, one level of the information variable
has less distributional information than the other level. This inde
pendent variable is labeled the information variable.


30
The statistical structure. The effects of the statistical struc
ture of the decision situation are studied by manipulating a distribu
tional information variable. This involves the specification of two
levels of statistical relationship among the two states. Since the
difficulty of the discrimination task increases as the area of dis
tributional overlap between the two states increases, one level of the
distribution variable will have a greater area of overlap than will
the other level. This independent variable is labeled the distribution
variable. All other distributional information items, if presented,
will be controlled as constants.
The cost structure. The effects of the cost structure of the
possible decision outcomes are studied by manipulating a singular in
formation variable. The manipulation of this variable involves the
specification of two levels of decision outcome relationships (in terms
of incurred costs). This independent variable is labeled the cost
variable. Since the importance of the discrimination task increases as
the costs associated with decision outcomes that involve decision errors
become unequal, the different levels of the cost variable will be
associated with different decision error costs. One level of the cost
variable is structured in favor of more variance investigations and the
other level of the cost variable is structured in favor of fewer vari
ance investigations. The only other singular information variable, the
actual results of the physical process and its cost variance, is the
primary experimental stimulus. This information item will be a random
variable whose distributions, given either of the two states of nature,
will be normally shaped with parameter values defined by the appropriate
level of the distribution variable.


31
Independent variables interaction. The research design of this
study is a factorial design in which the above three independent vari
ables, each at two levels, are fully crossed, thus producing eight (23)
independent variables combinations or treatments. The factorial
combination of the different levels of the independent variables adds
power to the research design by permitting the examination of the
effects of interactions upon the dependent variables.
The process variables
The effects of an individual's information processing and learn
ing efficiency are studied using certain observation variables measured
on a continuous (ratio) scale. Continuous measurement facilitates the
use of these variables as both independent and dependent variables.
When used as independent variables the measures are treated as random
components rather than as discrete classification levels. The effects
are studied using three major types of variables: 1) individual de
cision model sensitivity (relative effect of a variable response range),
2) individual decision criteria (relative effect of conservative ad
justment), and 3) relative initial decision anchor and final decision
anchor relationships. The first two types of variables are derived
from model parameters within the theory of signal detection.
A summary of the general research design
The general research design is depicted in Table 1. The design
is presented in terms of the dependent and independent variables to be
included within the research. More specific extensions arid


32
TABLE 1
GENERAL RESEARCH DESIGN IN TERMS OF EXPERIMENTAL VARIABLES
Dependent Variables
Independent Variables
Individual
Long-Run
Decision
Efficiency
Individual
Information
Processing and
Learning
Efficiency
Contents of the available infor
mation set
X
X
Structure of the decision situation:
Statistical relationships
X
X
Decision outcome relationships
X
X
Individual information processing
and learning efficiency
X
Note: An "X" within a cell indicates that the relationship of the
variables concerned are included within the experimental design.


33
operationalizations of these concepts are discussed in a later section
of this chapter. The objective at this point is to summarize in general
terms the relationships which will be included within the experimental
design.
General Research Methodology
Research methodology pertains to the general procedures or
methods employed in conducting research. Two general methods are
employed in this research-- simulation and laboratory experimentation.
The major objective of the simulation is to produce hypotheses which
will predict the behavior of human decision makers within the decision
situation assumed by the simulation. The simulation is based upon
assumptions derived from the conceptual development and from the gen
eral task environment of the laboratory experimentation. These
assumptions and the task environment are discussed in greater detail
in a later section of this chapter. The major objective of the
laboratory experiment is to test the conceptual development (through
the hypotheses derived by the simulation) of the effects of the vari
ous independent variables on the various observation variables.
The general and specific research design employed in the labora
tory experiment is the same as that employed within the simulation.
The major difference between the simulation and the laboratory experi
ment is the use of human subjects. This difference creates two major
sources of incompatability betv/een the simulation and the laboratory
experiment. First, the simulation makes certain assumptions concerning
human behavior and applies these assumptions consistently. Within


34
the laboratory experiment such behavior is not necessarily applied
consistently. Consequently, greater variance of results is expected
within the laboratory experiment than within the simulation. Second,
it should be noted that certain variables are largely affected by
individual attributes. Without a theory of the effect of individual
attributes any simulation of these variables would be arbitrary.
Accordingly, not all of the experimental variables are included in the
simulation. Hypotheses- concerning the variables not present within
the simulation either are derived from limited concepts concerning the
effects of individual attributes or are stated in an exploratory manner
(no expected difference).
General Experimental Environment and Task
The standard cost variance investigation situation studied within
this research is set in an environment of a manufacturing company.
More specifically, the subjects are asked to assume the role of the
operational manager of an assembly department which assembles a single
product, a metal folding chair. The operating efficiency of the
assembly department is determined completely by the labor efficiency of
the assembly workers.
Each subject receives a sequential series of standard cost vari
ance reports and is asked for each report to decide whether to investi
gate or not to investigate the reported labor efficiency variance.
The efficiency of a subject's decision performance and the amount of
payment he will receive for participating in the experiment is based
upon the total investigation decisions cost which he incurs over the
series of variance reports.


35
Each standard variance report is concerned with the results of a
single job-order to produce a constant number of chairs and reports
only aggregate (overall assembly department) results. Each report
contains the aggregate standard assembly time allowed per chair, the
actual assembly time incurred per chair, the overall labor efficiency
variance per chair, the total number of chairs produced, and the costs
associated with each possible decision in combination with each
possible state of nature. All time units are presented in minutes.
The singular information contained in a variance report is independent
of that contained in previous variance reports.
The experiment is conducted in two phases-- a training phase
and an experimental phase. The training phase consists of three
contiguous sessions in which the subject learns his role and presumably
develops his decision strategy. Performance feedback is given at the #
completion of each training session. The experimental phase consists
of a single session in which the subject receives a series of variance
reports similar to those presented in his training session. In this
phase no performance feedback is given until after the completion of
the entire experiment. The subject is paid according to his performance
in the experimental phase.3
Operationalization of Variables
This research employs the following three decision situation
independent variables, each measured using a discrete classification:
3Greater detail concerning the experimental environment, task, and
procedures is presented in Chapter V.


36
1) the information variable, 2) the distribution variable, and 3) the
cost variable. As indicated previously, each is varied across two levels.
The individual process variable types, measured on a continuous scale,
include the following: 1) individual decision model sensitivity, 2)
individual decision criteria, and 3) relative initial and final decision
anchor relationships. The major dependent variable is the individual
long-run decision efficiency (in terms of incurred costs). The
following discussion describes each of the variables or variable types
as they are employed in this study.
The information variable
The first level, labeled Jl_, is derived from the set of infor
mation assumed to come from individual experience with the physical
system. It includes the following items: 1) the historically derived
portion of time in which the process has been found to fall in each of
the two states, 2) the assumption that the random variable of interest
(actual minutes incurred per chair or its associated standard vari
ance) is normally distributed for both states, 3) the lowest observed
value of the random variable, 4) the highest observed value of the
random variable, 5) the maximum costs associated with each state, and
6) the minimum costs associated with each state.
The second level of the information variable, labeled 12^, includes
additional distributional information. In addition to the six items
contained in the II information set, the following two items are
included: 1) the mean of the random variable within each state, and
2) the standard deviation of the random variable within each state.


37
The distribution variable
Manipulation of the distribution variable involves two factors:
1)the distributional parameters of each state, and 2) the statistical
relationship between the states. The two levels of this variable are
generated through a change in the variance and a change in the stan
dardized distance between the means of the two states. The first
level of the distribution variable, labeled Sl_, has the following
parameters and relationships:
1) y1 = 36.0 actual minutes incurred per chair;
2) crn = a = 3) u12 = y + 1.5a = 40.5 actual minutes incurred per chair,
where, y = the mean of the jth state of nature (in-control = 1 and
* J
out-of-control = 2) given the ith distribution level
(SI = 1 and S2 = 2);
a.. = the standard deviation of the jth state of nature given
' J
the ith distribution level; and
= the standard deviation common to both states of nature
given the ith distribution level.
The second level of the distribution variable, labeled S2^, has
the following parameters and relationships:
1) y = 36.0 actual minutes incurred per chair;
2) 021 = 022 = 09 = 5.0 actual minutes incurred per chair; and
3) y2 = y21 + 1.80 = 45.0 actual minutes incurred per chair,
where y^., 0^, and 0^ are defined the same as in the SI level. The
third and fourth statistical moments of each state within each dis
tribution level are uncontrolled except that their deviations from a
normal distribution are minimized.


38
The cost variable
Given the two states of nature (the realization of which is un
known to the individual) and two possible decisions, there follows that
two types of errors can be made in reaching a decision. The first type
of decision error, labeled type A, is the decision to investigate the
physical process when the in-control state exists. The second type of
decision error, labeled type B, is the decision not to investigate when
the out-of-control state exists. The marginal cost of either error
type equals the cost that would have been incurred had the correct
decision been made minus the cost that was incurred by the incorrect
decision. Thus, the marginal costs of the two decision error types,
labeled MC^ and MCg, are:
= C(decision=not investigate | state=in-control) -
C(decision=investigate | state=in-control), and
MCg = C(decision=investigate | state=out-of-control) -
C(decision=not investigate | state=out-of-control),
where C( ) represents the decision cost for the situation within the
parentheses.
When the marginal cost of a type A error equals the marginal
cost of a type B error the cost structure should be ignored when making
an investigation decision. Of greater interest are situations in which
the marginal error costs not equal. In the present study, each level
of the cost variable takes one of the following forms: 1) the
marginal cost of a type B decision error equals three times the marginal
cost of a type A error (labeled level CJ_), and 2) the marginal cost
of a type A error equals three times the marginal cost of a type B
error (labeled level C2).


39
Individual decision model sensitivity variables
The sensitivity of the individual's decision model relative to the
decision situation is measured using the TSD parameter d. The theo
retical definition of d is:
d = (p2 Pj)/ = zi z2
where y. = the mean of the jth state of nature (in-control = 1 and
out-of-control = 2);
a = the common standard deviation of the states of nature; and
z. = the value of the normal distribution function associated
J
with the jth state of nature and any decision cutoff value
common to both states.
An empirical estimate of the d1 for an individual, labeled d 1,
is obtained using the individual's conditional probabilities P(decision=
investigate | state=out-of-control) and P(decision=investigate | state=
in-control) to calculate a subjective z2 and z2> As previously pointed
out, d' is relative to the decision situation. In particular, d is
relative to the distribution variable. To gain comparability across
situations the following measure is used:
DN, = di / d
where d^ is generated using optimal model k. The measure decreases
from a value of one as the result of several factors: 1) the individual
is inconsistent in his use of his cutoff value (i.e., there exists a
range around his cutoff value within which decisions are not made
using a strict relation to this cutoff value) or the individual makes
one or more temporary processing errors, and 2) the individual
utilizes more than one cutoff value.


40
Based upon a subjective analysis of each individual's decisions,
the measure DN^ can be adjusted for the effects of using multiple cut
off values. Defining d^a to be the individual's decision model sensi
tivity with the effects of multiple cutoff values eliminated:
DNAi = d\a / d'k
The difference DNA^- DN^ approaches zero as the effects of multiple
cutoff values decreases and becomes zero when the individual uses a
single cutoff value.
Individual decision criteria variables
The criteria the individual adopts in making his decisions are
measured using the TSD parameter 3. The theoretical definition of 3 is:
3 = d>(z2) / (z1)
where ( ) denotes the normal density function for the point in the
parentheses and 2. is defined the same as for the d'. variable.
J
An empirical estimate for the 3 of an individual, labeled 3 ,
is obtained using the z, and z2 values associated with the individual's
conditional probabilities employed in estimating dl. The measure 3..
is relative to the decision situation. In particular, it is relative
to those variables which affect the point on the decision axis which
the individual selects as his investigation decision cutoff value. A
measure of individual criteria comparable across situations is defined
as:
BN. = 3. / 3.
i i k
where 3^ is generated using optimal model k. The measure approaches a
value of one as 3.. approaches 3^ (i.e., as the individual's cutoff
decision value approaches the optimal model's cutoff decision value).


The measure approaches either a value of zero or positive infinity as
the result of several factors: 1) the individual does not process
properly the effects of the relative costs of the two types of decision
errors, 2) the individual does not process properly the effects of the
relative frequencies of the two states, and 3) the individual uses more
than one cutoff value.
Using the same subjective analysis as that used in adjusting the
DN measure, the BN^ measure can be adjusted for the effects of multipl
cutoff values. Defining B^ to be the individual's criteria measure
with the effects of multiple cutoff values eliminated:
BNA, B* / Bk
The difference BNA^- BN^ approaches a value of zero as the effects of
multiple cutoff values decrease and becomes zero when the individual
uses a single cutoff value.
Another measure derived from the BN- measure can be considered
a measure of individual conservatism. In this study, conservatism
refers to incomplete adjustment from an initial decision anchor towards
the optimal cutoff value. Non-conservatism refers to a more that
complete adjustment from the initial decision anchor past the optimal
cutoff value. The extent of conservatism is measured by the relative
distance between the individual's cutoff value and the optimal model's
cutoff value. Since the measure is dependent on the direction of ad
justment it is conditional upon the level of the cost variable. Using
BNC.¡ to denote tne extent of an individual's conservatism:
BNC.¡ = (3? B|<) / B^ given the Cl cost level; and
BNC.¡ = (b^ Bj) / B^ given the C2 cost level.


42
BNC.J approaches either positive infinity or a value of one as the extent
of conservatism increases, approaches a value of zero as the extent of
conservatism decreases, and approaches either a value of negative one
or negative infinity as the extent of non-conservatism increases.
Initial and final decision anchor variables
The relative relationships between an individual's initial and
final decision anchors can be measured in terms of the relative linear
distance between these two decision anchors. A measure of the relative
adjustment for individual i, labeled RA^, is conditional on the direc
tion of adjustment along the relevant decision axis. This direction
of adjustment is in turn conditional on the level of the cost variable.
Given the Cl cost level the initial decision anchor is greater than the
optimal model cutoff value, and given the C2 cost level the initial
decision anchor is less than the optimal model cutoff value. The RA^
measure is defined as:
RAi = (EDVi 0DVk) / (TDVi 0DVk) ,given the Cl cost level;
RA.j = (0DVk EDV.j) / (0DVk TDV.¡) ,given the C2 cost level,
where EDV^ = individual i's final (experiment) decision anchor;
TDV. = individual i's initial (training) decision anchor; and
0DVk = optimal model k's final (experiment) cutoff value.
The RA.j measure has the following relationships: 1) if no
adjustment occurs between the training and the experiment phases the
measure will equal a value of one, 2) if the adjustment is in the wrong
direction (away from the optimal model's experiment cutoff value) the
measure will be greater than a value of one, 3) if the adjustment is in


43
the proper direction but incomplete the measure will be less than a
value of one but greater than a value of zero, 4) if the adjustment
is in the proper direction and is complete the measure will equal a
value of zero, and 5) if the adjustment is in the proper direction but
more than complete the measure will be less than a value of zero.
Individual long-run decision efficiency
A major dependent variable of interest in this research is the
cost incurred as a result of the individual's variance investigation
decisions. The experimental objective function for all decision
situations is to minimize these costs (labeled investigation decisions
costs). The individual's long-run decision efficiency is measured in
terms of his minimization of the investigation decisions costs summed
over all decisions.
Given a pair of decision situations which hold constant all of
i
the independent variables except for the cost variable,the correct
decision within each situation will lead to different investigation
decisions costs. Consequently, absolute investigation decisions costs
are not comparable between decision situations: a relative measure
must be obtained. One such relative measure involves the comparison of
the individual's investigation decision costs with that of an optimal
model which uses the same information as that available to the indi
vidual. Denoting such a measure G-:
Gi: = / SMCk
where IC- = individual i's investigation decision cost for decision j;
MC^j = optimal model k's investigation decision cost for decision
j; and


44
SMC^ = the sum of optimal model k's investigation decisions costs
over all decisions (m in number).
The G measure can be classified into one of three submeasures
' J
where the classification is dependent upon the algebraic sign of G^j.
The classifications are 1) G.¡ > 0-- the individual's investigation
decision cost for decision j is greater than that of the optimal model,
2) G^j = 0-- the individual's investigation decision cost for decision
j is equal to that of the optimal model, and 3) G < 0-- the indi-
* \J
vidual's investigation decision cost for decision j is less than that
of the optimal model.
Summing those G.-s with identical algebraic signs, the summations
* 0
are defined as follows:
GP.j = the sum of those Gs with a positive algebraic sign;
GN^ = the sum of those G-s with a negative algebraic sign; and
GZ.j = the sum of those Gs with a value of zero.
It is easily shown that the following relation holds:
m
E G.. = GP, + GN, + GZ,
j=l J
where m equals the total number of decisions made by individual i. Since
the GZ.j summation equals zero, the sum of the G over all j = l,m
decisions, denoted G-, is:
G- = GP- + GN.
i i i
Conceptually, GP^ is the relative additional cost of those de
cisions made by individual i which are greater in value than that of
the optimal model. The GN^ is the relative savings of those decisions
made by individual i which are less in value than that of the optimal
model.


45
The anchoring and adjustment process may occur over three training
sessions and one experiment session. Using the first training session
as an estimate of the initial decision anchor and the experiment
session as an estimate of the final decision anchor, there remains two
training sessions within which the adjustment process itself can be
analyzed. In particular, the effects of the pattern of adjustments
within the training sessions are expected to influence significantly
the magnitude of an individual's experiment Gj measure. The adjustment
process can be measured in the training sessions by computing the
individual's Gj measure at the completion of each session. The
measures, labeled TG^, TG.¡0, and TG.¡3, are defined in the same manner
as the G. measure discussed above (with the exception that they are
computed from the appropriate training session data instead of the
experiment session data). The differences between these training
measures reflect the effects of the individual's adjustments in the
training sessions. These differences are defined as follows:
DTGii = TGn TGi2
DTGi2 = TGi2 TGi3
where, DTG-¡j = the difference between the G-¡ measure for the jth and
the jth + 1 training sessions of individual i; and
TG^j = the G.j measure for the jth training session of indi
vidual i.
Negative DTG-¡jvalues indicate an increase in the TGjj measure between
training sessions, and positive DTG-jj values indicate the reverse.
The effects of the directions of these adjustments should be
related to the effects of the final anchor on the individual's G-¡
measure (in the experiment). The effects of the adjustment measures


46
on tha experiment measure can be analyzed using a discrete classifica
tion of the direction of the adjustments. This classification of the
measure is accomplished using its algebraic sign. The second and
third training session results can be grouped into one of four
classes: PP(+,+), PM(+,-), MP(-,+), and MM(-,-). Subjects within
the MM classification have constantly increasing TG-jj measures and
subjects within the PP classification have constantly decreasing TG^j
measures.
Hypotheses Formation
As previously discussed, the formation of the experimental
hypotheses is accomplished primarily using the method of simulation.
Two general types of simulations are performed simulation of optimal
model performances and simulation of subjective investigation decision
performances. The first part of this section discusses the assumptions
and presents the results of the simulation of optimal model perform
ances. The second part of this section discusses certain assumptions
derived from the conceptual development that form the basis for the
simulation of subjective investigation decision performances. It also
presents the results of this simulation and the derivation of the ex
perimental hypotheses.
Simulation of Optimal Model Performances
The decision rule of the optimal model (given the objective func
tion of minimizing incurred costs) employed in this research is to
investigate the reported labor efficiency variance when:


47
Equation 1
where x = the actual minutes incurred per chair;
out = the state of out-of-control;
in
= the state of in-control;
Vin9N = the cost associated with not investigating an in-control
variance;
V-¡nsy = the cost associated with investigating an in-control
variance;
V0Ut,Y = the cost associated with investigating an out-of-control
variance; and
Vout,N = the cost associated with not investigating and
out-of-control variance.
The manipulation of the cost variable produces a constant cost ratio
at each of the two levels of this variable. When the cost variable
level is Cl the constant cost ratio is equal to one-third, and when the
cost variable level is C2 the constant cost ratio is equal to three.
The optimal model investigation decision rule expressed in equation 1
can be restate as:
LRq > PR.j CR Equation 2
where LRQ = the likelihood odds of the reported labor efficiency
variance being out-of-control (P(x|out)/P(x|in));
PR.j = the prior odds of the in-control state (P(in)/P(out)); and
CR = the constant cost ratio associated with the appropriate
level of the cost variable.
Assuming that both states of nature are normally distributed with equal
variance the .optimal model investigation decision rule expressed in


48
equation 2 can be restated in the following mathematical terms:
exp -(x y2)2/2cr2
: > PR-i CR
exp -(x \il)2/2a2
where x = actual minutes incurred per chair;
Vj = the mean of the in-control state;
y2 = the mean of the out-of-contrcl state;
a = the common standard deviation of both states; and
exp = the notation for the exponential funcition (e).
If the above equation was changed to an equality and solved for x the
result would be the optimal cutoff value in terms of actual minutes
incurred. This solution takes the general form:
G2(ln(PR-¡) + ln(CR))
x = + 0.50(y2 + y,) Equation 3
y2 yi
where x = the optimal cutoff value in terms of actual minutes incurred;
and
In = the notation for the natural logarithm function (loge).
With the decision rule in the form of equation 3 the parameters required
to fit the optimal model become identifiable. These parameters include
the mean of both states of nature, the common variance of the two
states, the prior odds of the in-control state, and the appropriate
value of the cost ratio.
With knowledge of the parameters required to fit the optimal
model the criteria employed for the simulation of optimal model per
formances can be discussed. The first criterion is that information
available to the optimal model be the same as the information avail
able to the individual. The information available to the individual
is manipulated by the levels of the information variable. Within the 12


49
information level the available information set contains statistical
estimates of all the parameters required to fit the optimal model.
However, within the II information level the available information
set does not contain all the required parameter estimates. Therefore,
within the II information level the optimal model must use the training
sessions to estimate the missing parameters. The parameter estimates,
both given and estimated, are presented in Table 2.
The second criterion is that the labor efficiency variance re
ports presented to the subjects in the experiment be the same reports
used in the simulation of optimal model performances. The appropriate
parameter estimates presented in Table 2 were used in the optimal
model decision rule expressed by equation 2 to simulate optimal model
investigation decisions. Table 3 presents some performance results
of this optimal model simulation.
Simulation of Subjective Investigation Decision Performances
This section discusses assumptions derived from the conceptual
development that form the basis for the simulation of subjective investi
gation decision performances. It also presents the results of this
simulation and the derivation of the experimental hypotheses.
Assumptions of simulation
The first assumption employed for the simulation of subjective
investigation decision performances is that the subjects will behave
as if they use the anchoring and adjustment heuristic during the
training phase. The anchoring and adjustment heuristic proposes that


TABLE 2
ESTIMATES OF THE PARAMETERS REQUIRED TO FIT THE OPTIMAL MODEL GIVEN THE VARIOUS CONDITIONS
Information
Experimental
Variable 11
Parameters
Information
Variable 12
Actual Underlying
Data Structure Parameters
Cost
Parameter
Distribution Variable
Distribution Variable
Distribution Variable
Variable
Estimate
SI
S2
SI
S2
SI
S2
35.99
36.26
36.00
36.00
35.975
35.943
y2
40.21
44.24
40.50
45.00
40.492
44.995
Cl
al
8.88
29.42
9.00
25.00
9.157
25.060
al
7.38
19.32
9.00
25.00
9.437
25.002
PRi
1.50
1.50
1.50
1.50
1.500
1.500
CR
0.33
0.33
0.33
0.33
0.333
0.333
C2a
CR
3.00
3.00
3.00
3.00
3.000
3.000
aThe parameter estimates
given the Cl cost level.
for u1 ,y2,cj2
,and PR.¡
given the C2
cost level are
the same as those


TABLE 3
RESULTS OF THE SIMULATION OF OPTIMAL MODEL PERFORMANCES GIVEN THE VARIOUS CONDITIONS
k Information
Distribution Cost
P(investigate
out-of-control)
P( investigate
in-control)
Optimal Cutoff
Value (actual
minutes incurred)
dk
6k
1
11
SI
Cl
0.875
0.383
36.90
1.447
0.539
2
11
SI
C2
0.400
0.033
41.29
1.581
5.206
3
11
S2
Cl
0.875
0.300
38.70
1.674
0.592
4
11
S2
C2
0.425
0.017
45.88
1.939
9.453
5
12
SI
Cl
0.900
0.400
36.86
1.535
0.454
6
12
SI
C2
0.400
0.033
41.26
1.581
5.206
7
12
S2
Cl
0.900
0.300
38.57
1.806
0.504
8
12
S2
C2
0.525
0.033
44.68
1.897
5.364
cn


52
a subject will select an initial anchor as a first decision approxi
mation and will adjust this decision anchor toward an optimally correct
point as he learns from his experiences with the states of nature.
However, the individual's adjustment process will not be sufficient:
i.e., he will tend to approach the optimally correct point but will not
adjust completely to that point.
The second assumption is that the location of the initial de
cision anchor on the actual minutes incurred continuum (axis) will be
conditional upon the level of the distribution variable. The reason
for this is that the initial decision anchor is expected to be located
at a central point between the means of the two states. Since the
mean of the out-of-control distribution shifts with the level of the
distribution variable, the central point between the means of the two
states also shifts with the level of the distribution variable. The
initial decision anchor used in the simulation of subjective investi
gation decision performances is the geometric intersection point of the
two states' distribution curves. This particular initial decision
anchor is used because it is the exact central point between the means
of the two states. However, it is selected for operational purposes
and is not necessary to the conceptual development. Any point of
central tendency would be satisfactory.
The final assumption is that the subjects' adjustment process
will be approximately equal over the various treatment conditions. The
adjustment process is defined as a linear movement along the actual
minutes incurred axis from the initial decision anchor towards the
appropriate optimal model decision cutoff. The magnitude of the linear
movement used in the simulation of subjective investigation decision


53
performances is 50 percent of the distance between the initial decision
anchor and the appropriate optimal decision value. The 50 percent
adjustment value, selected for operational purposes, is completely
arbitrary and is not necessary to the conceptual development. The only
restriction is that it be less than 100 percent. Furthermore, the
general assumption of equal adjustment over the various conditions is
not necessary to the conceptual development: the objective of this
general assumption is to facilitate a simple operationalization of the
simulation.
Training phase simulation and hypotheses formation
The first stage of the simulation involves simulating the
anchoring and adjustment process in the training sessions for each
combination of independent variables. A simulated subject sample size
of one is used for each condition (resulting in a total simulated sub
ject sample of eight). The results of this first stage simulation are
presented in graphical form in Figure 2. This figure depicts the
simulated initial decision anchors (points A), the simulated final
decision anchors (points B), and the simulated optimal cutoff values
(points C) for each of the eight combinations of independent variables.
The assumptions used in simulating the training phase can be
investigated using the results of the experimental training phase. The
assumptions are i) that the subjects will behave as if they used the
anchoring and adjustment heuristic, 2) that the initial decision anchor
will be located centrally between the means of the two states and will
be dependent on the level of the distribution variable, and 3) that the


II Information Level
Cost
Cl
Cost
C2
Distribution
SI
1 1 i
36.90 37.575 38.25
+
39.77 41.29
Distribution
S2
C B A
h
38.7 39.6 40.50
43.19
H
45.88
Distribution
SI
12 Information Level
Cost
Cl
B A
36.86 37.555 38.25
Cost
C2
39.755 41.26
C B A
Distribution 1 } 1
S2 38.57 39.535 40.50
+
42.59
H
44.68
Note: All scales are actual minutes incurred per chair produced.
FIGURE 2
SIMULATED INITIAL DECISION ANCHORS (POINTS A), FINAL DECISION
ANCHORS (POINTS B), AND OPTIMAL CUTOFF VALUES (POINTS C)


55
adjustment process will be approximately equal over the various con
ditions. The location of the initial decision anchor can be investigated
by examining the relationship between the individual's initial cutoff
value (in the first training session) and the intersection point of
the appropriate two distribution curves (the TDV.¡ measure). These
intersection points are conditional upon the level of the distribution
variable. Given the SI distribution level the intersection point is
38.25 actual minutes incurred, and given the S2 distribution level the
intersection point is 40.5 actual minutes incurred (see Figure 2 for a
graphical presentation of these points). The expected relationship is
that the mean initial decision anchor (TDV.) of the individuals, given
a level of the distribution variable, will not be significantly differ
ent from the intersection point of the appropriate two distribution
curves.
The assumption of approximately equal adjustment over the various
conditions can be investigated by examining the relationship between
the individual's adjustment over the complete experiment and the total
adjustment that would be required to reach the optimal model's decision
cutoff value. A measure of the relative adjustment for individual i,
RA.j, is conditional upon the direction of adjustment along the relevant
decision axis (i.e., the actual minutes incurred per chair produced).
The expected relationship is that the mean RA^ will not differ signifi
cantly between the levels of the various conditions.
Given the above simulation and discussion the following hypo
theses can be derived:
HI.la (TDVjISl) = 38.25.
The mean initial decision anchor given the SI distribution level


56
will equal the intersection point of the two distribution curves
within that level.
HI.lb (TDVi |S2) = 40.5.
The mean initial decision anchor given the S2 distribution level
will equal the intersection point of the two distribution curves
within that level.
HI.2a (TDV.]I1) = (157^12).
The mean initial decision anchor given the II information level
will equal the mean initial decision anchor given the 12 infor
mation level.
HI.2b (TDViiCl) = (TVi¡C2).
The mean initial decision anchor given the Cl cost level will
equal the mean initial decision anchor given the C2 cost level.
HI.3a (RAi111) = (RAi112).
The mean relative adjustment given the II information level will
equal the mean relative adjustment given the 12 information level.
HI.3b (RA^Sl) = (RAi ¡S2).
The mean relative adjustment given the SI distribution level will
equal the mean relative adjustment given the S2 distribution
level.
HI.3c (RAi|Cl) = (RA-|C2).
The mean relative adjustment given the Cl cost level will equal
the mean relative adjustment given the C2 cost level.
Individual decision model sensitivity hypotheses formation
The measures of individual decision model sensitivity should
generally be unaffected by the independent variables. These measures


57
have a greater relationship with individual decision processes than
with the specific conditions of a general decision situation. The
individual decision model sensitivity measure di should have a re
lationship with the distribution variable only. The theoretical d1
measure under the SI distribution level is 1.5 and under the S2 dis
tribution level is 1.8. Thus, the average dl under the S2 distribu
tion level should be greater than under the SI distribution level.
The DNA^ measure should have no systematic relationships with
the three independent variables. This measure of the relative de
viation of the individuals decision model sensitivity from that of
the optimal model's generally is due to individual decision incon
sistencies and temporary processing errors. If individuals are
randomized into the specific decision situations there are no a
priori reasons for expecting significant differences in this measure
due to the independent variables.
The difference between DNA^ and DN-¡ measures the effects of
multiple decision anchors upon the individual's relative decision
model sensitivity. Relationships between this difference and the in
dependent variables depend upon whether the independent variables are
related to the causes underlying an individual's use of more than one
cutoff value. A possible explanation involves the distribution vari
able. As the variance of the in-control state increases the absolute
distance between the lowest values of the random variable of interest
(actual minutes incurred) and the mean of the in-control state in
creases. In turn, as this distance increases subjects may be more
inclined to perceive that a second out-of-control distribution overlaps
the lower range of the in-control state. If such perceptions underly


58
the use of multiple cutoff values, the average difference between DNA.¡
and DN.j should differ between the two levels of the distribution vari
able. A second possible explanation involves the cost variable. As
the individual's final cutoff value increases, the absolute distance
between the individual's final cutoff value and the lowest values of
the random variable of interest increases. In turn, as this distance
increases subjects again may be more inclined to perceive that a sec
ond out-of-control distribution overlaps the lower range of the
in-control state. If such perceptions underly the use of multiple
cutoff values, the average difference between DNA- and DN.¡ should
differ between the two levels of the cost variable.
Based on the above discussion the effects of the independent
variables on the individual decision model sensitivity measures can
be hypothesized as follows:
H2.la (dj |SI) < (dj |S2).
Those subjects within the SI distribution level will have
significantly smaller mean djs then those subjects within the
S2 level.
H2.lb (di ¡II) = (d! ¡12).
Those subjects within the II information level and within the
12 information level will have equal mean d^s.
H2.1c (d* |Cl) = (d! |C2).
Those subjects within the Cl cost level and within the C2 cost
level will have equal mean d^s.
H2.2a (DNAilll) = (DA-j ¡ 12).
Those subjects within the II information level and within the 12
information level will have equal mean DNA-¡s.


59
H2.2b (DNA-j | SI) = (DNA-j |S2).
Those subjects within the SI distribution level and within the
S2 distribution level will have equal mean DNA-jS.
H2.2c (DA-,-1 Cl) = (DA-j IC2).
Those subjects within the Cl cost level and within the C2 cost
level will have equal mean DNA-¡s.
H2.3a (DNA-j-DN-j ¡SI) < (DNA-¡-DN-j |S2).
Those subjects within the SI distribution level will have
significantly smaller mean DNA-¡-DN-¡s than will those subjects
within the S2 distribution level.
H2.3b (DNAj-DNj|Cl) <(DNAi-DNi|C2).
Those subjects within the Cl cost level will have significantly
smaller mean DNA-¡-DN-¡s than will those subjects within the C2
cost level.
Individual decision criteria simulation and hypotheses formation
The variables which affect the BNj measure should be those factors
related to the individual's selection of a cutoff value. The most
significant variable affecting the individual's cutoff value should be
the cost variable. Within this variable the initial decision anchor
given the Cl level is much closer to the optimal cutoff value than is
the initial decision anchor given the C2 level. Consequently, the BN-j
measure should be closer to a value of one under the Cl level than
under the C2 level.
The difference between the BNA-¡ and the BNj measures should
relate to the same variables as those discussed for the difference


60
between the DNA.¡ and DN^ measures. The two possible explanations in
clude the distribution variable and the cost variable.
Theoretically, the more extreme the optimal cutoff value relative
to a central point between the distributions of the two states, the
higher should be the individual conservatism. The decision situations
with the most extreme optimal cutoff values are those within the C2
cost level; the situations with the least extreme optimal cutoff values
are those within the Cl cost level. Consequently, the BNCj measure
should be larger within the C2 cost level than within the Cl cost level.
The simulation and assumptions used in developing the training
phase hypotheses can be extended to enable formation of hypotheses con
cerning the individual decision criteria. The earlier simulation de
rived the following final decision cutoff values: 1) 37.555 actual
minutes incurred for the Cl cost level given the 12 information level
and the SI distribution level, and 2) 39.755 actual minutes incurred
for the C2 cost level given the 12 information level and the SI dis
tribution level. Given these cutoff values a f(e^) measure can be
calculated directly (the 3^ measure is used rather than the en- measure
due to the assumption of a single cutoff value). The functional
notation, f( ), is employed to indicate these estimates are based upon
a simulation rather than upon empirical observation. For this simu
lation the f(3?)s are as follows:
f(3*|Cl,Sl,I2) = 4>(zout) / f(B*|C-2,Sl,I2) = (zQut) / *(z.n) = 2.11774
where ( ) indicates the value of the normal density function at the
point in parentheses. Given the f(B^) measures the f(BNA^) measures
can be calculated. For this simulation the f(BNA^)s are as follows:


61
f(BNAi|Cl,SI,12) = f(B) / f(6k) = 1.41898
f(BNAi|C2,Sl,I2) = f(e?) / f(fJk) = 0.47017
where f($k) is the simulated measure at the cutoff value of the kth
optimal model.
This simulation can be extended to the S2 level of the distribu
tion variable given the 12 information level and can be extended to
both the SI and S2 levels of the distribution variable given the II
information level. The f(BNA-) measures obtained under the 12 infor
mation and S2 distribution conditions are as follows:
f(BNAiJCl,S2,12) = 1.42318
f(BNAi¡C2,S2,I2) = 0.47300
The f(BNA-) measures given the II information level are as follows:
f(BNAi|Cl,Sl,Il) = 1.40126
f(BNA -1C2,SI,11) = 0.46993
f(BNAi|Cl,S2,Il) = 1.38273
f(BNA.|C2,S2,Il) = 0.38094
Based on these f(BNA-) measures, hypotheses can be derived con
cerning the effects of the independent variables upon the BNA. measure.
The f(BNA^) measures can be averaged over the appropriate conditionals
to determine these effects. Averaging over all conditionals except the
information variable gives the following results:
f(BAiIII) = 0.90871
f(BNAi|12) = 0.94633
Although there is a difference in the f(BNA-) measures due to the infor
mation variable this difference is small compared to the standard error
of these estimates (the difference divided by the standard error is
0.03762/0.39157 = 0.0961). Therefore, a significant effect due to the
information variable would not be expected.


62
Averaging over all conditionals except the distribution variable
gives the following results:
f(BA-¡ 1 SI) = 0.94009
f(BNAi|S2) = 0.91496
Again, the difference in the f(BNAj) measures due to the distribution
variable is small compared to the standard error of the estimates (the
difference divided by the standard error is 0.02513/0.39174 = 0.C642).
A significant effect due to the distribution variable would not be
expected.
Averaging over all conditionals except the cost variable gives
the following results:
f (BA-j |Cl) = 1.40654
f(BA-¡ |C2) = 0.44851
The difference in the f(BNA-j) measures due to the cost variable is
large compared to the standard error of the estimates (the difference
divided by the standard error is 0.95803/0.02436 = 39.3308). Con
sequently, a significant effect would be expected for the cost
variable.
Another implication drawn from the f(BNAj) measures concerns the
homogeneity of variance for the measures given different levels of the
cost variable. The variance for the f(BNA-¡ | Cl) is .00034 and the vari
ance for the f(BNA-¡ |C2) is .00203. Using the F test for equal variances,
F=5.9344 (3 and 3 d.f.) which indicates the variances may not be equal.
The variance of (BNA-¡[C1) would be expected to be less than the vari
ance of (BNAj¡C2).
This simulation can be extended to enable the formation of
hypotheses concerning the effects of the independent variables on the


63
conservatism measure BNC-¡. This extension employs the f(B-j) and fU^)
measures used above. The results of this extended simulation averaged
over all conditionals except for the information variable are as follows:
f(BNCi | II) = 0.48328
f(BC.¡|I2) = 0.47475
The difference in the f(BNC^) measures due to the information variable
is small compared to the standard error of the estimates (the differ
ence divided by the standard error is 0.00853/0.06390 = 0.1335). A
significant effect due to the information variable would not be expected.
Averaging over all conditionals except for the distribution
variable gives the following results:
f(BCi|SI) = 0.47004
f (BNCi|S2) = 0.48799
The difference in the f (BNC.¡) measures due to the distribution vari
able is small compared to the standard error of the estimates (the
difference divided by the standard error is 0.01795/0.06357 = 0.2824).
A significant effect due to the distribution variable would not be
expected.
Finally, averaging over all conditionals except for the cost
variable gives the following results:
f(BC_¡ i Cl) = 0.40654
f(BCi|C2) = 0.55149
The difference in the f(BC.¡) measure due to the cost variable is large
compared to the standard error of the estimates (the difference divided
by the standard error is 0.14495/0.02436 = 5.9503). Consequently, a
significant effect would be expected for the cost variable.


64
The above discussion concerning the individual decision criteria
measures can be summarized as follows:
H3.1a (B.¡lll) = (BA-j 112).
The mean BNAj of those subjects within the II information level
and within the 12 information level will be equal.
H3.lb (BilSl) = (BA-¡ | S2).
The mean BNA-¡ of those subjects within the SI distribution level
and within the S2 distribution level will be equal.
H3.1c (BA-j | Cl) > (BA-¡ | C2).
Those subjects within the Cl cost level will have a significantly
larger mean BNA-j than will those subjects within the C2 level.
H3.2 a2(BNA-j [Cl) < a2(BNA-¡ | C2).
The BNA-¡ of those subjects within the Cl cost level will have a
significantly smaller variance than will the BNA-¡ of those sub
jects within the C2 cost level.
H3.3a (BCj ¡II) = (BCj 112).
The mean BNC-¡ of those subjects within the II information level
and within the 12 information level will be equal.
H3.3b (BCj|SI) = (BCj|S2).
The mean BNCj of those subjects within the SI distribution level
and within the S2 distribution level will be equal.
H3.3c (BCj|Cl) < (BCj j C2).
The mean BNCj of those subjects within the Cl cost level will be
significantly smaller than the mean BNCj of those subjects within
the C2 level.


65
H3.4a (BNAi-BNi|Sl) < (BNA.j-BN.¡S2).
The mean BNA.-BN^ of those subjects within the SI distribution
level will be significantly smaller than the mean BNA^-BN^ of
those subjects within the S2 distribution level.
H3.4b (BNA.-BN.|C1) <(BNAi-BN^|C2).
The mean BNA^-BN^ of those subjects within the Cl cost level
will be significantly smaller than the mean BNA.-BN. of those
subjects within the C2 cost level.
Individual long-run decision efficiency simulation and hypotheses
The simulation and assumptions used in developing the training
phase hypotheses enable the formation of hypotheses concerning the
G.j variable. The graph presented in Figure 3 depicts the relations of
the values derived within the training phase simulation given the
condition (12,SI). The interval between the final cutoff value and
the appropriate optimal model cutoff value represents the areas
associated with the 6. measure. For the Cl cost level the area of the
interval under the in-control distribution is the area associated with
the GN. measure; the area of the interval under the out-of-control
distribution is the area associated with the GP. measure. These areas
are the left-most shaded areas in Figure 3. For the C2 cost level the
area of the interval under the out-of-control distribution is the area
associated with the GN. measure; the area of the interval under the
in-control distribution is the area associated with the GP- measure.
These areas are the right-most shaded areas in Figure 3. The areas
under these curves for this simulation are as follows:


66
o
o
LD
o
LD
o
o
o
LO
LO
LD
LD
o
LO
o
CO
ID
C\J
LD
CM
'






LO
LD
r^
CO
cr>
o
H
CO
CO
CO
CO
CO
*3
FIGURE 3
GRAPH OF THE SIMULATION USED IN THE CONDITION
(12 INFORMATION, SI DISTRIBUTION)


67
Area(GN-j |Cl,SI,12) = 0.08483
Area(GP-¡ |Cl,SI, 12) = 0.05048
AreaiGNi|C2,S1,I2) = 0.19780
Area(GPi|C2,Sl,I2) = 0.06549
These areas must be adjusted for the relative frequency differ
ences and the relative cost differences between the various conditions.
If the in-control distribution is given a relative frequency weight of
one,then the relative frequency weight of the out-of-control distribu
tion would equal two-thirds (P(in-control) = 0.60 and P(out-of-control) =
0.40). The relative cost weights can be obtained directly from the
manipulation of the cost variable. Given the Cl cost level, errors
under the out-of-control distribution equal three times the errors under
the in-control distribution. Given the C2 cost level, errors under the
in-control distribution are equal to three times the errors under the
out-of-control distribution. The four areas associated with the
various G-¡ measures can be converted into relative relations by taking
into account both of these relative weights (frequency and cost). These
relative relations are as follows:
f(GN.¡ | Cl,S1,12) = (0.08483) (1) (1) = 0.08483
f(GP-j | Cl,SI, 12) = (0.05048) (2/3) (3) = 0.10096
f(GNi|C2,Sl,I2) = (0.19780)-(2/3)-(1) = 0.13187
f(GPi|C2,Sl,I2) = (0.06549) (1) (3) = 0.19647
The difference between the relative relations of GP.¡ and GNj
given a level of the cost variable indicates the relative effect of
these intervals upon the Gj measure. The differences for this simu
lation are as follows:


68
f(6i|Cl,Sl,I2) = f(GPi|C1,S1,12) f(GN^|C1,S1,I2)
= 0.10096 0.08483 = 0.01613
f(Gi |C2,S1,12) = f(GPi |C2,S1,12) f(GN .|C2,S1,12)
= 0.19647 0.13187 = 0.06460
This simulation can be extended to the S2 level of the distribu
tion variable given the 12 information level and can be extended to
both the SI and S2 distribution levels given the II information level.
The relative relations of the G- measure obtained under the 12 infor
mation and S2 distribution conditions are as follows:
f(Gi|Cl,S2,I2) = 0.07596 0.06385 = 0.01211
f(Gi|C2,S2,I2) = 0.15741 0.10639 = 0.05102
The relative relations for the G.¡ measure given the II information
level are as follows:
f(Gj|C1,S1,I1) = 0.09942 0.08230 = 0.01712
f(Gi|C2,S1,II) = 0.19629 0.13146 = 0.06483
f(G1|Cl,S2,Il) = 0.07246 0.05884 = 0.01362
f(Gi|C2,S2,Il) = 0.15342 0.14078 = 0.01264
Based upon these relative relations hypotheses can be derived
concerning the effects of the independent variables on the G^ measure.
These relations can be averaged over the appropriate conditionals to
determine the effects. Averaging over all conditionals except the
information variable gives the following results:
f(Gi|II) = 0.02705
f(CTj 112) = 0.03596
Although there is a difference in the f(G-) due to the information vari
able,this difference is small compared to the standard error of the
estimates (the difference divided by the standard error is


69
0.0089/0.01808 = 0.49226). Therefore, a significant effect due to the
information variable would not be expected.
Averaging over all conditionals except the distribution variable
gives the following results:
f((^ |SI) = 0.04152
f(G.|S2) = 0.02235
Although the difference between these measures is much larger than that
above, this difference is only slightly larger than the standard error
of the estimates (the difference divided by the standard error is
0.01832/0.01658 = 1.10495). The direction of the f(G^) differences
would be expected to be as predicted by the simulation but may not be
significant.
Averaging over all conditionals except the cost variable gives
the following results:
f(G.|Cl) = 0.01475
f(G.|C2) = 0.04827
The difference between these measures is large compared to the
standard error of the estimates (the difference divided by the standard
error is 0.03353/0.00951 = 3.5258). Consequently, a significant effect
would be expected for the cost variable.
These relative relations can be averaged over all but pairs of
conditionals. Such measures could indicate the presence of interaction
effects of the independent variables. The results of averaging the
pairwise conditionals indicate only one interaction of possible
significance-- the distribution by cost interaction. The f(G^) measures
conditional to these variables are as follows:
f (&.|SI,Cl) = 0.01663


70
f((x||Sl,C2) = 0.06471
f(Gi|S2,Cl) = 0.01287
f(Gi|S2,C2) = 0.03185
A graphic representation of this interaction is presented in Figure 4.
The difference between the two f(G.|Cl) measures is much smaller than
the difference between the two f(^|C2) measures. The differences are
0.00376 and 0.03286, respectively. This would indicate that the
distribution variable has a significant effect only when given the
C2 cost level. Such q result would explain the larger but not signifi
cant effect of the SI distribution level on the f(G^|S ) measures.
Other implications which can be derived from these relative
relations involve relationships between f(GP-) and f(GN^) measures.
Two such relationships appear to be of possible significance. Both
involve the ratio f(GP^)/f(GN^) (i.e., the effect of additional errors
relative to the effect of additional correct decisions). The results
of the first relationship are as follows:
f(GPi/GNi|C1) = 1.20480
f(GPi/GNi|C2) = 1.38337
The difference between these measures is large compared to the standard
error of the estimates (the difference divided by the standard error is
0.17857/0.07621 = 2.34313). Consequently, the GP./GN. ratio is ex
pected to be larger given the C2 cost level than given the Cl cost
level.
The results of the second relationship are as follows:
f(GPi/GNi|Sl,Cl) = 1.19908
f(GPi/GNi|Sl,C2) = 1.48206
f(GP./GiNi|S2,Cl) = 1.21057


71
f(G.)
Distribution
Variable
FIGURE 4
SIMULATED COST BY DISTRIBUTION VARIABLE INTERACTION ON f(G-¡)


72
f(GP j/GN-¡ |S2,C2) = 1.28465
The difference between the two f(GP-¡/GN-¡ |S1) is much smaller than the
difference between the two f (GP -¡/GN -j |S2). This would indicate the
presence of a distribution by cost interaction in which the effect of
the cost variable is significant only under the SI distribution level.
A graphical representation of this interaction is presented in Figure 5.
A final implication derived from these relative relations in
volves the homogeneity of variance for the f(G^) given the different
levels of the cost variable. The variance for the f(G^|Cl) is equal to
5.26 x 10-6 for this simulation, and the variance for the f(G^[C2) is
equal to 6.06 x 10"4 for this simulation. Using the F test for equal
variances, F=115.24 (3 and 3 d.f.) which indicates the variances are
not equal. Consequently, the variance of (G-¡ | Cl) is expected to be less
than the variance of (G-¡|C2).
Given the above simulation the effects of the independent vari
ables on the G-¡ variable can be hypothesized as follows:
H4.1 a2(Gj|Cl) < a2(Gi|C2).
The G-¡ of those subjects within the Cl cost level will have
significantly smaller variance than will the G-¡ of those subjects
within the C2 cost level.
H4.2a (GilCl) < (&||C2).
Those subjects within the Cl cost level will have a significantly
smaller mean Gj than will those subjects within the C2 level.
H4.2b (Gi|S2) < (Gi|Sl).
Those subjects within the S2 distribution level will have smaller
mean Gj than will those subjects within the SI level. Signifi
cance is not predicted due to the interaction effect with the
cost variable.


73
Distribution
Variable
FIGURE 5
SIMULATED COST BY DISTRIBUTION VARIABLE INTERACTION ON f(GPi/GNi)


74
H4.2c (G-j | II) = (Gi 112).
The mean G-j of those subjects within the II information level
and within the 12 information level will be equal.
H4.3 (Gi |SI,Cl) (Gj |S2,C1) < (Gi]Sl,C2) (G'i JS2,C2) .
There will be a significant interaction of the distribution and
cost variables in which the distribution variable will have no
significant effect given the Cl cost level but will have a
significant effect given the C2 cost level.
H4.4a (GP-j/GN-j |Cl) < (GP1/GNi |C2).
Those subjects within the Cl cost level will have a significantly
smaller mean GP-j/GNj than will those subjects within the C2 level.
H4.4 (GP-j/GN-j |C2,S2) (GP-j/GN-j | Cl,S2) <
(GP-j/GN-j ¡C2,S1) (GP-j/GN-j ] Cl,S1) .
There will be a significant interaction of the distribution and
cost variables in which the cost variable will have no signifi
cant effect given the S2 distribution level but will have a
significant effect given the SI level.
Individual long-run decision efficiency and the training phase
Since performance within the experiment is assumed to be a
function of the training adjustment process, it would be expected that
the subjects within the PP classification of training adjustment
pattern would have a lower average experiment G-¡ measure than those
subjects within the MM classification. Furthermore, the mean expected
Gj measure for the mixed classifications (PM and MP) would be expected
to fall between those of the non-mixed classifications. The mixed


75
classifications have one positive adjustment which would improve per
formance above that of subjects who make only negative adjustments. On
the other hand, the mixed classifications have one negative adjustment
which would decrease performance below that of subjects who make only
positive adjustments.
The following hypotheses predict the effects of the training
adjustment pattern on the Gi variable:
H4.5 (Gj|PP) < (G^|MM).
Those subjects within the PP classification of training adjust
ment pattern will have a significantly smaller mean G- measure
than those subjects within the MM classification.
H4.6 |PP) < (G.j |MP) = ((Tj |PM) < (GTj |MM).
Those subjects within the mixed classifications of training
adjustment pattern will have mean G^s that fall between those of
the non-mixed classifications. The mean G^s of the mixed classi
fications are not expected to be significantly different.


CHAPTER IV
THE EXPERIMENT
Experimental Environment
The decision situation studied within this research was standard
cost variance investigation decision making at the operational manage
ment level. The experimental setting was the simulated environment of
a manufacturing company, and the subjects were requested to assume the
role of an operational manager of an assembly department within this
environment. The assembly department assembled a single product, a
metal folding chair. Subjects were presented with a series of standard
cost variance reports for the department and were asked to decide for
each report whether the underlying physical process should be investi
gated to correct an out-of-control situation. Each series of variance
reports were cross-sectional in nature: i.e., all reports within a
series were assumed to have occurred simultaneously and were independent
of each other.
The physical process within the simulated environment was a
labor-paced process (Barefield, 1972): i.e., the operating efficiency
of the department was determined completely by the labor efficiency of
the workers. The labor efficiency standard (stated in terms of time
per unit assembled) was based on engineering estimates that allowed for
unavoidable labor inefficiencies and reasonable variation in worker
76


77
performance (i.e., the standards were currently attainable). The sub
jects were instructed to accept the labor efficiency standard as fair
in terms of control and performance goals.
The physical labor process of the department could be in one of
two mutually exclusive states of nature: either in-control or out-of
control. The overall labor process consisted of many individual
physical labor operations: the expected aggregate of these procedures
was represented by the overall labor efficiency standard. The variance
reports, however, included only the aggregate standard and labor
efficiency variance. The labor process was defined to be in-control
when all of the individual physical procedures were being performed as
expected. The labor process was defined to be out-of-control when one
or more of these procedures was not being performed as expected.
The task of the subject was to decide whether to investigate each
labor efficiency variance for its underlying causes. The purpose of
investigation was to facilitate correction of those individual pro
cedures which were not operating as expected. Tv/o assumptions were pro
vided to aid the subject's decision making process. First, if they
decided to investigate a variance and the labor process turned out to
be out-of-control, the process would be returned to the original
in-control state with certainty. Second, if they decided not to in
vestigate a variance and the labor process was out-of-control, the process
would remain out-of-control with certainty.
Various costs were associated with the variance investigation
decision. These costs depended upon the subject's decision (either in
vestigate or do not investigate) and upon the actual state of the labor
process (either in-control or out-of-control). Estimates of these


78
costs were presented to the subject on the face of each variance report
using the following format:
If Your
Investigation
Decision Is
And If The
Assembly Line
State Is
Then
Your
Costs
Are
Investigation
Production
Total
Yes
In-control
$
X
$
0.00
$ X
Yes
Out-of-control
$
Y
$
0.00
$ Y
No
In-control
$
0.00
$
0.00
$ 0.00
No
Out-of-control
$
0.00
$
Z
$ z
The letter X represented the estimated investigation cost when the
assembly line was in-control. The cost was related to the size of the
labor efficiency variance: the more negative (unfavorable) the variance
the larger the cost. The letter Y represented the estimated investi
gation cost when the assembly line was out-of-control. The cost was
related to the size of the labor efficiency variance: the more negative
(unfavorable) the variance the smaller the cost. Given any particular
labor efficiency variance, the investigation cost when out-of-control
(Y) was always larger than the investigation cost when in-control (X).
This was because the investigation cost included correction costs when
the assembly line was out-of-control. The letter Z represented the
estimated marginal production cost of operating the next period in the
out-of-control state. The cost was constant for all variance reports.
The subjects were told that their immediate supervisor, the pro
duct section manager, would evaluate their control performance in terms
of their minimization of both investigation and production costs above
the expected standard (labeled the total investigation decision cost).
A cash bonus was promised to the subjects, the size of the bonus being
contingent upon the extent to which they minimized their total investi
gation decisions cost. The measure of a subject's control performance,
labeled TIDCmi-n, was determined by summing the total investigation


79
decision costs incurred by the subject over the series of variance re
ports and dividing this sum by the sum of the total investigation de
cision costs incurred by an optimal model over the same series of vari
ance reports. The subject's cash bonus function, constant for all sub
jects, was inversely related to this measure. As TIDCmi-n approached
one, the payoff approached the maximum: as TIDCmi-n became larger than
one, the payoff approached the minimum.
Subjects
The subjects were 86 senior year undergraduate and master's level
graduate students enrolled in the business college at the University of
Florida. The subjects participated in the experiment during a two week
period; 47 participated in the first week and 39 participated in the
second week. A total of 92 subjects initially volunteered to participate
but six subjects failed to complete the experiment. The 86 subjects
who completed the experiment consisted of 63 males and 23 females.
Three subject selection criteria were applied: 1) the subject
must have completed an intermediate-level managerial accounting course,
2) the subject must have completed an introductory-level statistics
course, and 3) the subject must have earned an overall grade point
average (GPA) of at least 2.0 on a 4.0 scale.
Subjects initially were contacted within senior level and graduate
accounting classes. The contact was made by the experimenter giving a
brief oral presentation followed by passing sign-up sheets around each
class (a copy of the oral presentation is presented in Appendix B). To
motivate volunteering, the presentation focused on the student's pro
fessional responsibilities as future accountants and on the monetary


80
benefits that would accrue to those who volunteered. The subjects were
told that they would be given a cash payment for participating in the
experiment and that the amount of a subject's payment would depend upon
his performance. The minimum payment was set at $2.00 and the maximum
at $10.00.
Experimental Materials
The experimental materials included a background information book
let, variance investigation decision stimuli, heuristics questionnaire,
and motivations questionnaire.
Background Information
A background information booklet (see Appendix C) was designed
to provide the subjects with a common experimental environment. The
booklet provided the subject with general company information, general
product information, general manufacturing process information, and
specific assembly department information. The specific assembly depart
ment information included information concerning the employees, the
physical process, the accounting control system, the subject's task as
the operational manager, and the subject's performance evaluation as
the operational manager.
Variance Investigation Decisions
Various information constant over all decision trials within a
treatment condition was presented on a separate page prior to the start


81
of the decision trials. The extent of such information depended upon the
treatment condition but could include information such as the prior
probabilities of both states, the means of both states, the standard
deviations of both states, the shape of the distributions of both states,
the range of past labor efficiency variances, and the range of past
investigation costs. Copies of the prior information pages for two
specific treatment conditions are presented in Appendix D.
Each variance investigation decision trial consisted of the pre
sentation of a labor efficiency variance report and a subject's re
sponse to two questions. The questions were 1) would you investi
gate this reported variance, and 2) how strongly do you feel about your
decision? During the training phase each decision trial was followed
by feedback concerning the actual state of the assembly line and the
actual costs incurred for each possible decision given the actual state.
Decision trials were presented in booklets of 33 trials (each trial
included the report with questions followed by the feedback). Within
the experimental phase decision trials were presented in booklets of
50 trials (each trial included only the report with questions). In
both the training and the experimental phases, answer sheets were pro
vided for the subject to record his responses.
An example of a labor efficiency variance report with the set of
questions is presented in Figure 6. The format of the report and
questions was constant for all treatment conditions: the only variation
related to the distribution of the actual minutes incurred per chair
and the labor efficiency variance, and to the magnitude and structure
of the costs of investigation.


82
AMSECO
Metal Folding Chair Assembly Department
Labor Efficiency Variance Report For Job 5247
Standard Minutes
Allowed Per Chair
Actual Minutes
Incurred Per Chair
Labor Efficiency Total Chairs
Variance Per Chair Produced
36.0
44.0
-8.0
200
The Costs Associated With Investigation Are:
If Your
And If The
Then
Your Costs
Are
Investigation
Assembly Line
ieie 'A'* ir-irkkie + -kkk-kirk-kkk
Decision Is
State Is
Investigation
Production
Total
'kk'trirkirk'kir-kirk-k
**************
**-*******
Yes
In-Control
$ 28.33
$ 0.00
$ 28.33
Yes
0ut-0f-Control
$ 90.00
$ 0.00
$ 90.00
No
In-Control
$ 0.00
$ 0.00
$ 0.00
No
0u-0f-Control
S 0.00
$ 175.00
$ 175.00
Please answer the following questions placing your answers on the answer sheet:
A. Would you investigate this reported variance /circle the appropriate
response on the answer sheet/
NO
YES
B. How strongly do you feel about your decision /select a number between
0 and 100 which indicates the strength of your feeling and place this
number on the answer sheet/
0 10
20
30
40 50 60
70
80
90
100

*

* + *
k
*
k

-k kkiek kkkk-klrkkk kkkk kkkkkkk k -k kkkkk

k

*
k

k
k
k


Uncertain
Reasonably
Almost
Certain
Certain
FIGURE 6
VARIANCE INVESTIGATION DECISION TRIAL


83
Elicitation of Heuristics
The elicitation of each subject's heuristics was accomplished using
a predominately open-ended questionnaire (see Appendix E). The objec
tives of this questionnaire were 1) to determine the strategy the sub
ject used in answering the variance investigation questions within the
experimental phase, 2) to determine, where appropriate, the exact values
of any numerical rules used by the subject, 3) to determine whether the
subject thought he used his strategy consistently within the experimental
phase, 4) to determine the method or approach used by the subject in
forming his strategy within the training phase, and 5) to determine the
subjective importance attached to each information item presented to
the subject.
Elicitation of Subject Motivations
Subject motivations were elicited using a motivation questionnaire
(see Appendix F) developed by Snowball and Brown (1977). The question
naire is a ten item Likert-type scale which has submeasures for both
intrinsic and extrinsic motivation.
Experimental Procedures
Experimental procedures included assignment of subjects to
treatment conditions, adminstration of a training phase, administration
of an experimental phase, and final debriefing.


84
Assignment of Subjects to Treatment Conditions
Since each of the 86 subjects was assigned to one of eight groups,
randomization per se can not be relied upon to control for individual
attribute differences between groups. An alternative is to block the
randomization process on-indi vidual attribute dimensions assumed to
significantly affect the subject's information processing within the
task required by the experiment.
The literature relating to individual attributes has not pro
duced conclusive results concerning individual attributes effect on
decision processes. The most comprehensive study is that of Taylor and
Dunnette (1974) who analyzed the effect of 16 decision maker attributes
upon eight measures of predecisional, decision-point, and post decisional
processes within a personnel decision simulation experiment. Of con
siderable interest is that the relevant decision maker attributes
accounted for a generally small portion of the total decision variance
within the various processes (the range was from 8 percent to 33 percent).
The decision maker attribute with the greatest predictive capacity was
intelligence. In two of the predecisional processes (diagonosticity
and information processing rate) and in two of the decision-point and
post decisional processes (information retention and decision accuracy),
intelligence accounted for from one-half to almost all of the variance
explained by individual characteristics.
In the present study, the randomization process of assigning
subjects to treatment conditions was blocked on individual intelligence.
Ideally, individual intelligence should be measured using some validated
instrument (e.g., the Wesman Personnel Classification Test or the


85
Wechsler Adult Intelligence Scale). Due to resource limitations, however,
subject grade point average (GPA) was used as a surrogate for such a
measure. A median GPA was identified, and those subjects with a GPA
above the median were categorized as above average intelligence and
those subjects with a GPA below the median were categorized as average
intelligence. Each subject within an intelligence category then was
assigned randomly to one of the treatment conditions with the restric
tion that each intelligence group contributed an equal number of sub
jects to each condition.1 Upon assignment to a treatment condition each
subject received the background information booklet.
Training Phase
%
Each subject received training within the treatment condition to
*
which he was assigned. Training was conducted in groups of two subjects
within a 50 minute session administered by either the experimenter or
by an experimental assistant.2 Subjects were assigned randomly to the
experimental assistants subject to two restrictions. The restrictions
were that the experimental assistant was not currently the subject's
teacher for an academic course and that the subject's available time
coincided with that of the experimental assistant. Training of all
!The assignment process involved two procedures. First, the assignment
to the cost variable levels was by the week in which the subject par
ticipated in the experiment. Those subjects who participated during
the first week were assigned to the Cl cost level, and those subjects
who participated during the second week were assigned to the C2 cost
level. Second, the assignment to the information variable and the
distribution variable levels was by random selection based upon a ran
dom number table.
2The experimental assistants were not paid. The author believes, how
ever, that this is an exception to the aphorism that price reflects
value.


86
subjects (within each week) was completed over two contiguous days. The
training phase consisted of 99 decision trials with feedback. The
experimental materials used in the training phase were similar to those
used in the experimental phase. The decision trials with feedback were
presented in three booklets of 33 trials and the subject was provided
an answer sheet on which to record his responses. Additional perfor
mance feedback was given at the completion of each book of 33 decision
trials. This performance feedback was the subjects's TIDC^p for that
particular book.
Experimental Phase
The experimental session lasted one hour and was administered by
the experimenter. The experimental phase (within each week) was com
pleted over two contiguous days immediately following the training phase.
The experimental session consisted of three parts. The first
part was the presentation of 100 decision trials. The decision trials
were presented in two booklets of 50 trials each and the subject re
corded his responses on a separate answer sheet. The subjects were
allowed twenty minutes to complete the 100 decision trials. The second
part of the experimental phase was the elicitation of the subject's
heuristics. This elicitation was accomplished using the open-ended
questionnaire (see Appendix E). The third part of the experimental
phase was the administration of the motivation questionnaire (see
Appendix F).


87
Final Debriefing
Each subject's final performance measure (TIDCmi.n) for the
variance investigation decisions part of the experimental phase was
presented individually at a later date. At this time his cash pay
ment was determined, he was debriefed as to the purpose of the experi
ment, and any questions were answered. Additional data were collected
from those subjects not responding completely to the heuristic elicita
tion questionnaire.


CHAPTER V
ANALYSES AND RESULTS
Summary of Results
A summary of the analyses and results contained in this chapter
is presented in Table 4. The objective of this summary is to provide
a brief statement of each hypothesis together with the results of analy
ses for that hypothesis. Such a summary will facilitate the presenta
tion of the analyses and results and will serve as a reference for the
discussion and conclusions contained in the next chapter.
General Method of Analysis
Although several methods of analysis are employed, the most
prevalent method is analysis of variance using the model comparison
procedure (Appelbaum and Cramer, 1974; Lewis and Keren, 1977). This
method of analysis is employed due to the nonorthogonality of the data
structure. The problem of nonorthogonality arises in this instance as
a result of non-equal cell frequencies.
The model comparison procedure involves fitting a linear model
allowing for certain effects and then comparing the obtained fit to
that of a linear model which omits one or more of the effects. The
objective is to find the simplest model that adequately fits the data.
The procedure begins with the complete or full model (which allows for
88


TABLE 4
SUMMARY OF ANALYSES AND RESULTS
Hypotheses
Analysis
Method(s)
Test
Statistic(s)
Results
1.1a (TDVi1 SI)=38.5
Two-tailed Z test
Z=0.9912
(n.s.)
The mean initial decision anchor of
those subjects within the SI distribu
tion level was not significantly
different from 38,5
1.1b (TDVi|S2)=40.5
Two-tailed Z test
Z=1.5681
(p<.10)
The mean initial decision anchor of
those subjects within the S2 distribu
tion level was significantly less than
40.5
1.2a (TDV.|Il)=(TDVi
112)
Model comparison
Two-tailed Z test
F= 1.0944
Z=0.7007
(n.s.)
(n.s.)
The mean initial decision anchor of
those subjects within the reduced infor
mation level was not significantly dif
ferent from that of those subjects with
in the expanded information level.
1.2b (TDVi|Cl)=(TDVi
1 C2)
Model comparison
Two-tailed Z test
F=3.3931
Z=1.4362
(p<.10)
(p<.10)
The mean initial decision anchor of
those subjects within the Cl cost level
was significantly smaller than that of
those subjects within the C2 cost level.
1.3a (RA.|I1)=(RA.|I2)
Model comparison
Two-tailed Z test
F=0.0903
Z=0.4432
(n.s.)
(n.s.)
The mean relative decision anchor ad
justment of those subjects within the
reduced information level was not sig
nificantly different from that of those
subjects within the expanded information
level.


Full Text
subjects, who assumed the role of assembly department operational manager,
made labor efficiency variance investigation decisions based upon a series
of independent standard cost variance reports. The manipulated indepen
dent variables included contents of the available information set,
distributional properties of the states of control, and cost effects of
the investigation decision errors. Parameters of the psychological theory
of signal detection permitted measurement of sensitivity and criteria of
a subject's decision model. The subject's investigation decision costs
compared to normative investigation decision costs derived under similar
situations were employed as a relative measure of decision efficiency.
Overall, the implications of the conceptual framework were supported
by the obtained results. To explain the few major deviations from
expectations, an ex post hypothesis was introduced. This hypothesis
posits that the standard introduces a subjective adjustment bias. Depend
ing upon the direction of adjustment the standard subjectively inhibits
complete adjustment either to optimal decision values close to the
standard or to optimal decision values distant from the standard.
x


135
TABLE 21
MODEL COMPARISON PROCEDURE RESULTS FOR THE GPi/GN1 DEPENDENT VARIABLE
Model
SSe
d.f.(F)
F
Ful 1b
66.265
12,66
6.91a
Without 3-way interaction
67.538
1,66
1.27
Without distribution x cost
interaction
76.352
1,67
8.74a
Without information x distribution
interaction
67.607
1,67
0.07
Without information x cost
interaction
69.242
1,67
1.69
C
Without information
70.493
1,69
1.17
ap<.01
bfhe F value associated with the full
model is the
test of
the full
model and not a model comparison test,
c
The test of this model assumes the effects of those interactions
that include this variable are equal to zero.


140
increase at a rate greater than the relative additional decision savings
the subjects' variable response ranges increase.
The negative association between the G.¡ variable and the intrinsic
motivation factor indicates that as the relative net decision costs
increase the subjects' intrinsic motivation measures decrease. The same
association exists between the GP.¡ variable and the intrinsic motivation
factor: as the relative additional decision costs increase the subjects'
intrinsic motivation measures decrease. The GN.¡ variable and the in
trinsic motivation factor were associated positively: as the relative
additional decision savings increase the subjects' intrinsic motivation
measures increase.
Similar associations exist for the extrinsic (monetary) motivation
factor. Thus, as the relative additional decision costs increase (GP^
increases) the subjects' extrinsic (monetary) motivation measures de
crease. Conversely, as the relative additional decision savings in
crease (GN.¡ increases) the subjects' extrinsic (monetary) motivation
measures increase.
The negative association between the GP^ variable and the GPA-
variable indicates that as subjects' GPAs decrease relative additional
decision costs increase. Likewise, the positive association between the
GN.j variable and the GPA variable indicates that as subjects' GPAs in
crease relative additional decision savings increase.
Long-Run Decision Efficiency and Training Phase Performance
Hypothesis 4.5 predicted that those subjects in the PP classi
fication of training adjustment direction would have a signficantly
smaller mean G.¡ than those subjects in the MM classification. A t-test


81
of the decision trials. The extent of such information depended upon the
treatment condition but could include information such as the prior
probabilities of both states, the means of both states, the standard
deviations of both states, the shape of the distributions of both states,
the range of past labor efficiency variances, and the range of past
investigation costs. Copies of the prior information pages for two
specific treatment conditions are presented in Appendix D.
Each variance investigation decision trial consisted of the pre
sentation of a labor efficiency variance report and a subject's re
sponse to two questions. The questions were 1) would you investi
gate this reported variance, and 2) how strongly do you feel about your
decision? During the training phase each decision trial was followed
by feedback concerning the actual state of the assembly line and the
actual costs incurred for each possible decision given the actual state.
Decision trials were presented in booklets of 33 trials (each trial
included the report with questions followed by the feedback). Within
the experimental phase decision trials were presented in booklets of
50 trials (each trial included only the report with questions). In
both the training and the experimental phases, answer sheets were pro
vided for the subject to record his responses.
An example of a labor efficiency variance report with the set of
questions is presented in Figure 6. The format of the report and
questions was constant for all treatment conditions: the only variation
related to the distribution of the actual minutes incurred per chair
and the labor efficiency variance, and to the magnitude and structure
of the costs of investigation.


59
H2.2b (DNA-j | SI) = (DNAj |S2).
Those subjects within the SI distribution level and within the
S2 distribution level will have equal mean DNA-jS.
H2.2c (DA-¡ 1 Cl) = (DA-¡ IC2).
Those subjects within the Cl cost level and within the C2 cost
level will have equal mean DNA-¡s.
H2.3a (DNA-¡-DN-¡ ¡SI) < (DNAi-DNi |S2).
Those subjects within the SI distribution level will have
significantly smaller mean DNA-¡-DN-¡s than will those subjects
within the S2 distribution level.
H2.3b (DNAj-DN.¡ |Cl) <(DNAi-DNi | C2).
Those subjects within the Cl cost level will have significantly
smaller mean DNA-¡-DN-¡s than will those subjects within the C2
cost level.
Individual decision criteria simulation and hypotheses formation
The variables which affect the BN-j measure should be those factors
related to the individual's selection of a cutoff value. The most
significant variable affecting the individual's cutoff value should be
the cost variable. Within this variable the initial decision anchor
given the Cl level is much closer to the optimal cutoff value than is
the initial decision anchor given the C2 level. Consequently, the BN-j
measure should be closer to a value of one under the Cl level than
under the C2 level.
The difference between the BNA-¡ and the BN-j measures should
relate to the same variables as those discussed for the difference


TABLE 4
SUMMARY OF ANALYSES AND RESULTS
Hypotheses
Analysis
Method(s)
Test
Statistic(s)
Results
1.1a (TDVi1 SI)=38.5
Two-tailed Z test
Z=0.9912
(n.s.)
The mean initial decision anchor of
those subjects within the SI distribu
tion level was not significantly
different from 38,5
1.1b (TDVi|S2)=40.5
Two-tailed Z test
Z=1.5681
(p<.10)
The mean initial decision anchor of
those subjects within the S2 distribu
tion level was significantly less than
40.5
1.2a (TDV.|Il)=(TDVi
112)
Model comparison
Two-tailed Z test
F= 1.0944
Z=0.7007
(n.s.)
(n.s.)
The mean initial decision anchor of
those subjects within the reduced infor
mation level was not significantly dif
ferent from that of those subjects with
in the expanded information level.
1.2b (TDVi|Cl)=(TDV.
1 C2)
Model comparison
Two-tailed Z test
F=3.3931
Z=1.4362
(p<.10)
(p<.10)
The mean initial decision anchor of
those subjects within the Cl cost level
was significantly smaller than that of
those subjects within the C2 cost level.
1.3a (RAi|ll)=(RA.|l2)
Model comparison
Two-tailed Z test
F=0.0903
Z=0.4432
(n.s.)
(n.s.)
The mean relative decision anchor ad
justment of those subjects within the
reduced information level was not sig
nificantly different from that of those
subjects within the expanded information
level.


25
Available Information Set
The contents of the available information set refers to the infor
mation known by the individual prior to his decision. Such information
can be of two types-- singular or distributional (Tversky and Kahneman,
1977). Singular information consists of information that specifically
relates to the current decision. Distributional information consists
of information that relates to the relative frequencies and to the
statistical relationships among the states of nature.1 The difficulty
of the discrimination task may be affected by the presence or absence
of certain items of information. In general, the less information con
tained in the available set the more difficult the discrimination task,
for missing information required by a decision model must be estimated
by the individual. Compared to statistically derived estimates, these
subjective estimates are likely to have greater uncertainty and in
efficiency associated with them.
Within the variance investigation situation, the information con
tained on each variance report constitutes the singular information.
Two types of singular information are employed in this research-- the
actual results of the physical process and its variance from a standard,
and the marginal costs associated with each of the two possible decisions
in combination with each of the two possible states of nature. The
presence of both these singular information types is controlled as a
constant within this research. The presence or absence of specific
xThis definition of distributional information implicitly assumes that
the relative frequencies and statistical relationships are stable over
the relevant time frame. If these variables were non-stable, revised
estimates of their specific values would be required prior to each
decision, thus classifying them as singular information. This research
assumes that both of these variables are stable across all decisions.


86
subjects (within each week) was completed over two contiguous days. The
training phase consisted of 99 decision trials with feedback. The
experimental materials used in the training phase were similar to those
used in the experimental phase. The decision trials with feedback were
presented in three booklets of 33 trials and the subject was provided
an answer sheet on which to record his responses. Additional perfor
mance feedback was given at the completion of each book of 33 decision
trials. This performance feedback was the subjects's TIDC^^ for that
particular book.
Experimental Phase
The experimental session lasted one hour and was administered by
the experimenter. The experimental phase (within each week) was com
pleted over two contiguous days immediately following the training phase.
The experimental session consisted of three parts. The first
part was the presentation of 100 decision trials. The decision trials
were presented in two booklets of 50 trials each and the subject re
corded his responses on a separate answer sheet. The subjects were
allowed twenty minutes to complete the 100 decision trials. The second
part of the experimental phase was the elicitation of the subject's
heuristics. This elicitation was accomplished using the open-ended
questionnaire (see Appendix E). The third part of the experimental
phase was the administration of the motivation questionnaire (see
Appendix F).


117
interactions), the three motivation factors, and the GPA^ variable. The
three full models are:
- p + aj + Bk + Yi + 0Sjk + ctYj, + eykl + aSYjkl + £j
+ Ti + xi + 6i
The three model comparison procedure results are presented in
Table 14, and the F values associated with the sources of the reduced
models are presented in Table 15. The results indicate that the reduced
BN.j model contains the distribution variable, the cost variable, and
the extrinsic (monetary) motivation factor. Both the cost variable main
effect and the motivation factor are significant at the p<.01 level,
and the distribution variable main effect is significant at the p<.10
level. The reduced BNA- model contains the cost variable and the
extrinsic (monetary) motivation factor, and both are significant at the
p<.01 level. The reduced BNABN.¡ model contains the cost variable and
the GPA.j variable. The cost variable main effect is significant at the
pc.01 level, and the GPA.¡ variable is significant at the p<.10 level.
The means, variances, and sample sizes for the dependent variables
given the levels of the significant situation variables are included in
Table 15. The F test of equal variances indicated that the following
variances differ significantly: 1) within the BN- variable between the
levels of the distribution variable (F=1.51, p<.10), 2) within the BN-
variable between the levels of the cost variable (F=6.11, p<.01),
3) within the BNA^ variable between the levels of the cost variable
(F=9.36, p<.01), and 4) within the BNA BN^ variable between the levels
of the cost variable (F=39.74, p<.01). Using a Z test of equal means,
the means of the various levels were found to differ significantly
BNAi
BNA,-BN,


16
study by Magee and Dickhaut (1977). Using human subjects, Magee and
Dickhaut found support for the proposition that different manager per
formance (payoff) measures will affect the specific decision rules
(heuristics) employed by the decision maker (as a result of the decision
maker attempting to maximize his subjective utility).
Psychological Concepts
This section discusses the psychological concepts employed in
developing a conceptual framework of manager standard cost variance
investigation decisions. These concepts include the psychophysical
theory of signal detection and the human information processing con
cepts of decision heuristics.
Psychophysics
Psychophysics studies the relationships between physical and
psychological scales of measurement. Modern psychophysics adopts the
view that subjects can make meaningful evaluations of the magnitudes of
their sensory experiences, and therefore sensory magnitudes, as well
as physical magnitudes, can be quantified. One approach of modern
psychophysics is based upon the theory of signal detection (TSD). TSD
permits the separation of the decision maker's ability to discriminate
between classes of stimuli (sensitivity) from his motivational response
biases (decision criteria). Two comprehensive theoretical descriptions
of TSD are presented by Green and Swets (1974) and Egan (1975); general
surveys of the TSD theory are presented by Coombs et al. (1970), Watson
(1973), and Pastore and Scheirer (1974).


72
f(GPi/GN1¡S2,C2) = 1.28465
The difference between the two f(GP-j/GN-¡ ]SI) is much smaller than the
difference between the two f(GP-¡/GN-j |S2). This would indicate the
presence of a distribution by cost interaction in which the effect of
the cost variable is significant only under the SI distribution level.
A graphical representation of this interaction is presented in Figure 5.
A final implication derived from these relative relations in
volves the homogeneity of variance for the f(G^) given the different
levels of the cost variable. The variance for the f(G^|Cl) is equal to
5.26 x 10-6 for this simulation, and the variance for the f(G-j [C2) is
equal to 6.06 x 10"4 for this simulation. Using the F test for equal
variances, F=115.24 (3 and 3 d.f.) which indicates the variances are
not equal. Consequently, the variance of (G-j | Cl) is expected to be less
than the variance of (G-¡|C2).
Given the above simulation the effects of the independent vari
ables on the G-¡ variable can be hypothesized as follows:
H4.1 a2(G-j | Cl) < a2(G-j |C2).
The G-j of those subjects within the Cl cost level will have
significantly smaller variance than will the G-¡ of those subjects
within the C2 cost level.
H4.2a (GilCl) < (Of|C2).
Those subjects within the Cl cost level will have a significantly
smaller mean Gj than will those subjects within the C2 level.
H4.2b (G¡i|S2) < (Gi|Sl).
Those subjects within the S2 distribution level will have smaller
mean Gj than will those subjects within the SI level. Signifi
cance is not predicted due to the interaction effect with the
cost variable.


73
Distribution
Variable
FIGURE 5
SIMULATED COST BY DISTRIBUTION VARIABLE INTERACTION ON f(GPi/GNi)


181
Extent of Total Investi
gation Decisions Cost
Minimization Cash Bonus
1.021
to
1.030
$ 7.00
1.031
to
1.040
6.00
1.041
to
1.050
5.00
1.051
to
1.100
4.00
1.011
to
1.150
3.00
1.151
2.00
Again, I rely on your not discussing this experiment with other people.


104
TABLE 7
MODEL COMPARISON PROCEDURE RESULTS FOR THE RAj DEPENDENT VARIABLE
Model
SSe
d.f.(F)
F
a
Full
67.8716
8,73
0.58
Without 3-way interaction
67.8852
1,73
0.01
Without distribution x cost
interaction
67.8852
1,74
0.00
Without information x distribu
tion interaction
67.8889
1,74
0.00
Without information x cost
interaction
68.2633
1,74
0.41
b
Without information
68.3470
1,77
0.09
b
Without distribution
69.8016
1,77
1.73
b
Without cost
68.4208
1,77
0.17
a
The F value associated with the full
model is the
test of
the full
model and not a model comparison test.
I
The test of this model assumes the effects of those interactions
that include this variable are equal to zero.


38
The cost variable
Given the two states of nature (the realization of which is un
known to the individual) and two possible decisions, there follows that
two types of errors can be made in reaching a decision. The first type
of decision error, labeled type A, is the decision to investigate the
physical process when the in-control state exists. The second type of
decision error, labeled type B, is the decision not to investigate when
the out-of-control state exists. The marginal cost of either error
type equals the cost that would have been incurred had the correct
decision been made minus the cost that was incurred by the incorrect
decision. Thus, the marginal costs of the two decision error types,
labeled MC^ and MCg, are:
= C(decision=not investigate | state=in-control) -
C(decision=investigate | state=in-control), and
MCg = C(decision=investigate | state=out-of-control) -
C(decision=not investigate | state=out-of-control),
where C( ) represents the decision cost for the situation within the
parentheses.
When the marginal cost of a type A error equals the marginal
cost of a type B error the cost structure should be ignored when making
an investigation decision. Of greater interest are situations in which
the marginal error costs not equal. In the present study, each level
of the cost variable takes one of the following forms: 1) the
marginal cost of a type B decision error equals three times the marginal
cost of a type A error (labeled level CJ_), and 2) the marginal cost
of a type A error equals three times the marginal cost of a type B
error (labeled level C2).


Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
HUMAN INFORMATION PROCESSING FOR DECISIONS
TO INVESTIGATE COST VARIANCES
By
Clifton E. Brown
December 1978
Chairman: A. Rashad Abdel-khalik
Major Department: Accounting
Two major functions of management are those of planning and con
trol. Broadly defined, control is the management process which assures
that selected alternatives are implemented and executed in accordance
with plans. Important aspects of the control process concern analysis
and investigation of standard cost variances provided within accounting
reports. A substantial portion of the accounting variance investigation
literature has employed the normative model approach-- researchers have
created variance investigation models which a manager should use. Rarely
has attention been given to how the manager would interprete and integrate
information required by the various normative models.
The principal focus of this dissertation was on the effects of
specific situational variables upon a manager's information processing
for purposes of making variance investigation decisions. The specific
objectives were 1) tc develop a conceptual framework which will predict
effects cf specific situational variables on a manager's relative effi
ciency in information processing and in variance investigation decision
making, and 2) to empirically test some implications of this conceptual
framework.
viii


133
which the (G-j |S1 ,C1 )-(G.¡ |S2,C1) would be smaller than the (G^ [SI ,C2)-
(G-j |S2,C2). The results of the model comparison procedure indicated a
significant distribution by cost interaction (pc.Ol), with the
I SI ,C1)-(G-j ¡S2,C1) equal to 0.0074 and the (Gj |S1 ,C2)-(Gi |S2,C2)
equal to 0.0346. Hypothesis 4.2a predicted that (G^|C1) would be
significantly smaller than (G^|C2). A Z test of equal means indicated
that the means differ significantly (Z=2.4420, p<.01 one-tailed), with
the Cl level having the smaller mean. The lack of significance for
the cost variable main effect within the model comparison procedure
could be the result of the distribution by cost interaction. There
exists substantially less difference between the levels of the cost
variable when given S2 distribution level than when given the SI dis
tribution level. Hypothesis 4.2b predicted that (G.j|S2) would be
smaller than (G^|S1). A Z test of equal means indicated that the means
differ significantly (Z=1.7452, p<.05 one-tailed), with the S2 level
having the smaller mean. The lack of significance within the model
comparison procedure for the distribution main effect was predicted by
hypothesis 4.2b. Hypothesis 4.2c predicted that (G^|II) would not
differ significantly from (G^ j12). Using a Z test of equal means, the
difference was found to be not significant (Z=0.4528, n.s. two-tailed).
Hypotheses 4.4a and 4.4b relate to the ratio of the individual's
relative additional decision costs to the individual's relative addi
tional decision savings. The method of analysis was the model compari
son procedure where the dependent variable is the GP^/GN^ variable, and
the independent variables are the same as for the full G^ model. The
full GP.¡/GN.¡ model is:
Gpi/GN1 = + j + Bk + Y1 + aSjk + aYjl + 8Yk1 + a8Yjkl + +1 +
+ t + X. + 5.
1 l l


159
variance investigation decision behavior observed within this research.
Generalization of the modified conceptual development would involve
extensions aimed at reducing the limitations of the existing study.
The primary ex post concept developed within this research is
the decision behavior magnet view (incorporating the concept of sub
jective limits). Future research could directly test this concept.
Such research could involve variance investigation decision situations
in which the standard was not the same value as the mean of the
in-control state of nature. If this ex post concept is valid, then
subjects within situations involving convergent adjustment and a
strict standard should have less decision criteria conservatism than
subjects within situations involving convergent adjustment and a lax
standard. Similarly, subjects within situations involving divergent
adjustment and a lax standard should have less decision criteria
conservatism than subjects within situations involving divergent ad
justment and a strict standard. Additional research could test
whether the biases implied by the decision behavior magnet concept can
be reduced or eliminated by extended training within the particular
decision situation.
Generalizations of the conceptual development would require
extensions concerning both the experimental subjects and the experimental
environment. Extensions concerning the experimental subjects could
involve the use of non-students, perferably actual operational managers.
Such extensions could include managers within familiar experimental
environments and within non-familiar experimental environments. Be
yond extending the decision criteria conservatism effects, the central
question would be whether differential information processing


3
The second stage concerns the manager's investigation decision
criteria. Having arrived at a conclusion (albeit probabilistic) about
the distribution that generated the variance, the manager must integrate
and process various objective function parameters in order to arrive at
his variance investigation decision (these parameters can belong to
either the manager's objective function, the organization's objective
function, or both if the same). The potential investigation criteria
can be divided into two major categories: structural criteria and
behavioral criteria. The first category can be divided further into
structural cost criteria and structural probability criteria. The
structural cost criteria include such variables as the additional cost
of operating an out-of-control process (given that the process can be
returned to the original in-control state), and the costs of variance
investigation and correction (which can differ depending upon the
actual state that generated the variance). The structural probability
criteria include such variables as the probability that the source of
the variance is controllable (i.e., that it can be returned to the
original in-control state through managerial action), the probability
that the process will return to the in-control state without managerial
action, and the prior probabilities of each state distribution. The
behavioral criteria include such variables as the manager's perception
of the effect of the investigation upon his performance evaluation and
reward structure, and the manager's perception of the effect of the
decision upon employee performances and attitudes.
Although the variance investigation decision process has been
described as two stages, the manager may not actually utilize such a
sequential stage process. The manager's actual decision process is


39
Individual decision model sensitivity variables
The sensitivity of the individual's decision model relative to the
decision situation is measured using the TSD parameter d1. The theo
retical definition of d1 is:
d = (p2 Pj)/o = zi Z2
where y. = the mean of the jth state of nature (in-control = 1 and
out-of-control = 2);
a = the common standard deviation of the states of nature; and
z. = the value of the normal distribution function associated
J
with the jth state of nature and any decision cutoff value
common to both states.
An empirical estimate of the d1 for an individual, labeled d 1,
is obtained using the individual's conditional probabilities P(decision=
investigate | state=out-of-control) and P(decision=investigate | state=
in-control) to calculate a subjective z1 and z£. As previously pointed
out, d' is relative to the decision situation. In particular, d is
relative to the distribution variable. To gain comparability across
situations the following measure is used:
DN, / d¡
where d^ is generated using optimal model k. The measure decreases
from a value of one as the result of several factors: 1) the individual
is inconsistent in his use of his cutoff value (i.e., there exists a
range around his cutoff value within which decisions are not made
using a strict relation to this cutoff value) or the individual makes
one or more temporary processing errors, and 2) the individual
utilizes more than one cutoff value.


APPENDIX C
BACKGROUND INFORMATION BOOKLET
The following represents background information concerning the role
that you will be asked to assume in the business experiment for which
you have volunteered to participate. Please read this information
several times, familarizing yourself with the general aspects of this
role.
More detailed information together with specific instructions will be
presented at the inception of the actual experiment. You may use this
booklet when you participate in the experiment.
PLEASE DO NOT DISCUSS THIS EXPERIMENT WITH OTHER PEOPLE. If you have
questions contact Cl if Brown (Bryan 214E or telephone 2-0155).
My Training Session is:
Date:
Time:
Room:
My Experiment Session is:
Date: Time:
Room:
176


20
where <¡>( ) denotes the normal density function for the point in
parentheses and the z parameters are the same as defined for the d'
measure.
Although the assumption of equal variance normal distributions
is employed within this research, such an assumption is not necessary
to employ TSD. Egan (1975) demonstrates the use of TSD with exponential
distributions, chi-square distributions, Bernoulli distributions, and
Poisson distributions. Grier (1971) develops nonparametric measures
of discriminability and decision criteria.
Traditionally, psychophysics has employed TSD to study perceptual
processes: i.e., sensory processes such as audition and vision. Over
the last decade, however, TSD has been applied to conceptual processes.
These extensions to conceptual processes have included numerical
processing (Lieblich and Lieblich, 1969; Hammerton, 1970; Weissmar. et al.,
1975), medical diagnosis (Lusted, 1969; Lusted, 1971; Swets, 1972),
conceptual judgement (Ulehla et al., 1967a; Ulehla et al., 1967b), and
memory (Bernbach, 1967; Banks, 1970).
Human Information Processing
Human information processing (HIP), a subset of cognitive psy
chology, studies human judgement and decision making with particular
emphasis on the processing of information that determines these
activities. An area within HIP is the construction of models of human
decision making. The work within this area is typically classified
into schools of research which employ different paradigms, the two
major paradigms being the Bayesian and regression approaches (Slovic
and Lichtenstein, 1971). The TSD model is related to the Bayesian approach.


Copyright 1978
By
Clifton E.
Brown


113
significant at the pc.05 level, and the motivation factor is significant
at the p<.10 level.
Means, variances, and sample sizes of the three dependent variables
given the levels of the significant situation variables are included in
Table 12. Using the F test of equal variances, the following variances
were found to differ significantly: 1) within the DN.¡ variable between
the levels of the distribution variable (F=1.63, p<.10), 2) within the
DNA.j variable between the levels of the cost variable (F=2.57, pc.01),
and 3) within the DNA^-DN^ variable between the levels of the distribu
tion variable (F=2.20, pc.Ol). The Z test of equal means indicated that
the means of the variable levels differ significantly (pc.Ol) for each
of the situation variables found to be significant by the above sources
F values (see Table 12).
Hypothesis 2.2a predicted that (DNA-jll) would not differ signifi
cantly from (DNA.j|I2). A Z test v/as employed to test this hypothesis
and the difference was found to be not significant (Z=0.0010, n.s.
two-tailed). Hypothesis 2.2b predicted that (DNA-jSl) would not differ
significantly from (DAijS2). Again, a Z test indicated that the
difference was not significant (Z=0.4662, n.s. two-tailed). Hypothesis
2.2c predicted that (DNA^¡Cl) would not differ significantly from
(DNA^|C2). The results of the model comparison procedure and Z test
indicated that the difference was significant at the pc.Ol level, with
the Cl level having the greater mean.
Hypothesis 2.3a predicted that (DNA-DN^- jSl) would be significantly
smaller than (DNA-¡-DN.¡ |S2). Although the results of the model comparison
procedure and Z test indicated that there was a significant difference
between the means, the S2 distribution level had the smaller mean.


CHAPTER II
AREA OF INVESTIGATION
Managerial Accounting Concepts
This study relies on the synthesis of certain psychological con
cepts with certain accounting concepts. This section discusses the
accounting concepts: the nature of managerial decisions, managerial
task planning and control, and standard cost variance investigation.
Managerial Decisions
Since the information set (accounting report) exists to support
the manager's tasks, a conceptual framework of managerial decisions would
facilitate the accountant's information set selection. A conceptual
framework of managerial decisions should differentiate managerial de
cisions along dimensions that allow insights into the informational
needs of those decisions.
Anthony (1365) provides one such dimension along which he iden
tifies the purposes or orientation of managerial activities: strategic
planning, management control, and operational control. Another dimen
sion is provided by Simon (1966) who distinguishes between programmed
and nonprogrammed decisions. The underlying dimension is concerned
with the manner in which managers deal with their problems. The cri
teria for classifying a decision consist of the extent of structure
associated with the problem solving phases of the decision.
8


160
to the optimal model's total decision costs (GP-j) and the total
additional individual decision savings relative to the optimal model's
total decision costs (GN-¡). The hypotheses concerning these variables
were derived using a simulation which assumed a set of initial decision
anchors and assumed equal learning efficiency (in terms of relative
linear adjustments along the random variable axis). Given these
assumptions, specific variations in the decision situation were ex
pected to have specific effects on a subject's relative decision costs
(see Table 4 in Chapter V for a summary of these expected effects and
the obtained results). All of these expected decision situation effects
were supported, at least in part, by the obtained results. However,
the results of the individual decision model sensitivity and the indi
vidual decision criteria analyses indicated that learning efficiency
was not equal between the levels of the situation variables.
The support obtained for the hypotheses concerning decision
efficiency variables would suggest that the unequal learning efficiency
did not have a significant effect on the relative decision costs. The
majority of the unequal learning efficiency occurred between the levels
of the cost variable (the most significant effect, decision criteria
conservatism, was greater within the Cl cost level than within the C2
level). A closer examination of both the simulated and the obtained
cost variable effects on the G-¡ variable indicated that the obtained
effects were not as strong as the simulated effects (note that (Gj|C2)/
(GilCl) equalled 1.61 whereas f(Gj|C2)/f(Gj|Cl) equalled 3.27). The
simulated effects may be adjusted for unequal learning efficiency by
assuming the relative decision costs within the Cl cost level would be
increased by a factor of two (relative to the C2 cost level). After


This dissertation was submitted to the Graduate Faculty of the Department
of Accounting in the College of Business Administration and to the
Graduate Council, and was accepted as partial fulfillment of the require
ments for the degree of Doctor of Philosophy.
August 1978
Dean, Graduate School


187
6. Did you feel tense during the experiment?
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
k k k k k k k
Definitely Yes Probably Don't Probably No Definitely
Yes Yes Know No No
7. Did your desire to cooperate with the experimenter cause you to try
very hard?
*************************************************************
*******
Definitely No Probably Don't Probably Yes Definitely
No No Know Yes Yes
8. Do you think the money rewards caused you to try harder than you
would have without them?
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
k k k k k k k
Definitely Yes Probably Don't Probably No Definitely
Yes Yes Know No No
9. Did you enjoy the overall experience of participating in the
experiment?
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
k k k k k k k
Definitely No Probably Don't Probably Yes Definitely
No No Know Yes Yes
10.Did your desire to contribute to research knowledge cause you to try
very hard?
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
k k k k k k k
Definitely Yes Probably Don't Probably No Definitely
Yes Yes Know No No


18
Since the cells of this matrix are both exhaustive and mutually exclu
sive, the row-wise conditional probabilities must sum to one (this does
not necessarily hold for the sum of the column-wise conditional
probabilities). All parameters of the TSD model are derived from this
conditional probability matrix (see Appendix A for a more detailed
discussion of the conditional probability matrix including its rela
tionship with Bayes1 theorem).
In the single-interval task the subject analyzes the evidence
and classifies the stimulus into one of two categories according to
his criteria. The criteria are determined by his objective function.
The objective funcition of interest within this research is the
maximization of expected value. Assume that the subject has some value
(utility) fqr each of the four event outcomes. A payoff matrix of these
values related to the four outcomes is. given by the following (Egan,
1975):
DECISION
No
Yes
sn ^sn,Y ^sn,N
n vn,Y vn,N
The decision rule for maximizing the expected value (see Appendix A
for a derivation of this decision rule) is:
P(x|sn) P(n) Vn,W VV
P(xjn) P(sn) V$n,Y ^sn,N
At the point at which the above expression is an equality the subject
should be indifferent between saying "Yes" or saying "No." This point
can be considered the critical value of the likelihood ratio of the
observations, L(x0). The decision rule for saying "Yes" is expressed
by the relation L(x) > L(x0).


121
Relative Decision Criteria Conservatism
Hypothesis 3.3 relates to the effect of conservatism on the
relative individual decision criteria. The hypothesis proposes that
the greater the difference between the individual's initial decision
anchor and the optimal cutoff value, the higher the level of conser
vatism. The situation variable most related to such a difference is
the cost variable.
The method of analysis was the model comparison procedure where
the dependent variable is the BNC-¡ measure and the independent vari
ables are the three situation variables (with their interactions), the
three motivation factors, and the GPA-j variable. The full BNCj model
is:
BNCi = p + aj + ek + Y, + aSjk + Yji + BYk) + aSYjkl + Ej
+ t, + X, + ,
The model comparison procedure results are presented in Table
16, and the F values associated with the sources of the reduced BNCi
model are presented in Table 17. The results indicate that the re
duced BNCn- model contains the cost variable, the extrinsic (monetary)
motivation factor, and the GPA.¡ variable. Both the cost variable
main effect and the GPA^ variable are significant at the p<.05 level.
The extrinsic (monetary) motivation factor is significant at the
p<.01 level.
The means, variances, and sample sizes of the BNC^ measure given
the levels of the cost variable are included in Table 17. Using the
F test of equal variances, the variances of the BNC^ given the levels
of the cost variable were found to differ significantly (F=9.36, p<.01).


61
f(BNAi|Cl,Sl,I2) = f(6*) / f(6k) = 1.41898
f(BNAi|C2,Sl,I2) = f(e?) / f(f?k) = 0.47017
where f(ek) is the simulated measure at the cutoff value of the kth
optimal model.
This simulation can be extended to the S2 level of the distribu
tion variable given the 12 information level and can be extended to
both the SI and S2 levels of the distribution variable given the II
information level. The f(BNA-) measures obtained under the 12 infor
mation and S2 distribution conditions are as follows:
f(BNAiJCl,S2,12) = 1.42318
f(BNAi¡C2,S2,I2) = 0.47300
The f(BNA-) measures given the II information level are as follows:
f(BNAi|Cl,Sl,Il) = 1.40126
f(BNAi|C2,Sl,Il) = 0.46993
f(BNAi|Cl,S2,Il) = 1.38273
f(BNA.|C2,S2,Il) = 0.38094
Based on these f(BNA-) measures, hypotheses can be derived con
cerning the effects of the independent variables upon the BNA. measure.
The f(BNA^) measures can be averaged over the appropriate conditionals
to determine these effects. Averaging over all conditionals except the
information variable gives the following results:
f(BAiIII) = 0.90871
f(BAi|12) = 0.94633
Although there is a difference in the f(BNA-) measures due to the infor
mation variable this difference is small compared to the standard error
of these estimates (the difference divided by the standard error is
0.03762/0.39157 = 0.0961). Therefore, a significant effect due to the
information variable would not be expected.


122
TABLE 16
MODEL COMPARISON PROCEDURE RESULTS FOR THE BNC-j DEPENDENT VARIABLE
Model
SSe
d.f.(F)
F
b
Full
38.249
11,74
a
2.35
Without 3-way interaction
38.510
1,74
0.51
Without distribution x cost
interaction
39.291
1,75
1.52
Without information x distribu
tion interaction
38.543
1,75
0.06
Without information x cost
interaction
38.624
1,75
0.22
c
Without information
39.895
1,78
0.91
c
Without distribution
39.607
1,78
0.34
c
Without cost
42.778
1,78
a
6.61
a
p<.05
b
The F value associated with the full
model is the
test of
the full
model and not a model comparison test.
The test of this model assumes the effects of those interactions
that include this variable are equal to zero.


139
TABLE 23
CORRELATION COEFFICIENTS FOR INDIVIDUAL ATTRIBUTES WITH
LONG-RUN DECISION EFFICIENCY
Individual Attribute
G
GP
GN
GP /GN
DN
-0.7616
(.001)
-0.6528
(.001)
0.1868
(.085)
-0.6613
(.001)
Intrinsic motivation
-0.2854
(.008)
-0.3044
(.004)
0.2399
(.026)
-0.1758
(.121)
Extrinsic (monetary)
motivation
-0.1071
(.327)
-0.1887
(.081)
0.3018
(.005)
0.0185
(.871)
GPA
-0.1737
(.110)
-0.2526
(.019)
0.3377
(.002)
0.0861
(.450)
Note: The numbers within the parentheses are the exact -levels
associated with the rejection of the null hypothesis that the
correlation coefficient is equal to zero.


TABLE OF CONTENTS
ACKNOWLEDGEMENTS iii
ABSTRACT viii
CHAPTER I. INTRODUCTION AND OBJECTIVES 1
Variance Investigation Decision Processes 1
The Research Objectives A
Dissertation Organization 6
CHAPTER II. AREA OF INVESTIGATION 8
Managerial Accounting Concepts 8
Managerial Decisions 8
Management Task Planning and Control 9
Standard Cost Variance Investigation 11
Psychological Concepts 16
Psychophysics 16
Human Information Processing 20
CHAPTER III. RESEARCH METHODOLOGY AND DESIGN 23
General Conceptual Development 23
Decision Situation Structure 23
Available Information Set 25
Individual Information Processing Efficiency 26
Individual Ability to Expand the Information Set 27
General Research Design 28
Selection of Independent Variables 29


I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
A. Rashad Abdel-ffhali k, Chairman
Professor of Accounting
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Gary L. irHstrum
Associate Professor of Accounting
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Dougla<£^ T. STiowLalT1^-"
Assistant Professor of Accounting
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Richard A. Griggs ^ //
Assistant Professor of Psychology


TABLE 8
OBSERVED MEAN
EDV.jS AND
TDV.jS GIVEN THE
VARIOUS LEVELS OF THE
DISTRIBUTION AND COST VARIABLES
Distribution
Cl Cost
Variable Level
Expected Initial
C2 Cost
Variable Level
Variable
DV
TV
Decision Anchor
TDV
EDV
SI
37.83
38.16
38.25
38.62
39.01
S2
39.62
39.98
40.50
40.41
42.43
o
en


26
distributional information items (the statistical means, variances, and
distributional shapes of the two states) is manipulated as an indepen
dent variable. The presence of other distributional information items
(the relative frequencies and the allowed standard) is controlled as a
constant.
Individual Information Processing Efficiency
The individual's efficiency in processing the available infor
mation relates to the particular heuristics or strategies employed in
combining and weighting the various items of information. The term
efficiency implies a relationship between the individual's process
output (his decision) and a normatively correct or optimal decision.
The individual's decision and information processing performance can be
evaluated by comparing his performance against an optimal model.
Optimality refers to the best possible performance under given condi
tions. Since the optimal decision model relies on an incomplete infor
mation set rather than certain knowledge,even its performance can be
affected by both the structure of the situation and the contents of the
available information set.
The optimal models used in this research are a function of the
experimental environment: the single-interval procedure found within
the psychological theory of signal detection and the basic decision
situation of standard variance investigation. Within this environment
the optimal decision rule for minimizing (maximizing) the expected cost
(value) of a set of decisions is based upon an extension of Bayes'
theorem that takes into account the relative costs (values) of various


2
information provided. In particular, the manager's perception of the
standard setting process will have considerable influence upon his
variance investigation decision process. If the manager believes that
the standards are unrealistic (i.e., that the standards have been placed
too far from the in-control distribution) he may rescale either the
standards or the variances to correspond with his own perception of the
in-control distribution.2
If it is assumed that a manager accepts the standard setting pro
cess as realistic and does not rescale the standards or variances, his
variance investigation decisions may be viewed as the culmination of a
two-stage process. The first stage concerns the detection of the
particular distribution (e.g., in-control or out-of-control) that gen
erated the variance. The manager's performance of this task is a
function of the sensitivity of his decision process (model). This
sensitivity is affected by the structure of the particular situation
and by the manager's knowledge of this structure. The extent of the
manager's knowledge of the situation, in turn, is affected by the
available information (both contained within the variance report and
provided from other sources), and by the manager's ability to learn
from his experiences with the processes being controlled. In situations
where the controlled process distributions have some area of overlap
the results of a manager's detection process are probabilistic (i.e.,
prior to completing an investigation the existence of any given state
is uncertain). Furthermore, the greater the area of overlap, the more
difficult the discrimination task becomes.
2Within this context, an in-control distribution concerns statistical
congruence of production output and planned output in terms of
controllable resource utilization.


II Information Level
Cost
Cl
Cost
C2
Distribution
SI
1 1 i
36.90 37.575 38.25
+
39.77 41.29
Distribution
S2
C B A
h
38.7 39.6 40.50
43.19
H
45.88
Distribution
SI
12 Information Level
Cost
Cl
B A
+
36.86 37.555 38.25
Cost
C2
4
39.755 41.26
C B A
Distribution 1 } 1
S2 38.57 39.535 40.50
4-
42.59
4
44.68
Note: All scales are actual minutes incurred per chair produced.
FIGURE 2
SIMULATED INITIAL DECISION ANCHORS (POINTS A), FINAL DECISION
ANCHORS (POINTS B), AND OPTIMAL CUTOFF VALUES (POINTS C)


159
and thus parallel those obtained with the DNA^-DN^ measure. Both the
DNA^-DN^ and the BNA^-BN^ measures obtained results that indicated
greater multiple decision anchor usage when the subjects' initial de
cision anchors were closer to the standard (the SI distribution level
within the DNA^-DN^ measure and the Cl cost level within the BNA^-BN^
measure). No plausible explanation for these results is readily
apparent.
The obtained associations between the individual decision criteria
measures and the individual attributes of extrinsic (monetary) moti
vation and subject GPA have intuitive interpretation. As both the
extrinsic (monetary) motivation measures and the subjects' GPAs de
creased the level of conservatism increased. Furthermore, as the sub
jects' GPAs decreased the use of multiple decision anchors increased.
Of the two individual process variables (decision model sensitivity
and decision criteria), the decision criteria had the more signifi
cant affect on a subject's performance feedback measure and, in turn,
on the level of his monetary rewards. However, since motivation
pre-tests were not administered,causal relationships (i.e., decreased
motivation led to increased conservatism, or increased conservatism led
to decreased motivation) can not be claimed.
Individual Long-Run Decision Efficiency
The overall efficiency of an individual's decision making within
the experiment was measured using the G.¡ variable. The G- variable was
the individual's net additional decision costs relative to the optimal
model's total decision costs. The G^ variable was disaggregated into
two measures- the total additional individual decision costs relative


21
The Bayesian approach is a normative model specifying how a
decision should be made given certain internally consistent relation
ships among probabilistic beliefs. The basic beliefs of this approach
are that decisions should be based on subjective probabilities and
that these probabilities should be revised upon the receipt of addi
tional information in accordance with Bayes's theorem.
The major findings of Bayesian research are labeled conservatism.
The subjects, after receiving additional information, revise their
posterior probabilities in the same direction as the optimal model but
the revision is insufficient. Much of the research has focused on an
explanation of the cause of conservatism, the major explanations being
misperception (Peterson et al., 1968), misaggregation (DuCharme and
Peterson, 1968), and response bias (DuCharme, 1970).
Another area within HIP is the study of subjective information
processing principles and decision rules, labeled heuristics. A
heuristic, within this context, refers to a learned set of rules or
principles which are utilized by individuals in making the particular
decisions required of them. Research in this area has been concerned
with identifying systematic biases (relative to some definition of
optimal) of subjective heuristics within certain types of decision
tasks. Those information processing and decision rule biases iden
tified thus far have been labeled as an anchoring and adjustment
heuristic (Tversky and Kahneman, 1973), a representative heuristic
(Kahneman and Tversky, 1972; Swieringa et al., 1976), an avail-
ablility heuristic (Tversky and Kahneman, 1973), and the law of small
numbers (Tversky and Kahneman, 1971).


19
The critical value L(xQ) for this decision rule has two possible
values: a theoretical value which is a measure of the criteria of an
optimal (or ideal) subject and a subjective value which is a measure
of the criteria of an actual subject.
One set of TSD models assumes that both conditional probability
distributions are Gaussian; i.e., $(x|n) and $(x|sn) are normally
distributed. With the additional assumption of equal variance for both
distributions, the parameters which measure individual discrimination
sensitivity and individual decision criteria are labeled d1 and g,
respectively. The discriminability measure, d', has the following
theoretical definition:
where, ysn= the mean of the signal plus noise distribution;
vn = the mean of the noise alone distribution;
a = the standard deviation of both distributions;
zn = the value of the normal distribution function associated
with the noise alone distribution and any decision axis
cutoff value common to both distributions; and
z = the value of the normal distribution function associated
sn
with the signal plus noise distribution and any decision
axis cutoff value common to both distributions.
The d' measure theoretically is independent of the decision criteria
measure.
The decision criteria measure, B, has the following theoretical
definition:
B = i>(zsn) / i>(zn)


11
the task plans, and 2) facilitate management control evaluation of
operational control performance.
The operational control manager uses the specific task control
feedback to 1) decide whether the physical system performance is in
agreement with the specific task plan, and 2) to decide whether the
task can be restructured to bring performance back into agreement with
the plan (if performance and the plan differ). A task that can not be
restructured could have significance for the overall task control feed
back to the management control activity and possibly may lead to
modification of the original task plans. Such modification could in
turn have significance for the overall objective control feedback to
the strategic planning activity, and a modification of the original
overall objectives may follow.
Standard Cost Variance Investigation
Within the operational control activity the specific task plan
provides standards of performance in terms of expected component costs
and usages, and the desired physical outputs. The operational control
manager structures the tasks within the physical system and periodically
receives a standard cost variance report describing the system output
in terms of the task plan and the actual results. Upon receiving a vari
ance report the manager must decide the nature of the given variances.
Dopuch et al. (1967) present a classification of standard cost
variances that is based on the expected source of the variances. A
variance resulting from a random fluctuation of the physical system,
labeled a Type 1 variance, requires no operational control response if


180
If Your
Investigation
Decision Is
And If The
Assembly Line
State Is
Then
Your
Costs
Are
Investigation
Production
Total
Investigate
In-control
$
X
5
0.00
$ X
Investigate
Out-of-control
$
Y
$
0.00
$ Y
Not Investigate
In-control
$
0.00
$
0.00
$ 0.00
Not Investigate
Out-of-control
$
0.00
$ 175.00
$ 175.00
X represents the variable investigation cost when the assembly
line is in-control. This investigation cost is related to the size of
the labor efficiency variance; the more negative (unfavorable) the
variance the larger this cost.
Y represents the variable investigation cost when the assembly
line is out-of-control. This investigation cost also is related to the
size of the labor efficiency variance; the more negative (unfavorable)
the variance the smaller this cost. For any given labor efficiency
variance the investigation cost if the assembly line is out-of-control,
Y, is always larger than the investigation cost if the assembly line is
in-control, X. This is due to the investigation cost including the cost
of correcting the out-of-control labor procedure when the assembly line
is out-of-control.
The $175.00 cost is the expected marginal production cost of
operating next period in the out-of-control state. This cost is fixed
for all labor efficiency variance reports.
Your Performance Measure as the Operational Control Manager
The metal folding chair section manager (your immediate supervisor)
evaluates your control performance in terms of the minimization of both
investigation costs and production costs above the expected standard
(the total of these costs is called the Total Investigation Decisions
Cost). The extent to which your investigation decisions for a specified
period of time minimizes the Total Investigation Decision Cost is
determined as follows:
Extent of Total Investigation
Decisions Cost Minimization =
Total Investigation Decisions Cost of your decisions
Optimal Total Investigation Decisions Cost
The optimal Total Investigation Decisions Cost is determined by the
section manager at the end of the specified period of time. It i_s_
possible for your Total Investigation Decisions Cost to equal the optimal
Total Investigation Decisions Cost. AMSECO pays a cash bonus based on
this measure of the extent of Total Investigation Decisions Cost
minimization. The closer this measure is to one or less than one the
larger the cash bonus.
The cash bonus for 100 labor efficiency variance reports is as
fol 1ows:
Extent of Total Investi
gation Decisions Cost
Minimization
1.000
1.001 to 1.010
1.011 to 1.020
Cash Bonus
$ 10.00
9.00
8.00


175
2) A chance to earn up to $10.00 as payment for participating in
the experiment.
3) A chance to gain first-hand experience with an accounting
research project.
Your participation in the experiment would not involve a great
sacrifice of your time. The entire experiment requires only two
one-period sessions. The times of these sessions are flexible to meet
the requirements of your personal schedule. The first one-period
session, a training session, would be scheduled sometime on Monday or
Tuesday, April 10th and 11th. The second one-period session, the actual
experiment, would be scheduled sometime on Wednesday or Thursday,
April 12th and 13th.
I can assure you that your responses within the experiment will
be held strictly confidential. Only aggregate or anomymous results
will be made pub!ic.
I will pass around a sign-up sheet; I would appreciate it if
everyone would print their name and indicate with a check-mark whether
you are willing to participate. If you are willing to participate,
fill in your phone number and indicate some times that I can contact
you to schedule your participation.
I would like to thank you in advance for your time. The overall
success of this study is dependent upon a positive response from most
of you in this room. I am confident that as professionals you will
exceed my expectations.
Does anyone have any general questions?


7
aspects of human information processing such as the general heuristic
of anchoring and adjustment.
Chapter III synthesizes the general concepts of the previous
chapter and develops a conceptual framework of standard cost variance
investigation employing the methodology of the psychological concepts.
This conceptual framework was operationalized using simulation tech
niques, and experimental hypotheses were derived from the simulations.
The details of a laboratory experiment designed to test the
hypotheses derived from the conceptual variance investigation framework
are presented in Chapter IV. The experiment employed a between-subjects
design and was manipulated factorially using specific variance investi
gation situation variables. Modified parameters from the psychological
theory of signal detection were employed to measure subject's information
processing efficiency and decision model sensitivity.
The results obtained from the laboratory experiment are presented
in Chapter V. The model comparison procedure of non-orthogonal analysis
of variance was the primary method of analysis employed. Chapter VI
discusses the results, develops a modification (ex post) of the
original conceptual framework, and discusses some implications for
accounting and for future accounting research.


37
The distribution variable
Manipulation of the distribution variable involves two factors:
1)the distributional parameters of each state, and 2) the statistical
relationship between the states. The two levels of this variable are
generated through a change in the variance and a change in the stan
dardized distance between the means of the two states. The first
level of the distribution variable, labeled Sl_, has the following
parameters and relationships:
1) y1 = 36.0 actual minutes incurred per chair;
2) crn = a = Gj = 3.0 actual minutes incurred per chair; and
3) u12 = y + 1.5a, = 40.5 actual minutes incurred per chair,
where, y = the mean of the jth state of nature (in-control = 1 and
* J
out-of-control = 2) given the ith distribution level
(SI = 1 and S2 = 2);
a = the standard deviation of the jth state of nature given
' J
the ith distribution level; and
a = the standard deviation common to both states of nature
given the ith distribution level.
The second level of the distribution variable, labeled S2^, has
the following parameters and relationships:
1) y = 36.0 actual minutes incurred per chair;
2) a21 = a22 = a9 = 5.0 actual minutes incurred per chair; and
3) y2 = y21 + 1.8a = 45.0 actual minutes incurred per chair,
where y , a^-, and a^ are defined the same as in the SI level. The
third and fourth statistical moments of each state within each dis
tribution level are uncontrolled except that their deviations from a
normal distribution are minimized.


125
TABLE 18
CORRELATION COEFFICIENTS FOR EXTRINSIC (MONETARY) MOTIVATION AND
SUBJECT GPA WITH INDIVIDUAL DECISION CRITERIA VARIABLES
Model
Extrinsic (Monetary)
Motivation
Subject GPA
BN-¡
-0.1975 (.068)
-0.0468 (.669)
BNA-j
-0.1916 (.077)
-0.1000 (.360)
BNA-j-BN-j
-0.0564 (.606)
-0.1896 (.080)
BNCj
-0.2951 (.006)
-0.2346 (.030)
Note: The numbers within the parentheses are the exact a-levels
associated with the rejection of the null hypothesis that the
correlation coefficient is equal to zero.


102
of the distribution variable differ significantly (Z=7.6346, p<.01
two-tailed), with the S2 level having the larger mean. Differences in
the means of the levels of the cost variable are also significant
(Z=1.4362, p<.10 two-tailed), with the C2 level having the larger mean.
Hypothesis 1.1a predicted that (TDV^|S1) would not differ signifi
cantly from 38.25 actual minutes incurred. A single-sample Z test was
employed to test this hypothesis, and the difference was found to be not
significant (Z=0.9912, n.s. two-tailed). Hypothesis 1.1b predicted
that (TWi |S2) would not differ significantly from 40.5 actual minutes
incurred. A single-sample Z test indicated that the difference was
significant (Z=1.5681, p<.10 two-tailed). Hypothesis 1.2a predicted
that (TDV.j | II) would not differ significantly from (TDV^]I2). The Z
test used to test this hypothesis indicated no significant difference
(Z=0.7007, n.s. two-tailed). Hypothesis 1.2b predicted that (TDV^|Cl)
would not differ significantly from (TDV.¡|C2). This hypothesis was
tested by the above model comparison procedure and related Z test, and
the results indicated that the difference was significant (p<.10).
Decision Anchor Adjustment
Hypothesis 1.3 relates to an assumption made in the training phase
simulation: i.e., that subjects would have equal relative decision
anchor adjustment between the levels of the various experiment conditions.
The simulation arbitrarily used an adjustment of one-half the linear
distance between the appropriate expected initial decision anchor and
the appropriate optimal model decision cutoff value.
The method of analysis was the model comparison procedure where
the dependent variable is the RA^ measure and the independent variables


Hypotheses
Analysis
Method(s)
3.4b (BNA-¡-BN-,-1Cl) <
Model comparison
(BNAi-BN1|C2)
One-tailed Z test
4.1 a2(Gi|Cl) F test of equal
variances
4.2a (Gi|Cl)<(Gi|C2)
Sources F value
One-tailed Z test
4.2b (Gj1S2)<(Gi|SI)
Sources F value
One-tailed Z test
4.2c (Gi|Il)=(Gi|I2)
Model comparison
Two-tailed Z test
4 Continued
Test
Statistic(s)
F=11.4812 (pc.Ol)
Z=3.7090 (p<.01)
F=2.43 (p<.01)
F=2.56 (n.s.)
Z=2.442 (p<.01)
F=0.51 (n.s.)
Z=1.7452 (p<.05)
F=1.0488 (n.s.)
Z=0.4528 (n.s.)
Results
The mean multiple decision anchor effect
on relative decision criteria of those
subjects within the Cl cost level was
significantly larger than that of those
subjects within the C2 cost level.
The relative net decision costs variance
of those subjects within the Cl cost
level was significantly smaller than that
of those subjects within the C2 cost level.
The mean relative net decision costs of
those subjects within the C2 cost level
was larger than that of those subjects
within the Cl cost level.
The mean relative net decision costs of
those subjects within the SI distribution
level was larger than that of those sub
jects within the S2 distribution level.
The mean relative net decision costs of
those subjects within the reduced infor
mation level was not significantly dif
ferent than that of those subjects with
in the expanded information level.


100
TABLE 5
MODEL COMPARISON PROCEDURE RESULTS
FOR THE TDV-j
DEPENDENT
VARIABLE
Model
SSe
d.f.(F)
F
Ful 1d
91.6549
8,77
a
8.19
Without 3-way interaction
93.1248
1,77
1.23
Without distribution x cost
interaction
93.1426
1,78
0.01
Without information x distribu
tion interaction
93.1463
1,78
0.02
Without information x cost
interaction
93.7287
1,78
0.51
e
Without information
95.0374
1,81
1.09
e
Without distribution
164.6998
1,81
61.27a
. e
Without cost
97.6986
1,81
3.39C
a
p<.01
b
pc.05
Cp<.10
d
The F value associated with the full model is the test of the full
model and not a model comparison test.
e
The test of this model assumes the effects of those interactions
that include this variable are equal to zero.


127
models. Both the stated objective function and subject payoff function
involved the minimization of the decision costs. A subject's long-run
decision efficiency was measured in terms of his minimization of his
decision costs relative to those of an optimal model.
The method of analysis was the model comparison procedure where
the dependent variables are the GP.¡, and GN.¡ measures. The in
dependent variables are the three situation variables (with their
interactions), the DN.¡ variable, the three motivation factors, and the
GPA.j variable. Ideally, both the DN.¡ and BN^ variables should be
included in the model: however, the BN.¡ variable had significantly
greater association with the other independent variables than did the
DN.j variable (the R2 for the full BN.¡ model was 0.6474, whereas the
R2 for the full DN^ model was 0.2192). Inclusion of the BN.¡ variable
would produce a significant problem of multi colinearity. The three
full models are:
Gi
GPi = y + j + Bk + Y1 + a6jk + fjl + 6Ykl + a6Yjkl + *1 +
GN-- + t. + X. +6.
1 ill
The three model comparison procedure results are presented in
Table 19, and the F values associated with the sources of the reduced
models are presented in Table 20. The results indicate that the re
duced G. model contains the distribution by cost interaction and the
DN. variable. Both the interaction and the DN. variable are siqnifi-
l i a
cant at the p<.01 level. The reduced GP.. model contains the distribu
tion by cost interaction, the DN. variable, the intrinsic motivation
factor, the extrinsic (monetary) motivation factor, and the GPA.
variable. The interaction, extrinsic motivation factor, and GPA.
l
variable are significant at the p<.05 level. The interaction,


56
will equal the intersection point of the two distribution curves
within that level.
HI.lb (TV.|S2) = 40.5.
The mean initial decision anchor given the S2 distribution level
will equal the intersection point of the two distribution curves
within that level.
HI.2a (TDV.|I1) = (157^12).
The mean initial decision anchor given the II information level
will equal the mean initial decision anchor given the 12 infor
mation level.
HI.2b (TDViiCl) = (TVi|C2).
The mean initial decision anchor given the Cl cost level will
equal the mean initial decision anchor given the C2 cost level.
HI.3a (RAi111) = (RAi112).
The mean relative adjustment given the II information level will
equal the mean relative adjustment given the 12 information level.
HI.3b (RA^Sl) = (RAi ¡S2).
The mean relative adjustment given the SI distribution level will
equal the mean relative adjustment given the S2 distribution
level.
HI.3c (RAi|Cl) = (RA.|C2).
The mean relative adjustment given the Cl cost level will equal
the mean relative adjustment given the C2 cost level.
Individual decision model sensitivity hypotheses formation
The measures of individual decision model sensitivity should
generally be unaffected by the independent variables. These measures


APPENDIX E
SUBJECT HEURISTIC ELICITATION QUESTIONNAIRE
Please answer the following questions in the spaces provided. This is
not a test; there are no right or wrong answers. It is important that
you try to answer each question honestly. If you need additional space
use the back of this questionnaire.
1. What rule or set of rules did you use to make your investigation
decisions in the last 50 labor efficiency variance reports (the
second block)? Please include any numerical values that were part
of this rule or set of rules.
2. Do you feel you used the same rule or set of rules as you gave in
question one to make your variance investigation decisions in the
first 50 labor efficiency variance reports (the first block)? If
not the same, how does the rule or set of rules used in the first
50 reports differ from what you gave in question one? Please in
clude any numerical values used as part of the rules or sets of
rules that were different.
3.Rate the following items of information in terms of their importance
to you in making your variance investigation decisions (do not rate
the 'other' item unless you have specified some other information).
Circle the appropriate response.
A.The background information booklet.
irklrk-k'k'k-k'k'k-k'k'k'k-k'k'k'k'k-k'k'k'k'k'k'k'k'k'k'k'k'k'k'k'k-k'kk-k'k'k'k'k'k'k'k'k.'k-k
k
*



Not Slightly
Important Important
Moderately Very
Important Important
Extremely
Important
B. The information based upon past experience and past observation
which was presented in the front of each experiment booklet.
k'k-k'kk'kk-k'k-k'k-k-kkic'k'k'kkk-k'k-k'k'k-k-k'k'k'k'k'k'k-k'kk-k-k-k'k'k-k'k'k'k'k-k'k'k
* * *
Not Slightly Moderately Very Extremely
Important Important Important Important Important
C. The standard minutes allowed per chair.
**************************************************
k
k
*
*
*
Not Slightly
Important Important
Moderately Very
Important Important
Extremely
Important
184


TABLE 4-- Continued
Analysis Test
Hypotheses Method(s) Statistic(s) Resul ts
3.2 a2(BNA.|Cl) <
F test of equal
F-9.36 (p
K.01)
The adjusted relative decision criteria
variances
variance of those subjects within the Cl
cost level was significantly larger than
that of those subjects within the C2
cost level.
3.3a (MCi|Il) = (BCi|I2)
Model comparison
F=0.9075
(n.s.)
The mean decision criteria conservatism
of those subjects within the reduced
Two-tailed Z test
Z=0.6802
(n.s.)
information level was not significantly
different from that of those subjects
within the expanded information level.
3.3b (BNCi|Sl)=(BNC.|S2)
Model comparison
F=0.3377
(n.s.)
The mean decision criteria conservatism
of those subjects within the SI distribu-
Two-tailed Z test
Z=0.3295
(n.s.)
tion level was not significantly differ
ent from that of those subjects within
the S2 distribution level.
3.3c (BNCi|Cl)<(BNCi|C2)
Model comparison
F=6.6104
(p<.05)
The mean decision criteria conservatism
of those subjects within the Cl cost
One-tailed Z test
Z=2.5598
(pc.05)
level was significantly larger than that
of those subjects within the C2 cost
level.
3.4a (BNAi-BN.|Sl) <
Model comparison
F=0.0924
(n.s.)
The mean multiple decision anchor effect
on relative decision criteria of those
(BNAi-BNi|S2)
One-tailed Z test
Z=0.3892
(n.s.)
subjects within the SI distribution level
was not significantly different from that
of those subjects within the S2 distribu
tion level.
CO


149
Third, the obtained effects of the situation variables on the
relative decision costs were weaker than those predicted. The pre
dicted effects were based, in part, upon the assumption of equal
learning efficiency between the levels of the various situation vari
ables. However, as indicated earlier the obtained learning efficien
cies were not equal between various situation variable levels. A
plausible explanation for the weaker obtained effects of the situation
variables on the relative decision costs utilizes the unequal learning
efficiency results. That is, the incorporation of unequal learning
efficiencies within the prediction of the effects of the situation
variables on the relative decision costs will produce expected effects
with strengths similar to the obtained effects.
Finally, multiple decision anchor usage had greater effects upon
individual decision model sensitivity within situations where the state
distributions were closer together and had smaller variance. It had
greater effects on individual decision criteria within situations where
the decision costs favored making investigation decisions for smaller
(in absolute value) labor efficiency variances. Both of these results
are contrary to those predicted, and no explanations are offered.
Overall, the variable which had the largest impact on the relative
decision costs was the subjects' decision criteria conservatism. The
mean DNA^ measure over all subjects was 0.926. Considering that a
value of one would indicate variable response ranges were not used, the
subjects' overall decision model sensitiviy was approximately that of
the optimal model. The mean BNC^ over all subjects was 0.828. This
indicated that the average distance between the subjects' decision
criteria and the optimal model's decision criteria was 82.8 percent
of that of the optimal model. This was supported by the mean RA.¡ over


67
Area(GN-j |Cl,SI, 12) = 0.08483
Area(GP-j (Cl,S1,12) = 0.05048
AreaiGNi|C2,S1,I2) = 0.19780
Area(GPi|C2,Sl,I2) = 0.06549
These areas must be adjusted for the relative frequency differ
ences and the relative cost differences between the various conditions.
If the in-control distribution is given a relative frequency weight of
one,then the relative frequency weight of the out-of-control distribu
tion would equal two-thirds (P(in-control) = 0.60 and P(out-of-control) =
0.40). The relative cost weights can be obtained directly from the
manipulation of the cost variable. Given the Cl cost level, errors
under the out-of-control distribution equal three times the errors under
the in-control distribution. Given the C2 cost level, errors under the
in-control distribution are equal to three times the errors under the
out-of-control distribution. The four areas associated with the
various G-¡ measures can be converted into relative relations by taking
into account both of these relative weights (frequency and cost). These
relative relations are as follows:
f(GN.¡ | Cl,S1,12) = (0.08483) (1) (1) = 0.08483
f(GP-j | Cl,SI, 12) = (0.05048) (2/3) (3) = 0.10096
f(GNi|C2,Sl,I2) = (0.19780)-(2/3)-(1) = 0.13187
f(GPi|C2,Sl,I2) = (0.06549)(1)(3) = 0.19647
The difference between the relative relations of GP.¡ and GNj
given a level of the cost variable indicates the relative effect of
these intervals upon the Gj measure. The differences for this simu
lation are as follows:


22
This study will make particular use of the anchoring and
adjustment heuristic. In many situations, individuals first make
decisions by starting with an initial anchor (decision point) and then
adjust this initial anchor as they learn from their experiences. The
initial anchor can be suggested by the structure of the decision
situation, or can be the result of a partial computation or estimate.
Empirical tests involving the anchoring and adjustment heuristic in
dicate individuals do not sufficiently adjust their initial decision
point. That is, their adjustment is less than that which would allow
optimal processing of the available information (Slovic and Lichtenstein,
1971; Slovic, 1972; Alpert and Raiffa, 1968; Tversky and Kahneman,
1974).


130
extrinsic motivation factor, and GPA^ variable are significant at
the p<.05 level. The DN^ variable is significant at the p<.01 level
and the intrinsic motivation factor is significant at the p<.10 level.
The reduced GN.¡ model contains the intrinsic motivation factor, the
extrinsic (monetary) motivation factor, and the GPA^ variable. Both the
extrinsic motivation factor and the GPA.¡ variable are significant at the
pc.Ol level and the intrinsic motivation factor is significant at the
p<.05 level.
The means, variances, and sample sizes for the distribution by
cost interaction within the G^ and GP^ models are included in Table 20.
Figure 7 presents a graphic representation of both distribution by
cost interactions. Bartlett's test for homogeneity of variances in
dicted that the variances within the G^ model distribution by cost
interaction are significantly heterogeneous (x2=l0.66, p<.05 with
3 d.f.). The variances within the GP- model distribution by cost inter
action are considered homogeneous (x2=2.52, p<.25 with 3 d.f.).
Duncan's multiple range test of equal means indicated that the means
within the G^ model distribution by cost interaction form two groups:
the means within a group do not differ significantly (p=.01) but a
significant difference (p=.01) exists for the means between groups.
These groups are 1) (G.|S1,C2) and (G.¡ |S2,C2), and 2) |S2,C1),
(G\j |SI,C1), and (G.|S2,C2).
Hypothesis 4.1 predicted that the variance of (G^|C1) would be
significantly smaller tha n the variance of (G.j|C2). Using the F test
of equal variances, the variances differ significantly (F=2.43, pc.Ol),
with (G^ |Cl) having the smaller variance. Hypothesis 4.3 predicted a
significant distribution by cost interaction within the G- model in


TABLE 11
MODEL COMPARISON PROCEDURE RESULTS FOR THE DNi, DNAi, AND DNAi-DMi DEPENDENT VARIABLES
DNi
DNAi
DNAi-DMi
Model
SSe
d.f.(FT~
F
SSe
d.TTUT
F
SSe
d.f.(F)
F
Ful 1d
2.1852
11,74
4.66a
1.1753
11,74
2.03b
1.4200
11,74
0.97
Without 3-way interaction
2.2031
1,74
0.61
1.1758
1,74
0.03
1.4327
1,74
0.66
Without distribution x
cost interaction
2.2236
1,75
0.70
1.2105
1,75
2.22
1.4345
1,75
0.10
Without information x dis
tribution interaction
2.2032
1,75
0.00
1.1783
1,75
0.16
1.4343
1,75
0.09
Without information x
cost interaction
2.2123
1,75
0.31
1.1758
1,75
0.00
1.4412
1,75
0.45
. e
Without information
2.2324
1,78
0.00
1.2140
1,78
0.01
1.4453
1,78
0.01
. 6
Without distribution
2.4330
1,78
7.01a
1.2269
1,78
0.84
1.5566
1,78
6.01b
. e
Without cost
2.3748
1,78
4.98b
1.4522
1,78
15.31a
1.4575
1,78
0.66
V.Ol bp<.05 Cp<.10
^The F value associated with the full model is the test of the full model and not a model comparison test,
e
The test of this model assumes the effects of those interactions that include this variable are equal to
zero.


APPENDIX A
THEORY OF SIGNAL DETECTION DERIVATIONS
Relationship of Conditional Probability Matrix and Bayes1 Theorem
The unconditional probabilities of the four event outcomes can be
stated in terms of the conditional probabilities and the prior proba
bilities of the sn and n distributions: 1) the probability of a hit,
P(sn,Y) equals P(Y|sn)P(sn); 2) the probability of a miss, P(sn,N)
equals P(N |sn)P(sn); 3) the probability of a false alarm, P(n,Y) equals
P(Y|n)P(n); and 4) the probability of a correct rejection, P(n,N)
equals P(N|n)P(n).
Egan (1975, pp. 12-13) demonstrates the relationship between the
single-interval procedure and Bayes1 theorem. The posterior probability
of the event sn, given the discrete observation x, can be represented
as P(sn|x). Since the two events sn and n are exhaustive the addition
of P(sn|x) and P(njx) equals one. The posterior probability of the
event sn can be stated as P(sn,x) which equals:
P(x|sn)P(sn) = P(sn|x)P(x).
Rearranging the terms of this equality and expanding the P(x):
n/ i P(x|sn)P(sn)
P(sn x)= i¡!
P(x]sn)P(sn) + P(x[n)P(n)
This equation is a simple expression of Bayes' theorem. It can be ex
tended to the odds form, the posterior odds equals the prior odds times
the likelihood ratio of the observation x:
P(sn[x) P(sn) P(x|sn)
P(n|x) P(n) P(x|n)
172


192
Weissman, S. M.; Hollingworth, S. R.; Baird, J. C. "Psychophysical
Study of Numbers: III. Methodological Applications."
Psychological Research, 38 (1975), 97-115.
Zannetos, Z. A. "Standard Cost as a First Step to Probabilistic Control:
A Theoretical Justification, and Extension and Implications."
The Accounting Review, 39 (1964), 296-304.
Zigler, E. "Metatheoretical Issues in Developmental Psychology." In
Theories in Contemporary Psychology, M. Marx (Ed.). New York:
Macmillan, 1963.


BIOGRAPHICAL SKETCH
Clifton Edward Brown was born on May 17, 1947,in Kansas City,
Kansas. In June 1965 he graduated from St. Petersburg High School,
St. Petersburg, Florida. Mr. Brown enlisted in the United States Air
Force from 1965 through 1968, and was honorably separated with the
rank of Sergeant. He was married in December 1968 to Sandra Cornell.
Mr. Brown attended the University of South Florida from which
he graduated in June 1971 with a Bachelor of Arts degree majoring in
accounting. From graduation until 1974 Mr. Brown was employed as an
industrial accountant, during which time he held positions of con
troller and vice-president of finance. From September 1974 until
the present Mr. Brown has attended the University of Florida where
he has studied for a Doctor of Philosophy degree in business admin
istration (accounting).
While attending the University of Florida, Mr. Brown was the
recipent of the Price Waterhouse Foundation Grant, the American
Accounting Association Fellowship Grant, and the Haskin and Sells
Fellowship Grant. Mr. Brown was elected by the faculty as the
university representative to the 1975 American Accounting Association
Doctoral Consortium.
193


168
(or push) of the standard, subjectively limiting a complete (convergent)
adjustment to the optimal decision value. The decision behavior magnet
view represents a generalization of the subjective adjustment limit
concept. The subjective limit concept as presented earlier dealt with
decision criteria conservatism involving only convergent adjustment.
The question of whether decision makers could learn (either through
training or experience) to reduce the decision biases implied by the
decision behavior magnet concept remains unanswered. An alternative
approach, however, involves the standard setting process itself.
Previous research involving the nature of standards (e.g., strict
standards, currently attainable standards, lax standards) has primarily
dealt with the standard's motivational affects. Other things being
equal, the nature of the standard may affect the decision behavior
magnet biases.. Within situations requiring adjustment, lax standards
(which have values greater than the mean of the in-control state) may
reduce substantially divergent decision criteria conservatism. Within
situations requiring convergent adjustment, strict standards (which
have values less than the mean of the in-control state) may reduce
substantially convergent decision criteria conservatism.
Future Research
The implications of this study for future accounting research
take two general forms: modification of the existing conceptual
development and generalizations of the modified conceptual development.
Modifications of the existing conceptual development would involve
incorporating and testing the ex post explanations offered for certain


CHAPTER III
RESEARCH METHODOLOGY AND DESIGN
General Conceptual Development
A general problem confronting decision makers is that the indi
vidual must decide subjectively which state of nature is most probable
based upon some incomplete set of information. When an individual
deals repetitively with a similar situation his long-run decision
efficiency or "decision correctness" can be affected by 1) the struc
ture of the particular decision situation, 2) the contents of the set
of available information, 3) his efficiency in processing the available
information, and 4) his ability to expand the available information
through experience with, and observations of, the various states of
nature. Each of these four elements is discussed below with specific
reference to the objectives of this research.
Decision Situation Structure
The structure of the particular decision situation primarily de
pends upon several key variables. These variables include the number of
possible states of nature, the relative frequencies of the states, the
various statistical relationships among the states, and the relation
ships between the various decision outcomes (the costs incurred given a
specific decision and the existence of a specific state). Depending on
23


155
(DNA.j|C2). The subjective limit also should have had the effect of
skewing the measure away from those values which indicated larger
variable response ranges. The third moment (as expressed by the co
efficient of skewness) of the DNA- measure indicated that 1) the
skewness of the (DNA^|C1) distribution was positive (skewed away from
values which indicated larger variable response ranges), and 2) the
skewness of the (DNA-|C2) distribution was negative (skewed toward
values which indicated larger variable response ranges).
The multiple decision anchor variable (the DNA^-DN^ measure) was
affected primarily by the distribution variable. Those subjects within
the SI distribution level employed multiple decision anchors to a
larger extent than did those subjects within the S2 level. The hypo
theses posited a relation between multiple decision anchors and the
%
distance separating the lower tail of the in-control state from^the
subject's final decision anchor. It was expected that the larger this
distance (found within the C2 cost level and the S2 distribution level)
the greater would have been the subject's propensity to have perceived
a second out-of-control state below the existing in-control state
(and used multiple decision anchors). However, the obtained results
do not support this proposed relation.
The observed associations between the individual decision model
sensitivity measures and intrinsic motivation have intuitive inter
pretations. The intrinsic motivation factor measured experimental
task and experimental environment enjoyment as modified by decision
performance satisfaction. As the subjects use of multiple decision
anchors and variable response ranges increased, their intrinsic moti
vation appeared to decrease. Since the use of multiple decision


TABLE 4 Continued
2.3b
3.1a
3.1b
3.1c
Hypotheses
Analysis
Method(s)
Test
Statistic(s)
Results
(DNA-j-DN-j | Cl) <
(DNAi-DNi|C2)
Model comparison
One-tailed Z test
F=0.6618 (n.s.)
Z=0.8642 (n.s.)
The mean multiple decision anchor effect
on relative decision model sensitivity
of those subjects within the Cl cost
level was not significantly different
from that of those subjects within the
C2 cost level.
(BNAi|Il)=(BNAi|
12)
Model comparison
Two-tailed Z test
F=2.6738 (n.s.)
Z=0.7891 (n.s.)
The mean adjusted relative decision
criteria of those subjects within the
reduced information level was not sig
nificantly different from that of those
subjects within the expanded information
level.
(BNAi|Sl)=(BNAi|
S2)
Model comparison
Two-tailed Z test
F=1.9398 (n.s.)
Z=0.7048 (n.s.)
The mean adjusted relative decision
criteria of those subjects within the SI
distribution level was not significantly
different from that of those subjects
within the S2 distribution level.
(BNAi|Cl)>(BNAi|
C2)
Model comparison
One-tailed Z test
F=107.4386 (p<.01)
Z=10.6306 (p<.01)
The mean adjusted relative decision
criteria of those subjects within the Cl
cost level was significantly larger than
that of those subjects within the C2
cost level.


115
TABLE 13
CORRELATION COEFFICIENTS FOR INTRINSIC MOTIVATION WITH
INDIVIDUAL DECISION MODEL SENSITIVITY VARIABLES
Model
Intrinsic Motivation
d!
i
0.2045 (.059)
DN-j
0.2549 (.018)
DNA-¡
0.1724 (.113)
DNAj-DN-¡
-0.1672 (.124)
Note: The numbers within the parentheses are the exact a-levels
associated with the rejection of the null hypothesis that the
correlation coefficient is equal to zero.


138
significance for the cost variable main effect within the model com
parison procedure could be the result of the distribution by cost
interaction. The cost variable has a significant effect only when given
the SI distribution level.
Individual Attributes and Long-Run Decision Efficiency
The model comparison procedure results indicated that the DN-
variable, the intrinsic motivation factor, the extrinsic (monetary)
motivation factor, and the GPA variable had significant effects upon
various individual long-run decision efficiency measures.3 The
analysis of these effects required additional tests. A product-moment
correlation was computed for each of these individual attributes with
each of the dependent variables. The results of these correlations
are presented in Table 23.
The negative association between the variable and the DN-
variable indicates that as the relative net decision costs increase the
subjects' variable response ranges increase (DN- increases). The
same association exists between the GP.¡ variable and the DN.¡ variable:
as relative additional decision costs increase the subjects' variable
response ranges increase. The positive association between the GN^
variable and the DN.¡ variable indicates that as the relative additional
decision savings increase the subjects' variable response ranges decrease
(DN.j decreases). The GP^/GN^ variable and the DN^ variable were
associated negatively: as the relative additional decision costs
3See the last sections in this chapter for definitions and additional
analyses concerning these individual attributes.


TABLE 4-- Continued
Hypotheses
4.6 (Gi|PP)<(Gi|MP)=
(G-j | PM)<(G^1-1 MM)
Analysis Test
Method(s) Statistic(s)
Results
The mean relative net decision costs of
those subjects within the mixed training
session performances categories were less
than that of those subjects within the
constantly increasing training session
performances category and greater than
that of those subjects within the con
stantly decreasing training session
performances category.
LO
G">


62
Averaging over all conditionals except the distribution variable
gives the following results:
f(BA-¡ 1 SI) = 0.94009
f (BA-¡ 1S2) = 0.91496
Again, the difference in the f(BNAj) measures due to the distribution
variable is small compared to the standard error of the estimates (the
difference divided by the standard error is 0.02513/0.39174 = 0.C642).
A significant effect due to the distribution variable would not be
expected.
Averaging over all conditionals except the cost variable gives
the following results:
f(BA-¡ |Cl) = 1.40654
f(BA-¡ |C2) = 0.44851
The difference in the f(BNA-j) measures due to the cost variable is
large compared to the standard error of the estimates (the difference
divided by the standard error is 0.95803/0.02436 = 39.3308). Con
sequently, a significant effect would be expected for the cost
variable.
Another implication drawn from the f(BNAj) measures concerns the
homogeneity of variance for the measures given different levels of the
cost variable. The variance for the f(BNA-¡ | Cl) is .00034 and the vari
ance for the f(BNA-j |C2) is .00203. Using the F test for equal variances,
F=5.9344 (3 and 3 d.f.) which indicates the variances may not be equal.
The variance of (BNA-¡ [Cl) would be expected to be less than the vari
ance of (BNAj[C2).
This simulation can be extended to enable the formation of
hypotheses concerning the effects of the independent variables on the


TABLE 12
PARAMETER VALUES ASSOCIATED WITH THE REDUCED DNj, DNA-j, AND DNAj-DN-j MODELS
Source
F Values Associated
With
the Sources of the Reducec
1 Models
-DNi
DNi
DNAi
DNAi
d.f.
F
d.f.
F
d.f.
F
Distribution
1,79
7.10a
1,80
6.17 b
Cost
1,79
5.05b
1,80
15.52 a
Intrinsic motivation
1,79
8.64a
1,80
3.55C
1,80
3.59C
Extrinsic (non-monetary) motivation 1,79
1.68
1,80
1.38
1,80
0.24
Extrinsic (monetary) motivation 1,79
0.04
1,80
0.07
1,80
0.01
GPA
1,79
0.02
1,80
0.14
1,80
0.27
Means and Variances
Associated With
the Siqnificant
Variables
DNi
DNAi
DNAi-DNi
Variable Level N
Mean Variance
N
Mean
Variance
N
Mean
Variance
SI 44
0.8052 0.0391
44
0.9193
0.0183
44
0.1141
0.0249
S2 42
0.8845 0.0240
42
0.9329
0.0180
42
0.0484
0.0113
Cl 47
0.8789 0.0276
47
0.9726
0.0091
47
0.0938
0.0197
C2 39
0.8017 0.0369
39
0.8696
0.0234
39
0.0679
0.0185
9P<.01 bp<.05 Cp<.10


64
The above discussion concerning the individual decision criteria
measures can be summarized as follows:
H3.1a (BffjIll) = (BA-¡ 112).
The mean BNAj of those subjects within the II information level
and within the 12 information level will be equal.
H3.1b (BA-¡ |SI) = (BA-¡ |S2).
The mean BNAj of those subjects within the SI distribution level
and within the S2 distribution level will be equal.
H3.1c (BNA-j | Cl) > (BA-¡ | C2).
Those subjects within the Cl cost level will have a significantly
larger mean BNA-j than will those subjects within the C2 level.
H3.2 a2(BNA-j ¡C1) < a2(BNA-¡ | C2).
The BNA-¡ of those subjects within the Cl cost level will have a
significantly smaller variance than will the BNA-j of those sub
jects within the C2 cost level.
H3.3a (BCj ¡II) = (BCj 112).
The mean BNC-¡ of those subjects within the II information level
and within the 12 information level will be equal.
H3.3b (BNCj|Sl) = (BC-j |S2).
The mean BNC-j of those subjects within the SI distribution level
and within the S2 distribution level will be equal.
H3.3c (BCj|Cl) < (BCj j C2).
The mean BNCj of those subjects within the Cl cost level will be
significantly smaller than the mean BNCj of those subjects within
the C2 level.


46
on tha experiment measure can be analyzed using a discrete classifica
tion of the direction of the adjustments. This classification of the
measure is accomplished using its algebraic sign. The second and
third training session results can be grouped into one of four
classes: PP(+,+), PM(+,-), MP(-,+), and MMSubjects within
the MM classification have constantly increasing TG-jj measures and
subjects within the PP classification have constantly decreasing TG^j
measures.
Hypotheses Formation
As previously discussed, the formation of the experimental
hypotheses is accomplished primarily using the method of simulation.
Two general types of simulations are performed simulation of optimal
model performances and simulation of subjective investigation decision
performances. The first part of this section discusses the assumptions
and presents the results of the simulation of optimal model perform
ances. The second part of this section discusses certain assumptions
derived from the conceptual development that form the basis for the
simulation of subjective investigation decision performances. It also
presents the results of this simulation and the derivation of the ex
perimental hypotheses.
Simulation of Optimal Model Performances
The decision rule of the optimal model (given the objective func
tion of minimizing incurred costs) employed in this research is to
investigate the reported labor efficiency variance when:


137
GPi/GNi
Variable
FIGURE 8
DISTRIBUTION BY COST INTERACTION WITHIN THE GPi/GNi MODEL


TABLE 19
MODEL COMPARISON PROCEDURE RESULTS FOR THE G-¡, GP-¡, AND GN-¡ DEPENDENT VARIABLES
Gi
GPi
GNi
Model
SSe
d. f. ( F)
F
SSe
dTfTFT
F
SSe
-error-
F
c
Full
0.0809
12,73
12.08a
0.1824
12,73
a
7.46
0.0349
12,73
a
2.67
Without 3-way interaction
0.0811
1,73
0.17
0.1824
1,73
0.00
0.0352
1,73
0.52
Without distribution x
cost interaction
0.0895
1,74
a
7.71
0.1953
1,74
b
5.21
0.0356
1,74
0.96
Without information x dis
tribution interaction
0.0812
1,74
0.09
0.1824
1,74
0.01
0.0352
1,74
0.06
Without information x
cost interaction
0.0814
1,74
0.26
0.1826
1,74
0.07
0.0352
1,74
0.03
d
Without information
0.0826
1,76
1.05
0.1835
1,76
0.36
0.0357
1,77
0.05
d
Without distribution
0.0360
1,77
0.71
d
Without cost
0.0369
1,77
2.60
a b
p<.01 p<.05
c
The F value associated with the full model is the test of the full model and not a model comparison test,
d
The test of this model assumes the effects of those interactions that include this variable are equal to
zero.


97
all effects) and eliminates effects starting with the highest order
interactions. The model comparisons utilize the general linear model
theory (the method of least squares) as described by Scheffe" (1959)
and the F test as described by Lewis and Keren (1977) is used to test
the fit of the various models. Defining two models as ft and w, the
F test is:
d.f.(ft) SSe(w) SSe(ft)
F =
d.f. (aj-ft) SSe(ft)
where w is the more restricted model and the degrees of freedom associ
ated with the numerator and the denominator of the F statistic are
d.f.(co-ft) and d.f. (ft), respectively.
Given a dependent variable, the model comparison procedure is
used to find the simplest model which adequately fits the data. After
the simplest (or reduced) model is found the sources of this model are
presented with their corresponding F values. These source F values are
the same as model comparison F tests where the comparison models are the
reduced model and the reduced model without the corresponding source.
Within this research the source F values are also the same as the Type IV
sum of squares (Barr et al., 1976).
Unless otherwise specified, the a levels for the rejection of the
null hypothesis are pc.10 for main effects and p<.05 for interactions.
The selection of these a values is largely arbitrary: the values reflect
the preliminary nature of this research. Also, unless otherwise speci
fied, all 2-way interactions are tested against the without 3-way
interaction model, employing the assumption that the 3-way interaction
effect is equal to zero within the subject population. All the individual
attributes are retained in each model, employing the assumption that


35
Each standard variance report is concerned with the results of a
single job-order to produce a constant number of chairs and reports
only aggregate (overall assembly department) results. Each report
contains the aggregate standard assembly time allowed per chair, the
actual assembly time incurred per chair, the overall labor efficiency
variance per chair, the total number of chairs produced, and the costs
associated with each possible decision in combination with each
possible state of nature. All time units are presented in minutes.
The singular information contained in a variance report is independent
of that contained in previous variance reports.
The experiment is conducted in two phases-- a training phase
and an experimental phase. The training phase consists of three
contiguous sessions in which the subject learns his role and presumably
develops his decision strategy. Performance feedback is given at the #
completion of each training session. The experimental phase consists
of a single session in which the subject receives a series of variance
reports similar to those presented in his training session. In this
phase no performance feedback is given until after the completion of
the entire experiment. The subject is paid according to his performance
in the experimental phase.3
Operationalization of Variables
This research employs the following three decision situation
independent variables, each measured using a discrete classification:
3Greater detail concerning the experimental environment, task, and
procedures is presented in Chapter V.


116
the association between the motivation factor and individual decision
model sensitivity is only partially a result of the variable response
range (the remaining effect within the DNA^ measure). This drop in the
significance level could be explained by the negative association (albeit
an equally low significance level) between the motivation factor and
the DNA^-DN^ measure. As the use of multiple cutoff values decreases
(DNA^-DN^ decreases) the subjects' intrinsic motivation measures
increase.
Individual Decision Criteria
The following sections report the analyses and results of the
individual decision criteria variables. The variables include the
relative decision criteria, the relative decision criteria conservatism,
and the relationships between individual attributes and decision criteria.
Relative Decision Criteria
Hypotheses 3.1 and 3.2 relate to the adjusted relative individual
decision criteria. The adjustment concerns the elimination of the
effects of multiple cutoff values on individual decision criteria. The
variables which affect the adjusted relative decision criteria should be
those related to the individual's selection of a cutoff value. Of the
three situation variables the one most related to this selection is
the cost variable.
The method of analysis was the model comparison procedure where
the dependent variables are the BN.¡, BNA^, and BNA-BN- measures. The
independent variables are the three situation variables (with their


BIBLIOGRAPHY
Al pert, M., and Raiffa, H. "A Progress Report on the Training of
Probability Assessors." Unpublished manuscript, Harvard
University, 1968.
American Accounting Association. A Statement of Basic Accounting
Theory. Evanston, Ill.: The Association, 1966.
. "Report of the 1969-70, 1970-71 Committee on Managerial
Accounting." The Accounting Review Supplement, 47 (1972),
317-35.
Anthony, R. N. Planning and Control Systems: A Framework for
Analysis. Boston: Harvard University Press, 1965.
Appelbaum, M. I., and Cramer, E. M. "Some Problems in the Non-
orthoaonal Analysis of Variance." Psychological Bulletin,
81 (1974), 335-43.
Banks, W. P. "Signal Detection Theory and Human Memory." Psychological
Bulletin, 74 (1970), 81-99.
Barefield, R. M. "The Effect of Aggregation of Decision Making Success:
A Laboratory Study." Journal of Accounting Research, 10 (1972),
229-42.
Barr, A. J.; Goodnight, J. H.; Sail, J. P.; and Helwig, J. T.
A User's Guide to SAS 76. Raleigh, N.C.: SAS Institute, Inc.,
1976.
Bernbach, H. A. "Decision Processes in Memory." Psychological Review,
74 (1967), 462-80.
Bierman, H.; Fouraker, L. E.; and Jaedicke, R. K. "A Use Probability
and Statistics in Performance Evaluation." The Accounting
Review, 36 (1961), 409-17.
Coombs, C. H.; Dawes, R. M.; and Tversky, A. Mathematical Psychology:
An Elementary Introduction. Englewood Cliffs, N.J.:
Prentice-Hall, Inc., 1970.
188


143
clearly associated with any one of the previous sources. The rotated
factor pattern for the ten scale items is presented in Table 24.
Analyzing the eigenvalues of these factors, the sharpest break (or
decline in slope) occurred between the third and fourth factor. For
this reason only the first three factors are included within the
subject motivation analysis. The first three factors account for 62.1
percent of the variability in the motivation scale items.
The first motivation factor had heavy positive loadings on the
scale items "enjoy making the decisions," "satisfied with performance,"
and "enjoy overall experience of participating." The first motivation
factor was interpreted as basically measuring intrinsic motivation
stimulated by enjoyment of the experiment and the required task.
The second motivation factor had heavy positive loadings on the
scale items "desire to perform well," "desire to cooperate with the
experimenter," and "desire to contribute to research knowledge." This
factor was interpreted as basically measuring extrinsic motivation
stimulated by non-monetary sources.
The third motivation factor had heavy positive loadings on the
scale items "try less hard if money rewards halved" and "money rewards
caused you to try harder than without them." The third motivation
factor was interpreted as basically measuring extrinsic motivation
stimulated by monetary sources.
Using an orthogonal transformation matrix the individual scale
items were transformed into individual factor scores for each of the
first three motivation factors. Two methods of analysis were used to
test the association of subject motivation with the independent
variables: Duncan's multiple range test and Bartlett's test of homo
geneity of variance.


179
The
ment are:
per chair labor efficiency standards for the assembly depart-
Process
Standard Time Allowed
Assembly
Packaging
Total
The variance report contains
variance is reported only in
33
3
minutes
minutes
36 minutes
only the labor efficiency variance and this
aggregate form (i.e., the variance is not
broken into the assembly and packaging processes). The accounting
system is computerized and the variance report is in computer output
format.
The physical labor process of the assembly department can be in one
of two states; either in-control or out-of-control. The overall labor
process is made up of many individual physical labor procedures; the
expected aggregate of these procedures is represented by the labor
efficiency standard of 36 minutes per chair. The department is defined
to be in-control when all of these physical procedures are performed as
expected. The department is defined to be out-of-control when one or
more of these physical labor procedures are not performed as expected.
Since the overall labor process is made up of many individual labor
procedures it is possible for the department to be out-of-control while
at the same time the reported labor efficiency variance is small. This
would occur when some of the individual procedures are performed at
higher than normal efficiency while other procedures are performed at
subnormal (out-of-control) efficiency. It is also possible for the
department to be in-control while at the same time the reported labor
efficiency variance is large. This would occur when most of the indivd-
ual procedures are performed at slightly below normal efficiency. These
examples represent two extreme situations; the probability of occurance
for either of them is low when compared to the probabilities of occur
ance for situations that fall inbetween the two extreme examples.
Your Task as the Operational Control Manager
Based upon the variance report you must decide whether to investi
gate the particular labor efficiency variance for its underlying causes.
The purpose of investigation is to facilitate correcting those indivdual
labor procedures which are not operating as expected. The following
assumptions will aid your decision making:
1) If you decide to investigate the variance and the assembly
department turns out to be out-of-control, the department will
be returned to the original in-control state with certainty.
2) If you decide not to investigate the variance and the assembly
department turns out to be out-of-control, the department will
remain out-of-control with certainty.
There are certain marginal costs associated with the variance
investigation decision. These costs depend upon your decision (either
investigate or do not investigate) and upon the actual state of the
assembly department (either in-control or out-of-control). These mar
ginal costs have the following values and relationships:


132
0.03 1 J-
SI S2
Distribution
Variable
0.13
0.12
0.11
GPi 0.10
0.09
0.08
0.07
C2
Cl
h
SI S2
Distribution
Variable
Cost
Variable
Cost
Variable


87
Final Debriefing
Each subject's final performance measure (TIDCmi-n) for the
variance investigation decisions part of the experimental phase was
presented individually at a later date. At this time his cash pay
ment was determined, he was debriefed as to the purpose of the experi
ment, and any questions were answered. Additional data were collected
from those subjects not responding completely to the heuristic elicita
tion questionnaire.


TABLE 20
PARAMETER VALUES ASSOCIATED WITH THE REDUCED G.¡, GPi, AND GNj MODELS
F Values
Associated
! With the
Sources of
the Reduced
Models
Source
Gi
GPi
GNi
d.f.
F
d.f.
F
d.f.
F
Distribution
1,70
0.51
1,70
0.77
Cost
1,70
2.56
1,70
0.11
Distribution x cost
1,70
7.92a
1,70
5.39b
DNi
1,70
88.76a
1,70
46.39a
1,80
0.90
Intrinsic motivation
1,70
2.40
1,70
3.95C
1,80
5.00b
Extrinsic (non-monetary) motivation 1,70
0.00
1,70
0.01
1,80
0.05
Extrinsic (monetary) motivation
1,70
1.71
1,70
4.33b
1,80
7.59a
GPA
1,70
2.39
1,70
5.93b
1,80
9.82a
Means and
Variances
Associated
With the
Significant
Variables
... Gi
GP
j
Variable Level
N
Mean
Variance
N
Mean
Variance
si,
Cl
24
0.0504
0.00229
24
0.0987
0.00456
SI,
C2
20
0.0921
0.00371
20
0.1372
0.00538
S2,
Cl
23
0.0430
0.00097
23
0.0886
0.00297
S2,
C2
19
0.0575
0.00375
19
0.0934
0.00576
3p<.01
V.05
Cp<.10
ro
CO


106
The expected pattern given the SI distribution level was obtained. How
ever, the expected pattern given the S2 distribution level was not con
sistent with respect to the expected initial decision anchor (40.5).
Individual Decision Model Sensitivity
The following sections report the analyses and results of the
individual decision model sensitivity variables. The variables include
the d.! dependent variable, the relative decision model sensitivity, and
the relationships between intrinsic motivation and decision model
sensitivity.
The dj Dependent Variable
Hypothesis 2.1 relates to the fit of the subjects' responses to
the TSD model. The dependent variable d.! should have little relation
with the information and cost situation variables, but should have a
relation with the distribution variable. The relation with the dis
tribution variable is derived from the optimal model decision sensi
tivity measure, d'. Within the SI level of the distribution variable
the theoretical value of d^ is 1.5, and within the SI level of the dis
tribution variable the theoretical value of d.1 is 1.8.
k
The method of analysis was the model comparison procedure where
the dependent variable is the d.! measure and the independent variables
are the three situation variables (with their interactions), the three
motivation factors, and the GPA. variable. The full d! model is:
i i
di + j + Bk + Y, + aBjk + otYj, + BYkl + aByjkl + C, + T,


103
are the three situation variables (with their interactions) and the
GP/l- variable. The full RAj model is:
RAj = ii + a, + ^ + Y, + aeJk + oty-j, + BYkl + aBYjkl + 6,
The model comparison procedure results for this analysis are
presented in Table 7. The results indicate that the reduced model con-
taines only the p parameter: i.e., none of the situation variables had
a significant effect on an individual's relative decision anchor adjust
ment. Using the F test of equal variances, the variances between the
levels of the distribution variable and between the levels of the cost
variable were found to differ significantly (F=5.22, p<.01; F=6.23,
pc.01). The S2 distribution level and the Cl cost level had the greater
variances.
Hypothesis 1.3a predicted that (RA-|I1) would not differ signifi
cantly from (RA^|I2). A Z test indicated that the difference was not
significant (Z=0.4432, n.s. two-tailed). Hypothesis 1.3b predicted
that (RA.JS1) would not differ significantly from (RA^|S2). The
application of the Z test revealed no significant difference (Z=1.3052,
n.s. two-tailed). Hypothesis 1.3c predicted that (RA.¡|C1) would not
differ significantly from (RA-|C2). A Z test was employed to test this
hypothesis: again, the difference was found to be not significant
(Z=0.4865, n.s. two-tailed).
Table 8 presents the actual patterns of the TDVi and EDV.¡ measures
given the various levels of the distribution and cost variables. Given
that the TDV.¡ measures have incorporated some initial adjustment (during
the first training session), the expected TDVi and EDVi patterns are
1) (DV1|C1,S1) < (TVi|Cl,Sl) < 38.25 < (TDVi|C2,Sl) < (EDVi|C2,S1),
and 2) (EDV.|C1,S2) < (TDVi |Cl ,S2) < 40.5 < (TDVi |C2,S2) < (EDVj|C2,S2).


66
o
o
LO
o
LO
o
o
o
LO
LO
LO
LO
o
LO
o
CO
LO
C\J
r-.
LO
(NJ
'






LO
LO
r^
CO
CT>
o
CO
CO
CO
CO
CO
*3
FIGURE 3
GRAPH OF THE SIMULATION USED IN THE CONDITION
(12 INFORMATION, SI DISTRIBUTION)


13
the output when the system is known to be out-of-control, 3) the actual
output for the period, and 4) the deviation between the standard output
and the actual output for the period (the standard output variance).
The investigation decision criteria category can include such variables
as 1) the prior probabilities of the system's states, and 2) the
costs and benefits associated with each possible investigation decision
in combination with each possible system state.
Information relating to investigation decision criteria that is
not typically presented within the variance report includes 1) the
probability that the cause of the variance is controllable (assumed by
this research to equal one), 2) the probability that the system will
return to the in-control state without managerial action (assumed by
this research to equal zero), 3) the perceived effect of the investi
gation decision on manager performance evaluation, and 4) the perceived
effect of the investigation on employee performances and attitudes.
The variance investigation literature is concerned mainly with
the investigation significance of Type 1 and Type 2a variances. A
general approach described within the literature is that of the Shewhart
X chart procedure (Probst, 1971; Koehler, 1968; Luh, 1968; Jeurs, 1967;
Zannetos, 1964). This approach, based upon classical statistics,
involves sampling the system to construct an in-control mean and stan
dard deviation. Arbitrary control limits (generally plus or minus three
standard deviations) are used as the criteria for making a variance
investigation decision. Only one of the above studies considered the
effects of the information provided on the manager's investigation
decision. Probst (1971), in an industrial field study, found foremen
unwilling to accept the procedure as their investigation decision rule


183
The two assembly line state means, in minutes per chair, are:
In-control 36.0 Out-of-control 45.0
The two assembly line state standard deviations, in minutes per chair:
In-contorl 5.0 Out-of-control 5.0
The portion of time the assembly line was:
In-control 60% Out-of-control 40%
The actual minutes per chair for both states are normally distributed.
The most favorable labor efficiency variance was:
10.0 minutes per chair (26.0 actual minutes incurred per chair)
The most unfavorable labor efficiency variance was:
19.0 minutes per chair (55.0 actual minutes incurred per chair)
The maximum investigation cost when the assembly line state was:
In-control $ 98.75 Out-of-control $ 154.17
The minimum investigation cost when the assembly line state was:
In-control $ 62.50 Out-of-control $ 142.08


29
1) the effects of the decision situation structure, the available infor
mation set contents, the individual's information processing efficiency,
and the individual's learning efficiency on the individual's long-run
decision efficiency; and 2) the effects of the decision situation
structure and the available information set contents on the individual's
information processing and learning efficiency.
Selection of Independent Variables
The effects of two types of variables are of primary interest
within this study: these are the situation variables and the process
variables. The following discussion describes the selection of each
of the variables employed in this study.
The situation variables
The situation variables are the quantity of information available
to the individual prior to his decision, the statistical structure of
the two states of nature, and the cost structure of the possible decision
outcomes.
The available information set. The effects of the contents of
the available information set are studied by manipulating the presence
and absence of certain distributional information. This involves the
specification of two levels of available information set content. Since
the difficultly of the discrimination task may be increased by the
absence of certain information, one level of the information variable
has less distributional information than the other level. This inde
pendent variable is labeled the information variable.


108
TABLE 9
MODEL COMPARISON PROCEDURE RESULTS
FOR THE d'.
i
DEPENDENT
VARIABLE
Model
SSe
d.f.(F)
F
Full5
5.8767
11,74
4.66a
Without 3-way interaction
5.8926
1,74
0.20
Without distribution x cost
interaction
5.9032
1,75
0.13
Without information x distribu
tion interaction
5.8947
1,75
0.03
Without information x cost
interaction
5.8969
1,75
0.05
Without information0
5.9950
1,78
0.59
Without distribution0
9.4035
1,78
46.10a
Without cost0
5.9304
1,78
0.26
V.oi
bThe F value associated with the full
model is the
test of
the full
model and not a model comparison test,
c
The test of this model assumes the effects of those interactions
that include this variable are equal to zero.


Relative Decision Model Sensitivity 107
Intrinsic Motivation and Decision Model Sensitivity 114
Individual Decision Criteria 116
Relative Decision Criteria 116
Relative Decision Criteria Conservatism 121
Individual Attributes and Decision Criteria 124
Individual Long-Run Decision Efficiency 126
Relative Decision Costs 126
Individual Attributes and Long-Run Decision Efficiency ... 138
Long-Run Decision Efficiency and Training Phase Performance 140
Individual Attributes 141
Subject Grade Poir.t Average 141
Subject Motivations 142
CHAPTER VI. DISCUSSION OF RESULTS 146
Conceptual Development Revi si ted 146
Overall Results 148
Discussion of Results 150
Decision Anchor Selection and Adjustment 151
Individual Decision Model Sensitivity 153
Individual Decision Criteria 156
Individual Long-Run Decision Efficiency 159
Limitations 163
Implications for Accounting 166
Value of Additional Information 166
General Standard Setting Process 167
Future Research 168
APPENDIX A. THEORY OF SIGNAL DETECTION DERIVATIONS 172
vi


57
have a greater relationship with individual decision processes than
with the specific conditions of a general decision situation. The
individual decision model sensitivity measure dj should have a re
lationship with the distribution variable only. The theoretical d1
measure under the SI distribution level is 1.5 and under the S2 dis
tribution level is 1.8. Thus, the average dl under the S2 distribu
tion level should be greater than under the SI distribution level.
The DNA.¡ measure should have no systematic relationships with
the three independent variables. This measure of the relative de
viation of the individuals decision model sensitivity from that of
the optimal model's generally is due to individual decision incon
sistencies and temporary processing errors. If individuals are
randomized into the specific decision situations there are no a
priori reasons for expecting significant differences in this measure
due to the independent variables.
The difference between DNA^ and DN-¡ measures the effects of
multiple decision anchors upon the individual's relative decision
model sensitivity. Relationships between this difference and the in
dependent variables depend upon whether the independent variables are
related to the causes underlying an individual's use of more than one
cutoff value. A possible explanation involves the distribution vari
able. As the variance of the in-control state increases the absolute
distance between the lowest values of the random variable of interest
(actual minutes incurred) and the mean of the in-control state in
creases. In turn, as this distance increases subjects may be more
inclined to perceive that a second out-of-control distribution overlaps
the lower range of the in-control state. If such perceptions underly


85
Wechsler Adult Intelligence Scale). Due to resource limitations, however,
subject grade point average (GPA) was used as a surrogate for such a
measure. A median GPA was identified, and those subjects with a GPA
above the median were categorized as above average intelligence and
those subjects with a GPA below the median were categorized as average
intelligence. Each subject within an intelligence category then was
assigned randomly to one of the treatment conditions with the restric
tion that each intelligence group contributed an equal number of sub
jects to each condition.1 Upon assignment to a treatment condition each
subject received the background information booklet.
Training Phase
%
Each subject received training within the treatment condition to
*
which he was assigned. Training was conducted in groups of two subjects
within a 50 minute session administered by either the experimenter or
by an experimental assistant.2 Subjects were assigned randomly to the
experimental assistants subject to two restrictions. The restrictions
were that the experimental assistant was not currently the subject's
teacher for an academic course and that the subject's available time
coincided with that of the experimental assistant. Training of all
!The assignment process involved two procedures. First, the assignment
to the cost variable levels was by the week in which the subject par
ticipated in the experiment. Those subjects who participated during
the first week were assigned to the Cl cost level, and those subjects
who participated during the second week were assigned to the C2 cost
level. Second, the assignment to the information variable and the
distribution variable levels was by random selection based upon a ran
dom number table.
2The experimental assistants were not paid. The author believes, how
ever, that this is an exception to the aphorism that price reflects
value.


147
(in-control and out-of-control) for a given production labor process.
A random variable produced by the accountant provided evidence con
cerning state existance at particular points in time. This random
variable (the labor efficiency variance) had certain statistical
properties within the states of nature: the decision maker either
had or did not have knowledge of these particular properties. The
decision task was complicated by the existence of unequal values (costs)
associated with the various possible decision outcomes.
The form of the general heuristic proposed to describe decision
making behavior was that of anchoring and adjustment. Anchoring and
adjustment first requires the selection of an initial decision value
(anchor) from the continuum of possible decision values that are
located within the domain of the random variable (the labor efficiency
variance in the case of the present study). Once an initial decision
anchor is selected by the decision maker, training and experience
(learning) will lead to adjustments in the value of the anchor. Such
adjustments will be motivated primarily by feedback of various decision
performance measures. Should a particular decision performance measure
indicate less than satisfactory decision performance, the decision
maker will be motivated to adjust his decision anchor in an attempt to
improve decision performance measures in the future. The concept of
less than satisfactory decision performance assumes that the decision
maker has internalized the objective function from which the decision
performance measure was derived. Within this research the decision
performance objective function had a homomorphic relation with the
decision maker's payoff function.


APPENDIX F
SUBJECT MOTIVATION QUESTIONNAIRE
Please respond to the following questions by circling the appropriate
response. This is not a test; there are no right or wrong answers. It
is important that you try to answer each question honestly.
1. Suppose this same experiment was to be repeated on a number of future
occasions. How many more times would you be prepared to participate
in the experiment under circumstances similar to the present?
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk

*
*
* *
*
No 1
2
3
4 5
More
More More
More
More
More More
Than 5
Times Time
Ti mes
Times
Times Times
Times
Did your desire to perform well
in undertaking a challenging
task
cause you to try
very hard?
******************************* ******************************


k
* *
*
Definitely Yes
Probably
Don t
Probably No
Definitely
Yes
Yes
Know
No
No
Did you enjoy making the decisions requi
red in the experiment?
***************************************** ********************
* *
k
*
* *
*
Definitely No
Probably
Don' t
Probably Yes
Definitely
No
No
Know
Yes
Yes
Are you satisfied
with your performance
in the experiment?
*************************************************************
k k
*
*
k k
*
Definitely Yes
Probably
Don't
Probably No
Definitely
Yes
Yes
Know
No
No
Do you think you
would have tried less hard if the money rewards had
been halved?
*************************************************************
* *
*
*
* *
*
Definitely No
Probably
Don't
Probably Yes
Definitely
No
No
Know
Yes
Yes
186


32
TABLE 1
GENERAL RESEARCH DESIGN IN TERMS OF EXPERIMENTAL VARIABLES
Dependent Variables
Independent Variables
Individual
Long-Run
Decision
Efficiency
Individual
Information
Processing and
Learning
Efficiency
Contents of the available infor
mation set
X
X
Structure of the decision situation:
Statistical relationships
X
X
Decision outcome relationships
X
X
Individual information processing
and learning efficiency
X
Note: An "X" within a cell indicates that the relationship of the
variables concerned are included within the experimental design.


33
operationalizations of these concepts are discussed in a later section
of this chapter. The objective at this point is to summarize in general
terms the relationships which will be included within the experimental
design.
General Research Methodology
Research methodology pertains to the general procedures or
methods employed in conducting research. Two general methods are
employed in this research-- simulation and laboratory experimentation.
The major objective of the simulation is to produce hypotheses which
will predict the behavior of human decision makers within the decision
situation assumed by the simulation. The simulation is based upon
assumptions derived from the conceptual development and from the gen
eral task environment of the laboratory experimentation. These
assumptions and the task environment are discussed in greater detail
in a later section of this chapter. The major objective of the
laboratory experiment is to test the conceptual development (through
the hypotheses derived by the simulation) of the effects of the vari
ous independent variables on the various observation variables.
The general and specific research design employed in the labora
tory experiment is the same as that employed within the simulation.
The major difference between the simulation and the laboratory experi
ment is the use of human subjects. This difference creates two major
sources of incompatability betv/een the simulation and the laboratory
experiment. First, the simulation makes certain assumptions concerning
human behavior and applies these assumptions consistently. Within


34
the laboratory experiment such behavior is not necessarily applied
consistently. Consequently, greater variance of results is expected
within the laboratory experiment than within the simulation. Second,
it should be noted that certain variables are largely affected by
individual attributes. Without a theory of the effect of individual
attributes any simulation of these variables would be arbitrary.
Accordingly, not all of the experimental variables are included in the
simulation. Hypotheses- concerning the variables not present within
the simulation either are derived from limited concepts concerning the
effects of individual attributes or are stated in an exploratory manner
(no expected difference).
General Experimental Environment and Task
The standard cost variance investigation situation studied within
this research is set in an environment of a manufacturing company.
More specifically, the subjects are asked to assume the role of the
operational manager of an assembly department which assembles a single
product, a metal folding chair. The operating efficiency of the
assembly department is determined completely by the labor efficiency of
the assembly workers.
Each subject receives a sequential series of standard cost vari
ance reports and is asked for each report to decide whether to investi
gate or not to investigate the reported labor efficiency variance.
The efficiency of a subject's decision performance and the amount of
payment he will receive for participating in the experiment is based
upon the total investigation decisions cost which he incurs over the
series of variance reports.


40
Based upon a subjective analysis of each individual's decisions,
the measure DN^ can be adjusted for the effects of using multiple cut
off values. Defining d^a to be the individual's decision model sensi
tivity with the effects of multiple cutoff values eliminated:
DNAi = d\a / d'k
The difference DNA-- DN- approaches zero as the effects of multiple
cutoff values decreases and becomes zero when the individual uses a
single cutoff value.
Individual decision criteria variables
The criteria the individual adopts in making his decisions are
measured using the TSD parameter 3. The theoretical definition of 3 is:
3 = d>(z2) / where ( ) denotes the normal density function for the point in the
parentheses and z. is defined the same as for the d'. variable.
J
An empirical estimate for the 3 of an individual, labeled 3.5
is obtained using the z1 and z2 values associated with the individual's
conditional probabilities employed in estimating d^.. The measure 3..
is relative to the decision situation. In particular, it is relative
to those variables which affect the point on the decision axis which
the individual selects as his investigation decision cutoff value. A
measure of individual criteria comparable across situations is defined
as:
BN. = 3. / 3.
i i k
where 3^ is generated using optimal model k. The measure approaches a
value of one as 3_. approaches 3^ (i.e., as the individual's cutoff
decision value approaches the optimal model's cutoff decision value).


HUMAN INFORMATION PROCESSING FOR DECISIONS
TO INVESTIGATE COST VARIANCES
By
CLIFTON E. BROWN
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1978

Copyright 1978
By
Clifton E.
Brown

ACKNOWLEDGEMENTS
I am grateful to my supervisory committee; Rashad Abdel-khalik,
Douglas Snowball, Gary Hoi strum, and Richard Griggs. Without their
valuable contributions this dissertation would have had far more
obscurities.
I am indebted to my experimental assistants: Doug Snowball,
Gary Hoi strum, Ron Teichman, Ed Bailey, John Wragge, Robert Thompson,
Nancy Hetsko, Andy Judd, and Michael Gift (even though he was not on
time). Without their help the experimental procedures used within
this study would not have been possible.
Finally, my greatest debt is to my wife, Sandy. Without her
encouragement and support this dissertation never would have been done.

TABLE OF CONTENTS
ACKNOWLEDGEMENTS iii
ABSTRACT viii
CHAPTER I. INTRODUCTION AND OBJECTIVES 1
Variance Investigation Decision Processes 1
The Research Objectives A
Dissertation Organization 6
CHAPTER II. AREA OF INVESTIGATION 8
Managerial Accounting Concepts 8
Managerial Decisions 8
Management Task Planning and Control 9
Standard Cost Variance Investigation 11
Psychological Concepts 16
Psychophysics 16
Human Information Processing 20
CHAPTER III. RESEARCH METHODOLOGY AND DESIGN 23
General Conceptual Development 23
Decision Situation Structure 23
Available Information Set 25
Individual Information Processing Efficiency 26
Individual Ability to Expand the Information Set 27
General Research Design 28
Selection of Independent Variables 29

General Research Methodology 33
General Experimental Environment and Task 34
Operationalization of Variables 35
Hypothesis Formation 46
Simulation of Optimal Model Performances 46
Simulation of Investigation Decision Performances 49
CHAPTER IV. THE EXPERIMENT 76
Experimental Environment 76
Subjects 79
Experimental Materials 80
Background Information 80
Variance Investigation Decisions 80
Elicitation of Heuristics 83
%
Elicitation of Subject Motivations . . 83
Experimental Procedures 83
Assignment of Subjects to Treatment Conditions 84
Training Phase 85
Experimental Phase 86
Final Debriefing 87
CHAPTER V. ANALYSIS AND RESULTS 88
Summary of Results 88
General Method of Analysis 88
Training Phase Analysis and Results 98
Initial Decision Anchors 98
Decision Anchor Adjustment 102
Individual Decision Model Sensitivity 106
The dj Dependent Variable 106
v

Relative Decision Model Sensitivity 107
Intrinsic Motivation and Decision Model Sensitivity 114
Individual Decision Criteria 116
Relative Decision Criteria 116
Relative Decision Criteria Conservatism 121
Individual Attributes and Decision Criteria 124
Individual Long-Run Decision Efficiency 126
Relative Decision Costs 126
Individual Attributes and Long-Run Decision Efficiency ... 138
Long-Run Decision Efficiency and Training Phase Performance 140
Individual Attributes 141
Subject Grade Poir.t Average 141
Subject Motivations 142
CHAPTER VI. DISCUSSION OF RESULTS 146
Conceptual Development Revi si ted 146
Overall Results 148
Discussion of Results 150
Decision Anchor Selection and Adjustment 151
Individual Decision Model Sensitivity 153
Individual Decision Criteria 156
Individual Long-Run Decision Efficiency 159
Limitations 163
Implications for Accounting 166
Value of Additional Information 166
General Standard Setting Process 167
Future Research 168
APPENDIX A. THEORY OF SIGNAL DETECTION DERIVATIONS 172
vi

APPENDIX B. ORAL PRESENTATION TO ELICITE VOLUNTEER SUBJECTS .... 174
APPENDIX C. BACKGROUND INFORMATION BOOKLET 176
APPENDIX D. PRIOR INFORMATION SHEETS 182
APPENDIX E. SUBJECT HEURISTIC ELICITATION QUESTIONNAIRE 184
APPENDIX F. SUBJECT MOTIVATION QUESTIONNAIRE 186
BIBLIOGRAPHY 188
BIOGRAPHICAL SKETCH 193

Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
HUMAN INFORMATION PROCESSING FOR DECISIONS
TO INVESTIGATE COST VARIANCES
By
Clifton E. Brown
December 1978
Chairman: A. Rashad Abdel-khalik
Major Department: Accounting
Two major functions of management are those of planning and con
trol. Broadly defined, control is the management process which assures
that selected alternatives are implemented and executed in accordance
with plans. Important aspects of the control process concern analysis
and investigation of standard cost variances provided within accounting
reports. A substantial portion of the accounting variance investigation
literature has employed the normative model approach-- researchers have
created variance investigation models which a manager should use. Rarely
has attention been given to how the manager would interprete and integrate
information required by the various normative models.
The principal focus of this dissertation was on the effects of
specific situational variables upon a manager's information processing
for purposes of making variance investigation decisions. The specific
objectives were 1) tc develop a conceptual framework which will predict
effects cf specific situational variables on a manager's relative effi
ciency in information processing and in variance investigation decision
making, and 2) to empirically test some implications of this conceptual
framework.
viii

Variables that can affect the manager's variance investigation
decision process include 1) structure of the decision situation, 2)
contents of the available information set, 3) manager's information
processing efficiency, and 4) manager's learning efficiency. The
structure of the decision situation depends upon number possible states
of control, relative frequencies of the states, various statistical
relationships among the states, and various relationships between the
decision/state outcomes. The contents of the available information set
include information known by the manager prior to his decision. The
manager's information processing efficiency relates to the particular
strategies employed in combining and weighting various items of infor
mation. The manager's learning efficiency refers to the manager's
ability to learn from his experiences with the controlled process. Both
information processing and learning efficiency are the results of the
manager's particular decision strategies. These decision strategies
were expected to adapt a general form, referred to as anchoring and ad
justment. A natural starting point is used as a first approximation for
the investigation decision rule and is then adjusted as the manager
learns from his experiences.
Two experimental methods were used to derive and test implications
of the conceptual framework. Simulation techniques were used to opera
tionalize hypotheses and a laboratory experiment was used to test these
hypotheses.
The experimental environment was that of an assembly department
within a simulated manufacturing company. The assembly department as
sembled a single product and performance of this department was deter
mined completely by the assembly workers' labor efficiency. Student
IX

subjects, who assumed the role of assembly department operational manager,
made labor efficiency variance investigation decisions based upon a series
of independent standard cost variance reports. The manipulated indepen
dent variables included contents of the available information set,
distributional properties of the states of control, and cost effects of
the investigation decision errors. Parameters of the psychological theory
of signal detection permitted measurement of sensitivity and criteria of
a subject's decision model. The subject's investigation decision costs
compared to normative investigation decision costs derived under similar
situations were employed as a relative measure of decision efficiency.
Overall, the implications of the conceptual framework were supported
by the obtained results. To explain the few major deviations from
expectations, an ex post hypothesis was introduced. This hypothesis
posits that the standard introduces a subjective adjustment bias. Depend
ing upon the direction of adjustment the standard subjectively inhibits
complete adjustment either to optimal decision values close to the
standard or to optimal decision values distant from the standard.
x

CHAPTER I
INTRODUCTION AND OBJECTIVES
Business organizations generate internal accounting reports for
use by managers for purposes of evaluation and decision making.1 One
important evaluation process and decision task of management is that of
standard cost variance analysis and investigation. Based in part upon
standard cost variance reports generated by the internal accounting
system, managers estimate the likelihoods that various production pro
cesses remain under control and decide whether to investigate particu
lar variances. The present study concerns the effects of selected
variables on standard cost variance investigation decision making by
operational control managers. This chapter discusses general aspects
of the standard cost variance investigation decision process and the
variables which may affect the process and presents the research
objectives of the study.
Variance Investigation Decision Processes
An important factor in the evaluation and use of standard cost
variance reports is the manager's perception of the validity of the
xThe American Accounting Association 1966 Statement of Basic Accounting
Theory states that "the objective of accounting for internal use is to
provide information to persons within an organization that enables them
to make informed judgements and effective decisions which further the
organization's goals" (p. 38).
1

2
information provided. In particular, the manager's perception of the
standard setting process will have considerable influence upon his
variance investigation decision process. If the manager believes that
the standards are unrealistic (i.e., that the standards have been placed
too far from the in-control distribution) he may rescale either the
standards or the variances to correspond with his own perception of the
in-control distribution.2
If it is assumed that a manager accepts the standard setting pro
cess as realistic and does not rescale the standards or variances, his
variance investigation decisions may be viewed as the culmination of a
two-stage process. The first stage concerns the detection of the
particular distribution (e.g., in-control or out-of-control) that gen
erated the variance. The manager's performance of this task is a
function of the sensitivity of his decision process (model). This
sensitivity is affected by the structure of the particular situation
and by the manager's knowledge of this structure. The extent of the
manager's knowledge of the situation, in turn, is affected by the
available information (both contained within the variance report and
provided from other sources), and by the manager's ability to learn
from his experiences with the processes being controlled. In situations
where the controlled process distributions have some area of overlap
the results of a manager's detection process are probabilistic (i.e.,
prior to completing an investigation the existence of any given state
is uncertain). Furthermore, the greater the area of overlap, the more
difficult the discrimination task becomes.
2Within this context, an in-control distribution concerns statistical
congruence of production output and planned output in terms of
controllable resource utilization.

3
The second stage concerns the manager's investigation decision
criteria. Having arrived at a conclusion (albeit probabilistic) about
the distribution that generated the variance, the manager must integrate
and process various objective function parameters in order to arrive at
his variance investigation decision (these parameters can belong to
either the manager's objective function, the organization's objective
function, or both if the same). The potential investigation criteria
can be divided into two major categories: structural criteria and
behavioral criteria. The first category can be divided further into
structural cost criteria and structural probability criteria. The
structural cost criteria include such variables as the additional cost
of operating an out-of-control process (given that the process can be
returned to the original in-control state), and the costs of variance
investigation and correction (which can differ depending upon the
actual state that generated the variance). The structural probability
criteria include such variables as the probability that the source of
the variance is controllable (i.e., that it can be returned to the
original in-control state through managerial action), the probability
that the process will return to the in-control state without managerial
action, and the prior probabilities of each state distribution. The
behavioral criteria include such variables as the manager's perception
of the effect of the investigation upon his performance evaluation and
reward structure, and the manager's perception of the effect of the
decision upon employee performances and attitudes.
Although the variance investigation decision process has been
described as two stages, the manager may not actually utilize such a
sequential stage process. The manager's actual decision process is

4
labeled a heuristic. Within this context, heuristic refers to the
learned set of rules or principles that are utilized by the individual
in making the particular decisions required of him. The manager's
specific heuristic can be affected substantially by his individual
characteristics (e.g., intelligence, cognitive complexity and style,
decision process sensitivity, motivations, etc.). However, a general
heuristic, labeled anchoring and adjustment (Tversky and Kahnemar,
1974), is expected to describe the general form of the manager's vari
ance investigation decision process.3
The above examination of the manager's variance investigation
decision process indicates that the variables which can affect both
the process and the results of the process include 1) the structure
of the decision situation, 2) the contents of the available information
set, 3) the manager's information processing efficiency, and 4) the
manager's learning efficiency (from his experiences with the controlled
process). The manager's knowledge concerning the structure of the
decision situation may be limited by the contents of the available
information set and by his information processing and learning effici
encies. The accountant, to a large extent, has control over the con
tents of the available information set.
The Research Objectives
Many problems in accounting reduce to one of choosing among
alternative information sets that could be provided to a decision maker
(American Accounting Association, 1972). Two basic approaches to
3The anchoring and adjustment heuristic is described in greater detail
within Chapter II.

5
deciding on the information set to be presented within the cost variance
report have been advocated. First, the accountant can determine those
models which managers use in making variance investigation decisions and
provide an information set which would permit the implementation of
these models. Problems with this individual model approach are 1) the
possibility exists that different individual models use a wide range of
different information sets (which are not costless), and 2) the indi
vidual models may not be optimal (i.e., they could be inefficient with
respect to other decision models). The individual model approach cen
ters on the psychological question of what the manager is doing (or
would do) with the available (or additional) information. Second, the
accountant could create a normative model of the variance investigation
decision and provide the information required by that model. Problems
with the normative model approach are 1) the information provided may
not optimize the individual's investigation decisions (which may con
tinue to be made using the individual's model), and 2) the costs
associated with operationalization of the normative model may exceed
the benefits (cost savings as a result of more optimal investigation
decisions and greater congruence of manager goals with overall organi
zation goals) offered by the model. The normative model approach
centers on the analytical question of what the manager should be doing
(or should do) with the available (or additional) information.
A substantial portion of the accounting variance investigation
literature has focused upon the second approach-- the normative model
approach.4 However, rarely has attention been given to how the manager
4This literature is reviewed within Chapter II.

6
would interprete and integrate the information sets of the various
normative (optimal) variance investigation models. In many instances
the implicit assumption has been that the manager would process the
information with the same efficiency as the normative model.
The principal focus of this research is the effects of specific
situational variables on a manager's variance investigation decisions
(relative to the investigation decisions of an optimal model) and on
a manager's processing of available information (relative to the infor
mation processing of an optimal model). Specific objectives are:
1) To develop a conceptual framework which will predict the
effects of situational variables on a manager's relative
information processing efficiency and to empirically test
the implications of this conceptual framework.
2) To develop a conceptual framework which will predict the
effects of situational variables on a manager's relative
variance investigation decision efficiency and to
empirically test the implications of this conceptual framework.
Dissertation Organization
Chapter II presents certain general concepts from accounting and
psychology. These concepts are necessary for the development of the
specific environment studied in this research, and for the development
of a conceptual information processing and decision making framework
within this environment. The accounting concepts are concerned with the
nature of managerial decisions, managerial task planning and control,
and standard cost variance investigation. The psychological concepts
are concerned with the theory of signal detection and with cognitive

7
aspects of human information processing such as the general heuristic
of anchoring and adjustment.
Chapter III synthesizes the general concepts of the previous
chapter and develops a conceptual framework of standard cost variance
investigation employing the methodology of the psychological concepts.
This conceptual framework was operationalized using simulation tech
niques, and experimental hypotheses were derived from the simulations.
The details of a laboratory experiment designed to test the
hypotheses derived from the conceptual variance investigation framework
are presented in Chapter IV. The experiment employed a between-subjects
design and was manipulated factorially using specific variance investi
gation situation variables. Modified parameters from the psychological
theory of signal detection were employed to measure subject's information
processing efficiency and decision model sensitivity.
The results obtained from the laboratory experiment are presented
in Chapter V. The model comparison procedure of non-orthogonal analysis
of variance was the primary method of analysis employed. Chapter VI
discusses the results, develops a modification (ex post) of the
original conceptual framework, and discusses some implications for
accounting and for future accounting research.

CHAPTER II
AREA OF INVESTIGATION
Managerial Accounting Concepts
This study relies on the synthesis of certain psychological con
cepts with certain accounting concepts. This section discusses the
accounting concepts: the nature of managerial decisions, managerial
task planning and control, and standard cost variance investigation.
Managerial Decisions
Since the information set (accounting report) exists to support
the manager's tasks, a conceptual framework of managerial decisions would
facilitate the accountant's information set selection. A conceptual
framework of managerial decisions should differentiate managerial de
cisions along dimensions that allow insights into the informational
needs of those decisions.
Anthony (1365) provides one such dimension along which he iden
tifies the purposes or orientation of managerial activities: strategic
planning, management control, and operational control. Another dimen
sion is provided by Simon (1966) who distinguishes between programmed
and nonprogrammed decisions. The underlying dimension is concerned
with the manner in which managers deal with their problems. The cri
teria for classifying a decision consist of the extent of structure
associated with the problem solving phases of the decision.
8

9
A conceptual framework of managerial decisions that synthesizes
these dimensions is presented by Gorry and Morton (1971). Simon's
dimension is modified by replacing the terms "programmed" and "non-
programmed" with the terms "structured" and "unstructured" and adding
a third category labeled semi-structured. Gorry and Morton's conceptual
framework is used in this research as a model that will permit the
identification of a standard cost system within the overall managerial
decision framework.
Management Task Planning and Control
Standard cost systems present information sets to managers to
support various decisions concerning task planning and control.
Demski (1967) provides an excellent conceptual discussion of the manage
ment task planning and control process, and his approach (with some
modification) is adopted in this research.
Figure 1 presents a model of the management task planning and
control process. Environmental information, largely external to the
standard cost system, facilitates the planning of overall goals and
policies within the constraints imposed by the environment. Overall
objective control feedback 1) provides evidence for the ccntinuted
validity of overall assumptions made in forming the organization ob
jectives, and 2) facilitates strategic planning evaluation of organi
zation performance.
Both the environmental variables and the overall task control
feedback of the management control activity 1) provide evidence for
the continued validity of the overall task assumptions made in forming

Overall
Objective
Control
Strategic
Planning
Overall
Objective
Planning
Overal1
Task
Control
Management
Control
Overall and
Specific Task
Planning
Specific
Task
Control
V
Operational
Control
Specific Task
Structuring
Source: Modified from Demski, 1967.
Environment
Output
FIGURE 1
MANAGEMENT TASK PLANNING AND CONTROL PROCESS

11
the task plans, and 2) facilitate management control evaluation of
operational control performance.
The operational control manager uses the specific task control
feedback to 1) decide whether the physical system performance is in
agreement with the specific task plan, and 2) to decide whether the
task can be restructured to bring performance back into agreement with
the plan (if performance and the plan differ). A task that can not be
restructured could have significance for the overall task control feed
back to the management control activity and possibly may lead to
modification of the original task plans. Such modification could in
turn have significance for the overall objective control feedback to
the strategic planning activity, and a modification of the original
overall objectives may follow.
Standard Cost Variance Investigation
Within the operational control activity the specific task plan
provides standards of performance in terms of expected component costs
and usages, and the desired physical outputs. The operational control
manager structures the tasks within the physical system and periodically
receives a standard cost variance report describing the system output
in terms of the task plan and the actual results. Upon receiving a vari
ance report the manager must decide the nature of the given variances.
Dopuch et al. (1967) present a classification of standard cost
variances that is based on the expected source of the variances. A
variance resulting from a random fluctuation of the physical system,
labeled a Type 1 variance, requires no operational control response if

12
not statistically significant. Whether a variance resulting from a
change in the physical system, labeled a Type 2 variance, requires an
operational control response depends on whether the underlying cause of
the change is a temporary rather than a permanent phenomena. A tem
porary or controllable variance, labeled a Type 2a variance, is one
that operational control can correct in the future (i.e., the physical
system can be returned to the previous in-control condition). A per
manent or noncontrollable variance, labeled a Type 2b variance, is one
that operational control can not correct in the future.
If the manager decides that the variance is of Type 1 no action
is required. If, however, the manager decides that the variance is of
Type 2 he must then decide if the variance should be investigated for
its underlying causes. Should the variance be of Type 2b it would
have little operational control significance; should the variance be
of Type 2a the cause of the variance could be eliminated by restructur
ing the physical system. The present study confines itself to the
assumption that all variances which result from a change in the physical
system are controllable by the operational control activity. In other
words, it assumes that the standard cost variances have no significant
affect on either the strategic planning or the management control
activities.
The information set contained in a standard cost variance report
can be viewed as relating to one of two categories of information:
distributional properties of system states and investigation decision
criteria. The distributional properties category can include such
variables as 1) the mean and standard deviation of the output when the
system is known to be in-control, 2) the mean and standard deviation of

13
the output when the system is known to be out-of-control, 3) the actual
output for the period, and 4) the deviation between the standard output
and the actual output for the period (the standard output variance).
The investigation decision criteria category can include such variables
as 1) the prior probabilities of the system's states, and 2) the
costs and benefits associated with each possible investigation decision
in combination with each possible system state.
Information relating to investigation decision criteria that is
not typically presented within the variance report includes 1) the
probability that the cause of the variance is controllable (assumed by
this research to equal one), 2) the probability that the system will
return to the in-control state without managerial action (assumed by
this research to equal zero), 3) the perceived effect of the investi
gation decision on manager performance evaluation, and 4) the perceived
effect of the investigation on employee performances and attitudes.
The variance investigation literature is concerned mainly with
the investigation significance of Type 1 and Type 2a variances. A
general approach described within the literature is that of the Shewhart
X chart procedure (Probst, 1971; Koehler, 1968; Luh, 1968; Jeurs, 1967;
Zannetos, 1964). This approach, based upon classical statistics,
involves sampling the system to construct an in-control mean and stan
dard deviation. Arbitrary control limits (generally plus or minus three
standard deviations) are used as the criteria for making a variance
investigation decision. Only one of the above studies considered the
effects of the information provided on the manager's investigation
decision. Probst (1971), in an industrial field study, found foremen
unwilling to accept the procedure as their investigation decision rule

14
when the procedure was constructed objectively (the foremen indicated
the reason for not accepting the procedure was that they felt it
ignored their experience). However, Probst did not analyze his results
in terms of the performance efficiency of the foremen.
Other research has noted what are considered to be significant
drawbacks to the Shewhart X chart procedure: the failure to consider
the costs and benefits associated with the various investigation de
cision outcomes and the failure to consider information from prior
periods. Bierman et al. (1961) proposed the incorporation of investi
gation decision costs and benefits within a classical statistics frame
work. Thus managers may use statistical information to calculate the
probability that the variance reflects the in-control state, and com
bine this probability with the investigation decision costs and benefits.
Most of this information is assumed to be provided by the accountant
and accurately processed by the manager.
Kaplan (1975) and Jacobs (1978) have considered another deficiency
of the classical statistics approach-- the failure to use prior period
information. They analyze the use of the cummulative sum procedure,
whereby the cumulative sum of the variance is charted for each period.
Theoretically, under a stable state these sums should follow a random
walk, and any drift would indicate the system is out-of-control. Eval
uation of such a drift would be accomplished on the basis of information
provided by the accountant or derived from the manager's experience.
The economic cumulative sum procedure incorporates the effects of
estimated investigation decision costs within the drift evaluation.
Several studies in the variance investigation literature use
decision theory to construct a variance investigation model (Kaplan,

15
1969; Dyckman, 1969; Kaplan, 1975; Dittman and Prakash, 1978). If
all the parameters required by these models were available these
models could replace the manager as the variance investigation decision
maker. However, to the extent that not all the parameters are opera-
tionalizable (given seme cost constraint), the manager must continue
to make the variance investigation decisions using the information
provided by the accountant and by his own experiences.
A conclusion that can be drawn from the variance investigation
literature is that the additional information proposed for the use of
managers is becoming both diverse and complex. Furthermore, there is
a paucity of research relating to the manager's ability and efficiency
to interpret and integrate this additional information. Demski (1970,
1971) was one of the first to note the problem associated with the
decision-implementation interface (i.e., the effect of the control
system information on individual behavior and performance). Some
analytic research on this problem has followed. Using simulation
techniques, Magee (1976) compared average total cost for seven decision
rules under various operational conditions. Although the simpler
decision models (rules) tended to have larger average costs than the
more sophisticated decision models, the difference in average cost was
relatively small. Indeed, if model implementation and information costs
are considered the simpler models may be more efficient than the
sophisticated models. Magee also noted that because of the use of
different manager performance measures (average operating costs, average
number of months below standard, etc.) simpler decision models may re
sult in rational choices (i.e., choices that maximize the manager's
expected utility). Closely related to Magee's research is an empirical

16
study by Magee and Dickhaut (1977). Using human subjects, Magee and
Dickhaut found support for the proposition that different manager per
formance (payoff) measures will affect the specific decision rules
(heuristics) employed by the decision maker (as a result of the decision
maker attempting to maximize his subjective utility).
Psychological Concepts
This section discusses the psychological concepts employed in
developing a conceptual framework of manager standard cost variance
investigation decisions. These concepts include the psychophysical
theory of signal detection and the human information processing con
cepts of decision heuristics.
Psychophysics
Psychophysics studies the relationships between physical and
psychological scales of measurement. Modern psychophysics adopts the
view that subjects can make meaningful evaluations of the magnitudes of
their sensory experiences, and therefore sensory magnitudes, as well
as physical magnitudes, can be quantified. One approach of modern
psychophysics is based upon the theory of signal detection (TSD). TSD
permits the separation of the decision maker's ability to discriminate
between classes of stimuli (sensitivity) from his motivational response
biases (decision criteria). Two comprehensive theoretical descriptions
of TSD are presented by Green and Swets (1974) and Egan (1975); general
surveys of the TSD theory are presented by Coombs et al. (1970), Watson
(1973), and Pastore and Scheirer (1974).

17
The basic TSD experiment utilizes the single-interval procedure.
This procedure consists of a series of trials, each trial consisting of
an observation interval and a response interval. The possible stimulus
events during the observation interval are 1) the observation contains
a meaningful signal added to a background of noise (sn trial), and 2)
the observation contains only a background of noise (n trial). It is
assumed that each trial is independent of all other trials and that
the prior probabilities of n and sn are given and remain constant. The
background noise fluctuates at random from trial to trial; the stimulus
(usually a fixed level) is added to the noise. Therefore, the obser
vation fluctuates randomly from trial to trial. The task of the subject
is to detect whether the observation was generated by the signal plus
noise (sn) or by the noise alone (n) distribution. That is, in the
response interval the subject will respond with either "Yes, the signal
was present" (Y response), or "No, the signal was not present" (N re
sponse).
On any trial there exist four possible outcomes of the subject's
decision in conjunction with the actual distribution: 1) sn was pre
sented and the subject said "Yes" (a hit), 2) sn was presented and the
subject said "No" (a miss), 3) n was presented and the subject said
"Yes" (a false alarm), and 4) n was presented and the subject said
"No" (a correct rejection). A conditional probability matrix for a
series of these events is given by the following (Green and Swets, 1974):
RESPONSE
Yes No
^ sn P(Y|sn) P(N|sn)
rs
~ n P(Y|n) P(N|n)
UO

18
Since the cells of this matrix are both exhaustive and mutually exclu
sive, the row-wise conditional probabilities must sum to one (this does
not necessarily hold for the sum of the column-wise conditional
probabilities). All parameters of the TSD model are derived from this
conditional probability matrix (see Appendix A for a more detailed
discussion of the conditional probability matrix including its rela
tionship with Bayes1 theorem).
In the single-interval task the subject analyzes the evidence
and classifies the stimulus into one of two categories according to
his criteria. The criteria are determined by his objective function.
The objective funcition of interest within this research is the
maximization of expected value. Assume that the subject has some value
(utility) fqr each of the four event outcomes. A payoff matrix of these
values related to the four outcomes is. given by the following (Egan,
1975):
DECISION
No
Yes
sn ^sn,Y ^sn,N
n vn,Y vn,N
The decision rule for maximizing the expected value (see Appendix A
for a derivation of this decision rule) is:
P(x|sn) P(n) Vn,W VV
P(xjn) P(sn) V$n,Y ^sn,N
At the point at which the above expression is an equality the subject
should be indifferent between saying "Yes" or saying "No." This point
can be considered the critical value of the likelihood ratio of the
observations, L(x0). The decision rule for saying "Yes" is expressed
by the relation L(x) > L(x0).

19
The critical value L(xQ) for this decision rule has two possible
values: a theoretical value which is a measure of the criteria of an
optimal (or ideal) subject and a subjective value which is a measure
of the criteria of an actual subject.
One set of TSD models assumes that both conditional probability
distributions are Gaussian; i.e., $(x|n) and $(x|sn) are normally
distributed. With the additional assumption of equal variance for both
distributions, the parameters which measure individual discrimination
sensitivity and individual decision criteria are labeled d1 and g,
respectively. The discriminability measure, d', has the following
theoretical definition:
where, ysn= the mean of the signal plus noise distribution;
vn = the mean of the noise alone distribution;
a = the standard deviation of both distributions;
zn = the value of the normal distribution function associated
with the noise alone distribution and any decision axis
cutoff value common to both distributions; and
z = the value of the normal distribution function associated
sn
with the signal plus noise distribution and any decision
axis cutoff value common to both distributions.
The d' measure theoretically is independent of the decision criteria
measure.
The decision criteria measure, B, has the following theoretical
definition:
B = i>(zsn) / i>(zn)

20
where <¡>( ) denotes the normal density function for the point in
parentheses and the z parameters are the same as defined for the d'
measure.
Although the assumption of equal variance normal distributions
is employed within this research, such an assumption is not necessary
to employ TSD. Egan (1975) demonstrates the use of TSD with exponential
distributions, chi-square distributions, Bernoulli distributions, and
Poisson distributions. Grier (1971) develops nonparametric measures
of discriminability and decision criteria.
Traditionally, psychophysics has employed TSD to study perceptual
processes: i.e., sensory processes such as audition and vision. Over
the last decade, however, TSD has been applied to conceptual processes.
These extensions to conceptual processes have included numerical
processing (Lieblich and Lieblich, 1969; Hammerton, 1970; Weissmar. et al.,
1975), medical diagnosis (Lusted, 1969; Lusted, 1971; Swets, 1972),
conceptual judgement (Ulehla et al., 1967a; Ulehla et al., 1967b), and
memory (Bernbach, 1967; Banks, 1970).
Human Information Processing
Human information processing (HIP), a subset of cognitive psy
chology, studies human judgement and decision making with particular
emphasis on the processing of information that determines these
activities. An area within HIP is the construction of models of human
decision making. The work within this area is typically classified
into schools of research which employ different paradigms, the two
major paradigms being the Bayesian and regression approaches (Slovic
and Lichtenstein, 1971). The TSD model is related to the Bayesian approach.

21
The Bayesian approach is a normative model specifying how a
decision should be made given certain internally consistent relation
ships among probabilistic beliefs. The basic beliefs of this approach
are that decisions should be based on subjective probabilities and
that these probabilities should be revised upon the receipt of addi
tional information in accordance with Bayes's theorem.
The major findings of Bayesian research are labeled conservatism.
The subjects, after receiving additional information, revise their
posterior probabilities in the same direction as the optimal model but
the revision is insufficient. Much of the research has focused on an
explanation of the cause of conservatism, the major explanations being
misperception (Peterson et al., 1968), misaggregation (DuCharme and
Peterson, 1968), and response bias (DuCharme, 1970).
Another area within HIP is the study of subjective information
processing principles and decision rules, labeled heuristics. A
heuristic, within this context, refers to a learned set of rules or
principles which are utilized by individuals in making the particular
decisions required of them. Research in this area has been concerned
with identifying systematic biases (relative to some definition of
optimal) of subjective heuristics within certain types of decision
tasks. Those information processing and decision rule biases iden
tified thus far have been labeled as an anchoring and adjustment
heuristic (Tversky and Kahneman, 1973), a representative heuristic
(Kahneman and Tversky, 1972; Swieringa et al., 1976), an avail-
ablility heuristic (Tversky and Kahneman, 1973), and the law of small
numbers (Tversky and Kahneman, 1971).

22
This study will make particular use of the anchoring and
adjustment heuristic. In many situations, individuals first make
decisions by starting with an initial anchor (decision point) and then
adjust this initial anchor as they learn from their experiences. The
initial anchor can be suggested by the structure of the decision
situation, or can be the result of a partial computation or estimate.
Empirical tests involving the anchoring and adjustment heuristic in
dicate individuals do not sufficiently adjust their initial decision
point. That is, their adjustment is less than that which would allow
optimal processing of the available information (Slovic and Lichtenstein,
1971; Slovic, 1972; Alpert and Raiffa, 1968; Tversky and Kahneman,
1974).

CHAPTER III
RESEARCH METHODOLOGY AND DESIGN
General Conceptual Development
A general problem confronting decision makers is that the indi
vidual must decide subjectively which state of nature is most probable
based upon some incomplete set of information. When an individual
deals repetitively with a similar situation his long-run decision
efficiency or "decision correctness" can be affected by 1) the struc
ture of the particular decision situation, 2) the contents of the set
of available information, 3) his efficiency in processing the available
information, and 4) his ability to expand the available information
through experience with, and observations of, the various states of
nature. Each of these four elements is discussed below with specific
reference to the objectives of this research.
Decision Situation Structure
The structure of the particular decision situation primarily de
pends upon several key variables. These variables include the number of
possible states of nature, the relative frequencies of the states, the
various statistical relationships among the states, and the relation
ships between the various decision outcomes (the costs incurred given a
specific decision and the existence of a specific state). Depending on
23

24
the specific values which these variables may assume, the structure of
the particular decision situation can affect both the difficulty and
the importance of the individual's discrimination among the states. In
general, as the number of possible states increases and as the area of
distributional overlap of these states increases the discrimination task
becomes more difficult. Furthermore, the discrimination task becomes
more important (in terms of incurred costs) as the relative frequencies
of the states become equal and as the costs associated with the possible
decision outcomes which involve decision errors (an incorrect decision
given the existence of a specific state) become unequal.
The general situation employed in this research, selected for
its relevance to the accounting discipline, is the cost variance
investigation decision. In this situation two states of nature are
possible: 1) in-control (the underlying physical process described by
the standard cost variance report is functioning as planned), and 2)
out-of-control (the underlying physical process described by the stan
dard cost variance report is not functioning as planned). In reality,
the possible states of nature may be located on a continuum whose end
points are the states of in-control and out-of-control. Between these
two end points are any number of states that take the general form of
partially out-of-control or moving out-of-control. In the present
research the possible states of nature are confined to the two end
points. The relative frequencies of the two states are controlled as
constants with their values being close to equal. The statistical
relationships among the two states and the relationships between the
various decision outcomes are manipulated as independent variables.

25
Available Information Set
The contents of the available information set refers to the infor
mation known by the individual prior to his decision. Such information
can be of two types-- singular or distributional (Tversky and Kahneman,
1977). Singular information consists of information that specifically
relates to the current decision. Distributional information consists
of information that relates to the relative frequencies and to the
statistical relationships among the states of nature.1 The difficulty
of the discrimination task may be affected by the presence or absence
of certain items of information. In general, the less information con
tained in the available set the more difficult the discrimination task,
for missing information required by a decision model must be estimated
by the individual. Compared to statistically derived estimates, these
subjective estimates are likely to have greater uncertainty and in
efficiency associated with them.
Within the variance investigation situation, the information con
tained on each variance report constitutes the singular information.
Two types of singular information are employed in this research-- the
actual results of the physical process and its variance from a standard,
and the marginal costs associated with each of the two possible decisions
in combination with each of the two possible states of nature. The
presence of both these singular information types is controlled as a
constant within this research. The presence or absence of specific
xThis definition of distributional information implicitly assumes that
the relative frequencies and statistical relationships are stable over
the relevant time frame. If these variables were non-stable, revised
estimates of their specific values would be required prior to each
decision, thus classifying them as singular information. This research
assumes that both of these variables are stable across all decisions.

26
distributional information items (the statistical means, variances, and
distributional shapes of the two states) is manipulated as an indepen
dent variable. The presence of other distributional information items
(the relative frequencies and the allowed standard) is controlled as a
constant.
Individual Information Processing Efficiency
The individual's efficiency in processing the available infor
mation relates to the particular heuristics or strategies employed in
combining and weighting the various items of information. The term
efficiency implies a relationship between the individual's process
output (his decision) and a normatively correct or optimal decision.
The individual's decision and information processing performance can be
evaluated by comparing his performance against an optimal model.
Optimality refers to the best possible performance under given condi
tions. Since the optimal decision model relies on an incomplete infor
mation set rather than certain knowledge,even its performance can be
affected by both the structure of the situation and the contents of the
available information set.
The optimal models used in this research are a function of the
experimental environment: the single-interval procedure found within
the psychological theory of signal detection and the basic decision
situation of standard variance investigation. Within this environment
the optimal decision rule for minimizing (maximizing) the expected cost
(value) of a set of decisions is based upon an extension of Bayes'
theorem that takes into account the relative costs (values) of various

27
possible decision outcomes. The parameters required to fit the optimal
model include 1) the relative frequencies of the two states of nature,
2) the mean of each state, 3) the statistical variance of each state,
and 4) the costs associated with each of the two possible decisions in
combination with each of the two possible states of nature. Since the
individual's decisions are to be evaluated by comparisons with the out
puts of the optimal model, it would seem reasonable that the information
available to the optimal model be the same as the information available
to the individual. Consequently, for decision situations in which some
of the parameters required by the optimal model are not contained in the
available information set the optimal model must make estimates of the
missing parameters.
Individual Ability to Expand the Information Set
The individual's ability to expand the available information set
over time refers to his ability to learn from his experiences with the
states of nature. Such learning can occur through improved estimates
of unknown items of distributional information and through modifications
of information processing strategies to incorporate state relationships
which were unknown or undetected previously.
The general form of the expected individual information processing
strategy is described by the heuristic of anchoring and adjustment
(Tversky and Kahneman, 1974).2 A natural starting point, or anchor, is
used as the first approximation for the decision. This anchor is then
2A heuristic, within this context, refers to a learned set of rules or
principles which are utilized by an individual in making the particular
decisions required of him.

28
adjusted as the individual learns from his experiences with the states
of nature.
Within the present study the initial decision points or anchors
were expected to fall near the geometric intersection (either actual or
estimated, depending upon the available information) of the distribu
tional curves of the two states. Adjustments from these initial deci
sion points by the individual were expected to occur during the train
ing phase of the experiment as the individual gained experience with
the states and received feedback as to his performance (relative to an
optimal model). The extent of an individual's adjustments (his learn
ing efficiency) is measured using several variables. Each variable
measures the relative extent of adjustment (or lack of adjustment) from
the original decision anchor tov/ard the optimal value of that variable.
General Research Design
As stated previously, the focus of this research is on human
decision making and information processing within a particular decision
context-- that of standard cost variance investigation. Much of the
variance investigation literature within accounting has focused on how
the decision maker should integrate the available information and make
the investigation decision (see Kaplan, 1975 for a review of this
literature). Very little research has been concerned with how the
decision maker does accomplish these processes. The major objective
of this research is to study the processes used by the decision maker
in reaching variance investigation decisions. In particular, this re
search will examine within the conceptual framework discussed previously:

29
1) the effects of the decision situation structure, the available infor
mation set contents, the individual's information processing efficiency,
and the individual's learning efficiency on the individual's long-run
decision efficiency; and 2) the effects of the decision situation
structure and the available information set contents on the individual's
information processing and learning efficiency.
Selection of Independent Variables
The effects of two types of variables are of primary interest
within this study: these are the situation variables and the process
variables. The following discussion describes the selection of each
of the variables employed in this study.
The situation variables
The situation variables are the quantity of information available
to the individual prior to his decision, the statistical structure of
the two states of nature, and the cost structure of the possible decision
outcomes.
The available information set. The effects of the contents of
the available information set are studied by manipulating the presence
and absence of certain distributional information. This involves the
specification of two levels of available information set content. Since
the difficultly of the discrimination task may be increased by the
absence of certain information, one level of the information variable
has less distributional information than the other level. This inde
pendent variable is labeled the information variable.

30
The statistical structure. The effects of the statistical struc
ture of the decision situation are studied by manipulating a distribu
tional information variable. This involves the specification of two
levels of statistical relationship among the two states. Since the
difficulty of the discrimination task increases as the area of dis
tributional overlap between the two states increases, one level of the
distribution variable will have a greater area of overlap than will
the other level. This independent variable is labeled the distribution
variable. All other distributional information items, if presented,
will be controlled as constants.
The cost structure. The effects of the cost structure of the
possible decision outcomes are studied by manipulating a singular in
formation variable. The manipulation of this variable involves the
specification of two levels of decision outcome relationships (in terms
of incurred costs). This independent variable is labeled the cost
variable. Since the importance of the discrimination task increases as
the costs associated with decision outcomes that involve decision errors
become unequal, the different levels of the cost variable will be
associated with different decision error costs. One level of the cost
variable is structured in favor of more variance investigations and the
other level of the cost variable is structured in favor of fewer vari
ance investigations. The only other singular information variable, the
actual results of the physical process and its cost variance, is the
primary experimental stimulus. This information item will be a random
variable whose distributions, given either of the two states of nature,
will be normally shaped with parameter values defined by the appropriate
level of the distribution variable.

31
Independent variables interaction. The research design of this
study is a factorial design in which the above three independent vari
ables, each at two levels, are fully crossed, thus producing eight (23)
independent variables combinations or treatments. The factorial
combination of the different levels of the independent variables adds
power to the research design by permitting the examination of the
effects of interactions upon the dependent variables.
The process variables
The effects of an individual's information processing and learn
ing efficiency are studied using certain observation variables measured
on a continuous (ratio) scale. Continuous measurement facilitates the
use of these variables as both independent and dependent variables.
When used as independent variables the measures are treated as random
components rather than as discrete classification levels. The effects
are studied using three major types of variables: 1) individual de
cision model sensitivity (relative effect of a variable response range),
2) individual decision criteria (relative effect of conservative ad
justment), and 3) relative initial decision anchor and final decision
anchor relationships. The first two types of variables are derived
from model parameters within the theory of signal detection.
A summary of the general research design
The general research design is depicted in Table 1. The design
is presented in terms of the dependent and independent variables to be
included within the research. More specific extensions arid

32
TABLE 1
GENERAL RESEARCH DESIGN IN TERMS OF EXPERIMENTAL VARIABLES
Dependent Variables
Independent Variables
Individual
Long-Run
Decision
Efficiency
Individual
Information
Processing and
Learning
Efficiency
Contents of the available infor
mation set
X
X
Structure of the decision situation:
Statistical relationships
X
X
Decision outcome relationships
X
X
Individual information processing
and learning efficiency
X
Note: An "X" within a cell indicates that the relationship of the
variables concerned are included within the experimental design.

33
operationalizations of these concepts are discussed in a later section
of this chapter. The objective at this point is to summarize in general
terms the relationships which will be included within the experimental
design.
General Research Methodology
Research methodology pertains to the general procedures or
methods employed in conducting research. Two general methods are
employed in this research-- simulation and laboratory experimentation.
The major objective of the simulation is to produce hypotheses which
will predict the behavior of human decision makers within the decision
situation assumed by the simulation. The simulation is based upon
assumptions derived from the conceptual development and from the gen
eral task environment of the laboratory experimentation. These
assumptions and the task environment are discussed in greater detail
in a later section of this chapter. The major objective of the
laboratory experiment is to test the conceptual development (through
the hypotheses derived by the simulation) of the effects of the vari
ous independent variables on the various observation variables.
The general and specific research design employed in the labora
tory experiment is the same as that employed within the simulation.
The major difference between the simulation and the laboratory experi
ment is the use of human subjects. This difference creates two major
sources of incompatability betv/een the simulation and the laboratory
experiment. First, the simulation makes certain assumptions concerning
human behavior and applies these assumptions consistently. Within

34
the laboratory experiment such behavior is not necessarily applied
consistently. Consequently, greater variance of results is expected
within the laboratory experiment than within the simulation. Second,
it should be noted that certain variables are largely affected by
individual attributes. Without a theory of the effect of individual
attributes any simulation of these variables would be arbitrary.
Accordingly, not all of the experimental variables are included in the
simulation. Hypotheses- concerning the variables not present within
the simulation either are derived from limited concepts concerning the
effects of individual attributes or are stated in an exploratory manner
(no expected difference).
General Experimental Environment and Task
The standard cost variance investigation situation studied within
this research is set in an environment of a manufacturing company.
More specifically, the subjects are asked to assume the role of the
operational manager of an assembly department which assembles a single
product, a metal folding chair. The operating efficiency of the
assembly department is determined completely by the labor efficiency of
the assembly workers.
Each subject receives a sequential series of standard cost vari
ance reports and is asked for each report to decide whether to investi
gate or not to investigate the reported labor efficiency variance.
The efficiency of a subject's decision performance and the amount of
payment he will receive for participating in the experiment is based
upon the total investigation decisions cost which he incurs over the
series of variance reports.

35
Each standard variance report is concerned with the results of a
single job-order to produce a constant number of chairs and reports
only aggregate (overall assembly department) results. Each report
contains the aggregate standard assembly time allowed per chair, the
actual assembly time incurred per chair, the overall labor efficiency
variance per chair, the total number of chairs produced, and the costs
associated with each possible decision in combination with each
possible state of nature. All time units are presented in minutes.
The singular information contained in a variance report is independent
of that contained in previous variance reports.
The experiment is conducted in two phases-- a training phase
and an experimental phase. The training phase consists of three
contiguous sessions in which the subject learns his role and presumably
develops his decision strategy. Performance feedback is given at the #
completion of each training session. The experimental phase consists
of a single session in which the subject receives a series of variance
reports similar to those presented in his training session. In this
phase no performance feedback is given until after the completion of
the entire experiment. The subject is paid according to his performance
in the experimental phase.3
Operationalization of Variables
This research employs the following three decision situation
independent variables, each measured using a discrete classification:
3Greater detail concerning the experimental environment, task, and
procedures is presented in Chapter V.

36
1) the information variable, 2) the distribution variable, and 3) the
cost variable. As indicated previously, each is varied across two levels.
The individual process variable types, measured on a continuous scale,
include the following: 1) individual decision model sensitivity, 2)
individual decision criteria, and 3) relative initial and final decision
anchor relationships. The major dependent variable is the individual
long-run decision efficiency (in terms of incurred costs). The
following discussion describes each of the variables or variable types
as they are employed in this study.
The information variable
The first level, labeled Jl_, is derived from the set of infor
mation assumed to come from individual experience with the physical
system. It includes the following items: 1) the historically derived
portion of time in which the process has been found to fall in each of
the two states, 2) the assumption that the random variable of interest
(actual minutes incurred per chair or its associated standard vari
ance) is normally distributed for both states, 3) the lowest observed
value of the random variable, 4) the highest observed value of the
random variable, 5) the maximum costs associated with each state, and
6) the minimum costs associated with each state.
The second level of the information variable, labeled 12^, includes
additional distributional information. In addition to the six items
contained in the II information set, the following two items are
included: 1) the mean of the random variable within each state, and
2) the standard deviation of the random variable within each state.

37
The distribution variable
Manipulation of the distribution variable involves two factors:
1)the distributional parameters of each state, and 2) the statistical
relationship between the states. The two levels of this variable are
generated through a change in the variance and a change in the stan
dardized distance between the means of the two states. The first
level of the distribution variable, labeled Sl_, has the following
parameters and relationships:
1) y1 = 36.0 actual minutes incurred per chair;
2) crn = a = Gj = 3.0 actual minutes incurred per chair; and
3) u12 = y + 1.5a, = 40.5 actual minutes incurred per chair,
where, y = the mean of the jth state of nature (in-control = 1 and
* J
out-of-control = 2) given the ith distribution level
(SI = 1 and S2 = 2);
a = the standard deviation of the jth state of nature given
' J
the ith distribution level; and
a = the standard deviation common to both states of nature
given the ith distribution level.
The second level of the distribution variable, labeled S2^, has
the following parameters and relationships:
1) y = 36.0 actual minutes incurred per chair;
2) a21 = a22 = a9 = 5.0 actual minutes incurred per chair; and
3) y2 = y21 + 1.8a = 45.0 actual minutes incurred per chair,
where y , a^-, and a^ are defined the same as in the SI level. The
third and fourth statistical moments of each state within each dis
tribution level are uncontrolled except that their deviations from a
normal distribution are minimized.

38
The cost variable
Given the two states of nature (the realization of which is un
known to the individual) and two possible decisions, there follows that
two types of errors can be made in reaching a decision. The first type
of decision error, labeled type A, is the decision to investigate the
physical process when the in-control state exists. The second type of
decision error, labeled type B, is the decision not to investigate when
the out-of-control state exists. The marginal cost of either error
type equals the cost that would have been incurred had the correct
decision been made minus the cost that was incurred by the incorrect
decision. Thus, the marginal costs of the two decision error types,
labeled MC^ and MCg, are:
= C(decision=not investigate | state=in-control) -
C(decision=investigate | state=in-control), and
MCg = C(decision=investigate | state=out-of-control) -
C(decision=not investigate | state=out-of-control),
where C( ) represents the decision cost for the situation within the
parentheses.
When the marginal cost of a type A error equals the marginal
cost of a type B error the cost structure should be ignored when making
an investigation decision. Of greater interest are situations in which
the marginal error costs not equal. In the present study, each level
of the cost variable takes one of the following forms: 1) the
marginal cost of a type B decision error equals three times the marginal
cost of a type A error (labeled level CJ_), and 2) the marginal cost
of a type A error equals three times the marginal cost of a type B
error (labeled level C2).

39
Individual decision model sensitivity variables
The sensitivity of the individual's decision model relative to the
decision situation is measured using the TSD parameter d1. The theo
retical definition of d1 is:
d = (p2 Pj)/o = zi Z2
where y. = the mean of the jth state of nature (in-control = 1 and
out-of-control = 2);
a = the common standard deviation of the states of nature; and
z. = the value of the normal distribution function associated
J
with the jth state of nature and any decision cutoff value
common to both states.
An empirical estimate of the d1 for an individual, labeled d 1,
is obtained using the individual's conditional probabilities P(decision=
investigate | state=out-of-control) and P(decision=investigate | state=
in-control) to calculate a subjective z1 and z£. As previously pointed
out, d' is relative to the decision situation. In particular, d is
relative to the distribution variable. To gain comparability across
situations the following measure is used:
DN, / d¡
where d^ is generated using optimal model k. The measure decreases
from a value of one as the result of several factors: 1) the individual
is inconsistent in his use of his cutoff value (i.e., there exists a
range around his cutoff value within which decisions are not made
using a strict relation to this cutoff value) or the individual makes
one or more temporary processing errors, and 2) the individual
utilizes more than one cutoff value.

40
Based upon a subjective analysis of each individual's decisions,
the measure DN^ can be adjusted for the effects of using multiple cut
off values. Defining d^a to be the individual's decision model sensi
tivity with the effects of multiple cutoff values eliminated:
DNAi = d\a / d'k
The difference DNA-- DN- approaches zero as the effects of multiple
cutoff values decreases and becomes zero when the individual uses a
single cutoff value.
Individual decision criteria variables
The criteria the individual adopts in making his decisions are
measured using the TSD parameter 3. The theoretical definition of 3 is:
3 = d>(z2) / where ( ) denotes the normal density function for the point in the
parentheses and z. is defined the same as for the d'. variable.
J
An empirical estimate for the 3 of an individual, labeled 3.5
is obtained using the z1 and z2 values associated with the individual's
conditional probabilities employed in estimating d^.. The measure 3..
is relative to the decision situation. In particular, it is relative
to those variables which affect the point on the decision axis which
the individual selects as his investigation decision cutoff value. A
measure of individual criteria comparable across situations is defined
as:
BN. = 3. / 3.
i i k
where 3^ is generated using optimal model k. The measure approaches a
value of one as 3_. approaches 3^ (i.e., as the individual's cutoff
decision value approaches the optimal model's cutoff decision value).

The measure approaches either a value of zero or positive infinity as
the result of several factors: 1) the individual does not process
properly the effects of the relative costs of the two types of decision
errors, 2) the individual does not process properly the effects of the
relative frequencies of the two states, and 3) the individual uses more
than one cutoff value.
Using the same subjective analysis as that used in adjusting the
DN measure, the BN^ measure can be adjusted for the effects of multipl
cutoff values. Defining B^ to be the individual's criteria measure
with the effects of multiple cutoff values eliminated:
BNAi B* / Bk
The difference BNA^- BN^ approaches a value of zero as the effects of
multiple cutoff values decrease and becomes zero when the individual
uses a single cutoff value.
Another measure derived from the BN- measure can be considered
a measure of individual conservatism. In this study, conservatism
refers to incomplete adjustment from an initial decision anchor towards
the optimal cutoff value. Non-conservatism refers to a more that
complete adjustment from the initial decision anchor past the optimal
cutoff value. The extent of conservatism is measured by the relative
distance between the individual's cutoff value and the optimal model's
cutoff value. Since the measure is dependent on the direction of ad
justment it is conditional upon the level of the cost variable. Using
BNC.¡ to denote tne extent of an individual's conservatism:
BNC.¡ = (3? B|<) / B^ given the Cl cost level; and
BNC.j = (b^ Bj) / B^ given the C2 cost level.

42
BNC.J approaches either positive infinity or a value of one as the extent
of conservatism increases, approaches a value of zero as the extent of
conservatism decreases, and approaches either a value of negative one
or negative infinity as the extent of non-conservatism increases.
Initial and final decision anchor variables
The relative relationships between an individual's initial and
final decision anchors can be measured in terms of the relative linear
distance between these two decision anchors. A measure of the relative
adjustment for individual i, labeled RA^, is conditional on the direc
tion of adjustment along the relevant decision axis. This direction
of adjustment is in turn conditional on the level of the cost variable.
Given the Cl cost level the initial decision anchor is greater than the
optimal model cutoff value, and given the C2 cost level the initial
decision anchor is less than the optimal model cutoff value. The RA.
measure is defined as:
RAi = (EDVi 0DVk) / (TDVi 0DVk) ,given the Cl cost level;
RA.. = (0DVk EDV.¡) / (0DVk TDV.¡) ,given the C2 cost level,
where EDV^ = individual i's final (experiment) decision anchor;
TDV. = individual i's initial (training) decision anchor; and
0DVk = optimal model k's final (experiment) cutoff value.
The RA.j measure has the following relationships: 1) if no
adjustment occurs between the training and the experiment phases the
measure will equal a value of one, 2) if the adjustment is in the wrong
direction (away from the optimal model's experiment cutoff value) the
measure will be greater than a value of one, 3) if the adjustment is in

43
the proper direction but incomplete the measure will be less than a
value of one but greater than a value of zero, 4) if the adjustment
is in the proper direction and is complete the measure will equal a
value of zero, and 5) if the adjustment is in the proper direction but
more than complete the measure will be less than a value of zero.
Individual long-run decision efficiency
A major dependent variable of interest in this research is the
cost incurred as a result of the individual's variance investigation
decisions. The experimental objective function for all decision
situations is to minimize these costs (labeled investigation decisions
costs). The individual's long-run decision efficiency is measured in
terms of his minimization of the investigation decisions costs summed
over all decisions.
Given a pair of decision situations which hold constant all of
(
the independent variables except for the cost variable,the correct
decision within each situation will lead to different investigation
decisions costs. Consequently, absolute investigation decisions costs
are not comparable between decision situations: a relative measure
must be obtained. One such relative measure involves the comparison of
the individual's investigation decision costs with that of an optimal
model which uses the same information as that available to the indi
vidual. Denoting such a measure G-:
Gij = (ICi0 MCko> / SMCk
where IC- = individual i's investigation decision cost for decision j;
MC^j = optimal model k's investigation decision cost for decision
j; and

44
SMC= the sum of optimal model k's investigation decisions costs
over all decisions (m in number).
The G-. measure can be classified into one of three submeasures
' J
where the classification is dependent upon the algebraic sign of G.¡j.
The classifications are 1) G- > 0-- the individual's investigation
decision cost for decision j is greater than that of the optimal model,
2) G^j = 0-- the individual's investigation decision cost for decision
j is equal to that of the optimal model, and 3) G < 0-- the indi-
* J
vidual's investigation decision cost for decision j is less than that
of the optimal model.
Summing those G-s with identical algebraic signs, the summations
' 0
are defined as follows:
GP.j = the sum of those Gs with a positive algebraic sign;
GN^ = the sum of those Gs with a negative algebraic sign; and
GZ.j = the sum of those Gs with a value of zero.
It is easily shown that the following relation holds:
m
E G.. = GP, + GN, + GZ,
j=l J
where m equals the total number of decisions made by individual i. Since
the GZ.¡ summation equals zero, the sum of the G over all j = l,m
decisions, denoted G-, is:
G- = GP- + GN.
i i i
Conceptually, GP^ is the relative additional cost of those de
cisions made by individual i which are greater in value than that of
the optimal model. The GN^ is the relative savings of those decisions
made by individual i which are less in value than that of the optimal
model.

45
The anchoring and adjustment process may occur over three training
sessions and one experiment session. Using the first training session
as an estimate of the initial decision anchor and the experiment
session as an estimate of the final decision anchor, there remains two
training sessions within which the adjustment process itself can be
analyzed. In particular, the effects of the pattern of adjustments
within the training sessions are expected to influence significantly
the magnitude of an individual's experiment Gj measure. The adjustment
process can be measured in the training sessions by computing the
individual's Gj measure at the completion of each session. The
measures, labeled TG^, TG.¡0, and TG.¡3, are defined in the same manner
as the G. measure discussed above (with the exception that they are
computed from the appropriate training session data instead of the
experiment session data). The differences between these training
measures reflect the effects of the individual's adjustments in the
training sessions. These differences are defined as follows:
DTGii = TG11 TGi2
DTGi2 = TGi2 TGi3
where, DTGjj = the difference between the G-¡ measure for the jth and
the jth + 1 training sessions of individual i; and
TG^j = the G.j measure for the jth training session of indi
vidual i.
Negative DTGjjvalues indicate an increase in the TGjj measure between
training sessions, and positive DTG-jj values indicate the reverse.
The effects of the directions of these adjustments should be
related to the effects of the final anchor on the individual's G-j
measure (in the experiment). The effects of the adjustment measures

46
on tha experiment measure can be analyzed using a discrete classifica
tion of the direction of the adjustments. This classification of the
measure is accomplished using its algebraic sign. The second and
third training session results can be grouped into one of four
classes: PP(+,+), PM(+,-), MP(-,+), and MMSubjects within
the MM classification have constantly increasing TG-jj measures and
subjects within the PP classification have constantly decreasing TG^j
measures.
Hypotheses Formation
As previously discussed, the formation of the experimental
hypotheses is accomplished primarily using the method of simulation.
Two general types of simulations are performed simulation of optimal
model performances and simulation of subjective investigation decision
performances. The first part of this section discusses the assumptions
and presents the results of the simulation of optimal model perform
ances. The second part of this section discusses certain assumptions
derived from the conceptual development that form the basis for the
simulation of subjective investigation decision performances. It also
presents the results of this simulation and the derivation of the ex
perimental hypotheses.
Simulation of Optimal Model Performances
The decision rule of the optimal model (given the objective func
tion of minimizing incurred costs) employed in this research is to
investigate the reported labor efficiency variance when:

47
P(x|out) P(in) V1r|>N Vin>Y
Equation 1
where x = the actual minutes incurred per chair;
out = the state of out-of-control;
in
= the state of in-control;
Vin,N = ^e cost associated with not investigating an in-control
variance;
V-¡nsy = the cost associated with investigating an in-control
variance;
V0Ut,Y = the cost associated with investigating an out-of-control
variance; and
Vout,N = the cost associated with not investigating and
out-of-control variance.
The manipulation of the cost variable produces a constant cost ratio
at each of the two levels of this variable. When the cost variable
level is Cl the constant cost ratio is equal to one-third, and when the
cost variable level is C2 the constant cost ratio is equal to three.
The optimal model investigation decision rule expressed in equation 1
can be restate as:
LRq > PR.j CR Equation 2
where LRQ = the likelihood odds of the reported labor efficiency
variance being out-of-control (P(x|out)/P(x|in));
PR.j = the prior odds of the in-control state (P(in)/P(out)); and
CR = the constant cost ratio associated with the appropriate
level of the cost variable.
Assuming that both states of nature are normally distributed with equal
variance the .optimal model investigation decision rule expressed in

48
equation 2 can be restated in the following mathematical terms:
exp -(x y2)2/2cr2
: r > PR-i CR
exp -(x yj^cr2
where x = actual minutes incurred per chair;
Vj = the mean of the in-control state;
y2 = the mean of the out-of-contrcl state;
a = the common standard deviation of both states; and
exp = the notation for the exponential funcition (e).
If the above equation was changed to an equality and solved for x the
result would be the optimal cutoff value in terms of actual minutes
incurred. This solution takes the general form:
G2(ln(PR-¡) + ln(CR))
x = + 0.50(y2 + y,) Equation 3
y2 yi
where x = the optimal cutoff value in terms of actual minutes incurred;
and
In = the notation for the natural logarithm function (loge).
With the decision rule in the form of equation 3 the parameters required
to fit the optimal model become identifiable. These parameters include
the mean of both states of nature, the common variance of the two
states, the prior odds of the in-control state, and the appropriate
value of the cost ratio.
With knowledge of the parameters required to fit the optimal
model the criteria employed for the simulation of optimal model per
formances can be discussed. The first criterion is that information
available to the optimal model be the same as the information avail
able to the individual. The information available to the individual
is manipulated by the levels of the information variable. Within the 12

49
information level the available information set contains statistical
estimates of all the parameters required to fit the optimal model.
However, within the II information level the available information
set does not contain all the required parameter estimates. Therefore,
within the II information level the optimal model must use the training
sessions to estimate the missing parameters. The parameter estimates,
both given and estimated, are presented in Table 2.
The second criterion is that the labor efficiency variance re
ports presented to the subjects in the experiment be the same reports
used in the simulation of optimal model performances. The appropriate
parameter estimates presented in Table 2 were used in the optimal
model decision rule expressed by equation 2 to simulate optimal model
investigation decisions. Table 3 presents some performance results
of this optimal model simulation.
Simulation of Subjective Investigation Decision Performances
This section discusses assumptions derived from the conceptual
development that form the basis for the simulation of subjective investi
gation decision performances. It also presents the results of this
simulation and the derivation of the experimental hypotheses.
Assumptions of simulation
The first assumption employed for the simulation of subjective
investigation decision performances is that the subjects will behave
as if they use the anchoring and adjustment heuristic during the
training phase. The anchoring and adjustment heuristic proposes that

TABLE 2
ESTIMATES OF THE PARAMETERS REQUIRED TO FIT THE OPTIMAL MODEL GIVEN THE VARIOUS CONDITIONS
Information
Experimental
Variable 11
Parameters
Information
Variable 12
Actual Underlying
Data Structure Parameters
Cost
Parameter
Distribution Variable
Distribution Variable
Distribution Variable
Variable
Estimate
SI
S2
SI
S2
SI
S2
35.99
36.26
36.00
36.00
35.975
35.943
y2
40.21
44.24
40.50
45.00
40.492
44.995
Cl
CT?
8.88
29.42
9.00
25.00
9.157
25.060
7.38
19.32
9.00
25.00
9.437
25.002
PRi
1.50
1.50
1.50
1.50
1.500
1.500
CR
0.33
0.33
0.33
0.33
0.333
0.333
C2a
CR
3.00
3.00
3.00
3.00
3.000
3.000
aThe parameter estimates
given the Cl cost level.
for u1 ,y2,cj2
,and PR.¡
given the C2
cost level are
the same as those

TABLE 3
RESULTS OF THE SIMULATION OF OPTIMAL MODEL PERFORMANCES GIVEN THE VARIOUS CONDITIONS
k Information
Distribution Cost
P(investigate
out-of-control)
P( investigate
in-control)
Optimal Cutoff
Value (actual
minutes incurred)
dk
6k
1
11
SI
Cl
0.875
0.383
36.90
1.447
0.539
2
11
SI
C2
0.400
0.033
41.29
1.581
5.206
3
11
S2
Cl
0.875
0.300
38.70
1.674
0.592
4
11
S2
C2
0.425
0.017
45.88
1.939
9.453
5
12
SI
Cl
0.900
0.400
36.86
1.535
0.454
6
12
SI
C2
0.400
0.033
41.26
1.581
5.206
7
12
S2
Cl
0.900
0.300
38.57
1.806
0.504
8
12
S2
C2
0.525
0.033
44.68
1.897
5.364
cn

52
a subject will select an initial anchor as a first decision approxi
mation and will adjust this decision anchor toward an optimally correct
point as he learns from his experiences with the states of nature.
However, the individual's adjustment process will not be sufficient:
i.e., he will tend to approach the optimally correct point but will not
adjust completely to that point.
The second assumption is that the location of the initial de
cision anchor on the actual minutes incurred continuum (axis) will be
conditional upon the level of the distribution variable. The reason
for this is that the initial decision anchor is expected to be located
at a central point between the means of the two states. Since the
mean of the out-of-control distribution shifts with the level of the
distribution variable, the central point between the means of the two
states also shifts with the level of the distribution variable. The
initial decision anchor used in the simulation of subjective investi
gation decision performances is the geometric intersection point of the
two states' distribution curves. This particular initial decision
anchor is used because it is the exact central point between the means
of the two states. However, it is selected for operational purposes
and is not necessary to the conceptual development. Any point of
central tendency would be satisfactory.
The final assumption is that the subjects' adjustment process
will be approximately equal over the various treatment conditions. The
adjustment process is defined as a linear movement along the actual
minutes incurred axis from the initial decision anchor towards the
appropriate optimal model decision cutoff. The magnitude of the linear
movement used in the simulation of subjective investigation decision

53
performances is 50 percent of the distance between the initial decision
anchor and the appropriate optimal decision value. The 50 percent
adjustment value, selected for operational purposes, is completely
arbitrary and is not necessary to the conceptual development. The only
restriction is that it be less than 100 percent. Furthermore, the
general assumption of equal adjustment over the various conditions is
not necessary to the conceptual development: the objective of this
general assumption is to facilitate a simple operationalization of the
simulation.
Training phase simulation and hypotheses formation
The first stage of the simulation involves simulating the
anchoring and adjustment process in the training sessions for each
combination of independent variables. A simulated subject sample size
of one is used for each condition (resulting in a total simulated sub
ject sample of eight). The results of this first stage simulation are
presented in graphical form in Figure 2. This figure depicts the
simulated initial decision anchors (points A), the simulated final
decision anchors (points B), and the simulated optimal cutoff values
(points C) for each of the eight combinations of independent variables.
The assumptions used in simulating the training phase can be
investigated using the results of the experimental training phase. The
assumptions are i) that the subjects will behave as if they used the
anchoring and adjustment heuristic, 2) that the initial decision anchor
will be located centrally between the means of the two states and will
be dependent on the level of the distribution variable, and 3) that the

II Information Level
Cost
Cl
Cost
C2
Distribution
SI
1 1 i
36.90 37.575 38.25
+
39.77 41.29
Distribution
S2
C B A
h
38.7 39.6 40.50
43.19
H
45.88
Distribution
SI
12 Information Level
Cost
Cl
B A
+
36.86 37.555 38.25
Cost
C2
4
39.755 41.26
C B A
Distribution 1 } 1
S2 38.57 39.535 40.50
4-
42.59
4
44.68
Note: All scales are actual minutes incurred per chair produced.
FIGURE 2
SIMULATED INITIAL DECISION ANCHORS (POINTS A), FINAL DECISION
ANCHORS (POINTS B), AND OPTIMAL CUTOFF VALUES (POINTS C)

55
adjustment process will be approximately equal over the various con
ditions. The location of the initial decision anchor can be investigated
by examining the relationship between the individual's initial cutoff
value (in the first training session) and the intersection point of
the appropriate two distribution curves (the TDV.¡ measure). These
intersection points are conditional upon the level of the distribution
variable. Given the SI distribution level the intersection point is
38.25 actual minutes incurred, and given the S2 distribution level the
intersection point is 40.5 actual minutes incurred (see Figure 2 for a
graphical presentation of these points). The expected relationship is
that the mean initial decision anchor (TDV.) of the individuals, given
a level of the distribution variable, will not be significantly differ
ent from the intersection point of the appropriate two distribution
curves.
The assumption of approximately equal adjustment over the various
conditions can be investigated by examining the relationship between
the individual's adjustment over the complete experiment and the total
adjustment that would be required to reach the optimal model's decision
cutoff value. A measure of the relative adjustment for individual i,
RA.j, is conditional upon the direction of adjustment along the relevant
decision axis (i.e., the actual minutes incurred per chair produced).
The expected relationship is that the mean RA^ will not differ signifi
cantly between the levels of the various conditions.
Given the above simulation and discussion the following hypo
theses can be derived:
HI.la (TDV.|S1) = 38.25.
The mean initial decision anchor given the SI distribution level

56
will equal the intersection point of the two distribution curves
within that level.
HI.lb (TV.|S2) = 40.5.
The mean initial decision anchor given the S2 distribution level
will equal the intersection point of the two distribution curves
within that level.
HI.2a (TDV.|I1) = (157^12).
The mean initial decision anchor given the II information level
will equal the mean initial decision anchor given the 12 infor
mation level.
HI.2b (TDViiCl) = (TVi|C2).
The mean initial decision anchor given the Cl cost level will
equal the mean initial decision anchor given the C2 cost level.
HI.3a (RAi111) = (RAi112).
The mean relative adjustment given the II information level will
equal the mean relative adjustment given the 12 information level.
HI.3b (RA^Sl) = (RAi ¡S2).
The mean relative adjustment given the SI distribution level will
equal the mean relative adjustment given the S2 distribution
level.
HI.3c (RAi|Cl) = (RA.|C2).
The mean relative adjustment given the Cl cost level will equal
the mean relative adjustment given the C2 cost level.
Individual decision model sensitivity hypotheses formation
The measures of individual decision model sensitivity should
generally be unaffected by the independent variables. These measures

57
have a greater relationship with individual decision processes than
with the specific conditions of a general decision situation. The
individual decision model sensitivity measure dj should have a re
lationship with the distribution variable only. The theoretical d1
measure under the SI distribution level is 1.5 and under the S2 dis
tribution level is 1.8. Thus, the average dl under the S2 distribu
tion level should be greater than under the SI distribution level.
The DNA.¡ measure should have no systematic relationships with
the three independent variables. This measure of the relative de
viation of the individuals decision model sensitivity from that of
the optimal model's generally is due to individual decision incon
sistencies and temporary processing errors. If individuals are
randomized into the specific decision situations there are no a
priori reasons for expecting significant differences in this measure
due to the independent variables.
The difference between DNA^ and DN-¡ measures the effects of
multiple decision anchors upon the individual's relative decision
model sensitivity. Relationships between this difference and the in
dependent variables depend upon whether the independent variables are
related to the causes underlying an individual's use of more than one
cutoff value. A possible explanation involves the distribution vari
able. As the variance of the in-control state increases the absolute
distance between the lowest values of the random variable of interest
(actual minutes incurred) and the mean of the in-control state in
creases. In turn, as this distance increases subjects may be more
inclined to perceive that a second out-of-control distribution overlaps
the lower range of the in-control state. If such perceptions underly

58
the use of multiple cutoff values, the average difference between DNA.¡
and DN.¡ should differ between the two levels of the distribution vari
able. A second possible explanation involves the cost variable. As
the individual's final cutoff value increases, the absolute distance
between the individual's final cutoff value and the lowest values of
the random variable of interest increases. In turn, as this distance
increases subjects again may be more inclined to perceive that a sec
ond out-of-control distribution overlaps the lower range of the
in-control state. If such perceptions underly the use of multiple
cutoff values, the average difference between DNA- and DN.¡ should
differ between the two levels of the cost variable.
Based on the above discussion the effects of the independent
variables on the individual decision model sensitivity measures can
be hypothesized as follows:
H2.la (di |SI) < (di |S2).
Those subjects within the SI distribution level will have
significantly smaller mean djs then those subjects within the
S2 level.
H2.lb (di ¡II) = (di ¡12).
Those subjects within the II information level and within the
12 information level will have equal mean d^s.
H2.1c (di 1 Cl) = (di |C2).
Those subjects within the Cl cost level and within the C2 cost
level will have equal mean dis.
H2.2a (DNAilll) = (DA-j ] 12).
Those subjects within the II information level and within the 12
information level will have equal mean DNAjS.

59
H2.2b (DNA-j | SI) = (DNAj |S2).
Those subjects within the SI distribution level and within the
S2 distribution level will have equal mean DNA-jS.
H2.2c (DA-¡ 1 Cl) = (DA-¡ IC2).
Those subjects within the Cl cost level and within the C2 cost
level will have equal mean DNA-¡s.
H2.3a (DNA-¡-DN-¡ ¡SI) < (DNAi-DNi |S2).
Those subjects within the SI distribution level will have
significantly smaller mean DNA-¡-DN-¡s than will those subjects
within the S2 distribution level.
H2.3b (DNAj-DN.¡ |Cl) <(DNAi-DNi | C2).
Those subjects within the Cl cost level will have significantly
smaller mean DNA-¡-DN-¡s than will those subjects within the C2
cost level.
Individual decision criteria simulation and hypotheses formation
The variables which affect the BN-j measure should be those factors
related to the individual's selection of a cutoff value. The most
significant variable affecting the individual's cutoff value should be
the cost variable. Within this variable the initial decision anchor
given the Cl level is much closer to the optimal cutoff value than is
the initial decision anchor given the C2 level. Consequently, the BN-j
measure should be closer to a value of one under the Cl level than
under the C2 level.
The difference between the BNA-¡ and the BN-j measures should
relate to the same variables as those discussed for the difference

60
between the DNA.¡ and DN^ measures. The two possible explanations in
clude the distribution variable and the cost variable.
Theoretically, the more extreme the optimal cutoff value relative
to a central point between the distributions of the two states, the
higher should be the individual conservatism. The decision situations
with the most extreme optimal cutoff values are those within the C2
cost level; the situations with the least extreme optimal cutoff values
are those within the Cl cost level. Consequently, the BNCj measure
should be larger within the C2 cost level than within the Cl cost level.
The simulation and assumptions used in developing the training
phase hypotheses can be extended to enable formation of hypotheses con
cerning the individual decision criteria. The earlier simulation de
rived the following final decision cutoff values: 1) 37.555 actual
minutes incurred for the Cl cost level given the 12 information level
and the SI distribution level, and 2) 39.755 actual minutes incurred
for the C2 cost level given the 12 information level and the SI dis
tribution level. Given these cutoff values a f(e^) measure can be
calculated directly (the measure is used rather than the 3. measure
due to the assumption of a single cutoff value). The functional
notation, f( ), is employed to indicate these estimates are based upon
a simulation rather than upon empirical observation. For this simu
lation the f(3?)s are as follows:
f(B*|Cl,Sl,I2) = (zout) / <¡>(zin) = 0.70818
f(B*|C-2,Sl,I2) = (zQut) / *(zin) = 2.11774
where ( ) indicates the value of the normal density function at the
point in parentheses. Given the f(B^) measures the f(BNA^) measures
can be calculated. For this simulation the f(BNA^)s are as follows:

61
f(BNAi|Cl,Sl,I2) = f(6*) / f(6k) = 1.41898
f(BNAi|C2,Sl,I2) = f(e?) / f(f?k) = 0.47017
where f(ek) is the simulated measure at the cutoff value of the kth
optimal model.
This simulation can be extended to the S2 level of the distribu
tion variable given the 12 information level and can be extended to
both the SI and S2 levels of the distribution variable given the II
information level. The f(BNA-) measures obtained under the 12 infor
mation and S2 distribution conditions are as follows:
f(BNAiJCl,S2,12) = 1.42318
f(BNAi¡C2,S2,I2) = 0.47300
The f(BNA-) measures given the II information level are as follows:
f(BNAi|Cl,Sl,Il) = 1.40126
f(BNAi|C2,Sl,Il) = 0.46993
f(BNAi|Cl,S2,Il) = 1.38273
f(BNA.|C2,S2,Il) = 0.38094
Based on these f(BNA-) measures, hypotheses can be derived con
cerning the effects of the independent variables upon the BNA. measure.
The f(BNA^) measures can be averaged over the appropriate conditionals
to determine these effects. Averaging over all conditionals except the
information variable gives the following results:
f(BAiIII) = 0.90871
f(BAi|12) = 0.94633
Although there is a difference in the f(BNA-) measures due to the infor
mation variable this difference is small compared to the standard error
of these estimates (the difference divided by the standard error is
0.03762/0.39157 = 0.0961). Therefore, a significant effect due to the
information variable would not be expected.

62
Averaging over all conditionals except the distribution variable
gives the following results:
f(BA-¡ 1 SI) = 0.94009
f (BA-¡ 1S2) = 0.91496
Again, the difference in the f(BNAj) measures due to the distribution
variable is small compared to the standard error of the estimates (the
difference divided by the standard error is 0.02513/0.39174 = 0.C642).
A significant effect due to the distribution variable would not be
expected.
Averaging over all conditionals except the cost variable gives
the following results:
f(BA-¡ |Cl) = 1.40654
f(BA-¡ |C2) = 0.44851
The difference in the f(BNA-j) measures due to the cost variable is
large compared to the standard error of the estimates (the difference
divided by the standard error is 0.95803/0.02436 = 39.3308). Con
sequently, a significant effect would be expected for the cost
variable.
Another implication drawn from the f(BNAj) measures concerns the
homogeneity of variance for the measures given different levels of the
cost variable. The variance for the f(BNA-¡ | Cl) is .00034 and the vari
ance for the f(BNA-j |C2) is .00203. Using the F test for equal variances,
F=5.9344 (3 and 3 d.f.) which indicates the variances may not be equal.
The variance of (BNA-¡ [Cl) would be expected to be less than the vari
ance of (BNAj[C2).
This simulation can be extended to enable the formation of
hypotheses concerning the effects of the independent variables on the

63
conservatism measure BNC-¡. This extension employs the f(B-j) and fU^)
measures used above. The results of this extended simulation averaged
over all conditionals except for the information variable are as follows:
f(BNCi | II) = 0.48328
f(MCi 112) = 0.47475
The difference in the f(BNC^) measures due to the information variable
is small compared to the standard error of the estimates (the differ
ence divided by the standard error is 0.00853/0.06390 = 0.1335). A
significant effect due to the information variable would not be expected.
Averaging over all conditionals except for the distribution
variable gives the following results:
f (BCi|SI) = 0.47004
f (BICi |S2) = 0.48799
The difference in the f(BNC^) measures due to the distribution vari
able is small compared to the standard error of the estimates (the
difference divided by the standard error is 0.01795/0.06357 = 0.2824).
A significant effect due to the distribution variable would not be
expected.
Finally, averaging over all conditionals except for the cost
variable gives the following results:
f(BC_¡ i Cl) = 0.40654
f(BCi|C2) = 0.55149
The difference in the f(BNC.j) measure due to the cost variable is large
compared to the standard error of the estimates (the difference divided
by the standard error is 0.14495/0.02436 = 5.9503). Consequently, a
significant effect would be expected for the cost variable.

64
The above discussion concerning the individual decision criteria
measures can be summarized as follows:
H3.1a (BffjIll) = (BA-¡ 112).
The mean BNAj of those subjects within the II information level
and within the 12 information level will be equal.
H3.1b (BA-¡ |SI) = (BA-¡ |S2).
The mean BNAj of those subjects within the SI distribution level
and within the S2 distribution level will be equal.
H3.1c (BNA-j | Cl) > (BA-¡ | C2).
Those subjects within the Cl cost level will have a significantly
larger mean BNA-j than will those subjects within the C2 level.
H3.2 a2(BNA-j ¡C1) < a2(BNA-¡ | C2).
The BNA-¡ of those subjects within the Cl cost level will have a
significantly smaller variance than will the BNA-j of those sub
jects within the C2 cost level.
H3.3a (BCj ¡II) = (BCj 112).
The mean BNC-¡ of those subjects within the II information level
and within the 12 information level will be equal.
H3.3b (BNCj|Sl) = (BC-j |S2).
The mean BNC-j of those subjects within the SI distribution level
and within the S2 distribution level will be equal.
H3.3c (BCj|Cl) < (BCj j C2).
The mean BNCj of those subjects within the Cl cost level will be
significantly smaller than the mean BNCj of those subjects within
the C2 level.

65
H3.4a (BNAi-BNi|Sl) < (BNAi-BNi|S2).
The mean BNA.-BN.. of those subjects within the SI distribution
level will be significantly smaller than the mean BNA^-BN. of
those subjects within the S2 distribution level.
H3.4b (BNA.-BN.|Cl) <(BNAi-BN^|C2).
The mean BNA^-BN^ of those subjects within the Cl cost level
will be significantly smaller than the mean BNA.-BN. of those
subjects within the C2 cost level.
Individual long-run decision efficiency simulation and hypotheses
The simulation and assumptions used in developing the training
phase hypotheses enable the formation of hypotheses concerning the
G.j variable. The graph presented in Figure 3 depicts the relations of
the values derived within the training phase simulation given the
condition (12,SI). The interval between the final cutoff value and
the appropriate optimal model cutoff value represents the areas
associated with the measure. For the Cl cost level the area of the
interval under the in-control distribution is the area associated with
the GN measure; the area of the interval under the out-of-control
distribution is the area associated with the GP. measure. These areas
are the left-most shaded areas in Figure 3. For the C2 cost level the
area of the interval under the out-of-control distribution is the area
associated with the GN. measure; the area of the interval under the
in-control distribution is the area associated with the GP^ measure.
These areas are the right-most shaded areas in Figure 3. The areas
under these curves for this simulation are as follows:

66
o
o
LO
o
LO
o
o
o
LO
LO
LO
LO
o
LO
o
CO
LO
C\J
r-.
LO
(NJ
'






LO
LO
r^
CO
CT>
o
CO
CO
CO
CO
CO
*3
FIGURE 3
GRAPH OF THE SIMULATION USED IN THE CONDITION
(12 INFORMATION, SI DISTRIBUTION)

67
Area(GN-j |Cl,SI, 12) = 0.08483
Area(GP-j (Cl,S1,12) = 0.05048
AreaiGNi|C2,S1,I2) = 0.19780
Area(GPi|C2,Sl,I2) = 0.06549
These areas must be adjusted for the relative frequency differ
ences and the relative cost differences between the various conditions.
If the in-control distribution is given a relative frequency weight of
one,then the relative frequency weight of the out-of-control distribu
tion would equal two-thirds (P(in-control) = 0.60 and P(out-of-control) =
0.40). The relative cost weights can be obtained directly from the
manipulation of the cost variable. Given the Cl cost level, errors
under the out-of-control distribution equal three times the errors under
the in-control distribution. Given the C2 cost level, errors under the
in-control distribution are equal to three times the errors under the
out-of-control distribution. The four areas associated with the
various G-¡ measures can be converted into relative relations by taking
into account both of these relative weights (frequency and cost). These
relative relations are as follows:
f(GN.¡ | Cl,S1,12) = (0.08483) (1) (1) = 0.08483
f(GP-j | Cl,SI, 12) = (0.05048) (2/3) (3) = 0.10096
f(GNi|C2,Sl,I2) = (0.19780)-(2/3)-(1) = 0.13187
f(GPi|C2,Sl,I2) = (0.06549)(1)(3) = 0.19647
The difference between the relative relations of GP.¡ and GNj
given a level of the cost variable indicates the relative effect of
these intervals upon the Gj measure. The differences for this simu
lation are as follows:

68
f(GiI Cl ,S1,12) = f(GPj|Cl,Sl,I2) f(GNi|Cl,Sl,I2)
= 0.10096 0.08483 = 0.01613
f(G*|C2,S1,12) = f(GPi|C2,S1,12) f(GN.|C2,$l,I2)
= 0.19647 0.13187 = 0.06460
This simulation can be extended to the S2 level of the distribu
tion variable given the 12 information level and can be extended to
both the SI and S2 distribution levels given the II information level.
The relative relations of the G.¡ measure obtained under the 12 infor
mation and S2 distribution conditions are as follows:
f(Gi|Cl,S2,I2) = 0.07596 0.06385 = 0.01211
f(Gi|C2,S2,I2) = 0.15741 0.10639 = 0.05102
The relative relations for the G.¡ measure given the II information
level are as follows:
f(G1|Cl,SI,II) = 0.09942 0.08230 = 0.01712
f(Gi|C2,Sl,Il) = 0.19629 0.13146 = 0.06483
f(Gi|Cl,S2,Il) = 0.07246 0.05884 = 0.01362
f(Gi|C2,S2,Il) = 0.15342 0.14078 = 0.01264
Based upon these relative relations hypotheses can be derived
concerning the effects of the independent variables on the G^ measure.
These relations can be averaged over the appropriate conditionals to
determine the effects. Averaging over all conditionals except the
information variable gives the following results:
f(Gi|II) = 0.02705
f(G.¡|l2) = 0.03596
Although there is a difference in the f(G.j) due to the information vari
able,this difference is small compared to the standard error of the
estimates (the difference divided by the standard error is

69
0.0089/0.01808 = 0.49226). Therefore, a significant effect due to the
information variable would not be expected.
Averaging over all conditionals except the distribution variable
gives the following results:
fiG^Sl) = 0.04152
f(G.|S2) = 0.02235
Although the difference between these measures is much larger than that
above, this difference is only slightly larger than the standard error
of the estimates (the difference divided by the standard error is
0.01832/0.01658 = 1.10495). The direction of the f(G^) differences
would be expected to be as predicted by the simulation but may not be
significant.
Averaging over all conditionals except the cost variable gives
the following results:
f(G.|Cl) = 0.01475
f(Gi|C2) = 0.04827
The difference between these measures is large compared to the
standard error of the estimates (the difference divided by the standard
error is 0.03353/0.00951 = 3.5258). Consequently, a significant effect
would be expected for the cost variable.
These relative relations can be averaged over all but pairs of
conditionals. Such measures could indicate the presence of interaction
effects of the independent variables. The results of averaging the
pairwise conditionals indicate only one interaction of possible
significance-- the distribution by cost interaction. The f(G^) measures
conditional to these variables are as follows:
ftf^ISl.Cl) = 0.01663

70
f(G.¡|Sl,C2) = 0.06471
f(Gi|S2,Cl) = 0.01287
f(Gi|S2,C2) = 0.03185
A graphic representation of this interaction is presented in Figure 4.
The difference between the two f(G.|Cl) measures is much smaller than
the difference between the two f(G^|C2) measures. The differences are
0.00376 and 0.03286, respectively. This would indicate that the
distribution variable has a significant effect only when given the
C2 cost level. Such q result would explain the larger but not signifi
cant effect of the SI distribution level on the f(|S ) measures.
Other implications which can be derived from these relative
relations involve relationships between f(GP.¡) and f(GN.j) measures.
Two such relationships appear to be of possible significance. Both
involve the ratio f(GP^)/f(GN.) (i.e., the effect of additional errors
relative to the effect of additional correct decisions). The results
of the first relationship are as follows:
f(GPi/GNi|C1) = 1.20480
f(GPi/GNi¡C2) = 1.38337
The difference between these measures is large compared to the standard
error of the estimates (the difference divided by the standard error is
0.17857/0.07621 = 2.34313). Consequently, the GP^/GN^ ratio is ex
pected to be larger given the C2 cost level than given the Cl cost
level.
The results of the second relationship are as follows:
f(GPi/GNi|Sl,Cl) = 1.19908
f(GPi/GNi|Sl,C2) = 1.48206
f(GP./GNi|S2,Cl) = 1.21057

71
f(G.)
Distribution
Variable
FIGURE 4
SIMULATED COST BY DISTRIBUTION VARIABLE INTERACTION ON f(G.¡)

72
f(GPi/GN1¡S2,C2) = 1.28465
The difference between the two f(GP-j/GN-¡ ]SI) is much smaller than the
difference between the two f(GP-¡/GN-j |S2). This would indicate the
presence of a distribution by cost interaction in which the effect of
the cost variable is significant only under the SI distribution level.
A graphical representation of this interaction is presented in Figure 5.
A final implication derived from these relative relations in
volves the homogeneity of variance for the f(G^) given the different
levels of the cost variable. The variance for the f(G^|Cl) is equal to
5.26 x 10-6 for this simulation, and the variance for the f(G-j [C2) is
equal to 6.06 x 10"4 for this simulation. Using the F test for equal
variances, F=115.24 (3 and 3 d.f.) which indicates the variances are
not equal. Consequently, the variance of (G-j | Cl) is expected to be less
than the variance of (G-¡|C2).
Given the above simulation the effects of the independent vari
ables on the G-¡ variable can be hypothesized as follows:
H4.1 a2(G-j | Cl) < a2(G-j |C2).
The G-j of those subjects within the Cl cost level will have
significantly smaller variance than will the G-¡ of those subjects
within the C2 cost level.
H4.2a (GilCl) < (Of|C2).
Those subjects within the Cl cost level will have a significantly
smaller mean Gj than will those subjects within the C2 level.
H4.2b (G¡i|S2) < (Gi|Sl).
Those subjects within the S2 distribution level will have smaller
mean Gj than will those subjects within the SI level. Signifi
cance is not predicted due to the interaction effect with the
cost variable.

73
Distribution
Variable
FIGURE 5
SIMULATED COST BY DISTRIBUTION VARIABLE INTERACTION ON f(GPi/GNi)

74
H4.2c (G-j | II) = (Gi 112).
The mean G-j of those subjects within the II information level
and within the 12 information level will be equal.
H4.3 (G-jlSl.Cl) (Gi|S2,Cl) < (Gi]Sl,C2) (G.j|S2,C2) .
There will be a significant interaction of the distribution and
cost variables in which the distribution variable will have no
significant effect given the Cl cost level but will have a
significant effect given the C2 cost level.
H4.4a (GPj/GN.¡ | Cl) < (GP-¡/GN-¡ | C2).
Those subjects within the Cl cost level will have a significantly
smaller mean GP-j/GN-j than will those subjects within the C2 level.
H4.4 (GP-j/GN-j |C2,S2) (GP-j/GN-j | Cl,S2) <
(GP-j/GN-j ¡C2,S1) (GPi/GNi 1 C1,S1) .
There will be a significant interaction of the distribution and
cost variables in which the cost variable will have no signifi
cant effect given the S2 distribution level but will have a
significant effect given the SI level.
Individual long-run decision efficiency and the training phase
Since performance within the experiment is assumed to be a
function of the training adjustment process, it would be expected that
the subjects within the PP classification of training adjustment
pattern would have a lower average experiment G-¡ measure than those
subjects within the MM classification. Furthermore, the mean expected
Gj measure for the mixed classifications (PM and MP) would be expected
to fall between those of the non-mixed classifications. The mixed

75
classifications have one positive adjustment which would improve per
formance above that of subjects who make only negative adjustments. On
the other hand, the mixed classifications have one negative adjustment
which would decrease performance below that of subjects who make only
positive adjustments.
The following hypotheses predict the effects of the training
adjustment pattern on the Gi variable:
H4.5 (G^PP) < (G^ |MM).
Those subjects within the PP classification of training adjust
ment pattern will have a significantly smaller mean G- measure
than those subjects within the MM classification.
H4.6 (GjPP) < (G^MP) = |PM) < (G^MM).
Those subjects within the mixed classifications of training
adjustment pattern will have mean G^s that fall between those of
the non-mixed classifications. The mean G^s of the mixed classi
fications are not expected to be significantly different.

CHAPTER IV
THE EXPERIMENT
Experimental Environment
The decision situation studied within this research was standard
cost variance investigation decision making at the operational manage
ment level. The experimental setting was the simulated environment of
a manufacturing company, and the subjects were requested to assume the
role of an operational manager of an assembly department within this
environment. The assembly department assembled a single product, a
metal folding chair. Subjects were presented with a series of standard
cost variance reports for the department and were asked to decide for
each report whether the underlying physical process should be investi
gated to correct an out-of-control situation. Each series of variance
reports were cross-sectional in nature: i.e., all reports within a
series were assumed to have occurred simultaneously and were independent
of each other.
The physical process within the simulated environment was a
labor-paced process (Barefield, 1972): i.e., the operating efficiency
of the department was determined completely by the labor efficiency of
the workers. The labor efficiency standard (stated in terms of time
per unit assembled) was based on engineering estimates that allowed for
unavoidable labor inefficiencies and reasonable variation in worker
76

77
performance (i.e., the standards were currently attainable). The sub
jects were instructed to accept the labor efficiency standard as fair
in terms of control and performance goals.
The physical labor process of the department could be in one of
two mutually exclusive states of nature: either in-control or out-of
control. The overall labor process consisted of many individual
physical labor operations: the expected aggregate of these procedures
was represented by the overall labor efficiency standard. The variance
reports, however, included only the aggregate standard and labor
efficiency variance. The labor process was defined to be in-control
when all of the individual physical procedures were being performed as
expected. The labor process was defined to be out-of-control when one
or more of these procedures was not being performed as expected.
The task of the subject was to decide whether to investigate each
labor efficiency variance for its underlying causes. The purpose of
investigation was to facilitate correction of those individual pro
cedures which were not operating as expected. Tv/o assumptions were pro
vided to aid the subject's decision making process. First, if they
decided to investigate a variance and the labor process turned out to
be out-of-control, the process would be returned to the original
in-control state with certainty. Second, if they decided not to in
vestigate a variance and the labor process was out-of-control, the process
would remain out-of-control with certainty.
Various costs were associated with the variance investigation
decision. These costs depended upon the subject's decision (either in
vestigate or do not investigate) and upon the actual state of the labor
process (either in-control or out-of-control). Estimates of these

78
costs were presented to the subject on the face of each variance report
using the following format:
If Your
Investigation
Decision Is
And If The
Assembly Line
State Is
Then
Your
Costs
Are
Investigation
Production
Total
Yes
In-control
$
X
$
0.00
$ X
Yes
Out-of-control
$
Y
$
0.00
$ Y
No
In-control
$
0.00
$
0.00
$ 0.00
No
Out-of-control
$
0.00
$
Z
$ z
The letter X represented the estimated investigation cost when the
assembly line was in-control. The cost was related to the size of the
labor efficiency variance: the more negative (unfavorable) the variance
the larger the cost. The letter Y represented the estimated investi
gation cost when the assembly line was out-of-control. The cost was
related to the size of the labor efficiency variance: the more negative
(unfavorable) the variance the smaller the cost. Given any particular
labor efficiency variance, the investigation cost when out-of-control
(Y) was always larger than the investigation cost when in-control (X).
This was because the investigation cost included correction costs when
the assembly line was out-of-control. The letter Z represented the
estimated marginal production cost of operating the next period in the
out-of-control state. The cost was constant for all variance reports.
The subjects were told that their immediate supervisor, the pro
duct section manager, would evaluate their control performance in terms
of their minimization of both investigation and production costs above
the expected standard (labeled the total investigation decision cost).
A cash bonus was promised to the subjects, the size of the bonus being
contingent upon the extent to which they minimized their total investi
gation decisions cost. The measure of a subject's control performance,
labeled TIDC^^, was determined by summing the total investigation

79
decision costs incurred by the subject over the series of variance re
ports and dividing this sum by the sum of the total investigation de
cision costs incurred by an optimal model over the same series of vari
ance reports. The subject's cash bonus function, constant for all sub
jects, was inversely related to this measure. As TIDCmi-n approached
one, the payoff approached the maximum: as TIDCmi-n became larger than
one, the payoff approached the minimum.
Subjects
The subjects were 86 senior year undergraduate and master's level
graduate students enrolled in the business college at the University of
Florida. The subjects participated in the experiment during a two week
period; 47 participated in the first week and 39 participated in the
second week. A total of 92 subjects initially volunteered to participate
but six subjects failed to complete the experiment. The 86 subjects
who completed the experiment consisted of 63 males and 23 females.
Three subject selection criteria were applied: 1) the subject
must have completed an intermediate-level managerial accounting course,
2) the subject must have completed an introductory-level statistics
course, and 3) the subject must have earned an overall grade point
average (GPA) of at least 2.0 on a 4.0 scale.
Subjects initially were contacted within senior level and graduate
accounting classes. The contact was made by the experimenter giving a
brief oral presentation followed by passing sign-up sheets around each
class (a copy of the oral presentation is presented in Appendix B). To
motivate volunteering, the presentation focused on the student's pro
fessional responsibilities as future accountants and on the monetary

80
benefits that would accrue to those who volunteered. The subjects were
told that they would be given a cash payment for participating in the
experiment and that the amount of a subject's payment would depend upon
his performance. The minimum payment was set at $2.00 and the maximum
at $10.00.
Experimental Materials
The experimental materials included a background information book
let, variance investigation decision stimuli, heuristics questionnaire,
and motivations questionnaire.
Background Information
A background information booklet (see Appendix C) was designed
to provide the subjects with a common experimental environment. The
booklet provided the subject with general company information, general
product information, general manufacturing process information, and
specific assembly department information. The specific assembly depart
ment information included information concerning the employees, the
physical process, the accounting control system, the subject's task as
the operational manager, and the subject's performance evaluation as
the operational manager.
Variance Investigation Decisions
Various information constant over all decision trials within a
treatment condition was presented on a separate page prior to the start

81
of the decision trials. The extent of such information depended upon the
treatment condition but could include information such as the prior
probabilities of both states, the means of both states, the standard
deviations of both states, the shape of the distributions of both states,
the range of past labor efficiency variances, and the range of past
investigation costs. Copies of the prior information pages for two
specific treatment conditions are presented in Appendix D.
Each variance investigation decision trial consisted of the pre
sentation of a labor efficiency variance report and a subject's re
sponse to two questions. The questions were 1) would you investi
gate this reported variance, and 2) how strongly do you feel about your
decision? During the training phase each decision trial was followed
by feedback concerning the actual state of the assembly line and the
actual costs incurred for each possible decision given the actual state.
Decision trials were presented in booklets of 33 trials (each trial
included the report with questions followed by the feedback). Within
the experimental phase decision trials were presented in booklets of
50 trials (each trial included only the report with questions). In
both the training and the experimental phases, answer sheets were pro
vided for the subject to record his responses.
An example of a labor efficiency variance report with the set of
questions is presented in Figure 6. The format of the report and
questions was constant for all treatment conditions: the only variation
related to the distribution of the actual minutes incurred per chair
and the labor efficiency variance, and to the magnitude and structure
of the costs of investigation.

82
AMSECO
Metal Folding
Chair Assembly Department
Labor Efficiency
Variance Report
For Job 5247
Standard Minutes
Actual Minutes Labor
1 Efficiency
Total Chai
Allowed Per Chair
Incurred Per
Chair Variance Per Chair
Produced
36.0
44.0
-8.0
200
The Costs Associated With Investigation Are:
If Your
And If The
Then
Your Costs
Are
Investigation
Assembly Line
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
Decision Is
State Is
Investigation
Production
Total
4rkkk kkkkkk
kkkkkkkkkkkkkk
A-*********
kkkkkkk
kkkkkkkkkkkk
Yes
In-Control
$ 28.33
$ 0.00
$ 28.33
Yes
0ut-0f-Control
$ 90.00
$ 0.00
$ 90.00
No
In-Control
$ O.CO
$ 0.00
$ 0.00
No
0u-0f-Control
S 0.00
$ 175.00
$ 175.00
Please answer the following questions placing your answers on the answer sheet:
A. Would you investigate this reported variance /circle the appropriate
response on the answer sheet/
NO
YES
B. How strongly do you feel about your decision /select a number between
0 and 100 which indicates the strength of your feeling and place this
number on the answer sheet/
0 10
20
30
40 50 60
70
80
90
100

*
*
k k k
k
*
k

kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
k k
k

k k k
k

k
k
k
k
k
Uncertain
Reasonably
Almost
Certain
Certain
FIGURE 6
VARIANCE INVESTIGATION DECISION TRIAL

83
Elicitation of Heuristics
The elicitation of each subject's heuristics was accomplished using
a predominately open-ended questionnaire (see Appendix E). The objec
tives of this questionnaire were 1) to determine the strategy the sub
ject used in answering the variance investigation questions within the
experimental phase, 2) to determine, where appropriate, the exact values
of any numerical rules used by the subject, 3) to determine whether the
subject thought he used his strategy consistently within the experimental
phase, 4) to determine the method or approach used by the subject in
forming his strategy within the training phase, and 5) to determine the
subjective importance attached to each information item presented to
the subject.
Elicitation of Subject Motivations
Subject motivations were elicited using a motivation questionnaire
(see Appendix F) developed by Snowball and Brown (1977). The question
naire is a ten item Likert-type scale which has submeasures for both
intrinsic and extrinsic motivation.
Experimental Procedures
Experimental procedures included assignment of subjects to
treatment conditions, adminstration of a training phase, administration
of an experimental phase, and final debriefing.

84
Assignment of Subjects to Treatment Conditions
Since each of the 86 subjects was assigned to one of eight groups,
randomization per se can not be relied upon to control for individual
attribute differences between groups. An alternative is to block the
randomization process on-indi vidual attribute dimensions assumed to
significantly affect the subject's information processing within the
task required by the experiment.
The literature relating to individual attributes has not pro
duced conclusive results concerning individual attributes effect on
decision processes. The most comprehensive study is that of Taylor and
Dunnette (1974) who analyzed the effect of 16 decision maker attributes
upon eight measures of predecisional, decision-point, and post decisional
processes within a personnel decision simulation experiment. Of con
siderable interest is that the relevant decision maker attributes
accounted for a generally small portion of the total decision variance
within the various processes (the range was from 8 percent to 33 percent).
The decision maker attribute with the greatest predictive capacity was
intelligence. In two of the predecisional processes (diagonosticity
and information processing rate) and in two of the decision-point and
post decisional processes (information retention and decision accuracy),
intelligence accounted for from one-half to almost all of the variance
explained by individual characteristics.
In the present study, the randomization process of assigning
subjects to treatment conditions was blocked on individual intelligence.
Ideally, individual intelligence should be measured using some validated
instrument (e.g., the Wesman Personnel Classification Test or the

85
Wechsler Adult Intelligence Scale). Due to resource limitations, however,
subject grade point average (GPA) was used as a surrogate for such a
measure. A median GPA was identified, and those subjects with a GPA
above the median were categorized as above average intelligence and
those subjects with a GPA below the median were categorized as average
intelligence. Each subject within an intelligence category then was
assigned randomly to one of the treatment conditions with the restric
tion that each intelligence group contributed an equal number of sub
jects to each condition.1 Upon assignment to a treatment condition each
subject received the background information booklet.
Training Phase
%
Each subject received training within the treatment condition to
*
which he was assigned. Training was conducted in groups of two subjects
within a 50 minute session administered by either the experimenter or
by an experimental assistant.2 Subjects were assigned randomly to the
experimental assistants subject to two restrictions. The restrictions
were that the experimental assistant was not currently the subject's
teacher for an academic course and that the subject's available time
coincided with that of the experimental assistant. Training of all
!The assignment process involved two procedures. First, the assignment
to the cost variable levels was by the week in which the subject par
ticipated in the experiment. Those subjects who participated during
the first week were assigned to the Cl cost level, and those subjects
who participated during the second week were assigned to the C2 cost
level. Second, the assignment to the information variable and the
distribution variable levels was by random selection based upon a ran
dom number table.
2The experimental assistants were not paid. The author believes, how
ever, that this is an exception to the aphorism that price reflects
value.

86
subjects (within each week) was completed over two contiguous days. The
training phase consisted of 99 decision trials with feedback. The
experimental materials used in the training phase were similar to those
used in the experimental phase. The decision trials with feedback were
presented in three booklets of 33 trials and the subject was provided
an answer sheet on which to record his responses. Additional perfor
mance feedback was given at the completion of each book of 33 decision
trials. This performance feedback was the subjects's TIDC^^ for that
particular book.
Experimental Phase
The experimental session lasted one hour and was administered by
the experimenter. The experimental phase (within each week) was com
pleted over two contiguous days immediately following the training phase.
The experimental session consisted of three parts. The first
part was the presentation of 100 decision trials. The decision trials
were presented in two booklets of 50 trials each and the subject re
corded his responses on a separate answer sheet. The subjects were
allowed twenty minutes to complete the 100 decision trials. The second
part of the experimental phase was the elicitation of the subject's
heuristics. This elicitation was accomplished using the open-ended
questionnaire (see Appendix E). The third part of the experimental
phase was the administration of the motivation questionnaire (see
Appendix F).

87
Final Debriefing
Each subject's final performance measure (TIDCmi-n) for the
variance investigation decisions part of the experimental phase was
presented individually at a later date. At this time his cash pay
ment was determined, he was debriefed as to the purpose of the experi
ment, and any questions were answered. Additional data were collected
from those subjects not responding completely to the heuristic elicita
tion questionnaire.

CHAPTER V
ANALYSES AND RESULTS
Summary of Results
A summary of the analyses and results contained in this chapter
is presented in Table 4. The objective of this summary is to provide
a brief statement of each hypothesis together with the results of analy
ses for that hypothesis. Such a summary will facilitate the presenta
tion of the analyses and results and will serve as a reference for the
discussion and conclusions contained in the next chapter.
General Method of Analysis
Although several methods of analysis are employed, the most
prevalent method is analysis of variance using the model comparison
procedure (Appelbaum and Cramer, 1974; Lewis and Keren, 1977). This
method of analysis is employed due to the nonorthogonality of the data
structure. The problem of nonorthogonality arises in this instance as
a result of non-equal cell frequencies.
The model comparison procedure involves fitting a linear model
allowing for certain effects and then comparing the obtained fit to
that of a linear model which omits one or more of the effects. The
objective is to find the simplest model that adequately fits the data.
The procedure begins with the complete or full model (which allows for
88

TABLE 4
SUMMARY OF ANALYSES AND RESULTS
Hypotheses
Analysis
Method(s)
Test
Statistic(s)
Results
1.1a (TDVi1 SI)=38.5
Two-tailed Z test
Z=0.9912
(n.s.)
The mean initial decision anchor of
those subjects within the SI distribu
tion level was not significantly
different from 38,5
1.1b (TDVi|S2)=40.5
Two-tailed Z test
Z=1.5681
(p<.10)
The mean initial decision anchor of
those subjects within the S2 distribu
tion level was significantly less than
40.5
1.2a (TDV.|Il)=(TDVi
112)
Model comparison
Two-tailed Z test
F= 1.0944
Z=0.7007
(n.s.)
(n.s.)
The mean initial decision anchor of
those subjects within the reduced infor
mation level was not significantly dif
ferent from that of those subjects with
in the expanded information level.
1.2b (TDVi|Cl)=(TDV.
1 C2)
Model comparison
Two-tailed Z test
F=3.3931
Z=1.4362
(p<.10)
(p<.10)
The mean initial decision anchor of
those subjects within the Cl cost level
was significantly smaller than that of
those subjects within the C2 cost level.
1.3a (RAi|ll)=(RA.|l2)
Model comparison
Two-tailed Z test
F=0.0903
Z=0.4432
(n.s.)
(n.s.)
The mean relative decision anchor ad
justment of those subjects within the
reduced information level was not sig
nificantly different from that of those
subjects within the expanded information
level.

Hypotheses
Analysis
Method(s)
1.3b (Mi|Sl) = (Mi|S2)
Model comparison
Two-tailed Z test
1.3c (Mi|Cl) = (Mi|C2)
Model comparison
Two-tailed Z test
2.1a (cH ¡Sl)<(dT |S2)
Model comparison
One-tailed Z test
2.1b (34 | Il) = (cT'. 112)
Model comparison
Two-tailed Z test
2.1c (dt |Cl) = (d;. |C2)
Model comparison
Two-tailed Z test
ABLE 4-- Continued
Test
Statistic(s) Results
F=1.7308 (n.s.)
The mean relative decision anchor ad
Z=1.3054 (n.s.)
justment of those subjects within the
SI distribution level was not signifi
cantly different from that of those
subjects within the S2 distribution level
F=0.1735 (n.s.)
The mean relative decision anchor ad
Z=0.4865 (n.s.)
justment of those subjects within the
Cl cost level was not significantly dif
ferent from that of those subjects with
in the C2 cost level.
F=46.0985 (p<.01)
The mean decision model sensitivity of
those subjects within the SI distribu
Z=6.1144 (p<.01)
tion level was significantly smaller than
that of those subjects within the S2
distribution level.
F=0.5883 (n.s.)
The mean decision model sensitivity of
Z=0.4588 (n.s.)
those subjects within the reduced infor
mation level was not significantly dif
ferent from that of those subjects with
in the expanded information level.
F-0.2638 (n.s.)
The mean decision model sensitivity of
Z=0.3167 (n.s.)
those subjects within the Cl cost level
was not significantly different from
that of those subjects within the C2
cost level.

TABLE 4 Continued
Hypotheses
Analysis
Method(s)
Test
Statistic(s)
Results
2.2a (DNA-j | Il) = (DNAi
112)
Model comparison
Two-tailed Z test
F=0.0066 (n.s.)
ZO.OOIO (n.s.)
The mean adjusted relative decision
model sensitivity of those subjects
within the reduced information level was
not significantly different from that of
those subjects within the expanded
information level.
2.2b (DNA-j |SI) = (DNA-j
IS2)
Model comparison
Two-tailed Z test
F=0.8368 (n.s.)
Z=0.4662 (n.s.)
The mean adjusted relative decision
model sensitivity of those subjects
within the SI distribution level was not
significantly different from that of
those subjects within the S2 distribu
tion level.
2.2c (DNA-j | Cl) = (DNA-¡
|C2)
Model comparison
Two-tailed Z test
F=15.3114 (p<.01)
Z=3.6362 (p<.01)
The mean adjusted relative decision
model sensitivity of those subjects
within the Cl cost level was signifi
cantly larger than that of those sub
jects within the C2 cost level.
2.3a (DNA-¡-DNj |S1) <
(DNA-¡ DN-¡ |S2)
Model comparison
One-tailed Z test
F=6.0130 (p<.05)
Z=2.2737 (pc.05)
The mean multiple decision anchor effect
on relative decision model sensitivity
of those subjects within the SI distribu
tion was significantly larger than that
of those subjects within the S2 dis
tribution level.

TABLE 4 Continued
2.3b
3.1a
3.1b
3.1c
Hypotheses
Analysis
Method(s)
Test
Statistic(s)
Results
(DNA-j-DN-j | Cl) <
(DNAi-DNi|C2)
Model comparison
One-tailed Z test
F=0.6618 (n.s.)
Z=0.8642 (n.s.)
The mean multiple decision anchor effect
on relative decision model sensitivity
of those subjects within the Cl cost
level was not significantly different
from that of those subjects within the
C2 cost level.
(BNAi|Il)=(BNAi|
12)
Model comparison
Two-tailed Z test
F=2.6738 (n.s.)
Z=0.7891 (n.s.)
The mean adjusted relative decision
criteria of those subjects within the
reduced information level was not sig
nificantly different from that of those
subjects within the expanded information
level.
(BNAi|Sl)=(BNAi|
S2)
Model comparison
Two-tailed Z test
F=1.9398 (n.s.)
Z=0.7048 (n.s.)
The mean adjusted relative decision
criteria of those subjects within the SI
distribution level was not significantly
different from that of those subjects
within the S2 distribution level.
(BNAi|Cl)>(BNAi|
C2)
Model comparison
One-tailed Z test
F=107.4386 (p<.01)
Z=10.6306 (p<.01)
The mean adjusted relative decision
criteria of those subjects within the Cl
cost level was significantly larger than
that of those subjects within the C2
cost level.

TABLE 4-- Continued
Analysis Test
Hypotheses Method(s) Statistic(s) Resul ts
3.2 a2(BNA.|Cl) <
F test of equal
F-9.36 (p
K.01)
The adjusted relative decision criteria
variances
variance of those subjects within the Cl
cost level was significantly larger than
that of those subjects within the C2
cost level.
3.3a (MCi|Il) = (BCi|I2)
Model comparison
F=0.9075
(n.s.)
The mean decision criteria conservatism
of those subjects within the reduced
Two-tailed Z test
Z=0.6802
(n.s.)
information level was not significantly
different from that of those subjects
within the expanded information level.
3.3b (BNCi|Sl)=(BNC.|S2)
Model comparison
F=0.3377
(n.s.)
The mean decision criteria conservatism
of those subjects within the SI distribu-
Two-tailed Z test
Z=0.3295
(n.s.)
tion level was not significantly differ
ent from that of those subjects within
the S2 distribution level.
3.3c (BNCi|Cl)<(BNCi|C2)
Model comparison
F=6.6104
(p<.05)
The mean decision criteria conservatism
of those subjects within the Cl cost
One-tailed Z test
Z=2.5598
(pc.05)
level was significantly larger than that
of those subjects within the C2 cost
level.
3.4a (BNAi-BN.|Sl) <
Model comparison
F=0.0924
(n.s.)
The mean multiple decision anchor effect
on relative decision criteria of those
(BNAi-BNi|S2)
One-tailed Z test
Z=0.3892
(n.s.)
subjects within the SI distribution level
was not significantly different from that
of those subjects within the S2 distribu
tion level.
CO

Hypotheses
Analysis
Method(s)
3.4b (BNA-¡-BN-,-1Cl) <
Model comparison
(BNAi-BN1|C2)
One-tailed Z test
4.1 a2(Gi|Cl) F test of equal
variances
4.2a (Gi|Cl)<(Gi|C2)
Sources F value
One-tailed Z test
4.2b (Gj1S2)<(Gi|SI)
Sources F value
One-tailed Z test
4.2c (Gi|Il)=(Gi|I2)
Model comparison
Two-tailed Z test
4 Continued
Test
Statistic(s)
F=11.4812 (pc.Ol)
Z=3.7090 (p<.01)
F=2.43 (p<.01)
F=2.56 (n.s.)
Z=2.442 (p<.01)
F=0.51 (n.s.)
Z=1.7452 (p<.05)
F=1.0488 (n.s.)
Z=0.4528 (n.s.)
Results
The mean multiple decision anchor effect
on relative decision criteria of those
subjects within the Cl cost level was
significantly larger than that of those
subjects within the C2 cost level.
The relative net decision costs variance
of those subjects within the Cl cost
level was significantly smaller than that
of those subjects within the C2 cost level.
The mean relative net decision costs of
those subjects within the C2 cost level
was larger than that of those subjects
within the Cl cost level.
The mean relative net decision costs of
those subjects within the SI distribution
level was larger than that of those sub
jects within the S2 distribution level.
The mean relative net decision costs of
those subjects within the reduced infor
mation level was not significantly dif
ferent than that of those subjects with
in the expanded information level.

TABLE 4 Continued
Hypotheses
Analysis Test
Method(s) Statistic(s)
4.3 [(G^|S1,C1)-(G-¡|S2,C1)] Model comparison F=7.7124 (pc.Ol)
range test
4.4a (GPi/GN-j | Cl) <
Sources F value F=0.35 (n.s.)
(GP -¡/GN | C2)
One-tailed Z test Z=2.0198 (p<.01)
4.4b [ (GP -j/GN -j | C2, S2) -
Model comparison F=8.7438 (pc.Ol)
(GPi/GNi|Cl,S2)J
Duncan's multiple (a=.01)
<[(GP-j/GN-j |C2,S1)-
range test
(GPj/GN-j | Cl, SI)]
4.5 (Gj|PP)<(Gj|MM) One-tailed t test
t=l.303 (p<.10)
Results
There was a significant cost by distribu
tion interaction on the relative net
decision costs in which the C2 cost level
had a greater effect than the Cl cost
level while the distribution levels had
little effect.
The mean relative additional decision
costs and relative additional decision
savings ratio of those subjects within
the C2 cost level was larger than that
of those subjects within the Cl cost
level.
There was a significant cost by distribu
tion interaction on the relative addi
tional decision costs and relative addi
tional decision savings ratio in which
the cost levels had a significant effect
only when given the SI distribution level.
The mean relative net decision costs of
those subjects within the constantly
decreasing training session performances
category was significantly smaller than
that of those subjects within the con
stantly increasing training session per
formances category.

TABLE 4-- Continued
Hypotheses
4.6 (Gi|PP)<(Gi|MP)=
(G-j | PM)<(G^1-1 MM)
Analysis Test
Method(s) Statistic(s)
Results
The mean relative net decision costs of
those subjects within the mixed training
session performances categories were less
than that of those subjects within the
constantly increasing training session
performances category and greater than
that of those subjects within the con
stantly decreasing training session
performances category.
LO
G">

97
all effects) and eliminates effects starting with the highest order
interactions. The model comparisons utilize the general linear model
theory (the method of least squares) as described by Scheffe" (1959)
and the F test as described by Lewis and Keren (1977) is used to test
the fit of the various models. Defining two models as ft and w, the
F test is:
d.f.(ft) SSe(w) SSe(ft)
F =
d.f. (aj-ft) SSe(ft)
where w is the more restricted model and the degrees of freedom associ
ated with the numerator and the denominator of the F statistic are
d.f.(co-ft) and d.f. (ft), respectively.
Given a dependent variable, the model comparison procedure is
used to find the simplest model which adequately fits the data. After
the simplest (or reduced) model is found the sources of this model are
presented with their corresponding F values. These source F values are
the same as model comparison F tests where the comparison models are the
reduced model and the reduced model without the corresponding source.
Within this research the source F values are also the same as the Type IV
sum of squares (Barr et al., 1976).
Unless otherwise specified, the a levels for the rejection of the
null hypothesis are pc.10 for main effects and p<.05 for interactions.
The selection of these a values is largely arbitrary: the values reflect
the preliminary nature of this research. Also, unless otherwise speci
fied, all 2-way interactions are tested against the without 3-way
interaction model, employing the assumption that the 3-way interaction
effect is equal to zero within the subject population. All the individual
attributes are retained in each model, employing the assumption that

98
these effects are not necessarily equal to zero within the subject
population.
The parameters used by the various models as independent variables
are defined as follows:
U = the overall mean of the dependent variable,
oij= the effect of level j of the information variable,
8^= the effect of level k of the distribution variable,
y-|= the effect of level 1 of the cost variable,
ag. ccy.., By,-, agy. = the various interactions of the above
J K J I KI J KI
three variables,
ip = the effect of individual i's DN measure,
i
E, = the effect of individual i's intrinsic motivation measure,
i
t = the effect of individual i's extrinsic (non-monetary)
i
motivation measure,
# *
\^= the effect of individual i's extrinsic (monetary) motivation
measure, and
5.= the effect of individual i's GPA.
i
Training Phase Analyses and Results
The following sections report the analyses and results of the
training phase. The sections include the initial decision anchors and
the decision anchor adjustment.
Initial Decision Anchors
Hypotheses 1.1 and 1.2 relate to an assumption made in the con
ceptual development and simulation: i.e., that subjects would use a
point of central tendency (between the means of the appropriate states

99
of nature) as their initial decision anchor. The simulation arbitrarily
employed the geometric intersection of the appropriate state distribu
tion curves as the expected initial decision anchor. These expected
points are conditional upon the distribution variable: i.e., the
intersection point is 38.25 actual minutes incurred given the SI distribu
tion level and is 40.5 actual minutes incurred given the S2 distribution
level. The cutoff points employed within the first training session
(TDV.s) were used as an estimate of the subjects' initial decision
anchors.
The method of analysis was the model comparison procedure where
the dependent variable is the TDV^ measure and the independent variables
are the three situation variables (information, distribution, and cost)
with their interactions, and the GPA variable. The full TDV^ model is:
TDV1 = V + Oj + Bk + Y, + a6jk + aYj, + 6Ykl + ctSyjkl + Si
The model comparison procedure results for this analysis are
presented in Table 5, and the F values associated with the sources of
the reduced model are presented in Table 6. The results indicate that
the reduced model contains the distribution variable and the cost vari
able main effects. The distribution variable main effect is significant
at the p<.01 level and the cost variable main effect is significant at
the p<.10 level. The means, variances, and sample sizes of the TDV^
measure given the levels of these two independent variables are included
in Table 6. Using the F test of equal variances, the variances between
the levels of the distribution variable differ significantly (F=2.82,
pc.01), with the S2 level having the greater variance. The variances
between the levels of the cost variable do not differ significantly
(F=l.12, n.s.). Using the Z test of equal means, the means of the levels

100
TABLE 5
MODEL COMPARISON PROCEDURE RESULTS
FOR THE TDV-j
DEPENDENT
VARIABLE
Model
SSe
d.f.(F)
F
Ful 1d
91.6549
8,77
a
8.19
Without 3-way interaction
93.1248
1,77
1.23
Without distribution x cost
interaction
93.1426
1,78
0.01
Without information x distribu
tion interaction
93.1463
1,78
0.02
Without information x cost
interaction
93.7287
1,78
0.51
e
Without information
95.0374
1,81
1.09
e
Without distribution
164.6998
1,81
61.27a
. e
Without cost
97.6986
1,81
3.39C
a
p<.01
b
pc.05
Cp<.10
d
The F value associated with the full model is the test of the full
model and not a model comparison test.
e
The test of this model assumes the effects of those interactions
that include this variable are equal to zero.

101
TABLE 6
PARAMETER VALUES ASSOCIATED WITH THE REDUCED TDV-j MODEL
F Values
Associated With the
Sources of the Reduced Model
Source
d.f.
F
Distribution
1,81
60.843
Cost
1,81
b
3.59
GPA
1,81
0.02
Means and
Variances Associated
With the Significant Sources
Variable Level
N
Mean
Variance
SI
44
38.368
0.62548
S2
42
40.179
1.76465
Cl
47
39.053
1.86776
C2
39
39.492
2.09599
a
p<. 01
b
p<. 10

102
of the distribution variable differ significantly (Z=7.6346, p<.01
two-tailed), with the S2 level having the larger mean. Differences in
the means of the levels of the cost variable are also significant
(Z=1.4362, p<.10 two-tailed), with the C2 level having the larger mean.
Hypothesis 1.1a predicted that (TDV^|S1) would not differ signifi
cantly from 38.25 actual minutes incurred. A single-sample Z test was
employed to test this hypothesis, and the difference was found to be not
significant (Z=0.9912, n.s. two-tailed). Hypothesis 1.1b predicted
that (TWi |S2) would not differ significantly from 40.5 actual minutes
incurred. A single-sample Z test indicated that the difference was
significant (Z=1.5681, p<.10 two-tailed). Hypothesis 1.2a predicted
that (TDV.j | II) would not differ significantly from (TDV^]I2). The Z
test used to test this hypothesis indicated no significant difference
(Z=0.7007, n.s. two-tailed). Hypothesis 1.2b predicted that (TDV^|Cl)
would not differ significantly from (TDV.¡|C2). This hypothesis was
tested by the above model comparison procedure and related Z test, and
the results indicated that the difference was significant (p<.10).
Decision Anchor Adjustment
Hypothesis 1.3 relates to an assumption made in the training phase
simulation: i.e., that subjects would have equal relative decision
anchor adjustment between the levels of the various experiment conditions.
The simulation arbitrarily used an adjustment of one-half the linear
distance between the appropriate expected initial decision anchor and
the appropriate optimal model decision cutoff value.
The method of analysis was the model comparison procedure where
the dependent variable is the RA^ measure and the independent variables

103
are the three situation variables (with their interactions) and the
GP/l- variable. The full RAj model is:
RAj = ii + a, + ^ + Y, + aeJk + oty-j, + BYkl + aBYjkl + 6,
The model comparison procedure results for this analysis are
presented in Table 7. The results indicate that the reduced model con-
taines only the p parameter: i.e., none of the situation variables had
a significant effect on an individual's relative decision anchor adjust
ment. Using the F test of equal variances, the variances between the
levels of the distribution variable and between the levels of the cost
variable were found to differ significantly (F=5.22, p<.01; F=6.23,
pc.01). The S2 distribution level and the Cl cost level had the greater
variances.
Hypothesis 1.3a predicted that (RA-|I1) would not differ signifi
cantly from (RA^|I2). A Z test indicated that the difference was not
significant (Z=0.4432, n.s. two-tailed). Hypothesis 1.3b predicted
that (RA.JS1) would not differ significantly from (RA^|S2). The
application of the Z test revealed no significant difference (Z=1.3052,
n.s. two-tailed). Hypothesis 1.3c predicted that (RA.¡|C1) would not
differ significantly from (RA-|C2). A Z test was employed to test this
hypothesis: again, the difference was found to be not significant
(Z=0.4865, n.s. two-tailed).
Table 8 presents the actual patterns of the TDVi and EDV.¡ measures
given the various levels of the distribution and cost variables. Given
that the TDV.¡ measures have incorporated some initial adjustment (during
the first training session), the expected TDVi and EDVi patterns are
1) (DV1|C1,S1) < (TVi|Cl,Sl) < 38.25 < (TDVi|C2,Sl) < (EDVi|C2,S1),
and 2) (EDV.|C1,S2) < (TDVi |Cl ,S2) < 40.5 < (TDVi |C2,S2) < (EDVj|C2,S2).

104
TABLE 7
MODEL COMPARISON PROCEDURE RESULTS FOR THE RAj DEPENDENT VARIABLE
Model
SSe
d.f.(F)
F
a
Full
67.8716
8,73
0.58
Without 3-way interaction
67.8852
1,73
0.01
Without distribution x cost
interaction
67.8852
1,74
0.00
Without information x distribu
tion interaction
67.8889
1,74
0.00
Without information x cost
interaction
68.2633
1,74
0.41
b
Without information
68.3470
1,77
0.09
b
Without distribution
69.8016
1,77
1.73
b
Without cost
68.4208
1,77
0.17
a
The F value associated with the full
model is the
test of
the full
model and not a model comparison test.
I
The test of this model assumes the effects of those interactions
that include this variable are equal to zero.

TABLE 8
OBSERVED MEAN
EDV.jS AND
TDV.jS GIVEN THE
VARIOUS LEVELS OF THE
DISTRIBUTION AND COST VARIABLES
Distribution
Cl Cost
Variable Level
Expected Initial
C2 Cost
Variable Level
Variable
DV
TV
Decision Anchor
TDV
EDV
SI
37.83
38.16
38.25
38.62
39.01
S2
39.62
39.98
40.50
40.41
42.43
o
en

106
The expected pattern given the SI distribution level was obtained. How
ever, the expected pattern given the S2 distribution level was not con
sistent with respect to the expected initial decision anchor (40.5).
Individual Decision Model Sensitivity
The following sections report the analyses and results of the
individual decision model sensitivity variables. The variables include
the d.! dependent variable, the relative decision model sensitivity, and
the relationships between intrinsic motivation and decision model
sensitivity.
The dj Dependent Variable
Hypothesis 2.1 relates to the fit of the subjects' responses to
the TSD model. The dependent variable d.! should have little relation
with the information and cost situation variables, but should have a
relation with the distribution variable. The relation with the dis
tribution variable is derived from the optimal model decision sensi
tivity measure, d'. Within the SI level of the distribution variable
the theoretical value of d^ is 1.5, and within the SI level of the dis
tribution variable the theoretical value of d.1 is 1.8.
k
The method of analysis was the model comparison procedure where
the dependent variable is the d.! measure and the independent variables
are the three situation variables (with their interactions), the three
motivation factors, and the GPA. variable. The full d! model is:
i i
di + j + Bk + Y, + aBjk + otYj, + BYkl + aByjkl + C, + T,

107
The model comparison procedure results for this analysis are
presented in Table 9 and the F values associated with the sources of
the reduced model are presented in Table 10. The results indicate that
the d¡ reduced model contains only the distribution variable and the
intrinsic motivation factor. Both of these parameters are significant
at the p<.01 level. The means, variances, and sample sizes of the d^'
measure given the levels of the distribution variable are included in
Table 10. Applying the F test of equal variances, the variances be
tween the levels of this variable do not differ significantly (F=1.27,
n.s.). However, the Z test of equal means indicates that the means of
the levels of this variable differ siqnificantly (Z=6.1144, d<.01
two-tailed).
Hypothesis 2.1a predicted that (d.! |S1) would be siqnificantly
less than (d! |S2). This hypothesis was tested by the model comparison
procedure and Z test presented above, and the difference was found to
be siqnificant (p<.01), with the SI level having the smaller mean.
Hypothesis 2.1b predicted that (d.! |I1) would not differ significantly
from (d! |I2). This hypothesis was tested using a Z test, and the
difference was found to be not siqnificant (Z=0.4588, n.s. two-tailed).
Hypothesis 2.1c predicted that (d! |C1) would not differ significantly
from (dl |C2). A Z test indicated that the difference was not
significant (Z=0.3167, n.s. two-tailed).
Relative Decision Model Sensitivity
Hypotheses 2.2 and 2.3 relate to the adjusted relative deviation
of the individual's decision model sensitivity from the optimal model
sensitivity. The adjustment is concerned with the elimination of the

108
TABLE 9
MODEL COMPARISON PROCEDURE RESULTS
FOR THE d'.
i
DEPENDENT
VARIABLE
Model
SSe
d.f.(F)
F
Full5
5.8767
11,74
4.66a
Without 3-way interaction
5.8926
1,74
0.20
Without distribution x cost
interaction
5.9032
1,75
0.13
Without information x distribu
tion interaction
5.8947
1,75
0.03
Without information x cost
interaction
5.8969
1,75
0.05
Without information0
5.9950
1,78
0.59
Without distribution0
9.4035
1,78
46.10a
Without cost0
5.9304
1,78
0.26
V.oi
bThe F value associated with the full
model is the
test of
the full
model and not a model comparison test,
c
The test of this model assumes the effects of those interactions
that include this variable are equal to zero.

109
TABLE 10
PARAMETER VALUES ASSOCIATED WITH THE REDUCED d! MODEL
i
F Values Associated With the
Sources of the Reduced d1 Model
Source
d.f.
F
Distribution
1,80
46.52a
Intrinsic motivation
1,80
10.33a
Extrinsic (non-monetary)
motivation
1,80
1.80
Extrinsic (monetary)
motivation
1,80
0.06
GPA
1,80
0.20
Means and Variances Associated
With the Significant Sources
Variable Level N
Mean
Variance
SI 44
1.2325
0.09165
S2 42
1.6093
0.07200
p<.01

no
effects of multiple cutoff values on individual decision model sensi
tivity. Assuming random subject assignment to experimental situations,
the Dfij and DNA. variables should be independent of the effects of the
d' absolute magnitude differences (within the distribution variable).
The remaining effects are 1) the relative variable response range
within the DNA^ measure, and 2) the relative variable response range
and multiple cutoff value usage within the DN.¡ measure.
The methods of analysis were the model comparison procedures where
the dependent variables a"e the DN.¡, DNA^ and DNA^-DN- measures. The
independent variables are the three situation variables (with their
interactions), the three motivation factors, and the GPA.¡ variable.
The three full models are
DNi
DNAi u + j + 8k + ^ + 6jk + Tj! + + a6Yjk] + 5,
DNA--DN- + t. + X. +6.
ill
The three model comparison procedure results are presented in
Table 11 and the F values associated with the sources of the reduced
models are presented in Table 12. The results indicate that the reduced
DN.. model contains the distribution variable, the cost variable, and
the intrinsic motivation factor. Both the distribution variable main
effect and the motivation factor are significant at the p<.01 level,
and the cost variable is significant at the p<.05 level. The reduced
DNA. model contains the cost variable and the intrinsic motivation
factor. The cost variable main effect is significant at the p<.01
level and the motivation factor is significant at the pc.10 level.
The reduced DNA..-DN.. model contains the distribution variable and the
intrinsic motivation factor. The distribution variable main effect is

TABLE 11
MODEL COMPARISON PROCEDURE RESULTS FOR THE DNi, DNAi, AND DNAi-DMi DEPENDENT VARIABLES
DNi
DNAi
DNAi-DMi
Model
SSe
d.f.(FT~
F
SSe
d.TTUT
F
SSe
d.f.(F)
F
Ful 1d
2.1852
11,74
4.66a
1.1753
11,74
2.03b
1.4200
11,74
0.97
Without 3-way interaction
2.2031
1,74
0.61
1.1758
1,74
0.03
1.4327
1,74
0.66
Without distribution x
cost interaction
2.2236
1,75
0.70
1.2105
1,75
2.22
1.4345
1,75
0.10
Without information x dis
tribution interaction
2.2032
1,75
0.00
1.1783
1,75
0.16
1.4343
1,75
0.09
Without information x
cost interaction
2.2123
1,75
0.31
1.1758
1,75
0.00
1.4412
1,75
0.45
. e
Without information
2.2324
1,78
0.00
1.2140
1,78
0.01
1.4453
1,78
0.01
. 6
Without distribution
2.4330
1,78
7.01a
1.2269
1,78
0.84
1.5566
1,78
6.01b
. e
Without cost
2.3748
1,78
4.98b
1.4522
1,78
15.31a
1.4575
1,78
0.66
V.Ol bp<.05 Cp<.10
^The F value associated with the full model is the test of the full model and not a model comparison test,
e
The test of this model assumes the effects of those interactions that include this variable are equal to
zero.

TABLE 12
PARAMETER VALUES ASSOCIATED WITH THE REDUCED DNj, DNA-j, AND DNAj-DN-j MODELS
Source
F Values Associated
With
the Sources of the Reducec
1 Models
-DNi
DNi
DNAi
DNAi
d.f.
F
d.f.
F
d.f.
F
Distribution
1,79
7.10a
1,80
6.17 b
Cost
1,79
5.05b
1,80
15.52 a
Intrinsic motivation
1,79
8.64a
1,80
3.55C
1,80
3.59C
Extrinsic (non-monetary) motivation 1,79
1.68
1,80
1.38
1,80
0.24
Extrinsic (monetary) motivation 1,79
0.04
1,80
0.07
1,80
0.01
GPA
1,79
0.02
1,80
0.14
1,80
0.27
Means and Variances
Associated With
the Siqnificant
Variables
DNi
DNAi
DNAi-DNi
Variable Level N
Mean Variance
N
Mean
Variance
N
Mean
Variance
SI 44
0.8052 0.0391
44
0.9193
0.0183
44
0.1141
0.0249
S2 42
0.8845 0.0240
42
0.9329
0.0180
42
0.0484
0.0113
Cl 47
0.8789 0.0276
47
0.9726
0.0091
47
0.0938
0.0197
C2 39
0.8017 0.0369
39
0.8696
0.0234
39
0.0679
0.0185
9P<.01 bp<.05 Cp<.10

113
significant at the pc.05 level, and the motivation factor is significant
at the p<.10 level.
Means, variances, and sample sizes of the three dependent variables
given the levels of the significant situation variables are included in
Table 12. Using the F test of equal variances, the following variances
were found to differ significantly: 1) within the DN.¡ variable between
the levels of the distribution variable (F=1.63, p<.10), 2) within the
DNA.j variable between the levels of the cost variable (F=2.57, pc.01),
and 3) within the DNA^-DN^ variable between the levels of the distribu
tion variable (F=2.20, pc.Ol). The Z test of equal means indicated that
the means of the variable levels differ significantly (pc.Ol) for each
of the situation variables found to be significant by the above sources
F values (see Table 12).
Hypothesis 2.2a predicted that (DNA-jll) would not differ signifi
cantly from (DNA.j|I2). A Z test v/as employed to test this hypothesis
and the difference was found to be not significant (Z=0.0010, n.s.
two-tailed). Hypothesis 2.2b predicted that (DNA-jSl) would not differ
significantly from (DAijS2). Again, a Z test indicated that the
difference was not significant (Z=0.4662, n.s. two-tailed). Hypothesis
2.2c predicted that (DNA^¡Cl) would not differ significantly from
(DNA^|C2). The results of the model comparison procedure and Z test
indicated that the difference was significant at the pc.Ol level, with
the Cl level having the greater mean.
Hypothesis 2.3a predicted that (DNA-DN^- jSl) would be significantly
smaller than (DNA-¡-DN.¡ |S2). Although the results of the model comparison
procedure and Z test indicated that there was a significant difference
between the means, the S2 distribution level had the smaller mean.

114
Hypothesis 2.3b predicted that (DNA--DN-|Cl) would be significantly
smaller than (DNA--DN-|C2). A Z test was employed to test this hypothesis
and the difference was found to be not significant (Z=0.8642, n.s.
one-tailed).
Intrinsic Motivation and Decision Model Sensitivity
The model comparison procedure indicated that the intrinsic moti
vation factor had a significant effect on each of the four individual
decision model sensitivity dependent variables.1 The analysis of these
effects requires additional tests. A product-moment correlation was
computed for the intrinsic motivation factor with each of the four
dependent variables. The results of these correlations are presented
in Table* 13. The weakness of these correlations can be expected given
the relatively low explained variance associated with the various in
trinsic motivation measure sums of squares.
The positive association between intrinsic motivation and the d'
i
measure indicate that as the d1 measure increased in size (approached
i
the d1 measure or used a single cutoff value) the subjects' intrinsic
k
motivation measures increase. This result was confirmed by the posi
tive association between the motivation factor and the DN. measure.
As either the variable response range decreases or the use of multiple
cutoff values decreases (both are associated with increases in the DN-
measure) the subjects' intrinsic motivation measures increase. The
substantial drop in the level of significance for the association be
tween the intrinsic motivation and the DNA^ measures would suggest that
^ee the last section in this chapter for a definition and for additional
analyses concerning the intrinsic motivation factor.

115
TABLE 13
CORRELATION COEFFICIENTS FOR INTRINSIC MOTIVATION WITH
INDIVIDUAL DECISION MODEL SENSITIVITY VARIABLES
Model
Intrinsic Motivation
d!
i
0.2045 (.059)
DN-j
0.2549 (.018)
DNA-¡
0.1724 (.113)
DNAj-DN-¡
-0.1672 (.124)
Note: The numbers within the parentheses are the exact a-levels
associated with the rejection of the null hypothesis that the
correlation coefficient is equal to zero.

116
the association between the motivation factor and individual decision
model sensitivity is only partially a result of the variable response
range (the remaining effect within the DNA^ measure). This drop in the
significance level could be explained by the negative association (albeit
an equally low significance level) between the motivation factor and
the DNA^-DN^ measure. As the use of multiple cutoff values decreases
(DNA^-DN^ decreases) the subjects' intrinsic motivation measures
increase.
Individual Decision Criteria
The following sections report the analyses and results of the
individual decision criteria variables. The variables include the
relative decision criteria, the relative decision criteria conservatism,
and the relationships between individual attributes and decision criteria.
Relative Decision Criteria
Hypotheses 3.1 and 3.2 relate to the adjusted relative individual
decision criteria. The adjustment concerns the elimination of the
effects of multiple cutoff values on individual decision criteria. The
variables which affect the adjusted relative decision criteria should be
those related to the individual's selection of a cutoff value. Of the
three situation variables the one most related to this selection is
the cost variable.
The method of analysis was the model comparison procedure where
the dependent variables are the BN.¡, BNA^, and BNA-BN- measures. The
independent variables are the three situation variables (with their

117
interactions), the three motivation factors, and the GPA^ variable. The
three full models are:
- p + aj + Bk + Yi + 0Sjk + ctYj, + eykl + aSYjkl + £j
+ Ti + xi + 6i
The three model comparison procedure results are presented in
Table 14, and the F values associated with the sources of the reduced
models are presented in Table 15. The results indicate that the reduced
BN.j model contains the distribution variable, the cost variable, and
the extrinsic (monetary) motivation factor. Both the cost variable main
effect and the motivation factor are significant at the p<.01 level,
and the distribution variable main effect is significant at the p<.10
level. The reduced BNA- model contains the cost variable and the
extrinsic (monetary) motivation factor, and both are significant at the
p<.01 level. The reduced BNABN.¡ model contains the cost variable and
the GPA.j variable. The cost variable main effect is significant at the
pc.01 level, and the GPA.¡ variable is significant at the p<.10 level.
The means, variances, and sample sizes for the dependent variables
given the levels of the significant situation variables are included in
Table 15. The F test of equal variances indicated that the following
variances differ significantly: 1) within the BN- variable between the
levels of the distribution variable (F=1.51, p<.10), 2) within the BN-
variable between the levels of the cost variable (F=6.11, p<.01),
3) within the BNA^ variable between the levels of the cost variable
(F=9.36, p<.01), and 4) within the BNA BN^ variable between the levels
of the cost variable (F=39.74, p<.01). Using a Z test of equal means,
the means of the various levels were found to differ significantly
BNAi
BNA,-BN,

TABLE 14
MODEL COMPARISON PROCEDURE RESULTS FOR THE BN j, BNA -j, AND BNA-j-BN j DEPENDENT VARIABLES
BN i
BNAi
BNAi-BNi
Model
SSe
d.f.(F)
F
SSe
d.f.(F)
F
SSe
d.f.(F)
F
Ful 1 d
26.857
11,74
12.35 9
40.102
11,74
10.76a
8.299
11,74
2.47b
Without 3-way interaction
27.387
1,74
1.46
40.107
1,74
0.01
8.716
1,74
3.72C
Without distribution x
cost interaction
27.443
1,75
0.15
40.226
1,75
0.23
8.727
1,75
0.09
Without information x dis
tribution interaction
27.491
1,75
0.29
40.222
1,75
0.22
9.137
1,75
3.62C
Without information x
cost interaction
27.471
1,75
0.23
40.114
1,75
0.02
8.773
1,75
0.49
Without information6
28.415
1,78
2.23
41.707
1,78
2.61
9.258
1,78
0.43
Without distribution6
28.840
1,78
3.43C
41.359
1,78
1.94
9.217
1,78
0.09
g
Without cost
70.389
1,78
120.73a
95.941
1,78
107.44a
10.562
1,78
11.48a
3p<.01 bp<.05 Cp<.10
^The F value associated with the full model is the test of the full model and not a model comparison test,
e
The test of this model assumes the effects of those interactions that include this variable are equal to
zero.

TABLE 15
PARAMETER VALUES ASSOCIATED WITH THE REDUCED BN-j, BNA-¡, AND BNA-¡-BN-¡ MODELS
F Values Associated With the Sources of the Reduced Models
Source
BN i
BNAi
BNA
i-BNi
d. f.
F
d.f.
F
d.f.
F
Distribution
1,79
3.29
Cost
1,79
120.57a
1,80
105.60a
1,80
a
11.91
Intrinsic motivation
1,79
0.33
1,80
0.26
1,80
2.00
Extrinsic (non-i
nonetary) motivation 1,79
0.02
1,80
0.08
1,80
0.01
Extrinsic (monetary) motivation
1,79
10.16a
1,80
a
7.24
1,80
0.17
GPA
1,79
0.49
1,80
2.11
1,80
c
3.86
Means and Variances
Associated Witt
i the Significant
Variables
BN i
BNAi
BNAi-BNi
Variable Level
N
Mean
Variance
N
Mean
Variance
N
Mean
Variance
SI
44
1.0458
0.71977
44
1.1873
0.93408
44
0.0968
0.06872
S2
42
1.2272
1.08552
42
1.3572
1.54782
42
0.1280
0.20420
Cl
47
1.7777
0.63707
47
2.0046
0.96602
47
0.2269
0.21259
C2
39
0.3591
0.10422
39
0.3853
0.10326
39
0.0263
0.00535
p<.01
p<.05
p<. 10

120
(px.01) for each of the situation variables found to be significant by
the above sources F values. The sole exception was the distribution
variable within the BN^ measure. The means of the distribution levels
within the BHj variable do not differ significantly (Z=0.8830, n.s.
two-tailed).
Hypothesis 3.1a predicted that (BNA^jll) would not differ signifi
cantly from (BNA-112). A Z test indicated that the difference was not
significant (Z=0.7891, n.s. two-tailed). Hypothesis 3.1b predicted that
(BNA.|SI) would not differ significantly from (BNA^|S2). Again, a
Z test showed the difference to be not significant (Z=0.7048, n.s.
two-tailed). Hypothesis 3.1c predicted that (BNA-|Cl) would be signifi
cantly larger than (BNA-|C2). The model comparison procedure and Z test
indicated a significant difference at the p<.01 level, with the Cl level
having the larger mean. Hypothesis 3.2 predicted that the variance of
(BNA. |Cl) would be significantly smaller than the variance of (BNA1-|C2).
The F test of equal variances resulted in a finding of a significant
difference at the p<.01 level, with the C2 level having the smaller
variance.
Hypothesis 3.4a predicted that (BNA^-BN^|S1) would be signifi
cantly smaller than (BNABN.¡ |S2). A Z test of equal means indicated
that the difference was not significant (Z=0.3892, n.s. one-tailed):
however, the obtained direction is the same as the predicted direction
of this effect. Hypothesis 3.4b predicted that (BNAi-BNi¡C1) would be
significantly smaller than (BNA BN .j|C2). The model comparison pro
cedure and Z test indicated that the difference was significant at the
p<-01 level, with the C2 level having the smaller mean.

121
Relative Decision Criteria Conservatism
Hypothesis 3.3 relates to the effect of conservatism on the
relative individual decision criteria. The hypothesis proposes that
the greater the difference between the individual's initial decision
anchor and the optimal cutoff value, the higher the level of conser
vatism. The situation variable most related to such a difference is
the cost variable.
The method of analysis was the model comparison procedure where
the dependent variable is the BNC-¡ measure and the independent vari
ables are the three situation variables (with their interactions), the
three motivation factors, and the GPA-j variable. The full BNCj model
is:
BNCi = p + aj + ek + Y, + aSjk + Yji + BYk) + aSYjkl + Ej
+ t, + X, + ,
The model comparison procedure results are presented in Table
16, and the F values associated with the sources of the reduced BNCi
model are presented in Table 17. The results indicate that the re
duced BNCn- model contains the cost variable, the extrinsic (monetary)
motivation factor, and the GPA.¡ variable. Both the cost variable
main effect and the GPA^ variable are significant at the p<.05 level.
The extrinsic (monetary) motivation factor is significant at the
p<.01 level.
The means, variances, and sample sizes of the BNC^ measure given
the levels of the cost variable are included in Table 17. Using the
F test of equal variances, the variances of the BNC^ given the levels
of the cost variable were found to differ significantly (F=9.36, p<.01).

122
TABLE 16
MODEL COMPARISON PROCEDURE RESULTS FOR THE BNC-j DEPENDENT VARIABLE
Model
SSe
d.f.(F)
F
b
Full
38.249
11,74
a
2.35
Without 3-way interaction
38.510
1,74
0.51
Without distribution x cost
interaction
39.291
1,75
1.52
Without information x distribu
tion interaction
38.543
1,75
0.06
Without information x cost
interaction
38.624
1,75
0.22
c
Without information
39.895
1,78
0.91
c
Without distribution
39.607
1,78
0.34
c
Without cost
42.778
1,78
a
6.61
a
p<.05
b
The F value associated with the full
model is the
test of
the full
model and not a model comparison test.
The test of this model assumes the effects of those interactions
that include this variable are equal to zero.

123
TABLE 17
PARAMETER VALUES ASSOCIATED WITH THE REDUCED BNC-¡ MODEL
F Values Associated With the
Sources of the Reduced Model
Source
d.f.
F
Cost
1,80
6.88b
Intrinsic motivation
1,80
2.61
Extrinsic (non-monetary)
motivation
1,80
0.04
Extrinsic (monetary)
motivation
1,80
7.79a
GPA
1,80
4.90b
Means and Variances
Associated With the Significant Sources
Variable Level N
Mean
Variance
Cl
47
1.0046
C2
39
0.6147
0.96602
V.Ol
bp<.05
0.10326

124
The Z test of equal means indicated that the means of the cost levels
differ significantly (Z=2.5598, p<.05 two-tailed), with the C2 level
having the smaller mean.
Hypothesis 3.3a predicted that (BNC^|I1) would not differ signifi
cantly from (BNC.¡ 112). A Z test indicated that the difference was not
significant (Z=0.6802, n.s. two-tailed). Hypothesis 3.3b predicted that
(BNC^ | SI) would not differ significantly from (BNC.j|S2). Again, a Z
test indicated that the difference was not significant (Z=0.3295, n.s.
two-tailed). Hypothesis 3.3c predicted that (BNC^|Cl) would be signifi
cantly smaller than (BNC1-|C2). The model comparison procedure and Z
test indicated that the difference was significant at the p<.05 level,
with the C2 level having the smaller mean.
%
Individual Attributes and Decision Criteria
-
The model comparison procedure results indicated that the
extrinsic (monetary) motivation factor and the GPA^ variable had signifi
cant effects upon the various individual decision criteria variables.2
The analysis of these effects also required additional tests. A product
moment correlation was computed for the motivation factor and the GPA.¡
variable with each of the decision criteria dependent variables. The
results of these correlations are presented in Table 18.
The negative association between extrinsic (monetary) motivation
and the BN^ variable indicates that as the BNj measure increases the
subjects' extrinsic (monetary) motivation measures decrease. The BN^
2See the last sections in this chapter for definitions and for additional
analyses concerning the extrinsic (monetary) motivation factor and the
GPA.j variable.

125
TABLE 18
CORRELATION COEFFICIENTS FOR EXTRINSIC (MONETARY) MOTIVATION AND
SUBJECT GPA WITH INDIVIDUAL DECISION CRITERIA VARIABLES
Model
Extrinsic (Monetary)
Motivation
Subject GPA
BN-¡
-0.1975 (.068)
-0.0468 (.669)
BNA-j
-0.1916 (.077)
-0.1000 (.360)
BNA-j-BN-j
-0.0564 (.606)
-0.1896 (.080)
BNCj
-0.2951 (.006)
-0.2346 (.030)
Note: The numbers within the parentheses are the exact a-levels
associated with the rejection of the null hypothesis that the
correlation coefficient is equal to zero.

126
measure approaches infinity as the distance between the frj measure and
the 3^ measure increases. A similar association exists between the
motivation factor and the BNA-j measure (the BNA-j measure eliminates the
effects of multiple cutoff values). The BNA^-BN^ variable, which
measures the effect of multiple cutoff values upon individual decision
criteria, has no significant association with the motivation factor. The
BNC.J variable, which measures the effect of conservatism upon individual
decision criteria, has a stronger negative association with the moti
vation factor. As the level of conservatism increases (BNC.¡ increases)
subjects' extrinsic (monetary) motivation measures decrease.
The negative association between the GPA.¡ variable and the BNA^-BN^
variable indicates that as the BNA--BN.¡ variable increases the subjects'
GPA.¡s decrease (the BNA^-BN^ measure increases as multiple cutoff value
usage increases). The negative association between the GPA.¡ variable
and the BNC.¡ variable indicates that as the subjects' GPA^s decrease
the level of conservatism increases.
Individual Long-Run Decision Efficiency
The following sections report the analyses and results of the
individual long-run decision efficiency variables. The variables
include the relative decision costs, the relationships between individual
attributes and long-run decision efficiency, and the relationships be
tween long-run decision efficiency and training phase performance.
Relative Decision Costs
Hypotheses 4.1, 4.2, and 4.3 relate to the decision costs incurred
by the individuals relative to the decision costs incurred by the optimal

127
models. Both the stated objective function and subject payoff function
involved the minimization of the decision costs. A subject's long-run
decision efficiency was measured in terms of his minimization of his
decision costs relative to those of an optimal model.
The method of analysis was the model comparison procedure where
the dependent variables are the GP.¡, and GN.¡ measures. The in
dependent variables are the three situation variables (with their
interactions), the DN.¡ variable, the three motivation factors, and the
GPA.j variable. Ideally, both the DN.¡ and BN^ variables should be
included in the model: however, the BN.¡ variable had significantly
greater association with the other independent variables than did the
DN.j variable (the R2 for the full BN.¡ model was 0.6474, whereas the
R2 for the full DN^ model was 0.2192). Inclusion of the BN.¡ variable
would produce a significant problem of multi colinearity. The three
full models are:
Gi
GPi = y + j + Bk + Y1 + a6jk + fjl + 6Ykl + a6Yjkl + *1 +
GN-- + t. + X. +6.
1 ill
The three model comparison procedure results are presented in
Table 19, and the F values associated with the sources of the reduced
models are presented in Table 20. The results indicate that the re
duced G. model contains the distribution by cost interaction and the
DN. variable. Both the interaction and the DN. variable are siqnifi-
l i a
cant at the p<.01 level. The reduced GP.. model contains the distribu
tion by cost interaction, the DN. variable, the intrinsic motivation
factor, the extrinsic (monetary) motivation factor, and the GPA.
variable. The interaction, extrinsic motivation factor, and GPA.
l
variable are significant at the p<.05 level. The interaction,

TABLE 19
MODEL COMPARISON PROCEDURE RESULTS FOR THE G-¡, GP-¡, AND GN-¡ DEPENDENT VARIABLES
Gi
GPi
GNi
Model
SSe
d. f. ( F)
F
SSe
dTfTFT
F
SSe
-error-
F
c
Full
0.0809
12,73
12.08a
0.1824
12,73
a
7.46
0.0349
12,73
a
2.67
Without 3-way interaction
0.0811
1,73
0.17
0.1824
1,73
0.00
0.0352
1,73
0.52
Without distribution x
cost interaction
0.0895
1,74
a
7.71
0.1953
1,74
b
5.21
0.0356
1,74
0.96
Without information x dis
tribution interaction
0.0812
1,74
0.09
0.1824
1,74
0.01
0.0352
1,74
0.06
Without information x
cost interaction
0.0814
1,74
0.26
0.1826
1,74
0.07
0.0352
1,74
0.03
d
Without information
0.0826
1,76
1.05
0.1835
1,76
0.36
0.0357
1,77
0.05
d
Without distribution
0.0360
1,77
0.71
d
Without cost
0.0369
1,77
2.60
a b
p<.01 p<.05
c
The F value associated with the full model is the test of the full model and not a model comparison test,
d
The test of this model assumes the effects of those interactions that include this variable are equal to
zero.

TABLE 20
PARAMETER VALUES ASSOCIATED WITH THE REDUCED G.¡, GPi, AND GNj MODELS
F Values
Associated
! With the
Sources of
the Reduced
Models
Source
Gi
GPi
GNi
d.f.
F
d.f.
F
d.f.
F
Distribution
1,70
0.51
1,70
0.77
Cost
1,70
2.56
1,70
0.11
Distribution x cost
1,70
7.92a
1,70
5.39b
DNi
1,70
88.76a
1,70
46.39a
1,80
0.90
Intrinsic motivation
1,70
2.40
1,70
3.95C
1,80
5.00b
Extrinsic (non-monetary) motivation 1,70
0.00
1,70
0.01
1,80
0.05
Extrinsic (monetary) motivation
1,70
1.71
1,70
4.33b
1,80
7.59a
GPA
1,70
2.39
1,70
5.93b
1,80
9.82a
Means and
Variances
Associated
With the
Significant
Variables
... Gi
GP
j
Variable Level
N
Mean
Variance
N
Mean
Variance
si,
Cl
24
0.0504
0.00229
24
0.0987
0.00456
SI,
C2
20
0.0921
0.00371
20
0.1372
0.00538
S2,
Cl
23
0.0430
0.00097
23
0.0886
0.00297
S2,
C2
19
0.0575
0.00375
19
0.0934
0.00576
3p<.01
V.05
Cp<.10
ro
CO

130
extrinsic motivation factor, and GPA^ variable are significant at
the p<.05 level. The DN^ variable is significant at the p<.01 level
and the intrinsic motivation factor is significant at the p<.10 level.
The reduced GN.¡ model contains the intrinsic motivation factor, the
extrinsic (monetary) motivation factor, and the GPA^ variable. Both the
extrinsic motivation factor and the GPA.¡ variable are significant at the
pc.Ol level and the intrinsic motivation factor is significant at the
p<.05 level.
The means, variances, and sample sizes for the distribution by
cost interaction within the G^ and GP^ models are included in Table 20.
Figure 7 presents a graphic representation of both distribution by
cost interactions. Bartlett's test for homogeneity of variances in
dicted that the variances within the G^ model distribution by cost
interaction are significantly heterogeneous (x2=l0.66, p<.05 with
3 d.f.). The variances within the GP- model distribution by cost inter
action are considered homogeneous (x2=2.52, p<.25 with 3 d.f.).
Duncan's multiple range test of equal means indicated that the means
within the G^ model distribution by cost interaction form two groups:
the means within a group do not differ significantly (p=.01) but a
significant difference (p=.01) exists for the means between groups.
These groups are 1) (G.|S1,C2) and (G.¡ |S2,C2), and 2) |S2,C1),
(G\j |SI,C1), and (G.|S2,C2).
Hypothesis 4.1 predicted that the variance of (G^|C1) would be
significantly smaller tha n the variance of (G.j|C2). Using the F test
of equal variances, the variances differ significantly (F=2.43, pc.Ol),
with (G^ |Cl) having the smaller variance. Hypothesis 4.3 predicted a
significant distribution by cost interaction within the G- model in

FIGURE 7
DISTRIBUTION BY COST VARIABLE INTERACTION WITHIN
G1 AND GPi MODELS

132
0.03 1 J-
SI S2
Distribution
Variable
0.13
0.12
0.11
GPi 0.10
0.09
0.08
0.07
C2
Cl
h
SI S2
Distribution
Variable
Cost
Variable
Cost
Variable

133
which the (G-j |S1 ,C1 )-(G.¡ |S2,C1) would be smaller than the (G^ [SI ,C2)-
(G-j |S2,C2). The results of the model comparison procedure indicated a
significant distribution by cost interaction (pc.Ol), with the
I SI ,C1)-(G-j ¡S2,C1) equal to 0.0074 and the (Gj |S1 ,C2)-(Gi |S2,C2)
equal to 0.0346. Hypothesis 4.2a predicted that (G^|C1) would be
significantly smaller than (G^|C2). A Z test of equal means indicated
that the means differ significantly (Z=2.4420, p<.01 one-tailed), with
the Cl level having the smaller mean. The lack of significance for
the cost variable main effect within the model comparison procedure
could be the result of the distribution by cost interaction. There
exists substantially less difference between the levels of the cost
variable when given S2 distribution level than when given the SI dis
tribution level. Hypothesis 4.2b predicted that (G.j|S2) would be
smaller than (G^|S1). A Z test of equal means indicated that the means
differ significantly (Z=1.7452, p<.05 one-tailed), with the S2 level
having the smaller mean. The lack of significance within the model
comparison procedure for the distribution main effect was predicted by
hypothesis 4.2b. Hypothesis 4.2c predicted that (G^|II) would not
differ significantly from (G^ j12). Using a Z test of equal means, the
difference was found to be not significant (Z=0.4528, n.s. two-tailed).
Hypotheses 4.4a and 4.4b relate to the ratio of the individual's
relative additional decision costs to the individual's relative addi
tional decision savings. The method of analysis was the model compari
son procedure where the dependent variable is the GP^/GN^ variable, and
the independent variables are the same as for the full G^ model. The
full GP.¡/GN.¡ model is:
Gpi/GN1 = + j + Bk + Y1 + aSjk + aYjl + 8Yk1 + a8Yjkl + +1 +
+ t + X. + 5.
1 l l

134
The model comparison procedure results are presented in Table 21,
and the F values associated with the sources of the reduced model are
presented in Table 22. The results indicate that the reduced GP-¡/GM¡
model contains the distribution by cost interaction and the DMj variable,
both significant at the p<-01 level.
The means, variances, and sample sizes of the distribution by
cost interaction are included in Table 22. Figure 8 presents a graphic
representation of the distribution by cost interaction. Bartlett's test
of homogeneity of variances indicated that the variances within the
distribution by cost interaction are homogeneous (x2=1.068, p<.25 with
3 d.f.). Duncan's multiple range test of equal means indicated that
the means within the distribution by cost interaction form two over
lapping groups: the means within a group do not differ significantly
(p=.01), whereas the means between groups do differ significantly (p=.01).
These groups are 1) (GPi/GNi|C2,S1), (GPi/GNi|C2,S2), and
(GPi/GNi|C1,S2); and 2) (GPi/GNi|C1 ,S1), (GPi/GNi|C1.52), and
(GPi/GNi|C2,S2).
Hypothesis 4.4b predicted a significant distribution by cost
variable interaction in which the (GP^/GN^ |C2,S2)-(GPi/GN1-1Cl ,S2) would
be smaller than (GPi/GNi|C2,Sl)-(GPi/GNi|C1 ,S1). The results of the
model comparison procedure indicated a significant distribution by cost
interaction (p<.01). The (GPV/GN^ |C2,S2)-(GP^/GN1- |C1 ,S2) was equal to
0.0618, and the (GPi/GNi|C2,S1)-(GPi/GNi|C1,S1) was equal to 1.1086.
Hypothesis 4.4a predicted that (GP^/GN^|Cl) would be significantly
smaller than (GP^/GNj¡C2). Using a Z test of equal means, the difference
between the means was found to be significant (Z=2.0109, p<.01
one-tailed), with the Cl level having the smaller mean. The lack of

135
TABLE 21
MODEL COMPARISON PROCEDURE RESULTS FOR THE GPi/GN1 DEPENDENT VARIABLE
Model
SSe
d.f.(F)
F
Ful 1b
66.265
12,66
6.91a
Without 3-way interaction
67.538
1,66
1.27
Without distribution x cost
interaction
76.352
1,67
8.74a
Without information x distribution
interaction
67.607
1,67
0.07
Without information x cost
interaction
69.242
1,67
1.69
C
Without information
70.493
1,69
1.17
ap<.01
bfhe F value associated with the full
model is the
test of
the full
model and not a model comparison test,
c
The test of this model assumes the effects of those interactions
that include this variable are equal to zero.

136
TABLE 22
PARAMETER VALUES ASSOCIATED WITH THE REDUCED GPi/GNi MODEL
F Values Associated With the Sources of the Reduced Model
Source
d.f.
F
Distribution
1,70
0.64
Cost
1,70
0.35
Distribution
1,70
8.85a
DNj
1,70
59.01a
Intrinsic motivation
1,70
0.11
Extrinsic (non-monetary) motivation
1,70
0.54
Extrinsic (monetary) motivation
1,70
0.09
GPA
1,70
2.54
Means and Variances Associated With
the Siqnif
icant Sources
Variable Level
N
Mean
Variance
SI, Cl
24
2.0189
1.42035
SI, C2
19
3.1275
2.12453
S2, Cl
19
2.3372
2.09340
S2, C2
17
2.3990
1.71284
ap<.01

137
GPi/GNi
Variable
FIGURE 8
DISTRIBUTION BY COST INTERACTION WITHIN THE GPi/GNi MODEL

138
significance for the cost variable main effect within the model com
parison procedure could be the result of the distribution by cost
interaction. The cost variable has a significant effect only when given
the SI distribution level.
Individual Attributes and Long-Run Decision Efficiency
The model comparison procedure results indicated that the DN-
variable, the intrinsic motivation factor, the extrinsic (monetary)
motivation factor, and the GPA variable had significant effects upon
various individual long-run decision efficiency measures.3 The
analysis of these effects required additional tests. A product-moment
correlation was computed for each of these individual attributes with
each of the dependent variables. The results of these correlations
are presented in Table 23.
The negative association between the variable and the DN-
variable indicates that as the relative net decision costs increase the
subjects' variable response ranges increase (DN- increases). The
same association exists between the GP.¡ variable and the DN.¡ variable:
as relative additional decision costs increase the subjects' variable
response ranges increase. The positive association between the GN^
variable and the DN.¡ variable indicates that as the relative additional
decision savings increase the subjects' variable response ranges decrease
(DN.j decreases). The GP^/GN^ variable and the DN^ variable were
associated negatively: as the relative additional decision costs
3See the last sections in this chapter for definitions and additional
analyses concerning these individual attributes.

139
TABLE 23
CORRELATION COEFFICIENTS FOR INDIVIDUAL ATTRIBUTES WITH
LONG-RUN DECISION EFFICIENCY
Individual Attribute
G
GP
GN
GP /GN
DN
-0.7616
(.001)
-0.6528
(.001)
0.1868
(.085)
-0.6613
(.001)
Intrinsic motivation
-0.2854
(.008)
-0.3044
(.004)
0.2399
(.026)
-0.1758
(.121)
Extrinsic (monetary)
motivation
-0.1071
(.327)
-0.1887
(.081)
0.3018
(.005)
0.0185
(.871)
GPA
-0.1737
(.110)
-0.2526
(.019)
0.3377
(.002)
0.0861
(.450)
Note: The numbers within the parentheses are the exact -levels
associated with the rejection of the null hypothesis that the
correlation coefficient is equal to zero.

140
increase at a rate greater than the relative additional decision savings
the subjects' variable response ranges increase.
The negative association between the G.¡ variable and the intrinsic
motivation factor indicates that as the relative net decision costs
increase the subjects' intrinsic motivation measures decrease. The same
association exists between the GP.¡ variable and the intrinsic motivation
factor: as the relative additional decision costs increase the subjects'
intrinsic motivation measures decrease. The GN.¡ variable and the in
trinsic motivation factor were associated positively: as the relative
additional decision savings increase the subjects' intrinsic motivation
measures increase.
Similar associations exist for the extrinsic (monetary) motivation
factor. Thus, as the relative additional decision costs increase (GP^
increases) the subjects' extrinsic (monetary) motivation measures de
crease. Conversely, as the relative additional decision savings in
crease (GN.¡ increases) the subjects' extrinsic (monetary) motivation
measures increase.
The negative association between the GP^ variable and the GPA-
variable indicates that as subjects' GPAs decrease relative additional
decision costs increase. Likewise, the positive association between the
GN.j variable and the GPA variable indicates that as subjects' GPAs in
crease relative additional decision savings increase.
Long-Run Decision Efficiency and Training Phase Performance
Hypothesis 4.5 predicted that those subjects in the PP classi
fication of training adjustment direction would have a signficantly
smaller mean G.¡ than those subjects in the MM classification. A t-test

141
was used to test this hypothesis. The (GL¡|PP) was 0.0536 and the
(Gj|MM) was 0.0780. The F test for equal variances indicated that
the variances were equal (F=1.6378, n.s.). The t-test for the
hypothesis of equal means indicated that the hypothesis could be
rejected at the p<.10 level (t=l.303 with 41 d.f.).
Hypothesis 4.6 predicted that those subjects in the mixed
classifications (PM and MP) of training adjustment direction would
have mean measures between the non-mixed classifications. The (G.¡|PM)
was 0.0603 and the (G^(MP) was 0.0577. Since the hypothesis did not
predict significance (and because of the results of hypothesis 4.5),
these means were not tested statistically. The directional prediction
of hypothesis 4.6 was supported.
Individual Attributes
The following sections report the analyses and results of two
individual attributes: subject grade point average and subject moti
vations. The objective of both analyses was the determination of the
significance of any between-group variance of these attributes.
Subject Grade Point Average
Grade point averages (GPAs) of subjects were analyzed to determine
whether this individual attribute varied between groups of subjects. A
systematic variance of this attribute between groups could bias the
interpretation of the effects of the independent variables on particu
lar observation variables. Two methods of analysis were used to test
the association of GPA with the independent variables: Duncan's
multiple range test and Bartlett's test of homogeneity of variance.

142
Duncan's multiple range test was used to determine if the sub
jects' mean GPAs differed between the eight independent variables con
ditions. Given a completely between-subject assignment these eight
groups represented the results of the experimental sampling plan. The
results indicate that there were not significant differences (p=.01)
between the mean GPAs of these eight groups. Bartlett's test of
homogeneity of variance was used to test the variances of the GPAs for
these eight groups. The results indicate that the hypothesis of homo
geneity of variance could not be rejected (x2=7.6906 with 7 d.f., p<.50).
Overall, the results indicate that between-group variance should
not be affected significantly by differences between subjects' GPAs.
This does not indicate, however, that within-group variance will not be
affected by the differences between GPAs. The possible effects of
within-group GPA variance were presented above.
Subject Motivations
The objective underlying the analysis of subject motivations was
similar to that relating to subject GPAs: i.e., the analysis was
performed to determine whether motivations varied between groups of
subjects. The initial stage of this analysis consisted of reducing the
ten motivation scale items into several motivation factors using the
technique of factor analysis. The subjects' responses to the ten
motivation scale items were factor analyzed using a varimax rotation for
a four factor solution. The four factor solution was chosen on the
basis of prior beliefs concerning the general dimensions (or sources)
of subject motivations: intrinsic motivation, extrinsic non-monetary
motivation, extrinsic monetary motivation, and general motivation not

143
clearly associated with any one of the previous sources. The rotated
factor pattern for the ten scale items is presented in Table 24.
Analyzing the eigenvalues of these factors, the sharpest break (or
decline in slope) occurred between the third and fourth factor. For
this reason only the first three factors are included within the
subject motivation analysis. The first three factors account for 62.1
percent of the variability in the motivation scale items.
The first motivation factor had heavy positive loadings on the
scale items "enjoy making the decisions," "satisfied with performance,"
and "enjoy overall experience of participating." The first motivation
factor was interpreted as basically measuring intrinsic motivation
stimulated by enjoyment of the experiment and the required task.
The second motivation factor had heavy positive loadings on the
scale items "desire to perform well," "desire to cooperate with the
experimenter," and "desire to contribute to research knowledge." This
factor was interpreted as basically measuring extrinsic motivation
stimulated by non-monetary sources.
The third motivation factor had heavy positive loadings on the
scale items "try less hard if money rewards halved" and "money rewards
caused you to try harder than without them." The third motivation
factor was interpreted as basically measuring extrinsic motivation
stimulated by monetary sources.
Using an orthogonal transformation matrix the individual scale
items were transformed into individual factor scores for each of the
first three motivation factors. Two methods of analysis were used to
test the association of subject motivation with the independent
variables: Duncan's multiple range test and Bartlett's test of homo
geneity of variance.

144
. TABLE 24
VARIMAX ROTATED MOTIVATION FACTOR PATTERN AMD DESCRIPTION OF
MOTIVATION SCALE ITEMS
Scale Item
Factor 1
Factor 2
Factor 3
Factor 4
1
0.1399
0.2336
-0.2112
-0.7551
2
0.2610
0.7485
0.0327
0.0503
3
0.8287
0.2336
-0.0712
-0.0359
4
0.8001
-0.1721
0.2190
-0.0836
5
-0.0305
0.0197
0.9037
0.1436
6
-0.1229
0.2524
-0.0876
0.8044
7
0.0231
0.8198
0.0104
0.0174
8
0.0814
0.1075
0.8789
-0.0385
9
0.7324
0.3174
-0.0963
-0.3140
10
0.0148
0.7586
0.0848
-0.0441
Eigenvalue
2.6975
1.8552
1.6548
0.9528
Cum Eigenvalue
Portion
0.2700
0.4550
0.6210
0.7160
Description
of Motivation
Scale Items
1 Number of
times willing
to participate
in the future.
2 Desire to perform well in undertaking a challenging task.
3 Enjoy making the decisions.
4 Satisfied with your performance.
5 Would have tried less hard if money rewards had been halved.
6 Feel tense during the experiment.
7 Desire to cooperate with the experimenter.
8 Money rewards caused you to try harder than without them.
9 Enjoy the overall experience of participation.
10 Desire to contribute to research knowledge.

145
Duncan's multiple range test was used to determine if the subjects'
motivations differed between the eight independent variables conditions.
The results indicate that there were no significant differences (p=.01)
in the mean motivation factors between the eight groups. Bartlett's
test of homogeneity of variance was used to test the variances of the
three motivation factors for these eight groups. The results indicate
that the hypothesis of homogeneity of variance could not be rejected
for the second and third motivation factors (x2=10.9095 with 7 d.f.,
p<.25; x2=l.796 with 7 d.f., pc.95). The hypothesis of homogeneity
of variance could be rejected at the pc.10 level for the first moti
vation factor (x2=l2.7829 with 7 d.f.).
Overall, the results indicate that between-group variance should
not be affected significantly by differences between the subjects'
motivations. Again, the analysis does not indicate that within-group
variance will be unaffected by the differences between subjects'
motivations: these possible effects were presented above.

CHAPTER VI
DISCUSSION OF RESULTS
Conceptual Development Revisited
A brief restatement of the conceptual development presented in
Chapter III will serve as a basis for discussion of the results that
were reported in Chapter V. It was posited that an individual's
long-run decision efficiency would be affected by 1) the structure
of the particular decision situation, 2) the contents of the available
information set, 3) the individual's efficiency in processing the
available information, and 4) the individual's ability to expand the
available information through experience with the particular situation.
A general heuristic was expected to systematically account for a por
tion of the decision behavior of the subjects. A heuristic, within
this context, refers to a learned set of rules or principles which are
utilized by individuals in making the particular decisions required of
them. Since a heuristic is a learned response to some decision task
stimuli, individual characteristics can have considerable effect upon
the specific form of the heuristic. The conceptual development, how
ever, proposed that specific decision task stimuli would give rise to
general forms of heuristics which would be (at least partially) in
dependent of individual characteristies.
The decision task utilized in this research required the discrim
ination between the uncertain existence of two states of nature
146

147
(in-control and out-of-control) for a given production labor process.
A random variable produced by the accountant provided evidence con
cerning state existance at particular points in time. This random
variable (the labor efficiency variance) had certain statistical
properties within the states of nature: the decision maker either
had or did not have knowledge of these particular properties. The
decision task was complicated by the existence of unequal values (costs)
associated with the various possible decision outcomes.
The form of the general heuristic proposed to describe decision
making behavior was that of anchoring and adjustment. Anchoring and
adjustment first requires the selection of an initial decision value
(anchor) from the continuum of possible decision values that are
located within the domain of the random variable (the labor efficiency
variance in the case of the present study). Once an initial decision
anchor is selected by the decision maker, training and experience
(learning) will lead to adjustments in the value of the anchor. Such
adjustments will be motivated primarily by feedback of various decision
performance measures. Should a particular decision performance measure
indicate less than satisfactory decision performance, the decision
maker will be motivated to adjust his decision anchor in an attempt to
improve decision performance measures in the future. The concept of
less than satisfactory decision performance assumes that the decision
maker has internalized the objective function from which the decision
performance measure was derived. Within this research the decision
performance objective function had a homomorphic relation with the
decision maker's payoff function.

148
Overall Results
The overall results generally are supportive of the conceptual
development. The major deviations from the hypotheses derived from the
conceptual development and the simulations are summarized in this
section. Plausible explanations, consistent with the observed results,
are developed for those situations where the results deviated from the
hypotheses. These explanations, ex post in nature, have not been tested
by this research: they are offered as potential modifications to the
existing conceptual development and can be tested by future research.
First, within situations where the state distributions had wider
separation and the adjustment process involved divergence from the
standard, subjects' initial decision anchors were biased toward the
standard. The extent of this bias was not large and as the adjustment
process continued the bias was reduced substantially relative to the
adjustment bias of subjects within the opposite situations.
Second, variable response ranges were larger within situations
where the adjustment process involved divergence from the standard, and
decision criteria conservatism was larger within situations where the
adjustment process involved convergence toward the standard. Although
these effects were contrary to those predicted, the concept of a sub
jective adjustment limit provides a plausible explanation. The concept
proposes that subjects will perceive the standard as a subjective
adjustment limit and will be hesitant to approach the standard. This
subjective limit will have an effect where the situation involves
adjustment toward the standard. If this phenomenon does exist it would
affect both the variable response range and the decision criteria con
servatism in the same manner as the obtained results.

149
Third, the obtained effects of the situation variables on the
relative decision costs were weaker than those predicted. The pre
dicted effects were based, in part, upon the assumption of equal
learning efficiency between the levels of the various situation vari
ables. However, as indicated earlier the obtained learning efficien
cies were not equal between various situation variable levels. A
plausible explanation for the weaker obtained effects of the situation
variables on the relative decision costs utilizes the unequal learning
efficiency results. That is, the incorporation of unequal learning
efficiencies within the prediction of the effects of the situation
variables on the relative decision costs will produce expected effects
with strengths similar to the obtained effects.
Finally, multiple decision anchor usage had greater effects upon
individual decision model sensitivity within situations where the state
distributions were closer together and had smaller variance. It had
greater effects on individual decision criteria within situations where
the decision costs favored making investigation decisions for smaller
(in absolute value) labor efficiency variances. Both of these results
are contrary to those predicted, and no explanations are offered.
Overall, the variable which had the largest impact on the relative
decision costs was the subjects' decision criteria conservatism. The
mean DNA^ measure over all subjects was 0.926. Considering that a
value of one would indicate variable response ranges were not used, the
subjects' overall decision model sensitiviy was approximately that of
the optimal model. The mean BNC^ over all subjects was 0.828. This
indicated that the average distance between the subjects' decision
criteria and the optimal model's decision criteria was 82.8 percent
of that of the optimal model. This was supported by the mean RA.¡ over

150
all subjects: the overall mean RA- was 0.641 which indicated that the
subjects adjusted their decision anchors only 35.9 percent (1-RA^) of
the distance between their initial decision anchors and the optimal
model's decision value. However, even given this level of overall deci
sion criteria conservatism, the mean relative decision costs (G.¡) over
all subjects was 0.06 (the average subject's total decision costs were
6.0 percent greater than the optimal model's total decision costs).
The deviation of the G.¡ measure from a value of zero has been
attributed to the following factors: 1) variable response ranges,
2) multiple decision anchors, and 3) decision criteria conservatism.
Although each factor was affected differently by various situation
variables and individual attributes, the overall effects of these
factors on the relative decision costs indicate their relative import
ance. The overall effect of variable response ranges was shown to be
negligible. Ignoring the variable response range effects, the overall
relative decision efficiency measure (G-) can be adjusted into sub
measures which are affected by only one of the two remaining factors.
Eliminating the effects of multiple decision anchors, the adjusted
Gj measure was 0.045. This indicated that of the total deviation of
the subjects' decision costs from those of the optimal models, approxi
mately 75 percent was due to individual decision criteria conservatism
and approximately 25 percent was due to multiple decision anchors.
Discussion of Results
The following discussion of results is organized according to
decision processes and dependent variables. The discussion includes
decision anchor adjustment, individual decision model sensitivity,

151
individual decision criteria conservatism, and individual long-run
decision efficiency.
Decision Anchor Selection and Adjustment
The first aspect of the anchoring and adjustment heuristic is the
selection of the initial decision anchor. The formation of the hypo
theses concerned with this aspect was based upon the assumptions that
1) individuals (within a given level of distribution variable) would
tend to use a common initial decision anchor, and 2) the location of this
common anchor would be related to a measure of central tendency between
the two states of nature. This central tendency was operationalized as
the geometric intersection of the state distribution curves (the inter
section was conditional upon the level of the distribution variable).
The results indicated that the expected point of central tendency
was obtained for those subjects within the SI distribution level, but
that the expected point was statistically larger than that obtained
for those subjects within the S2 distribution level. The state means
were farther apart within the S2 distribution level than within the SI
level. As a consequence, the intersection of the state distribution
curves was located farther from the standard (the mean of the in-control
state) within the S2 distribution level than within the SI level. An
initial subjective bias provides a plausible explanation for the
estimate obtained within the S2 level. This bias would take the form of
a tendency to place greater weight on the standard as the intersection
of the distribution curves became larger (i.e., shifted to a higher
point on the actual minutes incurred axis).

152
An effect which differed from expectations was that the mean
initial decision anchor given the C2 cost level was significantly
larger than the mean initial decision anchor given the Cl level
(equality was predicted). A bias of the estimates of the initial de
cision anchors may explain this difference. It will be recalled that
the estimates had incorporated some initial adjustment (during the
first training session). Those subjects within the C2 cost level were
expected to adjust their initial decision anchors such that the value
(actual minutes incurred) increased: those subjects within the Cl cost
level were expected to adjust in the opposite direction. Initial ad
justment in the appropriate directions would produce the obtained cost
variable effect on the estimates of the initial decision anchors.
The second aspect of the anchoring and adjustment heuristic is
the adjustment process. The adjustment process was examined using a
measure of the individual's relative decision anchor adjustment. This
observation variable, RA-, was the linear distance ever which the
individual adjusted his decision anchor (the distance between his final
decision anchor and the optimal decision value) relative to the linear
distance over which the individual should have adjusted his decision
anchor (the distance between his initial decision anchor and the
optimal decision value). The results of analysis indicated that the
RA.j measure did not differ significantly between the levels of the
various situation variables. This was consistent with the predictions
from the training phase simulation. A major limitation of the measure
is that it is confined to the relative linear movement along the random
variable decision axis. As such the RAj measure is not sensitive to
the learning of various situation parameters and variables (e.g., the

153
relative frequencies of the two states and the relative decision error
costs), nor is it sensitive to the effects of various individual attri
butes (e.g., the individual's variable response range). Possible
effects of the situation variables on these aspects of the adjustment
process are examined in greater detail ( with measures of greater
sensitivity) in later sections relating to the individual decision
criteria and the individual decision model sensitivity.
Individual Decision Model Sensitivity
Individual decision model sensitivity was measured employing a
parameter of the TSD model, d'. This parameter measures the decision
sensitivity of the individual's model relative to a theoretical sensi
tivity. The factors which can affect an individual's decision model
sensitivity are 1) the use of a variable response range, and 2) the
use of multiple decision anchors. To isolate'these factors three
decision sensitivity measures were developed: 1) the DN-j measure
(which is affected by both factors), 2) the DNAj measure (which is
affected by only the variable response range), and 3) the DNAj-DNj
measure (which is affected by only the multiple decision anchor).
The variable response range (the DNA-j measure) was affected
primarily by the cost variable. That is, those subjects within the C2
cost level demonstrated larger variable response ranges than did those
subjects within the Cl cost level. The locations of the optimal de
cision values and the subjects' final decision anchors were closer to
the standard for the Cl cost level than for the C2 cost level (see
Figure 2, Chapter III). The subjective adjustment limit concept offered

154
earlier may explain the difference in effect of the cost variable on
the DNA.: measure. When the adjustment process involved convergence
toward the standard the subjects may have perceived the standard as a
limit to their adjustment process, a limit which they could have been
reluctant to approach. When the adjustment process involved divergence
from the standard the subjects did not have an objective value to per
ceive as a limit to their adjustment process. Whether a subject's
adjustment process involved convergence toward or divergence from the
standard depended upon the location of his initial decision anchor
relative to the optimal decision value. This factor was conditional
upon the cost variable. Given the Cl level the subject's initial
decision anchor was greater than the optimal decision value and the
adjustment process involved convergence toward the standard. Given
the C2 level the opposite held: i.e., the subject's initial decision
anchor was less than the optimal decision value and the adjustment
process involved divergence from the standard. As subjects' adjustments
within the Cl level converged toward the standard the subjective
limit of the standard could have acted as an intervening variable which
reduced the relative magnitude of the variable response range. As
subjects' adjustments within the C2 cost level diverged from the stan
dard no such subjective adjustment limit existed: therefore, the
relative magnitude of the variable response range could have increased.
The higher (statistical) moments of the DNA^ measure given the
levels of the cost variable were consistent with this concept of a
subjective adjustment limit. Such a limit should have had the effect
of reducing the variance of this measure. The test of the variances
indicated that (DNA^JCl) had a significantly smaller variance than

155
(DNA.j|C2). The subjective limit also should have had the effect of
skewing the measure away from those values which indicated larger
variable response ranges. The third moment (as expressed by the co
efficient of skewness) of the DNA- measure indicated that 1) the
skewness of the (DNA^|C1) distribution was positive (skewed away from
values which indicated larger variable response ranges), and 2) the
skewness of the (DNA-|C2) distribution was negative (skewed toward
values which indicated larger variable response ranges).
The multiple decision anchor variable (the DNA^-DN^ measure) was
affected primarily by the distribution variable. Those subjects within
the SI distribution level employed multiple decision anchors to a
larger extent than did those subjects within the S2 level. The hypo
theses posited a relation between multiple decision anchors and the
%
distance separating the lower tail of the in-control state from^the
subject's final decision anchor. It was expected that the larger this
distance (found within the C2 cost level and the S2 distribution level)
the greater would have been the subject's propensity to have perceived
a second out-of-control state below the existing in-control state
(and used multiple decision anchors). However, the obtained results
do not support this proposed relation.
The observed associations between the individual decision model
sensitivity measures and intrinsic motivation have intuitive inter
pretations. The intrinsic motivation factor measured experimental
task and experimental environment enjoyment as modified by decision
performance satisfaction. As the subjects use of multiple decision
anchors and variable response ranges increased, their intrinsic moti
vation appeared to decrease. Since the use of multiple decision

156
anchors adversely affects decision performance (as reported by the feed
back measure used in this research), their increased use would affect
intrinsic motivation through the performance satisfaction modifier.
Greater use of multiple decision anchors would result in poorer decision
performance feedback measures: these measures, in turn, could lower
the subject's decision performance satisfaction and thereby lower his
intrinsic motivation. As the variable response range increases in
relative size the number of decisions falling within this range increases.
This requires more guesswork or more complicated decision rules on the
part of the subjects, thereby increasing the difficulty (or the uncer
tainty) of the decision task. This increased task difficulty could
affect the subject's enjoyment of the decision task and thus lower his
intrinsic motivation.
Individual Decision Criteria
Individual decision criteria were measured using a parameter of
the TSD model, 6. This parameter measures the decision criteria used
by the individual independent of his decision model sensitivity. The
factors which can affect an individual's decision criteria are 1) the
individual's efficiency in processing (or learning) the effects of the
states' relative frequencies and the relative decision error costs,
and 2) the use of multiple decision anchors. To isolate these factors
several decision criteria measures were developed: 1) the BN- measure
(which is affected by both factors), 2) the BNA.¡ and BNC^ measures
(which are affected by only the individual's information processing
efficiency), and 3) the BNA^-BM^ measure (which is affected by multiple
decision anchors).

157
Those subjects within the Cl cost level produced significantly
larger mean BNA.¡ measures than did those subjects within the C2 level.
However, the BNA- measure can not be interpreted strictly as a measure
of the efficiency of individual information processing: rather, it is
a function of such a measure, the BNC^ measure. This is due to
characteristics of the data structure and the decision task. The 3,-
parameter is related monotonically to the random variable decision
axis (actual minutes incurred). Consequently, for those subjects with
in the Cl cost level the 3.¡ measure associated with the initial decision
anchor is larger than the Bk measure, and the expected adjustment pro
cess has the effect of reducing the 3- measure. For those subjects
within the C2 cost level the 3.¡ measure associated with the initial
decision anchor is smaller than the 3^ measure, and the expected ad
justment process has the effect of increasing the 3^ measure. There
fore, where BNA- is defined as the ratio of the final 3- measure and
the 3^ measure, a value greater than one is produced for those subjects
within the Cl cost level and a value less than one is produced for those
subjects within the C2 cost level. The BNC^ measure eliminates this
systematic relation with the value of one.
The efficiency of information processing (learning) was affected
primarily by the cost variable. Those subjects within the Cl cost
level were significantly more conservative (less efficient) in pro
cessing the information than were those subjects within the C2 cost
level. These results were the reverse of those predicted by the hypo
theses. The subjective limit concept may again be introduced as an
explanation. As subjects' adjustments within the Cl cost level con
verged toward the standard the subjective adjustment limit of the

158
standard could have acted as an intervening variable which increased
the level of relative conservatism (reduced the level of relative infor
mation processing efficiency). As subjects' adjustments within the C2
cost level diverged from the standard no such subjective adjustment
limit existed, thus the level of relative conservatism could have de
creased.
The higher (statistical) moments of the BNC.¡ measure given the
levels of the cost variable were consistent with this concept of a
subjective adjustment limit. The difference in variances of the
BNC.j measure was due to the difference in the ranges of the measure.
The range of (BNC^|C1) was 4.67 and the range of (BNC.¡|C2) was 1.57.
The subjective adjustment limit should have had the effect of skewing
the measure away from those values which indicated lower levels of
conservatism. The third moment (as expressed by the coefficient of
skewness) of the BNC^ measure indicated that 1) the skewness of the
(BNC^. ¡C1) distribution was positive (skewed away from values which
indicated lower levels of conservatism), and 2) the skewness of the
(BNC^|C2) distribution was negative (skewed toward values which in
dicated lower levels of conservatism).
The multiple decision anchor (the BNA--BN^ measure) was affected
primarly by the cost variable. Those subjects within the Cl cost
level employed multiple decision anchors to a larger extent than did
those subjects within the C2 cost level. The hypotheses posited a
relation between multiple decision anchors and the distance separating
the lower tail of the in-control state from the subject's final de
cision anchor. This relation was similar to that proposed in the
DNA^-DN^ measure. However, the results did not support the hypotheses

159
and thus parallel those obtained with the DNA^-DN^ measure. Both the
DNA^-DN^ and the BNA^-BN^ measures obtained results that indicated
greater multiple decision anchor usage when the subjects' initial de
cision anchors were closer to the standard (the SI distribution level
within the DNA^-DN^ measure and the Cl cost level within the BNA^-BN^
measure). No plausible explanation for these results is readily
apparent.
The obtained associations between the individual decision criteria
measures and the individual attributes of extrinsic (monetary) moti
vation and subject GPA have intuitive interpretation. As both the
extrinsic (monetary) motivation measures and the subjects' GPAs de
creased the level of conservatism increased. Furthermore, as the sub
jects' GPAs decreased the use of multiple decision anchors increased.
Of the two individual process variables (decision model sensitivity
and decision criteria), the decision criteria had the more signifi
cant affect on a subject's performance feedback measure and, in turn,
on the level of his monetary rewards. However, since motivation
pre-tests were not administered,causal relationships (i.e., decreased
motivation led to increased conservatism, or increased conservatism led
to decreased motivation) can not be claimed.
Individual Long-Run Decision Efficiency
The overall efficiency of an individual's decision making within
the experiment was measured using the G.¡ variable. The G- variable was
the individual's net additional decision costs relative to the optimal
model's total decision costs. The G^ variable was disaggregated into
two measures- the total additional individual decision costs relative

160
to the optimal model's total decision costs (GP-j) and the total
additional individual decision savings relative to the optimal model's
total decision costs (GN-¡). The hypotheses concerning these variables
were derived using a simulation which assumed a set of initial decision
anchors and assumed equal learning efficiency (in terms of relative
linear adjustments along the random variable axis). Given these
assumptions, specific variations in the decision situation were ex
pected to have specific effects on a subject's relative decision costs
(see Table 4 in Chapter V for a summary of these expected effects and
the obtained results). All of these expected decision situation effects
were supported, at least in part, by the obtained results. However,
the results of the individual decision model sensitivity and the indi
vidual decision criteria analyses indicated that learning efficiency
was not equal between the levels of the situation variables.
The support obtained for the hypotheses concerning decision
efficiency variables would suggest that the unequal learning efficiency
did not have a significant effect on the relative decision costs. The
majority of the unequal learning efficiency occurred between the levels
of the cost variable (the most significant effect, decision criteria
conservatism, was greater within the Cl cost level than within the C2
level). A closer examination of both the simulated and the obtained
cost variable effects on the G-¡ variable indicated that the obtained
effects were not as strong as the simulated effects (note that (Gj|C2)/
(GilCl) equalled 1.61 whereas f(Gj|C2)/f(Gj|Cl) equalled 3.27). The
simulated effects may be adjusted for unequal learning efficiency by
assuming the relative decision costs within the Cl cost level would be
increased by a factor of two (relative to the C2 cost level). After

161
this adjustment the simulated cost variable effects would have the same
strength as the obtained effects (i.e., f (G-j | C2 )/2 f (G^ | Cl) would equal
1.60).
The simulated and obtained distribution by cost variable inter
actions involving the relative decision costs had similar relationships
between their strengths: the effects of the obtained interaction were
not as strong as the effects of the simulated interaction (note that
[(G.¡ |S1,C2)-(G.¡ |S2,C2)] /[(^ |S1 ,C1 )-(Gi |S2,C1)] equalled 3.15 whereas
[f(Gi|Sl,C2)-f(Gi |S2,C2)] / [f¡SI,C1)-f(Oj |S2,C1)] equalled 8.74).
Again, if the simulated effects are adjusted for unequal learning
efficiency, then the simulated distribution by cost variable inter
actions have a strength similar to that of the obtained effects (i.e.,
[f(Gn. |Sl,C2)-f(Gi ¡S2,C2)] / [2-f (G^-1 SI ,C1 )-2*f (G.¡ |S2,C1)] would equal
4.37).
The simulated and obtained distribution by cost variable inter
action within the GP^/GN^ measure had opposite relationships between
their strength: the effects of the obtained interaction were stronger
than the effects of the simulated interaction (note that [(GP^/GN^|C2,S1)-
(GPi/GNi|C1,S1)] / [(GPi/GNi|C2,S2)-(GPi/GNi |C1 ,S2)] equalled 17.94
whereas [f(GPi/GNi|C2,S1)-f(GPi/GMi|C1,S1)] / [f(GP1/GNi|C2,S2)-
f(GP^/GN^jCl,S2)] equalled 3.82). To explain the difference in effect
strengths requires reference to two variables: multiple decision an
chors and unequal learning efficiency. If the GP.¡/GN.¡ measures are
adjusted for the effects of multiple decision anchors the strenqth of
the obtained interaction effect is 2.89 (using the above interaction
strength ratio). This reverses the relationships between the effect
strengths: i.e., the obtained interaction strength becomes weaker

162
than the predicted strength. The effects of unequal learning efficiency
were 1) an expansion of the interval between the final decision an
chor and the optimal decision value for those subjects with greater
decision criteria conservatism, and 2) the reduction of the interval
between the final decision anchor and the optimal decision value for
those subjects with less decision criteria conservatism. As the in
terval associated with those subjects within the Cl cost level expanded
(this level had greater decision criteria conservatism), both the
(GP-¡¡C1) and the (GN-j |Cl) became larger. However, the GP-¡ measure
became larger at a greater rate than did the GN-¡ measure (3GP-¡/9GN-¡>l).
As the interval associated with those subjects within the C2 cost level
contracted (this level had less decision criteria conservatism), both
the (GPj|C2) and the (GNj |C2) became smaller. However, the GP-¡ measure
became smaller at a greater rate than did the GN.¡ measure (3GP1-/3GN1->1).
Consequently, the ratio (GP^/GN^|Cl) increased as the level of conser
vatism increased, and the ratio (GP^/GNJC2) decreased as the level of
conservatism decreased. Incorporating these effects of unequal learning
efficiency into the simulated distribution by cost variable interaction
lowered the predicted strength of such an interaction.
The observed associations between the individual long-run decision
efficiency measures and the individual attributes may be interpreted
with some degree of intuitive satisfaction. As the subjects' variable
response ranges increased the relative decision costs increased (either
as a result of increased additional decision costs or decreased addi
tional decision savings). Furthermore, as the relative decision costs
increased both the subjects' intrinsic and extrinsic (monetary) moti
vation measures decreased. A similar association existed between the

163
subjects' GPAs and the relative decision costs: as the subjects' GPAs
increased either additional decision costs decreased or additional de
cision savings increased. It would be reasonable to assume that both
the DN. measure and the subject GPA are modifying variables on a sub
ject's decision cost performance. Since pretests were not administered
it was not possible to state the specific relations between the moti
vation factors and a subject's relative decision cost performance.
Li mitations
The major limitations of this study involve two general aspects;
the subjects employed within the experiment and the experimental
environment itself. Although these aspects are discussed separately,
in many instances the limitations they impose overlap one another.
These limitations affect primarily the external validity of this study.
There are several possible limitations involving the experimental
environment. First, the precision of subject performance feedback
during the training phase could be a limitation. The results obtained
in this study might be modified substantially if such accurate feedback
was not employed. The lack of decision performance difference between
the levels of the available information variable could be a direct
result of this feedback: i.e., the accuracy of the performance feedback
and its relationship with the optimal model may have replaced the need
for such additional statistical information.
Second, the background of the subjects in relation to the ex
perimental task is a possible limitation. Although the subjects re
ceived training in the experimental task, the primary source of their
knowledge concerning standard cost variance investigation could come

164
from the college classroom. Consequently, if they were not taught
(within the classroom) that situations exist in which investigation
decision values are located relatively close to the standard, then
greater decision criteria conservatism within these situations could
be the result of the lack of such knowledge. This suggests the
possibility of an availability bias (Tversky and Kahneman, 1973).
However, to the extent a manager must learn from his own experiences,
such a bias could exist in the real world.
A final limitation (discussed within this section) involves the
selection of the levels of the situation variables. Several situation
variables were set at a single level, while others were manipulated at
multiple levels. The important point is that only specific combina
tions of variable levels were studied within this research whereas an
infinite number of combinations are possible. Different variable levels
and different variable manipulations would create a difference in the
experimental environment which could produce results other than those
obtained in this study.
These experimental environment limitations do not mean, however,
that the concepts proposed within this study have not passed a meaning
ful empirical test. The objectives of experiments involving theory
or concept testing, and the relations between the nature created in the
experiment and the nature existing in the world at large are discussed
by Zigler (1963):
What the experimenter is saying is that if such and such holds
in the real world because of the principles expounded in the
particular theory under investigation, then such and such should
hold in the world which the experimenter has created. This
transatability is what gives theoretical import to experiments
which involve phenomena which, taken in isolation, not only
appear picayune but seem to have little relationship with what
one observes in nature, (pp. 353-4).

165
Finding the proposed concepts to hold within the experimental world
does not validate the claim that they hold within the real world. On
the other hand, finding the proposed concepts not to hold within the
experimental world would cast serious doubt on whether they hold in the
real world. Positive experimental results are a form of negative
assurance: the concepts have passed an initial empirical test and
remain (albeit somewhat scarred in most instances) potential candidates
for explaining real world phenomena.
The major limitation involving the subjects employed within this
research is that they were college students, predominantly accounting
majors. The obtained results and the conclusions based upon these
results are not generalizable beyond some unspecified population of
which the subjects are representative.
The selection of the type of subject to be employed within an
experiment is not independent of the experimental environment. Within
this research principal interest lies in the behavior of operational
managers within real world environments. However, given the selection
of the laboratory experimentation method (which implies the use of a
surrogate experimental environment), the selection of the type of
subject depends upon the answer to the question, "Which subject is a
better surrogate for a manager in a real world environment, a know
ledgeable student or the manager himself?" There are, of course, no
clear answers to this question. However, several aspects of the
situation favor the use of a knowledgeable student. First, the high
degree of abstraction within the experimental environment could have
made the task appear trivial to a manager and could have elicited
behavior not consistent with the principal situation. Second, the

156
experimental environment is closer to that of a test-taking situation
which is a more natural environment for a student than for a manager.
Implications For Accounting
Two general implications of this research for accounting are
discussed. These implications are the value of additional information,
and the general standard setting process.
Value of Additional Information
The manipulation of the information variable involved the quan
tity of information contained in the available information set, specifi
cally the presence or absence of various distribution information items.
The results indicated that the information variable did not have a
significant effect upon the individual relative decision costs. The
mean relative decision cost of subjects within the reduced information
level did not differ significantly from that of subjects within the
expanded information level. An implication of this result concerns the
net benefit (for the company) of providing the additional information
within the expanded information level. The additional information
within an actual environment is not costless. Consequently, if such
information is to be provided, other things being equal, the net
benefit of such an action should be positive. In this context, the net
benefit would be the marginal decision cost savings resulting from
improved manager variance investigation decisions (given the additional
information) less the costs associated with the production and dis
semination of the additional information. Within this research the lack

167
of a significant information variable effect implies that the net
benefit of providing additional information may not be positive.
However, it should be noted that the lack of an information variable
effect could be the result of either 1) that the subjects within the
reduced information level were able to estimate (with relative
efficiency) the missing information as a result of their experiences
with the decision task, or 2) that the subjects within the expanded
information level did not utilize the additional information efficiently.
General Standard Setting Process
Another implication of this research concerns the general standard
setting process. Within decision situations involving either con
vergent or divergent adjustment decision criteria,conservatism was de
fined as a less than complete adjustment toward the optimal decision
value. Convergent adjustment involved individual decision value
adjustment toward the standard, and divergent adjustment involved in
dividual decision value adjustment away from the standard. The stan
dard used within this research can be conceived of as a type of de
cision behavior magnet. Over certain ranges of its field (the rele
vant decision axis) the standard attracts variance investigation be
havior, and over other ranges of its field it repels variance investi
gation decision behavior. Decision criteria conservatism when the
adjustment process diverges from the standard could be the result of
the attraction (or pull) of the standard that restrains the individual
from making a complete (divergent) adjustment to the optimal decision
value. Decision criteria conservatism when the adjustment process
converges to the standard could be the result of the repelling force

168
(or push) of the standard, subjectively limiting a complete (convergent)
adjustment to the optimal decision value. The decision behavior magnet
view represents a generalization of the subjective adjustment limit
concept. The subjective limit concept as presented earlier dealt with
decision criteria conservatism involving only convergent adjustment.
The question of whether decision makers could learn (either through
training or experience) to reduce the decision biases implied by the
decision behavior magnet concept remains unanswered. An alternative
approach, however, involves the standard setting process itself.
Previous research involving the nature of standards (e.g., strict
standards, currently attainable standards, lax standards) has primarily
dealt with the standard's motivational affects. Other things being
equal, the nature of the standard may affect the decision behavior
magnet biases.. Within situations requiring adjustment, lax standards
(which have values greater than the mean of the in-control state) may
reduce substantially divergent decision criteria conservatism. Within
situations requiring convergent adjustment, strict standards (which
have values less than the mean of the in-control state) may reduce
substantially convergent decision criteria conservatism.
Future Research
The implications of this study for future accounting research
take two general forms: modification of the existing conceptual
development and generalizations of the modified conceptual development.
Modifications of the existing conceptual development would involve
incorporating and testing the ex post explanations offered for certain

159
variance investigation decision behavior observed within this research.
Generalization of the modified conceptual development would involve
extensions aimed at reducing the limitations of the existing study.
The primary ex post concept developed within this research is
the decision behavior magnet view (incorporating the concept of sub
jective limits). Future research could directly test this concept.
Such research could involve variance investigation decision situations
in which the standard was not the same value as the mean of the
in-control state of nature. If this ex post concept is valid, then
subjects within situations involving convergent adjustment and a
strict standard should have less decision criteria conservatism than
subjects within situations involving convergent adjustment and a lax
standard. Similarly, subjects within situations involving divergent
adjustment and a lax standard should have less decision criteria
conservatism than subjects within situations involving divergent ad
justment and a strict standard. Additional research could test
whether the biases implied by the decision behavior magnet concept can
be reduced or eliminated by extended training within the particular
decision situation.
Generalizations of the conceptual development would require
extensions concerning both the experimental subjects and the experimental
environment. Extensions concerning the experimental subjects could
involve the use of non-students, perferably actual operational managers.
Such extensions could include managers within familiar experimental
environments and within non-familiar experimental environments. Be
yond extending the decision criteria conservatism effects, the central
question would be whether differential information processing

170
efficiency (decision criteria conservatism) is a characteristic of only
training (learning) situations or remains a characteristic of
well-trained (non-learning) situations.
Extensions involving the experimental environment could take
several forms. First, they may relate to the levels of the situation
variables. Only certain situation variables were studied within this
research: extensions could study the affects of manipulating different
situation variables at different levels (e.g., more than two states of
nature, non-normal distributions, variable relative frequencies, vari
able decision cost functions, multiple standard cost variances).
Second, extensions could be made in relation to the experimental
task itself. Within this research the decision task was of a
cross-sectional nature: i.e., each decision trial was independent of
the previous decision trials. An extension could involve a decision
task of a time-series nature: i.e., where each decision trial would
be dependent on the previous decision trials. Such a decision task
extension would involve sequential decision making, where the physical
process is initially in some state of nature and over a series of
decision trials (variance reports) the process may or may not move to
another state of nature. The task of the decision maker would be to
detect the point in time in which the physical process shifts to an
other state.
The third and final extension discussed within this section con
cerns the nature of the experimental decisions. Within this research
the decisions involved the detection of state of nature for the purpose
of physical process control within cost minimization constraints. Be
yond the operational control level, standard variance reports can be

171
utilized as performance measures for the appropriate operational managers
and for the physical systems under their control. The nature of de
cisions of this type, although generally aggregated over longer time
periods, is that of adequate performance verses inadequate performance
rather than that of investigation verses non-investigation. Within the
current research, performance decisions were assumed implicitly within
the performance feedback measures given the subjects during the train
ing phase. An extension would involve the actual formation of such
decisions and, in turn, their effects upon operational control decisions.

APPENDIX A
THEORY OF SIGNAL DETECTION DERIVATIONS
Relationship of Conditional Probability Matrix and Bayes1 Theorem
The unconditional probabilities of the four event outcomes can be
stated in terms of the conditional probabilities and the prior proba
bilities of the sn and n distributions: 1) the probability of a hit,
P(sn,Y) equals P(Y|sn)P(sn); 2) the probability of a miss, P(sn,N)
equals P(N |sn)P(sn); 3) the probability of a false alarm, P(n,Y) equals
P(Y|n)P(n); and 4) the probability of a correct rejection, P(n,N)
equals P(N|n)P(n).
Egan (1975, pp. 12-13) demonstrates the relationship between the
single-interval procedure and Bayes1 theorem. The posterior probability
of the event sn, given the discrete observation x, can be represented
as P(sn|x). Since the two events sn and n are exhaustive the addition
of P(sn|x) and P(njx) equals one. The posterior probability of the
event sn can be stated as P(sn,x) which equals:
P(x|sn)P(sn) = P(sn|x)P(x).
Rearranging the terms of this equality and expanding the P(x):
n/ i P(x|sn)P(sn)
P(sn x)= i¡!
P(x]sn)P(sn) + P(x[n)P(n)
This equation is a simple expression of Bayes' theorem. It can be ex
tended to the odds form, the posterior odds equals the prior odds times
the likelihood ratio of the observation x:
P(sn[x) P(sn) P(x|sn)
P(n|x) P(n) P(x|n)
172

Egan (1975) further demonstrates that the posterior probability is
strictly monotone with the likelihood ratio of x, L(x); that is, when
L(x2) > LCXj) then P(sn|x2) > P(sn|x1).
173
Maximization of Expected Value Objective Function
The decision function for maximizing the expected value decision
goal is:
E(V) = P(Y!sn)P(sn)Vsn)Y + P(Y|n)P(n)VnjY
+ P(N|n)P(n)Vn>N + P(N|sn)P(sn)VsnjN
Given an observation x, the subject has two decision functions:
1) E(V|x,Y), the expected value given the observation and the
subject responds "Yes," and
2) E(V|x,N), the expected value give the observation and the
subject responds "No."
These two decision functions can be expressed as the sum of the values
associated with each decision weighted by the appropriate posterior
probability:
E(V|x,Y) = P(sn|x)VsrljY + P(n|x)VnjY, and
E(V j x,N) = P(n|x)VnjN + P(sn|x)VsnjN.
The subject should say "Yes" only if E(V|x,Y) > E(V|x,N). If the
righthand side of the above two equations are used in this inequality,
then the subject should say "Yes" only when:
P(sn|x)Vsn,y + P(n|x)VnY > P(n|x)VnjN + P(sn|x)Vsn>N
Rearranging the terms and expressing the posterior odds in terms of the
prior odds and the likelihood ratio of x results in the decision rule
for saying "Yes" in terms of the likelihood ratio of the observation:
P(x1sn) P(n) Vn,N Vn,Y
P(x n) > P(sn) 'vsn>Y VsM

APPENDIX B
ORAL PRESENTATION TO ELICITE VOLUNTEER SUBJECTS
I assume that most, if not all, of you intend to become account
ants. You should realize that accounting is more than an occupation,
it is a profession. As a member of a profession you have certain re
sponsibilities beyond those encountered in other occupations. One of
these responsibilities is the readiness to help expand the state of the
art, the level of knowledge, of the profession. Today you have an
opportunity to meet this responsibility. As part of my Ph.D. require
ments I am conducting an experiment which examines the impact upon
decision makers of certain accounting produced information. I am asking
that you volunteer to participate as subjects in this experiment.
The accounting information under study is the standard cost vari
ance report and, as a subject, you would be asked to assume the role of
an operations manager and to make certain decisions based upon standard
cost variance reports. After volunteering, greater detail concerning
the role and the required decisions will be provided in a short booklet
and in a training session.
Benefits to volunteering as a subject, beyond meeting part of your
professional responsibilities, include:
1) A chance to expand your knowledge of standard cost variance
reports and their utilization from the point of view of a
line manager.
174

175
2) A chance to earn up to $10.00 as payment for participating in
the experiment.
3) A chance to gain first-hand experience with an accounting
research project.
Your participation in the experiment would not involve a great
sacrifice of your time. The entire experiment requires only two
one-period sessions. The times of these sessions are flexible to meet
the requirements of your personal schedule. The first one-period
session, a training session, would be scheduled sometime on Monday or
Tuesday, April 10th and 11th. The second one-period session, the actual
experiment, would be scheduled sometime on Wednesday or Thursday,
April 12th and 13th.
I can assure you that your responses within the experiment will
be held strictly confidential. Only aggregate or anomymous results
will be made pub!ic.
I will pass around a sign-up sheet; I would appreciate it if
everyone would print their name and indicate with a check-mark whether
you are willing to participate. If you are willing to participate,
fill in your phone number and indicate some times that I can contact
you to schedule your participation.
I would like to thank you in advance for your time. The overall
success of this study is dependent upon a positive response from most
of you in this room. I am confident that as professionals you will
exceed my expectations.
Does anyone have any general questions?

APPENDIX C
BACKGROUND INFORMATION BOOKLET
The following represents background information concerning the role
that you will be asked to assume in the business experiment for which
you have volunteered to participate. Please read this information
several times, familarizing yourself with the general aspects of this
role.
More detailed information together with specific instructions will be
presented at the inception of the actual experiment. You may use this
booklet when you participate in the experiment.
PLEASE DO NOT DISCUSS THIS EXPERIMENT WITH OTHER PEOPLE. If you have
questions contact Cl if Brown (Bryan 214E or telephone 2-0155).
My Training Session is:
Date:
Time:
Room:
My Experiment Session is:
Date: Time:
Room:
176

177
You will be asked to assume the role of an assembly department
operational control manager and within this role you will be asked to
make certain decisions based upon information that will be presented to
you. The specific nature of these decisions will become clear as you
read this background information booklet.
General Company Information
Assume you are employed by the American Seating Company (AMSECO).
AMSECO manufactures and sells public seating and institutional furniture
to education, amusement, transportation and health care markets. The
company is one of the largest suppliers in each of its markets. Select
ed AMSECO financial statement items for the previous calendar year
include:
Net Sales
Cost of Sales
Net Income
Total Assets
Net Assets
$ 84,000,000
$ 67,250,000
$ 1,550,000
$ 60,465,000
$ 33,947,000
General Product Information
The assembly department of which you are the operational control
manager is engaged in the manufacturing of a metal folding chair. Since
the major market for the folding chair is the education market the chair
must meet stricter specifications than the typical home-use variety.
The chair is produced in three styles, the only difference between the
styles is the color; grey, brown, and black. The production of the
folding chair averages 1,000 chairs per week; this production accounts
for approximately 1.5 percent of AMSECO's net sales.
Metal Folding Chair Manufacturing Process
The folding chair is manufactured within three operational depart
ments using a sequential production system. The three departments are
fabrication, finishing, and assembly; you are the operational control
manager of the metal folding chair assembly department.
The fabrication department manufactures the parts required to make
the folding chair and transfers them to the finishing department in
large identical unit batches. The finishing department paints and fin
ishes the metal parts and groups them into kits, each kit containing
the parts necessary to assemble one folding chair. The kits are trans-
fered to the assembly department in 200 kit batches. The assembly
department assembles the folding chair kits in 200 chair job-orders,
packages each assembled chair, and transfers them to the product inven
tory department.
Metal Folding Chair Assembly Department
Employees
The assembly department employs one operational control manager
(yourself), one assembly supervisor, fifteen full-time assembly workers,
and two packagers. Two quality control inspectors work within the

178
assembly department; however, the costs associated with the inspectors
are not charged directly to the assembly department but rather to a
central product engineering and quality control department.
The assembly supervisor does not actually participate in the
physical assembly procedures but rather supervises the assembly workers,
the kit input to the assembly department, the packagers, and the
finished chair output to the product inventory department.
The operational control manager (yourself) has overall responsi
bility for assembly department product scheduling, physical assembly
process and procedures, and operational efficiency.
Physical Process
Five years ago the folding chair assembly was accomplished using
an assembly-line type process in which each assembly worker performed
a specific, repetitive task. An experiment at that time, however,
demonstrated that greater worker efficiency could be obtained if each
worker was responsible for the complete assembly of a chair. The
result was that the average time required to assemble a chair decreased
substantially. Consequently, the current assembly process consists
of each assembly worker assembling a complete chair from a kit. The
single largest drawback of this method is that the assembly line goes
out-of-control much more often than when the previous assembly method
was utilized. It is believed, however, that the net benefits of the
current method outweights this drawback.
The folding chair assembly is accomplished following a predeter
mined assembly sequence. After assembling a folding chair the worker
places it on a conveyor belt which takes it to the quality control
inspectors. The onsight quality control process consists of a visual
inspection of the chair's finish and a manual inspection of the chair's
operation. A chair which passes this inspection is folded and placed
on a conveyor belt which takes it to the packagers. A chair which
does not pass this inspection is placed in a rework area with a form
indicating the reason for inspection failure. Periodically the central
quality control department randomly selects a sample of finished chairs
and performs additional structural tests. The packagers box each folded
chair in a corregated cardboard carton, stamping the carton with the
color of the chair.
Accounting Control System
AMSECO utilizes a job-order costing and control system employing
standard costs. The folding chair is manufactured and controlled in
200 chair lots. The assembly department is charged only with material
usage and labor efficiency variances, not with material price and wage
rate variances. In the past the material usage variance has been
negligable and the operating performance of the assembly department has
been completely determined by the labor efficiency of the workers.
Engineering estimates and historical statistical studies have been
used to set the labor efficiency standards within the assembly depart
ment. These labor efficiency standards are reasonably attainable and
allow for unavoidable labor inefficiencies and reasonable variation in
worker performance. As the operational control manager you should
accept the labor efficiency standards in terms of fair control and
performance goals.

179
The
ment are:
per chair labor efficiency standards for the assembly depart-
Process
Standard Time Allowed
Assembly
Packaging
Total
The variance report contains
variance is reported only in
33
3
minutes
minutes
36 minutes
only the labor efficiency variance and this
aggregate form (i.e., the variance is not
broken into the assembly and packaging processes). The accounting
system is computerized and the variance report is in computer output
format.
The physical labor process of the assembly department can be in one
of two states; either in-control or out-of-control. The overall labor
process is made up of many individual physical labor procedures; the
expected aggregate of these procedures is represented by the labor
efficiency standard of 36 minutes per chair. The department is defined
to be in-control when all of these physical procedures are performed as
expected. The department is defined to be out-of-control when one or
more of these physical labor procedures are not performed as expected.
Since the overall labor process is made up of many individual labor
procedures it is possible for the department to be out-of-control while
at the same time the reported labor efficiency variance is small. This
would occur when some of the individual procedures are performed at
higher than normal efficiency while other procedures are performed at
subnormal (out-of-control) efficiency. It is also possible for the
department to be in-control while at the same time the reported labor
efficiency variance is large. This would occur when most of the indivd-
ual procedures are performed at slightly below normal efficiency. These
examples represent two extreme situations; the probability of occurance
for either of them is low when compared to the probabilities of occur
ance for situations that fall inbetween the two extreme examples.
Your Task as the Operational Control Manager
Based upon the variance report you must decide whether to investi
gate the particular labor efficiency variance for its underlying causes.
The purpose of investigation is to facilitate correcting those indivdual
labor procedures which are not operating as expected. The following
assumptions will aid your decision making:
1) If you decide to investigate the variance and the assembly
department turns out to be out-of-control, the department will
be returned to the original in-control state with certainty.
2) If you decide not to investigate the variance and the assembly
department turns out to be out-of-control, the department will
remain out-of-control with certainty.
There are certain marginal costs associated with the variance
investigation decision. These costs depend upon your decision (either
investigate or do not investigate) and upon the actual state of the
assembly department (either in-control or out-of-control). These mar
ginal costs have the following values and relationships:

180
If Your
Investigation
Decision Is
And If The
Assembly Line
State Is
Then
Your
Costs
Are
Investigation
Production
Total
Investigate
In-control
$
X
5
0.00
$ X
Investigate
Out-of-control
$
Y
$
0.00
$ Y
Not Investigate
In-control
$
0.00
$
0.00
$ 0.00
Not Investigate
Out-of-control
$
0.00
$ 175.00
$ 175.00
X represents the variable investigation cost when the assembly
line is in-control. This investigation cost is related to the size of
the labor efficiency variance; the more negative (unfavorable) the
variance the larger this cost.
Y represents the variable investigation cost when the assembly
line is out-of-control. This investigation cost also is related to the
size of the labor efficiency variance; the more negative (unfavorable)
the variance the smaller this cost. For any given labor efficiency
variance the investigation cost if the assembly line is out-of-control,
Y, is always larger than the investigation cost if the assembly line is
in-control, X. This is due to the investigation cost including the cost
of correcting the out-of-control labor procedure when the assembly line
is out-of-control.
The $175.00 cost is the expected marginal production cost of
operating next period in the out-of-control state. This cost is fixed
for all labor efficiency variance reports.
Your Performance Measure as the Operational Control Manager
The metal folding chair section manager (your immediate supervisor)
evaluates your control performance in terms of the minimization of both
investigation costs and production costs above the expected standard
(the total of these costs is called the Total Investigation Decisions
Cost). The extent to which your investigation decisions for a specified
period of time minimizes the Total Investigation Decision Cost is
determined as follows:
Extent of Total Investigation
Decisions Cost Minimization =
Total Investigation Decisions Cost of your decisions
Optimal Total Investigation Decisions Cost
The optimal Total Investigation Decisions Cost is determined by the
section manager at the end of the specified period of time. It i_s_
possible for your Total Investigation Decisions Cost to equal the optimal
Total Investigation Decisions Cost. AMSECO pays a cash bonus based on
this measure of the extent of Total Investigation Decisions Cost
minimization. The closer this measure is to one or less than one the
larger the cash bonus.
The cash bonus for 100 labor efficiency variance reports is as
fol 1ows:
Extent of Total Investi
gation Decisions Cost
Minimization
1.000
1.001 to 1.010
1.011 to 1.020
Cash Bonus
$ 10.00
9.00
8.00

181
Extent of Total Investi
gation Decisions Cost
Minimization Cash Bonus
1.021
to
1.030
$ 7.00
1.031
to
1.040
6.00
1.041
to
1.050
5.00
1.051
to
1.100
4.00
1.011
to
1.150
3.00
1.151
2.00
Again, I rely on your not discussing this experiment with other people.

APPENDIX D
PRIOR INFORMATION SHEETS
Experiment Condition (I1,S2,C2)
The following information is based upon past experience and past obser
vations of the assembly line. This information may be helpful to you
in making your decisions.
The portion of time the assembly line was:
In-control 60% Out-of-control 40%
The actual minutes per chair for both states are assumed to be
normally distributed.
The most favorable labor efficiency variance was:
10.0 minutes per chair (26.0 actual minutes incurred per chair)
The most unfavorable labor efficiency variance was:
19.0 minutes per chair (55.0 actual minutes incurred per chair)
The maximum investigation cost when the assembly line state was:
In-control $ 98.75 Out-of-control $ 154.17
The minimum investigation cost when the assembly line state was:
In-control $ 62.50 0ut-of-contro $ 142.08
Experiment Condition (I2,S2,C2)
The following information is based upon past experience and past obser
vations of the assembly line. This information may be helpful to you
in making your decisions.
182

183
The two assembly line state means, in minutes per chair, are:
In-control 36.0 Out-of-control 45.0
The two assembly line state standard deviations, in minutes per chair:
In-contorl 5.0 Out-of-control 5.0
The portion of time the assembly line was:
In-control 60% Out-of-control 40%
The actual minutes per chair for both states are normally distributed.
The most favorable labor efficiency variance was:
10.0 minutes per chair (26.0 actual minutes incurred per chair)
The most unfavorable labor efficiency variance was:
19.0 minutes per chair (55.0 actual minutes incurred per chair)
The maximum investigation cost when the assembly line state was:
In-control $ 98.75 Out-of-control $ 154.17
The minimum investigation cost when the assembly line state was:
In-control $ 62.50 Out-of-control $ 142.08

APPENDIX E
SUBJECT HEURISTIC ELICITATION QUESTIONNAIRE
Please answer the following questions in the spaces provided. This is
not a test; there are no right or wrong answers. It is important that
you try to answer each question honestly. If you need additional space
use the back of this questionnaire.
1. What rule or set of rules did you use to make your investigation
decisions in the last 50 labor efficiency variance reports (the
second block)? Please include any numerical values that were part
of this rule or set of rules.
2. Do you feel you used the same rule or set of rules as you gave in
question one to make your variance investigation decisions in the
first 50 labor efficiency variance reports (the first block)? If
not the same, how does the rule or set of rules used in the first
50 reports differ from what you gave in question one? Please in
clude any numerical values used as part of the rules or sets of
rules that were different.
3.Rate the following items of information in terms of their importance
to you in making your variance investigation decisions (do not rate
the 'other' item unless you have specified some other information).
Circle the appropriate response.
A.The background information booklet.
irklrk-k'k'k-k'k'k-k'k'k'k-k'k'k'k'k-k'k'k'k'k'k'k'k'k'k'k'k'k'k'k'k-k'kk-k'k'k'k'k'k'k'k'k.'k-k
k
*



Not Slightly
Important Important
Moderately Very
Important Important
Extremely
Important
B. The information based upon past experience and past observation
which was presented in the front of each experiment booklet.
k'k-k'kk'kk-k'k-k'k-k-kkic'k'k'kkk-k'k-k'k'k-k-k'k'k'k'k'k'k-k'kk-k-k-k'k'k-k'k'k'k'k-k'k'k
* * *
Not Slightly Moderately Very Extremely
Important Important Important Important Important
C. The standard minutes allowed per chair.
**************************************************
k
k
*
*
*
Not Slightly
Important Important
Moderately Very
Important Important
Extremely
Important
184

185
D. The actual minutes incurred per chair.
*************************************************
*****
Not Slightly Moderately Very Extremely
Important Important Important Important Important
E. The labor efficiency variance per chair.
*************************************************
*****
Not Slightly Moderately Very Extremely
Important Important Important Important Important
F. The total chairs produced.
*************************************************
Not Slightly Moderately Very Extremely
Important Important Important Important Important
G.The costs associated with investigation decisions.
k'k'k'k-k'k'k'k'k'k'k-k'k'kirk'k'k'k'k'k'k'k'k'k'k'k'k'k'k'k'k'k'k-k'k'k'k'kic'k'k'k'k'k'k'k'k'k
* * *
Not Slightly Moderately Very Extremely
Important Important Important Important Important
H.Other (please specify) .
*************************************************
*****
Not Slightly Moderately Very Extremely
Important Important Important Important Important
4. Recalling the training phase which you completed at an earlier time,
what rule or set of rules did you use to make your variance investi
gation decisions in the first 33 variance reports? Please include
any numerical values that were part of this rule or set of rules.
5. How did you initially form the rule or set of rules you gave in
question four? What lead you to use this particular rule or set of
rules?
6. If the rule or set of rules you gave in question four is different
from the rule or set of rules you gave in question one, what caused
you to make the change?
7. What other information items, if presented, do you feel would im
prove you variance investigation decisions?
8. Were your decisions influenced by discussions with other partici
pants in this experiment?

APPENDIX F
SUBJECT MOTIVATION QUESTIONNAIRE
Please respond to the following questions by circling the appropriate
response. This is not a test; there are no right or wrong answers. It
is important that you try to answer each question honestly.
1. Suppose this same experiment was to be repeated on a number of future
occasions. How many more times would you be prepared to participate
in the experiment under circumstances similar to the present?
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk

*
*
* *
*
No 1
2
3
4 5
More
More More
More
More
More More
Than 5
Times Time
Ti mes
Times
Times Times
Times
Did your desire to perform well
in undertaking a challenging
task
cause you to try
very hard?
******************************* ******************************


k
* *
*
Definitely Yes
Probably
Don t
Probably No
Definitely
Yes
Yes
Know
No
No
Did you enjoy making the decisions requi
red in the experiment?
***************************************** ********************
* *
k
*
* *
*
Definitely No
Probably
Don' t
Probably Yes
Definitely
No
No
Know
Yes
Yes
Are you satisfied
with your performance
in the experiment?
*************************************************************
k k
*
*
k k
*
Definitely Yes
Probably
Don't
Probably No
Definitely
Yes
Yes
Know
No
No
Do you think you
would have tried less hard if the money rewards had
been halved?
*************************************************************
* *
*
*
* *
*
Definitely No
Probably
Don't
Probably Yes
Definitely
No
No
Know
Yes
Yes
186

187
6. Did you feel tense during the experiment?
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
k k k k k k k
Definitely Yes Probably Don't Probably No Definitely
Yes Yes Know No No
7. Did your desire to cooperate with the experimenter cause you to try
very hard?
*************************************************************
*******
Definitely No Probably Don't Probably Yes Definitely
No No Know Yes Yes
8. Do you think the money rewards caused you to try harder than you
would have without them?
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
k k k k k k k
Definitely Yes Probably Don't Probably No Definitely
Yes Yes Know No No
9. Did you enjoy the overall experience of participating in the
experiment?
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
k k k k k k k
Definitely No Probably Don't Probably Yes Definitely
No No Know Yes Yes
10.Did your desire to contribute to research knowledge cause you to try
very hard?
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
k k k k k k k
Definitely Yes Probably Don't Probably No Definitely
Yes Yes Know No No

BIBLIOGRAPHY
Al pert, M., and Raiffa, H. "A Progress Report on the Training of
Probability Assessors." Unpublished manuscript, Harvard
University, 1968.
American Accounting Association. A Statement of Basic Accounting
Theory. Evanston, Ill.: The Association, 1966.
. "Report of the 1969-70, 1970-71 Committee on Managerial
Accounting." The Accounting Review Supplement, 47 (1972),
317-35.
Anthony, R. N. Planning and Control Systems: A Framework for
Analysis. Boston: Harvard University Press, 1965.
Appelbaum, M. I., and Cramer, E. M. "Some Problems in the Non-
orthoaonal Analysis of Variance." Psychological Bulletin,
81 (1974), 335-43.
Banks, W. P. "Signal Detection Theory and Human Memory." Psychological
Bulletin, 74 (1970), 81-99.
Barefield, R. M. "The Effect of Aggregation of Decision Making Success:
A Laboratory Study." Journal of Accounting Research, 10 (1972),
229-42.
Barr, A. J.; Goodnight, J. H.; Sail, J. P.; and Helwig, J. T.
A User's Guide to SAS 76. Raleigh, N.C.: SAS Institute, Inc.,
1976.
Bernbach, H. A. "Decision Processes in Memory." Psychological Review,
74 (1967), 462-80.
Bierman, H.; Fouraker, L. E.; and Jaedicke, R. K. "A Use Probability
and Statistics in Performance Evaluation." The Accounting
Review, 36 (1961), 409-17.
Coombs, C. H.; Dawes, R. M.; and Tversky, A. Mathematical Psychology:
An Elementary Introduction. Englewood Cliffs, N.J.:
Prentice-Hall, Inc., 1970.
188

189
Demski, J. S. "Analyzing the Effectiveness of the Traditional Standard
Cost Variance Model." Management Accounting, 48 (1967), 9-19.
. "The Decision Implementation Interface: Effects of Alterna
tive Performance Measurement Models." The Accounting Review,
45 (1970), 76-87.
. "Implementation Effects of Alternative Performance Mea
surement Models in a Multivariate Context." The Accounting Re-
view, 46 (1971), 768-78.
Dittman, D. A., and Prakash, P. "Cost Variance Investigation: Markovian
Control of Markov Processes." Journal of Accounting Research,
16 (1978), 14-25.
Dopuch, N.; Birnberg, J. G.; and Demski, J. S. "An Extension of
Standard Cost Variance Analysis." The Accounting Review,
42 (1967), 525-36.
Ducharme, W. M. "A Response Bias Explanation of Conservative Human
Inference." Journal of Experimental Psychology, 85 (1970),
66-74.
, and Peterson, C. R. "Intuitive Inference About Normally
Distributed Populations." Journal of Experimental Psychology,
78 (1968), 269-75.
Dyckman, T. R. "The Investigation of Cost Variances." Journal of
Accounting Research, 7 (1969), 215-44.
Egan, J. P. Signal Detection Theory and ROC Analysis. New York:
Academic Press, Inc., 1975.
Gorry, G. A., and Morton, M. S. S. "A Framework for Mnagement Infor
mation Systems." Sloan Management Review, Fall (1971), 55-70.
Green, D. M., and Swets, J. A. Signal Detection Theory and Psycho
physics. New York: Robert E. Krieger Publishing Co., 1974.
Grier, J. B. "Nonparametric Indexes for Sensitivity and Bias: Com
puting Formulas." Psychological Bulletin, 75 (1971), 424-9.
Hammertcn, M. "An Investigation into Changes in Decision Criteria and
Other Details of a Decision-Making Task." Psychonomic Science,
21 (1970), 203-4.
Jacobs, F. H. "An Evaluation of the Effectiveness of Some Cost Variance
Investigation Models." Journal of Accounting Research, 16
(1978), 190-203.
Juers, D. A. "Statistical Significance of Accounting Variances.
Management Accounting, 49 (1967), 20-5.

190
Kahneman, D., and Tversky, A. "Subjective Probability: A Judgement of
Representativeness." Cognitive Psychology, 5 (1972), 430-54.
Kaplan, R. S. "Optimal Investigation Strategies with Imperfect Infor
mation." Journal of Accounting Research, 7 (1969), 32-43.
. "The Significance and Investigation of Cost Variances: Survey
and Extensions." Journal of Accounting Research, 13 (1975),
311-37.
Koehler, R. W. "The Relevance of Probability Statistics to Accounting
Variance Control." Management Accounting, 50 (1968), 35-41.
Lewis, C., and Keren, G. "You Can't Have Your Cake and Eat It Too:
Some Considerations of the Error Term." Psychological
Bulletin, 84 (1977), 1150-4.
Lieblich, A., and Lieblich, I. "Arithmetical Estimation Under Condi
tions of Different Payoff Matrices." Psychonomic Science,
14 (1969), 87-8.
Luh, F. "Controlled Cost: An Operational Concept and Statistical
Approach to Standard Costing." The Accounting Review, 43
(1968), 123-32.
Lusted, L. B. "Perception of the Roentgen Image: Applications of
Signal Detectability Theory." Radiological Clinics of North
America, 3 (1969), 435-45.
. "Signal Detectability and Medical Decision-Making."
Science, 171 (1971), 1217-9.
Magee, R. P. "A Simulation Analysis of Alternative Cost Variance
Investigation Models." The Accounting Review, 51 (1976),
529-44.
, and Dickhaut, J. W. "Effects of Compensation Plans on
Heuristics in Cost Variance Investigations." Unpublished
manuscript, Northwestern University, 1977.
Pastore, R. E., and Scheirer, C. J. "Signal Detection Theory: Consid
erations for General Application." Psychological Bulletin,
81 (1974), 945-58.
Peterson, C. R.; DuCharme, W. M; and Edwards, W. "Sampling Dis
tributions and Probability Revisions." Journal of Experi-
mental Psychology, 76 (1968), 236-43.
Probst, F. R. "Probabilistic Cost Controls: A Behavioral Dimension."
The Accounting Review, 46 (1971), 113-8.
Scheffe", H. The Analysis of Variance. New York: John Wiley and
Sons, Inc., 1959.

191
Simon, H. A. The New Science of Management Decision. New York: Harper
and Row, Inc., 1966.
Slovic, P. "From Shakepeare to Simon: Speculations-- and Some Evidence--
About Man's Ability to Process Information." Oregon Research
Institute Bulletin, 12 (1972), 1-29.
, and Lichtenstein, S. "Comparison of Bayesian and Regression
Approaches to the Study of Information Processing in Judgement."
Organizational Behavior and Human Performance, 6 (1971),
649-744.
Snowball, D., and Brown, C. "Decision-Making Involving Sequential
Events: Some Effects of Disaggregated Data and Dispositions
Toward Risk." Unpublished manuscript, University of Florida, 1977.
Swets, J. A. "Signal Detection in Medical Diagonis." In Computer
Diagnosis and Diagnostic Methods, J. A. Jacquez (Ed.).
Springfield, IL: C. C. Thomas, 1972.
Swieringa, R. J.; Gibbins, M.; Larsson, L.; and Sweeney, J. L. "Ex
periments in the Heuristics of Human Information Processing."
Journal of Accounting Research: Supplement on Studies on
Human Information Processing In Accounting, 14 (1976), 159-87.
Taylor, R. N., and Dunnette, M. D. "Relative Contribution of Decision
Maker Attributes to Decision Processes." Organizational
Behavior and Human Performance, 12 (1974), 286-98.
Tversky, A., and Kahneman, D. "Belief in the Law of Small Numbers."
Psychological Bulletin, 76 (1971), 105-10.
. "Availability: A Heuristic for Judging Frequency and
Probabi1ity." Cognitive Psychology, 5 (1973), 207-32.
. "Judgement Under Uncertainty: Heuristics and Biases."
Science, 185 (1974), 1124-31. ~
. "Causal Schemata in Judegments Under Uncertainty." In
Progress in Social Psychology, M. Fishbein (Ed.). Hillsdale,
N.J.: Lawrence Erlbaum Associates, 1977.
Ulehla, Z. J.; Canges, L; Wackwitz, F. "Signal Detectability Theory
Applied to Conceptual Discrimination." Psychonomic Science,
8 (1967a), 221-2.
. "Integration of Conceptual Information." Psychonomic
Science, 8 (1967b), 223-4.
Watson, C. S. "Psychophysics." In Handbook of General Psychology,
B. B. Wolman (Ed.). New York: Prentice-Hall, Inc., 1973.

192
Weissman, S. M.; Hollingworth, S. R.; Baird, J. C. "Psychophysical
Study of Numbers: III. Methodological Applications."
Psychological Research, 38 (1975), 97-115.
Zannetos, Z. A. "Standard Cost as a First Step to Probabilistic Control:
A Theoretical Justification, and Extension and Implications."
The Accounting Review, 39 (1964), 296-304.
Zigler, E. "Metatheoretical Issues in Developmental Psychology." In
Theories in Contemporary Psychology, M. Marx (Ed.). New York:
Macmillan, 1963.

BIOGRAPHICAL SKETCH
Clifton Edward Brown was born on May 17, 1947,in Kansas City,
Kansas. In June 1965 he graduated from St. Petersburg High School,
St. Petersburg, Florida. Mr. Brown enlisted in the United States Air
Force from 1965 through 1968, and was honorably separated with the
rank of Sergeant. He was married in December 1968 to Sandra Cornell.
Mr. Brown attended the University of South Florida from which
he graduated in June 1971 with a Bachelor of Arts degree majoring in
accounting. From graduation until 1974 Mr. Brown was employed as an
industrial accountant, during which time he held positions of con
troller and vice-president of finance. From September 1974 until
the present Mr. Brown has attended the University of Florida where
he has studied for a Doctor of Philosophy degree in business admin
istration (accounting).
While attending the University of Florida, Mr. Brown was the
recipent of the Price Waterhouse Foundation Grant, the American
Accounting Association Fellowship Grant, and the Haskin and Sells
Fellowship Grant. Mr. Brown was elected by the faculty as the
university representative to the 1975 American Accounting Association
Doctoral Consortium.
193

I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
A. Rashad Abdel-ffhali k, Chairman
Professor of Accounting
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Gary L. irHstrum
Associate Professor of Accounting
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Dougla<£^ T. STiowLalT1^-"
Assistant Professor of Accounting
I certify that I have read this study and that in my opinion it
conforms to acceptable standards of scholarly presentation and is fully
adequate, in scope and quality, as a dissertation for the degree of
Doctor of Philosophy.
Richard A. Griggs ^ //
Assistant Professor of Psychology

This dissertation was submitted to the Graduate Faculty of the Department
of Accounting in the College of Business Administration and to the
Graduate Council, and was accepted as partial fulfillment of the require
ments for the degree of Doctor of Philosophy.
August 1978
Dean, Graduate School



47
P(x|out) P(in) V1r|>N Vin>Y
Equation 1
where x = the actual minutes incurred per chair;
out = the state of out-of-control;
in
= the state of in-control;
Vin,N = ^e cost associated with not investigating an in-control
variance;
V-¡nsy = the cost associated with investigating an in-control
variance;
V0Ut,Y = the cost associated with investigating an out-of-control
variance; and
Vout,N = the cost associated with not investigating and
out-of-control variance.
The manipulation of the cost variable produces a constant cost ratio
at each of the two levels of this variable. When the cost variable
level is Cl the constant cost ratio is equal to one-third, and when the
cost variable level is C2 the constant cost ratio is equal to three.
The optimal model investigation decision rule expressed in equation 1
can be restate as:
LRq > PR.j CR Equation 2
where LRQ = the likelihood odds of the reported labor efficiency
variance being out-of-control (P(x|out)/P(x|in));
PR.j = the prior odds of the in-control state (P(in)/P(out)); and
CR = the constant cost ratio associated with the appropriate
level of the cost variable.
Assuming that both states of nature are normally distributed with equal
variance the .optimal model investigation decision rule expressed in


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EPUS0L2VE_2XQ0O6 INGEST_TIME 2014-10-13T21:10:20Z PACKAGE AA00025728_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


142
Duncan's multiple range test was used to determine if the sub
jects' mean GPAs differed between the eight independent variables con
ditions. Given a completely between-subject assignment these eight
groups represented the results of the experimental sampling plan. The
results indicate that there were not significant differences (p=.01)
between the mean GPAs of these eight groups. Bartlett's test of
homogeneity of variance was used to test the variances of the GPAs for
these eight groups. The results indicate that the hypothesis of homo
geneity of variance could not be rejected (x2=7.6906 with 7 d.f., p<.50).
Overall, the results indicate that between-group variance should
not be affected significantly by differences between subjects' GPAs.
This does not indicate, however, that within-group variance will not be
affected by the differences between GPAs. The possible effects of
within-group GPA variance were presented above.
Subject Motivations
The objective underlying the analysis of subject motivations was
similar to that relating to subject GPAs: i.e., the analysis was
performed to determine whether motivations varied between groups of
subjects. The initial stage of this analysis consisted of reducing the
ten motivation scale items into several motivation factors using the
technique of factor analysis. The subjects' responses to the ten
motivation scale items were factor analyzed using a varimax rotation for
a four factor solution. The four factor solution was chosen on the
basis of prior beliefs concerning the general dimensions (or sources)
of subject motivations: intrinsic motivation, extrinsic non-monetary
motivation, extrinsic monetary motivation, and general motivation not


145
Duncan's multiple range test was used to determine if the subjects'
motivations differed between the eight independent variables conditions.
The results indicate that there were no significant differences (p=.01)
in the mean motivation factors between the eight groups. Bartlett's
test of homogeneity of variance was used to test the variances of the
three motivation factors for these eight groups. The results indicate
that the hypothesis of homogeneity of variance could not be rejected
for the second and third motivation factors (x2=10.9095 with 7 d.f.,
p<.25; x2=l.796 with 7 d.f., pc.95). The hypothesis of homogeneity
of variance could be rejected at the pc.10 level for the first moti
vation factor (x2=l2.7829 with 7 d.f.).
Overall, the results indicate that between-group variance should
not be affected significantly by differences between the subjects'
motivations. Again, the analysis does not indicate that within-group
variance will be unaffected by the differences between subjects'
motivations: these possible effects were presented above.


153
relative frequencies of the two states and the relative decision error
costs), nor is it sensitive to the effects of various individual attri
butes (e.g., the individual's variable response range). Possible
effects of the situation variables on these aspects of the adjustment
process are examined in greater detail ( with measures of greater
sensitivity) in later sections relating to the individual decision
criteria and the individual decision model sensitivity.
Individual Decision Model Sensitivity
Individual decision model sensitivity was measured employing a
parameter of the TSD model, d'. This parameter measures the decision
sensitivity of the individual's model relative to a theoretical sensi
tivity. The factors which can affect an individual's decision model
sensitivity are 1) the use of a variable response range, and 2) the
use of multiple decision anchors. To isolate'these factors three
decision sensitivity measures were developed: 1) the DN-j measure
(which is affected by both factors), 2) the DNAj measure (which is
affected by only the variable response range), and 3) the DNAj-DNj
measure (which is affected by only the multiple decision anchor).
The variable response range (the DNA-j measure) was affected
primarily by the cost variable. That is, those subjects within the C2
cost level demonstrated larger variable response ranges than did those
subjects within the Cl cost level. The locations of the optimal de
cision values and the subjects' final decision anchors were closer to
the standard for the Cl cost level than for the C2 cost level (see
Figure 2, Chapter III). The subjective adjustment limit concept offered


114
Hypothesis 2.3b predicted that (DNA--DN-|Cl) would be significantly
smaller than (DNA--DN-|C2). A Z test was employed to test this hypothesis
and the difference was found to be not significant (Z=0.8642, n.s.
one-tailed).
Intrinsic Motivation and Decision Model Sensitivity
The model comparison procedure indicated that the intrinsic moti
vation factor had a significant effect on each of the four individual
decision model sensitivity dependent variables.1 The analysis of these
effects requires additional tests. A product-moment correlation was
computed for the intrinsic motivation factor with each of the four
dependent variables. The results of these correlations are presented
in Table* 13. The weakness of these correlations can be expected given
the relatively low explained variance associated with the various in
trinsic motivation measure sums of squares.
The positive association between intrinsic motivation and the d'
i
measure indicate that as the d1 measure increased in size (approached
i
the d1 measure or used a single cutoff value) the subjects' intrinsic
k
motivation measures increase. This result was confirmed by the posi
tive association between the motivation factor and the DN. measure.
As either the variable response range decreases or the use of multiple
cutoff values decreases (both are associated with increases in the DN-
measure) the subjects' intrinsic motivation measures increase. The
substantial drop in the level of significance for the association be
tween the intrinsic motivation and the DNA^ measures would suggest that
^ee the last section in this chapter for a definition and for additional
analyses concerning the intrinsic motivation factor.


123
TABLE 17
PARAMETER VALUES ASSOCIATED WITH THE REDUCED BNC-¡ MODEL
F Values Associated With the
Sources of the Reduced Model
Source
d.f.
F
Cost
1,80
6.88b
Intrinsic motivation
1,80
2.61
Extrinsic (non-monetary)
motivation
1,80
0.04
Extrinsic (monetary)
motivation
1,80
7.79a
GPA
1,80
4.90b
Means and Variances
Associated With the Significant Sources
Variable Level N
Mean
Variance
Cl
47
1.0046
C2
39
0.6147
0.96602
V.Ol
bp<.05
0.10326


156
experimental environment is closer to that of a test-taking situation
which is a more natural environment for a student than for a manager.
Implications For Accounting
Two general implications of this research for accounting are
discussed. These implications are the value of additional information,
and the general standard setting process.
Value of Additional Information
The manipulation of the information variable involved the quan
tity of information contained in the available information set, specifi
cally the presence or absence of various distribution information items.
The results indicated that the information variable did not have a
significant effect upon the individual relative decision costs. The
mean relative decision cost of subjects within the reduced information
level did not differ significantly from that of subjects within the
expanded information level. An implication of this result concerns the
net benefit (for the company) of providing the additional information
within the expanded information level. The additional information
within an actual environment is not costless. Consequently, if such
information is to be provided, other things being equal, the net
benefit of such an action should be positive. In this context, the net
benefit would be the marginal decision cost savings resulting from
improved manager variance investigation decisions (given the additional
information) less the costs associated with the production and dis
semination of the additional information. Within this research the lack



PAGE 1

+80$1 ,1)250$7,21 352&(66,1* )25 '(&,6,216 72 ,19(67,*$7( &267 9$5,$1&(6 %\ &/,)721 ( %52:1 $ ',66(57$7,21 35(6(17(' 72 7+( *5$'8$7( &281&,/ 2) 7+( 81,9(56,7< 2) )/25,'$ ,1 3$57,$/ )8/),//0(17 2) 7+( 5(48,5(0(176 )25 7+( '(*5(( 2) '2&725 2) 3+,/2623+< 81,9(56,7< 2) )/25,'$

PAGE 2

&RS\ULJKW %\ &OLIWRQ ( %URZQ

PAGE 3

$&.12:/('*(0(176 DP JUDWHIXO WR P\ VXSHUYLVRU\ FRPPLWWHH 5DVKDG $EGHONKDOLN 'RXJODV 6QRZEDOO *DU\ +RL VWUXP DQG 5LFKDUG *ULJJV :LWKRXW WKHLU YDOXDEOH FRQWULEXWLRQV WKLV GLVVHUWDWLRQ ZRXOG KDYH KDG IDU PRUH REVFXULWLHV DP LQGHEWHG WR P\ H[SHULPHQWDO DVVLVWDQWV 'RXJ 6QRZEDOO *DU\ +RL VWUXP 5RQ 7HLFKPDQ (G %DLOH\ -RKQ :UDJJH 5REHUW 7KRPSVRQ 1DQF\ +HWVNR $QG\ -XGG DQG 0LFKDHO *LIW HYHQ WKRXJK KH ZDV QRW RQ WLPHf :LWKRXW WKHLU KHOS WKH H[SHULPHQWDO SURFHGXUHV XVHG ZLWKLQ WKLV VWXG\ ZRXOG QRW KDYH EHHQ SRVVLEOH )LQDOO\ P\ JUHDWHVW GHEW LV WR P\ ZLIH 6DQG\ :LWKRXW KHU HQFRXUDJHPHQW DQG VXSSRUW WKLV GLVVHUWDWLRQ QHYHU ZRXOG KDYH EHHQ GRQH

PAGE 4

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

PAGE 5

*HQHUDO 5HVHDUFK 0HWKRGRORJ\ *HQHUDO ([SHULPHQWDO (QYLURQPHQW DQG 7DVN 2SHUDWLRQDOL]DWLRQ RI 9DULDEOHV +\SRWKHVLV )RUPDWLRQ 6LPXODWLRQ RI 2SWLPDO 0RGHO 3HUIRUPDQFHV 6LPXODWLRQ RI ,QYHVWLJDWLRQ 'HFLVLRQ 3HUIRUPDQFHV &+$37(5 ,9 7+( (;3(5,0(17 ([SHULPHQWDO (QYLURQPHQW 6XEMHFWV ([SHULPHQWDO 0DWHULDOV %DFNJURXQG ,QIRUPDWLRQ 9DULDQFH ,QYHVWLJDWLRQ 'HFLVLRQV (OLFLWDWLRQ RI +HXULVWLFV b (OLFLWDWLRQ RI 6XEMHFW 0RWLYDWLRQV ([SHULPHQWDO 3URFHGXUHV $VVLJQPHQW RI 6XEMHFWV WR 7UHDWPHQW &RQGLWLRQV 7UDLQLQJ 3KDVH ([SHULPHQWDO 3KDVH )LQDO 'HEULHILQJ &+$37(5 9 $1$/<6,6 $1' 5(68/76 6XPPDU\ RI 5HVXOWV *HQHUDO 0HWKRG RI $QDO\VLV 7UDLQLQJ 3KDVH $QDO\VLV DQG 5HVXOWV ,QLWLDO 'HFLVLRQ $QFKRUV 'HFLVLRQ $QFKRU $GMXVWPHQW ,QGLYLGXDO 'HFLVLRQ 0RGHO 6HQVLWLYLW\ 7KH GM 'HSHQGHQW 9DULDEOH Y

PAGE 6

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

PAGE 7

$33(1',; % 25$/ 35(6(17$7,21 72 (/,&,7( 92/817((5 68%-(&76 $33(1',; & %$&.*5281' ,1)250$7,21 %22./(7 $33(1',; 35,25 ,1)250$7,21 6+((76 $33(1',; ( 68%-(&7 +(85,67,& (/,&,7$7,21 48(67,211$,5( $33(1',; ) 68%-(&7 027,9$7,21 48(67,211$,5( %,%/,2*5$3+< %,2*5$3+,&$/ 6.(7&+

PAGE 8

$EVWUDFW RI 'LVVHUWDWLRQ 3UHVHQWHG WR WKH *UDGXDWH &RXQFLO RI WKH 8QLYHUVLW\ RI )ORULGD LQ 3DUWLDO )XOILOOPHQW RI WKH 5HTXLUHPHQWV IRU WKH 'HJUHH RI 'RFWRU RI 3KLORVRSK\ +80$1 ,1)250$7,21 352&(66,1* )25 '(&,6,216 72 ,19(67,*$7( &267 9$5,$1&(6 %\ &OLIWRQ ( %URZQ 'HFHPEHU &KDLUPDQ $ 5DVKDG $EGHONKDOLN 0DMRU 'HSDUWPHQW $FFRXQWLQJ 7ZR PDMRU IXQFWLRQV RI PDQDJHPHQW DUH WKRVH RI SODQQLQJ DQG FRQn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nV LQIRUPDWLRQ SURFHVVLQJ IRU SXUSRVHV RI PDNLQJ YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV 7KH VSHFLILF REMHFWLYHV ZHUH f WF GHYHORS D FRQFHSWXDO IUDPHZRUN ZKLFK ZLOO SUHGLFW HIIHFWV FI VSHFLILF VLWXDWLRQDO YDULDEOHV RQ D PDQDJHUnV UHODWLYH HIILn FLHQF\ LQ LQIRUPDWLRQ SURFHVVLQJ DQG LQ YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ PDNLQJ DQG f WR HPSLULFDOO\ WHVW VRPH LPSOLFDWLRQV RI WKLV FRQFHSWXDO IUDPHZRUN YLLL

PAGE 9

9DULDEOHV WKDW FDQ DIIHFW WKH PDQDJHUnV YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ SURFHVV LQFOXGH f VWUXFWXUH RI WKH GHFLVLRQ VLWXDWLRQ f FRQWHQWV RI WKH DYDLODEOH LQIRUPDWLRQ VHW f PDQDJHUnV LQIRUPDWLRQ SURFHVVLQJ HIILFLHQF\ DQG f PDQDJHUnV OHDUQLQJ HIILFLHQF\ 7KH VWUXFWXUH RI WKH GHFLVLRQ VLWXDWLRQ GHSHQGV XSRQ QXPEHU SRVVLEOH VWDWHV RI FRQWURO UHODWLYH IUHTXHQFLHV RI WKH VWDWHV YDULRXV VWDWLVWLFDO UHODWLRQVKLSV DPRQJ WKH VWDWHV DQG YDULRXV UHODWLRQVKLSV EHWZHHQ WKH GHFLVLRQVWDWH RXWFRPHV 7KH FRQWHQWV RI WKH DYDLODEOH LQIRUPDWLRQ VHW LQFOXGH LQIRUPDWLRQ NQRZQ E\ WKH PDQDJHU SULRU WR KLV GHFLVLRQ 7KH PDQDJHUnV LQIRUPDWLRQ SURFHVVLQJ HIILFLHQF\ UHODWHV WR WKH SDUWLFXODU VWUDWHJLHV HPSOR\HG LQ FRPELQLQJ DQG ZHLJKWLQJ YDULRXV LWHPV RI LQIRUn PDWLRQ 7KH PDQDJHUnV OHDUQLQJ HIILFLHQF\ UHIHUV WR WKH PDQDJHUnV DELOLW\ WR OHDUQ IURP KLV H[SHULHQFHV ZLWK WKH FRQWUROOHG SURFHVV %RWK LQIRUPDWLRQ SURFHVVLQJ DQG OHDUQLQJ HIILFLHQF\ DUH WKH UHVXOWV RI WKH PDQDJHUnV SDUWLFXODU GHFLVLRQ VWUDWHJLHV 7KHVH GHFLVLRQ VWUDWHJLHV ZHUH H[SHFWHG WR DGDSW D JHQHUDO IRUP UHIHUUHG WR DV DQFKRULQJ DQG DGn MXVWPHQW $ QDWXUDO VWDUWLQJ SRLQW LV XVHG DV D ILUVW DSSUR[LPDWLRQ IRU WKH LQYHVWLJDWLRQ GHFLVLRQ UXOH DQG LV WKHQ DGMXVWHG DV WKH PDQDJHU OHDUQV IURP KLV H[SHULHQFHV 7ZR H[SHULPHQWDO PHWKRGV ZHUH XVHG WR GHULYH DQG WHVW LPSOLFDWLRQV RI WKH FRQFHSWXDO IUDPHZRUN 6LPXODWLRQ WHFKQLTXHV ZHUH XVHG WR RSHUDn WLRQDOL]H K\SRWKHVHV DQG D ODERUDWRU\ H[SHULPHQW ZDV XVHG WR WHVW WKHVH K\SRWKHVHV 7KH H[SHULPHQWDO HQYLURQPHQW ZDV WKDW RI DQ DVVHPEO\ GHSDUWPHQW ZLWKLQ D VLPXODWHG PDQXIDFWXULQJ FRPSDQ\ 7KH DVVHPEO\ GHSDUWPHQW DVn VHPEOHG D VLQJOH SURGXFW DQG SHUIRUPDQFH RI WKLV GHSDUWPHQW ZDV GHWHUn PLQHG FRPSOHWHO\ E\ WKH DVVHPEO\ ZRUNHUVn ODERU HIILFLHQF\ 6WXGHQW ,;

PAGE 10

VXEMHFWV ZKR DVVXPHG WKH UROH RI DVVHPEO\ GHSDUWPHQW RSHUDWLRQDO PDQDJHU PDGH ODERU HIILFLHQF\ YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV EDVHG XSRQ D VHULHV RI LQGHSHQGHQW VWDQGDUG FRVW YDULDQFH UHSRUWV 7KH PDQLSXODWHG LQGHSHQn GHQW YDULDEOHV LQFOXGHG FRQWHQWV RI WKH DYDLODEOH LQIRUPDWLRQ VHW GLVWULEXWLRQDO SURSHUWLHV RI WKH VWDWHV RI FRQWURO DQG FRVW HIIHFWV RI WKH LQYHVWLJDWLRQ GHFLVLRQ HUURUV 3DUDPHWHUV RI WKH SV\FKRORJLFDO WKHRU\ RI VLJQDO GHWHFWLRQ SHUPLWWHG PHDVXUHPHQW RI VHQVLWLYLW\ DQG FULWHULD RI D VXEMHFWnV GHFLVLRQ PRGHO 7KH VXEMHFWnV LQYHVWLJDWLRQ GHFLVLRQ FRVWV FRPSDUHG WR QRUPDWLYH LQYHVWLJDWLRQ GHFLVLRQ FRVWV GHULYHG XQGHU VLPLODU VLWXDWLRQV ZHUH HPSOR\HG DV D UHODWLYH PHDVXUH RI GHFLVLRQ HIILFLHQF\ 2YHUDOO WKH LPSOLFDWLRQV RI WKH FRQFHSWXDO IUDPHZRUN ZHUH VXSSRUWHG E\ WKH REWDLQHG UHVXOWV 7R H[SODLQ WKH IHZ PDMRU GHYLDWLRQV IURP H[SHFWDWLRQV DQ H[ SRVW K\SRWKHVLV ZDV LQWURGXFHG 7KLV K\SRWKHVLV SRVLWV WKDW WKH VWDQGDUG LQWURGXFHV D VXEMHFWLYH DGMXVWPHQW ELDV 'HSHQGn LQJ XSRQ WKH GLUHFWLRQ RI DGMXVWPHQW WKH VWDQGDUG VXEMHFWLYHO\ LQKLELWV FRPSOHWH DGMXVWPHQW HLWKHU WR RSWLPDO GHFLVLRQ YDOXHV FORVH WR WKH VWDQGDUG RU WR RSWLPDO GHFLVLRQ YDOXHV GLVWDQW IURP WKH VWDQGDUG [

PAGE 11

&+$37(5 ,1752'8&7,21 $1' 2%-(&7,9(6 %XVLQHVV RUJDQL]DWLRQV JHQHUDWH LQWHUQDO DFFRXQWLQJ UHSRUWV IRU XVH E\ PDQDJHUV IRU SXUSRVHV RI HYDOXDWLRQ DQG GHFLVLRQ PDNLQJ 2QH LPSRUWDQW HYDOXDWLRQ SURFHVV DQG GHFLVLRQ WDVN RI PDQDJHPHQW LV WKDW RI VWDQGDUG FRVW YDULDQFH DQDO\VLV DQG LQYHVWLJDWLRQ %DVHG LQ SDUW XSRQ VWDQGDUG FRVW YDULDQFH UHSRUWV JHQHUDWHG E\ WKH LQWHUQDO DFFRXQWLQJ V\VWHP PDQDJHUV HVWLPDWH WKH OLNHOLKRRGV WKDW YDULRXV SURGXFWLRQ SURn FHVVHV UHPDLQ XQGHU FRQWURO DQG GHFLGH ZKHWKHU WR LQYHVWLJDWH SDUWLFXn ODU YDULDQFHV 7KH SUHVHQW VWXG\ FRQFHUQV WKH HIIHFWV RI VHOHFWHG YDULDEOHV RQ VWDQGDUG FRVW YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ PDNLQJ E\ RSHUDWLRQDO FRQWURO PDQDJHUV 7KLV FKDSWHU GLVFXVVHV JHQHUDO DVSHFWV RI WKH VWDQGDUG FRVW YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ SURFHVV DQG WKH YDULDEOHV ZKLFK PD\ DIIHFW WKH SURFHVV DQG SUHVHQWV WKH UHVHDUFK REMHFWLYHV RI WKH VWXG\ 9DULDQFH ,QYHVWLJDWLRQ 'HFLVLRQ 3URFHVVHV $Q LPSRUWDQW IDFWRU LQ WKH HYDOXDWLRQ DQG XVH RI VWDQGDUG FRVW YDULDQFH UHSRUWV LV WKH PDQDJHUnV SHUFHSWLRQ RI WKH YDOLGLW\ RI WKH [7KH $PHULFDQ $FFRXQWLQJ $VVRFLDWLRQ 6WDWHPHQW RI %DVLF $FFRXQWLQJ 7KHRU\ VWDWHV WKDW WKH REMHFWLYH RI DFFRXQWLQJ IRU LQWHUQDO XVH LV WR SURYLGH LQIRUPDWLRQ WR SHUVRQV ZLWKLQ DQ RUJDQL]DWLRQ WKDW HQDEOHV WKHP WR PDNH LQIRUPHG MXGJHPHQWV DQG HIIHFWLYH GHFLVLRQV ZKLFK IXUWKHU WKH RUJDQL]DWLRQnV JRDOV S f

PAGE 12

LQIRUPDWLRQ SURYLGHG ,Q SDUWLFXODU WKH PDQDJHUnV SHUFHSWLRQ RI WKH VWDQGDUG VHWWLQJ SURFHVV ZLOO KDYH FRQVLGHUDEOH LQIOXHQFH XSRQ KLV YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ SURFHVV ,I WKH PDQDJHU EHOLHYHV WKDW WKH VWDQGDUGV DUH XQUHDOLVWLF LH WKDW WKH VWDQGDUGV KDYH EHHQ SODFHG WRR IDU IURP WKH LQFRQWURO GLVWULEXWLRQf KH PD\ UHVFDOH HLWKHU WKH VWDQGDUGV RU WKH YDULDQFHV WR FRUUHVSRQG ZLWK KLV RZQ SHUFHSWLRQ RI WKH LQFRQWURO GLVWULEXWLRQ ,I LW LV DVVXPHG WKDW D PDQDJHU DFFHSWV WKH VWDQGDUG VHWWLQJ SURn FHVV DV UHDOLVWLF DQG GRHV QRW UHVFDOH WKH VWDQGDUGV RU YDULDQFHV KLV YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV PD\ EH YLHZHG DV WKH FXOPLQDWLRQ RI D WZRVWDJH SURFHVV 7KH ILUVW VWDJH FRQFHUQV WKH GHWHFWLRQ RI WKH SDUWLFXODU GLVWULEXWLRQ HJ LQFRQWURO RU RXWRIFRQWUROf WKDW JHQn HUDWHG WKH YDULDQFH 7KH PDQDJHUnV SHUIRUPDQFH RI WKLV WDVN LV D IXQFWLRQ RI WKH VHQVLWLYLW\ RI KLV GHFLVLRQ SURFHVV PRGHOf 7KLV VHQVLWLYLW\ LV DIIHFWHG E\ WKH VWUXFWXUH RI WKH SDUWLFXODU VLWXDWLRQ DQG E\ WKH PDQDJHUnV NQRZOHGJH RI WKLV VWUXFWXUH 7KH H[WHQW RI WKH PDQDJHUnV NQRZOHGJH RI WKH VLWXDWLRQ LQ WXUQ LV DIIHFWHG E\ WKH DYDLODEOH LQIRUPDWLRQ ERWK FRQWDLQHG ZLWKLQ WKH YDULDQFH UHSRUW DQG SURYLGHG IURP RWKHU VRXUFHVf DQG E\ WKH PDQDJHUnV DELOLW\ WR OHDUQ IURP KLV H[SHULHQFHV ZLWK WKH SURFHVVHV EHLQJ FRQWUROOHG ,Q VLWXDWLRQV ZKHUH WKH FRQWUROOHG SURFHVV GLVWULEXWLRQV KDYH VRPH DUHD RI RYHUODS WKH UHVXOWV RI D PDQDJHUnV GHWHFWLRQ SURFHVV DUH SUREDELOLVWLF LH SULRU WR FRPSOHWLQJ DQ LQYHVWLJDWLRQ WKH H[LVWHQFH RI DQ\ JLYHQ VWDWH LV XQFHUWDLQf )XUWKHUPRUH WKH JUHDWHU WKH DUHD RI RYHUODS WKH PRUH GLIILFXOW WKH GLVFULPLQDWLRQ WDVN EHFRPHV :LWKLQ WKLV FRQWH[W DQ LQFRQWURO GLVWULEXWLRQ FRQFHUQV VWDWLVWLFDO FRQJUXHQFH RI SURGXFWLRQ RXWSXW DQG SODQQHG RXWSXW LQ WHUPV RI FRQWUROODEOH UHVRXUFH XWLOL]DWLRQ

PAGE 13

7KH VHFRQG VWDJH FRQFHUQV WKH PDQDJHUnV LQYHVWLJDWLRQ GHFLVLRQ FULWHULD +DYLQJ DUULYHG DW D FRQFOXVLRQ DOEHLW SUREDELOLVWLFf DERXW WKH GLVWULEXWLRQ WKDW JHQHUDWHG WKH YDULDQFH WKH PDQDJHU PXVW LQWHJUDWH DQG SURFHVV YDULRXV REMHFWLYH IXQFWLRQ SDUDPHWHUV LQ RUGHU WR DUULYH DW KLV YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ WKHVH SDUDPHWHUV FDQ EHORQJ WR HLWKHU WKH PDQDJHUnV REMHFWLYH IXQFWLRQ WKH RUJDQL]DWLRQnV REMHFWLYH IXQFWLRQ RU ERWK LI WKH VDPHf 7KH SRWHQWLDO LQYHVWLJDWLRQ FULWHULD FDQ EH GLYLGHG LQWR WZR PDMRU FDWHJRULHV VWUXFWXUDO FULWHULD DQG EHKDYLRUDO FULWHULD 7KH ILUVW FDWHJRU\ FDQ EH GLYLGHG IXUWKHU LQWR VWUXFWXUDO FRVW FULWHULD DQG VWUXFWXUDO SUREDELOLW\ FULWHULD 7KH VWUXFWXUDO FRVW FULWHULD LQFOXGH VXFK YDULDEOHV DV WKH DGGLWLRQDO FRVW RI RSHUDWLQJ DQ RXWRIFRQWURO SURFHVV JLYHQ WKDW WKH SURFHVV FDQ EH UHWXUQHG WR WKH RULJLQDO LQFRQWURO VWDWHf DQG WKH FRVWV RI YDULDQFH LQYHVWLJDWLRQ DQG FRUUHFWLRQ ZKLFK FDQ GLIIHU GHSHQGLQJ XSRQ WKH DFWXDO VWDWH WKDW JHQHUDWHG WKH YDULDQFHf 7KH VWUXFWXUDO SUREDELOLW\ FULWHULD LQFOXGH VXFK YDULDEOHV DV WKH SUREDELOLW\ WKDW WKH VRXUFH RI WKH YDULDQFH LV FRQWUROODEOH LH WKDW LW FDQ EH UHWXUQHG WR WKH RULJLQDO LQFRQWURO VWDWH WKURXJK PDQDJHULDO DFWLRQf WKH SUREDELOLW\ WKDW WKH SURFHVV ZLOO UHWXUQ WR WKH LQFRQWURO VWDWH ZLWKRXW PDQDJHULDO DFWLRQ DQG WKH SULRU SUREDELOLWLHV RI HDFK VWDWH GLVWULEXWLRQ 7KH EHKDYLRUDO FULWHULD LQFOXGH VXFK YDULDEOHV DV WKH PDQDJHUnV SHUFHSWLRQ RI WKH HIIHFW RI WKH LQYHVWLJDWLRQ XSRQ KLV SHUIRUPDQFH HYDOXDWLRQ DQG UHZDUG VWUXFWXUH DQG WKH PDQDJHUnV SHUFHSWLRQ RI WKH HIIHFW RI WKH GHFLVLRQ XSRQ HPSOR\HH SHUIRUPDQFHV DQG DWWLWXGHV $OWKRXJK WKH YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ SURFHVV KDV EHHQ GHVFULEHG DV WZR VWDJHV WKH PDQDJHU PD\ QRW DFWXDOO\ XWLOL]H VXFK D VHTXHQWLDO VWDJH SURFHVV 7KH PDQDJHUnV DFWXDO GHFLVLRQ SURFHVV LV

PAGE 14

ODEHOHG D KHXULVWLF :LWKLQ WKLV FRQWH[W KHXULVWLF UHIHUV WR WKH OHDUQHG VHW RI UXOHV RU SULQFLSOHV WKDW DUH XWLOL]HG E\ WKH LQGLYLGXDO LQ PDNLQJ WKH SDUWLFXODU GHFLVLRQV UHTXLUHG RI KLP 7KH PDQDJHUnV VSHFLILF KHXULVWLF FDQ EH DIIHFWHG VXEVWDQWLDOO\ E\ KLV LQGLYLGXDO FKDUDFWHULVWLFV HJ LQWHOOLJHQFH FRJQLWLYH FRPSOH[LW\ DQG VW\OH GHFLVLRQ SURFHVV VHQVLWLYLW\ PRWLYDWLRQV HWFf +RZHYHU D JHQHUDO KHXULVWLF ODEHOHG DQFKRULQJ DQG DGMXVWPHQW 7YHUVN\ DQG .DKQHPDU f LV H[SHFWHG WR GHVFULEH WKH JHQHUDO IRUP RI WKH PDQDJHUnV YDULn DQFH LQYHVWLJDWLRQ GHFLVLRQ SURFHVV 7KH DERYH H[DPLQDWLRQ RI WKH PDQDJHUnV YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ SURFHVV LQGLFDWHV WKDW WKH YDULDEOHV ZKLFK FDQ DIIHFW ERWK WKH SURFHVV DQG WKH UHVXOWV RI WKH SURFHVV LQFOXGH f WKH VWUXFWXUH RI WKH GHFLVLRQ VLWXDWLRQ f WKH FRQWHQWV RI WKH DYDLODEOH LQIRUPDWLRQ VHW f WKH PDQDJHUnV LQIRUPDWLRQ SURFHVVLQJ HIILFLHQF\ DQG f WKH PDQDJHUnV OHDUQLQJ HIILFLHQF\ IURP KLV H[SHULHQFHV ZLWK WKH FRQWUROOHG SURFHVVf 7KH PDQDJHUnV NQRZOHGJH FRQFHUQLQJ WKH VWUXFWXUH RI WKH GHFLVLRQ VLWXDWLRQ PD\ EH OLPLWHG E\ WKH FRQWHQWV RI WKH DYDLODEOH LQIRUPDWLRQ VHW DQG E\ KLV LQIRUPDWLRQ SURFHVVLQJ DQG OHDUQLQJ HIILFLn HQFLHV 7KH DFFRXQWDQW WR D ODUJH H[WHQW KDV FRQWURO RYHU WKH FRQn WHQWV RI WKH DYDLODEOH LQIRUPDWLRQ VHW 7KH 5HVHDUFK 2EMHFWLYHV 0DQ\ SUREOHPV LQ DFFRXQWLQJ UHGXFH WR RQH RI FKRRVLQJ DPRQJ DOWHUQDWLYH LQIRUPDWLRQ VHWV WKDW FRXOG EH SURYLGHG WR D GHFLVLRQ PDNHU $PHULFDQ $FFRXQWLQJ $VVRFLDWLRQ f 7ZR EDVLF DSSURDFKHV WR 7KH DQFKRULQJ DQG DGMXVWPHQW KHXULVWLF LV GHVFULEHG LQ JUHDWHU GHWDLO ZLWKLQ &KDSWHU ,,

PAGE 15

GHFLGLQJ RQ WKH LQIRUPDWLRQ VHW WR EH SUHVHQWHG ZLWKLQ WKH FRVW YDULDQFH UHSRUW KDYH EHHQ DGYRFDWHG )LUVW WKH DFFRXQWDQW FDQ GHWHUPLQH WKRVH PRGHOV ZKLFK PDQDJHUV XVH LQ PDNLQJ YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV DQG SURYLGH DQ LQIRUPDWLRQ VHW ZKLFK ZRXOG SHUPLW WKH LPSOHPHQWDWLRQ RI WKHVH PRGHOV 3UREOHPV ZLWK WKLV LQGLYLGXDO PRGHO DSSURDFK DUH f WKH SRVVLELOLW\ H[LVWV WKDW GLIIHUHQW LQGLYLGXDO PRGHOV XVH D ZLGH UDQJH RI GLIIHUHQW LQIRUPDWLRQ VHWV ZKLFK DUH QRW FRVWOHVVf DQG f WKH LQGLn YLGXDO PRGHOV PD\ QRW EH RSWLPDO LH WKH\ FRXOG EH LQHIILFLHQW ZLWK UHVSHFW WR RWKHU GHFLVLRQ PRGHOVf 7KH LQGLYLGXDO PRGHO DSSURDFK FHQn WHUV RQ WKH SV\FKRORJLFDO TXHVWLRQ RI ZKDW WKH PDQDJHU LV GRLQJ RU ZRXOG GRf ZLWK WKH DYDLODEOH RU DGGLWLRQDOf LQIRUPDWLRQ 6HFRQG WKH DFFRXQWDQW FRXOG FUHDWH D QRUPDWLYH PRGHO RI WKH YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ DQG SURYLGH WKH LQIRUPDWLRQ UHTXLUHG E\ WKDW PRGHO 3UREOHPV ZLWK WKH QRUPDWLYH PRGHO DSSURDFK DUH f WKH LQIRUPDWLRQ SURYLGHG PD\ QRW RSWLPL]H WKH LQGLYLGXDOnV LQYHVWLJDWLRQ GHFLVLRQV ZKLFK PD\ FRQn WLQXH WR EH PDGH XVLQJ WKH LQGLYLGXDOnV PRGHOf DQG f WKH FRVWV DVVRFLDWHG ZLWK RSHUDWLRQDOL]DWLRQ RI WKH QRUPDWLYH PRGHO PD\ H[FHHG WKH EHQHILWV FRVW VDYLQJV DV D UHVXOW RI PRUH RSWLPDO LQYHVWLJDWLRQ GHFLVLRQV DQG JUHDWHU FRQJUXHQFH RI PDQDJHU JRDOV ZLWK RYHUDOO RUJDQLn ]DWLRQ JRDOVf RIIHUHG E\ WKH PRGHO 7KH QRUPDWLYH PRGHO DSSURDFK FHQWHUV RQ WKH DQDO\WLFDO TXHVWLRQ RI ZKDW WKH PDQDJHU VKRXOG EH GRLQJ RU VKRXOG GRf ZLWK WKH DYDLODEOH RU DGGLWLRQDOf LQIRUPDWLRQ $ VXEVWDQWLDO SRUWLRQ RI WKH DFFRXQWLQJ YDULDQFH LQYHVWLJDWLRQ OLWHUDWXUH KDV IRFXVHG XSRQ WKH VHFRQG DSSURDFK WKH QRUPDWLYH PRGHO DSSURDFK +RZHYHU UDUHO\ KDV DWWHQWLRQ EHHQ JLYHQ WR KRZ WKH PDQDJHU 7KLV OLWHUDWXUH LV UHYLHZHG ZLWKLQ &KDSWHU ,,

PAGE 16

ZRXOG LQWHUSUHWH DQG LQWHJUDWH WKH LQIRUPDWLRQ VHWV RI WKH YDULRXV QRUPDWLYH RSWLPDOf YDULDQFH LQYHVWLJDWLRQ PRGHOV ,Q PDQ\ LQVWDQFHV WKH LPSOLFLW DVVXPSWLRQ KDV EHHQ WKDW WKH PDQDJHU ZRXOG SURFHVV WKH LQIRUPDWLRQ ZLWK WKH VDPH HIILFLHQF\ DV WKH QRUPDWLYH PRGHO 7KH SULQFLSDO IRFXV RI WKLV UHVHDUFK LV WKH HIIHFWV RI VSHFLILF VLWXDWLRQDO YDULDEOHV RQ D PDQDJHUnV YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV UHODWLYH WR WKH LQYHVWLJDWLRQ GHFLVLRQV RI DQ RSWLPDO PRGHOf DQG RQ D PDQDJHUnV SURFHVVLQJ RI DYDLODEOH LQIRUPDWLRQ UHODWLYH WR WKH LQIRUn PDWLRQ SURFHVVLQJ RI DQ RSWLPDO PRGHOf 6SHFLILF REMHFWLYHV DUH f 7R GHYHORS D FRQFHSWXDO IUDPHZRUN ZKLFK ZLOO SUHGLFW WKH HIIHFWV RI VLWXDWLRQDO YDULDEOHV RQ D PDQDJHUnV UHODWLYH LQIRUPDWLRQ SURFHVVLQJ HIILFLHQF\ DQG WR HPSLULFDOO\ WHVW WKH LPSOLFDWLRQV RI WKLV FRQFHSWXDO IUDPHZRUN f 7R GHYHORS D FRQFHSWXDO IUDPHZRUN ZKLFK ZLOO SUHGLFW WKH HIIHFWV RI VLWXDWLRQDO YDULDEOHV RQ D PDQDJHUnV UHODWLYH YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ HIILFLHQF\ DQG WR HPSLULFDOO\ WHVW WKH LPSOLFDWLRQV RI WKLV FRQFHSWXDO IUDPHZRUN 'LVVHUWDWLRQ 2UJDQL]DWLRQ &KDSWHU ,, SUHVHQWV FHUWDLQ JHQHUDO FRQFHSWV IURP DFFRXQWLQJ DQG SV\FKRORJ\ 7KHVH FRQFHSWV DUH QHFHVVDU\ IRU WKH GHYHORSPHQW RI WKH VSHFLILF HQYLURQPHQW VWXGLHG LQ WKLV UHVHDUFK DQG IRU WKH GHYHORSPHQW RI D FRQFHSWXDO LQIRUPDWLRQ SURFHVVLQJ DQG GHFLVLRQ PDNLQJ IUDPHZRUN ZLWKLQ WKLV HQYLURQPHQW 7KH DFFRXQWLQJ FRQFHSWV DUH FRQFHUQHG ZLWK WKH QDWXUH RI PDQDJHULDO GHFLVLRQV PDQDJHULDO WDVN SODQQLQJ DQG FRQWURO DQG VWDQGDUG FRVW YDULDQFH LQYHVWLJDWLRQ 7KH SV\FKRORJLFDO FRQFHSWV DUH FRQFHUQHG ZLWK WKH WKHRU\ RI VLJQDO GHWHFWLRQ DQG ZLWK FRJQLWLYH

PAGE 17

DVSHFWV RI KXPDQ LQIRUPDWLRQ SURFHVVLQJ VXFK DV WKH JHQHUDO KHXULVWLF RI DQFKRULQJ DQG DGMXVWPHQW &KDSWHU ,,, V\QWKHVL]HV WKH JHQHUDO FRQFHSWV RI WKH SUHYLRXV FKDSWHU DQG GHYHORSV D FRQFHSWXDO IUDPHZRUN RI VWDQGDUG FRVW YDULDQFH LQYHVWLJDWLRQ HPSOR\LQJ WKH PHWKRGRORJ\ RI WKH SV\FKRORJLFDO FRQFHSWV 7KLV FRQFHSWXDO IUDPHZRUN ZDV RSHUDWLRQDOL]HG XVLQJ VLPXODWLRQ WHFKn QLTXHV DQG H[SHULPHQWDO K\SRWKHVHV ZHUH GHULYHG IURP WKH VLPXODWLRQV 7KH GHWDLOV RI D ODERUDWRU\ H[SHULPHQW GHVLJQHG WR WHVW WKH K\SRWKHVHV GHULYHG IURP WKH FRQFHSWXDO YDULDQFH LQYHVWLJDWLRQ IUDPHZRUN DUH SUHVHQWHG LQ &KDSWHU ,9 7KH H[SHULPHQW HPSOR\HG D EHWZHHQVXEMHFWV GHVLJQ DQG ZDV PDQLSXODWHG IDFWRULDOO\ XVLQJ VSHFLILF YDULDQFH LQYHVWLn JDWLRQ VLWXDWLRQ YDULDEOHV 0RGLILHG SDUDPHWHUV IURP WKH SV\FKRORJLFDO WKHRU\ RI VLJQDO GHWHFWLRQ ZHUH HPSOR\HG WR PHDVXUH VXEMHFWnV LQIRUPDWLRQ SURFHVVLQJ HIILFLHQF\ DQG GHFLVLRQ PRGHO VHQVLWLYLW\ 7KH UHVXOWV REWDLQHG IURP WKH ODERUDWRU\ H[SHULPHQW DUH SUHVHQWHG LQ &KDSWHU 9 7KH PRGHO FRPSDULVRQ SURFHGXUH RI QRQRUWKRJRQDO DQDO\VLV RI YDULDQFH ZDV WKH SULPDU\ PHWKRG RI DQDO\VLV HPSOR\HG &KDSWHU 9, GLVFXVVHV WKH UHVXOWV GHYHORSV D PRGLILFDWLRQ H[ SRVWf RI WKH RULJLQDO FRQFHSWXDO IUDPHZRUN DQG GLVFXVVHV VRPH LPSOLFDWLRQV IRU DFFRXQWLQJ DQG IRU IXWXUH DFFRXQWLQJ UHVHDUFK

PAGE 18

&+$37(5 ,, $5($ 2) ,19(67,*$7,21 0DQDJHULDO $FFRXQWLQJ &RQFHSWV 7KLV VWXG\ UHOLHV RQ WKH V\QWKHVLV RI FHUWDLQ SV\FKRORJLFDO FRQn FHSWV ZLWK FHUWDLQ DFFRXQWLQJ FRQFHSWV 7KLV VHFWLRQ GLVFXVVHV WKH DFFRXQWLQJ FRQFHSWV WKH QDWXUH RI PDQDJHULDO GHFLVLRQV PDQDJHULDO WDVN SODQQLQJ DQG FRQWURO DQG VWDQGDUG FRVW YDULDQFH LQYHVWLJDWLRQ 0DQDJHULDO 'HFLVLRQV 6LQFH WKH LQIRUPDWLRQ VHW DFFRXQWLQJ UHSRUWf H[LVWV WR VXSSRUW WKH PDQDJHUnV WDVNV D FRQFHSWXDO IUDPHZRUN RI PDQDJHULDO GHFLVLRQV ZRXOG IDFLOLWDWH WKH DFFRXQWDQWnV LQIRUPDWLRQ VHW VHOHFWLRQ $ FRQFHSWXDO IUDPHZRUN RI PDQDJHULDO GHFLVLRQV VKRXOG GLIIHUHQWLDWH PDQDJHULDO GHn FLVLRQV DORQJ GLPHQVLRQV WKDW DOORZ LQVLJKWV LQWR WKH LQIRUPDWLRQDO QHHGV RI WKRVH GHFLVLRQV $QWKRQ\ f SURYLGHV RQH VXFK GLPHQVLRQ DORQJ ZKLFK KH LGHQn WLILHV WKH SXUSRVHV RU RULHQWDWLRQ RI PDQDJHULDO DFWLYLWLHV VWUDWHJLF SODQQLQJ PDQDJHPHQW FRQWURO DQG RSHUDWLRQDO FRQWURO $QRWKHU GLPHQn VLRQ LV SURYLGHG E\ 6LPRQ f ZKR GLVWLQJXLVKHV EHWZHHQ SURJUDPPHG DQG QRQSURJUDPPHG GHFLVLRQV 7KH XQGHUO\LQJ GLPHQVLRQ LV FRQFHUQHG ZLWK WKH PDQQHU LQ ZKLFK PDQDJHUV GHDO ZLWK WKHLU SUREOHPV 7KH FULn WHULD IRU FODVVLI\LQJ D GHFLVLRQ FRQVLVW RI WKH H[WHQW RI VWUXFWXUH DVVRFLDWHG ZLWK WKH SUREOHP VROYLQJ SKDVHV RI WKH GHFLVLRQ

PAGE 19

$ FRQFHSWXDO IUDPHZRUN RI PDQDJHULDO GHFLVLRQV WKDW V\QWKHVL]HV WKHVH GLPHQVLRQV LV SUHVHQWHG E\ *RUU\ DQG 0RUWRQ f 6LPRQnV GLPHQVLRQ LV PRGLILHG E\ UHSODFLQJ WKH WHUPV SURJUDPPHG DQG QRQ SURJUDPPHG ZLWK WKH WHUPV VWUXFWXUHG DQG XQVWUXFWXUHG DQG DGGLQJ D WKLUG FDWHJRU\ ODEHOHG VHPLVWUXFWXUHG *RUU\ DQG 0RUWRQnV FRQFHSWXDO IUDPHZRUN LV XVHG LQ WKLV UHVHDUFK DV D PRGHO WKDW ZLOO SHUPLW WKH LGHQWLILFDWLRQ RI D VWDQGDUG FRVW V\VWHP ZLWKLQ WKH RYHUDOO PDQDJHULDO GHFLVLRQ IUDPHZRUN 0DQDJHPHQW 7DVN 3ODQQLQJ DQG &RQWURO 6WDQGDUG FRVW V\VWHPV SUHVHQW LQIRUPDWLRQ VHWV WR PDQDJHUV WR VXSSRUW YDULRXV GHFLVLRQV FRQFHUQLQJ WDVN SODQQLQJ DQG FRQWURO 'HPVNL f SURYLGHV DQ H[FHOOHQW FRQFHSWXDO GLVFXVVLRQ RI WKH PDQDJHn PHQW WDVN SODQQLQJ DQG FRQWURO SURFHVV DQG KLV DSSURDFK ZLWK VRPH PRGLILFDWLRQf LV DGRSWHG LQ WKLV UHVHDUFK )LJXUH SUHVHQWV D PRGHO RI WKH PDQDJHPHQW WDVN SODQQLQJ DQG FRQWURO SURFHVV (QYLURQPHQWDO LQIRUPDWLRQ ODUJHO\ H[WHUQDO WR WKH VWDQGDUG FRVW V\VWHP IDFLOLWDWHV WKH SODQQLQJ RI RYHUDOO JRDOV DQG SROLFLHV ZLWKLQ WKH FRQVWUDLQWV LPSRVHG E\ WKH HQYLURQPHQW 2YHUDOO REMHFWLYH FRQWURO IHHGEDFN f SURYLGHV HYLGHQFH IRU WKH FFQWLQXWHG YDOLGLW\ RI RYHUDOO DVVXPSWLRQV PDGH LQ IRUPLQJ WKH RUJDQL]DWLRQ REn MHFWLYHV DQG f IDFLOLWDWHV VWUDWHJLF SODQQLQJ HYDOXDWLRQ RI RUJDQLn ]DWLRQ SHUIRUPDQFH %RWK WKH HQYLURQPHQWDO YDULDEOHV DQG WKH RYHUDOO WDVN FRQWURO IHHGEDFN RI WKH PDQDJHPHQW FRQWURO DFWLYLW\ f SURYLGH HYLGHQFH IRU WKH FRQWLQXHG YDOLGLW\ RI WKH RYHUDOO WDVN DVVXPSWLRQV PDGH LQ IRUPLQJ

PAGE 20

2YHUDOO 2EMHFWLYH &RQWURO 6WUDWHJLF 3ODQQLQJ 2YHUDOO 2EMHFWLYH 3ODQQLQJ 2YHUDO 7DVN &RQWURO 0DQDJHPHQW &RQWURO 2YHUDOO DQG 6SHFLILF 7DVN 3ODQQLQJ 6SHFLILF 7DVN &RQWURO f9 2SHUDWLRQDO &RQWURO 6SHFLILF 7DVN 6WUXFWXULQJ 6RXUFH 0RGLILHG IURP 'HPVNL (QYLURQPHQW 2XWSXW ),*85( 0$1$*(0(17 7$6. 3/$11,1* $1' &21752/ 352&(66

PAGE 21

WKH WDVN SODQV DQG f IDFLOLWDWH PDQDJHPHQW FRQWURO HYDOXDWLRQ RI RSHUDWLRQDO FRQWURO SHUIRUPDQFH 7KH RSHUDWLRQDO FRQWURO PDQDJHU XVHV WKH VSHFLILF WDVN FRQWURO IHHGEDFN WR f GHFLGH ZKHWKHU WKH SK\VLFDO V\VWHP SHUIRUPDQFH LV LQ DJUHHPHQW ZLWK WKH VSHFLILF WDVN SODQ DQG f WR GHFLGH ZKHWKHU WKH WDVN FDQ EH UHVWUXFWXUHG WR EULQJ SHUIRUPDQFH EDFN LQWR DJUHHPHQW ZLWK WKH SODQ LI SHUIRUPDQFH DQG WKH SODQ GLIIHUf $ WDVN WKDW FDQ QRW EH UHVWUXFWXUHG FRXOG KDYH VLJQLILFDQFH IRU WKH RYHUDOO WDVN FRQWURO IHHGn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n DQFH UHSRUW WKH PDQDJHU PXVW GHFLGH WKH QDWXUH RI WKH JLYHQ YDULDQFHV 'RSXFK HW DO f SUHVHQW D FODVVLILFDWLRQ RI VWDQGDUG FRVW YDULDQFHV WKDW LV EDVHG RQ WKH H[SHFWHG VRXUFH RI WKH YDULDQFHV $ YDULDQFH UHVXOWLQJ IURP D UDQGRP IOXFWXDWLRQ RI WKH SK\VLFDO V\VWHP ODEHOHG D 7\SH YDULDQFH UHTXLUHV QR RSHUDWLRQDO FRQWURO UHVSRQVH LI

PAGE 22

QRW VWDWLVWLFDOO\ VLJQLILFDQW :KHWKHU D YDULDQFH UHVXOWLQJ IURP D FKDQJH LQ WKH SK\VLFDO V\VWHP ODEHOHG D 7\SH YDULDQFH UHTXLUHV DQ RSHUDWLRQDO FRQWURO UHVSRQVH GHSHQGV RQ ZKHWKHU WKH XQGHUO\LQJ FDXVH RI WKH FKDQJH LV D WHPSRUDU\ UDWKHU WKDQ D SHUPDQHQW SKHQRPHQD $ WHPn SRUDU\ RU FRQWUROODEOH YDULDQFH ODEHOHG D 7\SH D YDULDQFH LV RQH WKDW RSHUDWLRQDO FRQWURO FDQ FRUUHFW LQ WKH IXWXUH LH WKH SK\VLFDO V\VWHP FDQ EH UHWXUQHG WR WKH SUHYLRXV LQFRQWURO FRQGLWLRQf $ SHUn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n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f WKH PHDQ DQG VWDQGDUG GHYLDWLRQ RI WKH RXWSXW ZKHQ WKH V\VWHP LV NQRZQ WR EH LQFRQWURO f WKH PHDQ DQG VWDQGDUG GHYLDWLRQ RI

PAGE 23

WKH RXWSXW ZKHQ WKH V\VWHP LV NQRZQ WR EH RXWRIFRQWURO f WKH DFWXDO RXWSXW IRU WKH SHULRG DQG f WKH GHYLDWLRQ EHWZHHQ WKH VWDQGDUG RXWSXW DQG WKH DFWXDO RXWSXW IRU WKH SHULRG WKH VWDQGDUG RXWSXW YDULDQFHf 7KH LQYHVWLJDWLRQ GHFLVLRQ FULWHULD FDWHJRU\ FDQ LQFOXGH VXFK YDULDEOHV DV f WKH SULRU SUREDELOLWLHV RI WKH V\VWHPnV VWDWHV DQG f WKH FRVWV DQG EHQHILWV DVVRFLDWHG ZLWK HDFK SRVVLEOH LQYHVWLJDWLRQ GHFLVLRQ LQ FRPELQDWLRQ ZLWK HDFK SRVVLEOH V\VWHP VWDWH ,QIRUPDWLRQ UHODWLQJ WR LQYHVWLJDWLRQ GHFLVLRQ FULWHULD WKDW LV QRW W\SLFDOO\ SUHVHQWHG ZLWKLQ WKH YDULDQFH UHSRUW LQFOXGHV f WKH SUREDELOLW\ WKDW WKH FDXVH RI WKH YDULDQFH LV FRQWUROODEOH DVVXPHG E\ WKLV UHVHDUFK WR HTXDO RQHf f WKH SUREDELOLW\ WKDW WKH V\VWHP ZLOO UHWXUQ WR WKH LQFRQWURO VWDWH ZLWKRXW PDQDJHULDO DFWLRQ DVVXPHG E\ WKLV UHVHDUFK WR HTXDO ]HURf f WKH SHUFHLYHG HIIHFW RI WKH LQYHVWLn JDWLRQ GHFLVLRQ RQ PDQDJHU SHUIRUPDQFH HYDOXDWLRQ DQG f WKH SHUFHLYHG HIIHFW RI WKH LQYHVWLJDWLRQ RQ HPSOR\HH SHUIRUPDQFHV DQG DWWLWXGHV 7KH YDULDQFH LQYHVWLJDWLRQ OLWHUDWXUH LV FRQFHUQHG PDLQO\ ZLWK WKH LQYHVWLJDWLRQ VLJQLILFDQFH RI 7\SH DQG 7\SH D YDULDQFHV $ JHQHUDO DSSURDFK GHVFULEHG ZLWKLQ WKH OLWHUDWXUH LV WKDW RI WKH 6KHZKDUW ; FKDUW SURFHGXUH 3UREVW .RHKOHU /XK -HXUV =DQQHWRV f 7KLV DSSURDFK EDVHG XSRQ FODVVLFDO VWDWLVWLFV LQYROYHV VDPSOLQJ WKH V\VWHP WR FRQVWUXFW DQ LQFRQWURO PHDQ DQG VWDQn GDUG GHYLDWLRQ $UELWUDU\ FRQWURO OLPLWV JHQHUDOO\ SOXV RU PLQXV WKUHH VWDQGDUG GHYLDWLRQVf DUH XVHG DV WKH FULWHULD IRU PDNLQJ D YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ 2QO\ RQH RI WKH DERYH VWXGLHV FRQVLGHUHG WKH HIIHFWV RI WKH LQIRUPDWLRQ SURYLGHG RQ WKH PDQDJHUnV LQYHVWLJDWLRQ GHFLVLRQ 3UREVW f LQ DQ LQGXVWULDO ILHOG VWXG\ IRXQG IRUHPHQ XQZLOOLQJ WR DFFHSW WKH SURFHGXUH DV WKHLU LQYHVWLJDWLRQ GHFLVLRQ UXOH

PAGE 24

ZKHQ WKH SURFHGXUH ZDV FRQVWUXFWHG REMHFWLYHO\ WKH IRUHPHQ LQGLFDWHG WKH UHDVRQ IRU QRW DFFHSWLQJ WKH SURFHGXUH ZDV WKDW WKH\ IHOW LW LJQRUHG WKHLU H[SHULHQFHf +RZHYHU 3UREVW GLG QRW DQDO\]H KLV UHVXOWV LQ WHUPV RI WKH SHUIRUPDQFH HIILFLHQF\ RI WKH IRUHPHQ 2WKHU UHVHDUFK KDV QRWHG ZKDW DUH FRQVLGHUHG WR EH VLJQLILFDQW GUDZEDFNV WR WKH 6KHZKDUW ; FKDUW SURFHGXUH WKH IDLOXUH WR FRQVLGHU WKH FRVWV DQG EHQHILWV DVVRFLDWHG ZLWK WKH YDULRXV LQYHVWLJDWLRQ GHn FLVLRQ RXWFRPHV DQG WKH IDLOXUH WR FRQVLGHU LQIRUPDWLRQ IURP SULRU SHULRGV %LHUPDQ HW DO f SURSRVHG WKH LQFRUSRUDWLRQ RI LQYHVWLn JDWLRQ GHFLVLRQ FRVWV DQG EHQHILWV ZLWKLQ D FODVVLFDO VWDWLVWLFV IUDPHn ZRUN 7KXV PDQDJHUV PD\ XVH VWDWLVWLFDO LQIRUPDWLRQ WR FDOFXODWH WKH SUREDELOLW\ WKDW WKH YDULDQFH UHIOHFWV WKH LQFRQWURO VWDWH DQG FRPn ELQH WKLV SUREDELOLW\ ZLWK WKH LQYHVWLJDWLRQ GHFLVLRQ FRVWV DQG EHQHILWV 0RVW RI WKLV LQIRUPDWLRQ LV DVVXPHG WR EH SURYLGHG E\ WKH DFFRXQWDQW DQG DFFXUDWHO\ SURFHVVHG E\ WKH PDQDJHU .DSODQ f DQG -DFREV f KDYH FRQVLGHUHG DQRWKHU GHILFLHQF\ RI WKH FODVVLFDO VWDWLVWLFV DSSURDFK WKH IDLOXUH WR XVH SULRU SHULRG LQIRUPDWLRQ 7KH\ DQDO\]H WKH XVH RI WKH FXPPXODWLYH VXP SURFHGXUH ZKHUHE\ WKH FXPXODWLYH VXP RI WKH YDULDQFH LV FKDUWHG IRU HDFK SHULRG 7KHRUHWLFDOO\ XQGHU D VWDEOH VWDWH WKHVH VXPV VKRXOG IROORZ D UDQGRP ZDON DQG DQ\ GULIW ZRXOG LQGLFDWH WKH V\VWHP LV RXWRIFRQWURO (YDOn XDWLRQ RI VXFK D GULIW ZRXOG EH DFFRPSOLVKHG RQ WKH EDVLV RI LQIRUPDWLRQ SURYLGHG E\ WKH DFFRXQWDQW RU GHULYHG IURP WKH PDQDJHUnV H[SHULHQFH 7KH HFRQRPLF FXPXODWLYH VXP SURFHGXUH LQFRUSRUDWHV WKH HIIHFWV RI HVWLPDWHG LQYHVWLJDWLRQ GHFLVLRQ FRVWV ZLWKLQ WKH GULIW HYDOXDWLRQ 6HYHUDO VWXGLHV LQ WKH YDULDQFH LQYHVWLJDWLRQ OLWHUDWXUH XVH GHFLVLRQ WKHRU\ WR FRQVWUXFW D YDULDQFH LQYHVWLJDWLRQ PRGHO .DSODQ

PAGE 25

'\FNPDQ .DSODQ 'LWWPDQ DQG 3UDNDVK f ,I DOO WKH SDUDPHWHUV UHTXLUHG E\ WKHVH PRGHOV ZHUH DYDLODEOH WKHVH PRGHOV FRXOG UHSODFH WKH PDQDJHU DV WKH YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ PDNHU +RZHYHU WR WKH H[WHQW WKDW QRW DOO WKH SDUDPHWHUV DUH RSHUD WLRQDOL]DEOH JLYHQ VHPH FRVW FRQVWUDLQWf WKH PDQDJHU PXVW FRQWLQXH WR PDNH WKH YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV XVLQJ WKH LQIRUPDWLRQ SURYLGHG E\ WKH DFFRXQWDQW DQG E\ KLV RZQ H[SHULHQFHV $ FRQFOXVLRQ WKDW FDQ EH GUDZQ IURP WKH YDULDQFH LQYHVWLJDWLRQ OLWHUDWXUH LV WKDW WKH DGGLWLRQDO LQIRUPDWLRQ SURSRVHG IRU WKH XVH RI PDQDJHUV LV EHFRPLQJ ERWK GLYHUVH DQG FRPSOH[ )XUWKHUPRUH WKHUH LV D SDXFLW\ RI UHVHDUFK UHODWLQJ WR WKH PDQDJHUnV DELOLW\ DQG HIILFLHQF\ WR LQWHUSUHW DQG LQWHJUDWH WKLV DGGLWLRQDO LQIRUPDWLRQ 'HPVNL f ZDV RQH RI WKH ILUVW WR QRWH WKH SUREOHP DVVRFLDWHG ZLWK WKH GHFLVLRQLPSOHPHQWDWLRQ LQWHUIDFH LH WKH HIIHFW RI WKH FRQWURO V\VWHP LQIRUPDWLRQ RQ LQGLYLGXDO EHKDYLRU DQG SHUIRUPDQFHf 6RPH DQDO\WLF UHVHDUFK RQ WKLV SUREOHP KDV IROORZHG 8VLQJ VLPXODWLRQ WHFKQLTXHV 0DJHH f FRPSDUHG DYHUDJH WRWDO FRVW IRU VHYHQ GHFLVLRQ UXOHV XQGHU YDULRXV RSHUDWLRQDO FRQGLWLRQV $OWKRXJK WKH VLPSOHU GHFLVLRQ PRGHOV UXOHVf WHQGHG WR KDYH ODUJHU DYHUDJH FRVWV WKDQ WKH PRUH VRSKLVWLFDWHG GHFLVLRQ PRGHOV WKH GLIIHUHQFH LQ DYHUDJH FRVW ZDV UHODWLYHO\ VPDOO ,QGHHG LI PRGHO LPSOHPHQWDWLRQ DQG LQIRUPDWLRQ FRVWV DUH FRQVLGHUHG WKH VLPSOHU PRGHOV PD\ EH PRUH HIILFLHQW WKDQ WKH VRSKLVWLFDWHG PRGHOV 0DJHH DOVR QRWHG WKDW EHFDXVH RI WKH XVH RI GLIIHUHQW PDQDJHU SHUIRUPDQFH PHDVXUHV DYHUDJH RSHUDWLQJ FRVWV DYHUDJH QXPEHU RI PRQWKV EHORZ VWDQGDUG HWFf VLPSOHU GHFLVLRQ PRGHOV PD\ UHn VXOW LQ UDWLRQDO FKRLFHV LH FKRLFHV WKDW PD[LPL]H WKH PDQDJHUnV H[SHFWHG XWLOLW\f &ORVHO\ UHODWHG WR 0DJHHnV UHVHDUFK LV DQ HPSLULFDO

PAGE 26

VWXG\ E\ 0DJHH DQG 'LFNKDXW f 8VLQJ KXPDQ VXEMHFWV 0DJHH DQG 'LFNKDXW IRXQG VXSSRUW IRU WKH SURSRVLWLRQ WKDW GLIIHUHQW PDQDJHU SHUn IRUPDQFH SD\RIIf PHDVXUHV ZLOO DIIHFW WKH VSHFLILF GHFLVLRQ UXOHV KHXULVWLFVf HPSOR\HG E\ WKH GHFLVLRQ PDNHU DV D UHVXOW RI WKH GHFLVLRQ PDNHU DWWHPSWLQJ WR PD[LPL]H KLV VXEMHFWLYH XWLOLW\f 3V\FKRORJLFDO &RQFHSWV 7KLV VHFWLRQ GLVFXVVHV WKH SV\FKRORJLFDO FRQFHSWV HPSOR\HG LQ GHYHORSLQJ D FRQFHSWXDO IUDPHZRUN RI PDQDJHU VWDQGDUG FRVW YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV 7KHVH FRQFHSWV LQFOXGH WKH SV\FKRSK\VLFDO WKHRU\ RI VLJQDO GHWHFWLRQ DQG WKH KXPDQ LQIRUPDWLRQ SURFHVVLQJ FRQn FHSWV RI GHFLVLRQ KHXULVWLFV 3V\FKRSK\VLFV 3V\FKRSK\VLFV VWXGLHV WKH UHODWLRQVKLSV EHWZHHQ SK\VLFDO DQG SV\FKRORJLFDO VFDOHV RI PHDVXUHPHQW 0RGHUQ SV\FKRSK\VLFV DGRSWV WKH YLHZ WKDW VXEMHFWV FDQ PDNH PHDQLQJIXO HYDOXDWLRQV RI WKH PDJQLWXGHV RI WKHLU VHQVRU\ H[SHULHQFHV DQG WKHUHIRUH VHQVRU\ PDJQLWXGHV DV ZHOO DV SK\VLFDO PDJQLWXGHV FDQ EH TXDQWLILHG 2QH DSSURDFK RI PRGHUQ SV\FKRSK\VLFV LV EDVHG XSRQ WKH WKHRU\ RI VLJQDO GHWHFWLRQ 76'f 76' SHUPLWV WKH VHSDUDWLRQ RI WKH GHFLVLRQ PDNHUnV DELOLW\ WR GLVFULPLQDWH EHWZHHQ FODVVHV RI VWLPXOL VHQVLWLYLW\f IURP KLV PRWLYDWLRQDO UHVSRQVH ELDVHV GHFLVLRQ FULWHULDf 7ZR FRPSUHKHQVLYH WKHRUHWLFDO GHVFULSWLRQV RI 76' DUH SUHVHQWHG E\ *UHHQ DQG 6ZHWV f DQG (JDQ f JHQHUDO VXUYH\V RI WKH 76' WKHRU\ DUH SUHVHQWHG E\ &RRPEV HW DO f :DWVRQ f DQG 3DVWRUH DQG 6FKHLUHU f

PAGE 27

7KH EDVLF 76' H[SHULPHQW XWLOL]HV WKH VLQJOHLQWHUYDO SURFHGXUH 7KLV SURFHGXUH FRQVLVWV RI D VHULHV RI WULDOV HDFK WULDO FRQVLVWLQJ RI DQ REVHUYDWLRQ LQWHUYDO DQG D UHVSRQVH LQWHUYDO 7KH SRVVLEOH VWLPXOXV HYHQWV GXULQJ WKH REVHUYDWLRQ LQWHUYDO DUH f WKH REVHUYDWLRQ FRQWDLQV D PHDQLQJIXO VLJQDO DGGHG WR D EDFNJURXQG RI QRLVH VQ WULDOf DQG f WKH REVHUYDWLRQ FRQWDLQV RQO\ D EDFNJURXQG RI QRLVH Q WULDOf ,W LV DVVXPHG WKDW HDFK WULDO LV LQGHSHQGHQW RI DOO RWKHU WULDOV DQG WKDW WKH SULRU SUREDELOLWLHV RI Q DQG VQ DUH JLYHQ DQG UHPDLQ FRQVWDQW 7KH EDFNJURXQG QRLVH IOXFWXDWHV DW UDQGRP IURP WULDO WR WULDO WKH VWLPXOXV XVXDOO\ D IL[HG OHYHOf LV DGGHG WR WKH QRLVH 7KHUHIRUH WKH REVHUn YDWLRQ IOXFWXDWHV UDQGRPO\ IURP WULDO WR WULDO 7KH WDVN RI WKH VXEMHFW LV WR GHWHFW ZKHWKHU WKH REVHUYDWLRQ ZDV JHQHUDWHG E\ WKH VLJQDO SOXV QRLVH VQf RU E\ WKH QRLVH DORQH Qf GLVWULEXWLRQ 7KDW LV LQ WKH UHVSRQVH LQWHUYDO WKH VXEMHFW ZLOO UHVSRQG ZLWK HLWKHU
PAGE 28

6LQFH WKH FHOOV RI WKLV PDWUL[ DUH ERWK H[KDXVWLYH DQG PXWXDOO\ H[FOXn VLYH WKH URZZLVH FRQGLWLRQDO SUREDELOLWLHV PXVW VXP WR RQH WKLV GRHV QRW QHFHVVDULO\ KROG IRU WKH VXP RI WKH FROXPQZLVH FRQGLWLRQDO SUREDELOLWLHVf $OO SDUDPHWHUV RI WKH 76' PRGHO DUH GHULYHG IURP WKLV FRQGLWLRQDO SUREDELOLW\ PDWUL[ VHH $SSHQGL[ $ IRU D PRUH GHWDLOHG GLVFXVVLRQ RI WKH FRQGLWLRQDO SUREDELOLW\ PDWUL[ LQFOXGLQJ LWV UHODn WLRQVKLS ZLWK %D\HV WKHRUHPf ,Q WKH VLQJOHLQWHUYDO WDVN WKH VXEMHFW DQDO\]HV WKH HYLGHQFH DQG FODVVLILHV WKH VWLPXOXV LQWR RQH RI WZR FDWHJRULHV DFFRUGLQJ WR KLV FULWHULD 7KH FULWHULD DUH GHWHUPLQHG E\ KLV REMHFWLYH IXQFWLRQ 7KH REMHFWLYH IXQFLWLRQ RI LQWHUHVW ZLWKLQ WKLV UHVHDUFK LV WKH PD[LPL]DWLRQ RI H[SHFWHG YDOXH $VVXPH WKDW WKH VXEMHFW KDV VRPH YDOXH XWLOLW\f ITU HDFK RI WKH IRXU HYHQW RXWFRPHV $ SD\RII PDWUL[ RI WKHVH YDOXHV UHODWHG WR WKH IRXU RXWFRPHV LV JLYHQ E\ WKH IROORZLQJ (JDQ f '(&,6,21 1R
PAGE 29

7KH FULWLFDO YDOXH /[f IRU WKLV GHFLVLRQ UXOH KDV WZR SRVVLEOH YDOXHV D WKHRUHWLFDO YDOXH ZKLFK LV D PHDVXUH RI WKH FULWHULD RI DQ RSWLPDO RU LGHDOf VXEMHFW DQG D VXEMHFWLYH YDOXH ZKLFK LV D PHDVXUH RI WKH FULWHULD RI DQ DFWXDO VXEMHFW 2QH VHW RI 76' PRGHOV DVVXPHV WKDW ERWK FRQGLWLRQDO SUREDELOLW\ GLVWULEXWLRQV DUH *DXVVLDQ LH [_Qf DQG [_VQf DUH QRUPDOO\ GLVWULEXWHG :LWK WKH DGGLWLRQDO DVVXPSWLRQ RI HTXDO YDULDQFH IRU ERWK GLVWULEXWLRQV WKH SDUDPHWHUV ZKLFK PHDVXUH LQGLYLGXDO GLVFULPLQDWLRQ VHQVLWLYLW\ DQG LQGLYLGXDO GHFLVLRQ FULWHULD DUH ODEHOHG G DQG J UHVSHFWLYHO\ 7KH GLVFULPLQDELOLW\ PHDVXUH Gn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n PHDVXUH WKHRUHWLFDOO\ LV LQGHSHQGHQW RI WKH GHFLVLRQ FULWHULD PHDVXUH 7KH GHFLVLRQ FULWHULD PHDVXUH KDV WKH IROORZLQJ WKHRUHWLFDO GHILQLWLRQ L!]VQf !]Qf

PAGE 30

ZKHUH c! f GHQRWHV WKH QRUPDO GHQVLW\ IXQFWLRQ IRU WKH SRLQW LQ SDUHQWKHVHV DQG WKH ] SDUDPHWHUV DUH WKH VDPH DV GHILQHG IRU WKH Gn PHDVXUH $OWKRXJK WKH DVVXPSWLRQ RI HTXDO YDULDQFH QRUPDO GLVWULEXWLRQV LV HPSOR\HG ZLWKLQ WKLV UHVHDUFK VXFK DQ DVVXPSWLRQ LV QRW QHFHVVDU\ WR HPSOR\ 76' (JDQ f GHPRQVWUDWHV WKH XVH RI 76' ZLWK H[SRQHQWLDO GLVWULEXWLRQV FKLVTXDUH GLVWULEXWLRQV %HUQRXOOL GLVWULEXWLRQV DQG 3RLVVRQ GLVWULEXWLRQV *ULHU f GHYHORSV QRQSDUDPHWULF PHDVXUHV RI GLVFULPLQDELOLW\ DQG GHFLVLRQ FULWHULD 7UDGLWLRQDOO\ SV\FKRSK\VLFV KDV HPSOR\HG 76' WR VWXG\ SHUFHSWXDO SURFHVVHV LH VHQVRU\ SURFHVVHV VXFK DV DXGLWLRQ DQG YLVLRQ 2YHU WKH ODVW GHFDGH KRZHYHU 76' KDV EHHQ DSSOLHG WR FRQFHSWXDO SURFHVVHV 7KHVH H[WHQVLRQV WR FRQFHSWXDO SURFHVVHV KDYH LQFOXGHG QXPHULFDO SURFHVVLQJ /LHEOLFK DQG /LHEOLFK +DPPHUWRQ :HLVVPDU HW DO f PHGLFDO GLDJQRVLV /XVWHG /XVWHG 6ZHWV f FRQFHSWXDO MXGJHPHQW 8OHKOD HW DO D 8OHKOD HW DO Ef DQG PHPRU\ %HUQEDFK %DQNV f +XPDQ ,QIRUPDWLRQ 3URFHVVLQJ +XPDQ LQIRUPDWLRQ SURFHVVLQJ +,3f D VXEVHW RI FRJQLWLYH SV\n FKRORJ\ VWXGLHV KXPDQ MXGJHPHQW DQG GHFLVLRQ PDNLQJ ZLWK SDUWLFXODU HPSKDVLV RQ WKH SURFHVVLQJ RI LQIRUPDWLRQ WKDW GHWHUPLQHV WKHVH DFWLYLWLHV $Q DUHD ZLWKLQ +,3 LV WKH FRQVWUXFWLRQ RI PRGHOV RI KXPDQ GHFLVLRQ PDNLQJ 7KH ZRUN ZLWKLQ WKLV DUHD LV W\SLFDOO\ FODVVLILHG LQWR VFKRROV RI UHVHDUFK ZKLFK HPSOR\ GLIIHUHQW SDUDGLJPV WKH WZR PDMRU SDUDGLJPV EHLQJ WKH %D\HVLDQ DQG UHJUHVVLRQ DSSURDFKHV 6ORYLF DQG /LFKWHQVWHLQ f 7KH 76' PRGHO LV UHODWHG WR WKH %D\HVLDQ DSSURDFK

PAGE 31

7KH %D\HVLDQ DSSURDFK LV D QRUPDWLYH PRGHO VSHFLI\LQJ KRZ D GHFLVLRQ VKRXOG EH PDGH JLYHQ FHUWDLQ LQWHUQDOO\ FRQVLVWHQW UHODWLRQn VKLSV DPRQJ SUREDELOLVWLF EHOLHIV 7KH EDVLF EHOLHIV RI WKLV DSSURDFK DUH WKDW GHFLVLRQV VKRXOG EH EDVHG RQ VXEMHFWLYH SUREDELOLWLHV DQG WKDW WKHVH SUREDELOLWLHV VKRXOG EH UHYLVHG XSRQ WKH UHFHLSW RI DGGLn WLRQDO LQIRUPDWLRQ LQ DFFRUGDQFH ZLWK %D\HVnV WKHRUHP 7KH PDMRU ILQGLQJV RI %D\HVLDQ UHVHDUFK DUH ODEHOHG FRQVHUYDWLVP 7KH VXEMHFWV DIWHU UHFHLYLQJ DGGLWLRQDO LQIRUPDWLRQ UHYLVH WKHLU SRVWHULRU SUREDELOLWLHV LQ WKH VDPH GLUHFWLRQ DV WKH RSWLPDO PRGHO EXW WKH UHYLVLRQ LV LQVXIILFLHQW 0XFK RI WKH UHVHDUFK KDV IRFXVHG RQ DQ H[SODQDWLRQ RI WKH FDXVH RI FRQVHUYDWLVP WKH PDMRU H[SODQDWLRQV EHLQJ PLVSHUFHSWLRQ 3HWHUVRQ HW DO f PLVDJJUHJDWLRQ 'X&KDUPH DQG 3HWHUVRQ f DQG UHVSRQVH ELDV 'X&KDUPH f $QRWKHU DUHD ZLWKLQ +,3 LV WKH VWXG\ RI VXEMHFWLYH LQIRUPDWLRQ SURFHVVLQJ SULQFLSOHV DQG GHFLVLRQ UXOHV ODEHOHG KHXULVWLFV $ KHXULVWLF ZLWKLQ WKLV FRQWH[W UHIHUV WR D OHDUQHG VHW RI UXOHV RU SULQFLSOHV ZKLFK DUH XWLOL]HG E\ LQGLYLGXDOV LQ PDNLQJ WKH SDUWLFXODU GHFLVLRQV UHTXLUHG RI WKHP 5HVHDUFK LQ WKLV DUHD KDV EHHQ FRQFHUQHG ZLWK LGHQWLI\LQJ V\VWHPDWLF ELDVHV UHODWLYH WR VRPH GHILQLWLRQ RI RSWLPDOf RI VXEMHFWLYH KHXULVWLFV ZLWKLQ FHUWDLQ W\SHV RI GHFLVLRQ WDVNV 7KRVH LQIRUPDWLRQ SURFHVVLQJ DQG GHFLVLRQ UXOH ELDVHV LGHQn WLILHG WKXV IDU KDYH EHHQ ODEHOHG DV DQ DQFKRULQJ DQG DGMXVWPHQW KHXULVWLF 7YHUVN\ DQG .DKQHPDQ f D UHSUHVHQWDWLYH KHXULVWLF .DKQHPDQ DQG 7YHUVN\ 6ZLHULQJD HW DO f DQ DYDLO DEOLOLW\ KHXULVWLF 7YHUVN\ DQG .DKQHPDQ f DQG WKH ODZ RI VPDOO QXPEHUV 7YHUVN\ DQG .DKQHPDQ f

PAGE 32

7KLV VWXG\ ZLOO PDNH SDUWLFXODU XVH RI WKH DQFKRULQJ DQG DGMXVWPHQW KHXULVWLF ,Q PDQ\ VLWXDWLRQV LQGLYLGXDOV ILUVW PDNH GHFLVLRQV E\ VWDUWLQJ ZLWK DQ LQLWLDO DQFKRU GHFLVLRQ SRLQWf DQG WKHQ DGMXVW WKLV LQLWLDO DQFKRU DV WKH\ OHDUQ IURP WKHLU H[SHULHQFHV 7KH LQLWLDO DQFKRU FDQ EH VXJJHVWHG E\ WKH VWUXFWXUH RI WKH GHFLVLRQ VLWXDWLRQ RU FDQ EH WKH UHVXOW RI D SDUWLDO FRPSXWDWLRQ RU HVWLPDWH (PSLULFDO WHVWV LQYROYLQJ WKH DQFKRULQJ DQG DGMXVWPHQW KHXULVWLF LQn GLFDWH LQGLYLGXDOV GR QRW VXIILFLHQWO\ DGMXVW WKHLU LQLWLDO GHFLVLRQ SRLQW 7KDW LV WKHLU DGMXVWPHQW LV OHVV WKDQ WKDW ZKLFK ZRXOG DOORZ RSWLPDO SURFHVVLQJ RI WKH DYDLODEOH LQIRUPDWLRQ 6ORYLF DQG /LFKWHQVWHLQ 6ORYLF $OSHUW DQG 5DLIID 7YHUVN\ DQG .DKQHPDQ f

PAGE 33

&+$37(5 ,,, 5(6($5&+ 0(7+2'2/2*< $1' '(6,*1 *HQHUDO &RQFHSWXDO 'HYHORSPHQW $ JHQHUDO SUREOHP FRQIURQWLQJ GHFLVLRQ PDNHUV LV WKDW WKH LQGLn YLGXDO PXVW GHFLGH VXEMHFWLYHO\ ZKLFK VWDWH RI QDWXUH LV PRVW SUREDEOH EDVHG XSRQ VRPH LQFRPSOHWH VHW RI LQIRUPDWLRQ :KHQ DQ LQGLYLGXDO GHDOV UHSHWLWLYHO\ ZLWK D VLPLODU VLWXDWLRQ KLV ORQJUXQ GHFLVLRQ HIILFLHQF\ RU GHFLVLRQ FRUUHFWQHVV FDQ EH DIIHFWHG E\ f WKH VWUXFn WXUH RI WKH SDUWLFXODU GHFLVLRQ VLWXDWLRQ f WKH FRQWHQWV RI WKH VHW RI DYDLODEOH LQIRUPDWLRQ f KLV HIILFLHQF\ LQ SURFHVVLQJ WKH DYDLODEOH LQIRUPDWLRQ DQG f KLV DELOLW\ WR H[SDQG WKH DYDLODEOH LQIRUPDWLRQ WKURXJK H[SHULHQFH ZLWK DQG REVHUYDWLRQV RI WKH YDULRXV VWDWHV RI QDWXUH (DFK RI WKHVH IRXU HOHPHQWV LV GLVFXVVHG EHORZ ZLWK VSHFLILF UHIHUHQFH WR WKH REMHFWLYHV RI WKLV UHVHDUFK 'HFLVLRQ 6LWXDWLRQ 6WUXFWXUH 7KH VWUXFWXUH RI WKH SDUWLFXODU GHFLVLRQ VLWXDWLRQ SULPDULO\ GHn SHQGV XSRQ VHYHUDO NH\ YDULDEOHV 7KHVH YDULDEOHV LQFOXGH WKH QXPEHU RI SRVVLEOH VWDWHV RI QDWXUH WKH UHODWLYH IUHTXHQFLHV RI WKH VWDWHV WKH YDULRXV VWDWLVWLFDO UHODWLRQVKLSV DPRQJ WKH VWDWHV DQG WKH UHODWLRQn VKLSV EHWZHHQ WKH YDULRXV GHFLVLRQ RXWFRPHV WKH FRVWV LQFXUUHG JLYHQ D VSHFLILF GHFLVLRQ DQG WKH H[LVWHQFH RI D VSHFLILF VWDWHf 'HSHQGLQJ RQ

PAGE 34

WKH VSHFLILF YDOXHV ZKLFK WKHVH YDULDEOHV PD\ DVVXPH WKH VWUXFWXUH RI WKH SDUWLFXODU GHFLVLRQ VLWXDWLRQ FDQ DIIHFW ERWK WKH GLIILFXOW\ DQG WKH LPSRUWDQFH RI WKH LQGLYLGXDOnV GLVFULPLQDWLRQ DPRQJ WKH VWDWHV ,Q JHQHUDO DV WKH QXPEHU RI SRVVLEOH VWDWHV LQFUHDVHV DQG DV WKH DUHD RI GLVWULEXWLRQDO RYHUODS RI WKHVH VWDWHV LQFUHDVHV WKH GLVFULPLQDWLRQ WDVN EHFRPHV PRUH GLIILFXOW )XUWKHUPRUH WKH GLVFULPLQDWLRQ WDVN EHFRPHV PRUH LPSRUWDQW LQ WHUPV RI LQFXUUHG FRVWVf DV WKH UHODWLYH IUHTXHQFLHV RI WKH VWDWHV EHFRPH HTXDO DQG DV WKH FRVWV DVVRFLDWHG ZLWK WKH SRVVLEOH GHFLVLRQ RXWFRPHV ZKLFK LQYROYH GHFLVLRQ HUURUV DQ LQFRUUHFW GHFLVLRQ JLYHQ WKH H[LVWHQFH RI D VSHFLILF VWDWHf EHFRPH XQHTXDO 7KH JHQHUDO VLWXDWLRQ HPSOR\HG LQ WKLV UHVHDUFK VHOHFWHG IRU LWV UHOHYDQFH WR WKH DFFRXQWLQJ GLVFLSOLQH LV WKH FRVW YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ ,Q WKLV VLWXDWLRQ WZR VWDWHV RI QDWXUH DUH SRVVLEOH f LQFRQWURO WKH XQGHUO\LQJ SK\VLFDO SURFHVV GHVFULEHG E\ WKH VWDQGDUG FRVW YDULDQFH UHSRUW LV IXQFWLRQLQJ DV SODQQHGf DQG f RXWRIFRQWURO WKH XQGHUO\LQJ SK\VLFDO SURFHVV GHVFULEHG E\ WKH VWDQn GDUG FRVW YDULDQFH UHSRUW LV QRW IXQFWLRQLQJ DV SODQQHGf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

PAGE 35

$YDLODEOH ,QIRUPDWLRQ 6HW 7KH FRQWHQWV RI WKH DYDLODEOH LQIRUPDWLRQ VHW UHIHUV WR WKH LQIRUn PDWLRQ NQRZQ E\ WKH LQGLYLGXDO SULRU WR KLV GHFLVLRQ 6XFK LQIRUPDWLRQ FDQ EH RI WZR W\SHV VLQJXODU RU GLVWULEXWLRQDO 7YHUVN\ DQG .DKQHPDQ f 6LQJXODU LQIRUPDWLRQ FRQVLVWV RI LQIRUPDWLRQ WKDW VSHFLILFDOO\ UHODWHV WR WKH FXUUHQW GHFLVLRQ 'LVWULEXWLRQDO LQIRUPDWLRQ FRQVLVWV RI LQIRUPDWLRQ WKDW UHODWHV WR WKH UHODWLYH IUHTXHQFLHV DQG WR WKH VWDWLVWLFDO UHODWLRQVKLSV DPRQJ WKH VWDWHV RI QDWXUH 7KH GLIILFXOW\ RI WKH GLVFULPLQDWLRQ WDVN PD\ EH DIIHFWHG E\ WKH SUHVHQFH RU DEVHQFH RI FHUWDLQ LWHPV RI LQIRUPDWLRQ ,Q JHQHUDO WKH OHVV LQIRUPDWLRQ FRQn WDLQHG LQ WKH DYDLODEOH VHW WKH PRUH GLIILFXOW WKH GLVFULPLQDWLRQ WDVN IRU PLVVLQJ LQIRUPDWLRQ UHTXLUHG E\ D GHFLVLRQ PRGHO PXVW EH HVWLPDWHG E\ WKH LQGLYLGXDO &RPSDUHG WR VWDWLVWLFDOO\ GHULYHG HVWLPDWHV WKHVH VXEMHFWLYH HVWLPDWHV DUH OLNHO\ WR KDYH JUHDWHU XQFHUWDLQW\ DQG LQn HIILFLHQF\ DVVRFLDWHG ZLWK WKHP :LWKLQ WKH YDULDQFH LQYHVWLJDWLRQ VLWXDWLRQ WKH LQIRUPDWLRQ FRQn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

PAGE 36

GLVWULEXWLRQDO LQIRUPDWLRQ LWHPV WKH VWDWLVWLFDO PHDQV YDULDQFHV DQG GLVWULEXWLRQDO VKDSHV RI WKH WZR VWDWHVf LV PDQLSXODWHG DV DQ LQGHSHQn GHQW YDULDEOH 7KH SUHVHQFH RI RWKHU GLVWULEXWLRQDO LQIRUPDWLRQ LWHPV WKH UHODWLYH IUHTXHQFLHV DQG WKH DOORZHG VWDQGDUGf LV FRQWUROOHG DV D FRQVWDQW ,QGLYLGXDO ,QIRUPDWLRQ 3URFHVVLQJ (IILFLHQF\ 7KH LQGLYLGXDOnV HIILFLHQF\ LQ SURFHVVLQJ WKH DYDLODEOH LQIRUn PDWLRQ UHODWHV WR WKH SDUWLFXODU KHXULVWLFV RU VWUDWHJLHV HPSOR\HG LQ FRPELQLQJ DQG ZHLJKWLQJ WKH YDULRXV LWHPV RI LQIRUPDWLRQ 7KH WHUP HIILFLHQF\ LPSOLHV D UHODWLRQVKLS EHWZHHQ WKH LQGLYLGXDOnV SURFHVV RXWSXW KLV GHFLVLRQf DQG D QRUPDWLYHO\ FRUUHFW RU RSWLPDO GHFLVLRQ 7KH LQGLYLGXDOnV GHFLVLRQ DQG LQIRUPDWLRQ SURFHVVLQJ SHUIRUPDQFH FDQ EH HYDOXDWHG E\ FRPSDULQJ KLV SHUIRUPDQFH DJDLQVW DQ RSWLPDO PRGHO 2SWLPDOLW\ UHIHUV WR WKH EHVW SRVVLEOH SHUIRUPDQFH XQGHU JLYHQ FRQGLn WLRQV 6LQFH WKH RSWLPDO GHFLVLRQ PRGHO UHOLHV RQ DQ LQFRPSOHWH LQIRUn PDWLRQ VHW UDWKHU WKDQ FHUWDLQ NQRZOHGJHHYHQ LWV SHUIRUPDQFH FDQ EH DIIHFWHG E\ ERWK WKH VWUXFWXUH RI WKH VLWXDWLRQ DQG WKH FRQWHQWV RI WKH DYDLODEOH LQIRUPDWLRQ VHW 7KH RSWLPDO PRGHOV XVHG LQ WKLV UHVHDUFK DUH D IXQFWLRQ RI WKH H[SHULPHQWDO HQYLURQPHQW WKH VLQJOHLQWHUYDO SURFHGXUH IRXQG ZLWKLQ WKH SV\FKRORJLFDO WKHRU\ RI VLJQDO GHWHFWLRQ DQG WKH EDVLF GHFLVLRQ VLWXDWLRQ RI VWDQGDUG YDULDQFH LQYHVWLJDWLRQ :LWKLQ WKLV HQYLURQPHQW WKH RSWLPDO GHFLVLRQ UXOH IRU PLQLPL]LQJ PD[LPL]LQJf WKH H[SHFWHG FRVW YDOXHf RI D VHW RI GHFLVLRQV LV EDVHG XSRQ DQ H[WHQVLRQ RI %D\HVn WKHRUHP WKDW WDNHV LQWR DFFRXQW WKH UHODWLYH FRVWV YDOXHVf RI YDULRXV

PAGE 37

SRVVLEOH GHFLVLRQ RXWFRPHV 7KH SDUDPHWHUV UHTXLUHG WR ILW WKH RSWLPDO PRGHO LQFOXGH f WKH UHODWLYH IUHTXHQFLHV RI WKH WZR VWDWHV RI QDWXUH f WKH PHDQ RI HDFK VWDWH f WKH VWDWLVWLFDO YDULDQFH RI HDFK VWDWH DQG f WKH FRVWV DVVRFLDWHG ZLWK HDFK RI WKH WZR SRVVLEOH GHFLVLRQV LQ FRPELQDWLRQ ZLWK HDFK RI WKH WZR SRVVLEOH VWDWHV RI QDWXUH 6LQFH WKH LQGLYLGXDOnV GHFLVLRQV DUH WR EH HYDOXDWHG E\ FRPSDULVRQV ZLWK WKH RXWn SXWV RI WKH RSWLPDO PRGHO LW ZRXOG VHHP UHDVRQDEOH WKDW WKH LQIRUPDWLRQ DYDLODEOH WR WKH RSWLPDO PRGHO EH WKH VDPH DV WKH LQIRUPDWLRQ DYDLODEOH WR WKH LQGLYLGXDO &RQVHTXHQWO\ IRU GHFLVLRQ VLWXDWLRQV LQ ZKLFK VRPH RI WKH SDUDPHWHUV UHTXLUHG E\ WKH RSWLPDO PRGHO DUH QRW FRQWDLQHG LQ WKH DYDLODEOH LQIRUPDWLRQ VHW WKH RSWLPDO PRGHO PXVW PDNH HVWLPDWHV RI WKH PLVVLQJ SDUDPHWHUV ,QGLYLGXDO $ELOLW\ WR ([SDQG WKH ,QIRUPDWLRQ 6HW 7KH LQGLYLGXDOn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f $ QDWXUDO VWDUWLQJ SRLQW RU DQFKRU LV XVHG DV WKH ILUVW DSSUR[LPDWLRQ IRU WKH GHFLVLRQ 7KLV DQFKRU LV WKHQ $ KHXULVWLF ZLWKLQ WKLV FRQWH[W UHIHUV WR D OHDUQHG VHW RI UXOHV RU SULQFLSOHV ZKLFK DUH XWLOL]HG E\ DQ LQGLYLGXDO LQ PDNLQJ WKH SDUWLFXODU GHFLVLRQV UHTXLUHG RI KLP

PAGE 38

DGMXVWHG DV WKH LQGLYLGXDO OHDUQV IURP KLV H[SHULHQFHV ZLWK WKH VWDWHV RI QDWXUH :LWKLQ WKH SUHVHQW VWXG\ WKH LQLWLDO GHFLVLRQ SRLQWV RU DQFKRUV ZHUH H[SHFWHG WR IDOO QHDU WKH JHRPHWULF LQWHUVHFWLRQ HLWKHU DFWXDO RU HVWLPDWHG GHSHQGLQJ XSRQ WKH DYDLODEOH LQIRUPDWLRQf RI WKH GLVWULEXn WLRQDO FXUYHV RI WKH WZR VWDWHV $GMXVWPHQWV IURP WKHVH LQLWLDO GHFLn VLRQ SRLQWV E\ WKH LQGLYLGXDO ZHUH H[SHFWHG WR RFFXU GXULQJ WKH WUDLQn LQJ SKDVH RI WKH H[SHULPHQW DV WKH LQGLYLGXDO JDLQHG H[SHULHQFH ZLWK WKH VWDWHV DQG UHFHLYHG IHHGEDFN DV WR KLV SHUIRUPDQFH UHODWLYH WR DQ RSWLPDO PRGHOf 7KH H[WHQW RI DQ LQGLYLGXDOnV DGMXVWPHQWV KLV OHDUQn LQJ HIILFLHQF\f LV PHDVXUHG XVLQJ VHYHUDO YDULDEOHV (DFK YDULDEOH PHDVXUHV WKH UHODWLYH H[WHQW RI DGMXVWPHQW RU ODFN RI DGMXVWPHQWf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f 9HU\ OLWWOH UHVHDUFK KDV EHHQ FRQFHUQHG ZLWK KRZ WKH GHFLVLRQ PDNHU GRHV DFFRPSOLVK WKHVH SURFHVVHV 7KH PDMRU REMHFWLYH RI WKLV UHVHDUFK LV WR VWXG\ WKH SURFHVVHV XVHG E\ WKH GHFLVLRQ PDNHU LQ UHDFKLQJ YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV ,Q SDUWLFXODU WKLV UHn VHDUFK ZLOO H[DPLQH ZLWKLQ WKH FRQFHSWXDO IUDPHZRUN GLVFXVVHG SUHYLRXVO\

PAGE 39

f WKH HIIHFWV RI WKH GHFLVLRQ VLWXDWLRQ VWUXFWXUH WKH DYDLODEOH LQIRUn PDWLRQ VHW FRQWHQWV WKH LQGLYLGXDOnV LQIRUPDWLRQ SURFHVVLQJ HIILFLHQF\ DQG WKH LQGLYLGXDOnV OHDUQLQJ HIILFLHQF\ RQ WKH LQGLYLGXDOnV ORQJUXQ GHFLVLRQ HIILFLHQF\ DQG f WKH HIIHFWV RI WKH GHFLVLRQ VLWXDWLRQ VWUXFWXUH DQG WKH DYDLODEOH LQIRUPDWLRQ VHW FRQWHQWV RQ WKH LQGLYLGXDOn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n SHQGHQW YDULDEOH LV ODEHOHG WKH LQIRUPDWLRQ YDULDEOH

PAGE 40

7KH VWDWLVWLFDO VWUXFWXUH 7KH HIIHFWV RI WKH VWDWLVWLFDO VWUXFn WXUH RI WKH GHFLVLRQ VLWXDWLRQ DUH VWXGLHG E\ PDQLSXODWLQJ D GLVWULEXn WLRQDO LQIRUPDWLRQ YDULDEOH 7KLV LQYROYHV WKH VSHFLILFDWLRQ RI WZR OHYHOV RI VWDWLVWLFDO UHODWLRQVKLS DPRQJ WKH WZR VWDWHV 6LQFH WKH GLIILFXOW\ RI WKH GLVFULPLQDWLRQ WDVN LQFUHDVHV DV WKH DUHD RI GLVn WULEXWLRQDO RYHUODS EHWZHHQ WKH WZR VWDWHV LQFUHDVHV RQH OHYHO RI WKH GLVWULEXWLRQ YDULDEOH ZLOO KDYH D JUHDWHU DUHD RI RYHUODS WKDQ ZLOO WKH RWKHU OHYHO 7KLV LQGHSHQGHQW YDULDEOH LV ODEHOHG WKH GLVWULEXWLRQ YDULDEOH $OO RWKHU GLVWULEXWLRQDO LQIRUPDWLRQ LWHPV LI SUHVHQWHG ZLOO EH FRQWUROOHG DV FRQVWDQWV 7KH FRVW VWUXFWXUH 7KH HIIHFWV RI WKH FRVW VWUXFWXUH RI WKH SRVVLEOH GHFLVLRQ RXWFRPHV DUH VWXGLHG E\ PDQLSXODWLQJ D VLQJXODU LQn IRUPDWLRQ YDULDEOH 7KH PDQLSXODWLRQ RI WKLV YDULDEOH LQYROYHV WKH VSHFLILFDWLRQ RI WZR OHYHOV RI GHFLVLRQ RXWFRPH UHODWLRQVKLSV LQ WHUPV RI LQFXUUHG FRVWVf 7KLV LQGHSHQGHQW YDULDEOH LV ODEHOHG WKH FRVW YDULDEOH 6LQFH WKH LPSRUWDQFH RI WKH GLVFULPLQDWLRQ WDVN LQFUHDVHV DV WKH FRVWV DVVRFLDWHG ZLWK GHFLVLRQ RXWFRPHV WKDW LQYROYH GHFLVLRQ HUURUV EHFRPH XQHTXDO WKH GLIIHUHQW OHYHOV RI WKH FRVW YDULDEOH ZLOO EH DVVRFLDWHG ZLWK GLIIHUHQW GHFLVLRQ HUURU FRVWV 2QH OHYHO RI WKH FRVW YDULDEOH LV VWUXFWXUHG LQ IDYRU RI PRUH YDULDQFH LQYHVWLJDWLRQV DQG WKH RWKHU OHYHO RI WKH FRVW YDULDEOH LV VWUXFWXUHG LQ IDYRU RI IHZHU YDULn DQFH LQYHVWLJDWLRQV 7KH RQO\ RWKHU VLQJXODU LQIRUPDWLRQ YDULDEOH WKH DFWXDO UHVXOWV RI WKH SK\VLFDO SURFHVV DQG LWV FRVW YDULDQFH LV WKH SULPDU\ H[SHULPHQWDO VWLPXOXV 7KLV LQIRUPDWLRQ LWHP ZLOO EH D UDQGRP YDULDEOH ZKRVH GLVWULEXWLRQV JLYHQ HLWKHU RI WKH WZR VWDWHV RI QDWXUH ZLOO EH QRUPDOO\ VKDSHG ZLWK SDUDPHWHU YDOXHV GHILQHG E\ WKH DSSURSULDWH OHYHO RI WKH GLVWULEXWLRQ YDULDEOH

PAGE 41

,QGHSHQGHQW YDULDEOHV LQWHUDFWLRQ 7KH UHVHDUFK GHVLJQ RI WKLV VWXG\ LV D IDFWRULDO GHVLJQ LQ ZKLFK WKH DERYH WKUHH LQGHSHQGHQW YDULn DEOHV HDFK DW WZR OHYHOV DUH IXOO\ FURVVHG WKXV SURGXFLQJ HLJKW f LQGHSHQGHQW YDULDEOHV FRPELQDWLRQV RU WUHDWPHQWV 7KH IDFWRULDO FRPELQDWLRQ RI WKH GLIIHUHQW OHYHOV RI WKH LQGHSHQGHQW YDULDEOHV DGGV SRZHU WR WKH UHVHDUFK GHVLJQ E\ SHUPLWWLQJ WKH H[DPLQDWLRQ RI WKH HIIHFWV RI LQWHUDFWLRQV XSRQ WKH GHSHQGHQW YDULDEOHV 7KH SURFHVV YDULDEOHV 7KH HIIHFWV RI DQ LQGLYLGXDOnV LQIRUPDWLRQ SURFHVVLQJ DQG OHDUQn LQJ HIILFLHQF\ DUH VWXGLHG XVLQJ FHUWDLQ REVHUYDWLRQ YDULDEOHV PHDVXUHG RQ D FRQWLQXRXV UDWLRf VFDOH &RQWLQXRXV PHDVXUHPHQW IDFLOLWDWHV WKH XVH RI WKHVH YDULDEOHV DV ERWK LQGHSHQGHQW DQG GHSHQGHQW YDULDEOHV :KHQ XVHG DV LQGHSHQGHQW YDULDEOHV WKH PHDVXUHV DUH WUHDWHG DV UDQGRP FRPSRQHQWV UDWKHU WKDQ DV GLVFUHWH FODVVLILFDWLRQ OHYHOV 7KH HIIHFWV DUH VWXGLHG XVLQJ WKUHH PDMRU W\SHV RI YDULDEOHV f LQGLYLGXDO GHn FLVLRQ PRGHO VHQVLWLYLW\ UHODWLYH HIIHFW RI D YDULDEOH UHVSRQVH UDQJHf f LQGLYLGXDO GHFLVLRQ FULWHULD UHODWLYH HIIHFW RI FRQVHUYDWLYH DGn MXVWPHQWf DQG f UHODWLYH LQLWLDO GHFLVLRQ DQFKRU DQG ILQDO GHFLVLRQ DQFKRU UHODWLRQVKLSV 7KH ILUVW WZR W\SHV RI YDULDEOHV DUH GHULYHG IURP PRGHO SDUDPHWHUV ZLWKLQ WKH WKHRU\ RI VLJQDO GHWHFWLRQ $ VXPPDU\ RI WKH JHQHUDO UHVHDUFK GHVLJQ 7KH JHQHUDO UHVHDUFK GHVLJQ LV GHSLFWHG LQ 7DEOH 7KH GHVLJQ LV SUHVHQWHG LQ WHUPV RI WKH GHSHQGHQW DQG LQGHSHQGHQW YDULDEOHV WR EH LQFOXGHG ZLWKLQ WKH UHVHDUFK 0RUH VSHFLILF H[WHQVLRQV DULG

PAGE 42

7$%/( *(1(5$/ 5(6($5&+ '(6,*1 ,1 7(506 2) (;3(5,0(17$/ 9$5,$%/(6 'HSHQGHQW 9DULDEOHV ,QGHSHQGHQW 9DULDEOHV ,QGLYLGXDO /RQJ5XQ 'HFLVLRQ (IILFLHQF\ ,QGLYLGXDO ,QIRUPDWLRQ 3URFHVVLQJ DQG /HDUQLQJ (IILFLHQF\ &RQWHQWV RI WKH DYDLODEOH LQIRUn PDWLRQ VHW ; ; 6WUXFWXUH RI WKH GHFLVLRQ VLWXDWLRQ 6WDWLVWLFDO UHODWLRQVKLSV ; ; 'HFLVLRQ RXWFRPH UHODWLRQVKLSV ; ; ,QGLYLGXDO LQIRUPDWLRQ SURFHVVLQJ DQG OHDUQLQJ HIILFLHQF\ ; 1RWH $Q ; ZLWKLQ D FHOO LQGLFDWHV WKDW WKH UHODWLRQVKLS RI WKH YDULDEOHV FRQFHUQHG DUH LQFOXGHG ZLWKLQ WKH H[SHULPHQWDO GHVLJQ

PAGE 43

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n HUDO WDVN HQYLURQPHQW RI WKH ODERUDWRU\ H[SHULPHQWDWLRQ 7KHVH DVVXPSWLRQV DQG WKH WDVN HQYLURQPHQW DUH GLVFXVVHG LQ JUHDWHU GHWDLO LQ D ODWHU VHFWLRQ RI WKLV FKDSWHU 7KH PDMRU REMHFWLYH RI WKH ODERUDWRU\ H[SHULPHQW LV WR WHVW WKH FRQFHSWXDO GHYHORSPHQW WKURXJK WKH K\SRWKHVHV GHULYHG E\ WKH VLPXODWLRQf RI WKH HIIHFWV RI WKH YDULn RXV LQGHSHQGHQW YDULDEOHV RQ WKH YDULRXV REVHUYDWLRQ YDULDEOHV 7KH JHQHUDO DQG VSHFLILF UHVHDUFK GHVLJQ HPSOR\HG LQ WKH ODERUDn WRU\ H[SHULPHQW LV WKH VDPH DV WKDW HPSOR\HG ZLWKLQ WKH VLPXODWLRQ 7KH PDMRU GLIIHUHQFH EHWZHHQ WKH VLPXODWLRQ DQG WKH ODERUDWRU\ H[SHULn PHQW LV WKH XVH RI KXPDQ VXEMHFWV 7KLV GLIIHUHQFH FUHDWHV WZR PDMRU VRXUFHV RI LQFRPSDWDELOLW\ EHWYHHQ WKH VLPXODWLRQ DQG WKH ODERUDWRU\ H[SHULPHQW )LUVW WKH VLPXODWLRQ PDNHV FHUWDLQ DVVXPSWLRQV FRQFHUQLQJ KXPDQ EHKDYLRU DQG DSSOLHV WKHVH DVVXPSWLRQV FRQVLVWHQWO\ :LWKLQ

PAGE 44

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f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n DQFH UHSRUWV DQG LV DVNHG IRU HDFK UHSRUW WR GHFLGH ZKHWKHU WR LQYHVWLn JDWH RU QRW WR LQYHVWLJDWH WKH UHSRUWHG ODERU HIILFLHQF\ YDULDQFH 7KH HIILFLHQF\ RI D VXEMHFWnV GHFLVLRQ SHUIRUPDQFH DQG WKH DPRXQW RI SD\PHQW KH ZLOO UHFHLYH IRU SDUWLFLSDWLQJ LQ WKH H[SHULPHQW LV EDVHG XSRQ WKH WRWDO LQYHVWLJDWLRQ GHFLVLRQV FRVW ZKLFK KH LQFXUV RYHU WKH VHULHV RI YDULDQFH UHSRUWV

PAGE 45

(DFK VWDQGDUG YDULDQFH UHSRUW LV FRQFHUQHG ZLWK WKH UHVXOWV RI D VLQJOH MRERUGHU WR SURGXFH D FRQVWDQW QXPEHU RI FKDLUV DQG UHSRUWV RQO\ DJJUHJDWH RYHUDOO DVVHPEO\ GHSDUWPHQWf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

PAGE 46

f WKH LQIRUPDWLRQ YDULDEOH f WKH GLVWULEXWLRQ YDULDEOH DQG f WKH FRVW YDULDEOH $V LQGLFDWHG SUHYLRXVO\ HDFK LV YDULHG DFURVV WZR OHYHOV 7KH LQGLYLGXDO SURFHVV YDULDEOH W\SHV PHDVXUHG RQ D FRQWLQXRXV VFDOH LQFOXGH WKH IROORZLQJ f LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ f LQGLYLGXDO GHFLVLRQ FULWHULD DQG f UHODWLYH LQLWLDO DQG ILQDO GHFLVLRQ DQFKRU UHODWLRQVKLSV 7KH PDMRU GHSHQGHQW YDULDEOH LV WKH LQGLYLGXDO ORQJUXQ GHFLVLRQ HIILFLHQF\ LQ WHUPV RI LQFXUUHG FRVWVf 7KH IROORZLQJ GLVFXVVLRQ GHVFULEHV HDFK RI WKH YDULDEOHV RU YDULDEOH W\SHV DV WKH\ DUH HPSOR\HG LQ WKLV VWXG\ 7KH LQIRUPDWLRQ YDULDEOH 7KH ILUVW OHYHO ODEHOHG -OB LV GHULYHG IURP WKH VHW RI LQIRUn PDWLRQ DVVXPHG WR FRPH IURP LQGLYLGXDO H[SHULHQFH ZLWK WKH SK\VLFDO V\VWHP ,W LQFOXGHV WKH IROORZLQJ LWHPV f WKH KLVWRULFDOO\ GHULYHG SRUWLRQ RI WLPH LQ ZKLFK WKH SURFHVV KDV EHHQ IRXQG WR IDOO LQ HDFK RI WKH WZR VWDWHV f WKH DVVXPSWLRQ WKDW WKH UDQGRP YDULDEOH RI LQWHUHVW DFWXDO PLQXWHV LQFXUUHG SHU FKDLU RU LWV DVVRFLDWHG VWDQGDUG YDULn DQFHf LV QRUPDOO\ GLVWULEXWHG IRU ERWK VWDWHV f WKH ORZHVW REVHUYHG YDOXH RI WKH UDQGRP YDULDEOH f WKH KLJKHVW REVHUYHG YDOXH RI WKH UDQGRP YDULDEOH f WKH PD[LPXP FRVWV DVVRFLDWHG ZLWK HDFK VWDWH DQG f WKH PLQLPXP FRVWV DVVRFLDWHG ZLWK HDFK VWDWH 7KH VHFRQG OHYHO RI WKH LQIRUPDWLRQ YDULDEOH ODEHOHG A LQFOXGHV DGGLWLRQDO GLVWULEXWLRQDO LQIRUPDWLRQ ,Q DGGLWLRQ WR WKH VL[ LWHPV FRQWDLQHG LQ WKH ,, LQIRUPDWLRQ VHW WKH IROORZLQJ WZR LWHPV DUH LQFOXGHG f WKH PHDQ RI WKH UDQGRP YDULDEOH ZLWKLQ HDFK VWDWH DQG f WKH VWDQGDUG GHYLDWLRQ RI WKH UDQGRP YDULDEOH ZLWKLQ HDFK VWDWH

PAGE 47

7KH GLVWULEXWLRQ YDULDEOH 0DQLSXODWLRQ RI WKH GLVWULEXWLRQ YDULDEOH LQYROYHV WZR IDFWRUV fWKH GLVWULEXWLRQDO SDUDPHWHUV RI HDFK VWDWH DQG f WKH VWDWLVWLFDO UHODWLRQVKLS EHWZHHQ WKH VWDWHV 7KH WZR OHYHOV RI WKLV YDULDEOH DUH JHQHUDWHG WKURXJK D FKDQJH LQ WKH YDULDQFH DQG D FKDQJH LQ WKH VWDQn GDUGL]HG GLVWDQFH EHWZHHQ WKH PHDQV RI WKH WZR VWDWHV 7KH ILUVW OHYHO RI WKH GLVWULEXWLRQ YDULDEOH ODEHOHG 6OB KDV WKH IROORZLQJ SDUDPHWHUV DQG UHODWLRQVKLSV f \ DFWXDO PLQXWHV LQFXUUHG SHU FKDLU f FUQ D "L DFWXDO PLQXWHV LQFXUUHG SHU FKDLU DQG f X \ D DFWXDO PLQXWHV LQFXUUHG SHU FKDLU ZKHUH \ f f WKH PHDQ RI WKH MWK VWDWH RI QDWXUH LQFRQWURO DQG r RXWRIFRQWURO f JLYHQ WKH LWK GLVWULEXWLRQ OHYHO 6, DQG 6 f D WKH VWDQGDUG GHYLDWLRQ RI WKH MWK VWDWH RI QDWXUH JLYHQ n WKH LWK GLVWULEXWLRQ OHYHO DQG WKH VWDQGDUG GHYLDWLRQ FRPPRQ WR ERWK VWDWHV RI QDWXUH JLYHQ WKH LWK GLVWULEXWLRQ OHYHO 7KH VHFRQG OHYHO RI WKH GLVWULEXWLRQ YDULDEOH ODEHOHG 6A KDV WKH IROORZLQJ SDUDPHWHUV DQG UHODWLRQVKLSV f \ DFWXDO PLQXWHV LQFXUUHG SHU FKDLU f DFWXDO PLQXWHV LQFXUUHG SHU FKDLU DQG f \f \ f DFWXDO PLQXWHV LQFXUUHG SHU FKDLU ZKHUH \A A DQG A DUH GHILQHG WKH VDPH DV LQ WKH 6, OHYHO 7KH WKLUG DQG IRXUWK VWDWLVWLFDO PRPHQWV RI HDFK VWDWH ZLWKLQ HDFK GLVn WULEXWLRQ OHYHO DUH XQFRQWUROOHG H[FHSW WKDW WKHLU GHYLDWLRQV IURP D QRUPDO GLVWULEXWLRQ DUH PLQLPL]HG

PAGE 48

7KH FRVW YDULDEOH *LYHQ WKH WZR VWDWHV RI QDWXUH WKH UHDOL]DWLRQ RI ZKLFK LV XQn NQRZQ WR WKH LQGLYLGXDOf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f &GHFLVLRQ LQYHVWLJDWH VWDWH LQFRQWUROf DQG 0&J &GHFLVLRQ LQYHVWLJDWH VWDWH RXWRIFRQWUROf &GHFLVLRQ QRW LQYHVWLJDWH VWDWH RXWRIFRQWUROf ZKHUH & f UHSUHVHQWV WKH GHFLVLRQ FRVW IRU WKH VLWXDWLRQ ZLWKLQ WKH SDUHQWKHVHV :KHQ WKH PDUJLQDO FRVW RI D W\SH $ HUURU HTXDOV WKH PDUJLQDO FRVW RI D W\SH % HUURU WKH FRVW VWUXFWXUH VKRXOG EH LJQRUHG ZKHQ PDNLQJ DQ LQYHVWLJDWLRQ GHFLVLRQ 2I JUHDWHU LQWHUHVW DUH VLWXDWLRQV LQ ZKLFK WKH PDUJLQDO HUURU FRVWV QRW HTXDO ,Q WKH SUHVHQW VWXG\ HDFK OHYHO RI WKH FRVW YDULDEOH WDNHV RQH RI WKH IROORZLQJ IRUPV f WKH PDUJLQDO FRVW RI D W\SH % GHFLVLRQ HUURU HTXDOV WKUHH WLPHV WKH PDUJLQDO FRVW RI D W\SH $ HUURU ODEHOHG OHYHO &-Bf DQG f WKH PDUJLQDO FRVW RI D W\SH $ HUURU HTXDOV WKUHH WLPHV WKH PDUJLQDO FRVW RI D W\SH % HUURU ODEHOHG OHYHO &f

PAGE 49

,QGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ YDULDEOHV 7KH VHQVLWLYLW\ RI WKH LQGLYLGXDOnV GHFLVLRQ PRGHO UHODWLYH WR WKH GHFLVLRQ VLWXDWLRQ LV PHDVXUHG XVLQJ WKH 76' SDUDPHWHU Gf 7KH WKHRn UHWLFDO GHILQLWLRQ RI Gf LV Gf S 3Mfp ]L ] ZKHUH \ WKH PHDQ RI WKH MWK VWDWH RI QDWXUH LQFRQWURO DQG RXWRIFRQWURO f D WKH FRPPRQ VWDQGDUG GHYLDWLRQ RI WKH VWDWHV RI QDWXUH DQG ] WKH YDOXH RI WKH QRUPDO GLVWULEXWLRQ IXQFWLRQ DVVRFLDWHG ZLWK WKH MWK VWDWH RI QDWXUH DQG DQ\ GHFLVLRQ FXWRII YDOXH FRPPRQ WR ERWK VWDWHV $Q HPSLULFDO HVWLPDWH RI WKH G IRU DQ LQGLYLGXDO ODEHOHG G LV REWDLQHG XVLQJ WKH LQGLYLGXDOnV FRQGLWLRQDO SUREDELOLWLHV 3GHFLVLRQ LQYHVWLJDWH VWDWH RXWRIFRQWUROf DQG 3GHFLVLRQ LQYHVWLJDWH VWDWH LQFRQWUROf WR FDOFXODWH D VXEMHFWLYH ] DQG ]! $V SUHYLRXVO\ SRLQWHG RXW Gn LV UHODWLYH WR WKH GHFLVLRQ VLWXDWLRQ ,Q SDUWLFXODU Gf LV UHODWLYH WR WKH GLVWULEXWLRQ YDULDEOH 7R JDLQ FRPSDUDELOLW\ DFURVV VLWXDWLRQV WKH IROORZLQJ PHDVXUH LV XVHG '1 GL G ZKHUH GA LV JHQHUDWHG XVLQJ RSWLPDO PRGHO N 7KH PHDVXUH GHFUHDVHV IURP D YDOXH RI RQH DV WKH UHVXOW RI VHYHUDO IDFWRUV f WKH LQGLYLGXDO LV LQFRQVLVWHQW LQ KLV XVH RI KLV FXWRII YDOXH LH WKHUH H[LVWV D UDQJH DURXQG KLV FXWRII YDOXH ZLWKLQ ZKLFK GHFLVLRQV DUH QRW PDGH XVLQJ D VWULFW UHODWLRQ WR WKLV FXWRII YDOXHf RU WKH LQGLYLGXDO PDNHV RQH RU PRUH WHPSRUDU\ SURFHVVLQJ HUURUV DQG f WKH LQGLYLGXDO XWLOL]HV PRUH WKDQ RQH FXWRII YDOXH

PAGE 50

%DVHG XSRQ D VXEMHFWLYH DQDO\VLV RI HDFK LQGLYLGXDOnV GHFLVLRQV WKH PHDVXUH '1A FDQ EH DGMXVWHG IRU WKH HIIHFWV RI XVLQJ PXOWLSOH FXWn RII YDOXHV 'HILQLQJ GAD WR EH WKH LQGLYLGXDOnV GHFLVLRQ PRGHO VHQVLn WLYLW\ ZLWK WKH HIIHFWV RI PXOWLSOH FXWRII YDOXHV HOLPLQDWHG '1$L G?D GnN 7KH GLIIHUHQFH '1$A '1A DSSURDFKHV ]HUR DV WKH HIIHFWV RI PXOWLSOH FXWRII YDOXHV GHFUHDVHV DQG EHFRPHV ]HUR ZKHQ WKH LQGLYLGXDO XVHV D VLQJOH FXWRII YDOXH ,QGLYLGXDO GHFLVLRQ FULWHULD YDULDEOHV 7KH FULWHULD WKH LQGLYLGXDO DGRSWV LQ PDNLQJ KLV GHFLVLRQV DUH PHDVXUHG XVLQJ WKH 76' SDUDPHWHU 7KH WKHRUHWLFDO GHILQLWLRQ RI LV G!]f M!]f ZKHUH M! f GHQRWHV WKH QRUPDO GHQVLW\ IXQFWLRQ IRU WKH SRLQW LQ WKH SDUHQWKHVHV DQG LV GHILQHG WKH VDPH DV IRU WKH Gn YDULDEOH $Q HPSLULFDO HVWLPDWH IRU WKH RI DQ LQGLYLGXDO ODEHOHG LV REWDLQHG XVLQJ WKH ] DQG ] YDOXHV DVVRFLDWHG ZLWK WKH LQGLYLGXDOnV FRQGLWLRQDO SUREDELOLWLHV HPSOR\HG LQ HVWLPDWLQJ GO 7KH PHDVXUH LV UHODWLYH WR WKH GHFLVLRQ VLWXDWLRQ ,Q SDUWLFXODU LW LV UHODWLYH WR WKRVH YDULDEOHV ZKLFK DIIHFW WKH SRLQW RQ WKH GHFLVLRQ D[LV ZKLFK WKH LQGLYLGXDO VHOHFWV DV KLV LQYHVWLJDWLRQ GHFLVLRQ FXWRII YDOXH $ PHDVXUH RI LQGLYLGXDO FULWHULD FRPSDUDEOH DFURVV VLWXDWLRQV LV GHILQHG DV %1 L L N ZKHUH A LV JHQHUDWHG XVLQJ RSWLPDO PRGHO N 7KH PHDVXUH DSSURDFKHV D YDOXH RI RQH DV DSSURDFKHV A LH DV WKH LQGLYLGXDOnV FXWRII GHFLVLRQ YDOXH DSSURDFKHV WKH RSWLPDO PRGHOnV FXWRII GHFLVLRQ YDOXHf

PAGE 51

7KH PHDVXUH DSSURDFKHV HLWKHU D YDOXH RI ]HUR RU SRVLWLYH LQILQLW\ DV WKH UHVXOW RI VHYHUDO IDFWRUV f WKH LQGLYLGXDO GRHV QRW SURFHVV SURSHUO\ WKH HIIHFWV RI WKH UHODWLYH FRVWV RI WKH WZR W\SHV RI GHFLVLRQ HUURUV f WKH LQGLYLGXDO GRHV QRW SURFHVV SURSHUO\ WKH HIIHFWV RI WKH UHODWLYH IUHTXHQFLHV RI WKH WZR VWDWHV DQG f WKH LQGLYLGXDO XVHV PRUH WKDQ RQH FXWRII YDOXH 8VLQJ WKH VDPH VXEMHFWLYH DQDO\VLV DV WKDW XVHG LQ DGMXVWLQJ WKH '1 f PHDVXUH WKH %1A PHDVXUH FDQ EH DGMXVWHG IRU WKH HIIHFWV RI PXOWLSO FXWRII YDOXHV 'HILQLQJ %A WR EH WKH LQGLYLGXDOnV FULWHULD PHDVXUH ZLWK WKH HIIHFWV RI PXOWLSOH FXWRII YDOXHV HOLPLQDWHG %1$ %r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nV FXWRII YDOXH DQG WKH RSWLPDO PRGHOnV FXWRII YDOXH 6LQFH WKH PHDVXUH LV GHSHQGHQW RQ WKH GLUHFWLRQ RI DGn MXVWPHQW LW LV FRQGLWLRQDO XSRQ WKH OHYHO RI WKH FRVW YDULDEOH 8VLQJ %1&c WR GHQRWH WQH H[WHQW RI DQ LQGLYLGXDOnV FRQVHUYDWLVP %1&c %_f %A JLYHQ WKH &O FRVW OHYHO DQG %1&c EA %Mf %A JLYHQ WKH & FRVW OHYHO

PAGE 52

%1&DSSURDFKHV HLWKHU SRVLWLYH LQILQLW\ RU D YDOXH RI RQH DV WKH H[WHQW RI FRQVHUYDWLVP LQFUHDVHV DSSURDFKHV D YDOXH RI ]HUR DV WKH H[WHQW RI FRQVHUYDWLVP GHFUHDVHV DQG DSSURDFKHV HLWKHU D YDOXH RI QHJDWLYH RQH RU QHJDWLYH LQILQLW\ DV WKH H[WHQW RI QRQFRQVHUYDWLVP LQFUHDVHV ,QLWLDO DQG ILQDO GHFLVLRQ DQFKRU YDULDEOHV 7KH UHODWLYH UHODWLRQVKLSV EHWZHHQ DQ LQGLYLGXDOnV LQLWLDO DQG ILQDO GHFLVLRQ DQFKRUV FDQ EH PHDVXUHG LQ WHUPV RI WKH UHODWLYH OLQHDU GLVWDQFH EHWZHHQ WKHVH WZR GHFLVLRQ DQFKRUV $ PHDVXUH RI WKH UHODWLYH DGMXVWPHQW IRU LQGLYLGXDO L ODEHOHG 5$A LV FRQGLWLRQDO RQ WKH GLUHFn WLRQ RI DGMXVWPHQW DORQJ WKH UHOHYDQW GHFLVLRQ D[LV 7KLV GLUHFWLRQ RI DGMXVWPHQW LV LQ WXUQ FRQGLWLRQDO RQ WKH OHYHO RI WKH FRVW YDULDEOH *LYHQ WKH &O FRVW OHYHO WKH LQLWLDO GHFLVLRQ DQFKRU LV JUHDWHU WKDQ WKH RSWLPDO PRGHO FXWRII YDOXH DQG JLYHQ WKH & FRVW OHYHO WKH LQLWLDO GHFLVLRQ DQFKRU LV OHVV WKDQ WKH RSWLPDO PRGHO FXWRII YDOXH 7KH 5$A PHDVXUH LV GHILQHG DV 5$L ('9L '9Nf 7'9L '9Nf JLYHQ WKH &O FRVW OHYHO 5$M '9N ('9Mf '9N 7'9cf JLYHQ WKH & FRVW OHYHO ZKHUH ('9A LQGLYLGXDO LnV ILQDO H[SHULPHQWf GHFLVLRQ DQFKRU 7'9 LQGLYLGXDO LnV LQLWLDO WUDLQLQJf GHFLVLRQ DQFKRU DQG '9N RSWLPDO PRGHO NnV ILQDO H[SHULPHQWf FXWRII YDOXH 7KH 5$M PHDVXUH KDV WKH IROORZLQJ UHODWLRQVKLSV f LI QR DGMXVWPHQW RFFXUV EHWZHHQ WKH WUDLQLQJ DQG WKH H[SHULPHQW SKDVHV WKH PHDVXUH ZLOO HTXDO D YDOXH RI RQH f LI WKH DGMXVWPHQW LV LQ WKH ZURQJ GLUHFWLRQ DZD\ IURP WKH RSWLPDO PRGHOnV H[SHULPHQW FXWRII YDOXHf WKH PHDVXUH ZLOO EH JUHDWHU WKDQ D YDOXH RI RQH f LI WKH DGMXVWPHQW LV LQ

PAGE 53

WKH SURSHU GLUHFWLRQ EXW LQFRPSOHWH WKH PHDVXUH ZLOO EH OHVV WKDQ D YDOXH RI RQH EXW JUHDWHU WKDQ D YDOXH RI ]HUR f LI WKH DGMXVWPHQW LV LQ WKH SURSHU GLUHFWLRQ DQG LV FRPSOHWH WKH PHDVXUH ZLOO HTXDO D YDOXH RI ]HUR DQG f LI WKH DGMXVWPHQW LV LQ WKH SURSHU GLUHFWLRQ EXW PRUH WKDQ FRPSOHWH WKH PHDVXUH ZLOO EH OHVV WKDQ D YDOXH RI ]HUR ,QGLYLGXDO ORQJUXQ GHFLVLRQ HIILFLHQF\ $ PDMRU GHSHQGHQW YDULDEOH RI LQWHUHVW LQ WKLV UHVHDUFK LV WKH FRVW LQFXUUHG DV D UHVXOW RI WKH LQGLYLGXDOnV YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV 7KH H[SHULPHQWDO REMHFWLYH IXQFWLRQ IRU DOO GHFLVLRQ VLWXDWLRQV LV WR PLQLPL]H WKHVH FRVWV ODEHOHG LQYHVWLJDWLRQ GHFLVLRQV FRVWVf 7KH LQGLYLGXDOn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nV LQYHVWLJDWLRQ GHFLVLRQ FRVWV ZLWK WKDW RI DQ RSWLPDO PRGHO ZKLFK XVHV WKH VDPH LQIRUPDWLRQ DV WKDW DYDLODEOH WR WKH LQGLn YLGXDO 'HQRWLQJ VXFK D PHDVXUH *L LFLM 0&NM! 60&N ZKHUH ,& LQGLYLGXDO LnV LQYHVWLJDWLRQ GHFLVLRQ FRVW IRU GHFLVLRQ M 0&AM RSWLPDO PRGHO NnV LQYHVWLJDWLRQ GHFLVLRQ FRVW IRU GHFLVLRQ M DQG

PAGE 54

60&A WKH VXP RI RSWLPDO PRGHO NnV LQYHVWLJDWLRQ GHFLVLRQV FRVWV RYHU DOO GHFLVLRQV P LQ QXPEHUf 7KH *f f PHDVXUH FDQ EH FODVVLILHG LQWR RQH RI WKUHH VXEPHDVXUHV n ZKHUH WKH FODVVLILFDWLRQ LV GHSHQGHQW XSRQ WKH DOJHEUDLF VLJQ RI *AM 7KH FODVVLILFDWLRQV DUH f *c f WKH LQGLYLGXDOnV LQYHVWLJDWLRQ GHFLVLRQ FRVW IRU GHFLVLRQ M LV JUHDWHU WKDQ WKDW RI WKH RSWLPDO PRGHO f *AM WKH LQGLYLGXDOnV LQYHVWLJDWLRQ GHFLVLRQ FRVW IRU GHFLVLRQ M LV HTXDO WR WKDW RI WKH RSWLPDO PRGHO DQG f *f f WKH LQGL r ?YLGXDOnV LQYHVWLJDWLRQ GHFLVLRQ FRVW IRU GHFLVLRQ M LV OHVV WKDQ WKDW RI WKH RSWLPDO PRGHO 6XPPLQJ WKRVH *V ZLWK LGHQWLFDO DOJHEUDLF VLJQV WKH VXPPDWLRQV r DUH GHILQHG DV IROORZV *3M WKH VXP RI WKRVH *f§V ZLWK D SRVLWLYH DOJHEUDLF VLJQ *1A WKH VXP RI WKRVH *fV ZLWK D QHJDWLYH DOJHEUDLF VLJQ DQG *=M WKH VXP RI WKRVH *f§V ZLWK D YDOXH RI ]HUR ,W LV HDVLO\ VKRZQ WKDW WKH IROORZLQJ UHODWLRQ KROGV P ( *3 *1 *= M O ZKHUH P HTXDOV WKH WRWDO QXPEHU RI GHFLVLRQV PDGH E\ LQGLYLGXDO L 6LQFH WKH *=M VXPPDWLRQ HTXDOV ]HUR WKH VXP RI WKH *f§ RYHU DOO M OP GHFLVLRQV GHQRWHG LV *3 *1 L L L &RQFHSWXDOO\ *3A LV WKH UHODWLYH DGGLWLRQDO FRVW RI WKRVH GHn FLVLRQV PDGH E\ LQGLYLGXDO L ZKLFK DUH JUHDWHU LQ YDOXH WKDQ WKDW RI WKH RSWLPDO PRGHO 7KH *1A LV WKH UHODWLYH VDYLQJV RI WKRVH GHFLVLRQV PDGH E\ LQGLYLGXDO L ZKLFK DUH OHVV LQ YDOXH WKDQ WKDW RI WKH RSWLPDO PRGHO

PAGE 55

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nV H[SHULPHQW *M PHDVXUH 7KH DGMXVWPHQW SURFHVV FDQ EH PHDVXUHG LQ WKH WUDLQLQJ VHVVLRQV E\ FRPSXWLQJ WKH LQGLYLGXDOnV *M PHDVXUH DW WKH FRPSOHWLRQ RI HDFK VHVVLRQ 7KH PHDVXUHV ODEHOHG 7*A 7*c DQG 7*c DUH GHILQHG LQ WKH VDPH PDQQHU DV WKH PHDVXUH GLVFXVVHG DERYH ZLWK WKH H[FHSWLRQ WKDW WKH\ DUH FRPSXWHG IURP WKH DSSURSULDWH WUDLQLQJ VHVVLRQ GDWD LQVWHDG RI WKH H[SHULPHQW VHVVLRQ GDWDf 7KH GLIIHUHQFHV EHWZHHQ WKHVH WUDLQLQJ PHDVXUHV UHIOHFW WKH HIIHFWV RI WKH LQGLYLGXDOnV DGMXVWPHQWV LQ WKH WUDLQLQJ VHVVLRQV 7KHVH GLIIHUHQFHV DUH GHILQHG DV IROORZV '7*LL 7*Q 7*L '7*L 7*L 7*L ZKHUH '7*cM WKH GLIIHUHQFH EHWZHHQ WKH *c PHDVXUH IRU WKH MWK DQG WKH MWK WUDLQLQJ VHVVLRQV RI LQGLYLGXDO L DQG 7*AM WKH *M PHDVXUH IRU WKH MWK WUDLQLQJ VHVVLRQ RI LQGLn YLGXDO L 1HJDWLYH '7*cMYDOXHV LQGLFDWH DQ LQFUHDVH LQ WKH 7*MM PHDVXUH EHWZHHQ WUDLQLQJ VHVVLRQV DQG SRVLWLYH '7*MM YDOXHV LQGLFDWH WKH UHYHUVH 7KH HIIHFWV RI WKH GLUHFWLRQV RI WKHVH DGMXVWPHQWV VKRXOG EH UHODWHG WR WKH HIIHFWV RI WKH ILQDO DQFKRU RQ WKH LQGLYLGXDOnV *c PHDVXUH LQ WKH H[SHULPHQWf 7KH HIIHFWV RI WKH DGMXVWPHQW PHDVXUHV

PAGE 56

RQ WKD H[SHULPHQW PHDVXUH FDQ EH DQDO\]HG XVLQJ D GLVFUHWH FODVVLILFDn WLRQ RI WKH GLUHFWLRQ RI WKH DGMXVWPHQWV 7KLV FODVVLILFDWLRQ RI WKH PHDVXUH LV DFFRPSOLVKHG XVLQJ LWV DOJHEUDLF VLJQ 7KH VHFRQG DQG WKLUG WUDLQLQJ VHVVLRQ UHVXOWV FDQ EH JURXSHG LQWR RQH RI IRXU FODVVHV 33f 30f 03f DQG 00f 6XEMHFWV ZLWKLQ WKH 00 FODVVLILFDWLRQ KDYH FRQVWDQWO\ LQFUHDVLQJ 7*MM PHDVXUHV DQG VXEMHFWV ZLWKLQ WKH 33 FODVVLILFDWLRQ KDYH FRQVWDQWO\ GHFUHDVLQJ 7*AM PHDVXUHV +\SRWKHVHV )RUPDWLRQ $V SUHYLRXVO\ GLVFXVVHG WKH IRUPDWLRQ RI WKH H[SHULPHQWDO K\SRWKHVHV LV DFFRPSOLVKHG SULPDULO\ XVLQJ WKH PHWKRG RI VLPXODWLRQ 7ZR JHQHUDO W\SHV RI VLPXODWLRQV DUH SHUIRUPHGf§ VLPXODWLRQ RI RSWLPDO PRGHO SHUIRUPDQFHV DQG VLPXODWLRQ RI VXEMHFWLYH LQYHVWLJDWLRQ GHFLVLRQ SHUIRUPDQFHV 7KH ILUVW SDUW RI WKLV VHFWLRQ GLVFXVVHV WKH DVVXPSWLRQV DQG SUHVHQWV WKH UHVXOWV RI WKH VLPXODWLRQ RI RSWLPDO PRGHO SHUIRUPn DQFHV 7KH VHFRQG SDUW RI WKLV VHFWLRQ GLVFXVVHV FHUWDLQ DVVXPSWLRQV GHULYHG IURP WKH FRQFHSWXDO GHYHORSPHQW WKDW IRUP WKH EDVLV IRU WKH VLPXODWLRQ RI VXEMHFWLYH LQYHVWLJDWLRQ GHFLVLRQ SHUIRUPDQFHV ,W DOVR SUHVHQWV WKH UHVXOWV RI WKLV VLPXODWLRQ DQG WKH GHULYDWLRQ RI WKH H[n SHULPHQWDO K\SRWKHVHV 6LPXODWLRQ RI 2SWLPDO 0RGHO 3HUIRUPDQFHV 7KH GHFLVLRQ UXOH RI WKH RSWLPDO PRGHO JLYHQ WKH REMHFWLYH IXQFn WLRQ RI PLQLPL]LQJ LQFXUUHG FRVWVf HPSOR\HG LQ WKLV UHVHDUFK LV WR LQYHVWLJDWH WKH UHSRUWHG ODERU HIILFLHQF\ YDULDQFH ZKHQ

PAGE 57

(TXDWLRQ ZKHUH [ WKH DFWXDO PLQXWHV LQFXUUHG SHU FKDLU RXW WKH VWDWH RI RXWRIFRQWURO LQ WKH VWDWH RI LQFRQWURO 9LQ1 WKH FRVW DVVRFLDWHG ZLWK QRW LQYHVWLJDWLQJ DQ LQFRQWURO YDULDQFH 9c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f &5 (TXDWLRQ ZKHUH /54 WKH OLNHOLKRRG RGGV RI WKH UHSRUWHG ODERU HIILFLHQF\ YDULDQFH EHLQJ RXWRIFRQWURO 3[_RXWf3[_LQff 35M WKH SULRU RGGV RI WKH LQFRQWURO VWDWH 3LQf3RXWff DQG &5 WKH FRQVWDQW FRVW UDWLR DVVRFLDWHG ZLWK WKH DSSURSULDWH OHYHO RI WKH FRVW YDULDEOH $VVXPLQJ WKDW ERWK VWDWHV RI QDWXUH DUH QRUPDOO\ GLVWULEXWHG ZLWK HTXDO YDULDQFH WKH RSWLPDO PRGHO LQYHVWLJDWLRQ GHFLVLRQ UXOH H[SUHVVHG LQ

PAGE 58

HTXDWLRQ FDQ EH UHVWDWHG LQ WKH IROORZLQJ PDWKHPDWLFDO WHUPV H[S [ \fFU f§ f§ 35L f &5 H[S [ ?LOfD ZKHUH [ DFWXDO PLQXWHV LQFXUUHG SHU FKDLU 9M WKH PHDQ RI WKH LQFRQWURO VWDWH \ WKH PHDQ RI WKH RXWRIFRQWUFO VWDWH D WKH FRPPRQ VWDQGDUG GHYLDWLRQ RI ERWK VWDWHV DQG H[S WKH QRWDWLRQ IRU WKH H[SRQHQWLDO IXQFLWLRQ Hf ,I WKH DERYH HTXDWLRQ ZDV FKDQJHG WR DQ HTXDOLW\ DQG VROYHG IRU [ WKH UHVXOW ZRXOG EH WKH RSWLPDO FXWRII YDOXH LQ WHUPV RI DFWXDO PLQXWHV LQFXUUHG 7KLV VROXWLRQ WDNHV WKH JHQHUDO IRUP *OQ35cf OQ&5ff [ \ \f (TXDWLRQ \ \L ZKHUH [ WKH RSWLPDO FXWRII YDOXH LQ WHUPV RI DFWXDO PLQXWHV LQFXUUHG DQG ,Q WKH QRWDWLRQ IRU WKH QDWXUDO ORJDULWKP IXQFWLRQ ORJHf :LWK WKH GHFLVLRQ UXOH LQ WKH IRUP RI HTXDWLRQ WKH SDUDPHWHUV UHTXLUHG WR ILW WKH RSWLPDO PRGHO EHFRPH LGHQWLILDEOH 7KHVH SDUDPHWHUV LQFOXGH WKH PHDQ RI ERWK VWDWHV RI QDWXUH WKH FRPPRQ YDULDQFH RI WKH WZR VWDWHV WKH SULRU RGGV RI WKH LQFRQWURO VWDWH DQG WKH DSSURSULDWH YDOXH RI WKH FRVW UDWLR :LWK NQRZOHGJH RI WKH SDUDPHWHUV UHTXLUHG WR ILW WKH RSWLPDO PRGHO WKH FULWHULD HPSOR\HG IRU WKH VLPXODWLRQ RI RSWLPDO PRGHO SHUn IRUPDQFHV FDQ EH GLVFXVVHG 7KH ILUVW FULWHULRQ LV WKDW LQIRUPDWLRQ DYDLODEOH WR WKH RSWLPDO PRGHO EH WKH VDPH DV WKH LQIRUPDWLRQ DYDLOn DEOH WR WKH LQGLYLGXDO 7KH LQIRUPDWLRQ DYDLODEOH WR WKH LQGLYLGXDO LV PDQLSXODWHG E\ WKH OHYHOV RI WKH LQIRUPDWLRQ YDULDEOH :LWKLQ WKH

PAGE 59

LQIRUPDWLRQ OHYHO WKH DYDLODEOH LQIRUPDWLRQ VHW FRQWDLQV VWDWLVWLFDO HVWLPDWHV RI DOO WKH SDUDPHWHUV UHTXLUHG WR ILW WKH RSWLPDO PRGHO +RZHYHU ZLWKLQ WKH ,, LQIRUPDWLRQ OHYHO WKH DYDLODEOH LQIRUPDWLRQ VHW GRHV QRW FRQWDLQ DOO WKH UHTXLUHG SDUDPHWHU HVWLPDWHV 7KHUHIRUH ZLWKLQ WKH ,, LQIRUPDWLRQ OHYHO WKH RSWLPDO PRGHO PXVW XVH WKH WUDLQLQJ VHVVLRQV WR HVWLPDWH WKH PLVVLQJ SDUDPHWHUV 7KH SDUDPHWHU HVWLPDWHV ERWK JLYHQ DQG HVWLPDWHG DUH SUHVHQWHG LQ 7DEOH 7KH VHFRQG FULWHULRQ LV WKDW WKH ODERU HIILFLHQF\ YDULDQFH UHn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n JDWLRQ GHFLVLRQ SHUIRUPDQFHV ,W DOVR SUHVHQWV WKH UHVXOWV RI WKLV VLPXODWLRQ DQG WKH GHULYDWLRQ RI WKH H[SHULPHQWDO K\SRWKHVHV $VVXPSWLRQV RI VLPXODWLRQ 7KH ILUVW DVVXPSWLRQ HPSOR\HG IRU WKH VLPXODWLRQ RI VXEMHFWLYH LQYHVWLJDWLRQ GHFLVLRQ SHUIRUPDQFHV LV WKDW WKH VXEMHFWV ZLOO EHKDYH DV LI WKH\ XVH WKH DQFKRULQJ DQG DGMXVWPHQW KHXULVWLF GXULQJ WKH WUDLQLQJ SKDVH 7KH DQFKRULQJ DQG DGMXVWPHQW KHXULVWLF SURSRVHV WKDW

PAGE 60

7$%/( (67,0$7(6 2) 7+( 3$5$0(7(56 5(48,5(' 72 ),7 7+( 237,0$/ 02'(/ *,9(1 7+( 9$5,286 &21',7,216 ,QIRUPDWLRQ ([SHULPHQWDO 9DULDEOH 3DUDPHWHUV ,QIRUPDWLRQ 9DULDEOH $FWXDO 8QGHUO\LQJ 'DWD 6WUXFWXUH 3DUDPHWHUV &RVW 3DUDPHWHU 'LVWULEXWLRQ 9DULDEOH 'LVWULEXWLRQ 9DULDEOH 'LVWULEXWLRQ 9DULDEOH 9DULDEOH (VWLPDWH 6, 6 6, 6 6, 6 \ &O DO DO 35L &5 &D &5 D7KH SDUDPHWHU HVWLPDWHV JLYHQ WKH &O FRVW OHYHO IRU X \FM DQG 35c JLYHQ WKH & FRVW OHYHO DUH WKH VDPH DV WKRVH

PAGE 61

7$%/( 5(68/76 2) 7+( 6,08/$7,21 2) 237,0$/ 02'(/ 3(5)250$1&(6 *,9(1 7+( 9$5,286 &21',7,216 N ,QIRUPDWLRQ 'LVWULEXWLRQ &RVW 3LQYHVWLJDWH RXWRIFRQWUROf 3 LQYHVWLJDWH LQFRQWUROf 2SWLPDO &XWRII 9DOXH DFWXDO PLQXWHV LQFXUUHGf GN N 6, &O 6, & 6 &O 6 & 6, &O 6, & 6 &O 6 & FQ

PAGE 62

D VXEMHFW ZLOO VHOHFW DQ LQLWLDO DQFKRU DV D ILUVW GHFLVLRQ DSSUR[Ln PDWLRQ DQG ZLOO DGMXVW WKLV GHFLVLRQ DQFKRU WRZDUG DQ RSWLPDOO\ FRUUHFW SRLQW DV KH OHDUQV IURP KLV H[SHULHQFHV ZLWK WKH VWDWHV RI QDWXUH +RZHYHU WKH LQGLYLGXDOnV DGMXVWPHQW SURFHVV ZLOO QRW EH VXIILFLHQW LH KH ZLOO WHQG WR DSSURDFK WKH RSWLPDOO\ FRUUHFW SRLQW EXW ZLOO QRW DGMXVW FRPSOHWHO\ WR WKDW SRLQW 7KH VHFRQG DVVXPSWLRQ LV WKDW WKH ORFDWLRQ RI WKH LQLWLDO GHn FLVLRQ DQFKRU RQ WKH DFWXDO PLQXWHV LQFXUUHG FRQWLQXXP D[LVf ZLOO EH FRQGLWLRQDO XSRQ WKH OHYHO RI WKH GLVWULEXWLRQ YDULDEOH 7KH UHDVRQ IRU WKLV LV WKDW WKH LQLWLDO GHFLVLRQ DQFKRU LV H[SHFWHG WR EH ORFDWHG DW D FHQWUDO SRLQW EHWZHHQ WKH PHDQV RI WKH WZR VWDWHV 6LQFH WKH PHDQ RI WKH RXWRIFRQWURO GLVWULEXWLRQ VKLIWV ZLWK WKH OHYHO RI WKH GLVWULEXWLRQ YDULDEOH WKH FHQWUDO SRLQW EHWZHHQ WKH PHDQV RI WKH WZR VWDWHV DOVR VKLIWV ZLWK WKH OHYHO RI WKH GLVWULEXWLRQ YDULDEOH 7KH LQLWLDO GHFLVLRQ DQFKRU XVHG LQ WKH VLPXODWLRQ RI VXEMHFWLYH LQYHVWLn JDWLRQ GHFLVLRQ SHUIRUPDQFHV LV WKH JHRPHWULF LQWHUVHFWLRQ SRLQW RI WKH WZR VWDWHVn GLVWULEXWLRQ FXUYHV 7KLV SDUWLFXODU LQLWLDO GHFLVLRQ DQFKRU LV XVHG EHFDXVH LW LV WKH H[DFW FHQWUDO SRLQW EHWZHHQ WKH PHDQV RI WKH WZR VWDWHV +RZHYHU LW LV VHOHFWHG IRU RSHUDWLRQDO SXUSRVHV DQG LV QRW QHFHVVDU\ WR WKH FRQFHSWXDO GHYHORSPHQW $Q\ SRLQW RI FHQWUDO WHQGHQF\ ZRXOG EH VDWLVIDFWRU\ 7KH ILQDO DVVXPSWLRQ LV WKDW WKH VXEMHFWVn DGMXVWPHQW SURFHVV ZLOO EH DSSUR[LPDWHO\ HTXDO RYHU WKH YDULRXV WUHDWPHQW FRQGLWLRQV 7KH DGMXVWPHQW SURFHVV LV GHILQHG DV D OLQHDU PRYHPHQW DORQJ WKH DFWXDO PLQXWHV LQFXUUHG D[LV IURP WKH LQLWLDO GHFLVLRQ DQFKRU WRZDUGV WKH DSSURSULDWH RSWLPDO PRGHO GHFLVLRQ FXWRII 7KH PDJQLWXGH RI WKH OLQHDU PRYHPHQW XVHG LQ WKH VLPXODWLRQ RI VXEMHFWLYH LQYHVWLJDWLRQ GHFLVLRQ

PAGE 63

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n MHFW VDPSOH RI HLJKWf 7KH UHVXOWV RI WKLV ILUVW VWDJH VLPXODWLRQ DUH SUHVHQWHG LQ JUDSKLFDO IRUP LQ )LJXUH 7KLV ILJXUH GHSLFWV WKH VLPXODWHG LQLWLDO GHFLVLRQ DQFKRUV SRLQWV $f WKH VLPXODWHG ILQDO GHFLVLRQ DQFKRUV SRLQWV %f DQG WKH VLPXODWHG RSWLPDO FXWRII YDOXHV SRLQWV &f IRU HDFK RI WKH HLJKW FRPELQDWLRQV RI LQGHSHQGHQW YDULDEOHV 7KH DVVXPSWLRQV XVHG LQ VLPXODWLQJ WKH WUDLQLQJ SKDVH FDQ EH LQYHVWLJDWHG XVLQJ WKH UHVXOWV RI WKH H[SHULPHQWDO WUDLQLQJ SKDVH 7KH DVVXPSWLRQV DUH Lf WKDW WKH VXEMHFWV ZLOO EHKDYH DV LI WKH\ XVHG WKH DQFKRULQJ DQG DGMXVWPHQW KHXULVWLF f WKDW WKH LQLWLDO GHFLVLRQ DQFKRU ZLOO EH ORFDWHG FHQWUDOO\ EHWZHHQ WKH PHDQV RI WKH WZR VWDWHV DQG ZLOO EH GHSHQGHQW RQ WKH OHYHO RI WKH GLVWULEXWLRQ YDULDEOH DQG f WKDW WKH

PAGE 64

,, ,QIRUPDWLRQ /HYHO &RVW &O &RVW & 'LVWULEXWLRQ 6, L 'LVWULEXWLRQ 6 & % $ K + 'LVWULEXWLRQ 6, ,QIRUPDWLRQ /HYHO &RVW &O % $ &RVW & & % $ 'LVWULEXWLRQ ` 6 + 1RWH $OO VFDOHV DUH DFWXDO PLQXWHV LQFXUUHG SHU FKDLU SURGXFHG ),*85( 6,08/$7(' ,1,7,$/ '(&,6,21 $1&+256 32,176 $f ),1$/ '(&,6,21 $1&+256 32,176 %f $1' 237,0$/ &872)) 9$/8(6 32,176 &f

PAGE 65

DGMXVWPHQW SURFHVV ZLOO EH DSSUR[LPDWHO\ HTXDO RYHU WKH YDULRXV FRQn GLWLRQV 7KH ORFDWLRQ RI WKH LQLWLDO GHFLVLRQ DQFKRU FDQ EH LQYHVWLJDWHG E\ H[DPLQLQJ WKH UHODWLRQVKLS EHWZHHQ WKH LQGLYLGXDOnV LQLWLDO FXWRII YDOXH LQ WKH ILUVW WUDLQLQJ VHVVLRQf DQG WKH LQWHUVHFWLRQ SRLQW RI WKH DSSURSULDWH WZR GLVWULEXWLRQ FXUYHV WKH 7'9c PHDVXUHf 7KHVH LQWHUVHFWLRQ SRLQWV DUH FRQGLWLRQDO XSRQ WKH OHYHO RI WKH GLVWULEXWLRQ YDULDEOH *LYHQ WKH 6, GLVWULEXWLRQ OHYHO WKH LQWHUVHFWLRQ SRLQW LV DFWXDO PLQXWHV LQFXUUHG DQG JLYHQ WKH 6 GLVWULEXWLRQ OHYHO WKH LQWHUVHFWLRQ SRLQW LV DFWXDO PLQXWHV LQFXUUHG VHH )LJXUH IRU D JUDSKLFDO SUHVHQWDWLRQ RI WKHVH SRLQWVf 7KH H[SHFWHG UHODWLRQVKLS LV WKDW WKH PHDQ LQLWLDO GHFLVLRQ DQFKRU 7'9f RI WKH LQGLYLGXDOV JLYHQ D OHYHO RI WKH GLVWULEXWLRQ YDULDEOH ZLOO QRW EH VLJQLILFDQWO\ GLIIHUn HQW IURP WKH LQWHUVHFWLRQ SRLQW RI WKH DSSURSULDWH WZR GLVWULEXWLRQ FXUYHV 7KH DVVXPSWLRQ RI DSSUR[LPDWHO\ HTXDO DGMXVWPHQW RYHU WKH YDULRXV FRQGLWLRQV FDQ EH LQYHVWLJDWHG E\ H[DPLQLQJ WKH UHODWLRQVKLS EHWZHHQ WKH LQGLYLGXDOnV DGMXVWPHQW RYHU WKH FRPSOHWH H[SHULPHQW DQG WKH WRWDO DGMXVWPHQW WKDW ZRXOG EH UHTXLUHG WR UHDFK WKH RSWLPDO PRGHOnV GHFLVLRQ FXWRII YDOXH $ PHDVXUH RI WKH UHODWLYH DGMXVWPHQW IRU LQGLYLGXDO L 5$M LV FRQGLWLRQDO XSRQ WKH GLUHFWLRQ RI DGMXVWPHQW DORQJ WKH UHOHYDQW GHFLVLRQ D[LV LH WKH DFWXDO PLQXWHV LQFXUUHG SHU FKDLU SURGXFHGf 7KH H[SHFWHG UHODWLRQVKLS LV WKDW WKH PHDQ 5$A ZLOO QRW GLIIHU VLJQLILn FDQWO\ EHWZHHQ WKH OHYHOV RI WKH YDULRXV FRQGLWLRQV *LYHQ WKH DERYH VLPXODWLRQ DQG GLVFXVVLRQ WKH IROORZLQJ K\SRn WKHVHV FDQ EH GHULYHG +,OD 7'9M,6Of 7KH PHDQ LQLWLDO GHFLVLRQ DQFKRU JLYHQ WKH 6, GLVWULEXWLRQ OHYHO

PAGE 66

ZLOO HTXDO WKH LQWHUVHFWLRQ SRLQW RI WKH WZR GLVWULEXWLRQ FXUYHV ZLWKLQ WKDW OHYHO +,OE 7'9L _6f 7KH PHDQ LQLWLDO GHFLVLRQ DQFKRU JLYHQ WKH 6 GLVWULEXWLRQ OHYHO ZLOO HTXDO WKH LQWHUVHFWLRQ SRLQW RI WKH WZR GLVWULEXWLRQ FXUYHV ZLWKLQ WKDW OHYHO +,D 7'9@,f Af 7KH PHDQ LQLWLDO GHFLVLRQ DQFKRU JLYHQ WKH ,, LQIRUPDWLRQ OHYHO ZLOO HTXDO WKH PHDQ LQLWLDO GHFLVLRQ DQFKRU JLYHQ WKH LQIRUn PDWLRQ OHYHO +,E 7'9LL&Of 79Lc&f 7KH PHDQ LQLWLDO GHFLVLRQ DQFKRU JLYHQ WKH &O FRVW OHYHO ZLOO HTXDO WKH PHDQ LQLWLDO GHFLVLRQ DQFKRU JLYHQ WKH & FRVW OHYHO +,D 5$Lf 5$Lf 7KH PHDQ UHODWLYH DGMXVWPHQW JLYHQ WKH ,, LQIRUPDWLRQ OHYHO ZLOO HTXDO WKH PHDQ UHODWLYH DGMXVWPHQW JLYHQ WKH LQIRUPDWLRQ OHYHO +,E 5$A6Of 5$L c6f 7KH PHDQ UHODWLYH DGMXVWPHQW JLYHQ WKH 6, GLVWULEXWLRQ OHYHO ZLOO HTXDO WKH PHDQ UHODWLYH DGMXVWPHQW JLYHQ WKH 6 GLVWULEXWLRQ OHYHO +,F 5$L_&Of 5$_&f 7KH PHDQ UHODWLYH DGMXVWPHQW JLYHQ WKH &O FRVW OHYHO ZLOO HTXDO WKH PHDQ UHODWLYH DGMXVWPHQW JLYHQ WKH & FRVW OHYHO ,QGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ K\SRWKHVHV IRUPDWLRQ 7KH PHDVXUHV RI LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ VKRXOG JHQHUDOO\ EH XQDIIHFWHG E\ WKH LQGHSHQGHQW YDULDEOHV 7KHVH PHDVXUHV

PAGE 67

KDYH D JUHDWHU UHODWLRQVKLS ZLWK LQGLYLGXDO GHFLVLRQ SURFHVVHV WKDQ ZLWK WKH VSHFLILF FRQGLWLRQV RI D JHQHUDO GHFLVLRQ VLWXDWLRQ 7KH LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ PHDVXUH GL VKRXOG KDYH D UHn ODWLRQVKLS ZLWK WKH GLVWULEXWLRQ YDULDEOH RQO\ 7KH WKHRUHWLFDO G PHDVXUH XQGHU WKH 6, GLVWULEXWLRQ OHYHO LV DQG XQGHU WKH 6 GLVn WULEXWLRQ OHYHO LV 7KXV WKH DYHUDJH GO XQGHU WKH 6 GLVWULEXn WLRQ OHYHO VKRXOG EH JUHDWHU WKDQ XQGHU WKH 6, GLVWULEXWLRQ OHYHO 7KH '1$A PHDVXUH VKRXOG KDYH QR V\VWHPDWLF UHODWLRQVKLSV ZLWK WKH WKUHH LQGHSHQGHQW YDULDEOHV 7KLV PHDVXUH RI WKH UHODWLYH GHn YLDWLRQ RI WKH LQGLYLGXDOfV GHFLVLRQ PRGHO VHQVLWLYLW\ IURP WKDW RI WKH RSWLPDO PRGHOnV JHQHUDOO\ LV GXH WR LQGLYLGXDO GHFLVLRQ LQFRQn VLVWHQFLHV DQG WHPSRUDU\ SURFHVVLQJ HUURUV ,I LQGLYLGXDOV DUH UDQGRPL]HG LQWR WKH VSHFLILF GHFLVLRQ VLWXDWLRQV WKHUH DUH QR D SULRUL UHDVRQV IRU H[SHFWLQJ VLJQLILFDQW GLIIHUHQFHV LQ WKLV PHDVXUH GXH WR WKH LQGHSHQGHQW YDULDEOHV 7KH GLIIHUHQFH EHWZHHQ '1$A DQG '1c PHDVXUHV WKH HIIHFWV RI PXOWLSOH GHFLVLRQ DQFKRUV XSRQ WKH LQGLYLGXDOnV UHODWLYH GHFLVLRQ PRGHO VHQVLWLYLW\ 5HODWLRQVKLSV EHWZHHQ WKLV GLIIHUHQFH DQG WKH LQn GHSHQGHQW YDULDEOHV GHSHQG XSRQ ZKHWKHU WKH LQGHSHQGHQW YDULDEOHV DUH UHODWHG WR WKH FDXVHV XQGHUO\LQJ DQ LQGLYLGXDOnV XVH RI PRUH WKDQ RQH FXWRII YDOXH $ SRVVLEOH H[SODQDWLRQ LQYROYHV WKH GLVWULEXWLRQ YDULn DEOH $V WKH YDULDQFH RI WKH LQFRQWURO VWDWH LQFUHDVHV WKH DEVROXWH GLVWDQFH EHWZHHQ WKH ORZHVW YDOXHV RI WKH UDQGRP YDULDEOH RI LQWHUHVW DFWXDO PLQXWHV LQFXUUHGf DQG WKH PHDQ RI WKH LQFRQWURO VWDWH LQn FUHDVHV ,Q WXUQ DV WKLV GLVWDQFH LQFUHDVHV VXEMHFWV PD\ EH PRUH LQFOLQHG WR SHUFHLYH WKDW D VHFRQG RXWRIFRQWURO GLVWULEXWLRQ RYHUODSV WKH ORZHU UDQJH RI WKH LQFRQWURO VWDWH ,I VXFK SHUFHSWLRQV XQGHUO\

PAGE 68

WKH XVH RI PXOWLSOH FXWRII YDOXHV WKH DYHUDJH GLIIHUHQFH EHWZHHQ '1$c DQG '1M VKRXOG GLIIHU EHWZHHQ WKH WZR OHYHOV RI WKH GLVWULEXWLRQ YDULn DEOH $ VHFRQG SRVVLEOH H[SODQDWLRQ LQYROYHV WKH FRVW YDULDEOH $V WKH LQGLYLGXDOnV ILQDO FXWRII YDOXH LQFUHDVHV WKH DEVROXWH GLVWDQFH EHWZHHQ WKH LQGLYLGXDOnV ILQDO FXWRII YDOXH DQG WKH ORZHVW YDOXHV RI WKH UDQGRP YDULDEOH RI LQWHUHVW LQFUHDVHV ,Q WXUQ DV WKLV GLVWDQFH LQFUHDVHV VXEMHFWV DJDLQ PD\ EH PRUH LQFOLQHG WR SHUFHLYH WKDW D VHFn RQG RXWRIFRQWURO GLVWULEXWLRQ RYHUODSV WKH ORZHU UDQJH RI WKH LQFRQWURO VWDWH ,I VXFK SHUFHSWLRQV XQGHUO\ WKH XVH RI PXOWLSOH FXWRII YDOXHV WKH DYHUDJH GLIIHUHQFH EHWZHHQ '1$ DQG '1c VKRXOG GLIIHU EHWZHHQ WKH WZR OHYHOV RI WKH FRVW YDULDEOH %DVHG RQ WKH DERYH GLVFXVVLRQ WKH HIIHFWV RI WKH LQGHSHQGHQW YDULDEOHV RQ WKH LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ PHDVXUHV FDQ EH K\SRWKHVL]HG DV IROORZV +OD GM _6,f GM _6f 7KRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO ZLOO KDYH VLJQLILFDQWO\ VPDOOHU PHDQ GMV WKHQ WKRVH VXEMHFWV ZLWKLQ WKH 6 OHYHO +OE GL c,,f G cf 7KRVH VXEMHFWV ZLWKLQ WKH ,, LQIRUPDWLRQ OHYHO DQG ZLWKLQ WKH LQIRUPDWLRQ OHYHO ZLOO KDYH HTXDO PHDQ GAV +F Gr _&Of G _&f 7KRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO DQG ZLWKLQ WKH & FRVW OHYHO ZLOO KDYH HTXDO PHDQ GAV +D '1$LOOOf '“$M c f 7KRVH VXEMHFWV ZLWKLQ WKH ,, LQIRUPDWLRQ OHYHO DQG ZLWKLQ WKH LQIRUPDWLRQ OHYHO ZLOO KDYH HTXDO PHDQ '1$cV

PAGE 69

+E '1$M 6,f '1$M _6f 7KRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO DQG ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO ZLOO KDYH HTXDO PHDQ '1$M6 +F '“$ &Of '“$M ,&f 7KRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO DQG ZLWKLQ WKH & FRVW OHYHO ZLOO KDYH HTXDO PHDQ '1$cV +D '1$M'1M c6,f '1$c'1M _6f 7KRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO ZLOO KDYH VLJQLILFDQWO\ VPDOOHU PHDQ '1$c'1cV WKDQ ZLOO WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO +E '1$M'1M_&Of '1$L'1L_&f 7KRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZLOO KDYH VLJQLILFDQWO\ VPDOOHU PHDQ '1$c'1cV WKDQ ZLOO WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO ,QGLYLGXDO GHFLVLRQ FULWHULD VLPXODWLRQ DQG K\SRWKHVHV IRUPDWLRQ 7KH YDULDEOHV ZKLFK DIIHFW WKH %1M PHDVXUH VKRXOG EH WKRVH IDFWRUV UHODWHG WR WKH LQGLYLGXDOnV VHOHFWLRQ RI D FXWRII YDOXH 7KH PRVW VLJQLILFDQW YDULDEOH DIIHFWLQJ WKH LQGLYLGXDOnV FXWRII YDOXH VKRXOG EH WKH FRVW YDULDEOH :LWKLQ WKLV YDULDEOH WKH LQLWLDO GHFLVLRQ DQFKRU JLYHQ WKH &O OHYHO LV PXFK FORVHU WR WKH RSWLPDO FXWRII YDOXH WKDQ LV WKH LQLWLDO GHFLVLRQ DQFKRU JLYHQ WKH & OHYHO &RQVHTXHQWO\ WKH %1M PHDVXUH VKRXOG EH FORVHU WR D YDOXH RI RQH XQGHU WKH &O OHYHO WKDQ XQGHU WKH & OHYHO 7KH GLIIHUHQFH EHWZHHQ WKH %1$c DQG WKH %1M PHDVXUHV VKRXOG UHODWH WR WKH VDPH YDULDEOHV DV WKRVH GLVFXVVHG IRU WKH GLIIHUHQFH

PAGE 70

EHWZHHQ WKH '1$c DQG '1A PHDVXUHV 7KH WZR SRVVLEOH H[SODQDWLRQV LQn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n FHUQLQJ WKH LQGLYLGXDO GHFLVLRQ FULWHULD 7KH HDUOLHU VLPXODWLRQ GHn ULYHG WKH IROORZLQJ ILQDO GHFLVLRQ FXWRII YDOXHV f DFWXDO PLQXWHV LQFXUUHG IRU WKH &O FRVW OHYHO JLYHQ WKH LQIRUPDWLRQ OHYHO DQG WKH 6, GLVWULEXWLRQ OHYHO DQG f DFWXDO PLQXWHV LQFXUUHG IRU WKH & FRVW OHYHO JLYHQ WKH LQIRUPDWLRQ OHYHO DQG WKH 6, GLVn WULEXWLRQ OHYHO *LYHQ WKHVH FXWRII YDOXHV D IHAf PHDVXUH FDQ EH FDOFXODWHG GLUHFWO\ WKH A PHDVXUH LV XVHG UDWKHU WKDQ WKH HQ PHDVXUH GXH WR WKH DVVXPSWLRQ RI D VLQJOH FXWRII YDOXHf 7KH IXQFWLRQDO QRWDWLRQ I f LV HPSOR\HG WR LQGLFDWH WKHVH HVWLPDWHV DUH EDVHG XSRQ D VLPXODWLRQ UDWKHU WKDQ XSRQ HPSLULFDO REVHUYDWLRQ )RU WKLV VLPXn ODWLRQ WKH I"fV DUH DV IROORZV Ir_&O6O,f !]RXWf IU]Qf I%r_&6O,f M!]4XWf r]Qf ZKHUH M! f LQGLFDWHV WKH YDOXH RI WKH QRUPDO GHQVLW\ IXQFWLRQ DW WKH SRLQW LQ SDUHQWKHVHV *LYHQ WKH I%Af PHDVXUHV WKH I%1$Af PHDVXUHV FDQ EH FDOFXODWHG )RU WKLV VLPXODWLRQ WKH I%1$AfV DUH DV IROORZV

PAGE 71

I%1$L_&O6,f I%pf INf I%1$L_&6O,f IH"f II-Nf ZKHUH INf LV WKH VLPXODWHG PHDVXUH DW WKH FXWRII YDOXH RI WKH NWK RSWLPDO PRGHO 7KLV VLPXODWLRQ FDQ EH H[WHQGHG WR WKH 6 OHYHO RI WKH GLVWULEXn WLRQ YDULDEOH JLYHQ WKH LQIRUPDWLRQ OHYHO DQG FDQ EH H[WHQGHG WR ERWK WKH 6, DQG 6 OHYHOV RI WKH GLVWULEXWLRQ YDULDEOH JLYHQ WKH ,, LQIRUPDWLRQ OHYHO 7KH I%1$f PHDVXUHV REWDLQHG XQGHU WKH LQIRUn PDWLRQ DQG 6 GLVWULEXWLRQ FRQGLWLRQV DUH DV IROORZV I%1$L-&O6f I%1$Lc&6,f 7KH I%1$f PHDVXUHV JLYHQ WKH ,, LQIRUPDWLRQ OHYHO DUH DV IROORZV I%1$L_&O6O,Of I%1$ &6,f I%1$L_&O6,Of I%1$_&6,Of %DVHG RQ WKHVH I%1$f PHDVXUHV K\SRWKHVHV FDQ EH GHULYHG FRQn FHUQLQJ WKH HIIHFWV RI WKH LQGHSHQGHQW YDULDEOHV XSRQ WKH %1$ PHDVXUH 7KH I%1$Af PHDVXUHV FDQ EH DYHUDJHG RYHU WKH DSSURSULDWH FRQGLWLRQDOV WR GHWHUPLQH WKHVH HIIHFWV $YHUDJLQJ RYHU DOO FRQGLWLRQDOV H[FHSW WKH LQIRUPDWLRQ YDULDEOH JLYHV WKH IROORZLQJ UHVXOWV I%“$L,,,f I%1$L_f $OWKRXJK WKHUH LV D GLIIHUHQFH LQ WKH I%1$f PHDVXUHV GXH WR WKH LQIRUn PDWLRQ YDULDEOH WKLV GLIIHUHQFH LV VPDOO FRPSDUHG WR WKH VWDQGDUG HUURU RI WKHVH HVWLPDWHV WKH GLIIHUHQFH GLYLGHG E\ WKH VWDQGDUG HUURU LV f 7KHUHIRUH D VLJQLILFDQW HIIHFW GXH WR WKH LQIRUPDWLRQ YDULDEOH ZRXOG QRW EH H[SHFWHG

PAGE 72

$YHUDJLQJ RYHU DOO FRQGLWLRQDOV H[FHSW WKH GLVWULEXWLRQ YDULDEOH JLYHV WKH IROORZLQJ UHVXOWV I%“$c 6,f I%1$L_6f $JDLQ WKH GLIIHUHQFH LQ WKH I%1$Mf PHDVXUHV GXH WR WKH GLVWULEXWLRQ YDULDEOH LV VPDOO FRPSDUHG WR WKH VWDQGDUG HUURU RI WKH HVWLPDWHV WKH GLIIHUHQFH GLYLGHG E\ WKH VWDQGDUG HUURU LV &f $ VLJQLILFDQW HIIHFW GXH WR WKH GLVWULEXWLRQ YDULDEOH ZRXOG QRW EH H[SHFWHG $YHUDJLQJ RYHU DOO FRQGLWLRQDOV H[FHSW WKH FRVW YDULDEOH JLYHV WKH IROORZLQJ UHVXOWV I %“$M _&Of I%“$c _&f 7KH GLIIHUHQFH LQ WKH I%1$Mf PHDVXUHV GXH WR WKH FRVW YDULDEOH LV ODUJH FRPSDUHG WR WKH VWDQGDUG HUURU RI WKH HVWLPDWHV WKH GLIIHUHQFH GLYLGHG E\ WKH VWDQGDUG HUURU LV f &RQn VHTXHQWO\ D VLJQLILFDQW HIIHFW ZRXOG EH H[SHFWHG IRU WKH FRVW YDULDEOH $QRWKHU LPSOLFDWLRQ GUDZQ IURP WKH I%1$Mf PHDVXUHV FRQFHUQV WKH KRPRJHQHLW\ RI YDULDQFH IRU WKH PHDVXUHV JLYHQ GLIIHUHQW OHYHOV RI WKH FRVW YDULDEOH 7KH YDULDQFH IRU WKH I%1$c &Of LV DQG WKH YDULn DQFH IRU WKH I%1$c _&f LV 8VLQJ WKH ) WHVW IRU HTXDO YDULDQFHV ) DQG GIf ZKLFK LQGLFDWHV WKH YDULDQFHV PD\ QRW EH HTXDO 7KH YDULDQFH RI %1$c>&f ZRXOG EH H[SHFWHG WR EH OHVV WKDQ WKH YDULn DQFH RI %1$Mc&f 7KLV VLPXODWLRQ FDQ EH H[WHQGHG WR HQDEOH WKH IRUPDWLRQ RI K\SRWKHVHV FRQFHUQLQJ WKH HIIHFWV RI WKH LQGHSHQGHQW YDULDEOHV RQ WKH

PAGE 73

FRQVHUYDWLVP PHDVXUH %1&c 7KLV H[WHQVLRQ HPSOR\V WKH I%Mf DQG I8Af PHDVXUHV XVHG DERYH 7KH UHVXOWV RI WKLV H[WHQGHG VLPXODWLRQ DYHUDJHG RYHU DOO FRQGLWLRQDOV H[FHSW IRU WKH LQIRUPDWLRQ YDULDEOH DUH DV IROORZV I%1&fL ,,f I%“&c_,f 7KH GLIIHUHQFH LQ WKH I%1&Af PHDVXUHV GXH WR WKH LQIRUPDWLRQ YDULDEOH LV VPDOO FRPSDUHG WR WKH VWDQGDUG HUURU RI WKH HVWLPDWHV WKH GLIIHUn HQFH GLYLGHG E\ WKH VWDQGDUG HUURU LV f $ VLJQLILFDQW HIIHFW GXH WR WKH LQIRUPDWLRQ YDULDEOH ZRXOG QRW EH H[SHFWHG $YHUDJLQJ RYHU DOO FRQGLWLRQDOV H[FHSW IRU WKH GLVWULEXWLRQ YDULDEOH JLYHV WKH IROORZLQJ UHVXOWV I%“&L_6,f I %1&L_6f 7KH GLIIHUHQFH LQ WKH I %1&cf PHDVXUHV GXH WR WKH GLVWULEXWLRQ YDULn DEOH LV VPDOO FRPSDUHG WR WKH VWDQGDUG HUURU RI WKH HVWLPDWHV WKH GLIIHUHQFH GLYLGHG E\ WKH VWDQGDUG HUURU LV f $ VLJQLILFDQW HIIHFW GXH WR WKH GLVWULEXWLRQ YDULDEOH ZRXOG QRW EH H[SHFWHG )LQDOO\ DYHUDJLQJ RYHU DOO FRQGLWLRQDOV H[FHSW IRU WKH FRVW YDULDEOH JLYHV WKH IROORZLQJ UHVXOWV I%“&Bc L &Of I%“&L_&f 7KH GLIIHUHQFH LQ WKH I%“&cf PHDVXUH GXH WR WKH FRVW YDULDEOH LV ODUJH FRPSDUHG WR WKH VWDQGDUG HUURU RI WKH HVWLPDWHV WKH GLIIHUHQFH GLYLGHG E\ WKH VWDQGDUG HUURU LV f &RQVHTXHQWO\ D VLJQLILFDQW HIIHFW ZRXOG EH H[SHFWHG IRU WKH FRVW YDULDEOH

PAGE 74

7KH DERYH GLVFXVVLRQ FRQFHUQLQJ WKH LQGLYLGXDO GHFLVLRQ FULWHULD PHDVXUHV FDQ EH VXPPDUL]HG DV IROORZV +D %“ƒcOOOf %“$M f 7KH PHDQ %1$M RI WKRVH VXEMHFWV ZLWKLQ WKH ,, LQIRUPDWLRQ OHYHO DQG ZLWKLQ WKH LQIRUPDWLRQ OHYHO ZLOO EH HTXDO +OE %“ƒLO6Of %“$c 6f 7KH PHDQ %1$c RI WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO DQG ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO ZLOO EH HTXDO +F %“$M &Of %“$c &f 7KRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZLOO KDYH D VLJQLILFDQWO\ ODUJHU PHDQ %1$M WKDQ ZLOO WKRVH VXEMHFWV ZLWKLQ WKH & OHYHO + D%1$M >&Of D%1$c &f 7KH %1$c RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZLOO KDYH D VLJQLILFDQWO\ VPDOOHU YDULDQFH WKDQ ZLOO WKH %1$c RI WKRVH VXEn MHFWV ZLWKLQ WKH & FRVW OHYHO +D %“&M c,,f %“&M f 7KH PHDQ %1&c RI WKRVH VXEMHFWV ZLWKLQ WKH ,, LQIRUPDWLRQ OHYHO DQG ZLWKLQ WKH LQIRUPDWLRQ OHYHO ZLOO EH HTXDO +E %“&M_6,f %“&M_6f 7KH PHDQ %1&M RI WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO DQG ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO ZLOO EH HTXDO +F %“&M_&Of %“&M M &f 7KH PHDQ %1&M RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZLOO EH VLJQLILFDQWO\ VPDOOHU WKDQ WKH PHDQ %1&M RI WKRVH VXEMHFWV ZLWKLQ WKH & OHYHO

PAGE 75

+D %1$L%1L_6Of %1$M%1c6f 7KH PHDQ %1$%1A RI WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO ZLOO EH VLJQLILFDQWO\ VPDOOHU WKDQ WKH PHDQ %1$A%1A RI WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO +E %1$%1_&f %1$L%1A_&f 7KH PHDQ %1$A%1A RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZLOO EH VLJQLILFDQWO\ VPDOOHU WKDQ WKH PHDQ %1$%1 RI WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO ,QGLYLGXDO ORQJUXQ GHFLVLRQ HIILFLHQF\ VLPXODWLRQ DQG K\SRWKHVHV 7KH VLPXODWLRQ DQG DVVXPSWLRQV XVHG LQ GHYHORSLQJ WKH WUDLQLQJ SKDVH K\SRWKHVHV HQDEOH WKH IRUPDWLRQ RI K\SRWKHVHV FRQFHUQLQJ WKH *M YDULDEOH 7KH JUDSK SUHVHQWHG LQ )LJXUH GHSLFWV WKH UHODWLRQV RI WKH YDOXHV GHULYHG ZLWKLQ WKH WUDLQLQJ SKDVH VLPXODWLRQ JLYHQ WKH FRQGLWLRQ 6,f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

PAGE 76

R R /' R /' R R R /2 /2 /' /' R /2 R &2 ,' &?/' &0 fn f f f f f f /2 /' UA &2 FU! R mf§+ &2 &2 &2 &2 &2 rf ),*85( *5$3+ 2) 7+( 6,08/$7,21 86(' ,1 7+( &21',7,21 ,1)250$7,21 6, ',675,%87,21f

PAGE 77

$UHD*1M _&O6,f $UHD*3c _&O6, f $UHDL*1L_&6,f $UHD*3L_&6O,f 7KHVH DUHDV PXVW EH DGMXVWHG IRU WKH UHODWLYH IUHTXHQF\ GLIIHUn HQFHV DQG WKH UHODWLYH FRVW GLIIHUHQFHV EHWZHHQ WKH YDULRXV FRQGLWLRQV ,I WKH LQFRQWURO GLVWULEXWLRQ LV JLYHQ D UHODWLYH IUHTXHQF\ ZHLJKW RI RQHWKHQ WKH UHODWLYH IUHTXHQF\ ZHLJKW RI WKH RXWRIFRQWURO GLVWULEXn WLRQ ZRXOG HTXDO WZRWKLUGV 3LQFRQWUROf DQG 3RXWRIFRQWUROf f 7KH UHODWLYH FRVW ZHLJKWV FDQ EH REWDLQHG GLUHFWO\ IURP WKH PDQLSXODWLRQ RI WKH FRVW YDULDEOH *LYHQ WKH &O FRVW OHYHO HUURUV XQGHU WKH RXWRIFRQWURO GLVWULEXWLRQ HTXDO WKUHH WLPHV WKH HUURUV XQGHU WKH LQFRQWURO GLVWULEXWLRQ *LYHQ WKH & FRVW OHYHO HUURUV XQGHU WKH LQFRQWURO GLVWULEXWLRQ DUH HTXDO WR WKUHH WLPHV WKH HUURUV XQGHU WKH RXWRIFRQWURO GLVWULEXWLRQ 7KH IRXU DUHDV DVVRFLDWHG ZLWK WKH YDULRXV *c PHDVXUHV FDQ EH FRQYHUWHG LQWR UHODWLYH UHODWLRQV E\ WDNLQJ LQWR DFFRXQW ERWK RI WKHVH UHODWLYH ZHLJKWV IUHTXHQF\ DQG FRVWf 7KHVH UHODWLYH UHODWLRQV DUH DV IROORZV I*1c &O6f f f f I*3M &O6, f f f f f f I*1L_&6O,f fff I*3L_&6O,f f f f f f 7KH GLIIHUHQFH EHWZHHQ WKH UHODWLYH UHODWLRQV RI *3c DQG *1M JLYHQ D OHYHO RI WKH FRVW YDULDEOH LQGLFDWHV WKH UHODWLYH HIIHFW RI WKHVH LQWHUYDOV XSRQ WKH *M PHDVXUH 7KH GLIIHUHQFHV IRU WKLV VLPXn ODWLRQ DUH DV IROORZV

PAGE 78

IL_&O6O,f I*3L_&6f I*1A_&6,f I*L _&6f I*3L _&6f I*1 _&6f 7KLV VLPXODWLRQ FDQ EH H[WHQGHG WR WKH 6 OHYHO RI WKH GLVWULEXn WLRQ YDULDEOH JLYHQ WKH LQIRUPDWLRQ OHYHO DQG FDQ EH H[WHQGHG WR ERWK WKH 6, DQG 6 GLVWULEXWLRQ OHYHOV JLYHQ WKH ,, LQIRUPDWLRQ OHYHO 7KH UHODWLYH UHODWLRQV RI WKH PHDVXUH REWDLQHG XQGHU WKH LQIRUn PDWLRQ DQG 6 GLVWULEXWLRQ FRQGLWLRQV DUH DV IROORZV I*L_&O6,f I*L_&6,f 7KH UHODWLYH UHODWLRQV IRU WKH *c PHDVXUH JLYHQ WKH ,, LQIRUPDWLRQ OHYHO DUH DV IROORZV I*M_&6,f I*L_&6,,f I*_&O6,Of I*L_&6,Of %DVHG XSRQ WKHVH UHODWLYH UHODWLRQV K\SRWKHVHV FDQ EH GHULYHG FRQFHUQLQJ WKH HIIHFWV RI WKH LQGHSHQGHQW YDULDEOHV RQ WKH *A PHDVXUH 7KHVH UHODWLRQV FDQ EH DYHUDJHG RYHU WKH DSSURSULDWH FRQGLWLRQDOV WR GHWHUPLQH WKH HIIHFWV $YHUDJLQJ RYHU DOO FRQGLWLRQDOV H[FHSW WKH LQIRUPDWLRQ YDULDEOH JLYHV WKH IROORZLQJ UHVXOWV I*L_,,f I&7M f $OWKRXJK WKHUH LV D GLIIHUHQFH LQ WKH I*f GXH WR WKH LQIRUPDWLRQ YDULn DEOHWKLV GLIIHUHQFH LV VPDOO FRPSDUHG WR WKH VWDQGDUG HUURU RI WKH HVWLPDWHV WKH GLIIHUHQFH GLYLGHG E\ WKH VWDQGDUG HUURU LV

PAGE 79

f 7KHUHIRUH D VLJQLILFDQW HIIHFW GXH WR WKH LQIRUPDWLRQ YDULDEOH ZRXOG QRW EH H[SHFWHG $YHUDJLQJ RYHU DOO FRQGLWLRQDOV H[FHSW WKH GLVWULEXWLRQ YDULDEOH JLYHV WKH IROORZLQJ UHVXOWV IA _6,f I*_6f $OWKRXJK WKH GLIIHUHQFH EHWZHHQ WKHVH PHDVXUHV LV PXFK ODUJHU WKDQ WKDW DERYH WKLV GLIIHUHQFH LV RQO\ VOLJKWO\ ODUJHU WKDQ WKH VWDQGDUG HUURU RI WKH HVWLPDWHV WKH GLIIHUHQFH GLYLGHG E\ WKH VWDQGDUG HUURU LV f 7KH GLUHFWLRQ RI WKH I*Af GLIIHUHQFHV ZRXOG EH H[SHFWHG WR EH DV SUHGLFWHG E\ WKH VLPXODWLRQ EXW PD\ QRW EH VLJQLILFDQW $YHUDJLQJ RYHU DOO FRQGLWLRQDOV H[FHSW WKH FRVW YDULDEOH JLYHV WKH IROORZLQJ UHVXOWV I*_&Of I*_&f 7KH GLIIHUHQFH EHWZHHQ WKHVH PHDVXUHV LV ODUJH FRPSDUHG WR WKH VWDQGDUG HUURU RI WKH HVWLPDWHV WKH GLIIHUHQFH GLYLGHG E\ WKH VWDQGDUG HUURU LV f &RQVHTXHQWO\ D VLJQLILFDQW HIIHFW ZRXOG EH H[SHFWHG IRU WKH FRVW YDULDEOH 7KHVH UHODWLYH UHODWLRQV FDQ EH DYHUDJHG RYHU DOO EXW SDLUV RI FRQGLWLRQDOV 6XFK PHDVXUHV FRXOG LQGLFDWH WKH SUHVHQFH RI LQWHUDFWLRQ HIIHFWV RI WKH LQGHSHQGHQW YDULDEOHV 7KH UHVXOWV RI DYHUDJLQJ WKH SDLUZLVH FRQGLWLRQDOV LQGLFDWH RQO\ RQH LQWHUDFWLRQ RI SRVVLEOH VLJQLILFDQFH WKH GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ 7KH I*Af PHDVXUHV FRQGLWLRQDO WR WKHVH YDULDEOHV DUH DV IROORZV I t_6,&Of

PAGE 80

I[__6O&f I*L_6&Of I*L_6&f $ JUDSKLF UHSUHVHQWDWLRQ RI WKLV LQWHUDFWLRQ LV SUHVHQWHG LQ )LJXUH 7KH GLIIHUHQFH EHWZHHQ WKH WZR I*_&Of PHDVXUHV LV PXFK VPDOOHU WKDQ WKH GLIIHUHQFH EHWZHHQ WKH WZR IA_&f PHDVXUHV 7KH GLIIHUHQFHV DUH DQG UHVSHFWLYHO\ 7KLV ZRXOG LQGLFDWH WKDW WKH GLVWULEXWLRQ YDULDEOH KDV D VLJQLILFDQW HIIHFW RQO\ ZKHQ JLYHQ WKH & FRVW OHYHO 6XFK T UHVXOW ZRXOG H[SODLQ WKH ODUJHU EXW QRW VLJQLILn FDQW HIIHFW RI WKH 6, GLVWULEXWLRQ OHYHO RQ WKH I*A_6 f PHDVXUHV 2WKHU LPSOLFDWLRQV ZKLFK FDQ EH GHULYHG IURP WKHVH UHODWLYH UHODWLRQV LQYROYH UHODWLRQVKLSV EHWZHHQ I*3f DQG I*1Af PHDVXUHV 7ZR VXFK UHODWLRQVKLSV DSSHDU WR EH RI SRVVLEOH VLJQLILFDQFH %RWK LQYROYH WKH UDWLR I*3AfI*1Af LH WKH HIIHFW RI DGGLWLRQDO HUURUV UHODWLYH WR WKH HIIHFW RI DGGLWLRQDO FRUUHFW GHFLVLRQVf 7KH UHVXOWV RI WKH ILUVW UHODWLRQVKLS DUH DV IROORZV I*3L*1L_&f I*3L*1L_&f 7KH GLIIHUHQFH EHWZHHQ WKHVH PHDVXUHV LV ODUJH FRPSDUHG WR WKH VWDQGDUG HUURU RI WKH HVWLPDWHV WKH GLIIHUHQFH GLYLGHG E\ WKH VWDQGDUG HUURU LV f &RQVHTXHQWO\ WKH *3*1 UDWLR LV H[n SHFWHG WR EH ODUJHU JLYHQ WKH & FRVW OHYHO WKDQ JLYHQ WKH &O FRVW OHYHO 7KH UHVXOWV RI WKH VHFRQG UHODWLRQVKLS DUH DV IROORZV I*3L*1L_6O&Of I*3L*1L_6O&f I*3*L1L_6&Of

PAGE 81

I*f 'LVWULEXWLRQ 9DULDEOH ),*85( 6,08/$7(' &267 %< ',675,%87,21 9$5,$%/( ,17(5$&7,21 21 I*cf

PAGE 82

I*3 M*1c _6&f 7KH GLIIHUHQFH EHWZHHQ WKH WZR I*3c*1c _6f LV PXFK VPDOOHU WKDQ WKH GLIIHUHQFH EHWZHHQ WKH WZR I *3 c*1 M _6f 7KLV ZRXOG LQGLFDWH WKH SUHVHQFH RI D GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ LQ ZKLFK WKH HIIHFW RI WKH FRVW YDULDEOH LV VLJQLILFDQW RQO\ XQGHU WKH 6, GLVWULEXWLRQ OHYHO $ JUDSKLFDO UHSUHVHQWDWLRQ RI WKLV LQWHUDFWLRQ LV SUHVHQWHG LQ )LJXUH $ ILQDO LPSOLFDWLRQ GHULYHG IURP WKHVH UHODWLYH UHODWLRQV LQn YROYHV WKH KRPRJHQHLW\ RI YDULDQFH IRU WKH I*Af JLYHQ WKH GLIIHUHQW OHYHOV RI WKH FRVW YDULDEOH 7KH YDULDQFH IRU WKH I*A_&Of LV HTXDO WR [ IRU WKLV VLPXODWLRQ DQG WKH YDULDQFH IRU WKH I*A>&f LV HTXDO WR [ IRU WKLV VLPXODWLRQ 8VLQJ WKH ) WHVW IRU HTXDO YDULDQFHV ) DQG GIf ZKLFK LQGLFDWHV WKH YDULDQFHV DUH QRW HTXDO &RQVHTXHQWO\ WKH YDULDQFH RI *c &Of LV H[SHFWHG WR EH OHVV WKDQ WKH YDULDQFH RI *c_&f *LYHQ WKH DERYH VLPXODWLRQ WKH HIIHFWV RI WKH LQGHSHQGHQW YDULn DEOHV RQ WKH *c YDULDEOH FDQ EH K\SRWKHVL]HG DV IROORZV + D*M_&Of D*L_&f 7KH *c RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZLOO KDYH VLJQLILFDQWO\ VPDOOHU YDULDQFH WKDQ ZLOO WKH *c RI WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO +D *LO&Of t__&f 7KRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZLOO KDYH D VLJQLILFDQWO\ VPDOOHU PHDQ *M WKDQ ZLOO WKRVH VXEMHFWV ZLWKLQ WKH & OHYHO +E *L_6f *L_6Of 7KRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO ZLOO KDYH VPDOOHU PHDQ *M WKDQ ZLOO WKRVH VXEMHFWV ZLWKLQ WKH 6, OHYHO 6LJQLILn FDQFH LV QRW SUHGLFWHG GXH WR WKH LQWHUDFWLRQ HIIHFW ZLWK WKH FRVW YDULDEOH

PAGE 83

'LVWULEXWLRQ 9DULDEOH ),*85( 6,08/$7(' &267 %< ',675,%87,21 9$5,$%/( ,17(5$&7,21 21 I*3L*1Lf

PAGE 84

+F *M ,,f *L f 7KH PHDQ *M RI WKRVH VXEMHFWV ZLWKLQ WKH ,, LQIRUPDWLRQ OHYHO DQG ZLWKLQ WKH LQIRUPDWLRQ OHYHO ZLOO EH HTXDO + *fL _6,&Of *M _6&f *L@6O&f *nL -6&f 7KHUH ZLOO EH D VLJQLILFDQW LQWHUDFWLRQ RI WKH GLVWULEXWLRQ DQG FRVW YDULDEOHV LQ ZKLFK WKH GLVWULEXWLRQ YDULDEOH ZLOO KDYH QR VLJQLILFDQW HIIHFW JLYHQ WKH &O FRVW OHYHO EXW ZLOO KDYH D VLJQLILFDQW HIIHFW JLYHQ WKH & FRVW OHYHO +D *3M*1M _&Of *3*1L _&f 7KRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZLOO KDYH D VLJQLILFDQWO\ VPDOOHU PHDQ *3M*1M WKDQ ZLOO WKRVH VXEMHFWV ZLWKLQ WKH & OHYHO + *3M*1M _&6f *3M*1M &O6f *3M*1M c&6f *3M*1M @ &O6f 7KHUH ZLOO EH D VLJQLILFDQW LQWHUDFWLRQ RI WKH GLVWULEXWLRQ DQG FRVW YDULDEOHV LQ ZKLFK WKH FRVW YDULDEOH ZLOO KDYH QR VLJQLILn FDQW HIIHFW JLYHQ WKH 6 GLVWULEXWLRQ OHYHO EXW ZLOO KDYH D VLJQLILFDQW HIIHFW JLYHQ WKH 6, OHYHO ,QGLYLGXDO ORQJUXQ GHFLVLRQ HIILFLHQF\ DQG WKH WUDLQLQJ SKDVH 6LQFH SHUIRUPDQFH ZLWKLQ WKH H[SHULPHQW LV DVVXPHG WR EH D IXQFWLRQ RI WKH WUDLQLQJ DGMXVWPHQW SURFHVV LW ZRXOG EH H[SHFWHG WKDW WKH VXEMHFWV ZLWKLQ WKH 33 FODVVLILFDWLRQ RI WUDLQLQJ DGMXVWPHQW SDWWHUQ ZRXOG KDYH D ORZHU DYHUDJH H[SHULPHQW *c PHDVXUH WKDQ WKRVH VXEMHFWV ZLWKLQ WKH 00 FODVVLILFDWLRQ )XUWKHUPRUH WKH PHDQ H[SHFWHG *M PHDVXUH IRU WKH PL[HG FODVVLILFDWLRQV 30 DQG 03f ZRXOG EH H[SHFWHG WR IDOO EHWZHHQ WKRVH RI WKH QRQPL[HG FODVVLILFDWLRQV 7KH PL[HG

PAGE 85

FODVVLILFDWLRQV KDYH RQH SRVLWLYH DGMXVWPHQW ZKLFK ZRXOG LPSURYH SHUn IRUPDQFH DERYH WKDW RI VXEMHFWV ZKR PDNH RQO\ QHJDWLYH DGMXVWPHQWV 2Q WKH RWKHU KDQG WKH PL[HG FODVVLILFDWLRQV KDYH RQH QHJDWLYH DGMXVWPHQW ZKLFK ZRXOG GHFUHDVH SHUIRUPDQFH EHORZ WKDW RI VXEMHFWV ZKR PDNH RQO\ SRVLWLYH DGMXVWPHQWV 7KH IROORZLQJ K\SRWKHVHV SUHGLFW WKH HIIHFWV RI WKH WUDLQLQJ DGMXVWPHQW SDWWHUQ RQ WKH *L YDULDEOH + *M_33f *A_00f 7KRVH VXEMHFWV ZLWKLQ WKH 33 FODVVLILFDWLRQ RI WUDLQLQJ DGMXVWn PHQW SDWWHUQ ZLOO KDYH D VLJQLILFDQWO\ VPDOOHU PHDQ PHDVXUH WKDQ WKRVH VXEMHFWV ZLWKLQ WKH 00 FODVVLILFDWLRQ + _33f *M _03f 7M _30f *7M _00f 7KRVH VXEMHFWV ZLWKLQ WKH PL[HG FODVVLILFDWLRQV RI WUDLQLQJ DGMXVWPHQW SDWWHUQ ZLOO KDYH PHDQ *AV WKDW IDOO EHWZHHQ WKRVH RI WKH QRQPL[HG FODVVLILFDWLRQV 7KH PHDQ *AV RI WKH PL[HG FODVVLn ILFDWLRQV DUH QRW H[SHFWHG WR EH VLJQLILFDQWO\ GLIIHUHQW

PAGE 86

&+$37(5 ,9 7+( (;3(5,0(17 ([SHULPHQWDO (QYLURQPHQW 7KH GHFLVLRQ VLWXDWLRQ VWXGLHG ZLWKLQ WKLV UHVHDUFK ZDV VWDQGDUG FRVW YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ PDNLQJ DW WKH RSHUDWLRQDO PDQDJHn PHQW OHYHO 7KH H[SHULPHQWDO VHWWLQJ ZDV WKH VLPXODWHG HQYLURQPHQW RI D PDQXIDFWXULQJ FRPSDQ\ DQG WKH VXEMHFWV ZHUH UHTXHVWHG WR DVVXPH WKH UROH RI DQ RSHUDWLRQDO PDQDJHU RI DQ DVVHPEO\ GHSDUWPHQW ZLWKLQ WKLV HQYLURQPHQW 7KH DVVHPEO\ GHSDUWPHQW DVVHPEOHG D VLQJOH SURGXFW D PHWDO IROGLQJ FKDLU 6XEMHFWV ZHUH SUHVHQWHG ZLWK D VHULHV RI VWDQGDUG FRVW YDULDQFH UHSRUWV IRU WKH GHSDUWPHQW DQG ZHUH DVNHG WR GHFLGH IRU HDFK UHSRUW ZKHWKHU WKH XQGHUO\LQJ SK\VLFDO SURFHVV VKRXOG EH LQYHVWLn JDWHG WR FRUUHFW DQ RXWRIFRQWURO VLWXDWLRQ (DFK VHULHV RI YDULDQFH UHSRUWV ZHUH FURVVVHFWLRQDO LQ QDWXUH LH DOO UHSRUWV ZLWKLQ D VHULHV ZHUH DVVXPHG WR KDYH RFFXUUHG VLPXOWDQHRXVO\ DQG ZHUH LQGHSHQGHQW RI HDFK RWKHU 7KH SK\VLFDO SURFHVV ZLWKLQ WKH VLPXODWHG HQYLURQPHQW ZDV D ODERUSDFHG SURFHVV %DUHILHOG f LH WKH RSHUDWLQJ HIILFLHQF\ RI WKH GHSDUWPHQW ZDV GHWHUPLQHG FRPSOHWHO\ E\ WKH ODERU HIILFLHQF\ RI WKH ZRUNHUV 7KH ODERU HIILFLHQF\ VWDQGDUG VWDWHG LQ WHUPV RI WLPH SHU XQLW DVVHPEOHGf ZDV EDVHG RQ HQJLQHHULQJ HVWLPDWHV WKDW DOORZHG IRU XQDYRLGDEOH ODERU LQHIILFLHQFLHV DQG UHDVRQDEOH YDULDWLRQ LQ ZRUNHU

PAGE 87

SHUIRUPDQFH LH WKH VWDQGDUGV ZHUH FXUUHQWO\ DWWDLQDEOHf 7KH VXEn MHFWV ZHUH LQVWUXFWHG WR DFFHSW WKH ODERU HIILFLHQF\ VWDQGDUG DV IDLU LQ WHUPV RI FRQWURO DQG SHUIRUPDQFH JRDOV 7KH SK\VLFDO ODERU SURFHVV RI WKH GHSDUWPHQW FRXOG EH LQ RQH RI WZR PXWXDOO\ H[FOXVLYH VWDWHV RI QDWXUH HLWKHU LQFRQWURO RU RXWRIn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n FHGXUHV ZKLFK ZHUH QRW RSHUDWLQJ DV H[SHFWHG 7YR DVVXPSWLRQV ZHUH SURn YLGHG WR DLG WKH VXEMHFWnV GHFLVLRQ PDNLQJ SURFHVV )LUVW LI WKH\ GHFLGHG WR LQYHVWLJDWH D YDULDQFH DQG WKH ODERU SURFHVV WXUQHG RXW WR EH RXWRIFRQWURO WKH SURFHVV ZRXOG EH UHWXUQHG WR WKH RULJLQDO LQFRQWURO VWDWH ZLWK FHUWDLQW\ 6HFRQG LI WKH\ GHFLGHG QRW WR LQn YHVWLJDWH D YDULDQFH DQG WKH ODERU SURFHVV ZDV RXWRIFRQWURO WKH SURFHVV ZRXOG UHPDLQ RXWRIFRQWURO ZLWK FHUWDLQW\ 9DULRXV FRVWV ZHUH DVVRFLDWHG ZLWK WKH YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ 7KHVH FRVWV GHSHQGHG XSRQ WKH VXEMHFWnV GHFLVLRQ HLWKHU LQn YHVWLJDWH RU GR QRW LQYHVWLJDWHf DQG XSRQ WKH DFWXDO VWDWH RI WKH ODERU SURFHVV HLWKHU LQFRQWURO RU RXWRIFRQWUROf (VWLPDWHV RI WKHVH

PAGE 88

FRVWV ZHUH SUHVHQWHG WR WKH VXEMHFW RQ WKH IDFH RI HDFK YDULDQFH UHSRUW XVLQJ WKH IROORZLQJ IRUPDW ,I
PAGE 89

GHFLVLRQ FRVWV LQFXUUHG E\ WKH VXEMHFW RYHU WKH VHULHV RI YDULDQFH UHn SRUWV DQG GLYLGLQJ WKLV VXP E\ WKH VXP RI WKH WRWDO LQYHVWLJDWLRQ GHn FLVLRQ FRVWV LQFXUUHG E\ DQ RSWLPDO PRGHO RYHU WKH VDPH VHULHV RI YDULn DQFH UHSRUWV 7KH VXEMHFWnV FDVK ERQXV IXQFWLRQ FRQVWDQW IRU DOO VXEn MHFWV ZDV LQYHUVHO\ UHODWHG WR WKLV PHDVXUH $V 7,'&PLQ DSSURDFKHG RQH WKH SD\RII DSSURDFKHG WKH PD[LPXP DV 7,'&PLQ EHFDPH ODUJHU WKDQ RQH WKH SD\RII DSSURDFKHG WKH PLQLPXP 6XEMHFWV 7KH VXEMHFWV ZHUH VHQLRU \HDU XQGHUJUDGXDWH DQG PDVWHUnV OHYHO JUDGXDWH VWXGHQWV HQUROOHG LQ WKH EXVLQHVV FROOHJH DW WKH 8QLYHUVLW\ RI )ORULGD 7KH VXEMHFWV SDUWLFLSDWHG LQ WKH H[SHULPHQW GXULQJ D WZR ZHHN SHULRG SDUWLFLSDWHG LQ WKH ILUVW ZHHN DQG SDUWLFLSDWHG LQ WKH VHFRQG ZHHN $ WRWDO RI VXEMHFWV LQLWLDOO\ YROXQWHHUHG WR SDUWLFLSDWH EXW VL[ VXEMHFWV IDLOHG WR FRPSOHWH WKH H[SHULPHQW 7KH VXEMHFWV ZKR FRPSOHWHG WKH H[SHULPHQW FRQVLVWHG RI PDOHV DQG IHPDOHV 7KUHH VXEMHFW VHOHFWLRQ FULWHULD ZHUH DSSOLHG f WKH VXEMHFW PXVW KDYH FRPSOHWHG DQ LQWHUPHGLDWHOHYHO PDQDJHULDO DFFRXQWLQJ FRXUVH f WKH VXEMHFW PXVW KDYH FRPSOHWHG DQ LQWURGXFWRU\OHYHO VWDWLVWLFV FRXUVH DQG f WKH VXEMHFW PXVW KDYH HDUQHG DQ RYHUDOO JUDGH SRLQW DYHUDJH *3$f RI DW OHDVW RQ D VFDOH 6XEMHFWV LQLWLDOO\ ZHUH FRQWDFWHG ZLWKLQ VHQLRU OHYHO DQG JUDGXDWH DFFRXQWLQJ FODVVHV 7KH FRQWDFW ZDV PDGH E\ WKH H[SHULPHQWHU JLYLQJ D EULHI RUDO SUHVHQWDWLRQ IROORZHG E\ SDVVLQJ VLJQXS VKHHWV DURXQG HDFK FODVV D FRS\ RI WKH RUDO SUHVHQWDWLRQ LV SUHVHQWHG LQ $SSHQGL[ %f 7R PRWLYDWH YROXQWHHULQJ WKH SUHVHQWDWLRQ IRFXVHG RQ WKH VWXGHQWnV SURn IHVVLRQDO UHVSRQVLELOLWLHV DV IXWXUH DFFRXQWDQWV DQG RQ WKH PRQHWDU\

PAGE 90

EHQHILWV WKDW ZRXOG DFFUXH WR WKRVH ZKR YROXQWHHUHG 7KH VXEMHFWV ZHUH WROG WKDW WKH\ ZRXOG EH JLYHQ D FDVK SD\PHQW IRU SDUWLFLSDWLQJ LQ WKH H[SHULPHQW DQG WKDW WKH DPRXQW RI D VXEMHFWnV SD\PHQW ZRXOG GHSHQG XSRQ KLV SHUIRUPDQFH 7KH PLQLPXP SD\PHQW ZDV VHW DW DQG WKH PD[LPXP DW ([SHULPHQWDO 0DWHULDOV 7KH H[SHULPHQWDO PDWHULDOV LQFOXGHG D EDFNJURXQG LQIRUPDWLRQ ERRNn OHW YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ VWLPXOL KHXULVWLFV TXHVWLRQQDLUH DQG PRWLYDWLRQV TXHVWLRQQDLUH %DFNJURXQG ,QIRUPDWLRQ $ EDFNJURXQG LQIRUPDWLRQ ERRNOHW VHH $SSHQGL[ &f ZDV GHVLJQHG WR SURYLGH WKH VXEMHFWV ZLWK D FRPPRQ H[SHULPHQWDO HQYLURQPHQW 7KH ERRNOHW SURYLGHG WKH VXEMHFW ZLWK JHQHUDO FRPSDQ\ LQIRUPDWLRQ JHQHUDO SURGXFW LQIRUPDWLRQ JHQHUDO PDQXIDFWXULQJ SURFHVV LQIRUPDWLRQ DQG VSHFLILF DVVHPEO\ GHSDUWPHQW LQIRUPDWLRQ 7KH VSHFLILF DVVHPEO\ GHSDUWn PHQW LQIRUPDWLRQ LQFOXGHG LQIRUPDWLRQ FRQFHUQLQJ WKH HPSOR\HHV WKH SK\VLFDO SURFHVV WKH DFFRXQWLQJ FRQWURO V\VWHP WKH VXEMHFWnV WDVN DV WKH RSHUDWLRQDO PDQDJHU DQG WKH VXEMHFWnV SHUIRUPDQFH HYDOXDWLRQ DV WKH RSHUDWLRQDO PDQDJHU 9DULDQFH ,QYHVWLJDWLRQ 'HFLVLRQV 9DULRXV LQIRUPDWLRQ FRQVWDQW RYHU DOO GHFLVLRQ WULDOV ZLWKLQ D WUHDWPHQW FRQGLWLRQ ZDV SUHVHQWHG RQ D VHSDUDWH SDJH SULRU WR WKH VWDUW

PAGE 91

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n VHQWDWLRQ RI D ODERU HIILFLHQF\ YDULDQFH UHSRUW DQG D VXEMHFWnV UHn VSRQVH WR WZR TXHVWLRQV 7KH TXHVWLRQV ZHUH f ZRXOG \RX LQYHVWLn JDWH WKLV UHSRUWHG YDULDQFH DQG f KRZ VWURQJO\ GR \RX IHHO DERXW \RXU GHFLVLRQ" 'XULQJ WKH WUDLQLQJ SKDVH HDFK GHFLVLRQ WULDO ZDV IROORZHG E\ IHHGEDFN FRQFHUQLQJ WKH DFWXDO VWDWH RI WKH DVVHPEO\ OLQH DQG WKH DFWXDO FRVWV LQFXUUHG IRU HDFK SRVVLEOH GHFLVLRQ JLYHQ WKH DFWXDO VWDWH 'HFLVLRQ WULDOV ZHUH SUHVHQWHG LQ ERRNOHWV RI WULDOV HDFK WULDO LQFOXGHG WKH UHSRUW ZLWK TXHVWLRQV IROORZHG E\ WKH IHHGEDFNf :LWKLQ WKH H[SHULPHQWDO SKDVH GHFLVLRQ WULDOV ZHUH SUHVHQWHG LQ ERRNOHWV RI WULDOV HDFK WULDO LQFOXGHG RQO\ WKH UHSRUW ZLWK TXHVWLRQVf ,Q ERWK WKH WUDLQLQJ DQG WKH H[SHULPHQWDO SKDVHV DQVZHU VKHHWV ZHUH SURn YLGHG IRU WKH VXEMHFW WR UHFRUG KLV UHVSRQVHV $Q H[DPSOH RI D ODERU HIILFLHQF\ YDULDQFH UHSRUW ZLWK WKH VHW RI TXHVWLRQV LV SUHVHQWHG LQ )LJXUH 7KH IRUPDW RI WKH UHSRUW DQG TXHVWLRQV ZDV FRQVWDQW IRU DOO WUHDWPHQW FRQGLWLRQV WKH RQO\ YDULDWLRQ UHODWHG WR WKH GLVWULEXWLRQ RI WKH DFWXDO PLQXWHV LQFXUUHG SHU FKDLU DQG WKH ODERU HIILFLHQF\ YDULDQFH DQG WR WKH PDJQLWXGH DQG VWUXFWXUH RI WKH FRVWV RI LQYHVWLJDWLRQ

PAGE 92

$06(&2 0HWDO )ROGLQJ &KDLU $VVHPEO\ 'HSDUWPHQW /DERU (IILFLHQF\ 9DULDQFH 5HSRUW )RU -RE 6WDQGDUG 0LQXWHV $OORZHG 3HU &KDLU $FWXDO 0LQXWHV ,QFXUUHG 3HU &KDLU /DERU (IILFLHQF\ 7RWDO &KDLUV 9DULDQFH 3HU &KDLU 3URGXFHG 7KH &RVWV $VVRFLDWHG :LWK ,QYHVWLJDWLRQ $UH ,I
PAGE 93

(OLFLWDWLRQ RI +HXULVWLFV 7KH HOLFLWDWLRQ RI HDFK VXEMHFWnV KHXULVWLFV ZDV DFFRPSOLVKHG XVLQJ D SUHGRPLQDWHO\ RSHQHQGHG TXHVWLRQQDLUH VHH $SSHQGL[ (f 7KH REMHFn WLYHV RI WKLV TXHVWLRQQDLUH ZHUH f WR GHWHUPLQH WKH VWUDWHJ\ WKH VXEn MHFW XVHG LQ DQVZHULQJ WKH YDULDQFH LQYHVWLJDWLRQ TXHVWLRQV ZLWKLQ WKH H[SHULPHQWDO SKDVH f WR GHWHUPLQH ZKHUH DSSURSULDWH WKH H[DFW YDOXHV RI DQ\ QXPHULFDO UXOHV XVHG E\ WKH VXEMHFW f WR GHWHUPLQH ZKHWKHU WKH VXEMHFW WKRXJKW KH XVHG KLV VWUDWHJ\ FRQVLVWHQWO\ ZLWKLQ WKH H[SHULPHQWDO SKDVH f WR GHWHUPLQH WKH PHWKRG RU DSSURDFK XVHG E\ WKH VXEMHFW LQ IRUPLQJ KLV VWUDWHJ\ ZLWKLQ WKH WUDLQLQJ SKDVH DQG f WR GHWHUPLQH WKH VXEMHFWLYH LPSRUWDQFH DWWDFKHG WR HDFK LQIRUPDWLRQ LWHP SUHVHQWHG WR WKH VXEMHFW (OLFLWDWLRQ RI 6XEMHFW 0RWLYDWLRQV 6XEMHFW PRWLYDWLRQV ZHUH HOLFLWHG XVLQJ D PRWLYDWLRQ TXHVWLRQQDLUH VHH $SSHQGL[ )f GHYHORSHG E\ 6QRZEDOO DQG %URZQ f 7KH TXHVWLRQn QDLUH LV D WHQ LWHP /LNHUWW\SH VFDOH ZKLFK KDV VXEPHDVXUHV IRU ERWK LQWULQVLF DQG H[WULQVLF PRWLYDWLRQ ([SHULPHQWDO 3URFHGXUHV ([SHULPHQWDO SURFHGXUHV LQFOXGHG DVVLJQPHQW RI VXEMHFWV WR WUHDWPHQW FRQGLWLRQV DGPLQVWUDWLRQ RI D WUDLQLQJ SKDVH DGPLQLVWUDWLRQ RI DQ H[SHULPHQWDO SKDVH DQG ILQDO GHEULHILQJ

PAGE 94

$VVLJQPHQW RI 6XEMHFWV WR 7UHDWPHQW &RQGLWLRQV 6LQFH HDFK RI WKH VXEMHFWV ZDV DVVLJQHG WR RQH RI HLJKW JURXSV UDQGRPL]DWLRQ SHU VH FDQ QRW EH UHOLHG XSRQ WR FRQWURO IRU LQGLYLGXDO DWWULEXWH GLIIHUHQFHV EHWZHHQ JURXSV $Q DOWHUQDWLYH LV WR EORFN WKH UDQGRPL]DWLRQ SURFHVV RQLQGL YLGXDO DWWULEXWH GLPHQVLRQV DVVXPHG WR VLJQLILFDQWO\ DIIHFW WKH VXEMHFWnV LQIRUPDWLRQ SURFHVVLQJ ZLWKLQ WKH WDVN UHTXLUHG E\ WKH H[SHULPHQW 7KH OLWHUDWXUH UHODWLQJ WR LQGLYLGXDO DWWULEXWHV KDV QRW SURn GXFHG FRQFOXVLYH UHVXOWV FRQFHUQLQJ LQGLYLGXDO DWWULEXWHV HIIHFW RQ GHFLVLRQ SURFHVVHV 7KH PRVW FRPSUHKHQVLYH VWXG\ LV WKDW RI 7D\ORU DQG 'XQQHWWH f ZKR DQDO\]HG WKH HIIHFW RI GHFLVLRQ PDNHU DWWULEXWHV XSRQ HLJKW PHDVXUHV RI SUHGHFLVLRQDO GHFLVLRQSRLQW DQG SRVW GHFLVLRQDO SURFHVVHV ZLWKLQ D SHUVRQQHO GHFLVLRQ VLPXODWLRQ H[SHULPHQW 2I FRQn VLGHUDEOH LQWHUHVW LV WKDW WKH UHOHYDQW GHFLVLRQ PDNHU DWWULEXWHV DFFRXQWHG IRU D JHQHUDOO\ VPDOO SRUWLRQ RI WKH WRWDO GHFLVLRQ YDULDQFH ZLWKLQ WKH YDULRXV SURFHVVHV WKH UDQJH ZDV IURP SHUFHQW WR SHUFHQWf 7KH GHFLVLRQ PDNHU DWWULEXWH ZLWK WKH JUHDWHVW SUHGLFWLYH FDSDFLW\ ZDV LQWHOOLJHQFH ,Q WZR RI WKH SUHGHFLVLRQDO SURFHVVHV GLDJRQRVWLFLW\ DQG LQIRUPDWLRQ SURFHVVLQJ UDWHf DQG LQ WZR RI WKH GHFLVLRQSRLQW DQG SRVW GHFLVLRQDO SURFHVVHV LQIRUPDWLRQ UHWHQWLRQ DQG GHFLVLRQ DFFXUDF\f LQWHOOLJHQFH DFFRXQWHG IRU IURP RQHKDOI WR DOPRVW DOO RI WKH YDULDQFH H[SODLQHG E\ LQGLYLGXDO FKDUDFWHULVWLFV ,Q WKH SUHVHQW VWXG\ WKH UDQGRPL]DWLRQ SURFHVV RI DVVLJQLQJ VXEMHFWV WR WUHDWPHQW FRQGLWLRQV ZDV EORFNHG RQ LQGLYLGXDO LQWHOOLJHQFH ,GHDOO\ LQGLYLGXDO LQWHOOLJHQFH VKRXOG EH PHDVXUHG XVLQJ VRPH YDOLGDWHG LQVWUXPHQW HJ WKH :HVPDQ 3HUVRQQHO &ODVVLILFDWLRQ 7HVW RU WKH

PAGE 95

:HFKVOHU $GXOW ,QWHOOLJHQFH 6FDOHf 'XH WR UHVRXUFH OLPLWDWLRQV KRZHYHU VXEMHFW JUDGH SRLQW DYHUDJH *3$f ZDV XVHG DV D VXUURJDWH IRU VXFK D PHDVXUH $ PHGLDQ *3$ ZDV LGHQWLILHG DQG WKRVH VXEMHFWV ZLWK D *3$ DERYH WKH PHGLDQ ZHUH FDWHJRUL]HG DV DERYH DYHUDJH LQWHOOLJHQFH DQG WKRVH VXEMHFWV ZLWK D *3$ EHORZ WKH PHGLDQ ZHUH FDWHJRUL]HG DV DYHUDJH LQWHOOLJHQFH (DFK VXEMHFW ZLWKLQ DQ LQWHOOLJHQFH FDWHJRU\ WKHQ ZDV DVVLJQHG UDQGRPO\ WR RQH RI WKH WUHDWPHQW FRQGLWLRQV ZLWK WKH UHVWULFn WLRQ WKDW HDFK LQWHOOLJHQFH JURXS FRQWULEXWHG DQ HTXDO QXPEHU RI VXEn MHFWV WR HDFK FRQGLWLRQ 8SRQ DVVLJQPHQW WR D WUHDWPHQW FRQGLWLRQ HDFK VXEMHFW UHFHLYHG WKH EDFNJURXQG LQIRUPDWLRQ ERRNOHW 7UDLQLQJ 3KDVH b (DFK VXEMHFW UHFHLYHG WUDLQLQJ ZLWKLQ WKH WUHDWPHQW FRQGLWLRQ WR A r ZKLFK KH ZDV DVVLJQHG 7UDLQLQJ ZDV FRQGXFWHG LQ JURXSV RI WZR VXEMHFWV ZLWKLQ D PLQXWH VHVVLRQ DGPLQLVWHUHG E\ HLWKHU WKH H[SHULPHQWHU RU E\ DQ H[SHULPHQWDO DVVLVWDQW 6XEMHFWV ZHUH DVVLJQHG UDQGRPO\ WR WKH H[SHULPHQWDO DVVLVWDQWV VXEMHFW WR WZR UHVWULFWLRQV 7KH UHVWULFWLRQV ZHUH WKDW WKH H[SHULPHQWDO DVVLVWDQW ZDV QRW FXUUHQWO\ WKH VXEMHFWnV WHDFKHU IRU DQ DFDGHPLF FRXUVH DQG WKDW WKH VXEMHFWnV DYDLODEOH WLPH FRLQFLGHG ZLWK WKDW RI WKH H[SHULPHQWDO DVVLVWDQW 7UDLQLQJ RI DOO 7KH DVVLJQPHQW SURFHVV LQYROYHG WZR SURFHGXUHV )LUVW WKH DVVLJQPHQW WR WKH FRVW YDULDEOH OHYHOV ZDV E\ WKH ZHHN LQ ZKLFK WKH VXEMHFW SDUn WLFLSDWHG LQ WKH H[SHULPHQW 7KRVH VXEMHFWV ZKR SDUWLFLSDWHG GXULQJ WKH ILUVW ZHHN ZHUH DVVLJQHG WR WKH &O FRVW OHYHO DQG WKRVH VXEMHFWV ZKR SDUWLFLSDWHG GXULQJ WKH VHFRQG ZHHN ZHUH DVVLJQHG WR WKH & FRVW OHYHO 6HFRQG WKH DVVLJQPHQW WR WKH LQIRUPDWLRQ YDULDEOH DQG WKH GLVWULEXWLRQ YDULDEOH OHYHOV ZDV E\ UDQGRP VHOHFWLRQ EDVHG XSRQ D UDQn GRP QXPEHU WDEOH 7KH H[SHULPHQWDO DVVLVWDQWV ZHUH QRW SDLG 7KH DXWKRU EHOLHYHV KRZn HYHU WKDW WKLV LV DQ H[FHSWLRQ WR WKH DSKRULVP WKDW SULFH UHIOHFWV YDOXH

PAGE 96

VXEMHFWV ZLWKLQ HDFK ZHHNf ZDV FRPSOHWHG RYHU WZR FRQWLJXRXV GD\V 7KH WUDLQLQJ SKDVH FRQVLVWHG RI GHFLVLRQ WULDOV ZLWK IHHGEDFN 7KH H[SHULPHQWDO PDWHULDOV XVHG LQ WKH WUDLQLQJ SKDVH ZHUH VLPLODU WR WKRVH XVHG LQ WKH H[SHULPHQWDO SKDVH 7KH GHFLVLRQ WULDOV ZLWK IHHGEDFN ZHUH SUHVHQWHG LQ WKUHH ERRNOHWV RI WULDOV DQG WKH VXEMHFW ZDV SURYLGHG DQ DQVZHU VKHHW RQ ZKLFK WR UHFRUG KLV UHVSRQVHV $GGLWLRQDO SHUIRUn PDQFH IHHGEDFN ZDV JLYHQ DW WKH FRPSOHWLRQ RI HDFK ERRN RI GHFLVLRQ WULDOV 7KLV SHUIRUPDQFH IHHGEDFN ZDV WKH VXEMHFWVnV 7,'&AS IRU WKDW SDUWLFXODU ERRN ([SHULPHQWDO 3KDVH 7KH H[SHULPHQWDO VHVVLRQ ODVWHG RQH KRXU DQG ZDV DGPLQLVWHUHG E\ WKH H[SHULPHQWHU 7KH H[SHULPHQWDO SKDVH ZLWKLQ HDFK ZHHNf ZDV FRPn SOHWHG RYHU WZR FRQWLJXRXV GD\V LPPHGLDWHO\ IROORZLQJ WKH WUDLQLQJ SKDVH 7KH H[SHULPHQWDO VHVVLRQ FRQVLVWHG RI WKUHH SDUWV 7KH ILUVW SDUW ZDV WKH SUHVHQWDWLRQ RI GHFLVLRQ WULDOV 7KH GHFLVLRQ WULDOV ZHUH SUHVHQWHG LQ WZR ERRNOHWV RI WULDOV HDFK DQG WKH VXEMHFW UHn FRUGHG KLV UHVSRQVHV RQ D VHSDUDWH DQVZHU VKHHW 7KH VXEMHFWV ZHUH DOORZHG WZHQW\ PLQXWHV WR FRPSOHWH WKH GHFLVLRQ WULDOV 7KH VHFRQG SDUW RI WKH H[SHULPHQWDO SKDVH ZDV WKH HOLFLWDWLRQ RI WKH VXEMHFWnV KHXULVWLFV 7KLV HOLFLWDWLRQ ZDV DFFRPSOLVKHG XVLQJ WKH RSHQHQGHG TXHVWLRQQDLUH VHH $SSHQGL[ (f 7KH WKLUG SDUW RI WKH H[SHULPHQWDO SKDVH ZDV WKH DGPLQLVWUDWLRQ RI WKH PRWLYDWLRQ TXHVWLRQQDLUH VHH $SSHQGL[ )f

PAGE 97

)LQDO 'HEULHILQJ (DFK VXEMHFWnV ILQDO SHUIRUPDQFH PHDVXUH 7,'&PLQf IRU WKH YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV SDUW RI WKH H[SHULPHQWDO SKDVH ZDV SUHVHQWHG LQGLYLGXDOO\ DW D ODWHU GDWH $W WKLV WLPH KLV FDVK SD\n PHQW ZDV GHWHUPLQHG KH ZDV GHEULHIHG DV WR WKH SXUSRVH RI WKH H[SHULn PHQW DQG DQ\ TXHVWLRQV ZHUH DQVZHUHG $GGLWLRQDO GDWD ZHUH FROOHFWHG IURP WKRVH VXEMHFWV QRW UHVSRQGLQJ FRPSOHWHO\ WR WKH KHXULVWLF HOLFLWDn WLRQ TXHVWLRQQDLUH

PAGE 98

&+$37(5 9 $1$/<6(6 $1' 5(68/76 6XPPDU\ RI 5HVXOWV $ VXPPDU\ RI WKH DQDO\VHV DQG UHVXOWV FRQWDLQHG LQ WKLV FKDSWHU LV SUHVHQWHG LQ 7DEOH 7KH REMHFWLYH RI WKLV VXPPDU\ LV WR SURYLGH D EULHI VWDWHPHQW RI HDFK K\SRWKHVLV WRJHWKHU ZLWK WKH UHVXOWV RI DQDO\n VHV IRU WKDW K\SRWKHVLV 6XFK D VXPPDU\ ZLOO IDFLOLWDWH WKH SUHVHQWDn WLRQ RI WKH DQDO\VHV DQG UHVXOWV DQG ZLOO VHUYH DV D UHIHUHQFH IRU WKH GLVFXVVLRQ DQG FRQFOXVLRQV FRQWDLQHG LQ WKH QH[W FKDSWHU *HQHUDO 0HWKRG RI $QDO\VLV $OWKRXJK VHYHUDO PHWKRGV RI DQDO\VLV DUH HPSOR\HG WKH PRVW SUHYDOHQW PHWKRG LV DQDO\VLV RI YDULDQFH XVLQJ WKH PRGHO FRPSDULVRQ SURFHGXUH $SSHOEDXP DQG &UDPHU /HZLV DQG .HUHQ f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

PAGE 99

7$%/( 6800$5< 2) $1$/<6(6 $1' 5(68/76 +\SRWKHVHV $QDO\VLV 0HWKRGVf 7HVW 6WDWLVWLFVf 5HVXOWV D 7'9L 6,f 7ZRWDLOHG = WHVW = QVf 7KH PHDQ LQLWLDO GHFLVLRQ DQFKRU RI WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXn WLRQ OHYHO ZDV QRW VLJQLILFDQWO\ GLIIHUHQW IURP E 7'9L_6f 7ZRWDLOHG = WHVW = Sf 7KH PHDQ LQLWLDO GHFLVLRQ DQFKRU RI WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXn WLRQ OHYHO ZDV VLJQLILFDQWO\ OHVV WKDQ D 7'9_,Of 7'9L f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW ) = QVf QVf 7KH PHDQ LQLWLDO GHFLVLRQ DQFKRU RI WKRVH VXEMHFWV ZLWKLQ WKH UHGXFHG LQIRUn PDWLRQ OHYHO ZDV QRW VLJQLILFDQWO\ GLIn IHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKn LQ WKH H[SDQGHG LQIRUPDWLRQ OHYHO E 7'9L_&Of 7'9L &f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW ) = Sf Sf 7KH PHDQ LQLWLDO GHFLVLRQ DQFKRU RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZDV VLJQLILFDQWO\ VPDOOHU WKDQ WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO D 5$_,f 5$_,f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW ) = QVf QVf 7KH PHDQ UHODWLYH GHFLVLRQ DQFKRU DGn MXVWPHQW RI WKRVH VXEMHFWV ZLWKLQ WKH UHGXFHG LQIRUPDWLRQ OHYHO ZDV QRW VLJn QLILFDQWO\ GLIIHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH H[SDQGHG LQIRUPDWLRQ OHYHO

PAGE 100

+\SRWKHVHV $QDO\VLV 0HWKRGVf E 0L_6Of 0L_6f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW F 0L_&Of 0L_&f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW D F7O c6OfG7 _6f 0RGHO FRPSDULVRQ 2QHWDLOHG = WHVW E ,Of F7W f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW F GW _&Of G _&f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW $%/( &RQWLQXHG 7HVW 6WDWLVWLFVf 5HVXOWV ) QVf 7KH PHDQ UHODWLYH GHFLVLRQ DQFKRU DGn = QVf MXVWPHQW RI WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO ZDV QRW VLJQLILn FDQWO\ GLIIHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO ) QVf 7KH PHDQ UHODWLYH GHFLVLRQ DQFKRU DGn = QVf MXVWPHQW RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZDV QRW VLJQLILFDQWO\ GLIn IHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKn LQ WKH & FRVW OHYHO ) Sf 7KH PHDQ GHFLVLRQ PRGHO VHQVLWLYLW\ RI WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXn = Sf WLRQ OHYHO ZDV VLJQLILFDQWO\ VPDOOHU WKDQ WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO ) QVf 7KH PHDQ GHFLVLRQ PRGHO VHQVLWLYLW\ RI = QVf WKRVH VXEMHFWV ZLWKLQ WKH UHGXFHG LQIRUn PDWLRQ OHYHO ZDV QRW VLJQLILFDQWO\ GLIn IHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKn LQ WKH H[SDQGHG LQIRUPDWLRQ OHYHO ) QVf 7KH PHDQ GHFLVLRQ PRGHO VHQVLWLYLW\ RI = QVf WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZDV QRW VLJQLILFDQWO\ GLIIHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO

PAGE 101

7$%/( f§ &RQWLQXHG +\SRWKHVHV $QDO\VLV 0HWKRGVf 7HVW 6WDWLVWLFVf 5HVXOWV D '1$L_OOf '1$L f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW ) QVf =222,2 QVf 7KH PHDQ DGMXVWHG UHODWLYH GHFLVLRQ PRGHO VHQVLWLYLW\ RI WKRVH VXEMHFWV ZLWKLQ WKH UHGXFHG LQIRUPDWLRQ OHYHO ZDV QRW VLJQLILFDQWO\ GLIIHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH H[SDQGHG LQIRUPDWLRQ OHYHO E '1$M _6,f '1$M ,6f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW ) QVf = QVf 7KH PHDQ DGMXVWHG UHODWLYH GHFLVLRQ PRGHO VHQVLWLYLW\ RI WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO ZDV QRW VLJQLILFDQWO\ GLIIHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXn WLRQ OHYHO F '1$M &Of '1$c _&f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW ) Sf = Sf 7KH PHDQ DGMXVWHG UHODWLYH GHFLVLRQ PRGHO VHQVLWLYLW\ RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZDV VLJQLILn FDQWO\ ODUJHU WKDQ WKDW RI WKRVH VXEn MHFWV ZLWKLQ WKH & FRVW OHYHO D '1$M'1M _6f '1$M '1c 6f 0RGHO FRPSDULVRQ 2QHWDLOHG = WHVW ) Sf = S f 7KH PHDQ PXOWLSOH GHFLVLRQ DQFKRU HIIHFW RQ UHODWLYH GHFLVLRQ PRGHO VHQVLWLYLW\ RI WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEX WLRQ ZDV VLJQLILFDQWO\ ODUJHU WKDQ WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVn WULEXWLRQ OHYHO

PAGE 102

+\SRWKHVHV $QDO\VLV 0HWKRGVf E '1$M'1M &Of '1$L'1L_&f 0RGHO FRPSDULVRQ 2QHWDLOHG = WHVW D %1$M ,Of %1$Qf 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW E %1$M 6Of %1$M 6f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW F %1$c &f!%1$&f 0RGHO FRPSDULVRQ 2QHWDLOHG = WHVW f§ &RQWLQXHG 7HVW 6WDWLVWLFVf 5HVXOWV ) QVf = QVf ) QVf = QVf ) QVf = QVf 7KH PHDQ PXOWLSOH GHFLVLRQ DQFKRU HIIHFW RQ UHODWLYH GHFLVLRQ PRGHO VHQVLWLYLW\ RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZDV QRW VLJQLILFDQWO\ GLIIHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO 7KH PHDQ DGMXVWHG UHODWLYH GHFLVLRQ FULWHULD RI WKRVH VXEMHFWV ZLWKLQ WKH UHGXFHG LQIRUPDWLRQ OHYHO ZDV QRW VLJn QLILFDQWO\ GLIIHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH H[SDQGHG LQIRUPDWLRQ OHYHO 7KH PHDQ DGMXVWHG UHODWLYH GHFLVLRQ FULWHULD RI WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO ZDV QRW VLJQLILFDQWO\ GLIIHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO ) Sf 7KH PHDQ DGMXVWHG UHODWLYH GHFLVLRQ FULWHULD RI WKRVH VXEMHFWV ZLWKLQ WKH &O = Sf FRVW OHYHO ZDV VLJQLILFDQWO\ ODUJHU WKDQ WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO

PAGE 103

7$%/( &RQWLQXHG $QDO\VLV 7HVW +\SRWKHVHV 0HWKRGVf 6WDWLVWLFVf 5HVXO WV D%1$L_&Of R%1$_&f ) WHVW RI HTXDO ) S .f 7KH DGMXVWHG UHODWLYH GHFLVLRQ FULWHULD YDULDQFHV YDULDQFH RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZDV VLJQLILFDQWO\ ODUJHU WKDQ WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO D %“&L_,Of %“&L_Of 0RGHO FRPSDULVRQ ) QVf 7KH PHDQ GHFLVLRQ FULWHULD FRQVHUYDWLVP RI WKRVH VXEMHFWV ZLWKLQ WKH UHGXFHG 7ZRWDLOHG = WHVW = QVf LQIRUPDWLRQ OHYHO ZDV QRW VLJQLILFDQWO\ GLIIHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH H[SDQGHG LQIRUPDWLRQ OHYHO E %1&L_6Of %1&_6f 0RGHO FRPSDULVRQ ) QVf 7KH PHDQ GHFLVLRQ FULWHULD FRQVHUYDWLVP RI WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEX 7ZRWDLOHG = WHVW = QVf WLRQ OHYHO ZDV QRW VLJQLILFDQWO\ GLIIHUn HQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO F %1&L_&Of%1&L_&f 0RGHO FRPSDULVRQ ) Sf 7KH PHDQ GHFLVLRQ FULWHULD FRQVHUYDWLVP RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW 2QHWDLOHG = WHVW = SFf OHYHO ZDV VLJQLILFDQWO\ ODUJHU WKDQ WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO D %1$L%1L_6,f 0RGHO FRPSDULVRQ ) QVf 7KH PHDQ PXOWLSOH GHFLVLRQ DQFKRU HIIHFW RQ UHODWLYH GHFLVLRQ FULWHULD RI WKRVH %1$L%1L_6f 2QHWDLOHG = WHVW = QVf VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO ZDV QRW VLJQLILFDQWO\ GLIIHUHQW IURP WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXn WLRQ OHYHO &2

PAGE 104

+\SRWKHVHV $QDO\VLV 0HWKRGVf E %1$M%1M _&Of 0RGHO FRPSDULVRQ %1$L%1_&f 2QHWDLOHG = WHVW D*L_&OfD*L_&f ) WHVW RI HTXDO YDULDQFHV D *L_&Of*L_&f 6RXUFHV ) YDOXH 2QHWDLOHG = WHVW E *M_6f*L_6,f 6RXUFHV ) YDOXH 2QHWDLOHG = WHVW F *L_,Of *L_,f 0RGHO FRPSDULVRQ 7ZRWDLOHG = WHVW f§ &RQWLQXHG 7HVW 6WDWLVWLFVf ) SF2Of = Sf ) Sf ) QVf = Sf ) QVf = Sf ) QVf = QVf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n MHFWV ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO 7KH PHDQ UHODWLYH QHW GHFLVLRQ FRVWV RI WKRVH VXEMHFWV ZLWKLQ WKH UHGXFHG LQIRUn PDWLRQ OHYHO ZDV QRW VLJQLILFDQWO\ GLIn IHUHQW WKDQ WKDW RI WKRVH VXEMHFWV ZLWKn LQ WKH H[SDQGHG LQIRUPDWLRQ OHYHO

PAGE 105

7$%/( f§ &RQWLQXHG +\SRWKHVHV $QDO\VLV 7HVW 0HWKRGVf 6WDWLVWLFVf >*A_6&f*c_6&f@ 0RGHO FRPSDULVRQ ) SF2Of U*_6O&f*_6&f@ 'XQFDQnV PXOWLSOH D f UDQJH WHVW D *3L*1L_&Of 6RXUFHV ) YDOXH ) QVf *3L*7UL_&f 2QHWDLOHG = WHVW = Sf E >*3*1M_&6f 0RGHO FRPSDULVRQ ) SF2Of *3L*1L_&O6f'XQFDQnV PXOWLSOH D f > *3 M*1 L & 6,f UDQJH WHVW *3M*1M &O 6,f@ *M_33f*M_00f 2QHWDLOHG W WHVW W O Sf 5HVXOWV 7KHUH ZDV D VLJQLILFDQW FRVW E\ GLVWULEXn WLRQ LQWHUDFWLRQ RQ WKH UHODWLYH QHW GHFLVLRQ FRVWV LQ ZKLFK WKH & FRVW OHYHO KDG D JUHDWHU HIIHFW WKDQ WKH &O FRVW OHYHO ZKLOH WKH GLVWULEXWLRQ OHYHOV KDG OLWWOH HIIHFW 7KH PHDQ UHODWLYH DGGLWLRQDO GHFLVLRQ FRVWV DQG UHODWLYH DGGLWLRQDO GHFLVLRQ VDYLQJV UDWLR RI WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO ZDV ODUJHU WKDQ WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO 7KHUH ZDV D VLJQLILFDQW FRVW E\ GLVWULEXn WLRQ LQWHUDFWLRQ RQ WKH UHODWLYH DGGLn WLRQDO GHFLVLRQ FRVWV DQG UHODWLYH DGGLn WLRQDO GHFLVLRQ VDYLQJV UDWLR LQ ZKLFK WKH FRVW OHYHOV KDG D VLJQLILFDQW HIIHFW RQO\ ZKHQ JLYHQ WKH 6, GLVWULEXWLRQ OHYHO 7KH PHDQ UHODWLYH QHW GHFLVLRQ FRVWV RI WKRVH VXEMHFWV ZLWKLQ WKH FRQVWDQWO\ GHFUHDVLQJ WUDLQLQJ VHVVLRQ SHUIRUPDQFHV FDWHJRU\ ZDV VLJQLILFDQWO\ VPDOOHU WKDQ WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH FRQn VWDQWO\ LQFUHDVLQJ WUDLQLQJ VHVVLRQ SHUn IRUPDQFHV FDWHJRU\

PAGE 106

7$%/( &RQWLQXHG +\SRWKHVHV *L_33f*L_03f *c _30f*c _00f $QDO\VLV 7HVW 0HWKRGVf 6WDWLVWLFVf 5HVXOWV 7KH PHDQ UHODWLYH QHW GHFLVLRQ FRVWV RI WKRVH VXEMHFWV ZLWKLQ WKH PL[HG WUDLQLQJ VHVVLRQ SHUIRUPDQFHV FDWHJRULHV ZHUH OHVV WKDQ WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH FRQVWDQWO\ LQFUHDVLQJ WUDLQLQJ VHVVLRQ SHUIRUPDQFHV FDWHJRU\ DQG JUHDWHU WKDQ WKDW RI WKRVH VXEMHFWV ZLWKLQ WKH FRQn VWDQWO\ GHFUHDVLQJ WUDLQLQJ VHVVLRQ SHUIRUPDQFHV FDWHJRU\ /2 FUW

PAGE 107

DOO HIIHFWVf DQG HOLPLQDWHV HIIHFWV VWDUWLQJ ZLWK WKH KLJKHVW RUGHU LQWHUDFWLRQV 7KH PRGHO FRPSDULVRQV XWLOL]H WKH JHQHUDO OLQHDU PRGHO WKHRU\ WKH PHWKRG RI OHDVW VTXDUHVf DV GHVFULEHG E\ 6FKHIIH f DQG WKH ) WHVW DV GHVFULEHG E\ /HZLV DQG .HUHQ f LV XVHG WR WHVW WKH ILW RI WKH YDULRXV PRGHOV 'HILQLQJ WZR PRGHOV DV IW DQG Z WKH ) WHVW LV GI IWf 66HRMf 66HIWf ) f GI DMIWf 66HIWf ZKHUH FR LV WKH PRUH UHVWULFWHG PRGHO DQG WKH GHJUHHV RI IUHHGRP DVVRFLn DWHG ZLWK WKH QXPHUDWRU DQG WKH GHQRPLQDWRU RI WKH ) VWDWLVWLF DUH GIFRIWf DQG GI IWf UHVSHFWLYHO\ *LYHQ D GHSHQGHQW YDULDEOH WKH PRGHO FRPSDULVRQ SURFHGXUH LV XVHG WR ILQG WKH VLPSOHVW PRGHO ZKLFK DGHTXDWHO\ ILWV WKH GDWD $IWHU WKH VLPSOHVW RU UHGXFHGf PRGHO LV IRXQG WKH VRXUFHV RI WKLV PRGHO DUH SUHVHQWHG ZLWK WKHLU FRUUHVSRQGLQJ ) YDOXHV 7KHVH VRXUFH ) YDOXHV DUH WKH VDPH DV PRGHO FRPSDULVRQ ) WHVWV ZKHUH WKH FRPSDULVRQ PRGHOV DUH WKH UHGXFHG PRGHO DQG WKH UHGXFHG PRGHO ZLWKRXW WKH FRUUHVSRQGLQJ VRXUFH :LWKLQ WKLV UHVHDUFK WKH VRXUFH ) YDOXHV DUH DOVR WKH VDPH DV WKH 7\SH ,9 VXP RI VTXDUHV %DUU HW DO f 8QOHVV RWKHUZLVH VSHFLILHG WKH D OHYHOV IRU WKH UHMHFWLRQ RI WKH QXOO K\SRWKHVLV DUH S IRU PDLQ HIIHFWV DQG S IRU LQWHUDFWLRQV 7KH VHOHFWLRQ RI WKHVH D YDOXHV LV ODUJHO\ DUELWUDU\ WKH YDOXHV UHIOHFW WKH SUHOLPLQDU\ QDWXUH RI WKLV UHVHDUFK $OVR XQOHVV RWKHUZLVH VSHFLn ILHG DOO ZD\ LQWHUDFWLRQV DUH WHVWHG DJDLQVW WKH ZLWKRXW ZD\ LQWHUDFWLRQ PRGHO HPSOR\LQJ WKH DVVXPSWLRQ WKDW WKH ZD\ LQWHUDFWLRQ HIIHFW LV HTXDO WR ]HUR ZLWKLQ WKH VXEMHFW SRSXODWLRQ $OO WKH LQGLYLGXDO DWWULEXWHV DUH UHWDLQHG LQ HDFK PRGHO HPSOR\LQJ WKH DVVXPSWLRQ WKDW

PAGE 108

WKHVH HIIHFWV DUH QRW QHFHVVDULO\ HTXDO WR ]HUR ZLWKLQ WKH VXEMHFW SRSXODWLRQ 7KH SDUDPHWHUV XVHG E\ WKH YDULRXV PRGHOV DV LQGHSHQGHQW YDULDEOHV DUH GHILQHG DV IROORZV 8 WKH RYHUDOO PHDQ RI WKH GHSHQGHQW YDULDEOH GM WKH HIIHFW RI OHYHO M RI WKH LQIRUPDWLRQ YDULDEOH A WKH HIIHFW RI OHYHO N RI WKH GLVWULEXWLRQ YDULDEOH
PAGE 109

RI QDWXUHf DV WKHLU LQLWLDO GHFLVLRQ DQFKRU 7KH VLPXODWLRQ DUELWUDULO\ HPSOR\HG WKH JHRPHWULF LQWHUVHFWLRQ RI WKH DSSURSULDWH VWDWH GLVWULEXn WLRQ FXUYHV DV WKH H[SHFWHG LQLWLDO GHFLVLRQ DQFKRU 7KHVH H[SHFWHG SRLQWV DUH FRQGLWLRQDO XSRQ WKH GLVWULEXWLRQ YDULDEOH LH WKH LQWHUVHFWLRQ SRLQW LV DFWXDO PLQXWHV LQFXUUHG JLYHQ WKH 6OL GLVWULEXn WLRQ OHYHO DQG LV DFWXDO PLQXWHV LQFXUUHG JLYHQ WKH 6 GLVWULEXWLRQ OHYHO 7KH FXWRII SRLQWV HPSOR\HG ZLWKLQ WKH ILUVW WUDLQLQJ VHVVLRQ 7'9Vf ZHUH XVHG DV DQ HVWLPDWH RI WKH VXEMHFWVn LQLWLDO GHFLVLRQ DQFKRUV 7KH PHWKRG RI DQDO\VLV ZDV WKH PRGHO FRPSDULVRQ SURFHGXUH ZKHUH WKH GHSHQGHQW YDULDEOH LV WKH 7'9A PHDVXUH DQG WKH LQGHSHQGHQW YDULDEOHV DUH WKH WKUHH VLWXDWLRQ YDULDEOHV LQIRUPDWLRQ GLVWULEXWLRQ DQG FRVWf ZLWK WKHLU LQWHUDFWLRQV DQG WKH *3$ YDULDEOH 7KH IXOO 7'9A PRGHO LV 7'9 9 2M HN < DMN &
PAGE 110

7$%/( 02'(/ &203$5,621 352&('85( 5(68/76 )25 7+( 7'9M '(3(1'(17 9$5,$%/( 0RGHO 66H GI)f ) )XO G D :LWKRXW ZD\ LQWHUDFWLRQ :LWKRXW GLVWULEXWLRQ [ FRVW LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ GLVWULEXn WLRQ LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ FRVW LQWHUDFWLRQ H :LWKRXW LQIRUPDWLRQ H :LWKRXW GLVWULEXWLRQ D H :LWKRXW FRVW & D S E SF &S G 7KH ) YDOXH DVVRFLDWHG ZLWK WKH IXOO PRGHO LV WKH WHVW RI WKH IXOO PRGHO DQG QRW D PRGHO FRPSDULVRQ WHVW H 7KH WHVW RI WKLV PRGHO DVVXPHV WKH HIIHFWV RI WKRVH LQWHUDFWLRQV WKDW LQFOXGH WKLV YDULDEOH DUH HTXDO WR ]HUR

PAGE 111

7$%/( 3$5$0(7(5 9$/8(6 $662&,$7(' :,7+ 7+( 5('8&(' 7'9M 02'(/ ) 9DOXHV $VVRFLDWHG :LWK WKH 6RXUFHV RI WKH 5HGXFHG 0RGHO 6RXUFH GI ) 'LVWULEXWLRQ &RVW E *3$ 0HDQV DQG 9DULDQFHV $VVRFLDWHG :LWK WKH 6LJQLILFDQW 6RXUFHV 9DULDEOH /HYHO 1 0HDQ 9DULDQFH 6, 6 &O & D S E S

PAGE 112

RI WKH GLVWULEXWLRQ YDULDEOH GLIIHU VLJQLILFDQWO\ = S WZRWDLOHGf ZLWK WKH 6 OHYHO KDYLQJ WKH ODUJHU PHDQ 'LIIHUHQFHV LQ WKH PHDQV RI WKH OHYHOV RI WKH FRVW YDULDEOH DUH DOVR VLJQLILFDQW = S WZRWDLOHGf ZLWK WKH & OHYHO KDYLQJ WKH ODUJHU PHDQ +\SRWKHVLV D SUHGLFWHG WKDW 7'9c_6f ZRXOG QRW GLIIHU VLJQLILn FDQWO\ IURP DFWXDO PLQXWHV LQFXUUHG $ VLQJOHVDPSOH = WHVW ZDV HPSOR\HG WR WHVW WKLV K\SRWKHVLV DQG WKH GLIIHUHQFH ZDV IRXQG WR EH QRW VLJQLILFDQW = QV WZRWDLOHGf +\SRWKHVLV E SUHGLFWHG WKDW 7:L _6f ZRXOG QRW GLIIHU VLJQLILFDQWO\ IURP DFWXDO PLQXWHV LQFXUUHG $ VLQJOHVDPSOH = WHVW LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV VLJQLILFDQW = S WZRWDLOHGf +\SRWKHVLV D SUHGLFWHG WKDW 7'9M ,,f ZRXOG QRW GLIIHU VLJQLILFDQWO\ IURP 7'9A@,f 7KH = WHVW XVHG WR WHVW WKLV K\SRWKHVLV LQGLFDWHG QR VLJQLILFDQW GLIIHUHQFH = QV WZRWDLOHGf +\SRWKHVLV E SUHGLFWHG WKDW 7'9A_&Of ZRXOG QRW GLIIHU VLJQLILFDQWO\ IURP 7'9c_&f 7KLV K\SRWKHVLV ZDV WHVWHG E\ WKH DERYH PRGHO FRPSDULVRQ SURFHGXUH DQG UHODWHG = WHVW DQG WKH UHVXOWV LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV VLJQLILFDQW Sf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

PAGE 113

DUH WKH WKUHH VLWXDWLRQ YDULDEOHV ZLWK WKHLU LQWHUDFWLRQVf DQG WKH *3$ YDULDEOH 7KH IXOO 5$M PRGHO LV 5$M LL D A < DH-N F
PAGE 114

7$%/( 02'(/ &203$5,621 352&('85( 5(68/76 )25 7+( 5$M '(3(1'(17 9$5,$%/( 0RGHO 66H GI)f ) D )XOO :LWKRXW ZD\ LQWHUDFWLRQ :LWKRXW GLVWULEXWLRQ [ FRVW LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ GLVWULEXn WLRQ LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ FRVW LQWHUDFWLRQ E :LWKRXW LQIRUPDWLRQ E :LWKRXW GLVWULEXWLRQ E :LWKRXW FRVW D 7KH ) YDOXH DVVRFLDWHG ZLWK WKH IXOO PRGHO LV WKH WHVW RI WKH IXOO PRGHO DQG QRW D PRGHO FRPSDULVRQ WHVW 7KH WHVW RI WKLV PRGHO DVVXPHV WKH HIIHFWV RI WKRVH LQWHUDFWLRQV WKDW LQFOXGH WKLV YDULDEOH DUH HTXDO WR ]HUR

PAGE 115

7$%/( 2%6(59(' 0($1 ('9M6 $1' 7'9M6 *,9(1 7+( 9$5,286 /(9(/6 2) 7+( ',675,%87,21 $1' &267 9$5,$%/(6 'LVWULEXWLRQ &O &RVW 9DULDEOH /HYHO ([SHFWHG ,QLWLDO & &RVW 9DULDEOH /HYHO 9DULDEOH ‹'9 79 'HFLVLRQ $QFKRU 7'9 ('9 6, 6 R HQ

PAGE 116

7KH H[SHFWHG SDWWHUQ JLYHQ WKH 6, GLVWULEXWLRQ OHYHO ZDV REWDLQHG +RZn HYHU WKH H[SHFWHG SDWWHUQ JLYHQ WKH 6 GLVWULEXWLRQ OHYHO ZDV QRW FRQn VLVWHQW ZLWK UHVSHFW WR WKH H[SHFWHG LQLWLDO GHFLVLRQ DQFKRU f ,QGLYLGXDO 'HFLVLRQ 0RGHO 6HQVLWLYLW\ 7KH IROORZLQJ VHFWLRQV UHSRUW WKH DQDO\VHV DQG UHVXOWV RI WKH LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ YDULDEOHV 7KH YDULDEOHV LQFOXGH WKH GM GHSHQGHQW YDULDEOH WKH UHODWLYH GHFLVLRQ PRGHO VHQVLWLYLW\ DQG WKH UHODWLRQVKLSV EHWZHHQ LQWULQVLF PRWLYDWLRQ DQG GHFLVLRQ PRGHO VHQVLWLYLW\ 7KH GM 'HSHQGHQW 9DULDEOH +\SRWKHVLV UHODWHV WR WKH ILW RI WKH VXEMHFWVn UHVSRQVHV WR WKH 76' PRGHO 7KH GHSHQGHQW YDULDEOH G VKRXOG KDYH OLWWOH UHODWLRQ ZLWK WKH LQIRUPDWLRQ DQG FRVW VLWXDWLRQ YDULDEOHV EXW VKRXOG KDYH D UHODWLRQ ZLWK WKH GLVWULEXWLRQ YDULDEOH 7KH UHODWLRQ ZLWK WKH GLVn WULEXWLRQ YDULDEOH LV GHULYHG IURP WKH RSWLPDO PRGHO GHFLVLRQ VHQVLn WLYLW\ PHDVXUH Gn :LWKLQ WKH 6, OHYHO RI WKH GLVWULEXWLRQ YDULDEOH WKH WKHRUHWLFDO YDOXH RI GA LV DQG ZLWKLQ WKH 6, OHYHO RI WKH GLVn WULEXWLRQ YDULDEOH WKH WKHRUHWLFDO YDOXH RI G LV N 7KH PHWKRG RI DQDO\VLV ZDV WKH PRGHO FRPSDULVRQ SURFHGXUH ZKHUH WKH GHSHQGHQW YDULDEOH LV WKH G PHDVXUH DQG WKH LQGHSHQGHQW YDULDEOHV DUH WKH WKUHH VLWXDWLRQ YDULDEOHV ZLWK WKHLU LQWHUDFWLRQVf WKH WKUHH PRWLYDWLRQ IDFWRUV DQG WKH *3$ YDULDEOH 7KH IXOO G PRGHO LV L L GL f fM %N < D_MN RW
PAGE 117

7KH PRGHO FRPSDULVRQ SURFHGXUH UHVXOWV IRU WKLV DQDO\VLV DUH SUHVHQWHG LQ 7DEOH DQG WKH ) YDOXHV DVVRFLDWHG ZLWK WKH VRXUFHV RI WKH UHGXFHG PRGHO DUH SUHVHQWHG LQ 7DEOH 7KH UHVXOWV LQGLFDWH WKDW WKH GM UHGXFHG PRGHO FRQWDLQV RQO\ WKH GLVWULEXWLRQ YDULDEOH DQG WKH LQWULQVLF PRWLYDWLRQ IDFWRU %RWK RI WKHVH SDUDPHWHUV DUH VLJQLILFDQW DW WKH S OHYHO 7KH PHDQV YDULDQFHV DQG VDPSOH VL]HV RI WKH GAn PHDVXUH JLYHQ WKH OHYHOV RI WKH GLVWULEXWLRQ YDULDEOH DUH LQFOXGHG LQ 7DEOH $SSO\LQJ WKH ) WHVW RI HTXDO YDULDQFHV WKH YDULDQFHV EHn WZHHQ WKH OHYHOV RI WKLV YDULDEOH GR QRW GLIIHU VLJQLILFDQWO\ ) QVf +RZHYHU WKH = WHVW RI HTXDO PHDQV LQGLFDWHV WKDW WKH PHDQV RI WKH OHYHOV RI WKLV YDULDEOH GLIIHU VLTQLILFDQWO\ = G WZRWDLOHGf +\SRWKHVLV D SUHGLFWHG WKDW G _6f ZRXOG EH VLTQLILFDQWO\ OHVV WKDQ G _6f 7KLV K\SRWKHVLV ZDV WHVWHG E\ WKH PRGHO FRPSDULVRQ SURFHGXUH DQG = WHVW SUHVHQWHG DERYH DQG WKH GLIIHUHQFH ZDV IRXQG WR EH VLTQLILFDQW Sf ZLWK WKH 6, OHYHO KDYLQJ WKH VPDOOHU PHDQ +\SRWKHVLV E SUHGLFWHG WKDW Gc @,f ZRXOG QRW GLIIHU VLJQLILFDQWO\ IURP G _,f 7KLV K\SRWKHVLV ZDV WHVWHG XVLQJ D = WHVW DQG WKH GLIIHUHQFH ZDV IRXQG WR EH QRW VLTQLILFDQW = QV WZRWDLOHGf +\SRWKHVLV F SUHGLFWHG WKDW G _&f ZRXOG QRW GLIIHU VLJQLILFDQWO\ IURP GO _&f $ = WHVW LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV QRW VLJQLILFDQW = QV WZRWDLOHGf 5HODWLYH 'HFLVLRQ 0RGHO 6HQVLWLYLW\ +\SRWKHVHV DQG UHODWH WR WKH DGMXVWHG UHODWLYH GHYLDWLRQ RI WKH LQGLYLGXDOnV GHFLVLRQ PRGHO VHQVLWLYLW\ IURP WKH RSWLPDO PRGHO VHQVLWLYLW\ 7KH DGMXVWPHQW LV FRQFHUQHG ZLWK WKH HOLPLQDWLRQ RI WKH

PAGE 118

7$%/( 02'(/ &203$5,621 352&('85( 5(68/76 )25 7+( Gn L '(3(1'(17 9$5,$%/( 0RGHO 66H GI)f ) )XO E D :LWKRXW ZD\ LQWHUDFWLRQ :LWKRXW GLVWULEXWLRQ [ FRVW LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ GLVWULEXn WLRQ LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ FRVW LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ :LWKRXW GLVWULEXWLRQ D :LWKRXW FRVW 9RL E7KH ) YDOXH DVVRFLDWHG ZLWK WKH IXOO PRGHO LV WKH WHVW RI WKH IXOO PRGHO DQG QRW D PRGHO FRPSDULVRQ WHVW F 7KH WHVW RI WKLV PRGHO DVVXPHV WKH HIIHFWV RI WKRVH LQWHUDFWLRQV WKDW LQFOXGH WKLV YDULDEOH DUH HTXDO WR ]HUR

PAGE 119

7$%/( 3$5$0(7(5 9$/8(6 $662&,$7(' :,7+ 7+( 5('8&(' G 02'(/ L ) 9DOXHV $VVRFLDWHG :LWK WKH 6RXUFHV RI WKH 5HGXFHG G 0RGHO 6RXUFH GI ) 'LVWULEXWLRQ D ,QWULQVLF PRWLYDWLRQ D ([WULQVLF QRQPRQHWDU\f PRWLYDWLRQ ([WULQVLF PRQHWDU\f PRWLYDWLRQ *3$ 0HDQV DQG 9DULDQFHV $VVRFLDWHG :LWK WKH 6LJQLILFDQW 6RXUFHV 9DULDEOH /HYHO 1 0HDQ 9DULDQFH 6, 6 92O

PAGE 120

QR HIIHFWV RI PXOWLSOH FXWRII YDOXHV RQ LQGLYLGXDO GHFLVLRQ PRGHO VHQVLn WLYLW\ $VVXPLQJ UDQGRP VXEMHFW DVVLJQPHQW WR H[SHULPHQWDO VLWXDWLRQV WKH 'ILM DQG '1$A YDULDEOHV VKRXOG EH LQGHSHQGHQW RI WKH HIIHFWV RI WKH Gn DEVROXWH PDJQLWXGH GLIIHUHQFHV ZLWKLQ WKH GLVWULEXWLRQ YDULDEOHf 7KH UHPDLQLQJ HIIHFWV DUH f WKH UHODWLYH YDULDEOH UHVSRQVH UDQJH ZLWKLQ WKH '1$A PHDVXUH DQG f WKH UHODWLYH YDULDEOH UHVSRQVH UDQJH DQG PXOWLSOH FXWRII YDOXH XVDJH ZLWKLQ WKH '1c PHDVXUH 7KH PHWKRGV RI DQDO\VLV ZHUH WKH PRGHO FRPSDULVRQ SURFHGXUHV ZKHUH WKH GHSHQGHQW YDULDEOHV DH WKH '1A '1$A DQG '1$c'1c PHDVXUHV 7KH LQGHSHQGHQW YDULDEOHV DUH WKH WKUHH VLWXDWLRQ YDULDEOHV ZLWK WKHLU LQWHUDFWLRQVf WKH WKUHH PRWLYDWLRQ IDFWRUV DQG WKH *3$c YDULDEOH 7KH WKUHH IXOO PRGHOV DUH '1L '1$L X fM N A fMN f7M D
PAGE 121

7$%/( 02'(/ &203$5,621 352&('85( 5(68/76 )25 7+( '1L '1$L $1' '1$L'0L '(3(1'(17 9$5,$%/(6 '1L '1$L '1$L'0L 0RGHO 66H G I)7a ) 66H G7787 ) 66H GI)f ) )XO G D E :LWKRXW ZD\ LQWHUDFWLRQ :LWKRXW GLVWULEXWLRQ [ FRVW LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ GLVn WULEXWLRQ LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ FRVW LQWHUDFWLRQ H :LWKRXW LQIRUPDWLRQ :LWKRXW GLVWULEXWLRQ D E H :LWKRXW FRVW E D 92O ES &S A7KH ) YDOXH DVVRFLDWHG ZLWK WKH IXOO PRGHO LV WKH WHVW RI WKH IXOO PRGHO DQG QRW D PRGHO FRPSDULVRQ WHVW H 7KH WHVW RI WKLV PRGHO DVVXPHV WKH HIIHFWV RI WKRVH LQWHUDFWLRQV WKDW LQFOXGH WKLV YDULDEOH DUH HTXDO WR ]HUR

PAGE 122

7$%/( 3$5$0(7(5 9$/8(6 $662&,$7(' :,7+ 7+( 5('8&(' '1M '1$L $1' '1$M'1M 02'(/6 6RXUFH ) 9DOXHV $VVRFLDWHG :LWK WKH 6RXUFHV RI WKH 5HGXFHF 0RGHOV '1L '1L '1$L '1$L GI ) GI ) GI ) 'LVWULEXWLRQ D E &RVW E D ,QWULQVLF PRWLYDWLRQ D & & ([WULQVLF QRQPRQHWDU\f PRWLYDWLRQ ([WULQVLF PRQHWDU\f PRWLYDWLRQ *3$ 0HDQV DQG 9DULDQFHV $VVRFLDWHG :LWK WKH 6LTQLILFDQW 9DULDEOHV '1L '1$L '1$L'1L 9DULDEOH /HYHO 1 0HDQ 9DULDQFH 1 0HDQ 9DULDQFH 1 0HDQ 9DULDQFH 6, 6 &O & DS ES &S

PAGE 123

VLJQLILFDQW DW WKH SF OHYHO DQG WKH PRWLYDWLRQ IDFWRU LV VLJQLILFDQW DW WKH S OHYHO 0HDQV YDULDQFHV DQG VDPSOH VL]HV RI WKH WKUHH GHSHQGHQW YDULDEOHV JLYHQ WKH OHYHOV RI WKH VLJQLILFDQW VLWXDWLRQ YDULDEOHV DUH LQFOXGHG LQ 7DEOH 8VLQJ WKH ) WHVW RI HTXDO YDULDQFHV WKH IROORZLQJ YDULDQFHV ZHUH IRXQG WR GLIIHU VLJQLILFDQWO\ f ZLWKLQ WKH '1c YDULDEOH EHWZHHQ WKH OHYHOV RI WKH GLVWULEXWLRQ YDULDEOH ) Sf f ZLWKLQ WKH '1$M YDULDEOH EHWZHHQ WKH OHYHOV RI WKH FRVW YDULDEOH ) SFf DQG f ZLWKLQ WKH '1$A'1A YDULDEOH EHWZHHQ WKH OHYHOV RI WKH GLVWULEXn WLRQ YDULDEOH ) SFf 7KH = WHVW RI HTXDO PHDQV LQGLFDWHG WKDW WKH PHDQV RI WKH YDULDEOH OHYHOV GLIIHU VLJQLILFDQWO\ SFf IRU HDFK RI WKH VLWXDWLRQ YDULDEOHV IRXQG WR EH VLJQLILFDQW E\ WKH DERYH VRXUFHV ) YDOXHV VHH 7DEOH f +\SRWKHVLV D SUHGLFWHG WKDW '1$_,f ZRXOG QRW GLIIHU VLJQLILn FDQWO\ IURP '1$A_,f $ = WHVW YDV HPSOR\HG WR WHVW WKLV K\SRWKHVLV DQG WKH GLIIHUHQFH ZDV IRXQG WR EH QRW VLJQLILFDQW = QV WZRWDLOHGf +\SRWKHVLV E SUHGLFWHG WKDW '1$AM6Of ZRXOG QRW GLIIHU VLJQLILFDQWO\ IURP '“$LM6f $JDLQ D = WHVW LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV QRW VLJQLILFDQW = QV WZRWDLOHGf +\SRWKHVLV F SUHGLFWHG WKDW '1$Ac&Of ZRXOG QRW GLIIHU VLJQLILFDQWO\ IURP '1$A_&f 7KH UHVXOWV RI WKH PRGHO FRPSDULVRQ SURFHGXUH DQG = WHVW LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV VLJQLILFDQW DW WKH SF OHYHO ZLWK WKH &O OHYHO KDYLQJ WKH JUHDWHU PHDQ +\SRWKHVLV D SUHGLFWHG WKDW '1$'1A M6Of ZRXOG EH VLJQLILFDQWO\ VPDOOHU WKDQ '1$A'1Ac6f $OWKRXJK WKH UHVXOWV RI WKH PRGHO FRPSDULVRQ SURFHGXUH DQG = WHVW LQGLFDWHG WKDW WKHUH ZDV D VLJQLILFDQW GLIIHUHQFH EHWZHHQ WKH PHDQV WKH 6 GLVWULEXWLRQ OHYHO KDG WKH VPDOOHU PHDQ

PAGE 124

+\SRWKHVLV E SUHGLFWHG WKDW '1$'1_&Of ZRXOG EH VLJQLILFDQWO\ VPDOOHU WKDQ '1$'1_&f $ = WHVW ZDV HPSOR\HG WR WHVW WKLV K\SRWKHVLV DQG WKH GLIIHUHQFH ZDV IRXQG WR EH QRW VLJQLILFDQW = QV RQHWDLOHGf ,QWULQVLF 0RWLYDWLRQ DQG 'HFLVLRQ 0RGHO 6HQVLWLYLW\ 7KH PRGHO FRPSDULVRQ SURFHGXUH LQGLFDWHG WKDW WKH LQWULQVLF PRWLn YDWLRQ IDFWRU KDG D VLJQLILFDQW HIIHFW RQ HDFK RI WKH IRXU LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ GHSHQGHQW YDULDEOHV 7KH DQDO\VLV RI WKHVH HIIHFWV UHTXLUHV DGGLWLRQDO WHVWV $ SURGXFWPRPHQW FRUUHODWLRQ ZDV FRPSXWHG IRU WKH LQWULQVLF PRWLYDWLRQ IDFWRU ZLWK HDFK RI WKH IRXU GHSHQGHQW YDULDEOHV 7KH UHVXOWV RI WKHVH FRUUHODWLRQV DUH SUHVHQWHG LQ 7DEOHr 7KH ZHDNQHVV RI WKHVH FRUUHODWLRQV FDQ EH H[SHFWHG JLYHQ WKH UHODWLYHO\ ORZ H[SODLQHG YDULDQFH DVVRFLDWHG ZLWK WKH YDULRXV LQn WULQVLF PRWLYDWLRQ PHDVXUH VXPV RI VTXDUHV 7KH SRVLWLYH DVVRFLDWLRQ EHWZHHQ LQWULQVLF PRWLYDWLRQ DQG WKH Gn L PHDVXUH LQGLFDWH WKDW DV WKH G PHDVXUH LQFUHDVHG LQ VL]H DSSURDFKHG L WKH G PHDVXUH RU XVHG D VLQJOH FXWRII YDOXHf WKH VXEMHFWVn LQWULQVLF N PRWLYDWLRQ PHDVXUHV LQFUHDVH 7KLV UHVXOW ZDV FRQILUPHG E\ WKH SRVLn WLYH DVVRFLDWLRQ EHWZHHQ WKH PRWLYDWLRQ IDFWRU DQG WKH '1 PHDVXUH $V HLWKHU WKH YDULDEOH UHVSRQVH UDQJH GHFUHDVHV RU WKH XVH RI PXOWLSOH FXWRII YDOXHV GHFUHDVHV ERWK DUH DVVRFLDWHG ZLWK LQFUHDVHV LQ WKH '1 PHDVXUHf WKH VXEMHFWVn LQWULQVLF PRWLYDWLRQ PHDVXUHV LQFUHDVH 7KH VXEVWDQWLDO GURS LQ WKH OHYHO RI VLJQLILFDQFH IRU WKH DVVRFLDWLRQ EHn WZHHQ WKH LQWULQVLF PRWLYDWLRQ DQG WKH '1$c PHDVXUHV ZRXOG VXJJHVW WKDW AHH WKH ODVW VHFWLRQ LQ WKLV FKDSWHU IRU D GHILQLWLRQ DQG IRU DGGLWLRQDO DQDO\VHV FRQFHUQLQJ WKH LQWULQVLF PRWLYDWLRQ IDFWRU

PAGE 125

7$%/( &255(/$7,21 &2()),&,(176 )25 ,175,16,& 027,9$7,21 :,7+ ,1',9,'8$/ '(&,6,21 02'(/ 6(16,7,9,7< 9$5,$%/(6 0RGHO ,QWULQVLF 0RWLYDWLRQ G L f '1M f '1$M f '1$M'1c f 1RWH 7KH QXPEHUV ZLWKLQ WKH SDUHQWKHVHV DUH WKH H[DFW DOHYHOV DVVRFLDWHG ZLWK WKH UHMHFWLRQ RI WKH QXOO K\SRWKHVLV WKDW WKH FRUUHODWLRQ FRHIILFLHQW LV HTXDO WR ]HUR

PAGE 126

WKH DVVRFLDWLRQ EHWZHHQ WKH PRWLYDWLRQ IDFWRU DQG LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ LV RQO\ SDUWLDOO\ D UHVXOW RI WKH YDULDEOH UHVSRQVH UDQJH WKH UHPDLQLQJ HIIHFW ZLWKLQ WKH '1$A PHDVXUHf 7KLV GURS LQ WKH VLJQLILFDQFH OHYHO FRXOG EH H[SODLQHG E\ WKH QHJDWLYH DVVRFLDWLRQ DOEHLW DQ HTXDOO\ ORZ VLJQLILFDQFH OHYHOf EHWZHHQ WKH PRWLYDWLRQ IDFWRU DQG WKH '1$A'1A PHDVXUH $V WKH XVH RI PXOWLSOH FXWRII YDOXHV GHFUHDVHV '1$A'1A GHFUHDVHVf WKH VXEMHFWVn LQWULQVLF PRWLYDWLRQ PHDVXUHV LQFUHDVH ,QGLYLGXDO 'HFLVLRQ &ULWHULD 7KH IROORZLQJ VHFWLRQV UHSRUW WKH DQDO\VHV DQG UHVXOWV RI WKH LQGLYLGXDO GHFLVLRQ FULWHULD YDULDEOHV 7KH YDULDEOHV LQFOXGH WKH UHODWLYH GHFLVLRQ FULWHULD WKH UHODWLYH GHFLVLRQ FULWHULD FRQVHUYDWLVP DQG WKH UHODWLRQVKLSV EHWZHHQ LQGLYLGXDO DWWULEXWHV DQG GHFLVLRQ FULWHULD 5HODWLYH 'HFLVLRQ &ULWHULD +\SRWKHVHV DQG UHODWH WR WKH DGMXVWHG UHODWLYH LQGLYLGXDO GHFLVLRQ FULWHULD 7KH DGMXVWPHQW FRQFHUQV WKH HOLPLQDWLRQ RI WKH HIIHFWV RI PXOWLSOH FXWRII YDOXHV RQ LQGLYLGXDO GHFLVLRQ FULWHULD 7KH YDULDEOHV ZKLFK DIIHFW WKH DGMXVWHG UHODWLYH GHFLVLRQ FULWHULD VKRXOG EH WKRVH UHODWHG WR WKH LQGLYLGXDOnV VHOHFWLRQ RI D FXWRII YDOXH 2I WKH WKUHH VLWXDWLRQ YDULDEOHV WKH RQH PRVW UHODWHG WR WKLV VHOHFWLRQ LV WKH FRVW YDULDEOH 7KH PHWKRG RI DQDO\VLV ZDV WKH PRGHO FRPSDULVRQ SURFHGXUH ZKHUH WKH GHSHQGHQW YDULDEOHV DUH WKH %1c %1$A DQG %1$A%1 PHDVXUHV 7KH LQGHSHQGHQW YDULDEOHV DUH WKH WKUHH VLWXDWLRQ YDULDEOHV ZLWK WKHLU

PAGE 127

LQWHUDFWLRQVf WKH WKUHH PRWLYDWLRQ IDFWRUV DQG WKH *3$A YDULDEOH 7KH WKUHH IXOO PRGHOV DUH S 2M %N
PAGE 128

7$%/( 02'(/ &203$5,621 352&('85( 5(68/76 )25 7+( %1M %1$ M $1' %1$M%1 M '(3(1'(17 9$5,$%/(6 %1 L %1$L %1$L%1L 0RGHO 66H G I )f ) 66H GI)f ) 66H GI)f ) )XO G D E :LWKRXW ZD\ LQWHUDFWLRQ & :LWKRXW GLVWULEXWLRQ [ FRVW LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ GLVn WULEXWLRQ LQWHUDFWLRQ & :LWKRXW LQIRUPDWLRQ [ FRVW LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ :LWKRXW GLVWULEXWLRQ & J :LWKRXW FRVW D D D S ES &S A7KH ) YDOXH DVVRFLDWHG ZLWK WKH IXOO PRGHO LV WKH WHVW RI WKH IXOO PRGHO DQG QRW D PRGHO FRPSDULVRQ WHVW H 7KH WHVW RI WKLV PRGHO DVVXPHV WKH HIIHFWV RI WKRVH LQWHUDFWLRQV WKDW LQFOXGH WKLV YDULDEOH DUH HTXDO WR ]HUR

PAGE 129

7$%/( 3$5$0(7(5 9$/8(6 $662&,$7(' :,7+ 7+( 5('8&(' %1c %1$c $1' %1$c%1c 02'(/6 ) 9DOXHV $VVRFLDWHG :LWK WKH 6RXUFHV RI WKH 5HGXFHG 0RGHOV 6RXUFH %1 L %1$L %1$ L%1L G I ) GI ) GI ) 'LVWULEXWLRQ & &RVW D D D ,QWULQVLF PRWLYDWLRQ ([WULQVLF QRQL QRQHWDU\f PRWLYDWLRQ ([WULQVLF PRQHWDU\f PRWLYDWLRQ D D *3$ F 0HDQV DQG 9DULDQFHV $VVRFLDWHG :LW L WKH 6LJQLILFDQW 9DULDEOHV %1 L %1$L %1$L%1L 9DULDEOH /HYHO 1 0HDQ 9DULDQFH 1 0HDQ 9DULDQFH 1 0HDQ 9DULDQFH 6, 6 &O & S S S

PAGE 130

S[f IRU HDFK RI WKH VLWXDWLRQ YDULDEOHV IRXQG WR EH VLJQLILFDQW E\ WKH DERYH VRXUFHV ) YDOXHV 7KH VROH H[FHSWLRQ ZDV WKH GLVWULEXWLRQ YDULDEOH ZLWKLQ WKH %1A PHDVXUH 7KH PHDQV RI WKH GLVWULEXWLRQ OHYHOV ZLWKLQ WKH %0 YDULDEOH GR QRW GLIIHU VLJQLILFDQWO\ = QV WZRWDLOHGf +\SRWKHVLV D SUHGLFWHG WKDW %1$MMOOf ZRXOG QRW GLIIHU VLJQLILn FDQWO\ IURP %1$f $ = WHVW LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV QRW VLJQLILFDQW = QV WZRWDLOHGf +\SRWKHVLV E SUHGLFWHG WKDW %1$_6,f ZRXOG QRW GLIIHU VLJQLILFDQWO\ IURP %1$A_6f $JDLQ D = WHVW VKRZHG WKH GLIIHUHQFH WR EH QRW VLJQLILFDQW = QV WZRWDLOHGf +\SRWKHVLV F SUHGLFWHG WKDW %1$_&Of ZRXOG EH VLJQLILn FDQWO\ ODUJHU WKDQ %1$-&f 7KH PRGHO FRPSDULVRQ SURFHGXUH DQG = WHVW LQGLFDWHG D VLJQLILFDQW GLIIHUHQFH DW WKH S OHYHO ZLWK WKH &O OHYHO KDYLQJ WKH ODUJHU PHDQ +\SRWKHVLV SUHGLFWHG WKDW WKH YDULDQFH RI %1$ c&f ZRXOG EH VLJQLILFDQWO\ VPDOOHU WKDQ WKH YDULDQFH RI %1$_&f 7KH ) WHVW RI HTXDO YDULDQFHV UHVXOWHG LQ D ILQGLQJ RI D VLJQLILFDQW GLIIHUHQFH DW WKH S OHYHO ZLWK WKH & OHYHO KDYLQJ WKH VPDOOHU YDULDQFH +\SRWKHVLV D SUHGLFWHG WKDW %1$A%1A_6f ZRXOG EH VLJQLILn FDQWO\ VPDOOHU WKDQ %1$f§%1M _6f $ = WHVW RI HTXDO PHDQV LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV QRW VLJQLILFDQW = QV RQHWDLOHGf KRZHYHU WKH REWDLQHG GLUHFWLRQ LV WKH VDPH DV WKH SUHGLFWHG GLUHFWLRQ RI WKLV HIIHFW +\SRWKHVLV E SUHGLFWHG WKDW %1$L%1Lc&f ZRXOG EH VLJQLILFDQWO\ VPDOOHU WKDQ %1$f§ %1 _&f 7KH PRGHO FRPSDULVRQ SURn FHGXUH DQG = WHVW LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV VLJQLILFDQW DW WKH S OHYHO ZLWK WKH & OHYHO KDYLQJ WKH VPDOOHU PHDQ

PAGE 131

5HODWLYH 'HFLVLRQ &ULWHULD &RQVHUYDWLVP +\SRWKHVLV UHODWHV WR WKH HIIHFW RI FRQVHUYDWLVP RQ WKH UHODWLYH LQGLYLGXDO GHFLVLRQ FULWHULD 7KH K\SRWKHVLV SURSRVHV WKDW WKH JUHDWHU WKH GLIIHUHQFH EHWZHHQ WKH LQGLYLGXDOnV LQLWLDO GHFLVLRQ DQFKRU DQG WKH RSWLPDO FXWRII YDOXH WKH KLJKHU WKH OHYHO RI FRQVHUn YDWLVP 7KH VLWXDWLRQ YDULDEOH PRVW UHODWHG WR VXFK D GLIIHUHQFH LV WKH FRVW YDULDEOH 7KH PHWKRG RI DQDO\VLV ZDV WKH PRGHO FRPSDULVRQ SURFHGXUH ZKHUH WKH GHSHQGHQW YDULDEOH LV WKH %1&c PHDVXUH DQG WKH LQGHSHQGHQW YDULn DEOHV DUH WKH WKUHH VLWXDWLRQ YDULDEOHV ZLWK WKHLU LQWHUDFWLRQVf WKH WKUHH PRWLYDWLRQ IDFWRUV DQG WKH *3$M YDULDEOH 7KH IXOO %1&M PRGHO LV %1&L S DM HN < DVMN DUM %
PAGE 132

7$%/( 02'(/ &203$5,621 352&('85( 5(68/76 )25 7+( %1&M '(3(1'(17 9$5,$%/( 0RGHO 66H GI)f ) E )XOO D :LWKRXW ZD\ LQWHUDFWLRQ :LWKRXW GLVWULEXWLRQ [ FRVW LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ GLVWULEXn WLRQ LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ FRVW LQWHUDFWLRQ F :LWKRXW LQIRUPDWLRQ F :LWKRXW GLVWULEXWLRQ F :LWKRXW FRVW D D S E 7KH ) YDOXH DVVRFLDWHG ZLWK WKH IXOO PRGHO LV WKH WHVW RI WKH IXOO PRGHO DQG QRW D PRGHO FRPSDULVRQ WHVW 7KH WHVW RI WKLV PRGHO DVVXPHV WKH HIIHFWV RI WKRVH LQWHUDFWLRQV WKDW LQFOXGH WKLV YDULDEOH DUH HTXDO WR ]HUR

PAGE 133

7$%/( 3$5$0(7(5 9$/8(6 $662&,$7(' :,7+ 7+( 5('8&(' %1&c 02'(/ ) 9DOXHV $VVRFLDWHG :LWK WKH 6RXUFHV RI WKH 5HGXFHG 0RGHO 6RXUFH GI ) &RVW E ,QWULQVLF PRWLYDWLRQ ([WULQVLF QRQPRQHWDU\f PRWLYDWLRQ ([WULQVLF PRQHWDU\f PRWLYDWLRQ D *3$ E 0HDQV DQG 9DULDQFHV $VVRFLDWHG :LWK WKH 6LJQLILFDQW 6RXUFHV 9DULDEOH /HYHO 1 0HDQ 9DULDQFH &O & 92O ES

PAGE 134

7KH = WHVW RI HTXDO PHDQV LQGLFDWHG WKDW WKH PHDQV RI WKH FRVW OHYHOV GLIIHU VLJQLILFDQWO\ = S WZRWDLOHGf ZLWK WKH & OHYHO KDYLQJ WKH VPDOOHU PHDQ +\SRWKHVLV D SUHGLFWHG WKDW %1&A_,f ZRXOG QRW GLIIHU VLJQLILn FDQWO\ IURP %1&c f $ = WHVW LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV QRW VLJQLILFDQW = QV WZRWDLOHGf +\SRWKHVLV E SUHGLFWHG WKDW %1&A 6,f ZRXOG QRW GLIIHU VLJQLILFDQWO\ IURP %1&M_6f $JDLQ D = WHVW LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV QRW VLJQLILFDQW = QV WZRWDLOHGf +\SRWKHVLV F SUHGLFWHG WKDW %1&c_&f ZRXOG EH VLJQLILn FDQWO\ VPDOOHU WKDQ %1&A_&f 7KH PRGHO FRPSDULVRQ SURFHGXUH DQG = WHVW LQGLFDWHG WKDW WKH GLIIHUHQFH ZDV VLJQLILFDQW DW WKH S OHYHO ZLWK WKH & OHYHO KDYLQJ WKH VPDOOHU PHDQ b ,QGLYLGXDO $WWULEXWHV DQG 'HFLVLRQ &ULWHULD f§ ‘ ‘ ‘ ‘ f§ ‘‘ ‘ ‘ ‘f§‘f§ f 7KH PRGHO FRPSDULVRQ SURFHGXUH UHVXOWV LQGLFDWHG WKDW WKH H[WULQVLF PRQHWDU\f PRWLYDWLRQ IDFWRU DQG WKH *3$c YDULDEOH KDG VLJQLILn FDQW HIIHFWV XSRQ WKH YDULRXV LQGLYLGXDO GHFLVLRQ FULWHULD YDULDEOHV 7KH DQDO\VLV RI WKHVH HIIHFWV DOVR UHTXLUHG DGGLWLRQDO WHVWV $ SURGXFW PRPHQW FRUUHODWLRQ ZDV FRPSXWHG IRU WKH PRWLYDWLRQ IDFWRU DQG WKH *3$c YDULDEOH ZLWK HDFK RI WKH GHFLVLRQ FULWHULD GHSHQGHQW YDULDEOHV 7KH UHVXOWV RI WKHVH FRUUHODWLRQV DUH SUHVHQWHG LQ 7DEOH 7KH QHJDWLYH DVVRFLDWLRQ EHWZHHQ H[WULQVLF PRQHWDU\f PRWLYDWLRQ DQG WKH %1A YDULDEOH LQGLFDWHV WKDW DV WKH %1A PHDVXUH LQFUHDVHV WKH VXEMHFWVn H[WULQVLF PRQHWDU\f PRWLYDWLRQ PHDVXUHV GHFUHDVH 7KH %1A 6HH WKH ODVW VHFWLRQV LQ WKLV FKDSWHU IRU GHILQLWLRQV DQG IRU DGGLWLRQDO DQDO\VHV FRQFHUQLQJ WKH H[WULQVLF PRQHWDU\f PRWLYDWLRQ IDFWRU DQG WKH *3$M YDULDEOH

PAGE 135

7$%/( &255(/$7,21 &2()),&,(176 )25 (;75,16,& 021(7$5
PAGE 136

PHDVXUH DSSURDFKHV LQILQLW\ DV WKH GLVWDQFH EHWZHHQ WKH IUM PHDVXUH DQG WKH %N PHDVXUH LQFUHDVHV $ VLPLODU DVVRFLDWLRQ H[LVWV EHWZHHQ WKH PRWLYDWLRQ IDFWRU DQG WKH %1$M PHDVXUH WKH %1$M PHDVXUH HOLPLQDWHV WKH HIIHFWV RI PXOWLSOH FXWRII YDOXHVf 7KH %1$A%1A YDULDEOH ZKLFK PHDVXUHV WKH HIIHFW RI PXOWLSOH FXWRII YDOXHV XSRQ LQGLYLGXDO GHFLVLRQ FULWHULD KDV QR VLJQLILFDQW DVVRFLDWLRQ ZLWK WKH PRWLYDWLRQ IDFWRU 7KH %1&YDULDEOH ZKLFK PHDVXUHV WKH HIIHFW RI FRQVHUYDWLVP XSRQ LQGLYLGXDO GHFLVLRQ FULWHULD KDV D VWURQJHU QHJDWLYH DVVRFLDWLRQ ZLWK WKH PRWLn YDWLRQ IDFWRU $V WKH OHYHO RI FRQVHUYDWLVP LQFUHDVHV %1&c LQFUHDVHVf VXEMHFWVn H[WULQVLF PRQHWDU\f PRWLYDWLRQ PHDVXUHV GHFUHDVH 7KH QHJDWLYH DVVRFLDWLRQ EHWZHHQ WKH *3$c YDULDEOH DQG WKH %1$%1A YDULDEOH LQGLFDWHV WKDW DV WKH %1$%1A YDULDEOH LQFUHDVHV WKH VXEMHFWVn *3$cV GHFUHDVH WKH %1$A%1A PHDVXUH LQFUHDVHV DV PXOWLSOH FXWRII YDOXH XVDJH LQFUHDVHVf 7KH QHJDWLYH DVVRFLDWLRQ EHWZHHQ WKH *3$c YDULDEOH DQG WKH %1&c YDULDEOH LQGLFDWHV WKDW DV WKH VXEMHFWVn *3$AV GHFUHDVH WKH OHYHO RI FRQVHUYDWLVP LQFUHDVHV ,QGLYLGXDO /RQJ5XQ 'HFLVLRQ (IILFLHQF\ 7KH IROORZLQJ VHFWLRQV UHSRUW WKH DQDO\VHV DQG UHVXOWV RI WKH LQGLYLGXDO ORQJUXQ GHFLVLRQ HIILFLHQF\ YDULDEOHV 7KH YDULDEOHV LQFOXGH WKH UHODWLYH GHFLVLRQ FRVWV WKH UHODWLRQVKLSV EHWZHHQ LQGLYLGXDO DWWULEXWHV DQG ORQJUXQ GHFLVLRQ HIILFLHQF\ DQG WKH UHODWLRQVKLSV EHn WZHHQ ORQJUXQ GHFLVLRQ HIILFLHQF\ DQG WUDLQLQJ SKDVH SHUIRUPDQFH 5HODWLYH 'HFLVLRQ &RVWV +\SRWKHVHV DQG UHODWH WR WKH GHFLVLRQ FRVWV LQFXUUHG E\ WKH LQGLYLGXDOV UHODWLYH WR WKH GHFLVLRQ FRVWV LQFXUUHG E\ WKH RSWLPDO

PAGE 137

PRGHOV %RWK WKH VWDWHG REMHFWLYH IXQFWLRQ DQG VXEMHFW SD\RII IXQFWLRQ LQYROYHG WKH PLQLPL]DWLRQ RI WKH GHFLVLRQ FRVWV $ VXEMHFWnV ORQJUXQ GHFLVLRQ HIILFLHQF\ ZDV PHDVXUHG LQ WHUPV RI KLV PLQLPL]DWLRQ RI KLV GHFLVLRQ FRVWV UHODWLYH WR WKRVH RI DQ RSWLPDO PRGHO 7KH PHWKRG RI DQDO\VLV ZDV WKH PRGHO FRPSDULVRQ SURFHGXUH ZKHUH WKH GHSHQGHQW YDULDEOHV DUH WKH *c *3c DQG *1c PHDVXUHV 7KH LQn GHSHQGHQW YDULDEOHV DUH WKH WKUHH VLWXDWLRQ YDULDEOHV ZLWK WKHLU LQWHUDFWLRQVf WKH '1c YDULDEOH WKH WKUHH PRWLYDWLRQ IDFWRUV DQG WKH *3$M YDULDEOH ,GHDOO\ ERWK WKH '1c DQG %1A YDULDEOHV VKRXOG EH LQFOXGHG LQ WKH PRGHO KRZHYHU WKH %1c YDULDEOH KDG VLJQLILFDQWO\ JUHDWHU DVVRFLDWLRQ ZLWK WKH RWKHU LQGHSHQGHQW YDULDEOHV WKDQ GLG WKH '1M YDULDEOH WKH 5 IRU WKH IXOO %1c PRGHO ZDV ZKHUHDV WKH 5 IRU WKH IXOO '1A PRGHO ZDV f ,QFOXVLRQ RI WKH %1c YDULDEOH ZRXOG SURGXFH D VLJQLILFDQW SUREOHP RI PXOWL FROLQHDULW\ 7KH WKUHH IXOO PRGHOV DUH *L *3L \ fM %N < DMN fIMO %
PAGE 138

7$%/( 02'(/ &203$5,621 352&('85( 5(68/76 )25 7+( *c *3c $1' *1c '(3(1'(17 9$5,$%/(6 *L *3L *1L 0RGHO 66H GI+ ) 66H fG7I7)7 ) 66H F7I;)7 ) F )XOO D D D :LWKRXW ZD\ LQWHUDFWLRQ :LWKRXW GLVWULEXWLRQ [ FRVW LQWHUDFWLRQ D E :LWKRXW LQIRUPDWLRQ [ GLVn WULEXWLRQ LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ FRVW LQWHUDFWLRQ G :LWKRXW LQIRUPDWLRQ G :LWKRXW GLVWULEXWLRQ G :LWKRXW FRVW D E S S F 7KH ) YDOXH DVVRFLDWHG ZLWK WKH IXOO PRGHO LV WKH WHVW RI WKH IXOO PRGHO DQG QRW D PRGHO FRPSDULVRQ WHVW G 7KH WHVW RI WKLV PRGHO DVVXPHV WKH HIIHFWV RI WKRVH LQWHUDFWLRQV WKDW LQFOXGH WKLV YDULDEOH DUH HTXDO WR ]HUR

PAGE 139

7$%/( 3$5$0(7(5 9$/8(6 $662&,$7(' :,7+ 7+( 5('8&(' *L *3L $1' *1L 02'(/6 ) 9DOXHV $VVRFLDWHG :LWK WKH 6RXUFHV RI WKH 5HGXFHG 0RGHOV 6RXUFH *L *3L *1L GI ) G I ) GI ) 'LVWULEXWLRQ &RVW 'LVWULEXWLRQ [ FRVW D E '1L D D ,QWULQVLF PRWLYDWLRQ & E ([WULQVLF QRQPRQHWDU\f PRWLYDWLRQ ([WULQVLF PRQHWDU\f PRWLYDWLRQ E D *3$ E D 0HDQV DQG 9DULDQFHV $VVRFLDWHG :LWK WKH 6LJQLILFDQW 9DULDEOHV *L *3 M 9DULDEOH /HYHO 1 0HDQ 9DULDQFH 1 0HDQ 9DULDQFH VL &O 6, & 6 &O 6 & DS 9 &S UR &2

PAGE 140

H[WULQVLF PRWLYDWLRQ IDFWRU DQG *3$A YDULDEOH DUH VLJQLILFDQW DW WKH S OHYHO 7KH '1 YDULDEOH LV VLJQLILFDQW DW WKH S OHYHO DQG WKH LQWULQVLF PRWLYDWLRQ IDFWRU LV VLJQLILFDQW DW WKH S OHYHO 7KH UHGXFHG *1c PRGHO FRQWDLQV WKH LQWULQVLF PRWLYDWLRQ IDFWRU WKH H[WULQVLF PRQHWDU\f PRWLYDWLRQ IDFWRU DQG WKH *3$ YDULDEOH %RWK WKH H[WULQVLF PRWLYDWLRQ IDFWRU DQG WKH *3$c YDULDEOH DUH VLJQLILFDQW DW WKH SF OHYHO DQG WKH LQWULQVLF PRWLYDWLRQ IDFWRU LV VLJQLILFDQW DW WKH S OHYHO 7KH PHDQV YDULDQFHV DQG VDPSOH VL]HV IRU WKH GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ ZLWKLQ WKH *A DQG *3A PRGHOV DUH LQFOXGHG LQ 7DEOH )LJXUH SUHVHQWV D JUDSKLF UHSUHVHQWDWLRQ RI ERWK GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQV %DUWOHWWnV WHVW IRU KRPRJHQHLW\ RI YDULDQFHV LQn GLFWHG WKDW WKH YDULDQFHV ZLWKLQ WKH *A PRGHO GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ DUH VLJQLILFDQWO\ KHWHURJHQHRXV [ O S ZLWK GIf 7KH YDULDQFHV ZLWKLQ WKH *3 PRGHO GLVWULEXWLRQ E\ FRVW LQWHUn DFWLRQ DUH FRQVLGHUHG KRPRJHQHRXV [ S ZLWK GIf 'XQFDQnV PXOWLSOH UDQJH WHVW RI HTXDO PHDQV LQGLFDWHG WKDW WKH PHDQV ZLWKLQ WKH *A PRGHO GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ IRUP WZR JURXSV WKH PHDQV ZLWKLQ D JURXS GR QRW GLIIHU VLJQLILFDQWO\ S f EXW D VLJQLILFDQW GLIIHUHQFH S f H[LVWV IRU WKH PHDQV EHWZHHQ JURXSV 7KHVH JURXSV DUH f *_6&f DQG *c _6&f DQG f _6&f *_6&f DQG *_6&f +\SRWKHVLV SUHGLFWHG WKDW WKH YDULDQFH RI *A_&f ZRXOG EH VLJQLILFDQWO\ VPDOOHU WKDQ WKH YDULDQFH RI *-&f 8VLQJ WKH ) WHVW RI HTXDO YDULDQFHV WKH YDULDQFHV GLIIHU VLJQLILFDQWO\ ) SF2Of ZLWK _&f KDYLQJ WKH VPDOOHU YDULDQFH +\SRWKHVLV SUHGLFWHG D VLJQLILFDQW GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ ZLWKLQ WKH PRGHO LQ

PAGE 141

),*85( ',675,%87,21 %< &267 9$5,$%/( ,17(5$&7,21 :,7+,1 $1' *3L 02'(/6

PAGE 142

6 6 'LVWULEXWLRQ 9DULDEOH *3L & &O K 6, 6 'LVWULEXWLRQ 9DULDEOH &RVW 9DULDEOH &RVW 9DULDEOH

PAGE 143

ZKLFK WKH *M _6 & f*Q_6&f ZRXOG EH VPDOOHU WKDQ WKH *A >6, &f *c_6&f 7KH UHVXOWV RI WKH PRGHO FRPSDULVRQ SURFHGXUH LQGLFDWHG D VLJQLILFDQW GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ SF2Of ZLWK WKH *L_6O&Of*L_6&Of HTXDO WR DQG WKH _6, fA _6&f HTXDO WR +\SRWKHVLV D SUHGLFWHG WKDW *A_&f ZRXOG EH VLJQLILFDQWO\ VPDOOHU WKDQ *M_&f $ = WHVW RI HTXDO PHDQV LQGLFDWHG WKDW WKH PHDQV GLIIHU VLJQLILFDQWO\ = S RQHWDLOHGf ZLWK WKH &O OHYHO KDYLQJ WKH VPDOOHU PHDQ 7KH ODFN RI VLJQLILFDQFH IRU WKH FRVW YDULDEOH PDLQ HIIHFW ZLWKLQ WKH PRGHO FRPSDULVRQ SURFHGXUH FRXOG EH WKH UHVXOW RI WKH GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ 7KHUH H[LVWV VXEVWDQWLDOO\ OHVV GLIIHUHQFH EHWZHHQ WKH OHYHOV RI WKH FRVW YDULDEOH ZKHQ JLYHQ 6 GLVWULEXWLRQ OHYHO WKDQ ZKHQ JLYHQ WKH 6, GLVn WULEXWLRQ OHYHO +\SRWKHVLV E SUHGLFWHG WKDW *M_6f ZRXOG EH VPDOOHU WKDQ *A_6f $ = WHVW RI HTXDO PHDQV LQGLFDWHG WKDW WKH PHDQV GLIIHU VLJQLILFDQWO\ = S RQHWDLOHGf ZLWK WKH 6 OHYHO KDYLQJ WKH VPDOOHU PHDQ 7KH ODFN RI VLJQLILFDQFH ZLWKLQ WKH PRGHO FRPSDULVRQ SURFHGXUH IRU WKH GLVWULEXWLRQ PDLQ HIIHFW ZDV SUHGLFWHG E\ K\SRWKHVLV E +\SRWKHVLV F SUHGLFWHG WKDW *A_,,f ZRXOG QRW GLIIHU VLJQLILFDQWO\ IURP *-,f 8VLQJ D = WHVW RI HTXDO PHDQV WKH GLIIHUHQFH ZDV IRXQG WR EH QRW VLJQLILFDQW = QV WZRWDLOHGf +\SRWKHVHV D DQG E UHODWH WR WKH UDWLR RI WKH LQGLYLGXDOnV UHODWLYH DGGLWLRQDO GHFLVLRQ FRVWV WR WKH LQGLYLGXDOnV UHODWLYH DGGLn WLRQDO GHFLVLRQ VDYLQJV 7KH PHWKRG RI DQDO\VLV ZDV WKH PRGHO FRPSDULn VRQ SURFHGXUH ZKHUH WKH GHSHQGHQW YDULDEOH LV WKH *3A*1A YDULDEOH DQG WKH LQGHSHQGHQW YDULDEOHV DUH WKH VDPH DV IRU WKH IXOO *A PRGHO 7KH IXOO *3A*1A PRGHO LV *9*1L f rM N
PAGE 144

7KH PRGHO FRPSDULVRQ SURFHGXUH UHVXOWV DUH SUHVHQWHG LQ 7DEOH DQG WKH ) YDOXHV DVVRFLDWHG ZLWK WKH VRXUFHV RI WKH UHGXFHG PRGHO DUH SUHVHQWHG LQ 7DEOH 7KH UHVXOWV LQGLFDWH WKDW WKH UHGXFHG *3A*0M PRGHO FRQWDLQV WKH GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ DQG WKH '0M YDULDEOH ERWK VLJQLILFDQW DW WKH S OHYHO 7KH PHDQV YDULDQFHV DQG VDPSOH VL]HV RI WKH GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ DUH LQFOXGHG LQ 7DEOH )LJXUH SUHVHQWV D JUDSKLF UHSUHVHQWDWLRQ RI WKH GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ %DUWOHWWnV WHVW RI KRPRJHQHLW\ RI YDULDQFHV LQGLFDWHG WKDW WKH YDULDQFHV ZLWKLQ WKH GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ DUH KRPRJHQHRXV [ S ZLWK GIf 'XQFDQnV PXOWLSOH UDQJH WHVW RI HTXDO PHDQV LQGLFDWHG WKDW WKH PHDQV ZLWKLQ WKH GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ IRUP WZR RYHUn ODSSLQJ JURXSV WKH PHDQV ZLWKLQ D JURXS GR QRW GLIIHU VLJQLILFDQWO\ S f ZKHUHDV WKH PHDQV EHWZHHQ JURXSV GR GLIIHU VLJQLILFDQWO\ S f 7KHVH JURXSV DUH f *3L*1L_&6f *3L*1L_&6f DQG *3L*1L_&6f DQG f *3L*1L_& 6f *3L*1L_&f DQG *3L*1L_&6f +\SRWKHVLV E SUHGLFWHG D VLJQLILFDQW GLVWULEXWLRQ E\ FRVW YDULDEOH LQWHUDFWLRQ LQ ZKLFK WKH *3A*1A _&6f*3L*1&O 6f ZRXOG EH VPDOOHU WKDQ *3L*1L_&6Of*3L*1L_& 6f 7KH UHVXOWV RI WKH PRGHO FRPSDULVRQ SURFHGXUH LQGLFDWHG D VLJQLILFDQW GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ Sf 7KH *3A*1A _&6f*3c*1c _& 6f ZDV HTXDO WR DQG WKH *3L*1L_&6f*3L*1L_&6f ZDV HTXDO WR +\SRWKHVLV D SUHGLFWHG WKDW *3A*1A_&Of ZRXOG EH VLJQLILFDQWO\ VPDOOHU WKDQ *3A*1M-&f 8VLQJ D = WHVW RI HTXDO PHDQV WKH GLIIHUHQFH EHWZHHQ WKH PHDQV ZDV IRXQG WR EH VLJQLILFDQW = S RQHWDLOHGf ZLWK WKH &O OHYHO KDYLQJ WKH VPDOOHU PHDQ 7KH ODFN RI

PAGE 145

7$%/( 02'(/ &203$5,621 352&('85( 5(68/76 )25 7+( *3L*1 '(3(1'(17 9$5,$%/( 0RGHO 66H GI)f ) )XO E D :LWKRXW ZD\ LQWHUDFWLRQ :LWKRXW GLVWULEXWLRQ [ FRVW LQWHUDFWLRQ D :LWKRXW LQIRUPDWLRQ [ GLVWULEXWLRQ LQWHUDFWLRQ :LWKRXW LQIRUPDWLRQ [ FRVW LQWHUDFWLRQ & :LWKRXW LQIRUPDWLRQ DS EIKH ) YDOXH DVVRFLDWHG ZLWK WKH IXOO PRGHO LV WKH WHVW RI WKH IXOO PRGHO DQG QRW D PRGHO FRPSDULVRQ WHVW F 7KH WHVW RI WKLV PRGHO DVVXPHV WKH HIIHFWV RI WKRVH LQWHUDFWLRQV WKDW LQFOXGH WKLV YDULDEOH DUH HTXDO WR ]HUR

PAGE 146

7$%/( 3$5$0(7(5 9$/8(6 $662&,$7(' :,7+ 7+( 5('8&(' *3L*1L 02'(/ ) 9DOXHV $VVRFLDWHG :LWK WKH 6RXUFHV RI WKH 5HGXFHG 0RGHO 6RXUFH GI ) 'LVWULEXWLRQ &RVW 'LVWULEXWLRQ D '1M D ,QWULQVLF PRWLYDWLRQ ([WULQVLF QRQPRQHWDU\f PRWLYDWLRQ ([WULQVLF PRQHWDU\f PRWLYDWLRQ *3$ 0HDQV DQG 9DULDQFHV $VVRFLDWHG :LWK WKH 6LTQLI LFDQW 6RXUFHV 9DULDEOH /HYHO 1 0HDQ 9DULDQFH 6, &O 6, & 6 &O 6 & DS

PAGE 147

c I 6, 6 'LVWULEXWLRQ 9DULDEOH ),*85( ',675,%87,21 %< &267 ,17(5$&7,21 :,7+,1 7+( &RVW 9DULDEOH *3L*1L 02'(/

PAGE 148

VLJQLILFDQFH IRU WKH FRVW YDULDEOH PDLQ HIIHFW ZLWKLQ WKH PRGHO FRPn SDULVRQ SURFHGXUH FRXOG EH WKH UHVXOW RI WKH GLVWULEXWLRQ E\ FRVW LQWHUDFWLRQ 7KH FRVW YDULDEOH KDV D VLJQLILFDQW HIIHFW RQO\ ZKHQ JLYHQ WKH 6, GLVWULEXWLRQ OHYHO ,QGLYLGXDO $WWULEXWHV DQG /RQJ5XQ 'HFLVLRQ (IILFLHQF\ 7KH PRGHO FRPSDULVRQ SURFHGXUH UHVXOWV LQGLFDWHG WKDW WKH '1 YDULDEOH WKH LQWULQVLF PRWLYDWLRQ IDFWRU WKH H[WULQVLF PRQHWDU\f PRWLYDWLRQ IDFWRU DQG WKH *3$ YDULDEOH KDG VLJQLILFDQW HIIHFWV XSRQ YDULRXV LQGLYLGXDO ORQJUXQ GHFLVLRQ HIILFLHQF\ PHDVXUHV 7KH DQDO\VLV RI WKHVH HIIHFWV UHTXLUHG DGGLWLRQDO WHVWV $ SURGXFWPRPHQW FRUUHODWLRQ ZDV FRPSXWHG IRU HDFK RI WKHVH LQGLYLGXDO DWWULEXWHV ZLWK HDFK RI WKH GHSHQGHQW YDULDEOHV 7KH UHVXOWV RI WKHVH FRUUHODWLRQV DUH SUHVHQWHG LQ 7DEOH 7KH QHJDWLYH DVVRFLDWLRQ EHWZHHQ WKH YDULDEOH DQG WKH '1 YDULDEOH LQGLFDWHV WKDW DV WKH UHODWLYH QHW GHFLVLRQ FRVWV LQFUHDVH WKH VXEMHFWVn YDULDEOH UHVSRQVH UDQJHV LQFUHDVH '1 LQFUHDVHVf 7KH VDPH DVVRFLDWLRQ H[LVWV EHWZHHQ WKH *3c YDULDEOH DQG WKH '1c YDULDEOH DV UHODWLYH DGGLWLRQDO GHFLVLRQ FRVWV LQFUHDVH WKH VXEMHFWVn YDULDEOH UHVSRQVH UDQJHV LQFUHDVH 7KH SRVLWLYH DVVRFLDWLRQ EHWZHHQ WKH *1A YDULDEOH DQG WKH '1c YDULDEOH LQGLFDWHV WKDW DV WKH UHODWLYH DGGLWLRQDO GHFLVLRQ VDYLQJV LQFUHDVH WKH VXEMHFWVn YDULDEOH UHVSRQVH UDQJHV GHFUHDVH '1M GHFUHDVHVf 7KH *3*1c YDULDEOH DQG WKH '1c YDULDEOH ZHUH DVVRFLDWHG QHJDWLYHO\ DV WKH UHODWLYH DGGLWLRQDO GHFLVLRQ FRVWV 6HH WKH ODVW VHFWLRQV LQ WKLV FKDSWHU IRU GHILQLWLRQV DQG DGGLWLRQDO DQDO\VHV FRQFHUQLQJ WKHVH LQGLYLGXDO DWWULEXWHV

PAGE 149

7$%/( &255(/$7,21 &2()),&,(176 )25 ,1',9,'8$/ $775,%87(6 :,7+ /21*581 '(&,6,21 ()),&,(1&< ,QGLYLGXDO $WWULEXWH *3 *1 *3 *1 '1 f f f f ,QWULQVLF PRWLYDWLRQ f f f f ([WULQVLF PRQHWDU\f PRWLYDWLRQ f f f f *3$ f f f f 1RWH 7KH QXPEHUV ZLWKLQ WKH SDUHQWKHVHV DUH WKH H[DFW FWOHYHOV DVVRFLDWHG ZLWK WKH UHMHFWLRQ RI WKH QXOO K\SRWKHVLV WKDW WKH FRUUHODWLRQ FRHIILFLHQW LV HTXDO WR ]HUR

PAGE 150

LQFUHDVH DW D UDWH JUHDWHU WKDQ WKH UHODWLYH DGGLWLRQDO GHFLVLRQ VDYLQJV WKH VXEMHFWVn YDULDEOH UHVSRQVH UDQJHV LQFUHDVH 7KH QHJDWLYH DVVRFLDWLRQ EHWZHHQ WKH *c YDULDEOH DQG WKH LQWULQVLF PRWLYDWLRQ IDFWRU LQGLFDWHV WKDW DV WKH UHODWLYH QHW GHFLVLRQ FRVWV LQFUHDVH WKH VXEMHFWVn LQWULQVLF PRWLYDWLRQ PHDVXUHV GHFUHDVH 7KH VDPH DVVRFLDWLRQ H[LVWV EHWZHHQ WKH *3c YDULDEOH DQG WKH LQWULQVLF PRWLYDWLRQ IDFWRU DV WKH UHODWLYH DGGLWLRQDO GHFLVLRQ FRVWV LQFUHDVH WKH VXEMHFWVn LQWULQVLF PRWLYDWLRQ PHDVXUHV GHFUHDVH 7KH *1c YDULDEOH DQG WKH LQn WULQVLF PRWLYDWLRQ IDFWRU ZHUH DVVRFLDWHG SRVLWLYHO\ DV WKH UHODWLYH DGGLWLRQDO GHFLVLRQ VDYLQJV LQFUHDVH WKH VXEMHFWVn LQWULQVLF PRWLYDWLRQ PHDVXUHV LQFUHDVH 6LPLODU DVVRFLDWLRQV H[LVW IRU WKH H[WULQVLF PRQHWDU\f PRWLYDWLRQ IDFWRU 7KXV DV WKH UHODWLYH DGGLWLRQDO GHFLVLRQ FRVWV LQFUHDVH *3A LQFUHDVHVf WKH VXEMHFWVn H[WULQVLF PRQHWDU\f PRWLYDWLRQ PHDVXUHV GHn FUHDVH &RQYHUVHO\ DV WKH UHODWLYH DGGLWLRQDO GHFLVLRQ VDYLQJV LQn FUHDVH *1M LQFUHDVHVf WKH VXEMHFWVn H[WULQVLF PRQHWDU\f PRWLYDWLRQ PHDVXUHV LQFUHDVH 7KH QHJDWLYH DVVRFLDWLRQ EHWZHHQ WKH *3A YDULDEOH DQG WKH *3$ YDULDEOH LQGLFDWHV WKDW DV VXEMHFWVn *3$V GHFUHDVH UHODWLYH DGGLWLRQDO GHFLVLRQ FRVWV LQFUHDVH /LNHZLVH WKH SRVLWLYH DVVRFLDWLRQ EHWZHHQ WKH *1M YDULDEOH DQG WKH *3$ YDULDEOH LQGLFDWHV WKDW DV VXEMHFWVn *3$V LQn FUHDVH UHODWLYH DGGLWLRQDO GHFLVLRQ VDYLQJV LQFUHDVH /RQJ5XQ 'HFLVLRQ (IILFLHQF\ DQG 7UDLQLQJ 3KDVH 3HUIRUPDQFH +\SRWKHVLV SUHGLFWHG WKDW WKRVH VXEMHFWV LQ WKH 33 FODVVLn ILFDWLRQ RI WUDLQLQJ DGMXVWPHQW GLUHFWLRQ ZRXOG KDYH D VLJQILFDQWO\ VPDOOHU PHDQ *c WKDQ WKRVH VXEMHFWV LQ WKH 00 FODVVLILFDWLRQ $ WWHVW

PAGE 151

ZDV XVHG WR WHVW WKLV K\SRWKHVLV 7KH */c_33f ZDV DQG WKH *M _00f ZDV 7KH ) WHVW IRU HTXDO YDULDQFHV LQGLFDWHG WKDW WKH YDULDQFHV ZHUH HTXDO ) QVf 7KH WWHVW IRU WKH K\SRWKHVLV RI HTXDO PHDQV LQGLFDWHG WKDW WKH K\SRWKHVLV FRXOG EH UHMHFWHG DW WKH S OHYHO W O ZLWK GIf +\SRWKHVLV SUHGLFWHG WKDW WKRVH VXEMHFWV LQ WKH PL[HG FODVVLILFDWLRQV 30 DQG 03f RI WUDLQLQJ DGMXVWPHQW GLUHFWLRQ ZRXOG KDYH PHDQ PHDVXUHV EHWZHHQ WKH QRQPL[HG FODVVLILFDWLRQV 7KH *c_30f ZDV DQG WKH *A03f ZDV 6LQFH WKH K\SRWKHVLV GLG QRW SUHGLFW VLJQLILFDQFH DQG EHFDXVH RI WKH UHVXOWV RI K\SRWKHVLV f WKHVH PHDQV ZHUH QRW WHVWHG VWDWLVWLFDOO\ 7KH GLUHFWLRQDO SUHGLFWLRQ RI K\SRWKHVLV ZDV VXSSRUWHG ,QGLYLGXDO $WWULEXWHV 7KH IROORZLQJ VHFWLRQV UHSRUW WKH DQDO\VHV DQG UHVXOWV RI WZR LQGLYLGXDO DWWULEXWHV VXEMHFW JUDGH SRLQW DYHUDJH DQG VXEMHFW PRWLn YDWLRQV 7KH REMHFWLYH RI ERWK DQDO\VHV ZDV WKH GHWHUPLQDWLRQ RI WKH VLJQLILFDQFH RI DQ\ EHWZHHQJURXS YDULDQFH RI WKHVH DWWULEXWHV 6XEMHFW *UDGH 3RLQW $YHUDJH *UDGH SRLQW DYHUDJHV *3$Vf RI VXEMHFWV ZHUH DQDO\]HG WR GHWHUPLQH ZKHWKHU WKLV LQGLYLGXDO DWWULEXWH YDULHG EHWZHHQ JURXSV RI VXEMHFWV $ V\VWHPDWLF YDULDQFH RI WKLV DWWULEXWH EHWZHHQ JURXSV FRXOG ELDV WKH LQWHUSUHWDWLRQ RI WKH HIIHFWV RI WKH LQGHSHQGHQW YDULDEOHV RQ SDUWLFXn ODU REVHUYDWLRQ YDULDEOHV 7ZR PHWKRGV RI DQDO\VLV ZHUH XVHG WR WHVW WKH DVVRFLDWLRQ RI *3$ ZLWK WKH LQGHSHQGHQW YDULDEOHV 'XQFDQnV PXOWLSOH UDQJH WHVW DQG %DUWOHWWnV WHVW RI KRPRJHQHLW\ RI YDULDQFH

PAGE 152

'XQFDQnV PXOWLSOH UDQJH WHVW ZDV XVHG WR GHWHUPLQH LI WKH VXEn MHFWVn PHDQ *3$V GLIIHUHG EHWZHHQ WKH HLJKW LQGHSHQGHQW YDULDEOHV FRQn GLWLRQV *LYHQ D FRPSOHWHO\ EHWZHHQVXEMHFW DVVLJQPHQW WKHVH HLJKW JURXSV UHSUHVHQWHG WKH UHVXOWV RI WKH H[SHULPHQWDO VDPSOLQJ SODQ 7KH UHVXOWV LQGLFDWH WKDW WKHUH ZHUH QRW VLJQLILFDQW GLIIHUHQFHV S f EHWZHHQ WKH PHDQ *3$V RI WKHVH HLJKW JURXSV %DUWOHWWnV WHVW RI KRPRJHQHLW\ RI YDULDQFH ZDV XVHG WR WHVW WKH YDULDQFHV RI WKH *3$V IRU WKHVH HLJKW JURXSV 7KH UHVXOWV LQGLFDWH WKDW WKH K\SRWKHVLV RI KRPRn JHQHLW\ RI YDULDQFH FRXOG QRW EH UHMHFWHG [ ZLWK GI Sf 2YHUDOO WKH UHVXOWV LQGLFDWH WKDW EHWZHHQJURXS YDULDQFH VKRXOG QRW EH DIIHFWHG VLJQLILFDQWO\ E\ GLIIHUHQFHV EHWZHHQ VXEMHFWVn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n UHVSRQVHV WR WKH WHQ PRWLYDWLRQ VFDOH LWHPV ZHUH IDFWRU DQDO\]HG XVLQJ D YDULPD[ URWDWLRQ IRU D IRXU IDFWRU VROXWLRQ 7KH IRXU IDFWRU VROXWLRQ ZDV FKRVHQ RQ WKH EDVLV RI SULRU EHOLHIV FRQFHUQLQJ WKH JHQHUDO GLPHQVLRQV RU VRXUFHVf RI VXEMHFW PRWLYDWLRQV LQWULQVLF PRWLYDWLRQ H[WULQVLF QRQPRQHWDU\ PRWLYDWLRQ H[WULQVLF PRQHWDU\ PRWLYDWLRQ DQG JHQHUDO PRWLYDWLRQ QRW

PAGE 153

FOHDUO\ DVVRFLDWHG ZLWK DQ\ RQH RI WKH SUHYLRXV VRXUFHV 7KH URWDWHG IDFWRU SDWWHUQ IRU WKH WHQ VFDOH LWHPV LV SUHVHQWHG LQ 7DEOH $QDO\]LQJ WKH HLJHQYDOXHV RI WKHVH IDFWRUV WKH VKDUSHVW EUHDN RU GHFOLQH LQ VORSHf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nV PXOWLSOH UDQJH WHVW DQG %DUWOHWWnV WHVW RI KRPRn JHQHLW\ RI YDULDQFH

PAGE 154

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

PAGE 155

'XQFDQnV PXOWLSOH UDQJH WHVW ZDV XVHG WR GHWHUPLQH LI WKH VXEMHFWVn PRWLYDWLRQV GLIIHUHG EHWZHHQ WKH HLJKW LQGHSHQGHQW YDULDEOHV FRQGLWLRQV 7KH UHVXOWV LQGLFDWH WKDW WKHUH ZHUH QR VLJQLILFDQW GLIIHUHQFHV S f LQ WKH PHDQ PRWLYDWLRQ IDFWRUV EHWZHHQ WKH HLJKW JURXSV %DUWOHWWnV WHVW RI KRPRJHQHLW\ RI YDULDQFH ZDV XVHG WR WHVW WKH YDULDQFHV RI WKH WKUHH PRWLYDWLRQ IDFWRUV IRU WKHVH HLJKW JURXSV 7KH UHVXOWV LQGLFDWH WKDW WKH K\SRWKHVLV RI KRPRJHQHLW\ RI YDULDQFH FRXOG QRW EH UHMHFWHG IRU WKH VHFRQG DQG WKLUG PRWLYDWLRQ IDFWRUV [ ZLWK GI S [ O ZLWK GI SFf 7KH K\SRWKHVLV RI KRPRJHQHLW\ RI YDULDQFH FRXOG EH UHMHFWHG DW WKH SF OHYHO IRU WKH ILUVW PRWLn YDWLRQ IDFWRU [ O ZLWK GIf 2YHUDOO WKH UHVXOWV LQGLFDWH WKDW EHWZHHQJURXS YDULDQFH VKRXOG QRW EH DIIHFWHG VLJQLILFDQWO\ E\ GLIIHUHQFHV EHWZHHQ WKH VXEMHFWVn PRWLYDWLRQV $JDLQ WKH DQDO\VLV GRHV QRW LQGLFDWH WKDW ZLWKLQJURXS YDULDQFH ZLOO EH XQDIIHFWHG E\ WKH GLIIHUHQFHV EHWZHHQ VXEMHFWVn PRWLYDWLRQV WKHVH SRVVLEOH HIIHFWV ZHUH SUHVHQWHG DERYH

PAGE 156

&+$37(5 9, ',6&866,21 2) 5(68/76 &RQFHSWXDO 'HYHORSPHQW 5HYLVLWHG $ EULHI UHVWDWHPHQW RI WKH FRQFHSWXDO GHYHORSPHQW SUHVHQWHG LQ &KDSWHU ,,, ZLOO VHUYH DV D EDVLV IRU GLVFXVVLRQ RI WKH UHVXOWV WKDW ZHUH UHSRUWHG LQ &KDSWHU 9 ,W ZDV SRVLWHG WKDW DQ LQGLYLGXDOnV ORQJUXQ GHFLVLRQ HIILFLHQF\ ZRXOG EH DIIHFWHG E\ f WKH VWUXFWXUH RI WKH SDUWLFXODU GHFLVLRQ VLWXDWLRQ f WKH FRQWHQWV RI WKH DYDLODEOH LQIRUPDWLRQ VHW f WKH LQGLYLGXDOnV HIILFLHQF\ LQ SURFHVVLQJ WKH DYDLODEOH LQIRUPDWLRQ DQG f WKH LQGLYLGXDOnV DELOLW\ WR H[SDQG WKH DYDLODEOH LQIRUPDWLRQ WKURXJK H[SHULHQFH ZLWK WKH SDUWLFXODU VLWXDWLRQ $ JHQHUDO KHXULVWLF ZDV H[SHFWHG WR V\VWHPDWLFDOO\ DFFRXQW IRU D SRUn WLRQ RI WKH GHFLVLRQ EHKDYLRU RI WKH VXEMHFWV $ KHXULVWLF ZLWKLQ WKLV FRQWH[W UHIHUV WR D OHDUQHG VHW RI UXOHV RU SULQFLSOHV ZKLFK DUH XWLOL]HG E\ LQGLYLGXDOV LQ PDNLQJ WKH SDUWLFXODU GHFLVLRQV UHTXLUHG RI WKHP 6LQFH D KHXULVWLF LV D OHDUQHG UHVSRQVH WR VRPH GHFLVLRQ WDVN VWLPXOL LQGLYLGXDO FKDUDFWHULVWLFV FDQ KDYH FRQVLGHUDEOH HIIHFW XSRQ WKH VSHFLILF IRUP RI WKH KHXULVWLF 7KH FRQFHSWXDO GHYHORSPHQW KRZn HYHU SURSRVHG WKDW VSHFLILF GHFLVLRQ WDVN VWLPXOL ZRXOG JLYH ULVH WR JHQHUDO IRUPV RI KHXULVWLFV ZKLFK ZRXOG EH DW OHDVW SDUWLDOO\f LQn GHSHQGHQW RI LQGLYLGXDO FKDUDFWHULVWLFV 7KH GHFLVLRQ WDVN XWLOL]HG LQ WKLV UHVHDUFK UHTXLUHG WKH GLVFULPn LQDWLRQ EHWZHHQ WKH XQFHUWDLQ H[LVWHQFH RI WZR VWDWHV RI QDWXUH

PAGE 157

LQFRQWURO DQG RXWRIFRQWUROf IRU D JLYHQ SURGXFWLRQ ODERU SURFHVV $ UDQGRP YDULDEOH SURGXFHG E\ WKH DFFRXQWDQW SURYLGHG HYLGHQFH FRQn FHUQLQJ VWDWH H[LVWDQFH DW SDUWLFXODU SRLQWV LQ WLPH 7KLV UDQGRP YDULDEOH WKH ODERU HIILFLHQF\ YDULDQFHf KDG FHUWDLQ VWDWLVWLFDO SURSHUWLHV ZLWKLQ WKH VWDWHV RI QDWXUH WKH GHFLVLRQ PDNHU HLWKHU KDG RU GLG QRW KDYH NQRZOHGJH RI WKHVH SDUWLFXODU SURSHUWLHV 7KH GHFLVLRQ WDVN ZDV FRPSOLFDWHG E\ WKH H[LVWHQFH RI XQHTXDO YDOXHV FRVWVf DVVRFLDWHG ZLWK WKH YDULRXV SRVVLEOH GHFLVLRQ RXWFRPHV 7KH IRUP RI WKH JHQHUDO KHXULVWLF SURSRVHG WR GHVFULEH GHFLVLRQ PDNLQJ EHKDYLRU ZDV WKDW RI DQFKRULQJ DQG DGMXVWPHQW $QFKRULQJ DQG DGMXVWPHQW ILUVW UHTXLUHV WKH VHOHFWLRQ RI DQ LQLWLDO GHFLVLRQ YDOXH DQFKRUf IURP WKH FRQWLQXXP RI SRVVLEOH GHFLVLRQ YDOXHV WKDW DUH ORFDWHG ZLWKLQ WKH GRPDLQ RI WKH UDQGRP YDULDEOH WKH ODERU HIILFLHQF\ YDULDQFH LQ WKH FDVH RI WKH SUHVHQW VWXG\f 2QFH DQ LQLWLDO GHFLVLRQ DQFKRU LV VHOHFWHG E\ WKH GHFLVLRQ PDNHU WUDLQLQJ DQG H[SHULHQFH OHDUQLQJf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nV SD\RII IXQFWLRQ

PAGE 158

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n LQLWLDO GHFLVLRQ DQFKRUV ZHUH ELDVHG WRZDUG WKH VWDQGDUG 7KH H[WHQW RI WKLV ELDV ZDV QRW ODUJH DQG DV WKH DGMXVWPHQW SURFHVV FRQWLQXHG WKH ELDV ZDV UHGXFHG VXEVWDQWLDOO\ UHODWLYH WR WKH DGMXVWPHQW ELDV RI VXEMHFWV ZLWKLQ WKH RSSRVLWH VLWXDWLRQV 6HFRQG YDULDEOH UHVSRQVH UDQJHV ZHUH ODUJHU ZLWKLQ VLWXDWLRQV ZKHUH WKH DGMXVWPHQW SURFHVV LQYROYHG GLYHUJHQFH IURP WKH VWDQGDUG DQG GHFLVLRQ FULWHULD FRQVHUYDWLVP ZDV ODUJHU ZLWKLQ VLWXDWLRQV ZKHUH WKH DGMXVWPHQW SURFHVV LQYROYHG FRQYHUJHQFH WRZDUG WKH VWDQGDUG $OWKRXJK WKHVH HIIHFWV ZHUH FRQWUDU\ WR WKRVH SUHGLFWHG WKH FRQFHSW RI D VXEn MHFWLYH DGMXVWPHQW OLPLW SURYLGHV D SODXVLEOH H[SODQDWLRQ 7KH FRQFHSW SURSRVHV WKDW VXEMHFWV ZLOO SHUFHLYH WKH VWDQGDUG DV D VXEMHFWLYH DGMXVWPHQW OLPLW DQG ZLOO EH KHVLWDQW WR DSSURDFK WKH VWDQGDUG 7KLV VXEMHFWLYH OLPLW ZLOO KDYH DQ HIIHFW ZKHUH WKH VLWXDWLRQ LQYROYHV DGMXVWPHQW WRZDUG WKH VWDQGDUG ,I WKLV SKHQRPHQRQ GRHV H[LVW LW ZRXOG DIIHFW ERWK WKH YDULDEOH UHVSRQVH UDQJH DQG WKH GHFLVLRQ FULWHULD FRQn VHUYDWLVP LQ WKH VDPH PDQQHU DV WKH REWDLQHG UHVXOWV

PAGE 159

7KLUG WKH REWDLQHG HIIHFWV RI WKH VLWXDWLRQ YDULDEOHV RQ WKH UHODWLYH GHFLVLRQ FRVWV ZHUH ZHDNHU WKDQ WKRVH SUHGLFWHG 7KH SUHn GLFWHG HIIHFWV ZHUH EDVHG LQ SDUW XSRQ WKH DVVXPSWLRQ RI HTXDO OHDUQLQJ HIILFLHQF\ EHWZHHQ WKH OHYHOV RI WKH YDULRXV VLWXDWLRQ YDULn DEOHV +RZHYHU DV LQGLFDWHG HDUOLHU WKH REWDLQHG OHDUQLQJ HIILFLHQn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f ODERU HIILFLHQF\ YDULDQFHV %RWK RI WKHVH UHVXOWV DUH FRQWUDU\ WR WKRVH SUHGLFWHG DQG QR H[SODQDWLRQV DUH RIIHUHG 2YHUDOO WKH YDULDEOH ZKLFK KDG WKH ODUJHVW LPSDFW RQ WKH UHODWLYH GHFLVLRQ FRVWV ZDV WKH VXEMHFWVn GHFLVLRQ FULWHULD FRQVHUYDWLVP 7KH PHDQ '1$A PHDVXUH RYHU DOO VXEMHFWV ZDV &RQVLGHULQJ WKDW D YDOXH RI RQH ZRXOG LQGLFDWH YDULDEOH UHVSRQVH UDQJHV ZHUH QRW XVHG WKH VXEMHFWVn RYHUDOO GHFLVLRQ PRGHO VHQVLWLYL\ ZDV DSSUR[LPDWHO\ WKDW RI WKH RSWLPDO PRGHO 7KH PHDQ %1&A RYHU DOO VXEMHFWV ZDV 7KLV LQGLFDWHG WKDW WKH DYHUDJH GLVWDQFH EHWZHHQ WKH VXEMHFWVn GHFLVLRQ FULWHULD DQG WKH RSWLPDO PRGHOnV GHFLVLRQ FULWHULD ZDV SHUFHQW RI WKDW RI WKH RSWLPDO PRGHO 7KLV ZDV VXSSRUWHG E\ WKH PHDQ 5$A RYHU

PAGE 160

DOO VXEMHFWV WKH RYHUDOO PHDQ 5$ ZDV ZKLFK LQGLFDWHG WKDW WKH VXEMHFWV DGMXVWHG WKHLU GHFLVLRQ DQFKRUV RQO\ SHUFHQW 5$Af RI WKH GLVWDQFH EHWZHHQ WKHLU LQLWLDO GHFLVLRQ DQFKRUV DQG WKH RSWLPDO PRGHOnV GHFLVLRQ YDOXH +RZHYHU HYHQ JLYHQ WKLV OHYHO RI RYHUDOO GHFLn VLRQ FULWHULD FRQVHUYDWLVP WKH PHDQ UHODWLYH GHFLVLRQ FRVWV *cf RYHU DOO VXEMHFWV ZDV WKH DYHUDJH VXEMHFWnV WRWDO GHFLVLRQ FRVWV ZHUH SHUFHQW JUHDWHU WKDQ WKH RSWLPDO PRGHOnV WRWDO GHFLVLRQ FRVWVf 7KH GHYLDWLRQ RI WKH PHDVXUH IURP D YDOXH RI ]HUR KDV EHHQ DWWULEXWHG WR WKH IROORZLQJ IDFWRUV f YDULDEOH UHVSRQVH UDQJHV f PXOWLSOH GHFLVLRQ DQFKRUV DQG f GHFLVLRQ FULWHULD FRQVHUYDWLVP $OWKRXJK HDFK IDFWRU ZDV DIIHFWHG GLIIHUHQWO\ E\ YDULRXV VLWXDWLRQ YDULDEOHV DQG LQGLYLGXDO DWWULEXWHV WKH RYHUDOO HIIHFWV RI WKHVH IDFWRUV RQ WKH UHODWLYH GHFLVLRQ FRVWV LQGLFDWH WKHLU UHODWLYH LPSRUWn DQFH 7KH RYHUDOO HIIHFW RI YDULDEOH UHVSRQVH UDQJHV ZDV VKRZQ WR EH QHJOLJLEOH ,JQRULQJ WKH YDULDEOH UHVSRQVH UDQJH HIIHFWV WKH RYHUDOO UHODWLYH GHFLVLRQ HIILFLHQF\ PHDVXUH *f FDQ EH DGMXVWHG LQWR VXEn PHDVXUHV ZKLFK DUH DIIHFWHG E\ RQO\ RQH RI WKH WZR UHPDLQLQJ IDFWRUV (OLPLQDWLQJ WKH HIIHFWV RI PXOWLSOH GHFLVLRQ DQFKRUV WKH DGMXVWHG *M PHDVXUH ZDV 7KLV LQGLFDWHG WKDW RI WKH WRWDO GHYLDWLRQ RI WKH VXEMHFWVn GHFLVLRQ FRVWV IURP WKRVH RI WKH RSWLPDO PRGHOV DSSUR[Ln PDWHO\ SHUFHQW ZDV GXH WR LQGLYLGXDO GHFLVLRQ FULWHULD FRQVHUYDWLVP DQG DSSUR[LPDWHO\ SHUFHQW ZDV GXH WR PXOWLSOH GHFLVLRQ DQFKRUV 'LVFXVVLRQ RI 5HVXOWV 7KH IROORZLQJ GLVFXVVLRQ RI UHVXOWV LV RUJDQL]HG DFFRUGLQJ WR GHFLVLRQ SURFHVVHV DQG GHSHQGHQW YDULDEOHV 7KH GLVFXVVLRQ LQFOXGHV GHFLVLRQ DQFKRU DGMXVWPHQW LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\

PAGE 161

LQGLYLGXDO GHFLVLRQ FULWHULD FRQVHUYDWLVP DQG LQGLYLGXDO ORQJUXQ GHFLVLRQ HIILFLHQF\ 'HFLVLRQ $QFKRU 6HOHFWLRQ DQG $GMXVWPHQW 7KH ILUVW DVSHFW RI WKH DQFKRULQJ DQG DGMXVWPHQW KHXULVWLF LV WKH VHOHFWLRQ RI WKH LQLWLDO GHFLVLRQ DQFKRU 7KH IRUPDWLRQ RI WKH K\SRn WKHVHV FRQFHUQHG ZLWK WKLV DVSHFW ZDV EDVHG XSRQ WKH DVVXPSWLRQV WKDW f LQGLYLGXDOV ZLWKLQ D JLYHQ OHYHO RI GLVWULEXWLRQ YDULDEOHf ZRXOG WHQG WR XVH D FRPPRQ LQLWLDO GHFLVLRQ DQFKRU DQG f WKH ORFDWLRQ RI WKLV FRPPRQ DQFKRU ZRXOG EH UHODWHG WR D PHDVXUH RI FHQWUDO WHQGHQF\ EHWZHHQ WKH WZR VWDWHV RI QDWXUH 7KLV FHQWUDO WHQGHQF\ ZDV RSHUDWLRQDOL]HG DV WKH JHRPHWULF LQWHUVHFWLRQ RI WKH VWDWH GLVWULEXWLRQ FXUYHV WKH LQWHUn VHFWLRQ ZDV FRQGLWLRQDO XSRQ WKH OHYHO RI WKH GLVWULEXWLRQ YDULDEOHf 7KH UHVXOWV LQGLFDWHG WKDW WKH H[SHFWHG SRLQW RI FHQWUDO WHQGHQF\ ZDV REWDLQHG IRU WKRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO EXW WKDW WKH H[SHFWHG SRLQW ZDV VWDWLVWLFDOO\ ODUJHU WKDQ WKDW REWDLQHG IRU WKRVH VXEMHFWV ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO 7KH VWDWH PHDQV ZHUH IDUWKHU DSDUW ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO WKDQ ZLWKLQ WKH 6, OHYHO $V D FRQVHTXHQFH WKH LQWHUVHFWLRQ RI WKH VWDWH GLVWULEXWLRQ FXUYHV ZDV ORFDWHG IDUWKHU IURP WKH VWDQGDUG WKH PHDQ RI WKH LQFRQWURO VWDWHf ZLWKLQ WKH 6 GLVWULEXWLRQ OHYHO WKDQ ZLWKLQ WKH 6, OHYHO $Q LQLWLDO VXEMHFWLYH ELDV SURYLGHV D SODXVLEOH H[SODQDWLRQ IRU WKH HVWLPDWH REWDLQHG ZLWKLQ WKH 6 OHYHO 7KLV ELDV ZRXOG WDNH WKH IRUP RI D WHQGHQF\ WR SODFH JUHDWHU ZHLJKW RQ WKH VWDQGDUG DV WKH LQWHUVHFWLRQ RI WKH GLVWULEXWLRQ FXUYHV EHFDPH ODUJHU LH VKLIWHG WR D KLJKHU SRLQW RQ WKH DFWXDO PLQXWHV LQFXUUHG D[LVf

PAGE 162

$Q HIIHFW ZKLFK GLIIHUHG IURP H[SHFWDWLRQV ZDV WKDW WKH PHDQ LQLWLDO GHFLVLRQ DQFKRU JLYHQ WKH & FRVW OHYHO ZDV VLJQLILFDQWO\ ODUJHU WKDQ WKH PHDQ LQLWLDO GHFLVLRQ DQFKRU JLYHQ WKH &O OHYHO HTXDOLW\ ZDV SUHGLFWHGf $ ELDV RI WKH HVWLPDWHV RI WKH LQLWLDO GHn FLVLRQ DQFKRUV PD\ H[SODLQ WKLV GLIIHUHQFH ,W ZLOO EH UHFDOOHG WKDW WKH HVWLPDWHV KDG LQFRUSRUDWHG VRPH LQLWLDO DGMXVWPHQW GXULQJ WKH ILUVW WUDLQLQJ VHVVLRQf 7KRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO ZHUH H[SHFWHG WR DGMXVW WKHLU LQLWLDO GHFLVLRQ DQFKRUV VXFK WKDW WKH YDOXH DFWXDO PLQXWHV LQFXUUHGf LQFUHDVHG WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZHUH H[SHFWHG WR DGMXVW LQ WKH RSSRVLWH GLUHFWLRQ ,QLWLDO DGn MXVWPHQW LQ WKH DSSURSULDWH GLUHFWLRQV ZRXOG SURGXFH WKH REWDLQHG FRVW YDULDEOH HIIHFW RQ WKH HVWLPDWHV RI WKH LQLWLDO GHFLVLRQ DQFKRUV 7KH VHFRQG DVSHFW RI WKH DQFKRULQJ DQG DGMXVWPHQW KHXULVWLF LV WKH DGMXVWPHQW SURFHVV 7KH DGMXVWPHQW SURFHVV ZDV H[DPLQHG XVLQJ D PHDVXUH RI WKH LQGLYLGXDOnV UHODWLYH GHFLVLRQ DQFKRU DGMXVWPHQW 7KLV REVHUYDWLRQ YDULDEOH 5$ ZDV WKH OLQHDU GLVWDQFH HYHU ZKLFK WKH LQGLYLGXDO DGMXVWHG KLV GHFLVLRQ DQFKRU WKH GLVWDQFH EHWZHHQ KLV ILQDO GHFLVLRQ DQFKRU DQG WKH RSWLPDO GHFLVLRQ YDOXHf UHODWLYH WR WKH OLQHDU GLVWDQFH RYHU ZKLFK WKH LQGLYLGXDO VKRXOG KDYH DGMXVWHG KLV GHFLVLRQ DQFKRU WKH GLVWDQFH EHWZHHQ KLV LQLWLDO GHFLVLRQ DQFKRU DQG WKH RSWLPDO GHFLVLRQ YDOXHf 7KH UHVXOWV RI DQDO\VLV LQGLFDWHG WKDW WKH 5$M PHDVXUH GLG QRW GLIIHU VLJQLILFDQWO\ EHWZHHQ WKH OHYHOV RI WKH YDULRXV VLWXDWLRQ YDULDEOHV 7KLV ZDV FRQVLVWHQW ZLWK WKH SUHGLFWLRQV IURP WKH WUDLQLQJ SKDVH VLPXODWLRQ $ PDMRU OLPLWDWLRQ RI WKH PHDVXUH LV WKDW LW LV FRQILQHG WR WKH UHODWLYH OLQHDU PRYHPHQW DORQJ WKH UDQGRP YDULDEOH GHFLVLRQ D[LV $V VXFK WKH 5$M PHDVXUH LV QRW VHQVLWLYH WR WKH OHDUQLQJ RI YDULRXV VLWXDWLRQ SDUDPHWHUV DQG YDULDEOHV HJ WKH

PAGE 163

UHODWLYH IUHTXHQFLHV RI WKH WZR VWDWHV DQG WKH UHODWLYH GHFLVLRQ HUURU FRVWVf QRU LV LW VHQVLWLYH WR WKH HIIHFWV RI YDULRXV LQGLYLGXDO DWWULn EXWHV HJ WKH LQGLYLGXDOnV YDULDEOH UHVSRQVH UDQJHf 3RVVLEOH HIIHFWV RI WKH VLWXDWLRQ YDULDEOHV RQ WKHVH DVSHFWV RI WKH DGMXVWPHQW SURFHVV DUH H[DPLQHG LQ JUHDWHU GHWDLO ZLWK PHDVXUHV RI JUHDWHU VHQVLWLYLW\f LQ ODWHU VHFWLRQV UHODWLQJ WR WKH LQGLYLGXDO GHFLVLRQ FULWHULD DQG WKH LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ ,QGLYLGXDO 'HFLVLRQ 0RGHO 6HQVLWLYLW\ ,QGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ ZDV PHDVXUHG HPSOR\LQJ D SDUDPHWHU RI WKH 76' PRGHO Gn 7KLV SDUDPHWHU PHDVXUHV WKH GHFLVLRQ VHQVLWLYLW\ RI WKH LQGLYLGXDOnV PRGHO UHODWLYH WR D WKHRUHWLFDO VHQVLn WLYLW\ 7KH IDFWRUV ZKLFK FDQ DIIHFW DQ LQGLYLGXDOnV GHFLVLRQ PRGHO VHQVLWLYLW\ DUH f WKH XVH RI D YDULDEOH UHVSRQVH UDQJH DQG f WKH XVH RI PXOWLSOH GHFLVLRQ DQFKRUV 7R LVRODWHnWKHVH IDFWRUV WKUHH GHFLVLRQ VHQVLWLYLW\ PHDVXUHV ZHUH GHYHORSHG f WKH '1M PHDVXUH ZKLFK LV DIIHFWHG E\ ERWK IDFWRUVf f WKH '1$M PHDVXUH ZKLFK LV DIIHFWHG E\ RQO\ WKH YDULDEOH UHVSRQVH UDQJHf DQG f WKH '1$M'1M PHDVXUH ZKLFK LV DIIHFWHG E\ RQO\ WKH PXOWLSOH GHFLVLRQ DQFKRUf 7KH YDULDEOH UHVSRQVH UDQJH WKH '1$M PHDVXUHf ZDV DIIHFWHG SULPDULO\ E\ WKH FRVW YDULDEOH 7KDW LV WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO GHPRQVWUDWHG ODUJHU YDULDEOH UHVSRQVH UDQJHV WKDQ GLG WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO 7KH ORFDWLRQV RI WKH RSWLPDO GHn FLVLRQ YDOXHV DQG WKH VXEMHFWVn ILQDO GHFLVLRQ DQFKRUV ZHUH FORVHU WR WKH VWDQGDUG IRU WKH &O FRVW OHYHO WKDQ IRU WKH & FRVW OHYHO VHH )LJXUH &KDSWHU ,,,f 7KH VXEMHFWLYH DGMXVWPHQW OLPLW FRQFHSW RIIHUHG

PAGE 164

HDUOLHU PD\ H[SODLQ WKH GLIIHUHQFH LQ HIIHFW RI WKH FRVW YDULDEOH RQ WKH '1$ PHDVXUH :KHQ WKH DGMXVWPHQW SURFHVV LQYROYHG FRQYHUJHQFH WRZDUG WKH VWDQGDUG WKH VXEMHFWV PD\ KDYH SHUFHLYHG WKH VWDQGDUG DV D OLPLW WR WKHLU DGMXVWPHQW SURFHVV D OLPLW ZKLFK WKH\ FRXOG KDYH EHHQ UHOXFWDQW WR DSSURDFK :KHQ WKH DGMXVWPHQW SURFHVV LQYROYHG GLYHUJHQFH IURP WKH VWDQGDUG WKH VXEMHFWV GLG QRW KDYH DQ REMHFWLYH YDOXH WR SHUn FHLYH DV D OLPLW WR WKHLU DGMXVWPHQW SURFHVV :KHWKHU D VXEMHFWnV DGMXVWPHQW SURFHVV LQYROYHG FRQYHUJHQFH WRZDUG RU GLYHUJHQFH IURP WKH VWDQGDUG GHSHQGHG XSRQ WKH ORFDWLRQ RI KLV LQLWLDO GHFLVLRQ DQFKRU UHODWLYH WR WKH RSWLPDO GHFLVLRQ YDOXH 7KLV IDFWRU ZDV FRQGLWLRQDO XSRQ WKH FRVW YDULDEOH *LYHQ WKH &O OHYHO WKH VXEMHFWnV LQLWLDO GHFLVLRQ DQFKRU ZDV JUHDWHU WKDQ WKH RSWLPDO GHFLVLRQ YDOXH DQG WKH DGMXVWPHQW SURFHVV LQYROYHG FRQYHUJHQFH WRZDUG WKH VWDQGDUG *LYHQ WKH & OHYHO WKH RSSRVLWH KHOG LH WKH VXEMHFWnV LQLWLDO GHFLVLRQ DQFKRU ZDV OHVV WKDQ WKH RSWLPDO GHFLVLRQ YDOXH DQG WKH DGMXVWPHQW SURFHVV LQYROYHG GLYHUJHQFH IURP WKH VWDQGDUG $V VXEMHFWVn DGMXVWPHQWV ZLWKLQ WKH &O OHYHO FRQYHUJHG WRZDUG WKH VWDQGDUG WKH VXEMHFWLYH OLPLW RI WKH VWDQGDUG FRXOG KDYH DFWHG DV DQ LQWHUYHQLQJ YDULDEOH ZKLFK UHGXFHG WKH UHODWLYH PDJQLWXGH RI WKH YDULDEOH UHVSRQVH UDQJH $V VXEMHFWVn DGMXVWPHQWV ZLWKLQ WKH & FRVW OHYHO GLYHUJHG IURP WKH VWDQn GDUG QR VXFK VXEMHFWLYH DGMXVWPHQW OLPLW H[LVWHG WKHUHIRUH WKH UHODWLYH PDJQLWXGH RI WKH YDULDEOH UHVSRQVH UDQJH FRXOG KDYH LQFUHDVHG 7KH KLJKHU VWDWLVWLFDOf PRPHQWV RI WKH '1$A PHDVXUH JLYHQ WKH OHYHOV RI WKH FRVW YDULDEOH ZHUH FRQVLVWHQW ZLWK WKLV FRQFHSW RI D VXEMHFWLYH DGMXVWPHQW OLPLW 6XFK D OLPLW VKRXOG KDYH KDG WKH HIIHFW RI UHGXFLQJ WKH YDULDQFH RI WKLV PHDVXUH 7KH WHVW RI WKH YDULDQFHV LQGLFDWHG WKDW '1$A-&Of KDG D VLJQLILFDQWO\ VPDOOHU YDULDQFH WKDQ

PAGE 165

'1$M_&f 7KH VXEMHFWLYH OLPLW DOVR VKRXOG KDYH KDG WKH HIIHFW RI VNHZLQJ WKH PHDVXUH DZD\ IURP WKRVH YDOXHV ZKLFK LQGLFDWHG ODUJHU YDULDEOH UHVSRQVH UDQJHV 7KH WKLUG PRPHQW DV H[SUHVVHG E\ WKH FRn HIILFLHQW RI VNHZQHVVf RI WKH '1$ PHDVXUH LQGLFDWHG WKDW f WKH VNHZQHVV RI WKH '1$A_&f GLVWULEXWLRQ ZDV SRVLWLYH VNHZHG DZD\ IURP YDOXHV ZKLFK LQGLFDWHG ODUJHU YDULDEOH UHVSRQVH UDQJHVf DQG f WKH VNHZQHVV RI WKH '1$_&f GLVWULEXWLRQ ZDV QHJDWLYH VNHZHG WRZDUG YDOXHV ZKLFK LQGLFDWHG ODUJHU YDULDEOH UHVSRQVH UDQJHVf 7KH PXOWLSOH GHFLVLRQ DQFKRU YDULDEOH WKH '1$A'1A PHDVXUHf ZDV DIIHFWHG SULPDULO\ E\ WKH GLVWULEXWLRQ YDULDEOH 7KRVH VXEMHFWV ZLWKLQ WKH 6, GLVWULEXWLRQ OHYHO HPSOR\HG PXOWLSOH GHFLVLRQ DQFKRUV WR D ODUJHU H[WHQW WKDQ GLG WKRVH VXEMHFWV ZLWKLQ WKH 6 OHYHO 7KH K\SRn WKHVHV SRVLWHG D UHODWLRQ EHWZHHQ PXOWLSOH GHFLVLRQ DQFKRUV DQG WKH b GLVWDQFH VHSDUDWLQJ WKH ORZHU WDLO RI WKH LQFRQWURO VWDWH IURPAWKH VXEMHFWnV ILQDO GHFLVLRQ DQFKRU ,W ZDV H[SHFWHG WKDW WKH ODUJHU WKLV GLVWDQFH IRXQG ZLWKLQ WKH & FRVW OHYHO DQG WKH 6 GLVWULEXWLRQ OHYHOf WKH JUHDWHU ZRXOG KDYH EHHQ WKH VXEMHFWnV SURSHQVLW\ WR KDYH SHUFHLYHG D VHFRQG RXWRIFRQWURO VWDWH EHORZ WKH H[LVWLQJ LQFRQWURO VWDWH DQG XVHG PXOWLSOH GHFLVLRQ DQFKRUVf +RZHYHU WKH REWDLQHG UHVXOWV GR QRW VXSSRUW WKLV SURSRVHG UHODWLRQ 7KH REVHUYHG DVVRFLDWLRQV EHWZHHQ WKH LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ PHDVXUHV DQG LQWULQVLF PRWLYDWLRQ KDYH LQWXLWLYH LQWHUn SUHWDWLRQV 7KH LQWULQVLF PRWLYDWLRQ IDFWRU PHDVXUHG H[SHULPHQWDO WDVN DQG H[SHULPHQWDO HQYLURQPHQW HQMR\PHQW DV PRGLILHG E\ GHFLVLRQ SHUIRUPDQFH VDWLVIDFWLRQ $V WKH VXEMHFWVf XVH RI PXOWLSOH GHFLVLRQ DQFKRUV DQG YDULDEOH UHVSRQVH UDQJHV LQFUHDVHG WKHLU LQWULQVLF PRWLn YDWLRQ DSSHDUHG WR GHFUHDVH 6LQFH WKH XVH RI PXOWLSOH GHFLVLRQ

PAGE 166

DQFKRUV DGYHUVHO\ DIIHFWV GHFLVLRQ SHUIRUPDQFH DV UHSRUWHG E\ WKH IHHGn EDFN PHDVXUH XVHG LQ WKLV UHVHDUFKf WKHLU LQFUHDVHG XVH ZRXOG DIIHFW LQWULQVLF PRWLYDWLRQ WKURXJK WKH SHUIRUPDQFH VDWLVIDFWLRQ PRGLILHU *UHDWHU XVH RI PXOWLSOH GHFLVLRQ DQFKRUV ZRXOG UHVXOW LQ SRRUHU GHFLVLRQ SHUIRUPDQFH IHHGEDFN PHDVXUHV WKHVH PHDVXUHV LQ WXUQ FRXOG ORZHU WKH VXEMHFWnV GHFLVLRQ SHUIRUPDQFH VDWLVIDFWLRQ DQG WKHUHE\ ORZHU KLV LQWULQVLF PRWLYDWLRQ $V WKH YDULDEOH UHVSRQVH UDQJH LQFUHDVHV LQ UHODWLYH VL]H WKH QXPEHU RI GHFLVLRQV IDOOLQJ ZLWKLQ WKLV UDQJH LQFUHDVHV 7KLV UHTXLUHV PRUH JXHVVZRUN RU PRUH FRPSOLFDWHG GHFLVLRQ UXOHV RQ WKH SDUW RI WKH VXEMHFWV WKHUHE\ LQFUHDVLQJ WKH GLIILFXOW\ RU WKH XQFHUn WDLQW\f RI WKH GHFLVLRQ WDVN 7KLV LQFUHDVHG WDVN GLIILFXOW\ FRXOG DIIHFW WKH VXEMHFWnV HQMR\PHQW RI WKH GHFLVLRQ WDVN DQG WKXV ORZHU KLV LQWULQVLF PRWLYDWLRQ ,QGLYLGXDO 'HFLVLRQ &ULWHULD ,QGLYLGXDO GHFLVLRQ FULWHULD ZHUH PHDVXUHG XVLQJ D SDUDPHWHU RI WKH 76' PRGHO 7KLV SDUDPHWHU PHDVXUHV WKH GHFLVLRQ FULWHULD XVHG E\ WKH LQGLYLGXDO LQGHSHQGHQW RI KLV GHFLVLRQ PRGHO VHQVLWLYLW\ 7KH IDFWRUV ZKLFK FDQ DIIHFW DQ LQGLYLGXDOnV GHFLVLRQ FULWHULD DUH f WKH LQGLYLGXDOnV HIILFLHQF\ LQ SURFHVVLQJ RU OHDUQLQJf WKH HIIHFWV RI WKH VWDWHVn UHODWLYH IUHTXHQFLHV DQG WKH UHODWLYH GHFLVLRQ HUURU FRVWV DQG f WKH XVH RI PXOWLSOH GHFLVLRQ DQFKRUV 7R LVRODWH WKHVH IDFWRUV VHYHUDO GHFLVLRQ FULWHULD PHDVXUHV ZHUH GHYHORSHG f WKH %1 PHDVXUH ZKLFK LV DIIHFWHG E\ ERWK IDFWRUVf f WKH %1$c DQG %1&A PHDVXUHV ZKLFK DUH DIIHFWHG E\ RQO\ WKH LQGLYLGXDOnV LQIRUPDWLRQ SURFHVVLQJ HIILFLHQF\f DQG f WKH %1$A%1A PHDVXUH ZKLFK LV DIIHFWHG E\ PXOWLSOH GHFLVLRQ DQFKRUVf

PAGE 167

7KRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO SURGXFHG VLJQLILFDQWO\ ODUJHU PHDQ %1$c PHDVXUHV WKDQ GLG WKRVH VXEMHFWV ZLWKLQ WKH & OHYHO +RZHYHU WKH %1$ PHDVXUH FDQ QRW EH LQWHUSUHWHG VWULFWO\ DV D PHDVXUH RI WKH HIILFLHQF\ RI LQGLYLGXDO LQIRUPDWLRQ SURFHVVLQJ UDWKHU LW LV D IXQFWLRQ RI VXFK D PHDVXUH WKH %1&A PHDVXUH 7KLV LV GXH WR FKDUDFWHULVWLFV RI WKH GDWD VWUXFWXUH DQG WKH GHFLVLRQ WDVN 7KH SDUDPHWHU LV UHODWHG PRQRWRQLFDOO\ WR WKH UDQGRP YDULDEOH GHFLVLRQ D[LV DFWXDO PLQXWHV LQFXUUHGf &RQVHTXHQWO\ IRU WKRVH VXEMHFWV ZLWKn LQ WKH &O FRVW OHYHO WKH Q PHDVXUH DVVRFLDWHG ZLWK WKH LQLWLDO GHFLVLRQ DQFKRU LV ODUJHU WKDQ WKH %N PHDVXUH DQG WKH H[SHFWHG DGMXVWPHQW SURn FHVV KDV WKH HIIHFW RI UHGXFLQJ WKH PHDVXUH )RU WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO WKH c PHDVXUH DVVRFLDWHG ZLWK WKH LQLWLDO GHFLVLRQ DQFKRU LV VPDOOHU WKDQ WKH A PHDVXUH DQG WKH H[SHFWHG DGn MXVWPHQW SURFHVV KDV WKH HIIHFW RI LQFUHDVLQJ WKH A PHDVXUH 7KHUHn IRUH ZKHUH %1$c LV GHILQHG DV WKH UDWLR RI WKH ILQDO PHDVXUH DQG WKH N PHDVXUH D YDOXH JUHDWHU WKDQ RQH LV SURGXFHG IRU WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO DQG D YDOXH OHVV WKDQ RQH LV SURGXFHG IRU WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO 7KH %1&A PHDVXUH HOLPLQDWHV WKLV V\VWHPDWLF UHODWLRQ ZLWK WKH YDOXH RI RQH 7KH HIILFLHQF\ RI LQIRUPDWLRQ SURFHVVLQJ OHDUQLQJf ZDV DIIHFWHG SULPDULO\ E\ WKH FRVW YDULDEOH 7KRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO ZHUH VLJQLILFDQWO\ PRUH FRQVHUYDWLYH OHVV HIILFLHQWf LQ SURn FHVVLQJ WKH LQIRUPDWLRQ WKDQ ZHUH WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO 7KHVH UHVXOWV ZHUH WKH UHYHUVH RI WKRVH SUHGLFWHG E\ WKH K\SRn WKHVHV 7KH VXEMHFWLYH OLPLW FRQFHSW PD\ DJDLQ EH LQWURGXFHG DV DQ H[SODQDWLRQ $V VXEMHFWVn DGMXVWPHQWV ZLWKLQ WKH &O FRVW OHYHO FRQn YHUJHG WRZDUG WKH VWDQGDUG WKH VXEMHFWLYH DGMXVWPHQW OLPLW RI WKH

PAGE 168

VWDQGDUG FRXOG KDYH DFWHG DV DQ LQWHUYHQLQJ YDULDEOH ZKLFK LQFUHDVHG WKH OHYHO RI UHODWLYH FRQVHUYDWLVP UHGXFHG WKH OHYHO RI UHODWLYH LQIRUn PDWLRQ SURFHVVLQJ HIILFLHQF\f $V VXEMHFWVn DGMXVWPHQWV ZLWKLQ WKH & FRVW OHYHO GLYHUJHG IURP WKH VWDQGDUG QR VXFK VXEMHFWLYH DGMXVWPHQW OLPLW H[LVWHG WKXV WKH OHYHO RI UHODWLYH FRQVHUYDWLVP FRXOG KDYH GHn FUHDVHG 7KH KLJKHU VWDWLVWLFDOf PRPHQWV RI WKH %1&c PHDVXUH JLYHQ WKH OHYHOV RI WKH FRVW YDULDEOH ZHUH FRQVLVWHQW ZLWK WKLV FRQFHSW RI D VXEMHFWLYH DGMXVWPHQW OLPLW 7KH GLIIHUHQFH LQ YDULDQFHV RI WKH %1&M PHDVXUH ZDV GXH WR WKH GLIIHUHQFH LQ WKH UDQJHV RI WKH PHDVXUH 7KH UDQJH RI %1&A_&f ZDV DQG WKH UDQJH RI %1&A_&f ZDV 7KH VXEMHFWLYH DGMXVWPHQW OLPLW VKRXOG KDYH KDG WKH HIIHFW RI VNHZLQJ WKH PHDVXUH DZD\ IURP WKRVH YDOXHV ZKLFK LQGLFDWHG ORZHU OHYHOV RI FRQVHUYDWLVP 7KH WKLUG PRPHQW DV H[SUHVVHG E\ WKH FRHIILFLHQW RI VNHZQHVVf RI WKH %1&A PHDVXUH LQGLFDWHG WKDW f WKH VNHZQHVV RI WKH %1&A c&f GLVWULEXWLRQ ZDV SRVLWLYH VNHZHG DZD\ IURP YDOXHV ZKLFK LQGLFDWHG ORZHU OHYHOV RI FRQVHUYDWLVPf DQG f WKH VNHZQHVV RI WKH %1&A_&f GLVWULEXWLRQ ZDV QHJDWLYH VNHZHG WRZDUG YDOXHV ZKLFK LQn GLFDWHG ORZHU OHYHOV RI FRQVHUYDWLVPf 7KH PXOWLSOH GHFLVLRQ DQFKRU WKH %1$A%1A PHDVXUHf ZDV DIIHFWHG SULPDUO\ E\ WKH FRVW YDULDEOH 7KRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO HPSOR\HG PXOWLSOH GHFLVLRQ DQFKRUV WR D ODUJHU H[WHQW WKDQ GLG WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO 7KH K\SRWKHVHV SRVLWHG D UHODWLRQ EHWZHHQ PXOWLSOH GHFLVLRQ DQFKRUV DQG WKH GLVWDQFH VHSDUDWLQJ WKH ORZHU WDLO RI WKH LQFRQWURO VWDWH IURP WKH VXEMHFWnV ILQDO GHn FLVLRQ DQFKRU 7KLV UHODWLRQ ZDV VLPLODU WR WKDW SURSRVHG LQ WKH '1$A'1A PHDVXUH +RZHYHU WKH UHVXOWV GLG QRW VXSSRUW WKH K\SRWKHVHV

PAGE 169

DQG WKXV SDUDOOHO WKRVH REWDLQHG ZLWK WKH '1$A'1A PHDVXUH %RWK WKH '1$A'1A DQG WKH %1$A%1A PHDVXUHV REWDLQHG UHVXOWV WKDW LQGLFDWHG JUHDWHU PXOWLSOH GHFLVLRQ DQFKRU XVDJH ZKHQ WKH VXEMHFWVn LQLWLDO GHn FLVLRQ DQFKRUV ZHUH FORVHU WR WKH VWDQGDUG WKH 6, GLVWULEXWLRQ OHYHO ZLWKLQ WKH '1$A'1A PHDVXUH DQG WKH &O FRVW OHYHO ZLWKLQ WKH %1$A%1A PHDVXUHf 1R SODXVLEOH H[SODQDWLRQ IRU WKHVH UHVXOWV LV UHDGLO\ DSSDUHQW 7KH REWDLQHG DVVRFLDWLRQV EHWZHHQ WKH LQGLYLGXDO GHFLVLRQ FULWHULD PHDVXUHV DQG WKH LQGLYLGXDO DWWULEXWHV RI H[WULQVLF PRQHWDU\f PRWLn YDWLRQ DQG VXEMHFW *3$ KDYH LQWXLWLYH LQWHUSUHWDWLRQ $V ERWK WKH H[WULQVLF PRQHWDU\f PRWLYDWLRQ PHDVXUHV DQG WKH VXEMHFWVn *3$V GHn FUHDVHG WKH OHYHO RI FRQVHUYDWLVP LQFUHDVHG )XUWKHUPRUH DV WKH VXEn MHFWVn *3$V GHFUHDVHG WKH XVH RI PXOWLSOH GHFLVLRQ DQFKRUV LQFUHDVHG 2I WKH WZR LQGLYLGXDO SURFHVV YDULDEOHV GHFLVLRQ PRGHO VHQVLWLYLW\ DQG GHFLVLRQ FULWHULDf WKH GHFLVLRQ FULWHULD KDG WKH PRUH VLJQLILn FDQW DIIHFW RQ D VXEMHFWnV SHUIRUPDQFH IHHGEDFN PHDVXUH DQG LQ WXUQ RQ WKH OHYHO RI KLV PRQHWDU\ UHZDUGV +RZHYHU VLQFH PRWLYDWLRQ SUHWHVWV ZHUH QRW DGPLQLVWHUHGFDXVDO UHODWLRQVKLSV LH GHFUHDVHG PRWLYDWLRQ OHG WR LQFUHDVHG FRQVHUYDWLVP RU LQFUHDVHG FRQVHUYDWLVP OHG WR GHFUHDVHG PRWLYDWLRQf FDQ QRW EH FODLPHG ,QGLYLGXDO /RQJ5XQ 'HFLVLRQ (IILFLHQF\ 7KH RYHUDOO HIILFLHQF\ RI DQ LQGLYLGXDOnV GHFLVLRQ PDNLQJ ZLWKLQ WKH H[SHULPHQW ZDV PHDVXUHG XVLQJ WKH *c YDULDEOH 7KH YDULDEOH ZDV WKH LQGLYLGXDOnV QHW DGGLWLRQDO GHFLVLRQ FRVWV UHODWLYH WR WKH RSWLPDO PRGHOnV WRWDO GHFLVLRQ FRVWV 7KH *A YDULDEOH ZDV GLVDJJUHJDWHG LQWR WZR PHDVXUHV WKH WRWDO DGGLWLRQDO LQGLYLGXDO GHFLVLRQ FRVWV UHODWLYH

PAGE 170

WR WKH RSWLPDO PRGHOnV WRWDO GHFLVLRQ FRVWV *3cf DQG WKH WRWDO DGGLWLRQDO LQGLYLGXDO GHFLVLRQ VDYLQJV UHODWLYH WR WKH RSWLPDO PRGHOnV WRWDO GHFLVLRQ FRVWV *1Mf 7KH K\SRWKHVHV FRQFHUQLQJ WKHVH YDULDEOHV ZHUH GHULYHG XVLQJ D VLPXODWLRQ ZKLFK DVVXPHG D VHW RI LQLWLDO GHFLVLRQ DQFKRUV DQG DVVXPHG HTXDO OHDUQLQJ HIILFLHQF\ LQ WHUPV RI UHODWLYH OLQHDU DGMXVWPHQWV DORQJ WKH UDQGRP YDULDEOH D[LVf *LYHQ WKHVH DVVXPSWLRQV VSHFLILF YDULDWLRQV LQ WKH GHFLVLRQ VLWXDWLRQ ZHUH H[n SHFWHG WR KDYH VSHFLILF HIIHFWV RQ D VXEMHFWnV UHODWLYH GHFLVLRQ FRVWV VHH 7DEOH LQ &KDSWHU 9 IRU D VXPPDU\ RI WKHVH H[SHFWHG HIIHFWV DQG WKH REWDLQHG UHVXOWVf $OO RI WKHVH H[SHFWHG GHFLVLRQ VLWXDWLRQ HIIHFWV ZHUH VXSSRUWHG DW OHDVW LQ SDUW E\ WKH REWDLQHG UHVXOWV +RZHYHU WKH UHVXOWV RI WKH LQGLYLGXDO GHFLVLRQ PRGHO VHQVLWLYLW\ DQG WKH LQGLn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f $ FORVHU H[DPLQDWLRQ RI ERWK WKH VLPXODWHG DQG WKH REWDLQHG FRVW YDULDEOH HIIHFWV RQ WKH *c YDULDEOH LQGLFDWHG WKDW WKH REWDLQHG HIIHFWV ZHUH QRW DV VWURQJ DV WKH VLPXODWHG HIIHFWV QRWH WKDW *c_&f *LO&Of HTXDOOHG ZKHUHDV I*M_&fI*M_&Of HTXDOOHG f 7KH VLPXODWHG HIIHFWV PD\ EH DGMXVWHG IRU XQHTXDO OHDUQLQJ HIILFLHQF\ E\ DVVXPLQJ WKH UHODWLYH GHFLVLRQ FRVWV ZLWKLQ WKH &O FRVW OHYHO ZRXOG EH LQFUHDVHG E\ D IDFWRU RI WZR UHODWLYH WR WKH & FRVW OHYHOf $IWHU

PAGE 171

WKLV DGMXVWPHQW WKH VLPXODWHG FRVW YDULDEOH HIIHFWV ZRXOG KDYH WKH VDPH VWUHQJWK DV WKH REWDLQHG HIIHFWV LH I *M & fm I *A &Of ZRXOG HTXDO f 7KH VLPXODWHG DQG REWDLQHG GLVWULEXWLRQ E\ FRVW YDULDEOH LQWHUn DFWLRQV LQYROYLQJ WKH UHODWLYH GHFLVLRQ FRVWV KDG VLPLODU UHODWLRQVKLSV EHWZHHQ WKHLU VWUHQJWKV WKH HIIHFWV RI WKH REWDLQHG LQWHUDFWLRQ ZHUH QRW DV VWURQJ DV WKH HIIHFWV RI WKH VLPXODWHG LQWHUDFWLRQ QRWH WKDW >*_6&f*c _6&f@ >A _6 & f*L _6&f@ HTXDOOHG ZKHUHDV >I*L_6O&fI*L _6&f@ >I*cc6,&fI2M _6&f@ HTXDOOHG f $JDLQ LI WKH VLPXODWHG HIIHFWV DUH DGMXVWHG IRU XQHTXDO OHDUQLQJ HIILFLHQF\ WKHQ WKH VLPXODWHG GLVWULEXWLRQ E\ FRVW YDULDEOH LQWHUn DFWLRQV KDYH D VWUHQJWK VLPLODU WR WKDW RI WKH REWDLQHG HIIHFWV LH &I*c _6 &fIgL _6&f@ >rI*L_6O&OfrI*L_6&Of@ ZRXOG HTXDO f 7KH VLPXODWHG DQG REWDLQHG GLVWULEXWLRQ E\ FRVW YDULDEOH LQWHUn DFWLRQ ZLWKLQ WKH *3A*1A PHDVXUH KDG RSSRVLWH UHODWLRQVKLSV EHWZHHQ WKHLU VWUHQJWK WKH HIIHFWV RI WKH REWDLQHG LQWHUDFWLRQ ZHUH VWURQJHU WKDQ WKH HIIHFWV RI WKH VLPXODWHG LQWHUDFWLRQ QRWH WKDW >*3A*1A_&6f *3L*1L_&6f@ >*3L*1L_&6f*3L*1L_&O6f@ HTXDOOHG ZKHUHDV >I*3L*1L_&6fI*3L*0L_&6f@ >I*3*1L_&6f I*3A*1A_&O6f@ HTXDOOHG f 7R H[SODLQ WKH GLIIHUHQFH LQ HIIHFW VWUHQJWKV UHTXLUHV UHIHUHQFH WR WZR YDULDEOHV PXOWLSOH GHFLVLRQ DQn FKRUV DQG XQHTXDO OHDUQLQJ HIILFLHQF\ ,I WKH *3c*1c PHDVXUHV DUH DGMXVWHG IRU WKH HIIHFWV RI PXOWLSOH GHFLVLRQ DQFKRUV WKH VWUHQTWK RI WKH REWDLQHG LQWHUDFWLRQ HIIHFW LV XVLQJ WKH DERYH LQWHUDFWLRQ VWUHQJWK UDWLRf 7KLV UHYHUVHV WKH UHODWLRQVKLSV EHWZHHQ WKH HIIHFW VWUHQJWKV LH WKH REWDLQHG LQWHUDFWLRQ VWUHQJWK EHFRPHV ZHDNHU

PAGE 172

WKDQ WKH SUHGLFWHG VWUHQJWK 7KH HIIHFWV RI XQHTXDO OHDUQLQJ HIILFLHQF\ ZHUH f DQ H[SDQVLRQ RI WKH LQWHUYDO EHWZHHQ WKH ILQDO GHFLVLRQ DQn FKRU DQG WKH RSWLPDO GHFLVLRQ YDOXH IRU WKRVH VXEMHFWV ZLWK JUHDWHU GHFLVLRQ FULWHULD FRQVHUYDWLVP DQG f WKH UHGXFWLRQ RI WKH LQWHUYDO EHWZHHQ WKH ILQDO GHFLVLRQ DQFKRU DQG WKH RSWLPDO GHFLVLRQ YDOXH IRU WKRVH VXEMHFWV ZLWK OHVV GHFLVLRQ FULWHULD FRQVHUYDWLVP $V WKH LQn WHUYDO DVVRFLDWHG ZLWK WKRVH VXEMHFWV ZLWKLQ WKH &O FRVW OHYHO H[SDQGHG WKLV OHYHO KDG JUHDWHU GHFLVLRQ FULWHULD FRQVHUYDWLVPf ERWK WKH *3cc&f DQG WKH *1M _&Of EHFDPH ODUJHU +RZHYHU WKH *3M PHDVXUH EHFDPH ODUJHU DW D JUHDWHU UDWH WKDQ GLG WKH *1c PHDVXUH *3c*1c!f $V WKH LQWHUYDO DVVRFLDWHG ZLWK WKRVH VXEMHFWV ZLWKLQ WKH & FRVW OHYHO FRQWUDFWHG WKLV OHYHO KDG OHVV GHFLVLRQ FULWHULD FRQVHUYDWLVPf ERWK WKH *3M c&f DQG WKH *1M _&f EHFDPH VPDOOHU +RZHYHU WKH *3M PHDVXUH EHFDPH VPDOOHU DW D JUHDWHU UDWH WKDQ GLG WKH *1c PHDVXUH *3A*1A!f &RQVHTXHQWO\ WKH UDWLR *3*1_&Of LQFUHDVHG DV WKH OHYHO RI FRQVHUn YDWLVP LQFUHDVHG DQG WKH UDWLR *3A*1-&f GHFUHDVHG DV WKH OHYHO RI FRQVHUYDWLVP GHFUHDVHG ,QFRUSRUDWLQJ WKHVH HIIHFWV RI XQHTXDO OHDUQLQJ HIILFLHQF\ LQWR WKH VLPXODWHG GLVWULEXWLRQ E\ FRVW YDULDEOH LQWHUDFWLRQ ORZHUHG WKH SUHGLFWHG VWUHQJWK RI VXFK DQ LQWHUDFWLRQ 7KH REVHUYHG DVVRFLDWLRQV EHWZHHQ WKH LQGLYLGXDO ORQJUXQ GHFLVLRQ HIILFLHQF\ PHDVXUHV DQG WKH LQGLYLGXDO DWWULEXWHV PD\ EH LQWHUSUHWHG ZLWK VRPH GHJUHH RI LQWXLWLYH VDWLVIDFWLRQ $V WKH VXEMHFWVn YDULDEOH UHVSRQVH UDQJHV LQFUHDVHG WKH UHODWLYH GHFLVLRQ FRVWV LQFUHDVHG HLWKHU DV D UHVXOW RI LQFUHDVHG DGGLWLRQDO GHFLVLRQ FRVWV RU GHFUHDVHG DGGLn WLRQDO GHFLVLRQ VDYLQJVf )XUWKHUPRUH DV WKH UHODWLYH GHFLVLRQ FRVWV LQFUHDVHG ERWK WKH VXEMHFWVn LQWULQVLF DQG H[WULQVLF PRQHWDU\f PRWLn YDWLRQ PHDVXUHV GHFUHDVHG $ VLPLODU DVVRFLDWLRQ H[LVWHG EHWZHHQ WKH

PAGE 173

VXEMHFWVn *3$V DQG WKH UHODWLYH GHFLVLRQ FRVWV DV WKH VXEMHFWVn *3$V LQFUHDVHG HLWKHU DGGLWLRQDO GHFLVLRQ FRVWV GHFUHDVHG RU DGGLWLRQDO GHn FLVLRQ VDYLQJV LQFUHDVHG ,W ZRXOG EH UHDVRQDEOH WR DVVXPH WKDW ERWK WKH '1A PHDVXUH DQG WKH VXEMHFW *3$ DUH PRGLI\LQJ YDULDEOHV RQ D VXEn MHFWnV GHFLVLRQ FRVW SHUIRUPDQFH 6LQFH SUHWHVWV ZHUH QRW DGPLQLVWHUHG LW ZDV QRW SRVVLEOH WR VWDWH WKH VSHFLILF UHODWLRQV EHWZHHQ WKH PRWLn YDWLRQ IDFWRUV DQG D VXEMHFWn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n SHULPHQWDO WDVN LV D SRVVLEOH OLPLWDWLRQ $OWKRXJK WKH VXEMHFWV UHn FHLYHG WUDLQLQJ LQ WKH H[SHULPHQWDO WDVN WKH SULPDU\ VRXUFH RI WKHLU NQRZOHGJH FRQFHUQLQJ VWDQGDUG FRVW YDULDQFH LQYHVWLJDWLRQ FRXOG FRPH

PAGE 174

IURP WKH FROOHJH FODVVURRP &RQVHTXHQWO\ LI WKH\ ZHUH QRW WDXJKW ZLWKLQ WKH FODVVURRPf WKDW VLWXDWLRQV H[LVW LQ ZKLFK LQYHVWLJDWLRQ GHFLVLRQ YDOXHV DUH ORFDWHG UHODWLYHO\ FORVH WR WKH VWDQGDUG WKHQ JUHDWHU GHFLVLRQ FULWHULD FRQVHUYDWLVP ZLWKLQ WKHVH VLWXDWLRQV FRXOG EH WKH UHVXOW RI WKH ODFN RI VXFK NQRZOHGJH 7KLV VXJJHVWV WKH SRVVLELOLW\ RI DQ DYDLODELOLW\ ELDV 7YHUVN\ DQG .DKQHPDQ f +RZHYHU WR WKH H[WHQW D PDQDJHU PXVW OHDUQ IURP KLV RZQ H[SHULHQFHV VXFK D ELDV FRXOG H[LVW LQ WKH UHDO ZRUOG $ ILQDO OLPLWDWLRQ GLVFXVVHG ZLWKLQ WKLV VHFWLRQf LQYROYHV WKH VHOHFWLRQ RI WKH OHYHOV RI WKH VLWXDWLRQ YDULDEOHV 6HYHUDO VLWXDWLRQ YDULDEOHV ZHUH VHW DW D VLQJOH OHYHO ZKLOH RWKHUV ZHUH PDQLSXODWHG DW PXOWLSOH OHYHOV 7KH LPSRUWDQW SRLQW LV WKDW RQO\ VSHFLILF FRPELQDn WLRQV RI YDULDEOH OHYHOV ZHUH VWXGLHG ZLWKLQ WKLV UHVHDUFK ZKHUHDV DQ LQILQLWH QXPEHU RI FRPELQDWLRQV DUH SRVVLEOH 'LIIHUHQW YDULDEOH OHYHOV DQG GLIIHUHQW YDULDEOH PDQLSXODWLRQV ZRXOG FUHDWH D GLIIHUHQFH LQ WKH H[SHULPHQWDO HQYLURQPHQW ZKLFK FRXOG SURGXFH UHVXOWV RWKHU WKDQ WKRVH REWDLQHG LQ WKLV VWXG\ 7KHVH H[SHULPHQWDO HQYLURQPHQW OLPLWDWLRQV GR QRW PHDQ KRZHYHU WKDW WKH FRQFHSWV SURSRVHG ZLWKLQ WKLV VWXG\ KDYH QRW SDVVHG D PHDQLQJn IXO HPSLULFDO WHVW 7KH REMHFWLYHV RI H[SHULPHQWV LQYROYLQJ WKHRU\ RU FRQFHSW WHVWLQJ DQG WKH UHODWLRQV EHWZHHQ WKH QDWXUH FUHDWHG LQ WKH H[SHULPHQW DQG WKH QDWXUH H[LVWLQJ LQ WKH ZRUOG DW ODUJH DUH GLVFXVVHG E\ =LJOHU f :KDW WKH H[SHULPHQWHU LV VD\LQJ LV WKDW LI VXFK DQG VXFK KROGV LQ WKH UHDO ZRUOG EHFDXVH RI WKH SULQFLSOHV H[SRXQGHG LQ WKH SDUWLFXODU WKHRU\ XQGHU LQYHVWLJDWLRQ WKHQ VXFK DQG VXFK VKRXOG KROG LQ WKH ZRUOG ZKLFK WKH H[SHULPHQWHU KDV FUHDWHG 7KLV WUDQVDWDELOLW\ LV ZKDW JLYHV WKHRUHWLFDO LPSRUW WR H[SHULPHQWV ZKLFK LQYROYH SKHQRPHQD ZKLFK WDNHQ LQ LVRODWLRQ QRW RQO\ DSSHDU SLFD\XQH EXW VHHP WR KDYH OLWWOH UHODWLRQVKLS ZLWK ZKDW RQH REVHUYHV LQ QDWXUH SS f

PAGE 175

)LQGLQJ WKH SURSRVHG FRQFHSWV WR KROG ZLWKLQ WKH H[SHULPHQWDO ZRUOG GRHV QRW YDOLGDWH WKH FODLP WKDW WKH\ KROG ZLWKLQ WKH UHDO ZRUOG 2Q WKH RWKHU KDQG ILQGLQJ WKH SURSRVHG FRQFHSWV QRW WR KROG ZLWKLQ WKH H[SHULPHQWDO ZRUOG ZRXOG FDVW VHULRXV GRXEW RQ ZKHWKHU WKH\ KROG LQ WKH UHDO ZRUOG 3RVLWLYH H[SHULPHQWDO UHVXOWV DUH D IRUP RI QHJDWLYH DVVXUDQFH WKH FRQFHSWV KDYH SDVVHG DQ LQLWLDO HPSLULFDO WHVW DQG UHPDLQ DOEHLW VRPHZKDW VFDUUHG LQ PRVW LQVWDQFHVf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f WKH VHOHFWLRQ RI WKH W\SH RI VXEMHFW GHSHQGV XSRQ WKH DQVZHU WR WKH TXHVWLRQ :KLFK VXEMHFW LV D EHWWHU VXUURJDWH IRU D PDQDJHU LQ D UHDO ZRUOG HQYLURQPHQW D NQRZn OHGJHDEOH VWXGHQW RU WKH PDQDJHU KLPVHOI" 7KHUH DUH RI FRXUVH QR FOHDU DQVZHUV WR WKLV TXHVWLRQ +RZHYHU VHYHUDO DVSHFWV RI WKH VLWXDWLRQ IDYRU WKH XVH RI D NQRZOHGJHDEOH VWXGHQW )LUVW WKH KLJK GHJUHH RI DEVWUDFWLRQ ZLWKLQ WKH H[SHULPHQWDO HQYLURQPHQW FRXOG KDYH PDGH WKH WDVN DSSHDU WULYLDO WR D PDQDJHU DQG FRXOG KDYH HOLFLWHG EHKDYLRU QRW FRQVLVWHQW ZLWK WKH SULQFLSDO VLWXDWLRQ 6HFRQG WKH

PAGE 176

H[SHULPHQWDO HQYLURQPHQW LV FORVHU WR WKDW RI D WHVWWDNLQJ VLWXDWLRQ ZKLFK LV D PRUH QDWXUDO HQYLURQPHQW IRU D VWXGHQW WKDQ IRU D PDQDJHU ,PSOLFDWLRQV )RU $FFRXQWLQJ 7ZR JHQHUDO LPSOLFDWLRQV RI WKLV UHVHDUFK IRU DFFRXQWLQJ DUH GLVFXVVHG 7KHVH LPSOLFDWLRQV DUH WKH YDOXH RI DGGLWLRQDO LQIRUPDWLRQ DQG WKH JHQHUDO VWDQGDUG VHWWLQJ SURFHVV 9DOXH RI $GGLWLRQDO ,QIRUPDWLRQ 7KH PDQLSXODWLRQ RI WKH LQIRUPDWLRQ YDULDEOH LQYROYHG WKH TXDQn WLW\ RI LQIRUPDWLRQ FRQWDLQHG LQ WKH DYDLODEOH LQIRUPDWLRQ VHW VSHFLILn FDOO\ WKH SUHVHQFH RU DEVHQFH RI YDULRXV GLVWULEXWLRQ LQIRUPDWLRQ LWHPV 7KH UHVXOWV LQGLFDWHG WKDW WKH LQIRUPDWLRQ YDULDEOH GLG QRW KDYH D VLJQLILFDQW HIIHFW XSRQ WKH LQGLYLGXDO UHODWLYH GHFLVLRQ FRVWV 7KH PHDQ UHODWLYH GHFLVLRQ FRVW RI VXEMHFWV ZLWKLQ WKH UHGXFHG LQIRUPDWLRQ OHYHO GLG QRW GLIIHU VLJQLILFDQWO\ IURP WKDW RI VXEMHFWV ZLWKLQ WKH H[SDQGHG LQIRUPDWLRQ OHYHO $Q LPSOLFDWLRQ RI WKLV UHVXOW FRQFHUQV WKH QHW EHQHILW IRU WKH FRPSDQ\f RI SURYLGLQJ WKH DGGLWLRQDO LQIRUPDWLRQ ZLWKLQ WKH H[SDQGHG LQIRUPDWLRQ OHYHO 7KH DGGLWLRQDO LQIRUPDWLRQ ZLWKLQ DQ DFWXDO HQYLURQPHQW LV QRW FRVWOHVV &RQVHTXHQWO\ LI VXFK LQIRUPDWLRQ LV WR EH SURYLGHG RWKHU WKLQJV EHLQJ HTXDO WKH QHW EHQHILW RI VXFK DQ DFWLRQ VKRXOG EH SRVLWLYH ,Q WKLV FRQWH[W WKH QHW EHQHILW ZRXOG EH WKH PDUJLQDO GHFLVLRQ FRVW VDYLQJV UHVXOWLQJ IURP LPSURYHG PDQDJHU YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV JLYHQ WKH DGGLWLRQDO LQIRUPDWLRQf OHVV WKH FRVWV DVVRFLDWHG ZLWK WKH SURGXFWLRQ DQG GLVn VHPLQDWLRQ RI WKH DGGLWLRQDO LQIRUPDWLRQ :LWKLQ WKLV UHVHDUFK WKH ODFN

PAGE 177

RI D VLJQLILFDQW LQIRUPDWLRQ YDULDEOH HIIHFW LPSOLHV WKDW WKH QHW EHQHILW RI SURYLGLQJ DGGLWLRQDO LQIRUPDWLRQ PD\ QRW EH SRVLWLYH +RZHYHU LW VKRXOG EH QRWHG WKDW WKH ODFN RI DQ LQIRUPDWLRQ YDULDEOH HIIHFW FRXOG EH WKH UHVXOW RI HLWKHU f WKDW WKH VXEMHFWV ZLWKLQ WKH UHGXFHG LQIRUPDWLRQ OHYHO ZHUH DEOH WR HVWLPDWH ZLWK UHODWLYH HIILFLHQF\f WKH PLVVLQJ LQIRUPDWLRQ DV D UHVXOW RI WKHLU H[SHULHQFHV ZLWK WKH GHFLVLRQ WDVN RU f WKDW WKH VXEMHFWV ZLWKLQ WKH H[SDQGHG LQIRUPDWLRQ OHYHO GLG QRW XWLOL]H WKH DGGLWLRQDO LQIRUPDWLRQ HIILFLHQWO\ *HQHUDO 6WDQGDUG 6HWWLQJ 3URFHVV $QRWKHU LPSOLFDWLRQ RI WKLV UHVHDUFK FRQFHUQV WKH JHQHUDO VWDQGDUG VHWWLQJ SURFHVV :LWKLQ GHFLVLRQ VLWXDWLRQV LQYROYLQJ HLWKHU FRQn YHUJHQW RU GLYHUJHQW DGMXVWPHQW GHFLVLRQ FULWHULDFRQVHUYDWLVP ZDV GHn ILQHG DV D OHVV WKDQ FRPSOHWH DGMXVWPHQW WRZDUG WKH RSWLPDO GHFLVLRQ YDOXH &RQYHUJHQW DGMXVWPHQW LQYROYHG LQGLYLGXDO GHFLVLRQ YDOXH DGMXVWPHQW WRZDUG WKH VWDQGDUG DQG GLYHUJHQW DGMXVWPHQW LQYROYHG LQn GLYLGXDO GHFLVLRQ YDOXH DGMXVWPHQW DZD\ IURP WKH VWDQGDUG 7KH VWDQn GDUG XVHG ZLWKLQ WKLV UHVHDUFK FDQ EH FRQFHLYHG RI DV D W\SH RI GHn FLVLRQ EHKDYLRU PDJQHW 2YHU FHUWDLQ UDQJHV RI LWV ILHOG WKH UHOHn YDQW GHFLVLRQ D[LVf WKH VWDQGDUG DWWUDFWV YDULDQFH LQYHVWLJDWLRQ EHn KDYLRU DQG RYHU RWKHU UDQJHV RI LWV ILHOG LW UHSHOV YDULDQFH LQYHVWLn JDWLRQ GHFLVLRQ EHKDYLRU 'HFLVLRQ FULWHULD FRQVHUYDWLVP ZKHQ WKH DGMXVWPHQW SURFHVV GLYHUJHV IURP WKH VWDQGDUG FRXOG EH WKH UHVXOW RI WKH DWWUDFWLRQ RU SXOOf RI WKH VWDQGDUG WKDW UHVWUDLQV WKH LQGLYLGXDO IURP PDNLQJ D FRPSOHWH GLYHUJHQWf DGMXVWPHQW WR WKH RSWLPDO GHFLVLRQ YDOXH 'HFLVLRQ FULWHULD FRQVHUYDWLVP ZKHQ WKH DGMXVWPHQW SURFHVV FRQYHUJHV WR WKH VWDQGDUG FRXOG EH WKH UHVXOW RI WKH UHSHOOLQJ IRUFH

PAGE 178

RU SXVKf RI WKH VWDQGDUG VXEMHFWLYHO\ OLPLWLQJ D FRPSOHWH FRQYHUJHQWf DGMXVWPHQW WR WKH RSWLPDO GHFLVLRQ YDOXH 7KH GHFLVLRQ EHKDYLRU PDJQHW YLHZ UHSUHVHQWV D JHQHUDOL]DWLRQ RI WKH VXEMHFWLYH DGMXVWPHQW OLPLW FRQFHSW 7KH VXEMHFWLYH OLPLW FRQFHSW DV SUHVHQWHG HDUOLHU GHDOW ZLWK GHFLVLRQ FULWHULD FRQVHUYDWLVP LQYROYLQJ RQO\ FRQYHUJHQW DGMXVWPHQW 7KH TXHVWLRQ RI ZKHWKHU GHFLVLRQ PDNHUV FRXOG OHDUQ HLWKHU WKURXJK WUDLQLQJ RU H[SHULHQFHf WR UHGXFH WKH GHFLVLRQ ELDVHV LPSOLHG E\ WKH GHFLVLRQ EHKDYLRU PDJQHW FRQFHSW UHPDLQV XQDQVZHUHG $Q DOWHUQDWLYH DSSURDFK KRZHYHU LQYROYHV WKH VWDQGDUG VHWWLQJ SURFHVV LWVHOI 3UHYLRXV UHVHDUFK LQYROYLQJ WKH QDWXUH RI VWDQGDUGV HJ VWULFW VWDQGDUGV FXUUHQWO\ DWWDLQDEOH VWDQGDUGV OD[ VWDQGDUGVf KDV SULPDULO\ GHDOW ZLWK WKH VWDQGDUGnV PRWLYDWLRQDO DIIHFWV 2WKHU WKLQJV EHLQJ HTXDO WKH QDWXUH RI WKH VWDQGDUG PD\ DIIHFW WKH GHFLVLRQ EHKDYLRU PDJQHW ELDVHV :LWKLQ VLWXDWLRQV UHTXLULQJ DGMXVWPHQW OD[ VWDQGDUGV ZKLFK KDYH YDOXHV JUHDWHU WKDQ WKH PHDQ RI WKH LQFRQWURO VWDWHf PD\ UHGXFH VXEVWDQWLDOO\ GLYHUJHQW GHFLVLRQ FULWHULD FRQVHUYDWLVP :LWKLQ VLWXDWLRQV UHTXLULQJ FRQYHUJHQW DGMXVWPHQW VWULFW VWDQGDUGV ZKLFK KDYH YDOXHV OHVV WKDQ WKH PHDQ RI WKH LQFRQWURO VWDWHf PD\ UHGXFH VXEVWDQWLDOO\ FRQYHUJHQW GHFLVLRQ FULWHULD FRQVHUYDWLVP )XWXUH 5HVHDUFK 7KH LPSOLFDWLRQV RI WKLV VWXG\ IRU IXWXUH DFFRXQWLQJ UHVHDUFK WDNH WZR JHQHUDO IRUPV PRGLILFDWLRQ RI WKH H[LVWLQJ FRQFHSWXDO GHYHORSPHQW DQG JHQHUDOL]DWLRQV RI WKH PRGLILHG FRQFHSWXDO GHYHORSPHQW 0RGLILFDWLRQV RI WKH H[LVWLQJ FRQFHSWXDO GHYHORSPHQW ZRXOG LQYROYH LQFRUSRUDWLQJ DQG WHVWLQJ WKH H[ SRVW H[SODQDWLRQV RIIHUHG IRU FHUWDLQ

PAGE 179

YDULDQFH LQYHVWLJDWLRQ GHFLVLRQ EHKDYLRU REVHUYHG ZLWKLQ WKLV UHVHDUFK *HQHUDOL]DWLRQ RI WKH PRGLILHG FRQFHSWXDO GHYHORSPHQW ZRXOG LQYROYH H[WHQVLRQV DLPHG DW UHGXFLQJ WKH OLPLWDWLRQV RI WKH H[LVWLQJ VWXG\ 7KH SULPDU\ H[ SRVW FRQFHSW GHYHORSHG ZLWKLQ WKLV UHVHDUFK LV WKH GHFLVLRQ EHKDYLRU PDJQHW YLHZ LQFRUSRUDWLQJ WKH FRQFHSW RI VXEn MHFWLYH OLPLWVf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n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n \RQG H[WHQGLQJ WKH GHFLVLRQ FULWHULD FRQVHUYDWLVP HIIHFWV WKH FHQWUDO TXHVWLRQ ZRXOG EH ZKHWKHU GLIIHUHQWLDO LQIRUPDWLRQ SURFHVVLQJ

PAGE 180

HIILFLHQF\ GHFLVLRQ FULWHULD FRQVHUYDWLVPf LV D FKDUDFWHULVWLF RI RQO\ WUDLQLQJ OHDUQLQJf VLWXDWLRQV RU UHPDLQV D FKDUDFWHULVWLF RI ZHOOWUDLQHG QRQOHDUQLQJf VLWXDWLRQV ([WHQVLRQV LQYROYLQJ WKH H[SHULPHQWDO HQYLURQPHQW FRXOG WDNH VHYHUDO IRUPV )LUVW WKH\ PD\ UHODWH WR WKH OHYHOV RI WKH VLWXDWLRQ YDULDEOHV 2QO\ FHUWDLQ VLWXDWLRQ YDULDEOHV ZHUH VWXGLHG ZLWKLQ WKLV UHVHDUFK H[WHQVLRQV FRXOG VWXG\ WKH DIIHFWV RI PDQLSXODWLQJ GLIIHUHQW VLWXDWLRQ YDULDEOHV DW GLIIHUHQW OHYHOV HJ PRUH WKDQ WZR VWDWHV RI QDWXUH QRQQRUPDO GLVWULEXWLRQV YDULDEOH UHODWLYH IUHTXHQFLHV YDULn DEOH GHFLVLRQ FRVW IXQFWLRQV PXOWLSOH VWDQGDUG FRVW YDULDQFHVf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f WKH SURFHVV PD\ RU PD\ QRW PRYH WR DQRWKHU VWDWH RI QDWXUH 7KH WDVN RI WKH GHFLVLRQ PDNHU ZRXOG EH WR GHWHFW WKH SRLQW LQ WLPH LQ ZKLFK WKH SK\VLFDO SURFHVV VKLIWV WR DQn RWKHU VWDWH 7KH WKLUG DQG ILQDO H[WHQVLRQ GLVFXVVHG ZLWKLQ WKLV VHFWLRQ FRQn FHUQV WKH QDWXUH RI WKH H[SHULPHQWDO GHFLVLRQV :LWKLQ WKLV UHVHDUFK WKH GHFLVLRQV LQYROYHG WKH GHWHFWLRQ RI VWDWH RI QDWXUH IRU WKH SXUSRVH RI SK\VLFDO SURFHVV FRQWURO ZLWKLQ FRVW PLQLPL]DWLRQ FRQVWUDLQWV %Hn \RQG WKH RSHUDWLRQDO FRQWURO OHYHO VWDQGDUG YDULDQFH UHSRUWV FDQ EH

PAGE 181

XWLOL]HG DV SHUIRUPDQFH PHDVXUHV IRU WKH DSSURSULDWH RSHUDWLRQDO PDQDJHUV DQG IRU WKH SK\VLFDO V\VWHPV XQGHU WKHLU FRQWURO 7KH QDWXUH RI GHn FLVLRQV RI WKLV W\SH DOWKRXJK JHQHUDOO\ DJJUHJDWHG RYHU ORQJHU WLPH SHULRGV LV WKDW RI DGHTXDWH SHUIRUPDQFH YHUVHV LQDGHTXDWH SHUIRUPDQFH UDWKHU WKDQ WKDW RI LQYHVWLJDWLRQ YHUVHV QRQLQYHVWLJDWLRQ :LWKLQ WKH FXUUHQW UHVHDUFK SHUIRUPDQFH GHFLVLRQV ZHUH DVVXPHG LPSOLFLWO\ ZLWKLQ WKH SHUIRUPDQFH IHHGEDFN PHDVXUHV JLYHQ WKH VXEMHFWV GXULQJ WKH WUDLQn LQJ SKDVH $Q H[WHQVLRQ ZRXOG LQYROYH WKH DFWXDO IRUPDWLRQ RI VXFK GHFLVLRQV DQG LQ WXUQ WKHLU HIIHFWV XSRQ RSHUDWLRQDO FRQWURO GHFLVLRQV

PAGE 182

$33(1',; $ 7+(25< 2) 6,*1$/ '(7(&7,21 '(5,9$7,216 5HODWLRQVKLS RI &RQGLWLRQDO 3UREDELOLW\ 0DWUL[ DQG %D\HV 7KHRUHP 7KH XQFRQGLWLRQDO SUREDELOLWLHV RI WKH IRXU HYHQW RXWFRPHV FDQ EH VWDWHG LQ WHUPV RI WKH FRQGLWLRQDO SUREDELOLWLHV DQG WKH SULRU SUREDn ELOLWLHV RI WKH VQ DQG Q GLVWULEXWLRQV f WKH SUREDELOLW\ RI D KLW 3VQQf3Qf 7KLV HTXDWLRQ LV D VLPSOH H[SUHVVLRQ RI %D\HVn WKHRUHP ,W FDQ EH H[n WHQGHG WR WKH RGGV IRUP WKH SRVWHULRU RGGV HTXDOV WKH SULRU RGGV WLPHV WKH OLNHOLKRRG UDWLR RI WKH REVHUYDWLRQ [ 3VQ_[f B 3VQf 3[_VQf 3Q_[f 3Qf 3[_Qf

PAGE 183

(JDQ f IXUWKHU GHPRQVWUDWHV WKDW WKH SRVWHULRU SUREDELOLW\ LV VWULFWO\ PRQRWRQH ZLWK WKH OLNHOLKRRG UDWLR RI [ /[f WKDW LV ZKHQ /[f /&;Mf WKHQ 3VQ_[f 3VQ_[f 0D[LPL]DWLRQ RI ([SHFWHG 9DOXH 2EMHFWLYH )XQFWLRQ 7KH GHFLVLRQ IXQFWLRQ IRU PD[LPL]LQJ WKH H[SHFWHG YDOXH GHFLVLRQ JRDO LV (9f 3
PAGE 184

$33(1',; % 25$/ 35(6(17$7,21 72 (/,&,7( 92/817((5 68%-(&76 DVVXPH WKDW PRVW LI QRW DOO RI \RX LQWHQG WR EHFRPH DFFRXQWn DQWV
PAGE 185

f $ FKDQFH WR HDUQ XS WR DV SD\PHQW IRU SDUWLFLSDWLQJ LQ WKH H[SHULPHQW f $ FKDQFH WR JDLQ ILUVWKDQG H[SHULHQFH ZLWK DQ DFFRXQWLQJ UHVHDUFK SURMHFW
PAGE 186

$33(1',; & %$&.*5281' ,1)250$7,21 %22./(7 7KH IROORZLQJ UHSUHVHQWV EDFNJURXQG LQIRUPDWLRQ FRQFHUQLQJ WKH UROH WKDW \RX ZLOO EH DVNHG WR DVVXPH LQ WKH EXVLQHVV H[SHULPHQW IRU ZKLFK \RX KDYH YROXQWHHUHG WR SDUWLFLSDWH 3OHDVH UHDG WKLV LQIRUPDWLRQ VHYHUDO WLPHV IDPLODUL]LQJ \RXUVHOI ZLWK WKH JHQHUDO DVSHFWV RI WKLV UROH 0RUH GHWDLOHG LQIRUPDWLRQ WRJHWKHU ZLWK VSHFLILF LQVWUXFWLRQV ZLOO EH SUHVHQWHG DW WKH LQFHSWLRQ RI WKH DFWXDO H[SHULPHQW
PAGE 187


PAGE 188

DVVHPEO\ GHSDUWPHQW KRZHYHU WKH FRVWV DVVRFLDWHG ZLWK WKH LQVSHFWRUV DUH QRW FKDUJHG GLUHFWO\ WR WKH DVVHPEO\ GHSDUWPHQW EXW UDWKHU WR D FHQWUDO SURGXFW HQJLQHHULQJ DQG TXDOLW\ FRQWURO GHSDUWPHQW 7KH DVVHPEO\ VXSHUYLVRU GRHV QRW DFWXDOO\ SDUWLFLSDWH LQ WKH SK\VLFDO DVVHPEO\ SURFHGXUHV EXW UDWKHU VXSHUYLVHV WKH DVVHPEO\ ZRUNHUV WKH NLW LQSXW WR WKH DVVHPEO\ GHSDUWPHQW WKH SDFNDJHUV DQG WKH ILQLVKHG FKDLU RXWSXW WR WKH SURGXFW LQYHQWRU\ GHSDUWPHQW 7KH RSHUDWLRQDO FRQWURO PDQDJHU \RXUVHOIf KDV RYHUDOO UHVSRQVLn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n PLQHG DVVHPEO\ VHTXHQFH $IWHU DVVHPEOLQJ D IROGLQJ FKDLU WKH ZRUNHU SODFHV LW RQ D FRQYH\RU EHOW ZKLFK WDNHV LW WR WKH TXDOLW\ FRQWURO LQVSHFWRUV 7KH RQVLJKW TXDOLW\ FRQWURO SURFHVV FRQVLVWV RI D YLVXDO LQVSHFWLRQ RI WKH FKDLUnV ILQLVK DQG D PDQXDO LQVSHFWLRQ RI WKH FKDLUn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n PHQW 7KHVH ODERU HIILFLHQF\ VWDQGDUGV DUH UHDVRQDEO\ DWWDLQDEOH DQG DOORZ IRU XQDYRLGDEOH ODERU LQHIILFLHQFLHV DQG UHDVRQDEOH YDULDWLRQ LQ ZRUNHU SHUIRUPDQFH $V WKH RSHUDWLRQDO FRQWURO PDQDJHU \RX VKRXOG DFFHSW WKH ODERU HIILFLHQF\ VWDQGDUGV LQ WHUPV RI IDLU FRQWURO DQG SHUIRUPDQFH JRDOV

PAGE 189

7KH PHQW DUH SHU FKDLU ODERU HIILFLHQF\ VWDQGDUGV IRU WKH DVVHPEO\ GHSDUW 3URFHVV 6WDQGDUG 7LPH $OORZHG $VVHPEO\ 3DFNDJLQJ 7RWDO 7KH YDULDQFH UHSRUW FRQWDLQV YDULDQFH LV UHSRUWHG RQO\ LQ PLQXWHV PLQXWHV PLQXWHV RQO\ WKH ODERU HIILFLHQF\ YDULDQFH DQG WKLV DJJUHJDWH IRUP LH WKH YDULDQFH LV QRW EURNHQ LQWR WKH DVVHPEO\ DQG SDFNDJLQJ SURFHVVHVf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f HIILFLHQF\ ,W LV DOVR SRVVLEOH IRU WKH GHSDUWPHQW WR EH LQFRQWURO ZKLOH DW WKH VDPH WLPH WKH UHSRUWHG ODERU HIILFLHQF\ YDULDQFH LV ODUJH 7KLV ZRXOG RFFXU ZKHQ PRVW RI WKH LQGLYG XDO SURFHGXUHV DUH SHUIRUPHG DW VOLJKWO\ EHORZ QRUPDO HIILFLHQF\ 7KHVH H[DPSOHV UHSUHVHQW WZR H[WUHPH VLWXDWLRQV WKH SUREDELOLW\ RI RFFXUDQFH IRU HLWKHU RI WKHP LV ORZ ZKHQ FRPSDUHG WR WKH SUREDELOLWLHV RI RFFXUn DQFH IRU VLWXDWLRQV WKDW IDOO LQEHWZHHQ WKH WZR H[WUHPH H[DPSOHV
PAGE 190

,I
PAGE 191

([WHQW RI 7RWDO ,QYHVWLn JDWLRQ 'HFLVLRQV &RVW 0LQLPL]DWLRQ &DVK %RQXV WR WR WR WR WR $JDLQ UHO\ RQ \RXU QRW GLVFXVVLQJ WKLV H[SHULPHQW ZLWK RWKHU SHRSOH

PAGE 192

$33(1',; 35,25 ,1)250$7,21 6+((76 ([SHULPHQW &RQGLWLRQ ,6&f 7KH IROORZLQJ LQIRUPDWLRQ LV EDVHG XSRQ SDVW H[SHULHQFH DQG SDVW REVHU YDWLRQV RI WKH DVVHPEO\ OLQH 7KLV LQIRUPDWLRQ PD\ EH KHOSIXO WR \RX LQ PDNLQJ \RXU GHFLVLRQV 7KH SRUWLRQ RI WLPH WKH DVVHPEO\ OLQH ZDV ,QFRQWURO b 2XWRIFRQWURO b 7KH DFWXDO PLQXWHV SHU FKDLU IRU ERWK VWDWHV DUH DVVXPHG WR EH QRUPDOO\ GLVWULEXWHG 7KH PRVW IDYRUDEOH ODERU HIILFLHQF\ YDULDQFH ZDV PLQXWHV SHU FKDLU DFWXDO PLQXWHV LQFXUUHG SHU FKDLUf 7KH PRVW XQIDYRUDEOH ODERU HIILFLHQF\ YDULDQFH ZDV PLQXWHV SHU FKDLU DFWXDO PLQXWHV LQFXUUHG SHU FKDLUf 7KH PD[LPXP LQYHVWLJDWLRQ FRVW ZKHQ WKH DVVHPEO\ OLQH VWDWH ZDV ,QFRQWURO 2XWRIFRQWURO 7KH PLQLPXP LQYHVWLJDWLRQ FRVW ZKHQ WKH DVVHPEO\ OLQH VWDWH ZDV ,QFRQWURO XWRIFRQWUR ([SHULPHQW &RQGLWLRQ ,6&f 7KH IROORZLQJ LQIRUPDWLRQ LV EDVHG XSRQ SDVW H[SHULHQFH DQG SDVW REVHU YDWLRQV RI WKH DVVHPEO\ OLQH 7KLV LQIRUPDWLRQ PD\ EH KHOSIXO WR \RX LQ PDNLQJ \RXU GHFLVLRQV

PAGE 193

7KH WZR DVVHPEO\ OLQH VWDWH PHDQV LQ PLQXWHV SHU FKDLU DUH ,QFRQWURO 2XWRIFRQWURO 7KH WZR DVVHPEO\ OLQH VWDWH VWDQGDUG GHYLDWLRQV LQ PLQXWHV SHU FKDLU ,QFRQWRUO 2XWRIFRQWURO 7KH SRUWLRQ RI WLPH WKH DVVHPEO\ OLQH ZDV ,QFRQWURO b 2XWRIFRQWURO b 7KH DFWXDO PLQXWHV SHU FKDLU IRU ERWK VWDWHV DUH QRUPDOO\ GLVWULEXWHG 7KH PRVW IDYRUDEOH ODERU HIILFLHQF\ YDULDQFH ZDV PLQXWHV SHU FKDLU DFWXDO PLQXWHV LQFXUUHG SHU FKDLUf 7KH PRVW XQIDYRUDEOH ODERU HIILFLHQF\ YDULDQFH ZDV PLQXWHV SHU FKDLU DFWXDO PLQXWHV LQFXUUHG SHU FKDLUf 7KH PD[LPXP LQYHVWLJDWLRQ FRVW ZKHQ WKH DVVHPEO\ OLQH VWDWH ZDV ,QFRQWURO 2XWRIFRQWURO 7KH PLQLPXP LQYHVWLJDWLRQ FRVW ZKHQ WKH DVVHPEO\ OLQH VWDWH ZDV ,QFRQWURO 2XWRIFRQWURO

PAGE 194

$33(1',; ( 68%-(&7 +(85,67,& (/,&,7$7,21 48(67,211$,5( 3OHDVH DQVZHU WKH IROORZLQJ TXHVWLRQV LQ WKH VSDFHV SURYLGHG 7KLV LV QRW D WHVW WKHUH DUH QR ULJKW RU ZURQJ DQVZHUV ,W LV LPSRUWDQW WKDW \RX WU\ WR DQVZHU HDFK TXHVWLRQ KRQHVWO\ ,I \RX QHHG DGGLWLRQDO VSDFH XVH WKH EDFN RI WKLV TXHVWLRQQDLUH :KDW UXOH RU VHW RI UXOHV GLG \RX XVH WR PDNH \RXU LQYHVWLJDWLRQ GHFLVLRQV LQ WKH ODVW ODERU HIILFLHQF\ YDULDQFH UHSRUWV WKH VHFRQG EORFNf" 3OHDVH LQFOXGH DQ\ QXPHULFDO YDOXHV WKDW ZHUH SDUW RI WKLV UXOH RU VHW RI UXOHV 'R \RX IHHO \RX XVHG WKH VDPH UXOH RU VHW RI UXOHV DV \RX JDYH LQ TXHVWLRQ RQH WR PDNH \RXU YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV LQ WKH ILUVW ODERU HIILFLHQF\ YDULDQFH UHSRUWV WKH ILUVW EORFNf" ,I QRW WKH VDPH KRZ GRHV WKH UXOH RU VHW RI UXOHV XVHG LQ WKH ILUVW UHSRUWV GLIIHU IURP ZKDW \RX JDYH LQ TXHVWLRQ RQH" 3OHDVH LQn FOXGH DQ\ QXPHULFDO YDOXHV XVHG DV SDUW RI WKH UXOHV RU VHWV RI UXOHV WKDW ZHUH GLIIHUHQW 5DWH WKH IROORZLQJ LWHPV RI LQIRUPDWLRQ LQ WHUPV RI WKHLU LPSRUWDQFH WR \RX LQ PDNLQJ \RXU YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV GR QRW UDWH WKH nRWKHUn LWHP XQOHVV \RX KDYH VSHFLILHG VRPH RWKHU LQIRUPDWLRQf &LUFOH WKH DSSURSULDWH UHVSRQVH $7KH EDFNJURXQG LQIRUPDWLRQ ERRNOHW fNNOUNNnNnNNnNnNNnNLFnNNnNnNnNnNNnNnNnNnNnNnNnNnNnNnNnNnNnNnNnNnNnNfNNnNnNnNnNnNnNnNnNnNN fN r 1RW 6OLJKWO\ ,PSRUWDQW ,PSRUWDQW 0RGHUDWHO\ 9HU\ ,PSRUWDQW ,PSRUWDQW ([WUHPHO\ ,PSRUWDQW % 7KH LQIRUPDWLRQ EDVHG XSRQ SDVW H[SHULHQFH DQG SDVW REVHUYDWLRQ ZKLFK ZDV SUHVHQWHG LQ WKH IURQW RI HDFK H[SHULPHQW ERRNOHW PNnNNnNNnNNPNnNNnNPNNNLFnNnNnNNNNnNNnNfNNNnNnNnNnNnNnNNnNfNNNNnNnNnNnNnNnNnNNnNnN r r r r r 1RW 6OLJKWO\ 0RGHUDWHO\ 9HU\ ([WUHPHO\ ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW & 7KH VWDQGDUG PLQXWHV DOORZHG SHU FKDLU rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr fN fN r r r 1RW 6OLJKWO\ ,PSRUWDQW ,PSRUWDQW 0RGHUDWHO\ 9HU\ ,PSRUWDQW ,PSRUWDQW ([WUHPHO\ ,PSRUWDQW

PAGE 195

' 7KH DFWXDO PLQXWHV LQFXUUHG SHU FKDLU rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr rrrrr 1RW 6OLJKWO\ 0RGHUDWHO\ 9HU\ ([WUHPHO\ ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ( 7KH ODERU HIILFLHQF\ YDULDQFH SHU FKDLU rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr rrrrr 1RW 6OLJKWO\ 0RGHUDWHO\ 9HU\ ([WUHPHO\ ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ) 7KH WRWDO FKDLUV SURGXFHG rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr 1RW 6OLJKWO\ 0RGHUDWHO\ 9HU\ ([WUHPHO\ ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW *7KH FRVWV DVVRFLDWHG ZLWK LQYHVWLJDWLRQ GHFLVLRQV fNnNnNnNnNnNnNnNnNnNnNNnNnNLUNnNnNnNnNnNnNnNnNnNLUNnNnNnNnNnNnNnNNnNnNNnNLFnNnNnNnNnNnNnNnNnN r r r r r 1RW 6OLJKWO\ 0RGHUDWHO\ 9HU\ ([WUHPHO\ ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW +2WKHU SOHDVH VSHFLI\f rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr rrrrr 1RW 6OLJKWO\ 0RGHUDWHO\ 9HU\ ([WUHPHO\ ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW ,PSRUWDQW 5HFDOOLQJ WKH WUDLQLQJ SKDVH ZKLFK \RX FRPSOHWHG DW DQ HDUOLHU WLPH ZKDW UXOH RU VHW RI UXOHV GLG \RX XVH WR PDNH \RXU YDULDQFH LQYHVWLn JDWLRQ GHFLVLRQV LQ WKH ILUVW YDULDQFH UHSRUWV" 3OHDVH LQFOXGH DQ\ QXPHULFDO YDOXHV WKDW ZHUH SDUW RI WKLV UXOH RU VHW RI UXOHV +RZ GLG \RX LQLWLDOO\ IRUP WKH UXOH RU VHW RI UXOHV \RX JDYH LQ TXHVWLRQ IRXU" :KDW OHDG \RX WR XVH WKLV SDUWLFXODU UXOH RU VHW RI UXOHV" ,I WKH UXOH RU VHW RI UXOHV \RX JDYH LQ TXHVWLRQ IRXU LV GLIIHUHQW IURP WKH UXOH RU VHW RI UXOHV \RX JDYH LQ TXHVWLRQ RQH ZKDW FDXVHG \RX WR PDNH WKH FKDQJH" :KDW RWKHU LQIRUPDWLRQ LWHPV LI SUHVHQWHG GR \RX IHHO ZRXOG LPn SURYH \RX YDULDQFH LQYHVWLJDWLRQ GHFLVLRQV" :HUH \RXU GHFLVLRQV LQIOXHQFHG E\ GLVFXVVLRQV ZLWK RWKHU SDUWLFLn SDQWV LQ WKLV H[SHULPHQW"

PAGE 196

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r r r r r 1R 0RUH 0RUH 0RUH 0RUH 0RUH 0RUH 0RUH 7KDQ 7LPHV 7LPH 7L PHV 7LPHV 7LPHV 7LPHV 7LPHV 'LG \RXU GHVLUH WR SHUIRUP ZHOO LQ XQGHUWDNLQJ D FKDOOHQJLQJ WDVN FDXVH \RX WR WU\ YHU\ KDUG" rrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr N r r r 'HILQLWHO\
PAGE 197

'LG \RX IHHO WHQVH GXULQJ WKH H[SHULPHQW" NNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNNN N N N N N N N 'HILQLWHO\
PAGE 198

%,%/,2*5$3+< $O SHUW 0 DQG 5DLIID + $ 3URJUHVV 5HSRUW RQ WKH 7UDLQLQJ RI 3UREDELOLW\ $VVHVVRUV 8QSXEOLVKHG PDQXVFULSW +DUYDUG 8QLYHUVLW\ $PHULFDQ $FFRXQWLQJ $VVRFLDWLRQ $ 6WDWHPHQW RI %DVLF $FFRXQWLQJ 7KHRU\ (YDQVWRQ ,OO 7KH $VVRFLDWLRQ 5HSRUW RI WKH &RPPLWWHH RQ 0DQDJHULDO $FFRXQWLQJ 7KH $FFRXQWLQJ 5HYLHZ 6XSSOHPHQW f $QWKRQ\ 5 1 3ODQQLQJ DQG &RQWURO 6\VWHPV $ )UDPHZRUN IRU $QDO\VLV %RVWRQ +DUYDUG 8QLYHUVLW\ 3UHVV $SSHOEDXP 0 DQG &UDPHU ( 0 6RPH 3UREOHPV LQ WKH 1RQ RUWKRDRQDO $QDO\VLV RI 9DULDQFH 3V\FKRORJLFDO %XOOHWLQ f %DQNV : 3 6LJQDO 'HWHFWLRQ 7KHRU\ DQG +XPDQ 0HPRU\ 3V\FKRORJLFDO %XOOHWLQ f %DUHILHOG 5 0 7KH (IIHFW RI $JJUHJDWLRQ RI 'HFLVLRQ 0DNLQJ 6XFFHVV $ /DERUDWRU\ 6WXG\ -RXUQDO RI $FFRXQWLQJ 5HVHDUFK f %DUU $ *RRGQLJKW + 6DLO 3 DQG +HOZLJ 7 $ 8VHUnV *XLGH WR 6$6 5DOHLJK 1& 6$6 ,QVWLWXWH ,QF %HUQEDFK + $ 'HFLVLRQ 3URFHVVHV LQ 0HPRU\ 3V\FKRORJLFDO 5HYLHZ f %LHUPDQ + )RXUDNHU / ( DQG -DHGLFNH 5 $ 8VH 3UREDELOLW\ DQG 6WDWLVWLFV LQ 3HUIRUPDQFH (YDOXDWLRQ 7KH $FFRXQWLQJ 5HYLHZ f &RRPEV & + 'DZHV 5 0 DQG 7YHUVN\ $ 0DWKHPDWLFDO 3V\FKRORJ\ $Q (OHPHQWDU\ ,QWURGXFWLRQ (QJOHZRRG &OLIIV 13UHQWLFH+DOO ,QF

PAGE 199

'HPVNL 6 $QDO\]LQJ WKH (IIHFWLYHQHVV RI WKH 7UDGLWLRQDO 6WDQGDUG &RVW 9DULDQFH 0RGHO 0DQDJHPHQW $FFRXQWLQJ f 7KH 'HFLVLRQ ,PSOHPHQWDWLRQ ,QWHUIDFH (IIHFWV RI $OWHUQDn WLYH 3HUIRUPDQFH 0HDVXUHPHQW 0RGHOV 7KH $FFRXQWLQJ 5HYLHZ f ,PSOHPHQWDWLRQ (IIHFWV RI $OWHUQDWLYH 3HUIRUPDQFH 0HDn VXUHPHQW 0RGHOV LQ D 0XOWLYDULDWH &RQWH[W 7KH $FFRXQWLQJ 5H YLHZ f 'LWWPDQ $ DQG 3UDNDVK 3 &RVW 9DULDQFH ,QYHVWLJDWLRQ 0DUNRYLDQ &RQWURO RI 0DUNRY 3URFHVVHV -RXUQDO RI $FFRXQWLQJ 5HVHDUFK f 'RSXFK 1 %LUQEHUJ DQG 'HPVNL 6 $Q ([WHQVLRQ RI 6WDQGDUG &RVW 9DULDQFH $QDO\VLV 7KH $FFRXQWLQJ 5HYLHZ f 'X&KDUPH : 0 $ 5HVSRQVH %LDV ([SODQDWLRQ RI &RQVHUYDWLYH +XPDQ ,QIHUHQFH -RXUQDO RI ([SHULPHQWDO 3V\FKRORJ\ f DQG 3HWHUVRQ & 5 ,QWXLWLYH ,QIHUHQFH $ERXW 1RUPDOO\ 'LVWULEXWHG 3RSXODWLRQV -RXUQDO RI ([SHULPHQWDO 3V\FKRORJ\ f '\FNPDQ 7 5 7KH ,QYHVWLJDWLRQ RI &RVW 9DULDQFHV -RXUQDO RI $FFRXQWLQJ 5HVHDUFK f (JDQ 3 6LJQDO 'HWHFWLRQ 7KHRU\ DQG 52& $QDO\VLV 1HZ
PAGE 200

.DKQHPDQ DQG 7YHUVN\ $ 6XEMHFWLYH 3UREDELOLW\ $ -XGJHPHQW RI 5HSUHVHQWDWLYHQHVV &RJQLWLYH 3V\FKRORJ\ f .DSODQ 5 6 2SWLPDO ,QYHVWLJDWLRQ 6WUDWHJLHV ZLWK ,PSHUIHFW ,QIRUn PDWLRQ -RXUQDO RI $FFRXQWLQJ 5HVHDUFK f 7KH 6LJQLILFDQFH DQG ,QYHVWLJDWLRQ RI &RVW 9DULDQFHV 6XUYH\ DQG ([WHQVLRQV -RXUQDO RI $FFRXQWLQJ 5HVHDUFK f .RHKOHU 5 : 7KH 5HOHYDQFH RI 3UREDELOLW\ 6WDWLVWLFV WR $FFRXQWLQJ 9DULDQFH &RQWURO 0DQDJHPHQW $FFRXQWLQJ f /HZLV & DQG .HUHQ
PAGE 201

6LPRQ + $ 7KH 1HZ 6FLHQFH RI 0DQDJHPHQW 'HFLVLRQ 1HZ
PAGE 202

:HLVVPDQ 6 0 +ROOLQJZRUWK 6 5 %DLUG & 3V\FKRSK\VLFDO 6WXG\ RI 1XPEHUV ,,, 0HWKRGRORJLFDO $SSOLFDWLRQV 3V\FKRORJLFDO 5HVHDUFK f =DQQHWRV = $ 6WDQGDUG &RVW DV D )LUVW 6WHS WR 3UREDELOLVWLF &RQWURO $ 7KHRUHWLFDO -XVWLILFDWLRQ DQG ([WHQVLRQ DQG ,PSOLFDWLRQV 7KH $FFRXQWLQJ 5HYLHZ f =LJOHU ( 0HWDWKHRUHWLFDO ,VVXHV LQ 'HYHORSPHQWDO 3V\FKRORJ\ ,Q 7KHRULHV LQ &RQWHPSRUDU\ 3V\FKRORJ\ 0 0DU[ (Gf 1HZ
PAGE 203

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n WUROOHU DQG YLFHSUHVLGHQW RI ILQDQFH )URP 6HSWHPEHU XQWLO WKH SUHVHQW 0U %URZQ KDV DWWHQGHG WKH 8QLYHUVLW\ RI )ORULGD ZKHUH KH KDV VWXGLHG IRU D 'RFWRU RI 3KLORVRSK\ GHJUHH LQ EXVLQHVV DGPLQn LVWUDWLRQ DFFRXQWLQJf :KLOH DWWHQGLQJ WKH 8QLYHUVLW\ RI )ORULGD 0U %URZQ ZDV WKH UHFLSHQW RI WKH 3ULFH :DWHUKRXVH )RXQGDWLRQ *UDQW WKH $PHULFDQ $FFRXQWLQJ $VVRFLDWLRQ )HOORZVKLS *UDQW DQG WKH +DVNLQ DQG 6HOOV )HOORZVKLS *UDQW 0U %URZQ ZDV HOHFWHG E\ WKH IDFXOW\ DV WKH XQLYHUVLW\ UHSUHVHQWDWLYH WR WKH $PHULFDQ $FFRXQWLQJ $VVRFLDWLRQ 'RFWRUDO &RQVRUWLXP

PAGE 204

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f67LRZ,LDO7A $VVLVWDQW 3URIHVVRU RI $FFRXQWLQJ FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG WKDW LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWDEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ 4X4M $ *ULJJV A 5LFKDUG $VVLVWDQW 3URIHVVRU RI 3V\FKRORJ\

PAGE 205

7KLV GLVVHUWDWLRQ ZDV VXEPLWWHG WR WKH *UDGXDWH )DFXOW\ RI WKH 'HSDUWPHQW RI $FFRXQWLQJ LQ WKH &ROOHJH RI %XVLQHVV $GPLQLVWUDWLRQ DQG WR WKH *UDGXDWH &RXQFLO DQG ZDV DFFHSWHG DV SDUWLDO IXOILOOPHQW RI WKH UHTXLUHn PHQWV IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ $XJXVW 'HDQ *UDGXDWH 6FKRRO


TABLE 4 Continued
Hypotheses
Analysis Test
Method(s) Statistic(s)
4.3 [(G^|S1,C1)-(G-¡|S2,C1)] Model comparison F=7.7124 (pc.Ol)
range test
4.4a (GPi/GN-j | Cl) <
Sources F value F=0.35 (n.s.)
(GP -¡/GN | C2)
One-tailed Z test Z=2.0198 (p<.01)
4.4b [ (GP -j/GN -j | C2, S2) -
Model comparison F=8.7438 (pc.Ol)
(GPi/GNi|Cl,S2)J
Duncan's multiple (a=.01)
<[(GP-j/GN-j |C2,S1)-
range test
(GPj/GN-j | Cl, SI)]
4.5 (Gj|PP)<(Gj|MM) One-tailed t test
t=l.303 (p<.10)
Results
There was a significant cost by distribu
tion interaction on the relative net
decision costs in which the C2 cost level
had a greater effect than the Cl cost
level while the distribution levels had
little effect.
The mean relative additional decision
costs and relative additional decision
savings ratio of those subjects within
the C2 cost level was larger than that
of those subjects within the Cl cost
level.
There was a significant cost by distribu
tion interaction on the relative addi
tional decision costs and relative addi
tional decision savings ratio in which
the cost levels had a significant effect
only when given the SI distribution level.
The mean relative net decision costs of
those subjects within the constantly
decreasing training session performances
category was significantly smaller than
that of those subjects within the con
stantly increasing training session per
formances category.


5
deciding on the information set to be presented within the cost variance
report have been advocated. First, the accountant can determine those
models which managers use in making variance investigation decisions and
provide an information set which would permit the implementation of
these models. Problems with this individual model approach are 1) the
possibility exists that different individual models use a wide range of
different information sets (which are not costless), and 2) the indi
vidual models may not be optimal (i.e., they could be inefficient with
respect to other decision models). The individual model approach cen
ters on the psychological question of what the manager is doing (or
would do) with the available (or additional) information. Second, the
accountant could create a normative model of the variance investigation
decision and provide the information required by that model. Problems
with the normative model approach are 1) the information provided may
not optimize the individual's investigation decisions (which may con
tinue to be made using the individual's model), and 2) the costs
associated with operationalization of the normative model may exceed
the benefits (cost savings as a result of more optimal investigation
decisions and greater congruence of manager goals with overall organi
zation goals) offered by the model. The normative model approach
centers on the analytical question of what the manager should be doing
(or should do) with the available (or additional) information.
A substantial portion of the accounting variance investigation
literature has focused upon the second approach-- the normative model
approach.4 However, rarely has attention been given to how the manager
4This literature is reviewed within Chapter II.


152
An effect which differed from expectations was that the mean
initial decision anchor given the C2 cost level was significantly
larger than the mean initial decision anchor given the Cl level
(equality was predicted). A bias of the estimates of the initial de
cision anchors may explain this difference. It will be recalled that
the estimates had incorporated some initial adjustment (during the
first training session). Those subjects within the C2 cost level were
expected to adjust their initial decision anchors such that the value
(actual minutes incurred) increased: those subjects within the Cl cost
level were expected to adjust in the opposite direction. Initial ad
justment in the appropriate directions would produce the obtained cost
variable effect on the estimates of the initial decision anchors.
The second aspect of the anchoring and adjustment heuristic is
the adjustment process. The adjustment process was examined using a
measure of the individual's relative decision anchor adjustment. This
observation variable, RA-, was the linear distance ever which the
individual adjusted his decision anchor (the distance between his final
decision anchor and the optimal decision value) relative to the linear
distance over which the individual should have adjusted his decision
anchor (the distance between his initial decision anchor and the
optimal decision value). The results of analysis indicated that the
RA.j measure did not differ significantly between the levels of the
various situation variables. This was consistent with the predictions
from the training phase simulation. A major limitation of the measure
is that it is confined to the relative linear movement along the random
variable decision axis. As such the RAj measure is not sensitive to
the learning of various situation parameters and variables (e.g., the


4
labeled a heuristic. Within this context, heuristic refers to the
learned set of rules or principles that are utilized by the individual
in making the particular decisions required of him. The manager's
specific heuristic can be affected substantially by his individual
characteristics (e.g., intelligence, cognitive complexity and style,
decision process sensitivity, motivations, etc.). However, a general
heuristic, labeled anchoring and adjustment (Tversky and Kahnemar,
1974), is expected to describe the general form of the manager's vari
ance investigation decision process.3
The above examination of the manager's variance investigation
decision process indicates that the variables which can affect both
the process and the results of the process include 1) the structure
of the decision situation, 2) the contents of the available information
set, 3) the manager's information processing efficiency, and 4) the
manager's learning efficiency (from his experiences with the controlled
process). The manager's knowledge concerning the structure of the
decision situation may be limited by the contents of the available
information set and by his information processing and learning effici
encies. The accountant, to a large extent, has control over the con
tents of the available information set.
The Research Objectives
Many problems in accounting reduce to one of choosing among
alternative information sets that could be provided to a decision maker
(American Accounting Association, 1972). Two basic approaches to
3The anchoring and adjustment heuristic is described in greater detail
within Chapter II.


170
efficiency (decision criteria conservatism) is a characteristic of only
training (learning) situations or remains a characteristic of
well-trained (non-learning) situations.
Extensions involving the experimental environment could take
several forms. First, they may relate to the levels of the situation
variables. Only certain situation variables were studied within this
research: extensions could study the affects of manipulating different
situation variables at different levels (e.g., more than two states of
nature, non-normal distributions, variable relative frequencies, vari
able decision cost functions, multiple standard cost variances).
Second, extensions could be made in relation to the experimental
task itself. Within this research the decision task was of a
cross-sectional nature: i.e., each decision trial was independent of
the previous decision trials. An extension could involve a decision
task of a time-series nature: i.e., where each decision trial would
be dependent on the previous decision trials. Such a decision task
extension would involve sequential decision making, where the physical
process is initially in some state of nature and over a series of
decision trials (variance reports) the process may or may not move to
another state of nature. The task of the decision maker would be to
detect the point in time in which the physical process shifts to an
other state.
The third and final extension discussed within this section con
cerns the nature of the experimental decisions. Within this research
the decisions involved the detection of state of nature for the purpose
of physical process control within cost minimization constraints. Be
yond the operational control level, standard variance reports can be


CHAPTER VI
DISCUSSION OF RESULTS
Conceptual Development Revisited
A brief restatement of the conceptual development presented in
Chapter III will serve as a basis for discussion of the results that
were reported in Chapter V. It was posited that an individual's
long-run decision efficiency would be affected by 1) the structure
of the particular decision situation, 2) the contents of the available
information set, 3) the individual's efficiency in processing the
available information, and 4) the individual's ability to expand the
available information through experience with the particular situation.
A general heuristic was expected to systematically account for a por
tion of the decision behavior of the subjects. A heuristic, within
this context, refers to a learned set of rules or principles which are
utilized by individuals in making the particular decisions required of
them. Since a heuristic is a learned response to some decision task
stimuli, individual characteristics can have considerable effect upon
the specific form of the heuristic. The conceptual development, how
ever, proposed that specific decision task stimuli would give rise to
general forms of heuristics which would be (at least partially) in
dependent of individual characteristies.
The decision task utilized in this research required the discrim
ination between the uncertain existence of two states of nature
146


42
BNC.J approaches either positive infinity or a value of one as the extent
of conservatism increases, approaches a value of zero as the extent of
conservatism decreases, and approaches either a value of negative one
or negative infinity as the extent of non-conservatism increases.
Initial and final decision anchor variables
The relative relationships between an individual's initial and
final decision anchors can be measured in terms of the relative linear
distance between these two decision anchors. A measure of the relative
adjustment for individual i, labeled RA^, is conditional on the direc
tion of adjustment along the relevant decision axis. This direction
of adjustment is in turn conditional on the level of the cost variable.
Given the Cl cost level the initial decision anchor is greater than the
optimal model cutoff value, and given the C2 cost level the initial
decision anchor is less than the optimal model cutoff value. The RA.
measure is defined as:
RAi = (EDVi 0DVk) / (TDVi 0DVk) ,given the Cl cost level;
RA.. = (0DVk EDV.¡) / (0DVk TDV.¡) ,given the C2 cost level,
where EDV^ = individual i's final (experiment) decision anchor;
TDV. = individual i's initial (training) decision anchor; and
0DVk = optimal model k's final (experiment) cutoff value.
The RA.j measure has the following relationships: 1) if no
adjustment occurs between the training and the experiment phases the
measure will equal a value of one, 2) if the adjustment is in the wrong
direction (away from the optimal model's experiment cutoff value) the
measure will be greater than a value of one, 3) if the adjustment is in


120
(px.01) for each of the situation variables found to be significant by
the above sources F values. The sole exception was the distribution
variable within the BN^ measure. The means of the distribution levels
within the BHj variable do not differ significantly (Z=0.8830, n.s.
two-tailed).
Hypothesis 3.1a predicted that (BNA^jll) would not differ signifi
cantly from (BNA-112). A Z test indicated that the difference was not
significant (Z=0.7891, n.s. two-tailed). Hypothesis 3.1b predicted that
(BNA.|SI) would not differ significantly from (BNA^|S2). Again, a
Z test showed the difference to be not significant (Z=0.7048, n.s.
two-tailed). Hypothesis 3.1c predicted that (BNA-|Cl) would be signifi
cantly larger than (BNA-|C2). The model comparison procedure and Z test
indicated a significant difference at the p<.01 level, with the Cl level
having the larger mean. Hypothesis 3.2 predicted that the variance of
(BNA. |Cl) would be significantly smaller than the variance of (BNA1-|C2).
The F test of equal variances resulted in a finding of a significant
difference at the p<.01 level, with the C2 level having the smaller
variance.
Hypothesis 3.4a predicted that (BNA^-BN^|S1) would be signifi
cantly smaller than (BNABN.¡ |S2). A Z test of equal means indicated
that the difference was not significant (Z=0.3892, n.s. one-tailed):
however, the obtained direction is the same as the predicted direction
of this effect. Hypothesis 3.4b predicted that (BNAi-BNi¡C1) would be
significantly smaller than (BNA BN .j|C2). The model comparison pro
cedure and Z test indicated that the difference was significant at the
p<-01 level, with the C2 level having the smaller mean.


84
Assignment of Subjects to Treatment Conditions
Since each of the 86 subjects was assigned to one of eight groups,
randomization per se can not be relied upon to control for individual
attribute differences between groups. An alternative is to block the
randomization process on-indi vidual attribute dimensions assumed to
significantly affect the subject's information processing within the
task required by the experiment.
The literature relating to individual attributes has not pro
duced conclusive results concerning individual attributes effect on
decision processes. The most comprehensive study is that of Taylor and
Dunnette (1974) who analyzed the effect of 16 decision maker attributes
upon eight measures of predecisional, decision-point, and post decisional
processes within a personnel decision simulation experiment. Of con
siderable interest is that the relevant decision maker attributes
accounted for a generally small portion of the total decision variance
within the various processes (the range was from 8 percent to 33 percent).
The decision maker attribute with the greatest predictive capacity was
intelligence. In two of the predecisional processes (diagonosticity
and information processing rate) and in two of the decision-point and
post decisional processes (information retention and decision accuracy),
intelligence accounted for from one-half to almost all of the variance
explained by individual characteristics.
In the present study, the randomization process of assigning
subjects to treatment conditions was blocked on individual intelligence.
Ideally, individual intelligence should be measured using some validated
instrument (e.g., the Wesman Personnel Classification Test or the


CHAPTER I
INTRODUCTION AND OBJECTIVES
Business organizations generate internal accounting reports for
use by managers for purposes of evaluation and decision making.1 One
important evaluation process and decision task of management is that of
standard cost variance analysis and investigation. Based in part upon
standard cost variance reports generated by the internal accounting
system, managers estimate the likelihoods that various production pro
cesses remain under control and decide whether to investigate particu
lar variances. The present study concerns the effects of selected
variables on standard cost variance investigation decision making by
operational control managers. This chapter discusses general aspects
of the standard cost variance investigation decision process and the
variables which may affect the process and presents the research
objectives of the study.
Variance Investigation Decision Processes
An important factor in the evaluation and use of standard cost
variance reports is the manager's perception of the validity of the
xThe American Accounting Association 1966 Statement of Basic Accounting
Theory states that "the objective of accounting for internal use is to
provide information to persons within an organization that enables them
to make informed judgements and effective decisions which further the
organization's goals" (p. 38).
1


156
anchors adversely affects decision performance (as reported by the feed
back measure used in this research), their increased use would affect
intrinsic motivation through the performance satisfaction modifier.
Greater use of multiple decision anchors would result in poorer decision
performance feedback measures: these measures, in turn, could lower
the subject's decision performance satisfaction and thereby lower his
intrinsic motivation. As the variable response range increases in
relative size the number of decisions falling within this range increases.
This requires more guesswork or more complicated decision rules on the
part of the subjects, thereby increasing the difficulty (or the uncer
tainty) of the decision task. This increased task difficulty could
affect the subject's enjoyment of the decision task and thus lower his
intrinsic motivation.
Individual Decision Criteria
Individual decision criteria were measured using a parameter of
the TSD model, 6. This parameter measures the decision criteria used
by the individual independent of his decision model sensitivity. The
factors which can affect an individual's decision criteria are 1) the
individual's efficiency in processing (or learning) the effects of the
states' relative frequencies and the relative decision error costs,
and 2) the use of multiple decision anchors. To isolate these factors
several decision criteria measures were developed: 1) the BN- measure
(which is affected by both factors), 2) the BNA.¡ and BNC^ measures
(which are affected by only the individual's information processing
efficiency), and 3) the BNA^-BM^ measure (which is affected by multiple
decision anchors).


80
benefits that would accrue to those who volunteered. The subjects were
told that they would be given a cash payment for participating in the
experiment and that the amount of a subject's payment would depend upon
his performance. The minimum payment was set at $2.00 and the maximum
at $10.00.
Experimental Materials
The experimental materials included a background information book
let, variance investigation decision stimuli, heuristics questionnaire,
and motivations questionnaire.
Background Information
A background information booklet (see Appendix C) was designed
to provide the subjects with a common experimental environment. The
booklet provided the subject with general company information, general
product information, general manufacturing process information, and
specific assembly department information. The specific assembly depart
ment information included information concerning the employees, the
physical process, the accounting control system, the subject's task as
the operational manager, and the subject's performance evaluation as
the operational manager.
Variance Investigation Decisions
Various information constant over all decision trials within a
treatment condition was presented on a separate page prior to the start


124
The Z test of equal means indicated that the means of the cost levels
differ significantly (Z=2.5598, p<.05 two-tailed), with the C2 level
having the smaller mean.
Hypothesis 3.3a predicted that (BNC^|I1) would not differ signifi
cantly from (BNC.¡ 112). A Z test indicated that the difference was not
significant (Z=0.6802, n.s. two-tailed). Hypothesis 3.3b predicted that
(BNC^ | SI) would not differ significantly from (BNC.j|S2). Again, a Z
test indicated that the difference was not significant (Z=0.3295, n.s.
two-tailed). Hypothesis 3.3c predicted that (BNC^|Cl) would be signifi
cantly smaller than (BNC1-|C2). The model comparison procedure and Z
test indicated that the difference was significant at the p<.05 level,
with the C2 level having the smaller mean.
%
Individual Attributes and Decision Criteria
-
The model comparison procedure results indicated that the
extrinsic (monetary) motivation factor and the GPA^ variable had signifi
cant effects upon the various individual decision criteria variables.2
The analysis of these effects also required additional tests. A product
moment correlation was computed for the motivation factor and the GPA.¡
variable with each of the decision criteria dependent variables. The
results of these correlations are presented in Table 18.
The negative association between extrinsic (monetary) motivation
and the BN^ variable indicates that as the BNj measure increases the
subjects' extrinsic (monetary) motivation measures decrease. The BN^
2See the last sections in this chapter for definitions and for additional
analyses concerning the extrinsic (monetary) motivation factor and the
GPA.j variable.


36
1) the information variable, 2) the distribution variable, and 3) the
cost variable. As indicated previously, each is varied across two levels.
The individual process variable types, measured on a continuous scale,
include the following: 1) individual decision model sensitivity, 2)
individual decision criteria, and 3) relative initial and final decision
anchor relationships. The major dependent variable is the individual
long-run decision efficiency (in terms of incurred costs). The
following discussion describes each of the variables or variable types
as they are employed in this study.
The information variable
The first level, labeled Jl_, is derived from the set of infor
mation assumed to come from individual experience with the physical
system. It includes the following items: 1) the historically derived
portion of time in which the process has been found to fall in each of
the two states, 2) the assumption that the random variable of interest
(actual minutes incurred per chair or its associated standard vari
ance) is normally distributed for both states, 3) the lowest observed
value of the random variable, 4) the highest observed value of the
random variable, 5) the maximum costs associated with each state, and
6) the minimum costs associated with each state.
The second level of the information variable, labeled 12^, includes
additional distributional information. In addition to the six items
contained in the II information set, the following two items are
included: 1) the mean of the random variable within each state, and
2) the standard deviation of the random variable within each state.


79
decision costs incurred by the subject over the series of variance re
ports and dividing this sum by the sum of the total investigation de
cision costs incurred by an optimal model over the same series of vari
ance reports. The subject's cash bonus function, constant for all sub
jects, was inversely related to this measure. As TIDCmi-n approached
one, the payoff approached the maximum: as TIDCmi-n became larger than
one, the payoff approached the minimum.
Subjects
The subjects were 86 senior year undergraduate and master's level
graduate students enrolled in the business college at the University of
Florida. The subjects participated in the experiment during a two week
period; 47 participated in the first week and 39 participated in the
second week. A total of 92 subjects initially volunteered to participate
but six subjects failed to complete the experiment. The 86 subjects
who completed the experiment consisted of 63 males and 23 females.
Three subject selection criteria were applied: 1) the subject
must have completed an intermediate-level managerial accounting course,
2) the subject must have completed an introductory-level statistics
course, and 3) the subject must have earned an overall grade point
average (GPA) of at least 2.0 on a 4.0 scale.
Subjects initially were contacted within senior level and graduate
accounting classes. The contact was made by the experimenter giving a
brief oral presentation followed by passing sign-up sheets around each
class (a copy of the oral presentation is presented in Appendix B). To
motivate volunteering, the presentation focused on the student's pro
fessional responsibilities as future accountants and on the monetary


150
all subjects: the overall mean RA- was 0.641 which indicated that the
subjects adjusted their decision anchors only 35.9 percent (1-RA^) of
the distance between their initial decision anchors and the optimal
model's decision value. However, even given this level of overall deci
sion criteria conservatism, the mean relative decision costs (G.¡) over
all subjects was 0.06 (the average subject's total decision costs were
6.0 percent greater than the optimal model's total decision costs).
The deviation of the G.¡ measure from a value of zero has been
attributed to the following factors: 1) variable response ranges,
2) multiple decision anchors, and 3) decision criteria conservatism.
Although each factor was affected differently by various situation
variables and individual attributes, the overall effects of these
factors on the relative decision costs indicate their relative import
ance. The overall effect of variable response ranges was shown to be
negligible. Ignoring the variable response range effects, the overall
relative decision efficiency measure (G-) can be adjusted into sub
measures which are affected by only one of the two remaining factors.
Eliminating the effects of multiple decision anchors, the adjusted
Gj measure was 0.045. This indicated that of the total deviation of
the subjects' decision costs from those of the optimal models, approxi
mately 75 percent was due to individual decision criteria conservatism
and approximately 25 percent was due to multiple decision anchors.
Discussion of Results
The following discussion of results is organized according to
decision processes and dependent variables. The discussion includes
decision anchor adjustment, individual decision model sensitivity,


163
subjects' GPAs and the relative decision costs: as the subjects' GPAs
increased either additional decision costs decreased or additional de
cision savings increased. It would be reasonable to assume that both
the DN. measure and the subject GPA are modifying variables on a sub
ject's decision cost performance. Since pretests were not administered
it was not possible to state the specific relations between the moti
vation factors and a subject's relative decision cost performance.
Li mitations
The major limitations of this study involve two general aspects;
the subjects employed within the experiment and the experimental
environment itself. Although these aspects are discussed separately,
in many instances the limitations they impose overlap one another.
These limitations affect primarily the external validity of this study.
There are several possible limitations involving the experimental
environment. First, the precision of subject performance feedback
during the training phase could be a limitation. The results obtained
in this study might be modified substantially if such accurate feedback
was not employed. The lack of decision performance difference between
the levels of the available information variable could be a direct
result of this feedback: i.e., the accuracy of the performance feedback
and its relationship with the optimal model may have replaced the need
for such additional statistical information.
Second, the background of the subjects in relation to the ex
perimental task is a possible limitation. Although the subjects re
ceived training in the experimental task, the primary source of their
knowledge concerning standard cost variance investigation could come


9
A conceptual framework of managerial decisions that synthesizes
these dimensions is presented by Gorry and Morton (1971). Simon's
dimension is modified by replacing the terms "programmed" and "non-
programmed" with the terms "structured" and "unstructured" and adding
a third category labeled semi-structured. Gorry and Morton's conceptual
framework is used in this research as a model that will permit the
identification of a standard cost system within the overall managerial
decision framework.
Management Task Planning and Control
Standard cost systems present information sets to managers to
support various decisions concerning task planning and control.
Demski (1967) provides an excellent conceptual discussion of the manage
ment task planning and control process, and his approach (with some
modification) is adopted in this research.
Figure 1 presents a model of the management task planning and
control process. Environmental information, largely external to the
standard cost system, facilitates the planning of overall goals and
policies within the constraints imposed by the environment. Overall
objective control feedback 1) provides evidence for the ccntinuted
validity of overall assumptions made in forming the organization ob
jectives, and 2) facilitates strategic planning evaluation of organi
zation performance.
Both the environmental variables and the overall task control
feedback of the management control activity 1) provide evidence for
the continued validity of the overall task assumptions made in forming


165
Finding the proposed concepts to hold within the experimental world
does not validate the claim that they hold within the real world. On
the other hand, finding the proposed concepts not to hold within the
experimental world would cast serious doubt on whether they hold in the
real world. Positive experimental results are a form of negative
assurance: the concepts have passed an initial empirical test and
remain (albeit somewhat scarred in most instances) potential candidates
for explaining real world phenomena.
The major limitation involving the subjects employed within this
research is that they were college students, predominantly accounting
majors. The obtained results and the conclusions based upon these
results are not generalizable beyond some unspecified population of
which the subjects are representative.
The selection of the type of subject to be employed within an
experiment is not independent of the experimental environment. Within
this research principal interest lies in the behavior of operational
managers within real world environments. However, given the selection
of the laboratory experimentation method (which implies the use of a
surrogate experimental environment), the selection of the type of
subject depends upon the answer to the question, "Which subject is a
better surrogate for a manager in a real world environment, a know
ledgeable student or the manager himself?" There are, of course, no
clear answers to this question. However, several aspects of the
situation favor the use of a knowledgeable student. First, the high
degree of abstraction within the experimental environment could have
made the task appear trivial to a manager and could have elicited
behavior not consistent with the principal situation. Second, the


APPENDIX B. ORAL PRESENTATION TO ELICITE VOLUNTEER SUBJECTS .... 174
APPENDIX C. BACKGROUND INFORMATION BOOKLET 176
APPENDIX D. PRIOR INFORMATION SHEETS 182
APPENDIX E. SUBJECT HEURISTIC ELICITATION QUESTIONNAIRE 184
APPENDIX F. SUBJECT MOTIVATION QUESTIONNAIRE 186
BIBLIOGRAPHY 188
BIOGRAPHICAL SKETCH 193


126
measure approaches infinity as the distance between the frj measure and
the 3^ measure increases. A similar association exists between the
motivation factor and the BNA-j measure (the BNA-j measure eliminates the
effects of multiple cutoff values). The BNA^-BN^ variable, which
measures the effect of multiple cutoff values upon individual decision
criteria, has no significant association with the motivation factor. The
BNC.J variable, which measures the effect of conservatism upon individual
decision criteria, has a stronger negative association with the moti
vation factor. As the level of conservatism increases (BNC.¡ increases)
subjects' extrinsic (monetary) motivation measures decrease.
The negative association between the GPA.¡ variable and the BNA^-BN^
variable indicates that as the BNA--BN.¡ variable increases the subjects'
GPA.¡s decrease (the BNA^-BN^ measure increases as multiple cutoff value
usage increases). The negative association between the GPA.¡ variable
and the BNC.¡ variable indicates that as the subjects' GPA^s decrease
the level of conservatism increases.
Individual Long-Run Decision Efficiency
The following sections report the analyses and results of the
individual long-run decision efficiency variables. The variables
include the relative decision costs, the relationships between individual
attributes and long-run decision efficiency, and the relationships be
tween long-run decision efficiency and training phase performance.
Relative Decision Costs
Hypotheses 4.1, 4.2, and 4.3 relate to the decision costs incurred
by the individuals relative to the decision costs incurred by the optimal


158
standard could have acted as an intervening variable which increased
the level of relative conservatism (reduced the level of relative infor
mation processing efficiency). As subjects' adjustments within the C2
cost level diverged from the standard no such subjective adjustment
limit existed, thus the level of relative conservatism could have de
creased.
The higher (statistical) moments of the BNC.¡ measure given the
levels of the cost variable were consistent with this concept of a
subjective adjustment limit. The difference in variances of the
BNC.j measure was due to the difference in the ranges of the measure.
The range of (BNC^|C1) was 4.67 and the range of (BNC.¡|C2) was 1.57.
The subjective adjustment limit should have had the effect of skewing
the measure away from those values which indicated lower levels of
conservatism. The third moment (as expressed by the coefficient of
skewness) of the BNC^ measure indicated that 1) the skewness of the
(BNC^. ¡C1) distribution was positive (skewed away from values which
indicated lower levels of conservatism), and 2) the skewness of the
(BNC^|C2) distribution was negative (skewed toward values which in
dicated lower levels of conservatism).
The multiple decision anchor (the BNA--BN^ measure) was affected
primarly by the cost variable. Those subjects within the Cl cost
level employed multiple decision anchors to a larger extent than did
those subjects within the C2 cost level. The hypotheses posited a
relation between multiple decision anchors and the distance separating
the lower tail of the in-control state from the subject's final de
cision anchor. This relation was similar to that proposed in the
DNA^-DN^ measure. However, the results did not support the hypotheses


164
from the college classroom. Consequently, if they were not taught
(within the classroom) that situations exist in which investigation
decision values are located relatively close to the standard, then
greater decision criteria conservatism within these situations could
be the result of the lack of such knowledge. This suggests the
possibility of an availability bias (Tversky and Kahneman, 1973).
However, to the extent a manager must learn from his own experiences,
such a bias could exist in the real world.
A final limitation (discussed within this section) involves the
selection of the levels of the situation variables. Several situation
variables were set at a single level, while others were manipulated at
multiple levels. The important point is that only specific combina
tions of variable levels were studied within this research whereas an
infinite number of combinations are possible. Different variable levels
and different variable manipulations would create a difference in the
experimental environment which could produce results other than those
obtained in this study.
These experimental environment limitations do not mean, however,
that the concepts proposed within this study have not passed a meaning
ful empirical test. The objectives of experiments involving theory
or concept testing, and the relations between the nature created in the
experiment and the nature existing in the world at large are discussed
by Zigler (1963):
What the experimenter is saying is that if such and such holds
in the real world because of the principles expounded in the
particular theory under investigation, then such and such should
hold in the world which the experimenter has created. This
transatability is what gives theoretical import to experiments
which involve phenomena which, taken in isolation, not only
appear picayune but seem to have little relationship with what
one observes in nature, (pp. 353-4).


82
AMSECO
Metal Folding
Chair Assembly Department
Labor Efficiency
Variance Report
For Job 5247
Standard Minutes
Actual Minutes Labor
1 Efficiency
Total Chai
Allowed Per Chair
Incurred Per
Chair Variance Per Chair
Produced
36.0
44.0
-8.0
200
The Costs Associated With Investigation Are:
If Your
And If The
Then
Your Costs
Are
Investigation
Assembly Line
kkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
Decision Is
State Is
Investigation
Production
Total
4rkkk kkkkkk
kkkkkkkkkkkkkk
A-*********
kkkkkkk
kkkkkkkkkkkk
Yes
In-Control
$ 28.33
$ 0.00
$ 28.33
Yes
0ut-0f-Control
$ 90.00
$ 0.00
$ 90.00
No
In-Control
$ O.CO
$ 0.00
$ 0.00
No
0u-0f-Control
S 0.00
$ 175.00
$ 175.00
Please answer the following questions placing your answers on the answer sheet:
A. Would you investigate this reported variance /circle the appropriate
response on the answer sheet/
NO
YES
B. How strongly do you feel about your decision /select a number between
0 and 100 which indicates the strength of your feeling and place this
number on the answer sheet/
0 10
20
30
40 50 60
70
80
90
100

*
*
k k k
k
*
k

kkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkkk
k k
k

k k k
k

k
k
k
k
k
Uncertain
Reasonably
Almost
Certain
Certain
FIGURE 6
VARIANCE INVESTIGATION DECISION TRIAL


27
possible decision outcomes. The parameters required to fit the optimal
model include 1) the relative frequencies of the two states of nature,
2) the mean of each state, 3) the statistical variance of each state,
and 4) the costs associated with each of the two possible decisions in
combination with each of the two possible states of nature. Since the
individual's decisions are to be evaluated by comparisons with the out
puts of the optimal model, it would seem reasonable that the information
available to the optimal model be the same as the information available
to the individual. Consequently, for decision situations in which some
of the parameters required by the optimal model are not contained in the
available information set the optimal model must make estimates of the
missing parameters.
Individual Ability to Expand the Information Set
The individual's ability to expand the available information set
over time refers to his ability to learn from his experiences with the
states of nature. Such learning can occur through improved estimates
of unknown items of distributional information and through modifications
of information processing strategies to incorporate state relationships
which were unknown or undetected previously.
The general form of the expected individual information processing
strategy is described by the heuristic of anchoring and adjustment
(Tversky and Kahneman, 1974).2 A natural starting point, or anchor, is
used as the first approximation for the decision. This anchor is then
2A heuristic, within this context, refers to a learned set of rules or
principles which are utilized by an individual in making the particular
decisions required of him.


The measure approaches either a value of zero or positive infinity as
the result of several factors: 1) the individual does not process
properly the effects of the relative costs of the two types of decision
errors, 2) the individual does not process properly the effects of the
relative frequencies of the two states, and 3) the individual uses more
than one cutoff value.
Using the same subjective analysis as that used in adjusting the
DN measure, the BN^ measure can be adjusted for the effects of multipl
cutoff values. Defining B^ to be the individual's criteria measure
with the effects of multiple cutoff values eliminated:
BNAi B* / Bk
The difference BNA^- BN^ approaches a value of zero as the effects of
multiple cutoff values decrease and becomes zero when the individual
uses a single cutoff value.
Another measure derived from the BN- measure can be considered
a measure of individual conservatism. In this study, conservatism
refers to incomplete adjustment from an initial decision anchor towards
the optimal cutoff value. Non-conservatism refers to a more that
complete adjustment from the initial decision anchor past the optimal
cutoff value. The extent of conservatism is measured by the relative
distance between the individual's cutoff value and the optimal model's
cutoff value. Since the measure is dependent on the direction of ad
justment it is conditional upon the level of the cost variable. Using
BNC.¡ to denote tne extent of an individual's conservatism:
BNC.¡ = (3? B|<) / B^ given the Cl cost level; and
BNC.j = (b^ Bj) / B^ given the C2 cost level.


APPENDIX B
ORAL PRESENTATION TO ELICITE VOLUNTEER SUBJECTS
I assume that most, if not all, of you intend to become account
ants. You should realize that accounting is more than an occupation,
it is a profession. As a member of a profession you have certain re
sponsibilities beyond those encountered in other occupations. One of
these responsibilities is the readiness to help expand the state of the
art, the level of knowledge, of the profession. Today you have an
opportunity to meet this responsibility. As part of my Ph.D. require
ments I am conducting an experiment which examines the impact upon
decision makers of certain accounting produced information. I am asking
that you volunteer to participate as subjects in this experiment.
The accounting information under study is the standard cost vari
ance report and, as a subject, you would be asked to assume the role of
an operations manager and to make certain decisions based upon standard
cost variance reports. After volunteering, greater detail concerning
the role and the required decisions will be provided in a short booklet
and in a training session.
Benefits to volunteering as a subject, beyond meeting part of your
professional responsibilities, include:
1) A chance to expand your knowledge of standard cost variance
reports and their utilization from the point of view of a
line manager.
174


60
between the DNA.¡ and DN^ measures. The two possible explanations in
clude the distribution variable and the cost variable.
Theoretically, the more extreme the optimal cutoff value relative
to a central point between the distributions of the two states, the
higher should be the individual conservatism. The decision situations
with the most extreme optimal cutoff values are those within the C2
cost level; the situations with the least extreme optimal cutoff values
are those within the Cl cost level. Consequently, the BNCj measure
should be larger within the C2 cost level than within the Cl cost level.
The simulation and assumptions used in developing the training
phase hypotheses can be extended to enable formation of hypotheses con
cerning the individual decision criteria. The earlier simulation de
rived the following final decision cutoff values: 1) 37.555 actual
minutes incurred for the Cl cost level given the 12 information level
and the SI distribution level, and 2) 39.755 actual minutes incurred
for the C2 cost level given the 12 information level and the SI dis
tribution level. Given these cutoff values a f(e^) measure can be
calculated directly (the measure is used rather than the 3. measure
due to the assumption of a single cutoff value). The functional
notation, f( ), is employed to indicate these estimates are based upon
a simulation rather than upon empirical observation. For this simu
lation the f(3?)s are as follows:
f(B*|Cl,Sl,I2) = (zout) / <¡>(zin) = 0.70818
f(B*|C-2,Sl,I2) = (zQut) / *(zin) = 2.11774
where ( ) indicates the value of the normal density function at the
point in parentheses. Given the f(B^) measures the f(BNA^) measures
can be calculated. For this simulation the f(BNA^)s are as follows:


no
effects of multiple cutoff values on individual decision model sensi
tivity. Assuming random subject assignment to experimental situations,
the Dfij and DNA. variables should be independent of the effects of the
d' absolute magnitude differences (within the distribution variable).
The remaining effects are 1) the relative variable response range
within the DNA^ measure, and 2) the relative variable response range
and multiple cutoff value usage within the DN.¡ measure.
The methods of analysis were the model comparison procedures where
the dependent variables a"e the DN.¡, DNA^ and DNA^-DN- measures. The
independent variables are the three situation variables (with their
interactions), the three motivation factors, and the GPA.¡ variable.
The three full models are
DNi
DNAi u + j + 8k + ^ + 6jk + Tj! + + a6Yjk] + 5,
DNA--DN- + t. + X. +6.
ill
The three model comparison procedure results are presented in
Table 11 and the F values associated with the sources of the reduced
models are presented in Table 12. The results indicate that the reduced
DN.. model contains the distribution variable, the cost variable, and
the intrinsic motivation factor. Both the distribution variable main
effect and the motivation factor are significant at the p<.01 level,
and the cost variable is significant at the p<.05 level. The reduced
DNA. model contains the cost variable and the intrinsic motivation
factor. The cost variable main effect is significant at the p<.01
level and the motivation factor is significant at the pc.10 level.
The reduced DNA..-DN.. model contains the distribution variable and the
intrinsic motivation factor. The distribution variable main effect is


191
Simon, H. A. The New Science of Management Decision. New York: Harper
and Row, Inc., 1966.
Slovic, P. "From Shakepeare to Simon: Speculations-- and Some Evidence--
About Man's Ability to Process Information." Oregon Research
Institute Bulletin, 12 (1972), 1-29.
, and Lichtenstein, S. "Comparison of Bayesian and Regression
Approaches to the Study of Information Processing in Judgement."
Organizational Behavior and Human Performance, 6 (1971),
649-744.
Snowball, D., and Brown, C. "Decision-Making Involving Sequential
Events: Some Effects of Disaggregated Data and Dispositions
Toward Risk." Unpublished manuscript, University of Florida, 1977.
Swets, J. A. "Signal Detection in Medical Diagonis." In Computer
Diagnosis and Diagnostic Methods, J. A. Jacquez (Ed.).
Springfield, IL: C. C. Thomas, 1972.
Swieringa, R. J.; Gibbins, M.; Larsson, L.; and Sweeney, J. L. "Ex
periments in the Heuristics of Human Information Processing."
Journal of Accounting Research: Supplement on Studies on
Human Information Processing In Accounting, 14 (1976), 159-87.
Taylor, R. N., and Dunnette, M. D. "Relative Contribution of Decision
Maker Attributes to Decision Processes." Organizational
Behavior and Human Performance, 12 (1974), 286-98.
Tversky, A., and Kahneman, D. "Belief in the Law of Small Numbers."
Psychological Bulletin, 76 (1971), 105-10.
. "Availability: A Heuristic for Judging Frequency and
Probabi1ity." Cognitive Psychology, 5 (1973), 207-32.
. "Judgement Under Uncertainty: Heuristics and Biases."
Science, 185 (1974), 1124-31. ~
. "Causal Schemata in Judegments Under Uncertainty." In
Progress in Social Psychology, M. Fishbein (Ed.). Hillsdale,
N.J.: Lawrence Erlbaum Associates, 1977.
Ulehla, Z. J.; Canges, L; Wackwitz, F. "Signal Detectability Theory
Applied to Conceptual Discrimination." Psychonomic Science,
8 (1967a), 221-2.
. "Integration of Conceptual Information." Psychonomic
Science, 8 (1967b), 223-4.
Watson, C. S. "Psychophysics." In Handbook of General Psychology,
B. B. Wolman (Ed.). New York: Prentice-Hall, Inc., 1973.


78
costs were presented to the subject on the face of each variance report
using the following format:
If Your
Investigation
Decision Is
And If The
Assembly Line
State Is
Then
Your
Costs
Are
Investigation
Production
Total
Yes
In-control
$
X
$
0.00
$ X
Yes
Out-of-control
$
Y
$
0.00
$ Y
No
In-control
$
0.00
$
0.00
$ 0.00
No
Out-of-control
$
0.00
$
Z
$ z
The letter X represented the estimated investigation cost when the
assembly line was in-control. The cost was related to the size of the
labor efficiency variance: the more negative (unfavorable) the variance
the larger the cost. The letter Y represented the estimated investi
gation cost when the assembly line was out-of-control. The cost was
related to the size of the labor efficiency variance: the more negative
(unfavorable) the variance the smaller the cost. Given any particular
labor efficiency variance, the investigation cost when out-of-control
(Y) was always larger than the investigation cost when in-control (X).
This was because the investigation cost included correction costs when
the assembly line was out-of-control. The letter Z represented the
estimated marginal production cost of operating the next period in the
out-of-control state. The cost was constant for all variance reports.
The subjects were told that their immediate supervisor, the pro
duct section manager, would evaluate their control performance in terms
of their minimization of both investigation and production costs above
the expected standard (labeled the total investigation decision cost).
A cash bonus was promised to the subjects, the size of the bonus being
contingent upon the extent to which they minimized their total investi
gation decisions cost. The measure of a subject's control performance,
labeled TIDC^^, was determined by summing the total investigation


136
TABLE 22
PARAMETER VALUES ASSOCIATED WITH THE REDUCED GPi/GNi MODEL
F Values Associated With the Sources of the Reduced Model
Source
d.f.
F
Distribution
1,70
0.64
Cost
1,70
0.35
Distribution
1,70
8.85a
DNj
1,70
59.01a
Intrinsic motivation
1,70
0.11
Extrinsic (non-monetary) motivation
1,70
0.54
Extrinsic (monetary) motivation
1,70
0.09
GPA
1,70
2.54
Means and Variances Associated With
the Siqnif
icant Sources
Variable Level
N
Mean
Variance
SI, Cl
24
2.0189
1.42035
SI, C2
19
3.1275
2.12453
S2, Cl
19
2.3372
2.09340
S2, C2
17
2.3990
1.71284
ap<.01


70
f(G.¡|Sl,C2) = 0.06471
f(Gi|S2,Cl) = 0.01287
f(Gi|S2,C2) = 0.03185
A graphic representation of this interaction is presented in Figure 4.
The difference between the two f(G.|Cl) measures is much smaller than
the difference between the two f(G^|C2) measures. The differences are
0.00376 and 0.03286, respectively. This would indicate that the
distribution variable has a significant effect only when given the
C2 cost level. Such q result would explain the larger but not signifi
cant effect of the SI distribution level on the f(|S ) measures.
Other implications which can be derived from these relative
relations involve relationships between f(GP.¡) and f(GN.j) measures.
Two such relationships appear to be of possible significance. Both
involve the ratio f(GP^)/f(GN.) (i.e., the effect of additional errors
relative to the effect of additional correct decisions). The results
of the first relationship are as follows:
f(GPi/GNi|C1) = 1.20480
f(GPi/GNi¡C2) = 1.38337
The difference between these measures is large compared to the standard
error of the estimates (the difference divided by the standard error is
0.17857/0.07621 = 2.34313). Consequently, the GP^/GN^ ratio is ex
pected to be larger given the C2 cost level than given the Cl cost
level.
The results of the second relationship are as follows:
f(GPi/GNi|Sl,Cl) = 1.19908
f(GPi/GNi|Sl,C2) = 1.48206
f(GP./GNi|S2,Cl) = 1.21057


44
SMC= the sum of optimal model k's investigation decisions costs
over all decisions (m in number).
The G-. measure can be classified into one of three submeasures
' J
where the classification is dependent upon the algebraic sign of G.¡j.
The classifications are 1) G- > 0-- the individual's investigation
decision cost for decision j is greater than that of the optimal model,
2) G^j = 0-- the individual's investigation decision cost for decision
j is equal to that of the optimal model, and 3) G < 0-- the indi-
* J
vidual's investigation decision cost for decision j is less than that
of the optimal model.
Summing those G-s with identical algebraic signs, the summations
' 0
are defined as follows:
GP.j = the sum of those Gs with a positive algebraic sign;
GN^ = the sum of those Gs with a negative algebraic sign; and
GZ.j = the sum of those Gs with a value of zero.
It is easily shown that the following relation holds:
m
E G.. = GP, + GN, + GZ,
j=l J
where m equals the total number of decisions made by individual i. Since
the GZ.¡ summation equals zero, the sum of the G over all j = l,m
decisions, denoted G-, is:
G- = GP- + GN.
i i i
Conceptually, GP^ is the relative additional cost of those de
cisions made by individual i which are greater in value than that of
the optimal model. The GN^ is the relative savings of those decisions
made by individual i which are less in value than that of the optimal
model.


69
0.0089/0.01808 = 0.49226). Therefore, a significant effect due to the
information variable would not be expected.
Averaging over all conditionals except the distribution variable
gives the following results:
fiG^Sl) = 0.04152
f(G.|S2) = 0.02235
Although the difference between these measures is much larger than that
above, this difference is only slightly larger than the standard error
of the estimates (the difference divided by the standard error is
0.01832/0.01658 = 1.10495). The direction of the f(G^) differences
would be expected to be as predicted by the simulation but may not be
significant.
Averaging over all conditionals except the cost variable gives
the following results:
f(G.|Cl) = 0.01475
f(Gi|C2) = 0.04827
The difference between these measures is large compared to the
standard error of the estimates (the difference divided by the standard
error is 0.03353/0.00951 = 3.5258). Consequently, a significant effect
would be expected for the cost variable.
These relative relations can be averaged over all but pairs of
conditionals. Such measures could indicate the presence of interaction
effects of the independent variables. The results of averaging the
pairwise conditionals indicate only one interaction of possible
significance-- the distribution by cost interaction. The f(G^) measures
conditional to these variables are as follows:
ftf^ISl.Cl) = 0.01663


General Research Methodology 33
General Experimental Environment and Task 34
Operationalization of Variables 35
Hypothesis Formation 46
Simulation of Optimal Model Performances 46
Simulation of Investigation Decision Performances 49
CHAPTER IV. THE EXPERIMENT 76
Experimental Environment 76
Subjects 79
Experimental Materials 80
Background Information 80
Variance Investigation Decisions 80
Elicitation of Heuristics 83
%
Elicitation of Subject Motivations . . 83
Experimental Procedures 83
Assignment of Subjects to Treatment Conditions 84
Training Phase 85
Experimental Phase 86
Final Debriefing 87
CHAPTER V. ANALYSIS AND RESULTS 88
Summary of Results 88
General Method of Analysis 88
Training Phase Analysis and Results 98
Initial Decision Anchors 98
Decision Anchor Adjustment 102
Individual Decision Model Sensitivity 106
The dj Dependent Variable 106
v


Overall
Objective
Control
Strategic
Planning
Overall
Objective
Planning
Overal1
Task
Control
Management
Control
Overall and
Specific Task
Planning
Specific
Task
Control
V
Operational
Control
Specific Task
Structuring
Source: Modified from Demski, 1967.
Environment
Output
FIGURE 1
MANAGEMENT TASK PLANNING AND CONTROL PROCESS


75
classifications have one positive adjustment which would improve per
formance above that of subjects who make only negative adjustments. On
the other hand, the mixed classifications have one negative adjustment
which would decrease performance below that of subjects who make only
positive adjustments.
The following hypotheses predict the effects of the training
adjustment pattern on the Gi variable:
H4.5 (G^PP) < (G^ |MM).
Those subjects within the PP classification of training adjust
ment pattern will have a significantly smaller mean G- measure
than those subjects within the MM classification.
H4.6 (GjPP) < (G^MP) = |PM) < (G^MM).
Those subjects within the mixed classifications of training
adjustment pattern will have mean G^s that fall between those of
the non-mixed classifications. The mean G^s of the mixed classi
fications are not expected to be significantly different.


30
The statistical structure. The effects of the statistical struc
ture of the decision situation are studied by manipulating a distribu
tional information variable. This involves the specification of two
levels of statistical relationship among the two states. Since the
difficulty of the discrimination task increases as the area of dis
tributional overlap between the two states increases, one level of the
distribution variable will have a greater area of overlap than will
the other level. This independent variable is labeled the distribution
variable. All other distributional information items, if presented,
will be controlled as constants.
The cost structure. The effects of the cost structure of the
possible decision outcomes are studied by manipulating a singular in
formation variable. The manipulation of this variable involves the
specification of two levels of decision outcome relationships (in terms
of incurred costs). This independent variable is labeled the cost
variable. Since the importance of the discrimination task increases as
the costs associated with decision outcomes that involve decision errors
become unequal, the different levels of the cost variable will be
associated with different decision error costs. One level of the cost
variable is structured in favor of more variance investigations and the
other level of the cost variable is structured in favor of fewer vari
ance investigations. The only other singular information variable, the
actual results of the physical process and its cost variance, is the
primary experimental stimulus. This information item will be a random
variable whose distributions, given either of the two states of nature,
will be normally shaped with parameter values defined by the appropriate
level of the distribution variable.


144
. TABLE 24
VARIMAX ROTATED MOTIVATION FACTOR PATTERN AMD DESCRIPTION OF
MOTIVATION SCALE ITEMS
Scale Item
Factor 1
Factor 2
Factor 3
Factor 4
1
0.1399
0.2336
-0.2112
-0.7551
2
0.2610
0.7485
0.0327
0.0503
3
0.8287
0.2336
-0.0712
-0.0359
4
0.8001
-0.1721
0.2190
-0.0836
5
-0.0305
0.0197
0.9037
0.1436
6
-0.1229
0.2524
-0.0876
0.8044
7
0.0231
0.8198
0.0104
0.0174
8
0.0814
0.1075
0.8789
-0.0385
9
0.7324
0.3174
-0.0963
-0.3140
10
0.0148
0.7586
0.0848
-0.0441
Eigenvalue
2.6975
1.8552
1.6548
0.9528
Cum Eigenvalue
Portion
0.2700
0.4550
0.6210
0.7160
Description
of Motivation
Scale Items
1 Number of
times willing
to participate
in the future.
2 Desire to perform well in undertaking a challenging task.
3 Enjoy making the decisions.
4 Satisfied with your performance.
5 Would have tried less hard if money rewards had been halved.
6 Feel tense during the experiment.
7 Desire to cooperate with the experimenter.
8 Money rewards caused you to try harder than without them.
9 Enjoy the overall experience of participation.
10 Desire to contribute to research knowledge.


52
a subject will select an initial anchor as a first decision approxi
mation and will adjust this decision anchor toward an optimally correct
point as he learns from his experiences with the states of nature.
However, the individual's adjustment process will not be sufficient:
i.e., he will tend to approach the optimally correct point but will not
adjust completely to that point.
The second assumption is that the location of the initial de
cision anchor on the actual minutes incurred continuum (axis) will be
conditional upon the level of the distribution variable. The reason
for this is that the initial decision anchor is expected to be located
at a central point between the means of the two states. Since the
mean of the out-of-control distribution shifts with the level of the
distribution variable, the central point between the means of the two
states also shifts with the level of the distribution variable. The
initial decision anchor used in the simulation of subjective investi
gation decision performances is the geometric intersection point of the
two states' distribution curves. This particular initial decision
anchor is used because it is the exact central point between the means
of the two states. However, it is selected for operational purposes
and is not necessary to the conceptual development. Any point of
central tendency would be satisfactory.
The final assumption is that the subjects' adjustment process
will be approximately equal over the various treatment conditions. The
adjustment process is defined as a linear movement along the actual
minutes incurred axis from the initial decision anchor towards the
appropriate optimal model decision cutoff. The magnitude of the linear
movement used in the simulation of subjective investigation decision


24
the specific values which these variables may assume, the structure of
the particular decision situation can affect both the difficulty and
the importance of the individual's discrimination among the states. In
general, as the number of possible states increases and as the area of
distributional overlap of these states increases the discrimination task
becomes more difficult. Furthermore, the discrimination task becomes
more important (in terms of incurred costs) as the relative frequencies
of the states become equal and as the costs associated with the possible
decision outcomes which involve decision errors (an incorrect decision
given the existence of a specific state) become unequal.
The general situation employed in this research, selected for
its relevance to the accounting discipline, is the cost variance
investigation decision. In this situation two states of nature are
possible: 1) in-control (the underlying physical process described by
the standard cost variance report is functioning as planned), and 2)
out-of-control (the underlying physical process described by the stan
dard cost variance report is not functioning as planned). In reality,
the possible states of nature may be located on a continuum whose end
points are the states of in-control and out-of-control. Between these
two end points are any number of states that take the general form of
partially out-of-control or moving out-of-control. In the present
research the possible states of nature are confined to the two end
points. The relative frequencies of the two states are controlled as
constants with their values being close to equal. The statistical
relationships among the two states and the relationships between the
various decision outcomes are manipulated as independent variables.


49
information level the available information set contains statistical
estimates of all the parameters required to fit the optimal model.
However, within the II information level the available information
set does not contain all the required parameter estimates. Therefore,
within the II information level the optimal model must use the training
sessions to estimate the missing parameters. The parameter estimates,
both given and estimated, are presented in Table 2.
The second criterion is that the labor efficiency variance re
ports presented to the subjects in the experiment be the same reports
used in the simulation of optimal model performances. The appropriate
parameter estimates presented in Table 2 were used in the optimal
model decision rule expressed by equation 2 to simulate optimal model
investigation decisions. Table 3 presents some performance results
of this optimal model simulation.
Simulation of Subjective Investigation Decision Performances
This section discusses assumptions derived from the conceptual
development that form the basis for the simulation of subjective investi
gation decision performances. It also presents the results of this
simulation and the derivation of the experimental hypotheses.
Assumptions of simulation
The first assumption employed for the simulation of subjective
investigation decision performances is that the subjects will behave
as if they use the anchoring and adjustment heuristic during the
training phase. The anchoring and adjustment heuristic proposes that


171
utilized as performance measures for the appropriate operational managers
and for the physical systems under their control. The nature of de
cisions of this type, although generally aggregated over longer time
periods, is that of adequate performance verses inadequate performance
rather than that of investigation verses non-investigation. Within the
current research, performance decisions were assumed implicitly within
the performance feedback measures given the subjects during the train
ing phase. An extension would involve the actual formation of such
decisions and, in turn, their effects upon operational control decisions.


This dissertation was submitted to the Graduate Faculty of the Department
of Accounting in the College of Business Administration and to the
Graduate Council, and was accepted as partial fulfillment of the require
ments for the degree of Doctor of Philosophy.
August 1978
Dean, Graduate School


161
this adjustment the simulated cost variable effects would have the same
strength as the obtained effects (i.e., f (G-j | C2 )/2 f (G^ | Cl) would equal
1.60).
The simulated and obtained distribution by cost variable inter
actions involving the relative decision costs had similar relationships
between their strengths: the effects of the obtained interaction were
not as strong as the effects of the simulated interaction (note that
[(G.¡ |S1,C2)-(G.¡ |S2,C2)] /[(^ |S1 ,C1 )-(Gi |S2,C1)] equalled 3.15 whereas
[f(Gi|Sl,C2)-f(Gi |S2,C2)] / [f¡SI,C1)-f(Oj |S2,C1)] equalled 8.74).
Again, if the simulated effects are adjusted for unequal learning
efficiency, then the simulated distribution by cost variable inter
actions have a strength similar to that of the obtained effects (i.e.,
[f(Gn. |Sl,C2)-f(Gi ¡S2,C2)] / [2-f (G^-1 SI ,C1 )-2*f (G.¡ |S2,C1)] would equal
4.37).
The simulated and obtained distribution by cost variable inter
action within the GP^/GN^ measure had opposite relationships between
their strength: the effects of the obtained interaction were stronger
than the effects of the simulated interaction (note that [(GP^/GN^|C2,S1)-
(GPi/GNi|C1,S1)] / [(GPi/GNi|C2,S2)-(GPi/GNi |C1 ,S2)] equalled 17.94
whereas [f(GPi/GNi|C2,S1)-f(GPi/GMi|C1,S1)] / [f(GP1/GNi|C2,S2)-
f(GP^/GN^jCl,S2)] equalled 3.82). To explain the difference in effect
strengths requires reference to two variables: multiple decision an
chors and unequal learning efficiency. If the GP.¡/GN.¡ measures are
adjusted for the effects of multiple decision anchors the strenqth of
the obtained interaction effect is 2.89 (using the above interaction
strength ratio). This reverses the relationships between the effect
strengths: i.e., the obtained interaction strength becomes weaker


162
than the predicted strength. The effects of unequal learning efficiency
were 1) an expansion of the interval between the final decision an
chor and the optimal decision value for those subjects with greater
decision criteria conservatism, and 2) the reduction of the interval
between the final decision anchor and the optimal decision value for
those subjects with less decision criteria conservatism. As the in
terval associated with those subjects within the Cl cost level expanded
(this level had greater decision criteria conservatism), both the
(GP-¡¡C1) and the (GN-j |Cl) became larger. However, the GP-¡ measure
became larger at a greater rate than did the GN-¡ measure (3GP-¡/9GN-¡>l).
As the interval associated with those subjects within the C2 cost level
contracted (this level had less decision criteria conservatism), both
the (GPj|C2) and the (GNj |C2) became smaller. However, the GP-¡ measure
became smaller at a greater rate than did the GN.¡ measure (3GP1-/3GN1->1).
Consequently, the ratio (GP^/GN^|Cl) increased as the level of conser
vatism increased, and the ratio (GP^/GNJC2) decreased as the level of
conservatism decreased. Incorporating these effects of unequal learning
efficiency into the simulated distribution by cost variable interaction
lowered the predicted strength of such an interaction.
The observed associations between the individual long-run decision
efficiency measures and the individual attributes may be interpreted
with some degree of intuitive satisfaction. As the subjects' variable
response ranges increased the relative decision costs increased (either
as a result of increased additional decision costs or decreased addi
tional decision savings). Furthermore, as the relative decision costs
increased both the subjects' intrinsic and extrinsic (monetary) moti
vation measures decreased. A similar association existed between the


55
adjustment process will be approximately equal over the various con
ditions. The location of the initial decision anchor can be investigated
by examining the relationship between the individual's initial cutoff
value (in the first training session) and the intersection point of
the appropriate two distribution curves (the TDV.¡ measure). These
intersection points are conditional upon the level of the distribution
variable. Given the SI distribution level the intersection point is
38.25 actual minutes incurred, and given the S2 distribution level the
intersection point is 40.5 actual minutes incurred (see Figure 2 for a
graphical presentation of these points). The expected relationship is
that the mean initial decision anchor (TDV.) of the individuals, given
a level of the distribution variable, will not be significantly differ
ent from the intersection point of the appropriate two distribution
curves.
The assumption of approximately equal adjustment over the various
conditions can be investigated by examining the relationship between
the individual's adjustment over the complete experiment and the total
adjustment that would be required to reach the optimal model's decision
cutoff value. A measure of the relative adjustment for individual i,
RA.j, is conditional upon the direction of adjustment along the relevant
decision axis (i.e., the actual minutes incurred per chair produced).
The expected relationship is that the mean RA^ will not differ signifi
cantly between the levels of the various conditions.
Given the above simulation and discussion the following hypo
theses can be derived:
HI.la (TDV.|S1) = 38.25.
The mean initial decision anchor given the SI distribution level


58
the use of multiple cutoff values, the average difference between DNA.¡
and DN.¡ should differ between the two levels of the distribution vari
able. A second possible explanation involves the cost variable. As
the individual's final cutoff value increases, the absolute distance
between the individual's final cutoff value and the lowest values of
the random variable of interest increases. In turn, as this distance
increases subjects again may be more inclined to perceive that a sec
ond out-of-control distribution overlaps the lower range of the
in-control state. If such perceptions underly the use of multiple
cutoff values, the average difference between DNA- and DN.¡ should
differ between the two levels of the cost variable.
Based on the above discussion the effects of the independent
variables on the individual decision model sensitivity measures can
be hypothesized as follows:
H2.la (di |SI) < (di |S2).
Those subjects within the SI distribution level will have
significantly smaller mean djs then those subjects within the
S2 level.
H2.lb (di ¡II) = (di ¡12).
Those subjects within the II information level and within the
12 information level will have equal mean d^s.
H2.1c (di 1 Cl) = (di |C2).
Those subjects within the Cl cost level and within the C2 cost
level will have equal mean dis.
H2.2a (DNAilll) = (DA-j ] 12).
Those subjects within the II information level and within the 12
information level will have equal mean DNAjS.


Egan (1975) further demonstrates that the posterior probability is
strictly monotone with the likelihood ratio of x, L(x); that is, when
L(x2) > LCXj) then P(sn|x2) > P(sn|x1).
173
Maximization of Expected Value Objective Function
The decision function for maximizing the expected value decision
goal is:
E(V) = P(Y!sn)P(sn)Vsn)Y + P(Y|n)P(n)VnjY
+ P(N|n)P(n)Vn>N + P(N|sn)P(sn)VsnjN
Given an observation x, the subject has two decision functions:
1) E(V|x,Y), the expected value given the observation and the
subject responds "Yes," and
2) E(V|x,N), the expected value give the observation and the
subject responds "No."
These two decision functions can be expressed as the sum of the values
associated with each decision weighted by the appropriate posterior
probability:
E(V|x,Y) = P(sn|x)VsrljY + P(n|x)VnjY, and
E(V j x,N) = P(n|x)VnjN + P(sn|x)VsnjN.
The subject should say "Yes" only if E(V|x,Y) > E(V|x,N). If the
righthand side of the above two equations are used in this inequality,
then the subject should say "Yes" only when:
P(sn|x)Vsn,y + P(n|x)VnY > P(n|x)VnjN + P(sn|x)Vsn>N
Rearranging the terms and expressing the posterior odds in terms of the
prior odds and the likelihood ratio of x results in the decision rule
for saying "Yes" in terms of the likelihood ratio of the observation:
P(x1sn) P(n) Vn,N Vn,Y
P(x n) > P(sn) 'vsn>Y VsM


ACKNOWLEDGEMENTS
I am grateful to my supervisory committee; Rashad Abdel-khalik,
Douglas Snowball, Gary Hoi strum, and Richard Griggs. Without their
valuable contributions this dissertation would have had far more
obscurities.
I am indebted to my experimental assistants: Doug Snowball,
Gary Hoi strum, Ron Teichman, Ed Bailey, John Wragge, Robert Thompson,
Nancy Hetsko, Andy Judd, and Michael Gift (even though he was not on
time). Without their help the experimental procedures used within
this study would not have been possible.
Finally, my greatest debt is to my wife, Sandy. Without her
encouragement and support this dissertation never would have been done.


151
individual decision criteria conservatism, and individual long-run
decision efficiency.
Decision Anchor Selection and Adjustment
The first aspect of the anchoring and adjustment heuristic is the
selection of the initial decision anchor. The formation of the hypo
theses concerned with this aspect was based upon the assumptions that
1) individuals (within a given level of distribution variable) would
tend to use a common initial decision anchor, and 2) the location of this
common anchor would be related to a measure of central tendency between
the two states of nature. This central tendency was operationalized as
the geometric intersection of the state distribution curves (the inter
section was conditional upon the level of the distribution variable).
The results indicated that the expected point of central tendency
was obtained for those subjects within the SI distribution level, but
that the expected point was statistically larger than that obtained
for those subjects within the S2 distribution level. The state means
were farther apart within the S2 distribution level than within the SI
level. As a consequence, the intersection of the state distribution
curves was located farther from the standard (the mean of the in-control
state) within the S2 distribution level than within the SI level. An
initial subjective bias provides a plausible explanation for the
estimate obtained within the S2 level. This bias would take the form of
a tendency to place greater weight on the standard as the intersection
of the distribution curves became larger (i.e., shifted to a higher
point on the actual minutes incurred axis).


134
The model comparison procedure results are presented in Table 21,
and the F values associated with the sources of the reduced model are
presented in Table 22. The results indicate that the reduced GP-¡/GM¡
model contains the distribution by cost interaction and the DMj variable,
both significant at the p<-01 level.
The means, variances, and sample sizes of the distribution by
cost interaction are included in Table 22. Figure 8 presents a graphic
representation of the distribution by cost interaction. Bartlett's test
of homogeneity of variances indicated that the variances within the
distribution by cost interaction are homogeneous (x2=1.068, p<.25 with
3 d.f.). Duncan's multiple range test of equal means indicated that
the means within the distribution by cost interaction form two over
lapping groups: the means within a group do not differ significantly
(p=.01), whereas the means between groups do differ significantly (p=.01).
These groups are 1) (GPi/GNi|C2,S1), (GPi/GNi|C2,S2), and
(GPi/GNi|C1,S2); and 2) (GPi/GNi|C1 ,S1), (GPi/GNi|C1.52), and
(GPi/GNi|C2,S2).
Hypothesis 4.4b predicted a significant distribution by cost
variable interaction in which the (GP^/GN^ |C2,S2)-(GPi/GN1-1Cl ,S2) would
be smaller than (GPi/GNi|C2,Sl)-(GPi/GNi|C1 ,S1). The results of the
model comparison procedure indicated a significant distribution by cost
interaction (p<.01). The (GPV/GN^ |C2,S2)-(GP^/GN1- |C1 ,S2) was equal to
0.0618, and the (GPi/GNi|C2,S1)-(GPi/GNi|C1,S1) was equal to 1.1086.
Hypothesis 4.4a predicted that (GP^/GN^|Cl) would be significantly
smaller than (GP^/GNj¡C2). Using a Z test of equal means, the difference
between the means was found to be significant (Z=2.0109, p<.01
one-tailed), with the Cl level having the smaller mean. The lack of


185
D. The actual minutes incurred per chair.
*************************************************
*****
Not Slightly Moderately Very Extremely
Important Important Important Important Important
E. The labor efficiency variance per chair.
*************************************************
*****
Not Slightly Moderately Very Extremely
Important Important Important Important Important
F. The total chairs produced.
*************************************************
Not Slightly Moderately Very Extremely
Important Important Important Important Important
G.The costs associated with investigation decisions.
k'k'k'k-k'k'k'k'k'k'k-k'k'kirk'k'k'k'k'k'k'k'k'k'k'k'k'k'k'k'k'k'k-k'k'k'k'kic'k'k'k'k'k'k'k'k'k
* * *
Not Slightly Moderately Very Extremely
Important Important Important Important Important
H.Other (please specify) .
*************************************************
*****
Not Slightly Moderately Very Extremely
Important Important Important Important Important
4. Recalling the training phase which you completed at an earlier time,
what rule or set of rules did you use to make your variance investi
gation decisions in the first 33 variance reports? Please include
any numerical values that were part of this rule or set of rules.
5. How did you initially form the rule or set of rules you gave in
question four? What lead you to use this particular rule or set of
rules?
6. If the rule or set of rules you gave in question four is different
from the rule or set of rules you gave in question one, what caused
you to make the change?
7. What other information items, if presented, do you feel would im
prove you variance investigation decisions?
8. Were your decisions influenced by discussions with other partici
pants in this experiment?


Hypotheses
Analysis
Method(s)
1.3b (Mi|Sl) = (Mi|S2)
Model comparison
Two-tailed Z test
1.3c (Mi|Cl) = (Mi|C2)
Model comparison
Two-tailed Z test
2.1a (cH ¡Sl)<(dT |S2)
Model comparison
One-tailed Z test
2.1b (34 | Il) = (cT'. 112)
Model comparison
Two-tailed Z test
2.1c (dt |Cl) = (d;. |C2)
Model comparison
Two-tailed Z test
ABLE 4-- Continued
Test
Statistic(s) Results
F=1.7308 (n.s.)
The mean relative decision anchor ad
Z=1.3054 (n.s.)
justment of those subjects within the
SI distribution level was not signifi
cantly different from that of those
subjects within the S2 distribution level
F=0.1735 (n.s.)
The mean relative decision anchor ad
Z=0.4865 (n.s.)
justment of those subjects within the
Cl cost level was not significantly dif
ferent from that of those subjects with
in the C2 cost level.
F=46.0985 (p<.01)
The mean decision model sensitivity of
those subjects within the SI distribu
Z=6.1144 (p<.01)
tion level was significantly smaller than
that of those subjects within the S2
distribution level.
F=0.5883 (n.s.)
The mean decision model sensitivity of
Z=0.4588 (n.s.)
those subjects within the reduced infor
mation level was not significantly dif
ferent from that of those subjects with
in the expanded information level.
F-0.2638 (n.s.)
The mean decision model sensitivity of
Z=0.3167 (n.s.)
those subjects within the Cl cost level
was not significantly different from
that of those subjects within the C2
cost level.


48
equation 2 can be restated in the following mathematical terms:
exp -(x y2)2/2cr2
: r > PR-i CR
exp -(x yj^cr2
where x = actual minutes incurred per chair;
Vj = the mean of the in-control state;
y2 = the mean of the out-of-contrcl state;
a = the common standard deviation of both states; and
exp = the notation for the exponential funcition (e).
If the above equation was changed to an equality and solved for x the
result would be the optimal cutoff value in terms of actual minutes
incurred. This solution takes the general form:
G2(ln(PR-¡) + ln(CR))
x = + 0.50(y2 + y,) Equation 3
y2 yi
where x = the optimal cutoff value in terms of actual minutes incurred;
and
In = the notation for the natural logarithm function (loge).
With the decision rule in the form of equation 3 the parameters required
to fit the optimal model become identifiable. These parameters include
the mean of both states of nature, the common variance of the two
states, the prior odds of the in-control state, and the appropriate
value of the cost ratio.
With knowledge of the parameters required to fit the optimal
model the criteria employed for the simulation of optimal model per
formances can be discussed. The first criterion is that information
available to the optimal model be the same as the information avail
able to the individual. The information available to the individual
is manipulated by the levels of the information variable. Within the 12


83
Elicitation of Heuristics
The elicitation of each subject's heuristics was accomplished using
a predominately open-ended questionnaire (see Appendix E). The objec
tives of this questionnaire were 1) to determine the strategy the sub
ject used in answering the variance investigation questions within the
experimental phase, 2) to determine, where appropriate, the exact values
of any numerical rules used by the subject, 3) to determine whether the
subject thought he used his strategy consistently within the experimental
phase, 4) to determine the method or approach used by the subject in
forming his strategy within the training phase, and 5) to determine the
subjective importance attached to each information item presented to
the subject.
Elicitation of Subject Motivations
Subject motivations were elicited using a motivation questionnaire
(see Appendix F) developed by Snowball and Brown (1977). The question
naire is a ten item Likert-type scale which has submeasures for both
intrinsic and extrinsic motivation.
Experimental Procedures
Experimental procedures included assignment of subjects to
treatment conditions, adminstration of a training phase, administration
of an experimental phase, and final debriefing.


TABLE 3
RESULTS OF THE SIMULATION OF OPTIMAL MODEL PERFORMANCES GIVEN THE VARIOUS CONDITIONS
k Information
Distribution Cost
P(investigate
out-of-control)
P( investigate
in-control)
Optimal Cutoff
Value (actual
minutes incurred)
dk
6k
1
11
SI
Cl
0.875
0.383
36.90
1.447
0.539
2
11
SI
C2
0.400
0.033
41.29
1.581
5.206
3
11
S2
Cl
0.875
0.300
38.70
1.674
0.592
4
11
S2
C2
0.425
0.017
45.88
1.939
9.453
5
12
SI
Cl
0.900
0.400
36.86
1.535
0.454
6
12
SI
C2
0.400
0.033
41.26
1.581
5.206
7
12
S2
Cl
0.900
0.300
38.57
1.806
0.504
8
12
S2
C2
0.525
0.033
44.68
1.897
5.364
cn


45
The anchoring and adjustment process may occur over three training
sessions and one experiment session. Using the first training session
as an estimate of the initial decision anchor and the experiment
session as an estimate of the final decision anchor, there remains two
training sessions within which the adjustment process itself can be
analyzed. In particular, the effects of the pattern of adjustments
within the training sessions are expected to influence significantly
the magnitude of an individual's experiment Gj measure. The adjustment
process can be measured in the training sessions by computing the
individual's Gj measure at the completion of each session. The
measures, labeled TG^, TG.¡0, and TG.¡3, are defined in the same manner
as the G. measure discussed above (with the exception that they are
computed from the appropriate training session data instead of the
experiment session data). The differences between these training
measures reflect the effects of the individual's adjustments in the
training sessions. These differences are defined as follows:
DTGii = TG11 TGi2
DTGi2 = TGi2 TGi3
where, DTGjj = the difference between the G-¡ measure for the jth and
the jth + 1 training sessions of individual i; and
TG^j = the G.j measure for the jth training session of indi
vidual i.
Negative DTGjjvalues indicate an increase in the TGjj measure between
training sessions, and positive DTG-jj values indicate the reverse.
The effects of the directions of these adjustments should be
related to the effects of the final anchor on the individual's G-j
measure (in the experiment). The effects of the adjustment measures


109
TABLE 10
PARAMETER VALUES ASSOCIATED WITH THE REDUCED d! MODEL
i
F Values Associated With the
Sources of the Reduced d1 Model
Source
d.f.
F
Distribution
1,80
46.52a
Intrinsic motivation
1,80
10.33a
Extrinsic (non-monetary)
motivation
1,80
1.80
Extrinsic (monetary)
motivation
1,80
0.06
GPA
1,80
0.20
Means and Variances Associated
With the Significant Sources
Variable Level N
Mean
Variance
SI 44
1.2325
0.09165
S2 42
1.6093
0.07200
p<.01


53
performances is 50 percent of the distance between the initial decision
anchor and the appropriate optimal decision value. The 50 percent
adjustment value, selected for operational purposes, is completely
arbitrary and is not necessary to the conceptual development. The only
restriction is that it be less than 100 percent. Furthermore, the
general assumption of equal adjustment over the various conditions is
not necessary to the conceptual development: the objective of this
general assumption is to facilitate a simple operationalization of the
simulation.
Training phase simulation and hypotheses formation
The first stage of the simulation involves simulating the
anchoring and adjustment process in the training sessions for each
combination of independent variables. A simulated subject sample size
of one is used for each condition (resulting in a total simulated sub
ject sample of eight). The results of this first stage simulation are
presented in graphical form in Figure 2. This figure depicts the
simulated initial decision anchors (points A), the simulated final
decision anchors (points B), and the simulated optimal cutoff values
(points C) for each of the eight combinations of independent variables.
The assumptions used in simulating the training phase can be
investigated using the results of the experimental training phase. The
assumptions are i) that the subjects will behave as if they used the
anchoring and adjustment heuristic, 2) that the initial decision anchor
will be located centrally between the means of the two states and will
be dependent on the level of the distribution variable, and 3) that the


99
of nature) as their initial decision anchor. The simulation arbitrarily
employed the geometric intersection of the appropriate state distribu
tion curves as the expected initial decision anchor. These expected
points are conditional upon the distribution variable: i.e., the
intersection point is 38.25 actual minutes incurred given the SI distribu
tion level and is 40.5 actual minutes incurred given the S2 distribution
level. The cutoff points employed within the first training session
(TDV.s) were used as an estimate of the subjects' initial decision
anchors.
The method of analysis was the model comparison procedure where
the dependent variable is the TDV^ measure and the independent variables
are the three situation variables (information, distribution, and cost)
with their interactions, and the GPA variable. The full TDV^ model is:
TDV1 = V + Oj + Bk + Y, + a6jk + aYj, + 6Ykl + ctSyjkl + Si
The model comparison procedure results for this analysis are
presented in Table 5, and the F values associated with the sources of
the reduced model are presented in Table 6. The results indicate that
the reduced model contains the distribution variable and the cost vari
able main effects. The distribution variable main effect is significant
at the p<.01 level and the cost variable main effect is significant at
the p<.10 level. The means, variances, and sample sizes of the TDV^
measure given the levels of these two independent variables are included
in Table 6. Using the F test of equal variances, the variances between
the levels of the distribution variable differ significantly (F=2.82,
pc.01), with the S2 level having the greater variance. The variances
between the levels of the cost variable do not differ significantly
(F=l.12, n.s.). Using the Z test of equal means, the means of the levels


Variables that can affect the manager's variance investigation
decision process include 1) structure of the decision situation, 2)
contents of the available information set, 3) manager's information
processing efficiency, and 4) manager's learning efficiency. The
structure of the decision situation depends upon number possible states
of control, relative frequencies of the states, various statistical
relationships among the states, and various relationships between the
decision/state outcomes. The contents of the available information set
include information known by the manager prior to his decision. The
manager's information processing efficiency relates to the particular
strategies employed in combining and weighting various items of infor
mation. The manager's learning efficiency refers to the manager's
ability to learn from his experiences with the controlled process. Both
information processing and learning efficiency are the results of the
manager's particular decision strategies. These decision strategies
were expected to adapt a general form, referred to as anchoring and ad
justment. A natural starting point is used as a first approximation for
the investigation decision rule and is then adjusted as the manager
learns from his experiences.
Two experimental methods were used to derive and test implications
of the conceptual framework. Simulation techniques were used to opera
tionalize hypotheses and a laboratory experiment was used to test these
hypotheses.
The experimental environment was that of an assembly department
within a simulated manufacturing company. The assembly department as
sembled a single product and performance of this department was deter
mined completely by the assembly workers' labor efficiency. Student
IX


12
not statistically significant. Whether a variance resulting from a
change in the physical system, labeled a Type 2 variance, requires an
operational control response depends on whether the underlying cause of
the change is a temporary rather than a permanent phenomena. A tem
porary or controllable variance, labeled a Type 2a variance, is one
that operational control can correct in the future (i.e., the physical
system can be returned to the previous in-control condition). A per
manent or noncontrollable variance, labeled a Type 2b variance, is one
that operational control can not correct in the future.
If the manager decides that the variance is of Type 1 no action
is required. If, however, the manager decides that the variance is of
Type 2 he must then decide if the variance sh