Citation
Test anxiety and academic self-efficacy as predictors of cognitive task performance

Material Information

Title:
Test anxiety and academic self-efficacy as predictors of cognitive task performance
Creator:
Lax, Ruth Laurie, 1952-
Publication Date:
Language:
English
Physical Description:
vii, 114 leaves : ; 28 cm.

Subjects

Subjects / Keywords:
Academic achievement ( jstor )
Academic testing ( jstor )
Anagrams ( jstor )
Anxiety ( jstor )
Banduras ( jstor )
Cognitive models ( jstor )
Cognitive psychology ( jstor )
Emotionality ( jstor )
Modeling ( jstor )
Test anxiety ( jstor )
Clinical Psychology thesis Ph. D ( lcsh )
Dissertations, Academic -- Clinical Psychology -- UF ( lcsh )
Educational psychology ( lcsh )
Test anxiety ( lcsh )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1982.
Bibliography:
Bibliography: leaves 104-113.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Ruth Laurie Lax.

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:
028816344 ( ALEPH )
09384681 ( OCLC )
ABW4077 ( NOTIS )

Downloads

This item has the following downloads:


Full Text



fl..


TEST ANXIETY
AS PREDICTORS


AND ACADEMIC
OF COGNITIVE


SELF-EFFICACY
TASK PERFORMANCE


RUTH


LAURIE


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


































This


dedicated


ssertation


my parents,


Judith


and


Melvin


Lax.

















ACKNOWLEDGEMENT S


would


Nathan


like


. Perry,


thank


Jr.,


dissertation


providing


committee


support


chairman,


guidance


this


year,


graciously


taking


over


as chairman,


short


notice,


when


former


chairman,


Wiley


Rasbury,


obtained


a new


position


Michigan.


also


greatly


appre-


ciate


time


tha t


Dr. Rasbury


into


formulation


planning


was


stages


invaluable.


this


dissertation.


dissertation


committee,


His assistance


Walter


Cunningham,


Suzanne


Bennett


Johnson,


Dr. Hugh


. Davis,


Jr.,


Sandra


Damico


, provided


many


instructive


comments


, advice,


support.


am especially


grateful


many


hours


that


Randy


Carter


spent


with


reviewing


statistical


analyses,


providing


psychological


support


when


computer


programs


began


grow


rapidly


in number.


He always


made


himself


available,


even


when


schedule


was


extremely


tight,


which

the f


it almost


our


always


undergraduate


was.


students


am also


who


deeply


helped


grate-


me with


data


collection


coding.


was


a mammoth


task;










that


Registrar,


Robert


Fullington,


staff


, put


into


obtaining


aptitude


scores


me.


would


also


like


thank


Gary


Rushakoff,


a special


friend,


who


Apple


psychological


taught


Microcomputer.


support have been


outs


technical


text-editing


assi


invaluable


stance


over


past


a half


years.


Warren


Rice


was


a great


source


support.


allowed


use


Apple


Microcomputer


to write


this


dissertation.


He also


helped


make


work


environment


at the


Gainesville


V.A.


Medical


Center


very


conducive


speedy


completion


dissertation.


Finally,


would


like


express


my gratitude


family,


who


have


persevered


with


provided


support


throughout


entire


graduate


career.


It would


have


been


difficult


without


them.

















TABLE


OF CONTENTS


ACKNOWLEDGEMENTS


ABSTRACT


CHAPTER


INTRODUCTION


S S S 51


Cognitive


Emoti


ona


Concomitants


Test


Anxiety


atment


Studies


elf-effi


cacy


Theo


METHOD


Subj
Proc


ects
edure


Statisti


cal Analys


S S S SS S68


THREE

FOUR


RESULTS .

DISCUSSION .


APPENDICES


ANAGRAMS


TASK


INTERCORRELATION


MATRIX


ON THE


INDEPENDENT


AND


DEPENDENT


VARIABLES


FOR


MAIN


SAMPLE


(n=162)


INTERCORRELATION


AND


DEPENDENT


MATRIX


VARIABLES


ON THE


FOR


INDEPENDENT


TOTAL


SAMPLE


REFERENCES


BIOGRAPHICAL


SKETCH


S *


Page

















Abstract


Graduate
in Partial


of Dissertation


Council


Fulfillment


Degree


Doctor


Presented


University


Florida


he Requirements
of Philosophy


TEST


ANXIETY


AND


ACADEMIC


SELF-EFFICACY


AS PREDICTORS


OF COGNITIVE


TASK


PERFORMANCE


Ruth


Laurie


Lax


August


Chairman:


Major


Nathan


Department:


Perry,


Clinical


Jr.

Psychology


This


study


assessed


effectiveness


test


anxiety,


academic


self-efficacy,


aptitude


as predictors


cognitive


task


performance


under


neutral,


evaluative,


reassuring


task


instructions.


was


hypothe


sized


that


although


test


anxiety


been


found


to predict


cognitive


task


tions,

would


performance


under


variables


enhance


prediction.


reassuring


academic


was


neutral


task


self-efficacy


theorized


that


condi-


aptitude


a person


academic


self-efficacy,


or their


beli


their


ability


cope


with


or master


academic


situations,


would










cognitive


task


performance


beyond


knowledge


a person


level


sized


test


that


anxiety,


aptitude


alone.


would


In addition,


account


was


a significant


hypothe-


amount


variance


cognitive task


performance.


was


found


that


test


anxiety


predict


cognitive


task


performance,


as measured


an anagrams


task,


final


grade


General


Psychology.


However,


test


anxiety


predicted


rather t

efficacy


performance


han


under


differentially,


predicted


cognitive


task


instructional


as predicted.


task


Academic


performance,


conditions,


self-


in inter-


action


with


test


anxiety,


anagrams


task,


alone.


Aptitude


scores


predicted


grade


performance


but


anagram


task


performance.


results


were


considered


speculative,


because


predictor


variables


only


accounted


a small


percentage


variance


task


performance.


findings


were


discussed


terms


methodological


improvements


academic


self-efficacy


measure,


terms


of other


possible


relevant


predictor


variables.

















CHAPTER


INTRODUCTION


It has


been


generally


found


that


there


is a correla-


tion


between


aptitude


measures


(e.g.


I.Q.


measures,


Scholastic

elementary,


Aptitude


Test)


secondary,


and


academic


college


performance


populations


(Plant


Richardson,


1958;


Fishman


Pasanella,


1960;


Lavin,


1965;


Conry

1969;


Plant,


Cronbach,


1965;

1970;


Dudek,


Goldberg,


McClelland,


1973


Lester,

). The


Harris


correlations


tend


to be


somewhat


higher


elementary


high


school


years,


than


college


or older


populations


(Brody


Brody,

between


1976).


aptitude


lower


correlations


grades


in college


for

and


relationship


other


select


populations,


are


most


restricted


range


of aptitude


test


scores


found


these


populations.


While


aptitude


measures


account


a large


proportion


variance


in cognitive


performance,


other


factors


are


considered


important,


especially


populations


exhibiting


a narrow


range


of aptitude


scores.


These


factors


generally


fall


into


domains


of "personality"


"coanitive


style


or motivational


factors.


... .


most


v&











individuals


perform


most


poorly


on cognitive


tasks


under


conditions


evaluation


of high


evaluative


minimal.


stress,


reverse


optimally


tends


to be


when


true


test-anxious


individuals


(Mandler


Sarason,


1952


Sarason,

Eriksen,


Mandler

1964;


Craighill,


Russell


Sarason,


Ganzer,


1965;


1968;


Sarason,


Paul

1958,


1959,


1961).


Wher


eas


test


anxiety


been


found


to be


related


popular


to a cognitive


explanation


performance


cause


decrement,


this


historically


decrement


performance


been


autonomic


reactivity


or what


Liebert


Morris


research


(1967)


suggested


have


that


termed


"emotionality."


cognitive


components


Recent


test


anxiety,


rather


than


autonomic


reactivity,


have


greater


predictive validity


regard


to cognitive


performance


under


stressful


evaluative


circumstances


(Neale


Katahn,


1968;


Sarason


Stoops,


1978;


Deff


enbacher,


1978;


Dusek,


Mergler,


Kermis


1976


Denney,


1980).


These


cognitive


components


include


worry


over


performance


, and


comparison


with


how


well


others


are


doing.


rumination


over


alternatives.


being


preoccupi


ed with


such


things


feelings


inadequacy


, anticipation


punishment.


.oss of


status


-b .


esteem


uiiu











While


earlier


definitions


test


anxiety


empha-


sized


a common


denominator


"anxiety,


or emotional


autonomic


reactivity


(Wine,


1971) ,


this


common


denominator


does


appear


to capture


most


outstanding


differences


between


persons


who


score


at extremes


test


anxiety


measures. T

of cognitive


thrust


cognitive


hese


differences


structures


current


(Wine,


test


interpretations


appear


1980).


anxi


test


There


research


anxiety.


This


nature

change


toward


disserta-


tion


study


examines


relationship


test


anxiety


other

more


cognitive process

specifically, the


variables

cognitive


an effort


processes


to delineate


underlying


performance


decrement


test


anxi


ety.


Test


anxiety


research


initially


focused


(Sarason,


1961)


on emotional

relationship


reactivity


test


anxiety


mediating

academic


variable


performance.


the

More


recent


research


suggested


that


test


anxiety


strong


cognitive


component,


as well.


It has


been


proposed


that


this


cognitive


component


accounts


much


varlanc

review

factors


task


examines


test


performance


(Wine,


contribution


anxiety,


as both


1980).


This


cognitive


relate


to cognit


literature

emotional

ive task


performance.


Experimental


studies


will


be discussed


from


q0ryry 1


I v~rsP


nf rpsparch.


which


bear


uoon


this


;I rP ;1S


.


.











relevant


an understanding


relative


contribution


of cognitive


emotional


components,


test


anxiety.


Treatment


studies


have


utilized


either


cognitive,


emotionality


reduction


approaches


, or combinations


these


approaches,


to attempt


to alleviate


tes t


anxiety.


An examination


effectiveness


these


various


treatment


modaliti


will


provide


information


underlying


mediating


variable s


test


anxiety.


Cognitive
Concomitants


Emotional
Test Anxiety


"Direction


Attention"


Hypothesis


Wine


theory


1980,


test


1971)


developed


"evaluational"


a cognitive-attentional


anxiety,


"direction


of attention"


hypothesis.


proposed


that


performance


differences


high


test-anxious


individuals


reflect


basic


differences


their


task


atten-


tion


during


evaluative


tasks.


The


high


test-anxious


individual


self-preoccupied


hypothesized


worry


to divide


task


cues.


attention


between


test-anxious


individual


presumably


focused


more


fully


task-











thinking.


There


a significant


amount


evidence


from


experimental


that


treatment


a cognitive-attentional


literature


approach


which


test


suggests


anxiety


important


in understanding


construct,


perhaps


more ;so


than


traditional


autonomic


arousal


interpretation


test


anxiety


(Wine,


1971;


Sarason,


1975;


Denney,


1980).


Worry-Emotionality


Distinction


A closely


related


theoretical


development


to Wine


(1971)


"direction


Emotionality


attention"


distinction


hypothesis,


developed


Liebert


the

and


Worry-

Morris


(1967).


These


authors


have


analyzed


test


anxiety


into


components


Worry


which


focusing


attention


on concerns


about


performance,


consequences


negative


self-evaluation,


evaluation


one


s ability


relative


to others


, and,


Emotionality,


which


refers


affective-physiological


experience


generated


from


increased


there fore


autonomic


, takes


arousal


form


Interfering


of heightened


test


worry


anxiety,


and/or


emotionality.


Two


separate


self-report


scales


have


been


developed


measure


the


Worry


and


Emotionality


components


test











Inventory


Test


Anxiety


(Osterhouse,


1972).


Worry


items


include


not


feeling


confident


about


performance,


worrying


a great


deal,


thinking


much


brighter


others


are,


thinking


about


consequences


failure,


eling


as prepared


as possible.


Emotionality


items


include


so nervous


cannot


remember


facts,


heart


beating


faster,


upset


stomach


, d)


uneasy,


upset


feeling,


feeling


anxiety.


Worry


Emotionality


scores


Worry-Emotionality


Inventory


have


been


related


to performance


expectations


regarding


examinations,


actual


examination


performance


Several


studies


have


shown


that


Worry


was


invers


related


to performance


expectations


of college


high


school


students


Liebert


taking


Morris,


ssroom


1967


exams


Morris


ctor


Liebert,


Altman,


1970).


1969;


Perfor-


mance


expectancy


was


measured


having


subj


ects


assess


how well


Whereas


expec


they


Worry


stations


expected


was


to perform


consistently


, Emotionality


was


on a particular


related


inversely


test.


to performance


related


rformance


expectations


some


samples


(Doctor


Altman,


1969;


Morris


Liebert


, 1970;


egler,


Morris,


Liebert,


1968) ,


unrelated


others


(Liebert


Morris,


1967;


Morri


Liebert,


1970;


Spiegler


Morris


Liebert,


1Q9C~IR


Tn ~11


l~ ..s


n rn


I I I I I I II "


s amnl1 A S


MnrrT i


1111


f-.I


i ii


I











A similar


pattern


exists


relation


Worry


Emotionality


to actual


performance


on exams.


Morris


and


Liebert


(1970


reported


that


in a college


sample,


Worry,


but


Emotionality,


formed


a significant


inverse


relation-


ship


with


test


performance,


and


Worry-performance


relationship


was


stronger


than


Emotionality-performance


relationship


high


school


sample


, however,


both


Worry


Emotionality


were


inverse


related


to performance,


and

both


the

age


strengths

samples,


relationships


Worry


Emotionality


were


differ.


significantly


correlated


with


each


other,


consequently


shared


considerable


variance.


Partial


correlation


analyses


revealed,


however,


that


both


samples,


when


common


variance


between


Worry


Emotionality


was


controlled,


Worry


was


negatively


correlated


with


test


grades,


whereas


Emotionality


was


longer


significantly


related


test


performan


In five

Deffenbacher


Morris


unpublished

(1980)


Liebert


studies

found re


(1970)


involving


sults


college


simil


college

ar to t


sample.


students,


hose


Worry


was


consistently


negatively


related


exam


performance,


while


Emotional ty


related


on some


occasions


on others.


Again,
aF. an4-e..


partial


correlations


n; Wn44nF


T.7Cr


demonstrated

nTr4-4a1 1cr


that

nI- _-


when

Wnr rvT\


the

rnnti nnrii


ce.











significantly


correlated


with


performance.


This


suggests


that


when


common


variance


removed,


Worry,


but


Emotionality,


significantly


negatively


correlated


with


performance


on college


exams.


Deffenbacher


(1977)


also investigated


relation-


ship


of Worry


and


Emotionality


to performance


Miller


Analogies


Test


(MAT),


test


which


is important


because


is used


graduate


admissions,


but


which


little


preparation


possible


there


no specific


knowledge


that


some


people


have


access


over


others,


in regard


to performance


task).


Deffenbacher


found


that


both


Worry


Emotionality


MAT.


When


were


partial


negatively


correlations


related

were ca


to performance


Iculated,


previous


studies,


only


Worry


continued


to form


significant


relationship


with


performance


MAT.


Further


analysis


revealed


a complex


relationship


among


Worry


, Emotionality


, and


performance.


Worry


was


broadly


related


to performance.


Subjects


high


on Worry


solved


fewer


analogies


than


subjects


on Worry.


effects


Emotionality


varied


with


Worry


level.


At low


levels


Worry,


Emotionality


was


unrelated


to performance,


high


levels


Worry,


high


Emotionality


was


debilitative.


That


, the


negative


ects


of Emotionality


were


nested


rn1; 4-1 4


4-a lrha aC


Wtnrr -t


iah 1P


Wnrrv


cinn tr ibuted


r3n rr~


I











Morris


Liebert


(1969)


employed


Worry


Emotionality


a design


which


manipulated


level


stress


through


obvious


non-obvious


timing


and


difficulty


items


timed


subtests


Wechsler


Adult


Intelligence


Scale.


Worry


level


was


found


to interact


with


timing


difficulty


level.


Subjects


were


high


on Worry


performed


better


untimed


than


timed


condition.


reverse


was


true


subjects


on Worry.


They


performed


better


timed


condition.


difficulty


or timing


Emotionality


conditions


evel


was


interact


unrelated


with


per-


formance


this


study.


Finally,


Morris


Perez


(1972)


found


that


Worry,


Emotionality,


was


significantly


negatively


related


reading


test


performance.


Interestingly,


although


correlation


was


significant,


Emotionality


was


positively


correlated


with


performance.


appears


that


a fairly


consistent


pattern


results


from


emerges


Worry


correlated


research


Emotionality


are


(correlations


Worry


component


Liebert


significantly


range


which


from


consistently


Morris


consi


.76).


forms


group.


stently


However,


a nega-


tive

test


relationship

performance,


with

not


performance

Emotionality.


expectations


actual


findings


4-4~~~ -% 4n 4-f ir rnI n Un I


man r


SiF~t


Emotion-


,t


n


FmnL: Alla 1; Cir


---


n .


ara


.











Therefore,


expectancies


both


performance


, Emotionality


is not


measures


performance


as consistently


related


to outcome


measures


, as Worry


As Morris


Liebert


(1970)


point


failure


to find


consistent


differences


Emotionality,


among


performance


stress


level


expectancy


groups


groups.


, may


These


reflect


studies


a generally


used


high


actual


examinations


tasks


to rate


expectancy


against.


This


situation


have


provided


subjects


with


greater


stress


than


less familiar


tasks


lesser


importance.


high


stress


level


tasks


these


studies


may


have


obscured


differences


among


expectancy


levels,


ch might


have


been


seen


with


less immediately


relevant


tasks.


It is


also


conceivable


that


autonomic


symptoms


are


only


attends


subject


relevant


them.


attends


to performance


That


to his


affective


extent


Worry


that


level


reaction.


subject


high,


Worry


present,


affective


reaction


is not


attended


Thus


Emotionality


would


only


be related


to performance


under


high


Worry


levels.


Emotionality


variable


Morris


Liebert


(1969,


1970)


studies


refers


to self-report


of affective


reac-


tivity.


It is


important


examine


relationship


between


S S


I ~I


m











self-report


emotional


anxious


arousal,


arousal,


performance


physiological


measures,


indices


in high


test-anxious


subjects.


Holroyd


et al.


(1978)


study


found


that


high


test-anxious


individuals


performed


more


poorly


reported


higher


levels


anxious


arousal


worry


in an analog


testing


situation,


than


test- anxious


individuals.


This


consistent


with


previous


findings.


During


testing


situation,


actual


autonomic


arousal


was


monitored


continuously.


Interes


tingly


results


from


these


measures


indicated


that


obtained


differ


ences


performance


were


accompanied


corresponding


difference


autonomic


reactivity.


High


t-anxious


individuals


manife


sted


virtually


identical


tonic


phasic


electrodermal


activity


heart


rate


responses.


Only


heart


rate


variability,


which


appeared


to reflect


differences


cognitive


attentional


responses


test-


anxious


groups,


successfully


differentiated


high


test-anxious


individuals.


The results


lend


further


support


to the


theory


that


cognitive


performance


defi


cits


high


test-anxious


subjects


are


not


a result


of maladaptive


levels


autonomic


arousal.


Rather,


appears


that


test


anxiety


more


accurately


conceptualized


as a cognitive-


attentional


phenomenon,


as has


been


proposed


Wine


(1971).











Attention


to Task


Cues


an effort


to identify


cognitive-attentional


factors


related


to test


anxiety,


a group


of researchers


is examining


utilization


task


cues


high


test-anxious


subjects


purpose


research


assess


whether


high


test-anxious


subjects


diff


erentially


attend


task-relevant


stimuli.


Quality


performance


on a


task


depends


extent


to which


task-relevant


stimuli


are


attended


to and


utilized.


If relevant


stimuli


are


attended


then


performance


will


be impaired.


Current


research


area


test


anxiety


task


cue


utilization


been


generated


from


earlier


motiva-


tional


or drive


theory


of Tolman


(1948).


influence


motivation

(1948).


on cue


He listed


utilization


strong


was


recognized


motivational


Tolman


conditions"


as one


of several


causes


formation


narrow


cognitive


maps


in animals


humans.


Bartlett


(1950)


introduced


concept


"range


cues


utilized"


examine


effects


of motivation


culminated


task


performance.


an influential


paper


The

by E


early


research


terbrook


(1959).


main


point


of the Easterbrook


paper


1959)


was


that


"when


dire


action


of behavior is


constant,


increase


in drive











analysis


that


drive


either


facilitate


or inhibit


performance,


depending


importance


peripheral


cues


successful


performance.


Although


there


been


much


research


to indicate


generated


that


in this


test


area,


anxiety


initial


related


studies


to narrowed


seem


cue


utilization.


Interestingly,


test-anxious


subjects


appear

than h


to be


igh


more


affected


test-anxious


subj


ects.


presence


added


rformance


task

the


cues

low


test-anxious


person


facilitated


more


addition


relevant


cues,


and


debilitated


more


addition


irrele-


vant


cues.


That


performance


test-anxious


subjects


reflects


a wider


range


task


cue


utilization,


accordance


with


terbrook


hypothe


Wine,


1980).


West,


Lee,


Ander


son


(1969)


tested


children


who


scored


high


or low


on the


Test


Anxiety


Scale


Children


(Sarason,


Davidson,


Lighthall,


Waite,


Ruebush,


1960)


with


forty


arithmetic


"story"


problems


constructed


such


a way


that


some


contained


additional


irrelevant


information.


Although


highly


level


test-anxious


performance


subjects


overall


than


showed


a slightly


less


poorer


test-anxious


group,


their


performance


was


not


hindered


addition


of irrelevant


information.


The


test


-anxious


subj


ects


performed


more


efficiently


unembellished


items


than


fnnas


J!* -A- .5- .5-


elevant


cues.


Geen


(1976)


found


-L .L -1


I rll











task


required


subject


to remember


which


a set


several


stimuli


was


one


designated


correct


on each


trials.


two


conditions,


additional


cues


were


added


task.


task.


one


other


case,

case,


the

the


cues

cues


aided

were i


solution


relevant


potentially


distracting


task


solution.


High


test-anxious


subjects


were


helped


ess


addition


relevant


information


than


test-anxious


subjects.


They


were


also


less


hindered


their


recall,


insertion


irrelevant


information.


In both


cases


, subjects


who


scored


high


test


anxiety,


behaved


they


were


ess


affected


additional


cues


than


those


test


anxiety.


This


finding


consis


tent


with


an assumption


a restricted


range


cue


utilization.


Whatever


underlying


mechanism


narrowed


cue


utilization,


this


restriction


cue


use


appears


repre


sent


a shift


in attentional


bias


priorities


high


test-anxious


subj


ects.


Wine


1980)


Geen


(1980)


propose


that


"primary


task"


high


test-anxious


situation


task


subject


in terms


toward


assigned


of diversion


self.


Concentration


task.


attention


Wine


away


attention


describes


from


inwardly


toward


self


draws


attention


away


from


task-related


stimuli.


Hamilton


(1975)


proposed


an interesting


model


~4-4-an 4- I-t


rol maIQ/


-r, Wi nI


Shvnnnth~sis


a











capacity
task. TI


attention


either


selective


tion


left


ie person


over
may


to external,


narrowing
attention


to a small


demands


respond


paying


task-related


the a
or by


number


rea


focused


restricting


less


inputs,


for
atten-


of foci.


(Hamilton,


1975,


These


cue


utilization


studies


support


proposition


that


high


test-anxious


person


does


attend


to relevant


task


cues,


as much


test-anxious


person.


Further


res


earch


will


necessary


to d


determine


whether


attention


high


test-anxious


person


is diverted


to internal


concerns.


cue


utilization


research


does


support


view


that


cognitive


factors


are


relevant


an understanding


construct


test


anxiety.


This


should


prove


to be


a fruitful


area


research


furthering


understanding


cognitive


processing


test-anxious


populations.


Social-Evaluative


Variables


While


attentive


high


test-anxious


task-relevant


cues,


subject


may


may


more


be entirely


attentive


social-evaluative


cues


than


test-anxious


subject


Wine,


1980


Dusek


, 1980).


There


some


evidence


to support


this


notion


in experimental


modeling


research.


Most


this


r ra ar a r n n^ a- h o I


~srs r~r


colleagues


rCL Cld ~ r nk


kaan


SI


n


n = _











test


anxiety


vigilance


may


with


contribute


regard


to increased


to possibly


helpful


cues


test-anxious


environment
individuals


S. because


are


more


high


insecure


than
more
cues


other persons
interested a


behavior
(Sarason


strange
s and a


, they ma
nd active


situations


attitudes


from


et al., 1968,


be relatively
n attending to
nd in "borrowing"
models.


496-


497)


an initial


effects


associates.


study,


modeling


There


Sarason


test


was


et al.


anxiety


a significant


(1968)


studied


on learning


anxiety


paired


conditions


intera


action.


high


middle


test


anxiety


groups


demon-


stated


superior


performance


task,


situations


where


subjects


could


observe


models


learning


lists.


Subsequent


studies


explored


effects


of character-


istics


to models


models


, on


' problem-solving


cognitive


behaviors


performance


feedback


high


test-anxious


subjects.


Sarason


(1973a)


subjects


solve


difficult


anagrams


under


one


of several


experimental


modeling


conditions.


group


observed


a model


silently


solve


sample


problems.


Another r


observed


a model


describing


manipula-


tion


letters


while


solving


anagrams.


third


group


observed


a model


comment


on principles


underlying


success-


solution


no-model


anagrams


control


while


condition


working.


While


There


overall,


was

Slow


also a

test-


anxious


subj


ects


performed


better


than


high


test-anxious











condition


which


model


stated


problem-solving


principles.


Although


not


clear


why


task


performance


high


test-anxious


subje


is generally


inferior


test-anxious


subject


in part,


because


actual


or assumed


paucity


task-relevant


responses.


modeling


situation


which


provided


cognitive


strategies


may


have


given


high


test-anxious


subjects


enough


information


that


they


were


confident


enough


to become


ess


self-


preoccupied,


more


able


to guide


their


own


task-relevant


behavior.


While


these


studies


suggest


that


observation


a model


be beneficial


test-anxious


subject,


other


studies


which


have


examined


ects


exposure


models


who


succeed,


fail,


or are


self-derogatory


are


some-


what


more


difficult


to interpret.


Sarason


1972)


investigated


whether


persons


differing


test


anxiety


would


respond


differentially


to a neutral


model


vs.


a model


is self-


derogatory


regarding


performance.


was


expected


that


the opportunity


to observe


a model


would


have


a facilitative


effect


on all


subjects,


espec


ially


those


high


test


anxiety.


However


was


expe


cted


that


the high


test-anxious


group


would


show


this


facilitative


to a 1


esser


extent


under


condition


involving


self


-derogatory


model.


While


4-I---a.s-'- 1 C.-.414~ rr


CA nf


A~l


c~nun


Alt~H rl 11


mn~


t.t r~ n


1


^-











self-derogatory


model


evaluated


condition


derogated


an authority


herself


figure


observation


performance


to be a failure.


that


while


been


There


might


be a difference


between


observing


someone


who


self-


debasing


thinks


he is


failing


, vs.


someone


who


told


an authority


that


he has


failed


Consequently,


another


experiment


was


conducted


including


success


failure


conditions


determined


an authority


figure.


There


was


test


anxiety


treatment


interaction.


condition


which


model


fail


(given


success


feedback


authority),


anxious


subjects


' performance


was


superior


that


of a middle


test


anxiety


group.


failure


condition


failed


an authority


figure)


high


test-


anxious


group


performed


at a significantly


lowe r


level


than


high


test


anxiety


group


that


was


exposed


nonfailing


model.


test


anxi


group


performed


better


under


failed


mod


condition


than


non-failed


model


condition.


results


experiments


suggest


that


someone


observing


who


someone


is declared


derogating


himself


an authority


. observing


figure


to be


failure


can


have


very


different


effects


on performance


high


test-anxious


subject.


self-derogatory


condition


authority-failure


condition


were


both


apparently


mr~4-, TfZ+ infl i*~ Ut


fPqt -anxi nus


subi


ects.


mhC; tt~C; nn


I rir i(i


I l l


q











observation


of failure


feedback


a much


more


negative


effect,


this


group,


than


observation


a self-


derogatory


model.


Sarason


suggests


that


observing


someone


who


dissatisfied


with


himself


(se If-derogatory


condition),


convincing


an ob


server,


since


there


may


no objective


basis


self-debasing


reaction.


However,


observation


of a negative


authority


figure


is much


more


convincing,


certainly


would


affect


evaluational


anxiety


high


test-anxious


subject.


Another


explana-


tion


lack


of detrimental


impact


self-derogatory


model


lies


the construct


"coping


model,


" described


Meichenbaum


(1972).


self-debasing


model


have


been


seen


high


test-anxious


subject


someone


who


exper-


ces


problems


similar


those


subject,


actively


copes


with


overcomes


those


difficult


es.


Failure


eedback


, however,


is clearcut


evidence


a failure


coping


whether


efforts
*


a coping


(Wine,

model w


1980)


ras,


Sarason


fact,


1975)


facilitative


examined


cogni-


tive


task


performance.


Subj


ects


were


exposed


to a mod


admitted


experiencing


anxiety


described


ways


coping


with


or to


one


who


only


admitted


anxiety,


versus


a model


who


admit


anxiety.


Exposure


self-disclosing


coping


mod


was


highly


facilitative


S S C -


I ___, 1L,,,


-IIE-YIIIIA


I











attention-directing


performance


cognitive


test-anxious


strategies,


persons.


is beneficial


Performance


also


enhanced


there


evidence


that


model


successfully


coping


with


worry


tension


associated


with


test


anxiety.


observation


models


performing


tasks


does


seem


to have


powerful


effects


subsequent


task


performance


of high


test-anxious


subjects.


social-


evaluative


problem-


solving


information


can


have


either


itive


or negative


effects


depending


on the


characteristics


modeling


situation.


These


studies


suggest


that


high


test-anxious


subject


attend


cues


task


situation


other


than


immediate


task


cues


Social-evaluative


feedback


appears


to influence


cognitive


task


performance


high


test-


anxious


subject.


This


further


evidence


a cognitive-


attentional


component


test


anxiety.


Treatment


Studies


an effort


to further


examine


contribution


cognitive


emotional


factors


test


anxiety,


treatment


research


test


anxiety


will


be examined.


Treatment


approaches











effects


these


approaches


on both


cognitive


task


perfor-


mance and


self-report


test


anxiety.


Systematic


Desensitization


Most


early


treatment


literature


assumed


emotional


reactivity


to be


major


character


stic


test


anxiety


As


such


, much


early


treatment


literature


focused on systematic

simple demonstrations


desensiti


zation,


efficacy


progressing


systematic


from

desensi-


tization


(Katahn,


Stronger,


Cherry,


1966;


Garlington


Cotler


, 1968;


Paul,


1964)


through


systematic


manipulations


of the


desensitization


process


Freeling


Shemberg,


1970).


Systematic


desen


sitization


treatment


test


anxiety


basically


involves


pairing


deep


muscle


relax-


ation


with


instructed


imagination


aversive


scenes,


inhibit


autonomic


arousal


provide


a new


response,


to be conditioned


to previously


anxiety-arousing


stimuli.


tematic


desensiti


zation


procedure


followed


general


outline


used


Paul


1966)


one


earliest


demon-


stations


this


treatment


text


anxiety.


Studies


compared


individually-designed


vs.


standardized


desen


siti-


nation


hierarchies


Emery &


Krumboltz,


1967), s


ystema tic











Guerney,


1969) ,


and


other


variations


presenting


systematic


desensitization


treatment


(Aponte


Aponte,


1971;


Feeling


Shemberg,


1970


Suinn,


1968).


In general,


these


studies


have


found


that


systematic


desensitization


ec-


tive


alleviating


test


anxiety


when


self-report


measures


anxiety


level


are


used


as outcome


criteria.


However,


effects


sensitization


on academic


performance


are


quite


in grade


as consistent.


point


Some


average


studies


treated


demonstrate


groups


an increase


compared


control

1969; J


Cherry,


groups


ohnson


1966)


(Aponte


Secrest,


Others


Aponte,


1968;


showed


1971;


Donner


Katahn,


no increment


Guerney,


Strenger,


in academic


performance


Garlington


Cotler,


1968;


Emery


Krumboltz,


1967)


or did


use


performance


measures


as outcome


criteria


(Freeling


Shemberg,


1970;


Suinn,


1968) .


Allen


(1972)


states


fact


that


desensitization


studies


control


effects


of non-specific


therapeutic


factors,


which


might


enhance


treatment


benefits


some


studies


others.


These


non-specific


factors


include


heightened


subj


expectancy,


use


only


one


therapist,


use


first


semester


shmen


(have not yet


adapted


college


routine


may


interpret


their


stress


test


anxie


ty),


lack


appropriate


control


groups


waiting


list


control


arouo).


a .











in self-report


anxiety


level.


However,


an emotionality-


oriented


treatment


approach


be sufficient


con-


sistently


alter


poor


performance


under


academically-evaluative


conditions.


More


extensive,


methodologically-sound


research


effects


of systematic


desensitization


treatments


cognitive


performance,


is required,


to make


more


definitive


statements.


Self-Control


Treatment


Techniques


Denney


1980)


reviewed


a series


self-control


treatment


studi


es.


He orders


studies


on a continuum


on which


one


extreme


represents


approaches


focused


solely


on emotionality


reduction,


other


extreme


represents


approaches


emphasizing


totally


cognitive


approach.


labels


three


basic


types


studies


reviewed


are


following


applied


relaxation,


self-control


training


techniques,


cognitive


coping


technique


es.


one


moves


from


treatment


approaches


relying


heavily


on emotionality


reduction


(applied


relaxation)


more


cognitively


ented


treatments


(self-control


training


cognitive


coping),


there


incr


easing


emphasis


on in-vivo


application


training


(training


actual


anxiety-provoking


setting),


guided










client


can


effectively


apply


toward


management


reduction


test


anxiety


when


arises


in real


life


situations.


The more


cognitively-oriented


treatment


approaches


attend


rather


to thoughts,


than


worries


alleviation


negative


emotional


self-statements,


arousal.


Applied


Relaxation


Techniques


Applied


examples


relaxation


self-control


techniques


proce


dures


constitute


simplest


reduction


anxiety.


The procedures


classified


as applied


relaxation


techniques


share


three


common


features


These


procedures


introduce

advocate


ed with


a self-control


Goldfried


(1971).


rational

Clients


similar


are


usually


informed


that


purpose


treatment


to provide


them


with


skills


actively


coping


with


anxiety.


They


are


told


that


the relaxation


training


they


receive


method


helping


them


to bring


response


relaxation


under


voluntary


control


Further,


they


are


told


, that


with


greater


proficiency


at inducing


relaxation


they


will


able


to apply


relaxation


to reduce


feelings


of anxiety


stressful


situations


second


encountered


feature


outside


applied


treatment.


relaxation


techniques











for

have


inducing

been de


relaxation,


veloped,


although


supplementing


numerous


other


relaxation


wit


approaches

h biofeed-


back-assistance


(Romano


Cabianca,


1978


cue-controll


relaxation


(Russell


Sipich,


1973) ,


imag


exerci


ses


(Samuels

3) The


Samuels,


third


1975


feature


other


applied


supplementary


relaxation


exercises.


techniques


involves


training


application


of relaxation


within


stressful


settings


encountered


outside


treatment.


Typically,


these


studies,


the clients


are


merely


instructed


to begin


applying


relaxation


outside


treatment.


Self-control


Training


Techniques


Self-control


training


techniques


include


three


main


features


found


applied


relaxation


techniques.


However,


addition,


self-


control


training


techniques


include


guided


rehearsal,


which


introduced


after


relaxation


induction


training.


During


guided


rehearsal,


clients


are


confronted


with


some


type


of stressful


stimulus


presented


within


consultative


setting.


Usually


stressful


stimuli


are


evoked


use


imagery.


Thus,


there


is greater


personal


involvement


in cognitive


approach


es.











Cognitive


Coping


Techniques


Cognitive


cognitively


with


coping


-oriented


an additional


some


tive


cases


techniques


approaches,


repertoire


instead


feature


which


attempt


coping


relaxation


cognitive


coping


are


most


to provide


skills


skills.


techniques


clients


beyond,


distinc-


is some


form


of cognitive


restructuring.


Cognitive


restructuring


three


basic


objectives


Clients


are


persuaded


that


beliefs


that


they


ente


rtain


while


confronting


certain


situations


affect


emotional


reactions


they


have


these


situations,


They


are,


next,


encouraged


to identify


particular


negative


self-


statements


that


they


make


when


confronting


a stressful


situation,


Clients


are


helped


to formulate


positive


rational


self-statements


which


will


replace


attenuate


negative


their


stress


self-statements


evaluative


thereby


situation.


As discussed


earlier


in regard


to the systematic


desensitization


studies,


most


appli


relaxation


studies


(Russell,


Stratoudakis,


1976;


Miller,


March


June


etti,


, 1975;


McGlynn,


Russe


Wise,


Patterson,


1977;


Chiang-Liang,


Denney,


1976)


found


that


this


treatment


approach


demonstrated


test-anxiety.


same


changes


on self-report


generalization


could


measures


be applie











Deffenbacher


cognitive


Parks,


coping


1979;


techniques


Denney


Rupert,


(Meichenbaum,


1977),


1972;


Holroyd


1976;


Lavigne,


1974;


Wine,


1970;


Goldfried,


Linehan,


Smith,


1978).


treatment


Analyses


strategies


comparative


on cognitive


impact


performance


demon


these


states


quite


different


findings.


Only


one


out


three


of the


applied


relaxation


studies


resulted


in cognitive


performance


improvement.


was


found


that


four


eight


self-control


training


mance


studies


measures,


reviewed,


whereas


revealed


five


improvements


seve n


on perfor-


cognitive-coping


technique


studies


measures.


demonstrated


Thus,


cognitive


improvements


coping


in these


techniques,


types


which


incorporate


some


form


of cognitive


restructuring,


are


observed


to have


a higher


rate


success


than


self-control


training


or applied


relaxation


hniques,


terms


performance


improvement.


Both


Meichenbaum


(1972)


Holroyd


1976)


compared


cognitive


coping


treatment


approaches


directly


with


systematic


desensitization,


Meichenbaum'


which


"cognitiv


is an arousal r

e modification"


educationn

treatment


technique.

t procedure


combined


an insight-oriented


therapy


which


was


designed


make


test-anxious


subjects


aware


their


anxiety-engendering


thoughts,


with


a modified


desensitization


procedure


an~~~~n1 r~~~~naA = ~~~~Inn% fn- n.a. -nn hr h nA aln I;an


Cn


;Inv; a5~~


bmn 1 h~lL~;I


kntrr


~m~n


JT1











that


cognitive


modification


group


was


most


effective


significantly


reducing


test


anxiety


as ass


essed


test


performance


obtained


an analogue


test


situation,


self-


report


anxiety


immediately


after


treatment


one-month


follow-up,


grade


point


average.


There


was


ess


consistent,


general


improvement


desensitiza-


tion


treatment


group,


who


appear


ed significantly


more


improved


than


waiting


list


controls.


Holroyd


Meichenbaum


(1976)

(1972)


study

study


is actually a f

. Meichenbaum'


sollow-up

s results


study

are


somewhat


difficult


to interpret


because


cognitive


modification


treatment


actually


a combination


cognitive

"coping"


treatment


imagery)


(the


use


of self-statements


an emotional


arousal


treatment


cognitive

(the


use


of relaxation).


It is


clear


whether


positive


results


cognitive


modification


treatment


are


primarily


cognitive


or the


combination


aspects,


arousal


treatment


reduction


forms.


aspects,


Holroyd


sought


to separate


these


individual


treatment


components.


He compared


combined


effectiveness


arousal-cognitive


cognitive,


treatments


arousal


test


reduction,


anxiety.


cognitive


test-anxious


therapy


client


focu


s task


exclusively


-irrelevant


on eliminating


ruminations


.m4-4- f;nn+- nn~ fnii


rl i pnt s


were


taught


to emit


incom-


fn rrl e


I ilr











was


traditional


group


systematic


desensitization


paradigm


(Paul,


1966).


combined


treatment


focused


both


eliminating


test-anxious


client


task- irrelevant


ruminations


task


focus,


and


reducing


their


emotional


arousal.


There


was


also


a pseudotherapy


control


group


included


study,


assess


degree


of improvement


attributed


to nonspecific


treatment


effects.


results


demonstrated


that


cognitive


therapy


approach

treatment


was


significantly


approaches.


more


effective


treatment


gains


than

were


the other

seen in both


laboratory


performance


average.


measures


test


an analog


Only


significantly


test


cognitive


better


results


anxiety


(self-report),


situation,


therapy


than


approach


grade


test


point


demonstrated


pseudotherapy


control


procedure.


systematic


desensitization


group,


combined


produced


situation


treatment


similar


(digit


group,


outcomes,


symbol


and


pseudotherapy


except


test


performance),


analog


groups


test


where


combined


treatment


approach


was


superior


systematic


desensitization


pseudotherapy


approach.


results


suggest


that


while


systematic


desensitization


an effective


treatment


test


anxiety,


may


no more


ective


than


an equally


convincing


treatment


ritual


other


sort.


~~I r -. II) *I


rocan 1 *


tr* nm hi n a


annrn-rhb


a rP


s nm PW1 RI -


typ;ltmpnt


1 r **-











being


less


effective


than


cognitive


or systematic


desensi-


tization


treatments


alone


(Meichenbaum,


Gilmore,


Fedoravicius,


1971) ,


or more


effective


than


systematic


desensitization


alone


(Meichenbaum,


1972).


Holroyd


(1976)


suggests


that


results


obtained


with


combined


treatments


be influenced


procedural


variables


such


length


treatments,


order


treatments


specific


treatments


which


are


combined.


There


have


been


other


cognitive


coping


approaches


which


have


proven


to be successful


improving


academic


performance


reducing


self-report


test


anxiety.


Recently,


Harris


Johnson


1980)


have


compared


a unique


covert


modeling


approach,


to a self-control


desen


sitization


treatment,


a study-skills


training


group


(with


covert


modeling


group


and


the desensitization


group


receiving


some


study


skills


training,


as well).


Harris


Johnson


(1980)


hypothesized


to benefit

performing


more

the


that


an individual


from a modeling

behavior is the


would


procedure,

individual


be likely


himself


model

There-


fore


they


developed


treatment


in which


individual


utilizes


own


highly


personalized


images


self-competency


stressful


situations


cope


with


test


anxiety.


These


images


are


taken


from


life


situations


other


than


test-taking,


in which

nlavina


individual


in a music


recital


operates


successfully


or runnincr


a radio


for

show.


example


After











from


test-taking


situation.


This


procedure


allows


individual


to utilize


a standardized


Harris


"coping


Johnson


own


skills


found,


resources,


" package


similar


rather


imposed


to previous


than


upon


studies


having


him.


comparing


cognitive


approaches


to desensitization,


that


covert


modeling


procedure


was


superior


to both


self-control


desensiti


zation paradigm and


study-skills


training


group.


individualized


cover t


modeling


procedure


combined


with


study


skills


training


significantly


reduced


test-anxiety,


assessed


significantly


several

increased


self-report

academic D


measures.


performance,


It also

as assessed


grade


point


average.


self-control


desensitization


treatment


was


also


effective


reducing


self-reported


test


anxiety,


but


was


successful


improving


academic


per-


formance.


Subjects


study


skills


training


group


a small,


but


nonsignificant


increase


in grade


point


average,


show


reduction


in test-anxiety


on all


anxiety


measures.


Although


study


skills


training


procedure


presumably


prevented


a decrement


in academic


performance,


was


not


effective


in either


reducing


test-


anxiety


or significantly


increasing


grade


point


average.


Harris


Johnson


(1980)


study


clearly


indicates


that


a cognitive


coping


approach


treatment


in which


positive


- n .....~~~~ nfn 1 n 4-r nan 4-4t ir n;r o -,nr vnr n


0mlnm


rr n rr a I: Ilh


: Mr( ~~ M1I


rr rr An 1 a rr L n











treatment


also


serves


to highlight


individual


own


areas


competence.


Reinforcement


a sense


of mastery


appears


to be


a more


active


treatment


than


muscular


relaxation


Harris


(1980)


investigated


concept


individualized


coping


imagery


In more


detail


in a follow-up


study.


was


interested


effectiveness


different


imagery


instruc-


tion


Harris


procedures


compared


treatment


individualized


test


coping


anxiety.


imagery


Specifically


treatment


based


on non-academic


experiences


of competence


success


individualized


coping


imagery


treatment


based


on academic


experiences


competence


success.


Since


images


competence


utilized


Harris


Johnson


(1980)


study


were


non-academic


instruction


in coping


nature,

images


Harris

based


hypothesized

on previous a


that


academic


success


experiences


might


prove


even


more


effective


reducing


In addition

effective


test


anxiety


Harris


relaxation


improving


wanted


training


assess

would


academic


extent


enhance


performance.


to which

overall


treatment


ectiveness


individualized


coping


imagery


techniques.


Consequently,


she compared


four


treatment


techniques


non-academic


individualized


coping


imagery


with


relaxation


training,


non-academic


individual


t'lii nr r


1 m. cnarrty


144 -hnriil-


r0=1 ava i*-4 nn


tr n i na -


_31 I


academic r











relaxation


individualized


coping


imagery


group


without

control


relaxation


group


wer e


academic


significantly

c performance


different


(grade


from


point


average)


at post-test.


Pre-test


to post-test


treatment


changes


grade


point


average


were


significant


both


non-academic


imagery


groups


.05)


approached


significance


academic


coping


imagery


group


with


relaxation


.089).


academic


coping


imagery


treatment


without


relaxation


show


a significant


pre- treatment


to post-treatment


improvement


grade


point


average.


All


variations


individualized


coping


imagery


manifested


significant


decreases


test


anxiety


on self-report


measures.


apparent


from


results


this


study,


that


type


imagery


treatment


performance


instruction


effectiveness,


improvement.


used


treatment


particularly


academic


a bearing


in regard


imagery


to academic


condition


without


relaxation


proved


to enhance


treatment


effective-


ness.


Harris


(1980)


hypothesized


that the academic


coping


imagery

procedure


treatment

e because


without

academic


relaxation


images


was


were


a less


likely


effective


to elicit


fear


avoidance


rather


than


images


feelings


competence


these


subjects,


since


academic


experiences


are


so closely


linked


test-taking


situation.


Con-


-~~~~~~ ~ ~ ~ ~ -I -- -I..* ,re .


I~'.~.~


I I. I. 'I


I


C











have


allowed


subjects


to be


more


comfortable


about


academic


imagery,


since


relaxation


training


been


hypothesized


to reduce


distracting


thoughts


physiolog-


ical


imagery


responses


with


(Lang,


relaxation


1979) .


Hence,


treatment


was


academic


more


coping


efficacious


than


non-relaxation


group,


desensitization


hierarchy


was


completed


more


efficiently.


relaxation


component


significantly


effectiveness


non-academic


individualized


coping


imagery


paradigm,


presumably


because


subjects


were


initially


more


able


to visualize


non-academic


imagery


without


fear


or avoidance.


Harri


s (1980)


study


Harris


Johnson


(1980)


study


highlight


findings


previous


research,


that


while


both


emotionality


cognitively-oriented


treatments


are


effective


improving


self-report


test


anxiety,


cognitive


treatment


approaches


multimodal


approaches


whi


ch alleviate


cognitive


performance


deficits


test-anxious


person.


Further-


more


, the


Harris


(1980)


study


indicates


that


multimodal


approaches


necessary,


if a powerful


cognitive


coping


model


applie


such


non-academic


individualized


coping


imagery


paradigm


utilized


in that


study.











General


Conclusions


While


most


early


research


area


test


anxiety


was


focused


upon


debilitating


effects


test


anxiety


(Wine,


1971),


current


research


has begun


to investi-


gate


contribution


cognitive


affective


variables


understanding


construct


test


anxiety.


Research


in the


areas


task


cue


utilization,


social-evaluative


cues,


worry-emotionality


distinction,


treatment


approaches


to test

appear


anxiety,


to be


demonstrated


operative


that


relations


cognitive

ip between


factors

test


anxiety


cognitive


task


performance


in evaluative


situations.


High


test-anxious


subjects


appear


to be


less


attentive


task


cues


more


attentive


to social-evalua-


tive


cues.


Cognitive


treatment


approaches


which


address


negative


self-statements,


task-irrelevant


attentional


patterns,


or negative


imagery


test-anxious


individual


appear


to enhance


task


performance.


While


these


studies


have


anxie


suggested


are


that


related


cognitive


to cognitive


concomitants


task


test


performance,


description


a very


spec


global


ific


evel.


cognitive


That


concomitants


high


remains


test-anxious


individual


considered


to be negatively


self-preoccupied,


test-anxious


person


is presumably


k-oriented











as more


fully


understand


test


anxiety


as a cognitive


construct,


we need


to investigate


extent


to which


test


anxiety


related


to specific


cognitive-mediational


variables.


Self-Efficacy


Theory


cognitive-mediational


variable


which


been


examined primarily


relation


to phobic


behavior,


variable


originally


self-efficacy.


conceptualized


term


Bandura


"self-efficacy"


1977)


was


as a mastery


expectancy


variable


which


could


be a powerful


predi


ctor


current

change


basic


level


following


premise


behavior, a

therapeutic


theory


s well


as a predictor


intervention.


that


In fact,


client


behavior

the


s efficacy


expectations


are


central


to behavior


change.


theory


pro-


poses


that


expectations


of personal


effi


cacy


"determine


whether


coping


behavior


will


be initiated,


much


effort


will


expanded,


long


will


sus


tainted


face


obstacles


aversive


circumstance


s" (Bandura,


1977,


191).


Through


experiences


of mastery


provided


via


various


thera-


peutic


modes,


there


enhancement


self-efficacy











gain


corrective


experiences


that


reinforce


their


sense


self-efficacy


(Bandura,


1977).


When


individuals


cease


their


coping


efforts


prematurely,


they


retain


their


debilitating


expectations


thus,


choice


fears


presumed


of behavioral


efinit


to affect


settings


ely.


coping


and


Perceived


efforts,


activities.


self-efficacy


as well


Bandura


(1977)


emphasizes


that


expect


station


sole


deter-


minant


behavior.


Expectation


of mastery,


alone,


will


produce


desired


performance


component


capabilities


are


lacking.


In addition,


there


are


many


tasks


people


can


do successfully,


which


are


performed,


because


adequate


incentives


are


not


present


However,


given


appropriate


skills


adequate


incentives,


efficacy


expectations


are


seen


Bandura


colleagues


(Bandura


Adams,


1977


Bandura,


Adams,


Beyer


, 1977),


as a major


deter-


minant


of people


s choice


of activity


employment


their


coping


efforts,


their


persistence


coping


effort


when


dealing


with


stressful


situations.


In Bandura


theoretical


analysis


, expectations


self-efficacy


are


based


on four


sources


information:


performance


accomplishments,


vicarious


experience


, 3)


verbal


persuasion,


emotional


arousal.


Bandura


(1977)


predicted


that


therapeutic


interventions


employing


enactive


- t I


, I _


t I











interpretive


treatments


desensitization).


Bandura,


or emotional


Adams,


arousal


Beyer


(relaxation,


(1977)


des

shi


signed an

p between


compare


experiment


self-efficacy

effectiveness


test

and


hypothesized


behavioral


enactive


relation-


change,


vicarious


exper-


ience


in enhancing


efficacy


expectations.


Severely


snake-


phobic


subjects


were


ass


signed


to a part


cipant


modeling


condition,


traditional


modeling


condition,


or a no-


treatment


control


group.


magnitude


subj


ects


self-efficacy


was


assesse


d by


giving


each


subject


a list


of the


tasks


used


on a behavioral


avoidance


pretest


(related


to approaching


a snake).


subjects


were


asked


to rate


tasks


they


expected


to perform


successfully.


stre


ngth


self-efficacy


was


asses


asking


each


subject


to rate


expectancy


success


each


task


terms


of a 100


point


scale


divided


into


intervals.


generality


of self-effi


cacv


was


ass


essed


asking


subjects


to rate


level


(magnitude)


and


strength


their


expectations


with


regard


to successfully


carrying


same


behavioral


task


with


an unfamiliar


snake.


results


supported


hypotheses


that


partic


ipant


modeling


technique,


with


performance


mastery


component,


instilled


high


est,


strongest,


most


generalized


afF4 t-nn~~~v a nr a r. an .4 nrA n' rn0 1n r ie 4rn4mn4


nFC; ~3 ~~t


Fh t 1 hGld rl


tf; a 3r; nrr a


Crr r a 3 CM an C


tnn


r-%.Fy











that


behavior.


self-efficacy


As a further


accurately


extension


predict


theory,


future


Bandura


Adams


(1977)


examined


role


of reducing


emotional


arousal


through


desensitization,


on self


-efficacy.


results


arousal


supported


through


hypotheses


desensitization


that


would


eliminating


enhance


emotional


subjects'


levels


self-efficacy.


was


also


found


that


level


self-efficacy


varne


d for


each


subject,


that


higher


stronger


level


self


-effi


cacy


induced


treatment


, the


greater


reduction


in avoidance


behavior.


level


self-efficacy


closely


corresponded


to approach


behavior,


and


accurately


predicted


subsequent


behavior


similar


dissimilar


tasks.


summary,


Bandura


Adams


(1977)


Bandura,


Adams,


Beyer


(1977)


demon-


strated


that


successes


reported


varying


techniques


of enactive


performance)


mastery,


vicarious


experience,


desen


sitization


can


readily


adequately


have


explained


also demonstrated


through


that


self-efficacy


level


theory.


efficacy


They


estab-


lished


intervention


significantly


affects


therapeutic


success


in overcoming


fearful


avoidant


behavior.


res


ults


suggest


that


efficacy


actually


a better


predictor


of future


behavior,


than


is past


behavior.


It is


more


--











1977,


Bandura,


Adams,


Hardy,


Howell


in press).


seems


likely


that


level


self-efficacy


might


prove


an exce


llent


predictor


of academic


success


and


career


choice.


Extending


Bandura


s res


ults


achievement


situation,


one


would


assume


that


students


with


levels


academic


self-efficacy


would


fear


avoid


subj


ects


extracurricular


outside


their


activities


repertoire


which


of coping


they


skills.


believed

These


lay

students


might


res


trict


themselves


to academic


programs


perceived


as requiring


lesser


abilities.


They


might


choose


careers


which


are


lower


in socio-


economic


career


status.


They


would


be predicted


to have


poorer


academic


per-


formance


school


Lalonde


(1979


extended


Bandura


s (1977)


theory


investigating


achievement-ori


relationship


ente


between


d behavior.


self


hypothesis


-efficacy


that


expectations


of self-efficacy


would


play


a major


role


poor

rest


grades,


ricted


res


career


tricted

r choice


future

(lower


academic

socio-


expectations,


economic


career


choice)


test


hypotheses


linking


academic


perfor-


mance,


as well


efficacy ex

to academic


academic


pectations,


career


a measure


achievement-oriented


expectations,


of self-efficacy


situations


pertaining


was











Beyer


asking


(1977)


subjects


research.


to rate


This


their


simplified


efficacy


method


expectations


successfully


completing


certain


tasks


was


effective


snake


phobics.


However,


was


so easy


to determine


efficacy


expectations


achievement


situations,


since


subjects


do not


have


a particular


identifiable


tasks


which


indicate


level


their


avoidance


behavior.


Lalonde


(1979) ,


therefore,


investigated


self-efficacy,


related


to achievement


behavior


, by


three


stages


research:


development


a paper


and


pencil


self-report


measure


academic


self-efficacy,


establishment


discriminant


and


convergent


validities


, 3)


undertaking


of a validational


experimental


study.


In order


to develop


items


Measure


Academic


Self


-Efficacy


(MASE),


Bandura


s self-efficacy


theory


research


were


carefully


examined


as background


data


defining


behavioral


characteristics


strongly


weakly


efficacious


people.


To define


as clearly


possible


behavioral


corre


latest


self


-efficacy,


twenty

staff)


judge

were


(graduate


given


psycho logy


a written


students


definition


effica


teaching

cy as well


theor


etical


statements


empirical


results


Bandura


colleagues.


judges


were


asked











classroom


behavior


ensure


that


generality


students


' academic


self-e


fficacy


was


assessed


(the


extent


to which


feelings


effi


cacy


generalize


various


academic


situations).


Bandura


(1977)


stated


that


efficacy


expec-


tancies


varied


along


three


dimensions


magnitude,


generality,


and


strength,


that


an adequate


analy


S'S


of expectancy


must


include


three


dimensions.


Efficacy


strength


generality


were


asse


ssed


employing


a five-


point


response


scale,


tapping


a wide


variety


academic


activities


behaviors.


Magnitude


of efficacy


was


more


difficult


assess


achievement-oriented


situations,


lend


since


themselves


performance


task


achievement


to rank

s relate


situations


ordering


to snakes.


do not


difficulty,

Thus, a mIT


as easily


as do


lagni tude


component


was


included


this


scale.


A reliable


scale


was


deve loped


from


an initial


item


pool


items.


final


form


Measure


Academic


Self-Efficacy


items)


(MASE)


a Hoyt


1941)


estimate


internal


reliability


scale


homog


eneity


.95,


a standard


error


measurement


t-retest


reliability


MASE


was


ove r


a one


month


period.


Convergent


discriminant


validity


MASE


was


assessed


with


a number


measures.


was


predicted


that


-I,-1 .- T1a nnCIO 17


iMsi or


tail-


nrrrP


nt-P ft


nhYlr~l


bth~ll


4m r











inte


llectual


achievement


(Intellectual


Achievement


Respon-


sibility


Questionnaire)


Crandall


, Katkovsky,


Crandall,


1965) ,


career


aspirations


expectations


coded


socioeconomic


self-concept


status


(Piers


post


-Harri


high


school


Children


career


s Self


plans,


-Concept


Scale)


ers


, 1969)


was


also


hypothesized


that


self


-efficacy


would


correlate


negatively


with


social


anxiety


social


Avoidance


stress


Scal


Watson


Friend,


1969


test


anxiety


(Test


Anxiety


Scal


, Sarason,


1958)


Finally,


an experiment


was


designed


to behaviorally


validate


MASE


testing


hypotheses


derived


from


self-


efficacy


theory


anagrams


task


was


designed


test


hypoth


eses


that


highly-efficacious


high


school


students


would


persist


longer


an anagrams


task


than


would


ess-


efficacious


students


highly-


efficacious


students


would


persi


st longer


face


of failure


than


would


ess-


effi


cacious


students


forced


failure


experienced


earlier


in the


anagrams


task


would


have


a greater


ect


on le


SS-


efficacious


students


, than


would


failure


experienced


later


These


hypotheses


were


based


on hypothe


ses


from


self


-efficacy


theory


which


state


that


highly


self


-efficacious


people


will


ers


longer


ected


at overcoming


failure,


whereas


difficulties


ess


are


self-efficacious


less


persons


are


more


ected


-,


early


experiences


of failure.


Persis-


w


w --











total


anagrams


number


anagrams


returned


proportion


anagrams


skipped,


working


until


forced


to stop


, willing-


ness to

working


return

(if f


to skipped


orced


anagrams,


to finish


prematurely


willingness

y). Failure


to keep

was


defined


terms


of the


early


vs.


late


presentation


of hard


anagrams.


anagrams


That is, an


presented


early


before


failure


medium


experience


easy


hard


anagrams.


late


failure


experience


hard


anagrams


presented


following


presentation


of medium


easy


anagrams.


terms


convergent


discriminant


validity


analyses,

predicted,


the M

with


ASE


correlated


grade


point


moderately


average,


scholast


positively,

ic aptitude,


willingness


to accept


responsibility


itive


intellectual


achievement,


self-concept.


terms


of discriminant


validity


MASE


correlated


moderately


negatively


with


social


and


test


anxiety.


This


dicted


from


self-


efficacy


theory


that


Bandura


(1977)


views


anxiety


efficacy


antithetical.


cannot


be anxious


about


one'


ability


efficacious


at the


same


time.


stronger


efficacy


expectations,


greater


reduction


anxiety


avoidance


behavior.


The

between


hypotheses

academic se


concerning


if-efficacy


proposed

career e


relationship


expectations,


f-*-1 blf r 0 r in l3


Sr'h nri1


ni anc


ciinnnrfsa


i ch~r


jj n n


~rtPYP


n


N











In addition,


with


effects


verbal


intelligence


removed,

positive


academic

predictor


efficacy


grad


scores

e point


remained

average


but

post-


secondary


school


plans.


Analyses


experimental


task


results


failed


support t


correlate


hypo the


with


that


sistence


academic


self-efficacy


an anagrams


task,


would


except


one


instance.


This


was


variable


willingness,


part


subjects


forced


to finish


early,


to return


to skipped


anagrams.


More


highly-efficacious


subjects


indicated


this


willingness.


terms


of overall


performance,


highly-


efficacious


subjects


correctly


solved


more


hard


anagrams


esis


than


concerning


the less-efficacious


effects


subjects


of encountering


failure


hypoth-

earlier


on in


task


was


supported.


Also,


interaction


effects


efficacy


experimental


condition


were


found,


indicating


that less-effica


cious


subjects


were


not,


general,


more


affected


failure


than


high-efficacious


subjects


as hypothesized.


summary


, the


data


sugge


that,


terms


internal


consistency


test-retest


reliability,


MASE


reliable

validity


provide


instrument.

correlations


suDDOrt


The

with


... ....


convergent


other


construct


discriminant


self-report


validity


instruments,


e MASE.










self-efficacy,


task


difficulty,


persistence.


A number


of explanations


were


provided


post-hoc


these


results.


most


cogent


explanation


appears


to be


that


duration


time


quately


provided


measure


task


persistence.


was


A large


sufficient


proportion


to ade-


(71%)


subjects


were


forced


to finish


prematurely


minute


time


limit


This


suggests


that


40 minutes


may


have


been


sufficient


time


to discriminate


between


high- and


low-effica


students


terms


task


persistence.


However,


study


find


that


highly-efficacious


subjects


performed


better


hard


anagrams


than


efficacious


subjects,


exhibited


more


willingness


to return


to unfini


shed


(skipped


anagrams.


Overall,


results


study


provide


tentative


support


aspects


Bandura


(1977)


theory


of self-


efficacy.


Correlations


between


MASE


scores,


other


self-


report


indices,


teacher


ratings,


indicated


that


academically


highly-efficacious


students


are


more


willing


to accept


responsibility


their


academic


successes


than


failures.


Bandura


1977)


stated


that


self-efficacy


was


enhanced


as res


ulting


only

from


when

their


subjects


own


viewed


ability,


their


rather


successes


than


function


of contextual


factor s


such


as chance,


luck,


ease


task.


Following


this


reasoning,


was


also











MASE


also


correlated


positively


with


grade


point


average,


formal


post-secondary education plans,


plans


to enroll


a degree-granting


institution,


career


expectations


with


higher


career


socio-economic


possibilities


indices


(according


presented


to a ranking


subjects).


Finally,


with


verbal


intelligence


factor


removed,


MASE


correlated


at a low,


but


positive


level


with


grade


point


average


post


sec


ondary


school


plans.


These


res


ults


suggest


that


mastery


expectancy


variable


self


-efficacy


relevant


to prediction


of academic


success


planning.


res


ults


also


relate


work


McCl


elland


(1973),


Tait


(1973) ,


Taylor


(1964),


who


found


that


non-intellec-


tual


factors


play


a role


academic


vocational


success.


Since


academic


self-efficacy


an expectancy


variable,


important


to distingui


sh it


from


other


expectancy


variables.


Maier


Rotter


Seligman


s (1966)


(1976


locus


learned


of control


theory


helplessness


s theory


seen


somewhat


analogous


to Bandura


s (1977)


self-


efficacy


theory


Measure


Academic


Self-Efficacy.


Locus


measure


of control


outcome


learned


expectancies,


help


essness,


rather


than


however,


mastery


both


expec-


tancies


Rotter


theory


concerned


with


a person'


belief


that


outcomes


are


determined


either


one


s own


RC~~~~~~*iln flnq rnf0 ilq g ll


(inf n-rna


nr hv


tnr~pe


~ ~ f 1' on s


nrlfci~a


I











outcomes


are


internally


controlled,


may


feel


a sense


efficacy,


they


believe


that


they


lack


skills


required


to perform


activity.


Maier


Seligman


s (1976)


theory


learned


helplessness,


Hopelessness


Scale


(Beck


et al.,


1974)


resemble


concept


self-


efficacy,


are


differ-


from


self-efficacy


MASE.


Learned


helplessness


also


an outcome


expectancy


variable.


It states


that


when


people


cannot


what


happens


them,


they


become


apathetic


or "learned


helpless"


stop


trying


cope


(Lalonde,


1979)


People


may


give


trying


because


they


lack


a sense


efficacy


in achieving


required


performances,


or they


may


give


trying


because


they


expect


their


efforts


to produce


no results


in an unresponsive


or punishing


environment.


To change


efficacy-based


futility


requires


development


a sense


personal


effectiveness.


change


outcome-based


futility


requires


changing


social


environment


competencies


that


they


people


already


are


rewarded


possess.


Bandura


using


suggests


that


considering


both


effi


cacy


outcome


expec


tancies


would


produce


best


prediction


of behavior


(Lalonde,


1979).


concept


achievement


motivation


developed


I1QCQII


- 1 -


M..' al 1 n4l-fL


*nhor


M~~lhll


~1


--A-


n3n


c-irn











issue


concern


over


competition


with


a standard


excellence


, and


concern


with


success


or failure


in an


academic


task.


A subject


may


describe


either


positive


negative


feelings


towar d


achievement


achievement


motive

Thus,


index,


high


and


need


still


score


achievement


high


ers


need

any


achievement.


concerns


regarding


achievement,


including


fear


failure,


or disappointment


at lack


success.


absence


achievement


concerns


indicates


achievement


motivation.


conceptualization


academic


self-efficacy,


high

tion


self-efficacy


toward


achievement


seen


terms


situations.


a positive


That


orienta-


person


a mastery


expectancy


in achievement


situations.


Low


self-efficacy


is concerned


with


achievement


failure,


that


person


perceives


a lack


personal


coping


skills


achievement


situation.


Concern


over


competitive


situations


and


academic


failure


would,


therefore,


fall


into


domain


self-efficacy.


Those


who


are


indifferent


to achievement


situations


would


fall


into


category


self-e


fficacy.


These


two


concepts


achievement


motivation,


academic


self-efficacy,


thus,


appear


to differ


theoretical


level.


They


also differ


methodologically,


that


achievement


motivation


-a~~~~ ~ ~ ~ -k A ntr.n r lE S l *f


V's1;


a nrniective


AFln


1 m ~ ~P TV


Y-LIU~J


kr











themes.


This


a radi


call


different


methodology


from


that


proposed


Lalonde


(1979


assessing


academic


self-


efficacy.


research


strongly


and


suggested


treatment


that


literature


cognitive


test


concomitants


anxiety


test


anxiety


are


important


in predicting


performance


in evaluative


situations.


Although


emotional


arousal


factors


appear


to be related


to self-report


test


anxiety,


these


factors


apparently


are


less


predictive


performance


differences


high


test-anxious


individuals.


At this


point,


cognitive


factors


which


underlie


test


anxiety


have


only


been


generally


identified.


In order


more


adequately


understand


cognitive


concomitants


test


anxiety,


necessary


to identify


specific


cognitive


variables


which


mediate


observed


performance


differences.


Bandura


proposed


a cognitive-mediating


variable,


self-efficacy,


which


appears


to predict


extent


to which


individuals


approach


or avoid


task


situations.


This


variable


also


have


relevance


understanding


cognitive


factors


underlying


test


anxiety.


Lalonde


s study


(1979


begun


to delineate


role


academic


self-efficacy


as a relevant


variable


prediction


academic


success


career


planning.











their


academic


successes,


have


career


expectations


future


plans


associated


with


greater


socio-economic


academic


achievement.


In her


study,


Lalonde


also


found


that


test


anxiety,


as measured


Test


Anxiety


Scale,


correlated


correlation


these


study


indicate'


variables,


extends


.36 with

ed some


but


work


academic s

overlapping


identity.


Lalonde,


elf-efficacy.


variance


This


examining


This


across


sertation


rela-


tionship


test


anxi


academic


self-efficacy


predictors


task


performance,


under


various


conditions.


It is


cognitive


proposed


that


concomitant


self-efficacy


test


anxiety.


is a relevant


appears


reflect


some


core


problems


test-anxious


individual


academically-evaluative


situation.


Wine


(1971)


others


(Sarason,


1975;


Meichenbaum,


1972)


have


hypothesized,


test-anxious


person


demonstrates


a performance


decrement


evaluative


situation,


because


he fails


to focus


relevant


parts


task.


Instead


he ruminates


on his


perceived


incompetence


test


situation.


Since


concept


of academic


self-efficacy


deals


with


a person'


belief


that


he has


does


not


have)


coping


skills


to master


academic


situations,


appears


that


academic


self-efficacy


is certainly


related,


nnrin* anr+-n 1


1 atra7 -


Ir trr -A-v


ac nen -f


test


anxiety


m


L-


j Ft ^-











may,


thus,


provide


some


further


information


about


modus


operandi


test-anxious


individual.


This


role


dissertation


of academic


study


self-e


an attempt


fficacy


to delineate


relates


multidimensional


construct


test


anxiety.


study


determines,


via


a multivariate


predictor


model,


extent


to which


a measure


of academic


self-efficacy,


oper-


ationalized


academic


anxiety.


as a performance


situations,


Furthermore,


mastery


related


it determines


expectancy

a measure


whether


variable


test


these


variables


jointly


singly


can


predict


task


performance


under


evaluative,


reassuring,


neutral


task


conditions.


This


study


thus,


an attempt


to establish


whether


self-


efficacy


extent


is a cognitive


to which


concomitant


there


test


non-overlapping


anxiety,


variance


across


these


variables,


in predi


cting


to cognitive


performance.


It also


determines


extent


to which


both


variables


predict


performance


under


evaluatively-stressful


conditions,


the extent to


which


this


prediction


generalizes


to non-stressful


question


conditions.


That


of whether these variables


addr


are


esses


generalized


predictors


of academic


performance,


or only


relevant


under


evaluatively-


stressful


nmto n


academic


task


40c4-


conditions.


rain nil C ~ flV1 04-1 ciiria -


tohi rh


1,r I


~nui PSt~


rPtrl nlle










related


to performance


differences


in high


test-anxious


subje


cts.


Many


studies


manipulated


instructional


conditions,


while


holding


task


demands


constant


(Sarason,


1972).


These


studies


have


generally


involved


comparisons


"ego- involving"


instructions--


these


instructions


inform


subject


that


task


performance


level


capabilities


reflective


minimal


task


intellectual


instructions--instructions


which


specifically


directions


address


solving


cognitive


an anagrams


task


task);


reassur-


ance


instructions--instructions


which


reassure


subj


ect


that


task


performance


will


be used


to evaluate


him.


Frequently,


under


this


last


condition,


subject


told


that


performance


anonymous,


only


important


that


examiner


can


examine


average


group


performance


data


(Wine,


1971;


Sarason,


1958;


Sarason


Gan


zer,


1962,


1963;


Russell


Sarason,


1965


Ganzer,


1968).


Highly


t-anxious


subjects


tend


to perform


more


poorly


following


highly


evaluative


"ego- involving"


instru


actions,


than


following


reassuring


"anonymous"


instructions.


reverse


tends


to be


true


test-anxious


subjects.


Following


highly


evaluative


instructions,


test-anxious


subjects


perform


better


than


high


test-anxious


subjects.


However,


following


reassuring


"anonymous"


instructions,


high










or minimal


task


instruction


condition,


effects


test


anxiety


task


performance,


are


wiped


out.


High


test-anxious


intermediate


subjects


between


their


perform


at equivalent


performances


levels,

highly-


evaluative


and


reassuring


conditions.


Thus,


these


particular


studies


indicate


that


tes t


anxi


appears


to be a relevant


variable


cognitive


task


performance,


task instructions


create


an evaluative


situation,


opposite


evaluative


situation


, that


, a reassuring


situation,


where


individual


performances


not


count.


However,


task


instructions


are


neutral,


minus


evaluative


component,


high


test-anxious


subjects


perform


at equivalent


levels


on a cognitive


task.


There


is only


one


study


whi


ch has


examined


self-efficacy


relation


task


conditions.


This


behavioral


validation


study,


from


Lalonde


(1979)


dissertation.


This


study


manipulate


task


instructions


manner


above.


described


In other


test


words,


anxi


subjects


studies


were


described


presented


with


instructions


varying


amount


of evaluative


emphasis o

to develop


n performance.


a failure


experience


study

e, by


did,

prese


however,

nting su


attempt


objects


with

were


anagrams

presented


varying


to subjects


degrees


in various


difficultyy

orders:


Anagrams

hard.











failure


than less-efficacious


students.


Presumably,


condition


with


hard


anagrams


presented


first


would


produce

early f


most


failure.


difficult


results


task


situation


indicated


providing


that,


terms


persistence,


the less-efficacious


students


were


more


affected


failure


experimental


experiences.


conditions,


However,


highly-effi


over


cacious


all

students


solved


more


hard


anagrams


Thus,


creation


a "forced"


failure


experience


differentially


affect


low- and


high-efficacy


students.


In conclu-


sion,


this


study


suggests


that


task


difficulty


might


differ-


entially


affect


performance


high


self-


efficacious


produced


students,


pres


but


entation


a forced


of hard


failure


anagrams


experience


early


task


will


differentially


affect


performance


groups


of subjects.


strength


results


this


study


are


somewhat


diminished


fact


that


easy


and


medium


anagram


sets


were


extremely


easy


both


groups,


thus


reducing


discriminative


power


anagram


variable.


terms


proportions


anagrams


solved,


100%


of all


of the


easy


medium


anagrams


were


solved


both


groups.


Unfortunately,


this


study


is not


comparable


~n~r O+~ a4-ia 1CC t~1iini-im~ni nii REc 1-n dr


1 nclf-reii+-i nnc\

~Cii~;ae


tirh; ~h


f~ct


fcre~


Invi8 C~f











Lalonde


study


regarding


differential


performance


highly-efficacious


and less-efficacious


students


under


varying


evaluative


conditions.


data


suggest,


however,


that


there


be overall


cognitive


performance


differences


between


low- and


high-efficacy


students,


but


specific


failure


experiences


will


differentially


performance


groups.


Hypothe


ses


regarding


differential


performance


vs.


high-e


fficacious


students


can


be generated


from


Bandura


expected


that


theory.


there


From


would


Bandura


an over


s theory, it i

11 performance


difference


on an academically-related


cognitive


task,


however,


manipulation


evaluative


vs.


non-evaluative


instructions


would


produce


differential


performance


results,


such


that


effects


efficacy


variable


would


only


be significant


under


one


or the


other


condition.


There


are


several


aspects


of Bandura


theory


which


would


suggest


that


an experimental


manipulation


would


produce


to Bandura


differential


s theory,


results.


expectations


First


all,


personal


effi


according


cacy


are


derived


from


four


sources


information:


performance


accomplishments,


vicarious


experience,


verbal


persuasion,


biological


data


(Bandura,


1977)


more


dependable


F~xa~r~ n t ~S


Snnlrne


effi cacv


infnrmai nn -


ar rea tP r


1 _I 1r


I 1 T-


I J 1











could


considered


persuasion.


to fall


subject


into

led,


the

via


category


suggestion,


verbal

to believe


that


problem-solving


task,


example


anagrams


task


use


d in the


test


anxiety


studies,


a measure


intelligence


or ability.


Thus,


subject


presumably


becomes


persuasion


ego-involved.


According


a relatively


poor


to Bandura


source


(1977) ,


providing


verbal


change


perceived


self


-efficacy.


Efficacy


expectations


induced


verbal


persuasion


are


likely


to be weaker


than


those


arising


from


one'


own


accompli


shments.


This


fact


that


verbal


persuasions


do not


provide


an "authentic


experiential


base"


beliefs


(Bandura,


1977,


198).


If subjects


already


have


a long


history


of failure


cope


with


academically-evaluative


situations,


expectations


induced


suggestion


are


likely


to dramatically


alter


performance


provided


in a positive


individual


or negative


a long


direction.


history


Thus,


failure


success


at academically-related


imposition


a bri


laboratory


manipulation


unlikely


to reduce


greatly


increase


relationship


self-efficacy


performance.


task


conditions


suggest


an academically-


difficult


task,


regardless


of whether


instructions


are


neutral,


reassuring,


or evaluative,


perceived


self


compe-


4-on~~~~~~rn r; a -~m 3-A r n rFrnmnr


rj; 11


rnl~Carl


Ca


Canna











total


pattern


strongly


experiences

efficacious


in which the

person, after


failures

efficacy


occur.

expec-


stations


are


developed


through


repeated


success,


negative


reduced


impact


(Bandura,


instructions


are


of occasional


1977).


unlikely


failures


same


to induce


likely


token,


feelings


to be


,"te su n


coping


mastery


in a person


with


poor


self-efficacy


long


standing,


academic


situations.


Thus,


variable


of academic


self-efficacy,


task


instructions


are


likely


to signifi-


cantly


reduce


increase


relationship


between


self-


efficacy


cognitive


task


performance.


In the


present


study,


test


anxiety


self-efficacy


are


both


used


to predict


performance


under


evaluative,


neutral,


reassuring


task


ins


tructions.


According


research


findings


discussed


above


in regard


test


anxi


theoretical


issues


presented


terms


self-efficacy,


test


anxiety


will


interact


with


instruc-


tional


condition


when


predicting


performance


on a


cognitive


interact


task.


with


is expec


instructional


that


self-efficacy


condition


when


will


predicting


performance


on a cognitive


task


It is expected


that


test


anxiety


will


account


a significant


amount


variance


in performance


under


evaluative


reassuring


* Sr~Lwrr~


3 a ~; '


^ ^


snnhtrnC


V r-


rrrr~


nn


I-











on a cognitive


task


regardless


task


instructions


presented


to the


subj


ect.


If this


differential


prediction


test


anxiety


academic


self


-efficacy


supported,


this


suggests


that


these


are


different,


but


overlapping


dimensions,


performance,


which


beyond


facilitate


knowledge


prediction


of ability


cognitive


level.


cognitive


task


which


was


presented


subjects


was


a series


anagrams.


This


laboratory


analog


task


was


chosen


because


been


shown


to be


a sensitive


measure


of cognitive


task


performance


differences


related


test


anxiety


under


evaluative


reassuring


instructional


conditions


(Sarason


Stoops,


1978;


Sarason,


1972;


Russell


Sarason,


1965).


In addition,


final


grades


each


subjects


their


General


Psychology


class


were


obtained,


see


whether


test


anxiety


self


-efficacy


would


also


predict


to a cognitive


performance


measure


which


meaning


side


of the


laboratory


chology


measure


was


because


setting.


chosen


was


final


external


one


grade


cognitive


performance


General


performance


measures


which


subjects


would


have


in common.


Even


though


subjects


were


same


sections,


there


was


some


similarity


type


of materials


level


of difficulty


nrss~nt sSI


.1n 1


ii i rr


p~fe


erih ~


nth rs r


rrr~~aa


~nltr E~


|











final


grade


General


Psychology


was


readily


available


use


investigator.


hypotheses


proposed


this


study,


given


existing


literature


theory,


were


following:


Academic


self-efficacy


would


be moderately


negatively


correlated


with


test


anxiety.


Test

under


anxiety


would


evaluative


predict


anagram


reassuring


tas


performance

k instructions,


but


not


as eff


actively


under


neutral


instructions,


in accordance


Self-efficacy


under


with


would


evaluative,


test


predict


neutral,


anxiety


anagram


literature.


performance


reassuring


task


instructions,


in accordance


with


Bandura'


theory


Cognitive


ability


level,


as means


ured


an aptitude


test,


would


account


a significant


amount


variance


predicting


to cognitive


task


performance,


provided


there


were


suffi


cient


variation


in aptitude


scores.


range


aptitude


scores


was


reduced,


then


predictive


power


cognitive


ability


would


be reduced,


well.


Test


anxiety,


self-efficacy,


aptitud e


scores,


would


Psychology,


be predictive


as measured


of performance


subject


General


s final


















CHAPTER


TWO


METHOD


Subjects


subjects


were


undergraduate


students


enrolled


introductory


psychology


course


sec


tions


(PSY


2013)


University


Florida


during


Fall


Semester,


1981.


They


were


required


to participate


in experimental


studies


in order


to complete


their


course


requirement.


Procedure


subjects


signed


experimental


sessions,


which


were


held


one


week


apart.


They


were


required


complete


both


sessions


in order


receive


credit


participation


study.


Those


subjects


who


completed


both


sessions


were


included


study


(n=234 .


inves


tigator


was


assisted


data


collection


sessions


four


undergraduate


psyc


hology


majors


, who


administered


niipqth-i nnnsi r


iir1'1 nir


Ifi ,rf*


Efla f


i nn s


cnnndnntf~i


!" | rj


T~T


r Kr| ,











During


initial


testing


session,


subjects


received


Test Anxiety


Scale


Measure


Academic


Self


-Efficacy


The


Test


Anxi


Scal


, developed


Sarason


(1972) ,


iSa


item


scale


frequently


used


assess


test


anxiety.


true/fal


test,


yields


one


total


score


Test


-retest


reliabiliti


over


.80 have


been


reported


interval


several


weeks


Sarason, 1978)


on a sample


female


undergraduates.


Wagaman,


Cormi


Cormier


(1975)


reported


test


-retest


reliability


coefficient


This


test


been


utilized


identify


high


and


test-anxious


individual


most


research


conducted


Sarason


colleagues


(Sarason,


1968,


1972


, 1973a,


1975;


Sarason,


Pederson,


Nyman,


1968;


Sarason


Stoops


, 1978).


It has


been


used


many


other


rese


arch


studies


area


test


anxie


(Wine,


1971;


Spielberger,


Anton,


Bedell,


1976;


Wine,


1980)


subj


ects


were


also


admini


stereo


Measure


Academi


Self


-Efficacy


(MASE)


developed


Lalonde


(1979)


a 64


item


test


designed


assess


self-efficacy


in academic


situations


The


Hoyt


estimate


internal


consistency


reliability


test


was


found


to be


.95,


while


test-retest


reliability,


over


a one


month


period,


t~~~~ina~Er -IrC ri-I U1n*gnii n


a CacnTnl


MBS'I;1


wLI


I rilu


ill nn-l


*I no


w r


Sr i r











response


were


controlled


having


equal


numbers


negatively


infrequency


positively


items


worded


scale.


items,


Social


inclusion


desirability


a response


was


controlled


employing


Jackson


s (1966)


stringent


correlated


social


item


more


selection


highly


desirability


procedure.


with


scale


Only


efficacy


were


included


those


scale


in the


items


than


final


which


with


scale.


MASE


contains


a Likert-type


format.


subjects


were


required


assess


their


response


to each


item


on a


five


point


scale,


ranging


from


"never


or rarely


true"


to "almost


always


or always


true.


This


test


took


approxi-


mately


10-15


minutes


to complete,


entire


initial


testing


session


took


approximately


30 minutes.


At the


time


initial


ting


session,


subjects


also


gave


their


written


permit


ssion


to release


inve


stigator,


their


verbal


quantitative Scholastic


Aptitude


Test


scores


and


final


grades


in PSY


2013


(Intro-


ductory


Psychology).


Scholastic


Aptitude


Test


(SAT)


scores


were


obtained


from


Regi


star


s office,


grades


were


obtained


from


individual


instructors.


end of


initial


ting


session,


subjects


were


told


to return


an additional


testing


sess


ion,


one


week


later.


S th 1tnr


4-At *4-* n I


F; n~t


cr ~m: ~ :


nCrrC: n*r


nnnA


Ilrnn


c


1











several


Sarason,


test


1973a,


anxiety

1973b;


studies

Russell,


(Sarason,


Sarason,


Stoops,


1965;


1978;


Crager,


1959).


Sargent


series


anagrams


cons


anagrams


which


are


divided


into


three


difficulty


levels


easy,


medium,


hard


Prior


beginning


actual


dissertation


study,


a pilot


study


been


conducted,


to select


anagrams which

challenge, yet


of performance.


would

not s


be difficult


o difficult


In order


enough


to present


as to provide


assess


level


no range


of difficulty


Sargent


(1940)


anagrams


this


college


population,


52 subjects


were


administered


of Sargent


s anagrams.


They


were


given


unlimited


time


to complete


anagrams;


however,


they


were


assesse


d for


amount


time


took


to complete


13 most


difficult


anagrams.


This


was


done


order


to provide


an estimate


an appropriate


time


limit


presentation


anagrams


eventually


selected.


percentage


of subj


ects


solving


each


anagram


correctly


was


computed.


was


found


that,


this


pilot


sample,


not

most


of Sargent


"most

most


difficult.


difficult"

difficult a


anagrams


.nagrams


were


this


sample


were


solved


to 69%


the sample.


Eight


- t -- -


II


Lt


L


-Y- 1.----


r urr\


*











completion


of Sargent'


s "hard"


anagrams.


was


found


that


they


took


a mean


time


43 minutes


to complete


these


anagrams.


After


this


initial


pilot


study


was


completed,


additional

31 subjects


pilot

were


study

given


was

all


conducted.


aspects


For

the


this


full


study,


ssertation


study,


including


both


stionnaire


sess


(first


sess


ion)


experimental


ses


sion


(second


sess


ion)


purpose


second


pilot


study


was


to essentially


experimental


material


instructions,


time


limits


to be


used


full


dissertation


study


actual


was


dissertation


following


study,


manner.


Prior


second


session


second


sess


ion,


subjects


were


randomly


ass


signed


one


three


experimental


conditions


neutral


, reassuring


, or evaluative


instructions.


When


they


attended


second


session


they


were


task


taken


under


to separate


one


rooms


three


given


instructional


anagrams


conditions.


instructions


were


administered


undergraduate


assis-


tants.


Condition


one,


"neutral"


instructional


condition,


involved


following


instructions


(adapted


from


Russell


Sarason,


1965).


These


instructions


were


read


aloud


- -- 1~ 2 -- -- -











This
probl
give
anagr
mixed
ingfu
from
words


part


em-s
you
am i
up,
i wo
each
. S


answers
column.
the back


have


olvin
to so


saw
and
rd.
anag
olve


the
g t
lyve
ord
you
Onl
ram
the


on the
If you
s of th


minutes


pr
n
e


experiment


ask.
are
with
r tas
y one
, and
anag
oblem
eeds
answe


to sol


The p
called
the 1
k is t
word
they
rams,
Sheet
cratch
r shee
ve the


involves


robl
Sana
ette
o ma
can
are
and
s in
pap


ts.


ems


gr
rs
ke
be
al
wr
t
er
Y


a
will


ams. An
all
a mean-
made
1 common
ite the
he answer
, use
ou will


anagram


task.


subjects


"evaluative"


condition


received


same


basic


instructions


given


subjects


"neutral"


condition,


plus


an added


evaluative


component:


This


problem
give yo
anagram
mixed u
ingful
from ea
words.
answers
column.
backs o
are a m
abstrac
above a
student
particu
difficu
You wil
task.


part


'u
L
PU
p
w
W
.C


experiment


solvin
to so
is a w
, and
ord.
h anag


Solve
on the
If yo
the a
asure
terms
erage
have
ar ana
ty, so
have


g tas
lve a
ord w
your
Only
ram,
the a
prob
u nee
answer
of th
. Hi
intel
been
grams
you
45 mi


k.
re c
ith
task
one
and
nagr
lem
d sc
she
e ab
ch s


The
all
the
is
wor
the
ams
she


ch
ch


ligenc
found
quick
should
nutes


involves


problem
ed anag
letter
to mak
d can b
y are a
, and w
ets in
ch pape
SThes


to
Sst
,nd
sol
and
yve
sol


ms


th
ude
mos
ve
wi
lit
ve


will


ams. An
all
a mean-
made
1 common
ite the
he answer
, use the
anagrams
ink in
nts of
t college
these
th little
tle trouble.
the anagram


subjects


"reassuring"


condition


were


also


given


same


basic


instructions


as subjects


"neutral"


r


I











This


part


problem-


experiment


solving


task.


involves


problems


will


give


you


anagram


to solve


a word


are


called


with


anagrams.
letters all


mixed


word.


your
Only


anagram,
Solve th


problem


task
one w


they


or


anagrams,


s to
d can


are


and


sheets


make


a meaningful


be made
common


write


answer


from


each


words.
answers
column.


you


need


answer


anagrams


scratch


sheets


before


paper,


Most


, or have


use
you


backs


have


worked


probably
n them.


seen


These


they
books
not f
some
this
will


anagrams


are


harder


work


than


or magazines.


min


sh all


happens
find th


45 minutes


same


most
Cons


them,


anagrams
, don't


very


worry


anagrams


to solve


way,
, have


equently,


you


however,
seen in
you may


may


find


difficult.


about


easy.


one


will


have


anagrams


anagrams


anagrams


task


success fully


was


complex


scored


ted.


total


As shown


number


above,


subjects


were


given


minutes


to solve


anagrams


task.


time


limit


was


selected


from


second


pilot


study


discussed


previously


that


study,


subjects


been


given


minutes


to solve


anagrams,


although


there


was


an acceptable


range


of performance,


some


anagrams


were


solved


only


of the


very


subj


subj


ects,


ects


"ELSAUX"


"E IVARR"


was


was


solved


solved


only


the subjects


In general,


these


thirteen


words


appeared


slightly


the original


more

pilot


difficult

group.


ese


Therefore


subjects


than


actual


U
a- a an .5.. .5.- a 4.- -- -I- 1~ -4- J-1- -- -


~,,!3,3


Lt


,L., 3..


L


.,,,~,


,,


A


,,,


L











Statistical


Anal


analy


ses


were


conducted


using


Statistical


Analysis


stem


SAS)


(Helwig,


Council,


1979)


packaged


programs


multiple


regress


analy


ses.


primary


programs


utilized


were


GLM


(general


linear


mode


Stepwise


(stepwise


regression).


Prior


regress


analy


ses


being


conducted,


was


necessary


to randomly


select


subj


ects


a main


sample


a validation


sample.


Out


subjects,


there


were


21 subjects


were


missing


data


independent


variables

or self-e


either


fficacy


aptitude

scores.


scores


Most


test


these


anxiety

subjects


scores,

were


missing


scores


various


reasons


(foreign


students,


junior


college


transfers,


etc.) .


Four


subj ects


were


missing


test


anxiety


or self efficacy


scores,


to lack


of comply


etion


stionnaires


These


21 subje


were


generally


excluded


from


regression


analy


ses


, although


their


scores


were


include


in means


analy


ses


Pearson


R correlations.


An additional


three


subj


ects


were


included


anagram


analy


ses


, but


exc


luded


grade


analy


ses


because


they


have


grades


2013


(Intro-


ductory


Psychology)


course


remaining


subjects


were


randomly


selected











randomization


was


accomplished


computer,


using


a random


seed


procedure


from


system


(Helwig


Council,


1979).


The initial


analy


ses


involved


testing


original


proposed


dependent


models


, one


variable,


anagram


one


performance


grade


performance


dependent


variable.


After


ese


models


were


tested,


stepwise


regre


ssion


was


conducted


to establish


best


models


both


grade


anagram.


These


analyses


included


variables


from


original


proposed


models


as well


one


additional


acting


with


because


international


self-efficacy.


was


hypothesized


variable,


This


that


test


variable


test


anxi


was


anxiety


inter-


included


might


interact


with


self-efficacy


predicting


performance


dependent


measures,


it might


enhance


model


beyond


the original


proposed


linear


variables.


Finally,


analyses


variance


were


conducted


establish


whether


this


study


would


replicate


findings


of previous


studies


which


employed


similar


manipulations


test


anxiety


instructions,


yet


utilized


an anal


variance


model


rather


than


a linear


regression


model


data


analyses.


















CHAPTER


THREE


RESULTS


Descriptive


statistics


were


calculated


continuous


variables


study


test


anxi


ety,


self-


efficacy,

quantitati


Scholastic A

ve combined,


Latitude


Test


verbal


(SAT)

alone,


verbal


quantitative


alone,


grade,


anagram


Table


In addition,


Pearson


r corre


lation


matri


ces


were


calculated


on all


quantitative


variables


, for


combined


main


validation


sample


n=213


(Table


As hypothesis


zed,


test


anxiety


correlated


moderately


negatively


with


self-efficacy,


.0001)


Initially,


multiple


regrets


sion


analy


ses


were


conducted


both


grade


anagram


dependent


variables.


initial


model


include


group


instructional


condition),


test a

tative


anxiety


anxiety,


SATQ)


, group


self-efficacy,


interactions

self-efficacy


verbal

between

, group


(SATV), S

group and

and SATV,


quanti-


test

and


group


SATQ


initial


analysis


, using


anagram


dependent


TPW Vh al-^^ a


#alllt;


r' n~: nnC n~


tt~r; ~klnE


Ck~C


I


tnn


r







7).












0 0 0 0 0 0
0 0 0 0 0 0
*I-4 p-I
0) a) 0 0 ci
U) U> ci 0 N r-4
H (N N N
.0

$4
'U
00 g g g
.44 0 0
C r4r4
di CIU 0 0 H 0
-H> ~-D en
ri (N (N
ci
Q4
U,
0
Fe
C '00
(U $4-H 0 (N H a) 0
w en ci (N
Ii rem S S S
C C-H N N (N N en
ci 0> C" a) a)
'0 4,0)
C
H U)
ci 0
C
iU H
C H N (N 0
5.4 Eu S S S S
o (1) tO 0 0 a) en -.
H N (N
LA U]
(5) 0)
o H
ri
44 0)
(I) V
--a (N (N '.0 '0 a)
44 Z en ci H H ci 44
ri ri ri -4
44 -1-i
U)

U
-ri U) -H
4-) > 4-i
0.1 -H U]
-H >1 4-' Eu
$4 >1 U Eu H
o 44 0 44 0
CO C) C) U v-I -H
C) H ri H Eu ~I.J C)
o .0 X '4-4 .0 C U)
Eu C '4-i S-i 'U
r1 U) II
$4 I > 01 $4 C)
44 4-' I I Fe H
























a)
Ca)
ore

Go"
SI.'
00"
A
cow
Ow
r-4-r-I
(U
.,.4rd
etc
C',
WV
4J*n
o#Q
0
002
Uc
4-JO)
a)
4C
*r144
a)
Ott,
'do
rUm
WV
010)
4-'
O
0
COH
CC'4
Oh


'-40
a)
WO)
$4-

'U
r-4U2
'-4
"cm
('I


4-1(/2
0)
Lw"
we'
CJ2H
"C


a)
X '0
CO
9 -rP
M -
0) CL
04.'


lr14











self-efficacy


.05)


SATV


.02)


, SATQ


.03) .


main


effect


group


instructional


condition


was


not


significant,


neither


were


interaction


ects


test


anxiety


group,


self


-efficacy


group,


SATV


group,


SATQ


group.


r2 for


this


model


was


.14.


same


process


ting


original


proposed


model


was


also carried


grade


variable.


model


tested


included


five


main


variables


group,


test


anxiety,


self-efficacy


, SATV,


SATQ


, as well


interaction


group


with


other


four


variables.


In this


initial


model,


none


variables


reached


significance


.05)


level


or better.


this


model


was


.13.


Validation


analy


ses


were


conducted


both


anagram


model


grade


model.


For


anagram


model,


observed


scores


validation


sample


(n=51)


were


correlated


with


their


expected


scores


Based


regression


equation


generated


from


main


sample.


Pearson


r correlation


was


.12,


which


was


signi


ficant


indicating


that


results


with


this


sample


of n=162


be stable


when


this


model


replicated


with


other


samples


of equivalent


size.


Validation


analyses


were


also


conducted


grade


wtn~alr. H aa r'y a A a a ra r rT -^ no tr 4A-.4 a Inn 1


mn rl a 1


aka


hkEd m~d~


~ ~mn 1 n


ah rd ~


rnr


tno











Pearson


r correlation


was


.34,


which


was


significant


level


of probability.


This


model


did


replicate


validation


sample.


Additional


analyses


were


also


conducted


with


reduced


models


both


grade


anagram


including


main


effects


only).


In both


anagram


grade


models,


main


effects


group


interaction


group


with


other


variables


were


eliminated


from


models.


tests


were


conducted,


comparing


grade


reduced


anagram,


models


results


full


were


models.


significant,


both


indi-


eating


that


presence


group


main


effects


group


interactions


not


significantly


enhance


models.


Hence,


they


were


removed


from


additional


analyses.


Since


values


both


anagram


model


grade


model


were


relatively


low,


and .13,


respectively,


appeared


that


initial


propo


sed variables


accounted


only


a relatively


small


proportion


variance


performance.


It appeared


likely


that


other


variables


might


account


Other


more


variables


variance


currently


provide


model


better


might


models.


correlate


more


highly


with


anagram


grade


performance.


appeared


appropriate,


after


testing


initial


models,


build


a better


model


through


stepwise


regression


n~~~~~~~~~~m tnC nnAa -a t ,-.atynar 4 r


ram--


n rh ~n~rr rh


ck n


hF


hf


I7CI











between


variables.


Since


text


anxiety


self-efficacy


were


hypothesized


to jointly


predict


cognitive


task


per-


formance,


was


additionally


hypoth


esized


that


interaction


test


anxiety


self-efficacy


might


predict


cognitive


performance


more


effectively


than


test


anxiety


self-efficacy


operating


within


model


linear


fashion.


In this


stepwise


proc


edure


main


effect


variables


were


included,


i.e. ,


test


anxiety,


self-efficacy,


SATV,


SATQ.


anxiety


In addition,


self-efficacy


interaction


TESTSE)


was


test


included


model,


as discussed


above.


In order


to estimate


"best"


model,


was


necessary


use


available


data.


Hence,


main


stepwise


this


sample


regression


case,


validation


was


subjects


run


were


sample


were


complete


included


merged,


sample.


stepwise


analysis


grade,


subjects


were


included


stepwis e


analysis


anagram.


A stepwise


regression


analyst


complete


sample


was


conducted


anagram


dependent


variable


(n=213).


stepwise


analysis


yielded


variables


as significant


predictors


anagram


task


performance.


These


variables


were


test anxiety,


interaction


test


anxiety


with












this


model


was


.04.


overall


F value


was


210)


=4.54,


.01.


stepwise


regression


analysis


grade


on a


sample


n=210


three


subj ects


main


sample


have


grade


data),


variables


were


significant


predictors


grade


performance.


These


variables


were


test


anxiety


SATQ.


model


was


=.07.


The overall


value


this


model


was


207)


.74,


.0006


Test


anxiety


was


a significant


predi


ctor


variable


.02,


SATQ


was


significant


Validation


analy


ses


were


conducted


stepwise


regression


model


developed


from


complete


sample


(one


grades,


one


anagram)


First,


model


anagram


were


grade


used


developed


a standard


multiple


full


stepwise


regression


regression,


performed


main


samples


Then


these


models


were


validated


smaller


validation


sample


samples


were


merged


final


regression


res


ults


are


reported.


anagram


model


(test


anxiety


TESTSE


predi


ctors


main


sample


was


validated


validation


sample


correlation


between


the expected


score


observed


score


was


r=.ll


, which


was


non-s


significant.


validation


sample


was


combined


with


main


sample,


--


.9 -n m .- -


~rr~lrn:a


t nla~


9


I


I


mt


-











same


analyses


described


above


anagram


were


also


done


with


grade


dependent


variable.


stepwise

grade pe


regression


rformance


found


was


test


that


anxiety


best

SATQ.


model

The


predicting

standard


regress


procedure


(GLM)


was


used


to analyse


this


model


main


sample.


This


model


was


then


validated


cross-validation


sample.


expected


value


grade


was


correlated


with


observed


value,


validation


sample,


Pearson


r was


found


to be


which


was


significant


.03.


Hence,


this


model


was


validated.


validation


sample


was


then


combined


with


main


sample,


and


a multiple


regression


analysis


was


formed


combined


sample


211) .


results


yielded


an r2


.07.


Test


anxi


contributed


significantly


to the


model,


with


a probability


.03.


SATQ


also


contributed


significantly


mod


with


a probability


.02.


In addition,


several


analy


ses


of variance


were


conducted


on smaller


to determine


whether


results


this


study


would


replicate


previous


test


anxiety


studies


which


employed


an analysis


variance


in statistical


analyses.


first


analysis


variance


examined


instruction


variable


, group,


test


anxiety,


in a 3


. S


* S


nfl ~ .4--- -S -d~ l -t I-* n4 AA ,-


- n ,


r=.30,


~~lt~: A I1L~


m: r3 rl 1


mn~Al


)d A II A n


* I


Illk











condition.


test


anxiety


classification,


high


test-anxious


subjects


were


drawn


from


upper


lower


score


distribution.


score


ranges


three


Test


Anxiety


Scale


groups


were


high


test


anxiety,


anxiety,


21-37,


0-10.


middle


These


test


cutoff


anxiety,


scores


11-20,


corresponded


test


almost


identically


those


used


Sarason


(1972).


In this


analysis


variance


with


group


test


anxiety


independent


variables ,


anagram


performance


dependent


variable,


overall


F value


was


significant,


there


were


no significant


main


effects


or interactions,


although


group


main


effect


approached


significance


.07).


same


analysis


variance


was


also


conducted


grade


dependent


variable.


In this


analysis


overall


value


was


significant,


218)


.03)


There


was


a signifi-


cant


main


effect


test


anxiety


.01).


Further


analyses


compare


were

e the


conducted with

individual tes


Duncan


anxiety


s Multiple


means.


Range

The r


Test


results


indicated


that


high test anxiety


group


received


signifi-


cantly


lower


grades


.05)


than


middle


test


anxiety


groups


, but


these


groups


significantly


differ


from


each


other.


nfl~~h AA~ n1a rii tmA r


r~nneiiint~rd


~n~lt~~; d


W;IC


tl~Y1


n











anxiety


groups


with


same


cutoff


scores


as above.


self-efficacy


measure,


subjects


were


classified


into


high


and


self-efficacy


groups,


were


upper


and


lower


score


stribution.


remaining


subjects


were


classified


into


middle


self-efficacy


groups.


score


ranges


three


self-efficacy


groups


were


follows


high


self-efficacy,


or greater,


middle


self-efficacy


, 222-261,


self-efficacy,


lower.

with a


Again,


inagram


as in

the


previous


dependent


analysis


variable,


variance


overall


value


was


significant,


and


there


were


no significant


main


effects


interactions.


When


grade


was


used


dependent


variable,


overall


F value


was


significant;


however,


there


was


a significant


main


effect


tes t


anxiety


.01)


Duncan


s Multiple


Range


Test


indicated


again


that


lower


high


grades


test-anxious


than


middle


group


received


significantly


test-anxious


group


.05).


















CHAPTER


FOUR


DISCUSSION


Test


anxiety


self-efficacy


correlated


significantly


a moderate


negative


relationship.


correlation


was


very


similar


correlation


coefficient


found


Lalonde


dissertation


study,


r=-


(Lalonde,


1979).


These


results


indicate


that


test


anxiety


self-efficacy


are


moderately


related,


identical


constructs.


In the


initial


model,


with


anagram


dependent


variable,


test


anxiety,


self-efficacy,


verbal,


quantitative,


However,


were


significant


instructional


condition


predictors


(evaluative,


of performance.


reassuring,


neutral)


was


neither


significant


as a main


effect,


nor


interaction


with


other


variables.


Therefore,


hypothesis


that


test


anxiety


would


interact


with


task


instructions,


was


supported


data.


was


predicted


that


self-efficacy


would


interact


with


instructions,


this


was


supported


these


results.


was


also


hypothesized


that


test


anxiety,


self-


efficacy,


verbal,


quantitative,


would


predict











Validation


anal


yses


conducted


on both


original


anagram


both.


grade


original


models


yielded


anagram


model,


different


based


res


on 162


ults


subjects,


cross-validate,


indicating


that


results


this


model


be stable


when


replicated


with


a comparably


sized


sample.


However,


if the


data


from


main


sample


are


merged


with


data


from


validation


sample,


resultant


regression


model,


based


on 213


subjects,


prove


to be stable.


Validation


analyses


conducted


original


grade


model


indicated


that


this


original


model


validate


smaller


validation


sample.


This


result


indicates


that


this


model


would


be stable


replicated


in another


sample.


main


results


study,


based


on the stepwise


regression


with


anagram


dependent


variable,


indicated


that


tes t


anxiety,


interaction


test


anxiety


with


self-efficacy,

performance.


were

Test


both si

anxiety,


.gnificant


a main


predictors

effect, r


anagram


elated


negatively


to performance.


stepwise


regression


analysis


grades,


test


anxiety


SATQ


were


found


to be significant


predictors


of grades


received


introductory


psychology


course.


fact


that


test


anxiety


interact


with


instructional


i a nntrarv


s.vtra lo


,


~n nrl ~ f i nn


1~1


I L











instructional


manipulation


indicated


inconclusive


results.


Some


studies


have


found


a main


effect


test


anxiety,


with


high


test-anxious


subjects


doing


more


poorly


than


test-anxious


subjects,


but


test


anxiety


instruction


interaction


Sarason, Mandler,


Craighill,


1952


Mandler


Sarason,


1965).


Sarason,

Although


Crager,


test


1959;


anxiety


Russell


instruction


interaction


has been


found


some


researchers,


an unequivocal


finding.


It could


be speculated


that


current


study


find


test


anxiety


instruction


interaction


because


method


data


analysis.


previous


studies


utilized


tical


an analysis


analysis,


of variance


classifying


ANOVA)


subjects


model


into


in statis-


high


test


anxiety


groups


median


split)


or three


test


anxiety


groups


low,


middle,


high).


current


study


treated


data


continuous


form


analyzed


data


within


multiple


regression


framework.


In order


answer


this


question,


the data


from


this


study


were


analyzed


ANOVA


model


to make


analyses


more


comparable


previous


studies.


analysis


of variance


test


anxiety


instruc-


tions,


with


anagram


dependent


variable


found


- 2 2 .C I -- -- ~ n -n


i-I, nr


nn a; nni P4 t-mn


mt, *L


TiT











type


of instruction,


as related


to anagrami


performance.


On the


other


hand,


there


was


a significant


relationship


between


test


anxiety


grade


performance,


with


high


t-anxious


subjects


doing


significantly


more


poorly


than


middle


or low


test-anxious


subjects


their


general


psychol-


course.


When


analyses


were


conducted


using


test


anxiety,


self-efficacy,


instructions


as independent


variables


ANOVA


neither


significant


model,


anagram


similar


analysis


overall,


results


nor


grade


were


grade


analysis


obtained.


analysis


found


that


Although


were


test


anxiety


was


grade


analyses


significantly


analy


that


ses


test


related


replicated


anxiety


to grade


findings


a negative


performance.


regression


relationship


school


performance.


These


analyses


also


confirmed


previous


test


anxiety


studies


which


have


found


that


test


anxiety


a negative


effect


on cognitive


task


performance


(Wine,


1971) .


With


anagram


dependent


variable,


however,


no significant


results


were


found.


This


study


replicate


those


previous


test


anxiety


studi


which


found


test


Sarason,


anxiety


1973b)


instruction


however, it


interaction


replicate


(Sarason,


those


1961;


studies


which


found


a ge


neral


negative


relationship


between


test


anxiety


performance


(Russell


Sarason,


1965;


Crager,


, nr n


a a- 4 -- n'-4 -..-.... -'*l~. f % U* ~ 'tn


CI 112*: CI FIL~


~IIILYI~L1


~~lyl


A


^ fa











interaction.


It also


appears


from


these


ANOVA


analyses,


that


test


grade


anxiety.


measure


This


more


sensitive


fortunately


effects


case,


because


anagram


grade


task


is a meaningful


ANOVA


a contrived


measure


with


analyses


laboratory


performance


anagram


measure,


school.


task


replicate


findings


from


multiple


regression


model,


that


test


anxiety


self-efficacy


interact


to predict


performance.


is possible


that


results


replicate


because


effect


small,


ANOVA


analysis


not


sens


itive


enough


to pick


this


relation-


ship.


Cohen


(1968)


proposed


that


linear


regression


analyses


increase


size


of statistical


effects


over


ANOVA


models


36%.


Although


there


are


major


limitations


interpre-


stations


to be made


from


regression


analyses


anagram


performance,


model,


given


some


amount


culative


variance


tentative


accounted


interpretations


can


be made


from


finding


an interaction


between


test


anxiety


self-efficacy


Although


only


tentative


finding,


this


interaction


is rather


interesting


worth


discussing.


interaction


between


test


anxiety


self-


efficacy


can


be analyze


ed by


breaking


into


subs


ec-


U -a -1 at. a aII e -- -A--a


__Y *


- cc ~ ~


n!


rrru


LL


L











effect


test-anxiety.


In order


interpret


these


effects,


one


takes


original


regression


equation


total


sample


(n=213


, setting


equation


zero,


one


can


compare


effect


a one


unit


change


test-anxiety


with


anxiety


no change


test-anxiety,


on performance.


Similarly,


assess


one


effect


can


test


compare


effect


a one


unit


change


self-efficacy


with


change


self-efficacy,


assess


effect


of self-


efficacy


on performance.


test


anxiety


effect


was


found,


as follows:


A
Ani


=7.6518-


.2375


TA)+


.0012


TA*SE)


.6518-


A A
Anl-An2=-


SE=.2375/


.2375(TA+1)+. 0012((TA+1)*(SE)


.2375+. 0012(


.0012


SE=197


(TA=


test-anxiety,


SE=self-efficacy


n=expected


anagram


score)


one


unit


change


test


anxiety


score


is expected


corres


pond


to a change


in anagram


score


.2375


.0012


(self-efficacy).


one


solves


equation


(setting


equation


to zero),


it is


found


that


at a self-e


cacy


score


test


anxiety


no eff


ect.


When


self-


efficacy


scores


exceed


, then


effect


test


anxiety











These


results


suggest


that


two


persons


with


high


self-efficacy,


one


relatively


higher


test


anxiety,


then


that


person


will


have


better


anagram


performance.


Conversely,


subjects


with


self-efficacy,


one


higher


tes t


anxiety


relative


other,


then


anagram


performance


person


is expect


magnitude


to be lower


expected


diff


than


other


erence


will


nec


essarily


depend


actual


placement


self-


efficacy


scores


within


distribution.


Test


anxiety


can,


therefore,


effect


have


either


depending


a relatively


level


positive


self-efficac


or negative

y. At higher


levels


self-efficacy,


test


anxiety


a positive


effect,


at lower


levels


of self-efficacy,


test


anxiety


negative


effect.


These


results


suggest


that


when


a person


high


self-efficacy,


i.e.,


a strong


beli


their


ability


persist


to master


academic


situations,


then


worry


they


experience


regarding


tests


other


evaluative


situations


facilitative


of performance.


other


hand,


a person


self-efficacy,


or believes


that


they


lack


coping


skills


to master


academic


situations,


then


their


worry


preoccupation


regarding


testing


situations


is likely


to be debilitating


to performance.


less


SC0 1 f-off A mr4 ii


- --- .1--


nar nn l ILa t *ri 0~J ~ ~


tacf


nLlrl!nne


~rlrr~~r~


I


I











depending


upon


person


s general


expectancies


regarding


their


ability


to successfully


cope


with


academic


tasks.


Although


general


trend


literature


been


to perceive


test


anxiety


as debilitative,


theorists


related


areas


have


suggested


that


"worry"


can


have


positive


aspects.


Janis


(1958)


has discussed


"work


worrying"


as a means


dealing


effectively


with


a challenging


reality


mobilizing


approaching


situation.


strategies


difficult


Worry


can


ahead


be utilized


time,


situation.


to help


Janis


as a means


solve


theorized


that


when


worry


involved


experienced


mental


cognitions


rehearsal


expectations,


adaptively,


which

which


person


produces

function


becomes


reality-based

as a source o


hope


reassurance.


course


, worry


can


also


become


debilitative.


Marmor


1958)


makes


a distinction


between


"realistic"


"neurotic"


worrying.


He describes


reali


stic


worry


defensive


function


purpose


which


to deal


either


with


to ward


painful


an anticipated


consequences


real


one


trauma


already


experienced.


" He


theor


izes


that


when


this


realistic


worry


successful,


mental


activity


involved


leads


to action


mastery


engage


of the


"neurotic


threat


situation.


worry,


these


However,


people


some


are


people


unable


achi iv7


" I nn Pr


m~sf PI*V II t











which


clinically
ation)--or


problem


given
place


identical
labeled


else


with


what


as obsessional


effort


to deal


an intellectual


up entirely, and
to the emotional


regress
level,


is usually


rumin-


with


level


takes


in which


anxiety,
feature.


that


bound
The


or unbound,


clinical


an anxiety


picture


state,


the
may


phobia,


dominant


then


or a


conversion


1958,


hysteria.


(Marmor,


in Janis,


. 377)


Although


the clinical


state


Marmor


is describing


might


more


extreme


than


that


test-anxious


college


student,


description


"realistic"


"neuroti


worry


appears


to be somewhat


analogous


way


test


anxiety


might


operate,


either


an adaptive


or maladaptive


fashion.


In 1960,


Alpert


Haber


conceived


idea


that


test


anxiety


could


have


both


debilitative


cilitative


components.


They


developed


a scale,


Achievement


Anxi


Test


(AAT) ,


which


attempted


measure


facilitative


debilitative


Facilitating


test


Anxi


anxiety.


test


Scale,


consisted


Debilitating


Anxiety


Scale.


Facilitating


Anxiety


Scale


correlated


positively


with


academic


rformance


grade


point


average


Debilitating


Anxiety


Scale


correlated


negative


with


academic


performance.


Using


multiple


corre


lations


pre-


di ct


performance,


Alpert


Haber


(1960)


found


that


using


both


the plus


minus


sca


together


enhanced


performance











investigator'


s a


prior


conceptualization


test


anxiety


as a linear


a negative


dimension,


fashion.


res


affecting


ults


performance


current


solely


study


suggest


that


test


anxiety


may


a more


complex


dimension,


interacting


with


another


cognitive-mediational


variable,


self


-efficacy,


to affect


performance


either


positive


negative


ways.


These


results


suggest


that


a highly


self-efficacious


person,


some


maladapti


cognitive

ve effects


processes


test


may


occur


anxiety,


which


counteract


channel


person


into


productive,


positive,


self-referential


thinking.


This


positive


thinking


may


then


serve


as a motivator


adaptive


performance.


In individuals


with


self-efficacy,


degree


test


anxiety


increases,


negative


effect


test


anxiety


on performance


becomes


greater.


Therefore,


combined


effects


self-efficacy


test


anxiety


on cognitive


processes


are


presumed


to be maladaptive


in nature,


leading


to decreased


performance.


Although


several


interpretations


have


been


generated


regarding the interaction


test


anxiety


self-efficacy,


must


be remembered


that


above


interpretations


relate


to only


international


effect.


also


necessary


to interpret


effect


of self-efficacy


on performance,


4- 'v.4a S fl j t~%t' n4- rt aFCar,4 4anv e4.~ n,, n-I


CaE: nn


:nCn


aFCnnC


ak; a


an nhtr ~C











change


self-efficacy,


on anagram


performance


(the


algebra


abbreviated


this


case


A
Anl=7.6518-

Anl-An2=0.0


.2375 (TA)+0.0012 (TA) (SE)


012(TA)


Self-Efficacy


effect=0


.0012


This


result


can


be interpreted


saying


that


there


is a


positive


effect


of self-efficacy


on anagram


performance,


this


effect


increases


with


increasing


test


anxiety.


Self-efficacy


a greater


effect


on performance


at higher


levels

of test


test


anxiety.


anxiety, an

For some


a lesser


reason,


effect


self


at lower


-efficacy


levels


appears


interact


more


strongly


with


test


anxiety


predicting


performance,


at high


at higher


levels


levels


test anxiety,


test


a person


anxiety.


is more


Perhaps,


self-


focused


concerned


with


well


they


are


going


perform.


Self-referential


thought


may,


therefore,


have


greater


bearing


on performance


these


individual


Consequently


, it


is more


cruc


account


their


level


of self


-effi


cacy.


At lower


levels


test


anxiety,


person


invested


in how


well


they


will


task,


so it does


t matter


what


they


believe


they


can


on academic


tasks


There


may


more


a ~~~~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ L A.-- ---a I1 .C Iba. 5-a Aa1.


of a discrepancy
a a3 ^


L1 1: Y


,,,,,, 1


,L,,.s











variables


may


more


significant


predictors


of performance,


case


test-anxious


persons.


Al though


these


results


are


quite


interesting,


they


must


remain


highly


speculative


small


amount


variance

residuals


accounted


earlier


* this

models


model.

had s


pattern


suggested


that


other


variables

results.


might


Perhaps


systemati


other


cally


cognitive


accounting


factors,


currently


this


Other


model,


are


variables,


important


as persi


predicting


stence


task


task,


performance.


or how


much


subject


cared


about


his performance,


or the


quantity


have


negative


self-


a stronger


statements


made


on performance


subjects


than


those


variables


systematically


measured


this


study.


Harris


Johnson


(1980)


Harris


(1980


research


cited


earlier


suggests


that


mental


imagery


an important


factor


effects


test


on performance.


anxiety-evoking


Harris


situations,


high


test-anxious


individuals


tend


to have


negative


imagery,


which


vivid


persistent


When


Harris


(1980)


trained


subjects


to vi


suali


images


of competence


based


previous


personal


success


experience


individuals


manif


ested


a reduction


test


anxie


improved


a a a


such


effect


(1980)


anxiety


found that under


,,











An additional


self-efficacy


relationship


limiting


predictor


these


factor


model


variables


test


finding


anagram


anxiety/


that


performance


replicate


grade


analy


ses.


final


regression


model


grade


yielded


test


anxiety


SATQ


significant

an addition


predictors.

1 dependent


rationale


variable


was


for

see


using

if t


grade


variables


selected


to predict


anagram


performance,


a laboratory


analog


task,


would


academic


generalize


setting.


self-efficacy


does


to a meaningful


From these

not predict


results,

grade p


measure


in an


appears


performance ,


that

although


test


anxiety


does.


This


suggests


that


particular


model


generated


generalize


to predict


this


anagram


external


performance


measure


does


entirely


achievement.


Test


anxiety


related


to performance


in both


circumstances


however,


tes t


anxiety


effect


different


under


each


circumstance.


relation


ship


between


test


anxiety


anagram


performance


nonlinear


(positive


under


some


circumstances


negative


under


others).


relationship


test


anxiety


to grade


performance


is negative.


more


test


anxiety


a person


, the


more


likely


it will


have


debilitating


measure


effect


on performance.


of performance,


test


Thus,


anxiety


with


appears


grade


to have


- .


,,


-










situation,


creating


"evaluative"


situational


cues,


elicits


those


effects


predicted


test


anxiety


under


"evaluative"


conditions.


In addition,


as might


expected,


aptitude


scores


appear


to have


a strong


effect


prediction


grades.


this


case


, SATQ


significantly


predicts


grades


General


Psychology.


It is unclear


self-efficacy


predictive


anagram


performance


predictive


grades.


Additional


studies


will


necessary


answer


this


question.


finding


that


"best"


predictor


model


anagram


performance


identical


"best"


model


grade


performance,


anagram


does


model.


totally


suggest


negate


that


results


either


found


model


limited


generalizability,


or that


external


measure,


itself,


have


been


adequate


to bring


potential


effects


test


anxiety


self-efficacy.


From


a methodo-


logical


pers


pective,


both


grade


measure


anagram


measure


some


limitations.


anagram


measure


only


used


anagrams


previous


studies


(the


typical


which


may


number


have


anagrams


limited


used


sensitivity


measure.


In addition,


grade


measure,


final


grade


General


Psychology


only


a potential


eight


data


points,


also


possibly


limiting


sensitivity


of this


measure.




Full Text

PAGE 1

7(67 $1;,(7< $1' $&$'(0,& 6(/)()),&$&< $6 35(',&7256 2) &2*1,7,9( 7$6. 3(5)250$1&( %< 587+ /$85,( /$; $ ',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

7KLV GLVVHUWDWLRQ GHGLFDWHG WR P\ SDUHQWV -XGLWK DQG 0HOYLQ /D[

PAGE 3

$&.12:/('*(0(176 ZRXOG OLNH WR WKDQN P\ GLVVHUWDWLRQ FRPPLWWHH FKDLUPDQ 'U 1DWKDQ : 3HUU\ -U IRU SURYLGLQJ VXSSRUW DQG JXLGDQFH WKLV \HDU DQG IRU JUDFLRXVO\ WDNLQJ RYHU DV FKDLUPDQ RQ VKRUW QRWLFH ZKHQ P\ IRUPHU FKDLUPDQ 'U :LOH\ & 5DVEXU\ REWDLQHG D QHZ SRVLWLRQ LQ 0LFKLJDQ DOVR JUHDWO\ DSSUHn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n IXO WR WKH IRXU XQGHUJUDGXDWH VWXGHQWV ZKR KHOSHG PH ZLWK GDWD FROOHFWLRQ DQG FRGLQJ ,W ZDV D PDPPRWK WDVN DQG 3DWW\ %HOO (ULQ %R\HU 9LYLDQ 0RQWHV DQG $GDP 5RVHQEHUJ ZHUH KDUG ZRUNLQJ DQG HQWKXVLDVWLF WKURXJKRXW WKH HQWLUH SURFHVV DOVR JUHDWO\ DSSUHFLDWH WKH WLPH DQG HQHUJ\ ‘ f f

PAGE 4

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

PAGE 5

7$%/( 2) &217(176 $&.12:/('*(0(176 $%675$&7 &+$37(5 21( ,1752'8&7,21 &RJQLWLYH DQG (PRWLRQDO &RQFRPLWDQWV RI 7HVW $Q[LHW\ 7UHDWPHQW 6WXGLHV 6HOIHIILFDF\ 7KHRU\ 7:2 0(7+2' 6XEMHFWV 3URFHGXUH 6WDWLVWLFDO $QDO\VLV 7+5(( 5(68/76 )285 ',6&866,21 $33(1',&(6 $ $1$*5$06 7$6. % ,17(5&255(/$7,21 0$75,; 21 7+( ,1'(3(1'(17 $1' '(3(1'(17 9$5,$%/(6 )25 7+( 0$,1 6$03/( Q f & ,17(5&255(/$7,21 0$75,; 21 7+( ,1'(3(1'(17 $1' '(3(1'(17 9$5,$%/(6 )25 727$/ 6$03/( 5()(5(1&(6 %,2*5$3+,&$/ 6.(7&+ Y

PAGE 6

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n WLRQV WKH YDULDEOHV RI DFDGHPLF VHOIHIILFDF\ DQG DSWLWXGH ZRXOG HQKDQFH SUHGLFWLRQ ,W ZDV WKHRUL]HG WKDW D SHUVRQnV DFDGHPLF VHOIHIILFDF\ RU WKHLU EHOLHI LQ WKHLU DELOLW\ WR FRSH ZLWK RU PDVWHU DFDGHPLF VLWXDWLRQV ZRXOG EH UHODWHG WR FRJQLWLYH DVSHFWV RI WHVW DQ[LHW\ QHJDWLYH VHOIVWDWHPHQWV QHJDWLYH LPDJHU\f $FDGHPLF VHOIHIILFDF\ ZDV WKHUHIRUH K\SRWKHVL]HG WR HQKDQFH SUHGLFWLRQ RI YL

PAGE 7

FRJQLWLYH WDVN SHUIRUPDQFH EH\RQG NQRZOHGJH RI D SHUVRQnV OHYHO RI WHVW DQ[LHW\ DORQH ,Q DGGLWLRQ LW ZDV K\SRWKHn VL]HG WKDW DSWLWXGH ZRXOG DFFRXQW IRU D VLJQLILFDQW DPRXQW RI YDULDQFH LQ FRJQLWLYH WDVN SHUIRUPDQFH ,W ZDV IRXQG WKDW WHVW DQ[LHW\ GLG SUHGLFW FRJQLWLYH WDVN SHUIRUPDQFH DV PHDVXUHG E\ DQ DQDJUDPV WDVN DQG ILQDO JUDGH LQ *HQHUDO 3V\FKRORJ\ +RZHYHU WHVW DQ[LHW\ SUHGLFWHG SHUIRUPDQFH XQGHU DOO WDVN LQVWUXFWLRQDO FRQGLWLRQV UDWKHU WKDQ GLIIHUHQWLDOO\ DV SUHGLFWHG $FDGHPLF VHOI HIILFDF\ SUHGLFWHG FRJQLWLYH WDVN SHUIRUPDQFH LQ LQWHUn DFWLRQ ZLWK WHVW DQ[LHW\ RQ WKH DQDJUDPV WDVN DORQH $SWLWXGH VFRUHV SUHGLFWHG JUDGH SHUIRUPDQFH EXW QRW DQDJUDP WDVN SHUIRUPDQFH 7KH UHVXOWV ZHUH FRQVLGHUHG DV VSHFXODWLYH EHFDXVH WKH SUHGLFWRU YDULDEOHV RQO\ DFFRXQWHG IRU D VPDOO SHUFHQWDJH RI WKH YDULDQFH LQ WDVN SHUIRUPDQFH 7KH ILQGLQJV ZHUH GLVFXVVHG LQ WHUPV RI PHWKRGRORJLFDO LPSURYHPHQWV LQ WKH DFDGHPLF VHOIHIILFDF\ PHDVXUH DQG LQ WHUPV RI RWKHU SRVVLEOH UHOHYDQW SUHGLFWRU YDULDEOHV ‘ f 9,

PAGE 8

&+$37(5 21( ,1752'8&7,21 ,W KDV EHHQ JHQHUDOO\ IRXQG WKDW WKHUH LV D FRUUHODn WLRQ EHWZHHQ DSWLWXGH PHDVXUHV HJ ,4 PHDVXUHV 6FKRODVWLF $SWLWXGH 7HVWf DQG DFDGHPLF SHUIRUPDQFH LQ HOHPHQWDU\ VHFRQGDU\ DQG FROOHJH SRSXODWLRQV 3ODQW t 5LFKDUGVRQ )LVKPDQ t 3DVDQHOOD /DYLQ &RQU\ t 3ODQW 'XGHN *ROGEHUJ /HVWHU t +DUULV &URQEDFK 0F&OHOODQG f 7KH FRUUHODWLRQV WHQG WR EH VRPHZKDW KLJKHU LQ WKH HOHPHQWDU\ DQG KLJK VFKRRO \HDUV WKDQ LQ FROOHJH RU ROGHU SRSXODWLRQV %URG\ t %URG\ f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n PHQWDO UHVHDUFK KDV GHPRQVWUDWHG WKDW KLJKO\ WHVWDQ[LRXV

PAGE 9

LQGLYLGXDOV SHUIRUP PRVW SRRUO\ RQ FRJQLWLYH WDVNV XQGHU FRQGLWLRQV RI KLJK HYDOXDWLYH VWUHVV DQG RSWLPDOO\ ZKHQ HYDOXDWLRQ LV PLQLPDO 7KH UHYHUVH WHQGV WR EH WUXH IRU ORZ WHVWDQ[LRXV LQGLYLGXDOV 0DQGOHU t 6DUDVRQ 6DUDVRQ 0DQGOHU t &UDLJKLOO *DQ]HU 3DXO t (ULNVHQ 5XVVHOO t 6DUDVRQ 6DUDVRQ f :KHUHDV WHVW DQ[LHW\ KDV EHHQ IRXQG WR EH UHODWHG WR D FRJQLWLYH SHUIRUPDQFH GHFUHPHQW WKH KLVWRULFDOO\ SRSXODU H[SODQDWLRQ IRU WKH FDXVH RI WKLV GHFUHPHQW LQ SHUIRUPDQFH KDV EHHQ DXWRQRPLF UHDFWLYLW\ RU ZKDW /LHEHUW DQG 0RUULV f KDYH WHUPHG HPRWLRQDOLW\ 5HFHQW UHVHDUFK VXJJHVWHG WKDW FRJQLWLYH FRPSRQHQWV RI WHVW DQ[LHW\ UDWKHU WKDQ DXWRQRPLF UHDFWLYLW\ PD\ KDYH JUHDWHU SUHGLFWLYH YDOLGLW\ LQ UHJDUG WR FRJQLWLYH SHUIRUPDQFH XQGHU VWUHVVIXO HYDOXDWLYH FLUFXPVWDQFHV 1HDOH t .DWDKQ 6DUDVRQ t 6WRRSV 'HIIHQEDFKHU 'XVHN 0HUJOHU t .HUPLV 'HQQH\ f 7KHVH FRJQLWLYH FRPSRQHQWV LQFOXGH Df ZRUU\ RYHU SHUIRUPDQFH DQG FRPSDULVRQ ZLWK KRZ ZHOO RWKHUV DUH GRLQJ Ef UXPLQDWLRQ RYHU DOWHUQDWLYHV Ff EHLQJ SUHRFFXSLHG ZLWK VXFK WKLQJV DV IHHOLQJV RI LQDGHTXDF\ DQWLFLSDWLRQ RI SXQLVKPHQW DQG ORVV RI VWDWXV DQG HVWHHP 0HLFKHQEDXP 0DQGOHU t :DWVRQ 0DUOHWW t :DWVRQ f

PAGE 10

:KLOH HDUOLHU GHILQLWLRQV RI WHVW DQ[LHW\ KDG HPSKDn VL]HG D FRPPRQ GHQRPLQDWRU RI DQ[LHW\ RU HPRWLRQDO RU DXWRQRPLF UHDFWLYLW\ :LQH f WKLV FRPPRQ GHQRPLQDWRU GRHV QRW DSSHDU WR FDSWXUH WKH PRVW RXWVWDQGLQJ GLIIHUHQFHV EHWZHHQ SHUVRQV ZKR VFRUH DW H[WUHPHV RQ WHVW DQ[LHW\ 7KHVH GLIIHUHQFHV DSSHDU WR OLH LQ WKH QDWXUH RI FRJQLWLYH VWUXFWXUHV :LQH f 7KHUH LV D FKDQJH LQ WKH WKUXVW RI FXUUHQW WHVW DQ[LHW\ UHVHDUFK WRZDUG FRJQLWLYH LQWHUSUHWDWLRQV RI WHVW DQ[LHW\ 7KLV GLVVHUWDn WLRQ VWXG\ H[DPLQHV WKH UHODWLRQVKLS RI WHVW DQ[LHW\ WR RWKHU FRJQLWLYH SURFHVV YDULDEOHV LQ DQ HIIRUW WR GHOLQHDWH PRUH VSHFLILFDOO\ WKH FRJQLWLYH SURFHVVHV XQGHUO\LQJ SHUIRUPDQFH GHFUHPHQW LQ WHVW DQ[LHW\ 7HVW DQ[LHW\ UHVHDUFK LQLWLDOO\ IRFXVHG 6DUDVRQ f RQ HPRWLRQDO UHDFWLYLW\ DV WKH PHGLDWLQJ YDULDEOH LQ WKH UHODWLRQVKLS RI WHVW DQ[LHW\ WR DFDGHPLF SHUIRUPDQFH 0RUH UHFHQW UHVHDUFK KDV VXJJHVWHG WKDW WHVW DQ[LHW\ KDV D VWURQJ FRJQLWLYH FRPSRQHQW DV ZHOO ,W KDV EHHQ SURSRVHG WKDW WKLV FRJQLWLYH FRPSRQHQW DFFRXQWV IRU PXFK RI WKH YDULDQFH LQ WDVN SHUIRUPDQFH :LQH f 7KLV OLWHUDWXUH UHYLHZ H[DPLQHV WKH FRQWULEXWLRQ RI FRJQLWLYH DQG HPRWLRQDO IDFWRUV LQ WHVW DQ[LHW\ DV ERWK UHODWH WR FRJQLWLYH WDVN SHUIRUPDQFH ([SHULPHQWDO VWXGLHV ZLOO EH GLVFXVVHG IURP VHYHUDO GLYHUVH DUHDV RI UHVHDUFK ZKLFK EHDU XSRQ WKLV ,Q DGGLWLRQ WKH WUHDWPHQW OLWHUDWXUH LQ WHVW DQ[LHW\ ZLOO EH H[DPLQHG 7KH WUHDWPHQW OLWHUDWXUH LV

PAGE 11

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f KDV GHYHORSHG D FRJQLWLYHDWWHQWLRQDO WKHRU\ RI WHVW RU HYDOXDWLRQDO DQ[LHW\ WKH GLUHFWLRQ RI DWWHQWLRQ K\SRWKHVLV 6KH KDV SURSRVHG WKDW WKH SHUIRUPDQFH GLIIHUHQFHV RI ORZ DQG KLJK WHVWDQ[LRXV LQGLYLGXDOV UHIOHFW EDVLF GLIIHUHQFHV LQ WKHLU WDVN DWWHQn WLRQ GXULQJ HYDOXDWLYH WDVNV 7KH KLJK WHVWDQ[LRXV LQGLYLGXDO LV K\SRWKHVL]HG WR GLYLGH KLV DWWHQWLRQ EHWZHHQ VHOISUHRFFXSLHG ZRUU\ DQG WDVN FXHV 7KH ORZ WHVWDQ[LRXV LQGLYLGXDO LV SUHVXPDEO\ IRFXVHG PRUH IXOO\ RQ WKH WDVNn UHOHYDQW FXHV 7KH KLJK WHVWDQ[LRXV SHUVRQ WKXV KDV D WHQGHQF\ WR GLUHFW KLV DWWHQWLRQ LQWHUQDOO\ UDWKHU WKDQ WR WKH WDVN +H IRFXVHV RQ VHOIHYDOXDWLYH DQG VHOIGHSUHFDWRU\

PAGE 12

WKLQNLQJ 7KHUH LV D VLJQLILFDQW DPRXQW RI HYLGHQFH IURP WKH H[SHULPHQWDO DQG WUHDWPHQW OLWHUDWXUH ZKLFK VXJJHVWV WKDW D FRJQLWLYHDWWHQWLRQDO DSSURDFK WR WHVW DQ[LHW\ LV LPSRUWDQW LQ XQGHUVWDQGLQJ WKH FRQVWUXFW SHUKDSV PRUH VR WKDQ D WUDGLWLRQDO DXWRQRPLF DURXVDO LQWHUSUHWDWLRQ RI WHVW DQ[LHW\ :LQH 6DUDVRQ 'HQQH\ f :RUU\(PRWLRQDOLW\ 'LVWLQFWLRQ $ FORVHO\ UHODWHG WKHRUHWLFDO GHYHORSPHQW WR :LQHnV f GLUHFWLRQ RI DWWHQWLRQ K\SRWKHVLV LV WKH :RUU\ (PRWLRQDOLW\ GLVWLQFWLRQ GHYHORSHG E\ /LHEHUW DQG 0RUULV f 7KHVH DXWKRUV KDYH DQDO\]HG WHVW DQ[LHW\ LQWR WKH WZR FRPSRQHQWV RI f :RUU\ ZKLFK LV WKH IRFXVLQJ RI DWWHQWLRQ RQ FRQFHUQV DERXW SHUIRUPDQFH FRQVHTXHQFHV QHJDWLYH VHOIHYDOXDWLRQ HYDOXDWLRQ RI RQHnV DELOLW\ UHODWLYH WR RWKHUV DQG f (PRWLRQDOLW\ ZKLFK UHIHUV WR WKH DIIHFWLYHSK\VLRORJLFDO H[SHULHQFH JHQHUDWHG IURP LQFUHDVHG DXWRQRPLF DURXVDO ,QWHUIHULQJ WHVW DQ[LHW\ WKHUHIRUH WDNHV WKH IRUP RI KHLJKWHQHG ZRUU\ DQGRU HPRWLRQDOLW\ 7ZR VHSDUDWH VHOIUHSRUW VFDOHV KDYH EHHQ GHYHORSHG WR PHDVXUH WKH :RUU\ DQG (PRWLRQDOLW\ FRPSRQHQWV RI WHVW DQ[LHW\ 2QH LV WKH WHQ LWHP :RUU\(PRWLRQDOLW\ ,QYHQWRU\ /LHEHUW t 0RUULV f 7KH RWKHU VFDOH LV WKH LWHP

PAGE 13

,QYHQWRU\ RI 7HVW $Q[LHW\ 2VWHUKRXVH f :RUU\ LWHPV LQFOXGH Df QRW IHHOLQJ FRQILGHQW DERXW SHUIRUPDQFH Ef ZRUU\LQJ D JUHDW GHDO Ff WKLQNLQJ KRZ PXFK EULJKWHU RWKHUV DUH Gf WKLQNLQJ DERXW WKH FRQVHTXHQFHV RI IDLOXUH Hf IHHOLQJ QRW DV SUHSDUHG DV SRVVLEOH (PRWLRQDOLW\ LWHPV LQFOXGH Df VR QHUYRXV FDQQRW UHPHPEHU IDFWV Ef KHDUW EHDWLQJ IDVWHU Ff XSVHW VWRPDFK Gf XQHDV\ XSVHW IHHOLQJ Hf IHHOLQJ DQ[LHW\ :RUU\ DQG (PRWLRQDOLW\ VFRUHV RQ WKH :RUU\(PRWLRQDOLW\ ,QYHQWRU\ KDYH EHHQ UHODWHG WR SHUIRUPDQFH H[SHFWDWLRQV UHJDUGLQJ H[DPLQDWLRQV DQG DFWXDO H[DPLQDWLRQ SHUIRUPDQFH 6HYHUDO VWXGLHV KDYH VKRZQ WKDW :RUU\ ZDV LQYHUVHO\ UHODWHG WR SHUIRUPDQFH H[SHFWDWLRQV RI FROOHJH DQG KLJK VFKRRO VWXGHQWV WDNLQJ FODVVURRP H[DPV 'RFWRU t $OWPDQ /LHEHUW t 0RUULV 0RUULV t /LHEHUW f 3HUIRUn PDQFH H[SHFWDQF\ ZDV PHDVXUHG E\ KDYLQJ VXEMHFWV DVVHVV KRZ ZHOO WKH\ H[SHFWHG WR SHUIRUP RQ D SDUWLFXODU WHVW :KHUHDV :RUU\ ZDV FRQVLVWHQWO\ UHODWHG WR SHUIRUPDQFH H[SHFWDWLRQV (PRWLRQDOLW\ ZDV LQYHUVHO\ UHODWHG WR SHUIRUPDQFH H[SHFWDWLRQV LQ VRPH VDPSOHV 'RFWRU t $OWPDQ 0RUULV t /LHEHUW 6SLHJOHU 0RUULV t /LHEHUW f EXW XQUHODWHG LQ RWKHUV /LHEHUW t 0RUULV 0RUULV t /LHEHUW 6SLHJOHU 0RUULV t /LHEHUW f ,Q DOO EXW RQH RI WKH VDPSOHV LQ WKH 0RUULV DQG /LHEHUW f VWXGLHV :RUU\ FRUUHODWHG PRUH QHJDWLYHO\ ZLWK SHUIRUPDQFH H[SHFWDWLRQV WKDQ (PRWLRQDOLW\

PAGE 14

$ VLPLODU SDWWHUQ H[LVWV IRU WKH UHODWLRQ RI :RUU\ DQG (PRWLRQDOLW\ WR DFWXDO SHUIRUPDQFH RQ H[DPV 0RUULV DQG /LHEHUW f UHSRUWHG WKDW LQ D FROOHJH VDPSOH :RUU\ EXW QRW (PRWLRQDOLW\ IRUPHG D VLJQLILFDQW LQYHUVH UHODWLRQn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f IRXQG UHVXOWV VLPLODU WR WKRVH RI WKH 0RUULV DQG /LHEHUW f FROOHJH VDPSOH :RUU\ ZDV FRQVLVWHQWO\ QHJDWLYHO\ UHODWHG WR H[DP SHUIRUPDQFH ZKLOH (PRWLRQDOLW\ UHODWHG RQ VRPH RFFDVLRQV DQG QRW RQ RWKHUV $JDLQ SDUWLDO FRUUHODWLRQV GHPRQVWUDWHG WKDW ZKHQ WKH HIIHFWV RI (PRWLRQDOLW\ ZHUH SDUWLDOOHG RXW :RUU\ FRQWLQXHG WR IRUP D VLJQLILFDQW QHJDWLYH FRUUHODWLRQ ZLWK SHUIRUPDQFH +RZHYHU ZKHQ :RUU\ ZDV SDUWLDOOHG RXW (PRWLRQDOLW\ ZDV

PAGE 15

QRW VLJQLILFDQWO\ FRUUHODWHG ZLWK SHUIRUPDQFH 7KLV VXJJHVWV WKDW ZKHQ WKH FRPPRQ YDULDQFH LV UHPRYHG :RUU\ EXW QRW (PRWLRQDOLW\ LV VLJQLILFDQWO\ DQG QHJDWLYHO\ FRUUHODWHG ZLWK SHUIRUPDQFH RQ FROOHJH H[DPV 'HIIHQEDFKHU f DOVR LQYHVWLJDWHG WKH UHODWLRQn VKLS RI :RUU\ DQG (PRWLRQDOLW\ WR SHUIRUPDQFH RQ WKH 0LOOHU $QDORJLHV 7HVW 0$7f D WHVW ZKLFK LV LPSRUWDQW EHFDXVH LW LV XVHG IRU JUDGXDWH DGPLVVLRQV EXW IRU ZKLFK OLWWOH SUHSDUDWLRQ LV SRVVLEOH LH WKHUH LV QR VSHFLILF NQRZOHGJH WKDW VRPH SHRSOH PD\ KDYH DFFHVV WR RYHU RWKHUV LQ UHJDUG WR SHUIRUPDQFH RQ WKH WDVNf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

PAGE 16

0RUULV DQG /LHEHUW f HPSOR\HG :RUU\ DQG (PRWLRQDOLW\ LQ D GHVLJQ ZKLFK PDQLSXODWHG OHYHO RI VWUHVV WKURXJK REYLRXV DQG QRQREYLRXV WLPLQJ DQG GLIILFXOW\ RI LWHPV IRU WKH WLPHG VXEWHVWV RI WKH :HFKVOHU $GXOW ,QWHOOLJHQFH 6FDOH :RUU\ OHYHO ZDV IRXQG WR LQWHUDFW ZLWK WLPLQJ DQG GLIILFXOW\ OHYHO 6XEMHFWV ZKR ZHUH KLJK RQ :RUU\ SHUIRUPHG EHWWHU LQ WKH XQWLPHG WKDQ WKH WLPHG FRQGLWLRQ 7KH UHYHUVH ZDV WUXH IRU VXEMHFWV ORZ RQ :RUU\ 7KH\ SHUIRUPHG EHWWHU RQ WKH WLPHG FRQGLWLRQ (PRWLRQDOLW\ OHYHO GLG QRW LQWHUDFW ZLWK GLIILFXOW\ RU WLPLQJ FRQGLWLRQV DQG ZDV XQUHODWHG WR SHUn IRUPDQFH LQ WKLV VWXG\ )LQDOO\ 0RUULV DQG 3HUH] f IRXQG WKDW :RUU\ EXW QRW (PRWLRQDOLW\ ZDV VLJQLILFDQWO\ QHJDWLYHO\ UHODWHG WR UHDGLQJ WHVW SHUIRUPDQFH ,QWHUHVWLQJO\ DOWKRXJK WKH FRUUHODWLRQ ZDV QRW VLJQLILFDQW (PRWLRQDOLW\ ZDV SRVLWLYHO\ FRUUHODWHG ZLWK SHUIRUPDQFH ,W DSSHDUV WKDW D IDLUO\ FRQVLVWHQW SDWWHUQ RI UHVXOWV HPHUJHV IURP WKH UHVHDUFK RI WKH /LHEHUW DQG 0RUULV JURXS :RUU\ DQG (PRWLRQDOLW\ DUH VLJQLILFDQWO\ DQG FRQVLVWHQWO\ FRUUHODWHG FRUUHODWLRQV UDQJH IURP WR f +RZHYHU LW LV WKH :RUU\ FRPSRQHQW ZKLFK FRQVLVWHQWO\ IRUPV D QHJDn WLYH UHODWLRQVKLS ZLWK SHUIRUPDQFH H[SHFWDWLRQV DQG DFWXDO WHVW SHUIRUPDQFH QRW (PRWLRQDOLW\ 7KH ILQGLQJV IRU (PRWLRQDOLW\ DUH PXFK PRUH LQFRQVLVWHQW DQG PL[HG (PRWLRQn DOLW\ ZDV XQUHODWHG UHODWHG RQO\ ZLWKLQ FHUWDLQ VWUDWD RI :RUU\ RU QHJDWLYHO\ UHODWHG WR SHUIRUPDQFH PHDVXUHV

PAGE 17

7KHUHIRUH IRU ERWK SHUIRUPDQFH PHDVXUHV DQG SHUIRUPDQFH H[SHFWDQFLHV (PRWLRQDOLW\ LV QRW DV FRQVLVWHQWO\ UHODWHG WR RXWFRPH PHDVXUHV DV :RUU\ $V 0RUULV DQG /LHEHUW f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f VWXGLHV UHIHUV WR VHOIUHSRUW RI DIIHFWLYH UHDFn WLYLW\ ,W LV LPSRUWDQW WR H[DPLQH WKH UHODWLRQVKLS EHWZHHQ VHOIUHSRUW RI HPRWLRQDOLW\ DQG DFWXDO SK\VLRORJLFDO LQGLFHV RI HPRWLRQDO UHDFWLYLW\ +ROUR\G :HVWEURRN :ROI DQG %DGKRUQ f DWWHPSWHG WR DVVHVV WKH UHODWLRQVKLS EHWZHHQ

PAGE 18

VHOIUHSRUW RI DQ[LRXV DURXVDO SK\VLRORJLFDO LQGLFHV RI HPRWLRQDO DURXVDO DQG SHUIRUPDQFH PHDVXUHV LQ KLJK DQG ORZ WHVWDQ[LRXV VXEMHFWV 7KH +ROUR\G HW DO f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f

PAGE 19

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n WLRQDO RU GULYH WKHRU\ RI 7ROPDQ f 7KH LQIOXHQFH RI PRWLYDWLRQ RQ FXH XWLOL]DWLRQ ZDV UHFRJQL]HG E\ 7ROPDQ f +H OLVWHG VWURQJ PRWLYDWLRQDO FRQGLWLRQV DV RQH RI VHYHUDO FDXVHV RI WKH IRUPDWLRQ RI QDUURZ FRJQLWLYH PDSV LQ DQLPDOV DQG KXPDQV %DUWOHWW f LQWURGXFHG WKH FRQFHSW RI UDQJH RI FXHV XWLOL]HG WR H[DPLQH WKH HIIHFWV RI PRWLYDWLRQ RQ WDVN SHUIRUPDQFH 7KH HDUO\ UHVHDUFK FXOPLQDWHG LQ DQ LQIOXHQWLDO SDSHU E\ (DVWHUEURRN f 7KH PDLQ SRLQW RI WKH (DVWHUEURRN SDSHU f ZDV WKDW ZKHQ GLUHFWLRQ RI EHKDYLRU LV FRQVWDQW LQFUHDVH LQ GULYH LQ WKLV FDVH DQ[LHW\ GULYHf LV DVVRFLDWHG ZLWK D UHGXFWLRQ LQ WKH UDQJH RI FXH XVH 2QH LPSOLFDWLRQ RI (DVWHUEURRNnV

PAGE 20

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n YDQW FXHV 7KDW LV WKH SHUIRUPDQFH RI ORZ WHVWDQ[LRXV VXEMHFWV UHIOHFWV D ZLGHU UDQJH RI WDVN FXH XWLOL]DWLRQ LQ DFFRUGDQFH ZLWK WKH (DVWHUEURRN K\SRWKHVLV :LQH f :HVW /HH DQG $QGHUVRQ f WHVWHG FKLOGUHQ ZKR KDG VFRUHG KLJK RU ORZ RQ WKH 7HVW $Q[LHW\ 6FDOH IRU &KLOGUHQ 6DUDVRQ 'DYLGVRQ /LJKWKDOO :DLWH t 5XHEXVK f ZLWK IRUW\ DULWKPHWLF VWRU\ SUREOHPV FRQVWUXFWHG LQ VXFK D ZD\ WKDW VRPH FRQWDLQHG DGGLWLRQDO LUUHOHYDQW LQIRUPDWLRQ $OWKRXJK WKH KLJKO\ WHVWDQ[LRXV VXEMHFWV VKRZHG D VOLJKWO\ SRRUHU OHYHO RI SHUIRUPDQFH RYHUDOO WKDQ WKH OHVV WHVWDQ[LRXV JURXS WKHLU SHUIRUPDQFH ZDV QRW KLQGHUHG E\ WKH DGGLWLRQ RI LUUHOHYDQW LQIRUPDWLRQ 7KH ORZ WHVWDQ[LRXV VXEMHFWV SHUIRUPHG PRUH HIILFLHQWO\ RQ WKH XQHPEHOOLVKHG LWHPV WKDQ RQ WKH RQHV FRQWDLQLQJ LUUHOHYDQW FXHV *HHQ f IRXQG IXUWKHU HYLGHQFH RI WHVW DQ[LHW\ DV D VRXUFH RI UHVWULFWHG FXH XWLOL]DWLRQ LQ DQ H[SHULPHQW LQYROYLQJ VHULDO OHDUQLQJ

PAGE 21

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f DQG *HHQ f SURSRVH WKDW WKH SULPDU\ WDVN RI WKH KLJK WHVWDQ[LRXV VXEMHFW LV QRW WKH DVVLJQHG WDVN :LQH GHVFULEHV WKH VLWXDWLRQ LQ WHUPV RI GLYHUVLRQ RI DWWHQWLRQ DZD\ IURP WKH WDVN DQG WRZDUG WKH VHOI &RQFHQWUDWLRQ RI DWWHQWLRQ LQZDUGO\ WRZDUG WKH VHOI GUDZV DWWHQWLRQ DZD\ IURP WDVNUHODWHG VWLPXOL +DPLOWRQ f KDV SURSRVHG DQ LQWHUHVWLQJ PRGHO RI DWWHQWLRQ UHODWHG WR :LQHnV K\SRWKHVLV :KHQHYHU D SHUVRQnV SULRULWLHV IURP WKH DYDLOn DEOH OLPLWHG VWRUHf JLYH SUHFHGHQFH WR WKH GHPDQGV RI HQGXULQJ GLVSRVLWLRQV OHVV VSDUH

PAGE 22

FDSDFLW\ LV OHIW RYHU IRU GHPDQGV RI WKH WDVN 7KH SHUVRQ PD\ UHVSRQG E\ SD\LQJ OHVV DWWHQWLRQ WR H[WHUQDO WDVNUHODWHG LQSXWV HLWKHU E\ QDUURZLQJ WKH DUHD IRFXVVHG IRU VHOHFWLYH DWWHQWLRQ RU E\ UHVWULFWLQJ DWWHQn WLRQ WR D VPDOO QXPEHU RI IRFL +DPLOWRQ SS f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f 7KHUH LV VRPH HYLGHQFH WR VXSSRUW WKLV QRWLRQ LQ H[SHULPHQWDO PRGHOLQJ UHVHDUFK 0RVW RI WKLV UHVHDUFK KDV EHHQ FRQGXFWHG E\ 6DUDVRQ DQG KLV FROOHDJXHV 6DUDVRQ D DQG 6DUDVRQ 3HGHUVRQ t 1\PDQ f 6DUDVRQ 3HGHUVRQ DQG 1\PDQ f UHDVRQHG WKDW

PAGE 23

WHVW DQ[LHW\ PD\ FRQWULEXWH WR LQFUHDVHG YLJLODQFH ZLWK UHJDUG WR SRVVLEO\ KHOSIXO FXHV LQ WKH HQYLURQPHQW EHFDXVH KLJK WHVWDQ[LRXV LQGLYLGXDOV DUH PRUH LQVHFXUH WKDQ RWKHU SHUVRQV WKH\ PD\ EH UHODWLYHO\ PRUH LQWHUHVWHG DQG DFWLYH LQ DWWHQGLQJ WR FXHV LQ VWUDQJH VLWXDWLRQV DQG LQ ERUURZLQJ EHKDYLRUV DQG DWWLWXGHV IURP PRGHOV 6DUDVRQ HW DO SS f ,Q DQ LQLWLDO VWXG\ 6DUDVRQ HW DO f VWXGLHG WKH HIIHFWV RI PRGHOLQJ DQG WHVW DQ[LHW\ RQ OHDUQLQJ SDLUHG DVVRFLDWHV 7KHUH ZDV D VLJQLILFDQW DQ[LHW\ E\ FRQGLWLRQV LQWHUDFWLRQ 7KH KLJK DQG PLGGOH WHVW DQ[LHW\ JURXSV GHPRQn VWUDWHG VXSHULRU SHUIRUPDQFH RQ WKH WDVN LQ WKH VLWXDWLRQV ZKHUH WKH VXEMHFWV FRXOG REVHUYH PRGHOV OHDUQLQJ WKH OLVWV 6XEVHTXHQW VWXGLHV H[SORUHG WKH HIIHFWV RI FKDUDFWHUn LVWLFV RI PRGHOVn SUREOHPVROYLQJ EHKDYLRUV DQG IHHGEDFN WR PRGHOV RQ WKH FRJQLWLYH SHUIRUPDQFH RI KLJK DQG ORZ WHVWDQ[LRXV VXEMHFWV 6DUDVRQ Df KDG VXEMHFWV VROYH GLIILFXOW DQDJUDPV XQGHU RQH RI VHYHUDO H[SHULPHQWDO PRGHOLQJ FRQGLWLRQV 2QH JURXS REVHUYHG D PRGHO VLOHQWO\ VROYH VDPSOH SUREOHPV $QRWKHU REVHUYHG D PRGHO GHVFULELQJ WKH PDQLSXODn WLRQ RI OHWWHUV ZKLOH VROYLQJ WKH DQDJUDPV $ WKLUG JURXS REVHUYHG D PRGHO FRPPHQW RQ SULQFLSOHV XQGHUO\LQJ WKH VXFFHVVn IXO VROXWLRQ RI DQDJUDPV ZKLOH ZRUNLQJ 7KHUH ZDV DOVR D QRPRGHO FRQWURO FRQGLWLRQ :KLOH RYHUDOO WKH ORZ WHVW DQ[LRXV VXEMHFWV SHUIRUPHG EHWWHU WKDQ WKH KLJK WHVWDQ[LRXV VXEMHFWV WKHUH ZDV DQ DQ[LHW\ E\ H[SHULPHQWDO FRQGLWLRQ LQWHUDFWLRQ 7KH KLJK WHVWDQ[LRXV VXEMHFWV ZHUH VXSHULRU LQ SHUIRUPDQFH WR WKH ORZ WHVWDQ[LRXV VXEMHFWV XQGHU WKH

PAGE 24

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n ZKDW PRUH GLIILFXOW WR LQWHUSUHW 6DUDVRQ f LQYHVWLJDWHG ZKHWKHU SHUVRQV LQ WHVW DQ[LHW\ ZRXOG UHVSRQG GLIIHUHQWLDOO\ WR D QHXWUDO PRGHO YV D PRGHO ZKR LV VHOIn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

PAGE 25

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f DQ[LRXV VXEMHFWVn SHUIRUPDQFH ZDV VXSHULRU WR WKDW RI D PLGGOH DQG ORZ WHVW DQ[LHW\ JURXS )RU WKH IDLOXUH FRQGLWLRQ IDLOHG E\ DQ DXWKRULW\ ILJXUHf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

PAGE 26

REVHUYDWLRQ RI IDLOXUH IHHGEDFN KDG D PXFK PRUH QHJDWLYH HIIHFW RQ WKLV JURXS WKDQ WKH REVHUYDWLRQ RI D VHOIn GHURJDWRU\ PRGHO 6DUDVRQ VXJJHVWV WKDW REVHUYLQJ VRPHRQH ZKR LV GLVVDWLVILHG ZLWK KLPVHOI VHOIGHURJDWRU\ FRQGLWLRQf PD\ QRW EH FRQYLQFLQJ WR DQ REVHUYHU VLQFH WKHUH PD\ EH QR REMHFWLYH EDVLV IRU WKH VHOIGHEDVLQJ UHDFWLRQ +RZHYHU WKH REVHUYDWLRQ RI D QHJDWLYH DXWKRULW\ ILJXUH LV PXFK PRUH FRQYLQFLQJ DQG FHUWDLQO\ ZRXOG DIIHFW WKH HYDOXDWLRQDO DQ[LHW\ RI WKH KLJK WHVWDQ[LRXV VXEMHFW $QRWKHU H[SODQDn WLRQ IRU WKH ODFN RI GHWULPHQWDO LPSDFW RI WKH VHOIGHURJDWRU\ PRGHO OLHV LQ WKH FRQVWUXFW RI WKH FRSLQJ PRGHO GHVFULEHG E\ 0HLFKHQEDXP f 7KH VHOIGHEDVLQJ PRGHO PD\ KDYH EHHQ VHHQ E\ WKH KLJK WHVWDQ[LRXV VXEMHFW DV VRPHRQH ZKR H[SHUn LHQFHV SUREOHPV VLPLODU WR WKRVH RI WKH VXEMHFW EXW ZKR DFWLYHO\ FRSHV ZLWK DQG RYHUFRPHV WKRVH GLIILFXOWLHV )DLOXUH IHHGEDFN KRZHYHU LV FOHDUFXW HYLGHQFH RI D IDLOXUH LQ FRSLQJ HIIRUWV :LQH f 6DUDVRQ f H[DPLQHG ZKHWKHU D FRSLQJ PRGHO ZDV LQ IDFW IDFLOLWDWLYH RI FRJQLn WLYH WDVN SHUIRUPDQFH 6XEMHFWV ZHUH H[SRVHG WR D PRGHO ZKR DGPLWWHG H[SHULHQFLQJ DQ[LHW\ DQG GHVFULEHG ZD\V RI FRSLQJ ZLWK LW RU WR RQH ZKR RQO\ DGPLWWHG WR WKH DQ[LHW\ YHUVXV D PRGHO ZKR GLG QRW DGPLW WR DQ\ DQ[LHW\ ([SRVXUH WR WKH VHOIGLVFORVLQJ FRSLQJ PRGHO ZDV KLJKO\ IDFLOLWDWLYH RI WKH WHVWDQ[LRXV VXEMHFWVn FRJQLWLYH WDVN SHUIRUPDQFH 7KH UHVXOWV RI WKHVH VWXGLHV VXJJHVW LQ JHQHUDO WKDW H[SRVXUH WR PRGHOV ZKR DUH WDVNRULHQWHG DQG ZKR SURYLGH

PAGE 27

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

PAGE 28

HIIHFWV RI WKHVH DSSURDFKHV RQ ERWK FRJQLWLYH WDVN SHUIRUn PDQFH DQG VHOIUHSRUW RI WHVW DQ[LHW\ 6\VWHPDWLF 'HVHQVLWL]DWLRQ 0RVW RI WKH HDUO\ WUHDWPHQW OLWHUDWXUH DVVXPHG HPRWLRQDO UHDFWLYLW\ WR EH WKH PDMRU FKDUDFWHULVWLF RI WHVW DQ[LHW\ $V VXFK PXFK RI WKH HDUO\ WUHDWPHQW OLWHUDWXUH IRFXVHG RQ V\VWHPDWLF GHVHQVLWL]DWLRQ SURJUHVVLQJ IURP VLPSOH GHPRQVWUDWLRQV RI WKH HIILFDF\ RI V\VWHPDWLF GHVHQVLn WL]DWLRQ .DWDKQ 6WUHQJHU t &KHUU\ *DUOLQJWRQ t &RWLHU 3DXO f WKURXJK V\VWHPDWLF PDQLSXODWLRQV RI WKH GHVHQVLWL]DWLRQ SURFHVV HJ )UHHOLQJ t 6KHPEHUJ f 6\VWHPDWLF GHVHQVLWL]DWLRQ WUHDWPHQW IRU WHVW DQ[LHW\ EDVLFDOO\ LQYROYHV WKH SDLULQJ RI GHHS PXVFOH UHOD[n DWLRQ ZLWK WKH LQVWUXFWHG LPDJLQDWLRQ RI DYHUVLYH VFHQHV WR LQKLELW DXWRQRPLF DURXVDO DQG SURYLGH D QHZ UHVSRQVH WR EH FRQGLWLRQHG WR SUHYLRXVO\ DQ[LHW\DURXVLQJ VWLPXOL 7KH V\VWHPDWLF GHVHQVLWL]DWLRQ SURFHGXUH IROORZHG WKH JHQHUDO RXWOLQH XVHG E\ 3DXO f LQ RQH RI WKH HDUOLHVW GHPRQn VWUDWLRQV RI WKLV WUHDWPHQW IRU WH[W DQ[LHW\ 6WXGLHV FRPSDUHG LQGLYLGXDOO\GHVLJQHG YV VWDQGDUGL]HG GHVHQVLWLn ]DWLRQ KLHUDUFKLHV (PHU\ t .UXPEROW] f V\VWHPDWLF GHVHQVLWL]DWLRQ YV UHOD[DWLRQ WKHUDS\ -RKQVRQ t 6HFUHVW f DXWRPDWHG YV OLYHWKHUDSLVW GHVHQVLWL]DWLRQ 'RQQHU

PAGE 29

6W *XHUQH\ f DQG RWKHU YDULDWLRQV RI SUHVHQWLQJ WKH V\VWHPDWLF GHVHQVLWL]DWLRQ WUHDWPHQW $SRQWH t $SRQWH )UHHOLQJ t 6KHPEHUJ 6XLQQ f ,Q JHQHUDO WKHVH VWXGLHV KDYH IRXQG WKDW V\VWHPDWLF GHVHQVLWL]DWLRQ LV HIIHFn WLYH LQ DOOHYLDWLQJ WHVW DQ[LHW\ ZKHQ VHOIUHSRUW PHDVXUHV RI DQ[LHW\ OHYHO DUH XVHG DV RXWFRPH FULWHULD +RZHYHU WKH HIIHFWV RI GHVHQVLWL]DWLRQ RQ DFDGHPLF SHUIRUPDQFH DUH QRW TXLWH DV FRQVLVWHQW 6RPH VWXGLHV GHPRQVWUDWH DQ LQFUHDVH LQ JUDGH SRLQW DYHUDJH IRU WUHDWHG JURXSV FRPSDUHG WR FRQWURO JURXSV $SRQWH t $SRQWH 'RQQHU t *XHUQH\ -RKQVRQ t 6HFUHVW DQG .DWDKQ 6WUHQJHU t &KHUU\ f 2WKHUV VKRZHG QR LQFUHPHQW LQ DFDGHPLF SHUIRUPDQFH *DUOLQJWRQ t &RWLHU (PHU\ t .UXPEROW] f RU GLG QRW XVH SHUIRUPDQFH PHDVXUHV DV RXWFRPH FULWHULD )UHHOLQJ t 6KHPEHUJ 6XLQQ f $OOHQ f VWDWHV WKH IDFW WKDW WKH GHVHQVLWL]DWLRQ VWXGLHV GLG QRW FRQWURO IRU WKH HIIHFWV RI QRQVSHFLILF WKHUDSHXWLF IDFWRUV ZKLFK PLJKW HQKDQFH WUHDWPHQW EHQHILWV LQ VRPH VWXGLHV DQG QRW RWKHUV 7KHVH QRQVSHFLILF IDFWRUV LQFOXGH KHLJKWHQHG VXEMHFW H[SHFWDQF\ WKH XVH RI RQO\ RQH WKHUDSLVW WKH XVH RI ILUVW VHPHVWHU IUHVKPHQ KDYH QRW \HW DGDSWHG WR FROOHJH URXWLQH DQG PD\ LQWHUSUHW WKHLU VWUHVV DV WHVW DQ[LHW\f DQG WKH ODFN RI DSSURSULDWH FRQWURO JURXSV HJ ZDLWLQJ OLVW FRQWURO JURXSf 7KH V\VWHPDWLF GHVHQVLWL]DWLRQ UHVHDUFK GRHV GHPRQVWUDWH WKDW DQ HPRWLRQDOLW\ UHGXFWLRQ WHFKQLTXH FDQ HIIHFW FKDQJHV

PAGE 30

LQ VHOIUHSRUW RI DQ[LHW\ OHYHO +RZHYHU DQ HPRWLRQDOLW\ RULHQWHG WUHDWPHQW DSSURDFK PD\ QRW EH VXIILFLHQW WR FRQn VLVWHQWO\ DOWHU SRRU SHUIRUPDQFH XQGHU DFDGHPLFDOO\HYDOXDWLYH FRQGLWLRQV 0RUH H[WHQVLYH PHWKRGRORJLFDOO\VRXQG UHVHDUFK RQ WKH HIIHFWV RI V\VWHPDWLF GHVHQVLWL]DWLRQ WUHDWPHQWV RQ FRJQLWLYH SHUIRUPDQFH LV UHTXLUHG WR PDNH PRUH GHILQLWLYH VWDWHPHQWV 6HOI&RQWURO 7UHDWPHQW 7HFKQLTXHV 'HQQH\ f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f WR PRUH FRJQLWLYHO\ RULHQWHG WUHDWPHQWV VHOIFRQWURO WUDLQLQJ DQG FRJQLWLYH FRSLQJf WKHUH LV LQFUHDVLQJ HPSKDVLV RQ LQYLYR DSSOLFDWLRQ WUDLQLQJ WUDLQLQJ LQ WKH DFWXDO DQ[LHW\SURYRNLQJ VHWWLQJf JXLGHG UHKHDUVDO DQG FRJQLWLYH UHVWUXFWXULQJ $OO RI WKHVH DSSURDFKHV HPSKDVL]H WKH DWWDLQPHQW RI FRSLQJ VNLOOV ZKLFK

PAGE 31

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f 7KHVH SURFHGXUHV DUH LQWURGXFHG ZLWK D VHOIFRQWURO UDWLRQDOH VLPLODU WR WKH RQH DGYRFDWHG E\ *ROGIULHG f &OLHQWV DUH XVXDOO\ LQIRUPHG WKDW WKH SXUSRVH RI WKH WUHDWPHQW LV WR SURYLGH WKHP ZLWK VNLOOV IRU DFWLYHO\ FRSLQJ ZLWK DQ[LHW\ 7KH\ DUH WROG WKDW WKH UHOD[DWLRQ WUDLQLQJ WKH\ UHFHLYH LV D PHWKRG IRU KHOSLQJ WKHP WR EULQJ WKH UHVSRQVH RI UHOD[DWLRQ XQGHU YROXQWDU\ FRQWURO )XUWKHU WKH\ DUH WROG WKDW ZLWK JUHDWHU SURILFLHQF\ DW LQGXFLQJ UHOD[DWLRQ WKH\ ZLOO EH DEOH WR DSSO\ UHOD[DWLRQ WR UHGXFH IHHOLQJV RI DQ[LHW\ LQ VWUHVVIXO VLWXDWLRQV HQFRXQWHUHG RXWVLGH RI WUHDWPHQW f 7KH VHFRQG IHDWXUH RI WKH DSSOLHG UHOD[DWLRQ WHFKQLTXHV LV WUDLQLQJ LQ WKH LQGXFWLRQ RI UHOD[DWLRQ 3URJUHVVLYH PXVFOH UHOD[DWLRQ H[HUFLVHV 5LPP t 0DVWHUV f DGDSWHG IURP -DFREVHQnV f ZRUN DUH WKH PRVW FRPPRQ PHWKRG

PAGE 32

IRU LQGXFLQJ UHOD[DWLRQ DOWKRXJK QXPHURXV RWKHU DSSURDFKHV KDYH EHHQ GHYHORSHG VXSSOHPHQWLQJ UHOD[DWLRQ ZLWK ELRIHHGn EDFNDVVLVWDQFH 5RPDQR t &DELDQFD f FXHFRQWUROOHG UHOD[DWLRQ 5XVVHOO t 6LSLFK f DQG LPDJHU\ H[HUFLVHV 6DPXHOV t 6DPXHOV f DQG RWKHU VXSSOHPHQWDU\ H[HUFLVHV f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

PAGE 33

&RJQLWLYH &RSLQJ 7HFKQLTXHV &RJQLWLYH FRSLQJ WHFKQLTXHV ZKLFK DUH WKH PRVW FRJQLWLYHO\RULHQWHG DSSURDFKHV DWWHPSW WR SURYLGH FOLHQWV ZLWK DQ DGGLWLRQDO UHSHUWRLUH RI FRSLQJ VNLOOV EH\RQG RU LQ VRPH FDVHV LQVWHDG RI UHOD[DWLRQ VNLOOV 7KH GLVWLQFn WLYH IHDWXUH RI WKH FRJQLWLYH FRSLQJ WHFKQLTXHV LV VRPH IRUP RI FRJQLWLYH UHVWUXFWXULQJ &RJQLWLYH UHVWUXFWXULQJ KDV WKUHH EDVLF REMHFWLYHV f &OLHQWV DUH SHUVXDGHG WKDW WKH EHOLHIV WKDW WKH\ HQWHUWDLQ ZKLOH FRQIURQWLQJ FHUWDLQ VLWXDWLRQV DIIHFW WKH HPRWLRQDO UHDFWLRQV WKH\ KDYH LQ WKHVH VLWXDWLRQV f 7KH\ DUH QH[W HQFRXUDJHG WR LGHQWLI\ WKH SDUWLFXODU QHJDWLYH VHOIVWDWHPHQWV WKDW WKH\ PDNH ZKHQ FRQIURQWLQJ D VWUHVVIXO VLWXDWLRQ f &OLHQWV DUH KHOSHG WR IRUPXODWH SRVLWLYH DQG UDWLRQDO VHOIVWDWHPHQWV ZKLFK ZLOO UHSODFH WKH QHJDWLYH VHOIVWDWHPHQWV DQG WKHUHE\ DWWHQXDWH WKHLU VWUHVV LQ WKH HYDOXDWLYH VLWXDWLRQ $V GLVFXVVHG HDUOLHU LQ UHJDUG WR WKH V\VWHPDWLF GHVHQVLWL]DWLRQ VWXGLHV PRVW RI WKH DSSOLHG UHOD[DWLRQ VWXGLHV 5XVVHOO 0LOOHU t -XQH 5XVVHOO :LVH t 6WUDWRXGDNLV 0DUFKHWWL 0F*O\QQ t 3DWWHUVRQ DQG &KLDQJ/LDQJ t 'HQQH\ f IRXQG WKDW WKLV WUHDWPHQW DSSURDFK GHPRQVWUDWHG FKDQJHV RQ VHOIUHSRUW PHDVXUHV RI WHVWDQ[LHW\ 7KH VDPH JHQHUDOL]DWLRQ FRXOG EH DSSOLHG WR WKH UHVXOWV RI WKH VHOIFRQWURO WUDLQLQJ WHFKQLTXHV 'HIIHQEDFKHU 0DWKLV t 0LFKDHOV =HPRUH

PAGE 34

'HIIHQEDFKHU t 3DUNV 'HQQH\ t 5XSHUW f DQG WKH FRJQLWLYH FRSLQJ WHFKQLTXHV 0HLFKHQEDXP +ROUR\G /DYLJQH :LQH *ROGIULHG /LQHKDQ t 6PLWK f $QDO\VHV RI WKH FRPSDUDWLYH LPSDFW RI WKHVH WUHDWPHQW VWUDWHJLHV RQ FRJQLWLYH SHUIRUPDQFH GHPRQVWUDWHV TXLWH GLIIHUHQW ILQGLQJV 2QO\ RQH RXW RI WKUHH RI WKH DSSOLHG UHOD[DWLRQ VWXGLHV UHVXOWHG LQ FRJQLWLYH SHUIRUPDQFH LPSURYHPHQW ,W ZDV IRXQG WKDW IRXU RI WKH HLJKW VHOIFRQWURO WUDLQLQJ VWXGLHV UHYLHZHG UHYHDOHG LPSURYHPHQWV RQ SHUIRUn PDQFH PHDVXUHV ZKHUHDV ILYH RI WKH VHYHQ FRJQLWLYHFRSLQJ IW WHFKQLTXH VWXGLHV GHPRQVWUDWHG LPSURYHPHQWV LQ WKHVH W\SHV RI PHDVXUHV 7KXV WKH FRJQLWLYH FRSLQJ WHFKQLTXHV ZKLFK LQFRUSRUDWH VRPH IRUP RI FRJQLWLYH UHVWUXFWXULQJ DUH REVHUYHG WR KDYH D KLJKHU UDWH RI VXFFHVV WKDQ VHOIFRQWURO WUDLQLQJ RU DSSOLHG UHOD[DWLRQ WHFKQLTXHV LQ WHUPV RI SHUIRUPDQFH LPSURYHPHQW %RWK 0HLFKHQEDXP f DQG +ROUR\G f FRPSDUHG FRJQLWLYH FRSLQJ WUHDWPHQW DSSURDFKHV GLUHFWO\ ZLWK V\VWHPDWLF GHVHQVLWL]DWLRQ ZKLFK LV DQ DURXVDO UHGXFWLRQ WHFKQLTXH 0HLFKHQEDXPnV FRJQLWLYH PRGLILFDWLRQ WUHDWPHQW SURFHGXUH FRPELQHG DQ LQVLJKWRULHQWHG WKHUDS\ ZKLFK ZDV GHVLJQHG WR PDNH WHVWDQ[LRXV VXEMHFWV DZDUH RI WKHLU DQ[LHW\HQJHQGHULQJ WKRXJKWV ZLWK D PRGLILHG GHVHQVLWL]DWLRQ SURFHGXUH ZKLFK HPSOR\HG Df FRSLQJ LPDJHU\ RQ KRZ WR KDQGOH DQ[LHW\ DQG Ef VHOILQVWUXFWLRQDO WUDLQLQJ WR DWWHQG WR WKH WDVN DQG QRW UXPLQDWH DERXW RQHVHOI 5HVXOWV RI WKH VWXG\ LQGLFDWHG

PAGE 35

WKDW WKH FRJQLWLYH PRGLILFDWLRQ JURXS ZDV PRVW HIIHFWLYH LQ VLJQLILFDQWO\ UHGXFLQJ WHVW DQ[LHW\ DV DVVHVVHG E\ Df WHVW SHUIRUPDQFH REWDLQHG LQ DQ DQDORJXH WHVW VLWXDWLRQ Ef VHOI UHSRUW RI DQ[LHW\ LPPHGLDWHO\ DIWHU WUHDWPHQW DQG DW D RQHPRQWK IROORZXS DQG Ff JUDGH SRLQW DYHUDJH 7KHUH ZDV OHVV FRQVLVWHQW EXW JHQHUDO LPSURYHPHQW E\ WKH GHVHQVLWL]Dn WLRQ WUHDWPHQW JURXS ZKR DSSHDUHG VLJQLILFDQWO\ PRUH LPSURYHG WKDQ WKH ZDLWLQJ OLVW FRQWUROV 7KH +ROUR\G f VWXG\ LV DFWXDOO\ D IROORZXS VWXG\ RQ WKH 0HLFKHQEDXP f VWXG\ 0HLFKHQEDXPnV UHVXOWV DUH VRPHZKDW GLIILFXOW WR LQWHUSUHW EHFDXVH KLV FRJQLWLYH PRGLILFDWLRQ WUHDWPHQW LV DFWXDOO\ D FRPELQDWLRQ RI D FRJQLWLYH WUHDWPHQW WKH XVH RI VHOIVWDWHPHQWV DQG FRJQLWLYH FRSLQJ LPDJHU\f DQG DQ HPRWLRQDO DURXVDO WUHDWPHQW WKH XVH RI UHOD[DWLRQf ,W LV QRW FOHDU ZKHWKHU WKH SRVLWLYH UHVXOWV RI WKH FRJQLWLYH PRGLILFDWLRQ WUHDWPHQW DUH SULPDULO\ GXH WR WKH FRJQLWLYH DVSHFWV WKH DURXVDO UHGXFWLRQ DVSHFWV RU WKH FRPELQDWLRQ RI WKH WZR WUHDWPHQW IRUPV +ROUR\G VRXJKW WR VHSDUDWH RXW WKHVH LQGLYLGXDO WUHDWPHQW FRPSRQHQWV +H FRPSDUHG WKH HIIHFWLYHQHVV RI FRJQLWLYH DURXVDO UHGXFWLRQ DQG FRPELQHG DURXVDOFRJQLWLYH WUHDWPHQWV IRU WHVW DQ[LHW\ 7KH FRJQLWLYH WKHUDS\ IRFXVHG H[FOXVLYHO\ RQ HOLPLQDWLQJ WKH WHVWDQ[LRXV FOLHQWnV WDVNLUUHOHYDQW UXPLQDWLRQV DQG DWWHQWLRQDO IRFXV 7KH FOLHQWV ZHUH WDXJKW WR HPLW LQFRPn SDWLEOH VHOIVWDWHPHQWV GHVLJQHG WR IDFLOLWDWH WDVNDWWHQWLRQ LQ VWUHVVIXO VLWXDWLRQV 7KH HPRWLRQDO DURXVDO WUHDWPHQW

PAGE 36

ZDV WKH WUDGLWLRQDO JURXS V\VWHPDWLF GHVHQVLWL]DWLRQ SDUDGLJP 3DXO f 7KH FRPELQHG WUHDWPHQW IRFXVHG RQ ERWK HOLPLQDWLQJ WKH WHVWDQ[LRXV FOLHQWnV WDVNLUUHOHYDQW UXPLQDWLRQV DQG WDVN IRFXV DQG UHGXFLQJ WKHLU HPRWLRQDO DURXVDO 7KHUH ZDV DOVR D SVHXGRWKHUDS\ FRQWURO JURXS LQFOXGHG LQ WKH VWXG\ WR DVVHVV WKH GHJUHH RI LPSURYHPHQW DWWULEXWHG WR QRQVSHFLILF WUHDWPHQW HIIHFWV 7KH UHVXOWV GHPRQVWUDWHG WKDW WKH FRJQLWLYH WKHUDS\ DSSURDFK ZDV VLJQLILFDQWO\ PRUH HIIHFWLYH WKDQ WKH RWKHU WUHDWPHQW DSSURDFKHV 7KH WUHDWPHQW JDLQV ZHUH VHHQ LQ ERWK ODERUDWRU\ PHDVXUHV RI WHVW DQ[LHW\ VHOIUHSRUWf WHVW SHUIRUPDQFH LQ DQ DQDORJ WHVW VLWXDWLRQ DQG JUDGH SRLQW DYHUDJH 2QO\ WKH FRJQLWLYH WKHUDS\ DSSURDFK GHPRQVWUDWHG VLJQLILFDQWO\ EHWWHU UHVXOWV WKDQ WKH SVHXGRWKHUDS\ FRQWURO SURFHGXUH 7KH V\VWHPDWLF GHVHQVLWL]DWLRQ JURXS WKH FRPELQHG WUHDWPHQW JURXS DQG WKH SVHXGRWKHUDS\ JURXSV SURGXFHG VLPLODU RXWFRPHV H[FHSW IRU WKH DQDORJ WHVW VLWXDWLRQ GLJLW V\PERO WHVW SHUIRUPDQFHf ZKHUH WKH FRPELQHG WUHDWPHQW DSSURDFK ZDV VXSHULRU WR WKH V\VWHPDWLF GHVHQVLWL]DWLRQ DQG WKH SVHXGRWKHUDS\ DSSURDFK 7KH UHVXOWV VXJJHVW WKDW ZKLOH V\VWHPDWLF GHVHQVLWL]DWLRQ LV DQ HIIHFWLYH WUHDWPHQW IRU WHVW DQ[LHW\ LW PD\ EH QR PRUH HIIHFWLYH WKDQ DQ HTXDOO\ FRQYLQFLQJ WUHDWPHQW ULWXDO RI DQ\ RWKHU VRUW 7KH UHVXOWV IRU WKH FRPELQHG WUHDWPHQW DSSURDFK DUH VRPHZKDW SX]]OLQJ 2WKHU VWXGLHV XVLQJ FRPELQHG WUHDWPHQWV KDYH \LHOGHG FRQIOLFWLQJ UHVXOWV ZLWK WKH FRPELQHG WUHDWPHQW

PAGE 37

EHLQJ OHVV HIIHFWLYH WKDQ FRJQLWLYH RU O WL]DWLRQ WUHDWPHQWV DORQH 0HLFKHQEDXP *LOPRUH t )HGRUDYLFLXV f RU PRUH HIIHFWLYH WKDQ V\VWHPDWLF GHVHQVLWL]DWLRQ DORQH 0HLFKHQEDXP f +ROUR\G f VXJJHVWV WKDW WKH UHVXOWV REWDLQHG ZLWK FRPELQHG WUHDWPHQWV PD\ EH LQIOXHQFHG E\ SURFHGXUDO YDULDEOHV VXFK DV OHQJWK RI WUHDWPHQWV RUGHU RI WUHDWPHQWV DQG WKH WUHDWPHQWV ZKLFK DUH FRPELQHG 7KHUH KDYH EHHQ RWKHU FRJQLWLYH FRSLQJ DSSURDFKHV ZKLFK KDYH SURYHQ WR EH VXFFHVVIXO LQ LPSURYLQJ DFDGHPLF SHUIRUPDQFH DQG UHGXFLQJ VHOIUHSRUW RI WHVW DQ[LHW\ 5HFHQWO\ +DUULV DQG -RKQVRQ f KDYH FRPSDUHG D XQLTXH FRYHUW PRGHOLQJ DSSURDFK WR D VHOIFRQWURO GHVHQVLWL]DWLRQ WUHDWPHQW DQG D VWXG\VNLOOV WUDLQLQJ JURXS ZLWK WKH FRYHUW PRGHOLQJ JURXS DQG WKH GHVHQVLWL]DWLRQ JURXS UHFHLYLQJ VRPH VWXG\ VNLOOV WUDLQLQJ DV ZHOOf +DUULV DQG -RKQVRQ f K\SRWKHVL]HG WKDW DQ LQGLYLGXDO ZRXOG EH OLNHO\ WR EHQHILW PRUH IURP D PRGHOLQJ SURFHGXUH LI WKH PRGHO SHUIRUPLQJ WKH EHKDYLRU LV WKH LQGLYLGXDO KLPVHOI 7KHUHn IRUH WKH\ GHYHORSHG D WUHDWPHQW LQ ZKLFK WKH LQGLYLGXDO XWLOL]HV KLV RZQ KLJKO\ SHUVRQDOL]HG LPDJHV RI VHOIFRPSHWHQF\ LQ VWUHVVIXO VLWXDWLRQV WR FRSH ZLWK KLV WHVW DQ[LHW\ 7KHVH LPDJHV DUH WDNHQ IURP OLIH VLWXDWLRQV RWKHU WKDQ WHVWWDNLQJ LQ ZKLFK WKH LQGLYLGXDO RSHUDWHV VXFFHVVIXOO\ IRU H[DPSOH SOD\LQJ LQ D PXVLF UHFLWDO RU UXQQLQJ D UDGLR VKRZ $IWHU WKH LQGLYLGXDO GHYHORSV KLV RZQ XQLTXH VXFFHVV LPDJHU\ WKHVH LPDJHV DUH SDLUHG ZLWK DQ[LHW\HOLFLWLQJ LPDJHV WDNHQ

PAGE 38

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n IRUPDQFH 6XEMHFWV LQ WKH VWXG\ VNLOOV WUDLQLQJ JURXS KDG D VPDOO EXW QRQVLJQLILFDQW LQFUHDVH LQ JUDGH SRLQW DYHUDJH EXW GLG QRW VKRZ UHGXFWLRQ LQ WHVWDQ[LHW\ RQ DOO RI WKH DQ[LHW\ PHDVXUHV $OWKRXJK WKH VWXG\ VNLOOV WUDLQLQJ SURFHGXUH SUHVXPDEO\ SUHYHQWHG D GHFUHPHQW LQ DFDGHPLF SHUIRUPDQFH LW ZDV QRW HIIHFWLYH LQ HLWKHU UHGXFLQJ WHVW DQ[LHW\ RU VLJQLILFDQWO\ LQFUHDVLQJ JUDGH SRLQW DYHUDJH 7KH +DUULV DQG -RKQVRQ f VWXG\ FOHDUO\ LQGLFDWHV WKDW D FRJQLWLYH FRSLQJ DSSURDFK WR WUHDWPHQW LQ ZKLFK SRVLWLYH LPDJHU\ VXSSODQWV QHJDWLYH DQ[LHW\SURYRNLQJ LPDJHU\ LV KLJKO\ HIIHFWLYH LQ DOWHULQJ ERWK VHOIUHSRUW RI WHVW DQ[LHW\ DQG LPSRUWDQWO\ DFDGHPLF SHUIRUPDQFH 7KLV

PAGE 39

WUHDWPHQW DOVR VHUYHV WR KLJKOLJKW WR WKH LQGLYLGXDO KLV RZQ DUHDV RI FRPSHWHQFH 5HLQIRUFHPHQW RI D VHQVH RI PDVWHU\ DSSHDUV WR EH D PRUH HIIHFWLYH WUHDWPHQW WKDQ PXVFXODU UHOD[DWLRQ +DUULV f LQYHVWLJDWHG WKH FRQFHSW RI LQGLYLGXDOL]HG FRSLQJ LPDJHU\ LQ PRUH GHWDLO LQ D IROORZXS VWXG\ 6KH ZDV LQWHUHVWHG LQ WKH HIIHFWLYHQHVV RI GLIIHUHQW LPDJHU\ LQVWUXFn WLRQ SURFHGXUHV LQ WKH WUHDWPHQW RI WHVW DQ[LHW\ 6SHFLILFDOO\ +DUULV FRPSDUHG LQGLYLGXDOL]HG FRSLQJ LPDJHU\ WUHDWPHQW EDVHG RQ QRQDFDGHPLF H[SHULHQFHV RI FRPSHWHQFH DQG VXFFHVV WR LQGLYLGXDOL]HG FRSLQJ LPDJHU\ WUHDWPHQW EDVHG RQ DFDGHPLF H[SHULHQFHV RI FRPSHWHQFH DQG VXFFHVV 6LQFH WKH LPDJHV RI FRPSHWHQFH XWLOL]HG LQ WKH +DUULV DQG -RKQVRQ f VWXG\ ZHUH QRQDFDGHPLF LQ QDWXUH +DUULV K\SRWKHVL]HG WKDW LQVWUXFWLRQ LQ FRSLQJ LPDJHV EDVHG RQ SUHYLRXV DFDGHPLF VXFFHVV H[SHULHQFHV PLJKW SURYH HYHQ PRUH HIIHFWLYH LQ UHGXFLQJ WHVW DQ[LHW\ DQG LPSURYLQJ DFDGHPLF SHUIRUPDQFH ,Q DGGLWLRQ +DUULV ZDQWHG WR DVVHVV WKH H[WHQW WR ZKLFK HIIHFWLYH UHOD[DWLRQ WUDLQLQJ ZRXOG HQKDQFH WKH RYHUDOO WUHDWPHQW HIIHFWLYHQHVV RI WKH LQGLYLGXDOL]HG FRSLQJ LPDJHU\ WHFKQLTXHV &RQVHTXHQWO\ VKH FRPSDUHG IRXU WUHDWPHQW WHFKQLTXHV f QRQDFDGHPLF LQGLYLGXDOL]HG FRSLQJ LPDJHU\ ZLWK UHOD[DWLRQ WUDLQLQJ f QRQDFDGHPLF LQGLYLGXDOL]HG FRSLQJ LPDJHU\ ZLWKRXW UHOD[DWLRQ WUDLQLQJ f DFDGHPLF FRSLQJ LPDJHU\ ZLWK UHOD[DWLRQ WUDLQLQJ DQG f DFDGHPLF FRSLQJ LPDJHU\ ZLWKRXW UHOD[DWLRQ WUDLQLQJ 7KH VWXG\ IRXQG WKDW RQO\ WKH DFDGHPLF FRSLQJ LPDJHU\ JURXS FRPELQHG ZLWK

PAGE 40

UHOD[DWLRQ DQG WKH LQGLYLGXDOL]HG FRSLQJ LPDJHU\ JURXS ZLWKRXW UHOD[DWLRQ ZHUH VLJQLILFDQWO\ GLIIHUHQW IURP WKH FRQWURO JURXS LQ DFDGHPLF SHUIRUPDQFH JUDGH SRLQW DYHUDJHf DW SRVWWHVW 3UHWHVW WR SRVWWHVW WUHDWPHQW FKDQJHV LQ JUDGH SRLQW DYHUDJH ZHUH VLJQLILFDQW IRU ERWK QRQDFDGHPLF LPDJHU\ JURXSV S f DQG DSSURDFKHG VLJQLILFDQFH IRU WKH DFDGHPLF FRSLQJ LPDJHU\ JURXS ZLWK UHOD[DWLRQ S f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f K\SRWKHVL]HG WKDW WKH DFDGHPLF FRSLQJ LPDJHU\ WUHDWPHQW ZLWKRXW UHOD[DWLRQ ZDV D OHVV HIIHFWLYH SURFHGXUH EHFDXVH DFDGHPLF LPDJHV ZHUH OLNHO\ WR HOLFLW IHDU DQG DYRLGDQFH UDWKHU WKDQ LPDJHV DQG IHHOLQJV RI FRPSHWHQFH LQ WKHVH VXEMHFWV VLQFH DFDGHPLF H[SHULHQFHV DUH VR FORVHO\ OLQNHG WR WKH WHVWWDNLQJ VLWXDWLRQ &RQn VHTXHQWO\ WKH\ KDG PRUH GLIILFXOW\ HYRNLQJ LPDJHV RI FRPSHWHQFH UHJDUGLQJ DFDGHPLF H[SHULHQFHV 7KH DGGLWLRQ RI UHOD[DWLRQ WR WKH DFDGHPLF FRSLQJ LPDJHU\ SDUDGLJP PD\

PAGE 41

KDYH DOORZHG WKH VXEMHFWV WR EH PRUH FRPIRUWDEOH DERXW DFDGHPLF LPDJHU\ VLQFH UHOD[DWLRQ WUDLQLQJ KDV EHHQ K\SRWKHVL]HG WR UHGXFH GLVWUDFWLQJ WKRXJKWV DQG SK\VLRORJn LFDO UHVSRQVHV /DQJ f +HQFH WKH DFDGHPLF FRSLQJ LPDJHU\ ZLWK UHOD[DWLRQ WUHDWPHQW ZDV PRUH HIILFDFLRXV WKDQ WKH QRQUHOD[DWLRQ JURXS DQG WKH GHVHQVLWL]DWLRQ KLHUDUFK\ ZDV FRPSOHWHG PRUH HIILFLHQWO\ 7KH UHOD[DWLRQ FRPSRQHQW GLG QRW VLJQLILFDQWO\ DGG WR WKH HIIHFWLYHQHVV RI WKH QRQDFDGHPLF LQGLYLGXDOL]HG FRSLQJ LPDJHU\ SDUDGLJP SUHVXPDEO\ EHFDXVH WKH VXEMHFWV ZHUH LQLWLDOO\ PRUH DEOH WR YLVXDOL]H QRQDFDGHPLF LPDJHU\ ZLWKRXW IHDU RU DYRLGDQFH 7KH +DUULV f VWXG\ DQG WKH +DUULV DQG -RKQVRQ f VWXG\ KLJKOLJKW WKH ILQGLQJV RI SUHYLRXV UHVHDUFK WKDW ZKLOH ERWK HPRWLRQDOLW\ DQG FRJQLWLYHO\RULHQWHG WUHDWPHQWV DUH HIIHFWLYH LQ LPSURYLQJ VHOIUHSRUW RI WHVW DQ[LHW\ LW LV WKH FRJQLWLYH WUHDWPHQW DSSURDFKHV DQG PXOWLPRGDO DSSURDFKHV ZKLFK DOOHYLDWH WKH FRJQLWLYH SHUIRUPDQFH GHILFLWV RI WKH WHVWDQ[LRXV SHUVRQ )XUWKHUn PRUH WKH +DUULV f VWXG\ LQGLFDWHV WKDW PXOWLPRGDO DSSURDFKHV PD\ QRW EH QHFHVVDU\ LI D SRZHUIXO FRJQLWLYH FRSLQJ PRGHO LV DSSOLHG VXFK DV WKH QRQDFDGHPLF LQGLYLGXDOL]HG FRSLQJ LPDJHU\ SDUDGLJP XWLOL]HG LQ WKDW V WXG\

PAGE 42

*HQHUDO &RQFOXVLRQV :KLOH PRVW RI WKH HDUO\ UHVHDUFK LQ WKH DUHD RI WHVW DQ[LHW\ ZDV IRFXVHG XSRQ WKH GHELOLWDWLQJ HIIHFWV RI WHVW DQ[LHW\ :LQH f FXUUHQW UHVHDUFK KDV EHJXQ WR LQYHVWLn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n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f ,Q RUGHU WR PRUH HIIHFWLYHO\ SURSRVH FRJQLWLYHO\RULHQWHG WUHDWPHQW VWUDWHJLHV DV ZHOO

PAGE 43

DV PRUH IXOO\ XQGHUVWDQG WHVW DQ[LHW\ DV D FRJQLWLYH FRQVWUXFW ZH QHHG WR LQYHVWLJDWH WKH H[WHQW WR ZKLFK WHVW DQ[LHW\ LV UHODWHG WR VSHFLILF FRJQLWLYHPHGLDWLRQDO YDULDEOHV 6HOI(IILFDF\ 7KHRU\ 2QH FRJQLWLYHPHGLDWLRQDO YDULDEOH ZKLFK KDV EHHQ H[DPLQHG SULPDULO\ LQ UHODWLRQ WR SKRELF EHKDYLRU LV WKH YDULDEOH RI VHOIHIILFDF\ 7KH WHUP VHOIHIILFDF\ ZDV RULJLQDOO\ FRQFHSWXDOL]HG E\ %DQGXUD f DV D PDVWHU\ H[SHFWDQF\ YDULDEOH ZKLFK FRXOG EH D SRZHUIXO SUHGLFWRU RI FXUUHQW OHYHO RI EHKDYLRU DV ZHOO DV D SUHGLFWRU RI EHKDYLRU FKDQJH IROORZLQJ WKHUDSHXWLF LQWHUYHQWLRQ ,Q IDFW WKH EDVLF SUHPLVH RI WKH WKHRU\ LV WKDW WKH FOLHQWnV HIILFDF\ H[SHFWDWLRQV DUH FHQWUDO WR EHKDYLRU FKDQJH 7KH WKHRU\ SURn SRVHV WKDW H[SHFWDWLRQV RI SHUVRQDO HIILFDF\ GHWHUPLQH ZKHWKHU FRSLQJ EHKDYLRU ZLOO EH LQLWLDWHG KRZ PXFK HIIRUW ZLOO EH H[SDQGHG DQG KRZ ORQJ LW ZLOO EH VXVWDLQHG LQ WKH IDFH RI REVWDFOHV DQG DYHUVLYH FLUFXPVWDQFHV %DQGXUD S f 7KURXJK H[SHULHQFHV RI PDVWHU\ SURYLGHG YLD YDULRXV WKHUDn SHXWLF PRGHV WKHUH LV HQKDQFHPHQW RI VHOIHIILFDF\ DQG FRUUHVSRQGLQJ UHGXFWLRQ LQ WDVN RU VLWXDWLRQDO DYRLGDQFH EHKDYLRU 7KRVH ZKR SHUVLVW LQ VXEMHFWLYHO\ WKUHDWHQLQJ DFWLYLWLHV WKDW WXUQ RXW WR EH UHODWLYHO\ VDIH

PAGE 44

JDLQ FRUUHFWLYH H[SHULHQFHV WKDW UHLQIRUFH WKHLU VHQVH RI VHOIHIILFDF\ %DQGXUD f :KHQ LQGLYLGXDOV FHDVH WKHLU FRSLQJ HIIRUWV SUHPDWXUHO\ WKH\ UHWDLQ WKHLU GHELOLWDWLQJ H[SHFWDWLRQV DQG IHDUV LQGHILQLWHO\ 3HUFHLYHG VHOIHIILFDF\ LV WKXV SUHVXPHG WR DIIHFW FRSLQJ HIIRUWV DV ZHOO DV FKRLFH RI EHKDYLRUDO VHWWLQJV DQG DFWLYLWLHV %DQGXUD f HPSKDVL]HV WKDW H[SHFWDWLRQ LV QRW WKH VROH GHWHUn PLQDQW RI EHKDYLRU ([SHFWDWLRQ RI PDVWHU\ DORQH ZLOO QRW SURGXFH GHVLUHG SHUIRUPDQFH LI WKH FRPSRQHQW FDSDELOLWLHV DUH ODFNLQJ ,Q DGGLWLRQ WKHUH DUH PDQ\ WDVNV SHRSOH FDQ GR VXFFHVVIXOO\ ZKLFK DUH QRW SHUIRUPHG EHFDXVH DGHTXDWH LQFHQWLYHV DUH QRW SUHVHQW +RZHYHU JLYHQ DSSURSULDWH VNLOOV DQG DGHTXDWH LQFHQWLYHV HIILFDF\ H[SHFWDWLRQV DUH VHHQ E\ %DQGXUD DQG KLV FROOHDJXHV %DQGXUD t $GDPV %DQGXUD $GDPV t %H\HU f DV D PDMRU GHWHUn PLQDQW RI SHRSOHnV FKRLFH RI DFWLYLWLHV WKH HPSOR\PHQW RI WKHLU FRSLQJ HIIRUWV DQG WKHLU SHUVLVWHQFH DW WKH FRSLQJ HIIRUW ZKHQ GHDOLQJ ZLWK VWUHVVIXO VLWXDWLRQV ,Q %DQGXUDnV WKHRUHWLFDO DQDO\VLV H[SHFWDWLRQV RI VHOIHIILFDF\ DUH EDVHG RQ IRXU VRXUFHV RI LQIRUPDWLRQ f SHUIRUPDQFH DFFRPSOLVKPHQWV f YLFDULRXV H[SHULHQFH f YHUEDO SHUVXDVLRQ f HPRWLRQDO DURXVDO %DQGXUD f SUHGLFWHG WKDW WKHUDSHXWLF LQWHUYHQWLRQV HPSOR\LQJ HQDFWLYH LH SHUIRUPDQFH DFFRPSOLVKPHQWVf WHFKQLTXHV ZRXOG EH PRUH HIIHFWLYH LQ UDLVLQJ VXEMHFWVn VHOIHIILFDF\ H[SHFWDWLRQV WKDQ ZRXOG EH WKH LQWHUYHQWLRQV HPSOR\LQJ YLFDULRXV H[SHULHQFH PRGHOLQJf YHUEDO SHUVXDVLRQ VHOILQVWUXFWLRQ

PAGE 45

LQWHUSUHWLYH WUHDWPHQWVf RU HPRWLRQDO DURXVDO UHOD[DWLRQ GHVHQVLWL]DWLRQf %DQGXUD $GDPV DQG %H\HU f GHVLJQHG DQ H[SHULPHQW WR WHVW WKH K\SRWKHVL]HG UHODWLRQn VKLS EHWZHHQ VHOIHIILFDF\ DQG EHKDYLRUDO FKDQJH DQG WR FRPSDUH WKH HIIHFWLYHQHVV RI HQDFWLYH DQG YLFDULRXV H[SHUn LHQFH LQ HQKDQFLQJ HIILFDF\ H[SHFWDWLRQV 6HYHUHO\ VQDNH SKRELF VXEMHFWV ZHUH DVVLJQHG WR D SDUWLFLSDQW PRGHOLQJ FRQGLWLRQ D WUDGLWLRQDO PRGHOLQJ FRQGLWLRQ RU D QRn WUHDWPHQW FRQWURO JURXS 7KH PDJQLWXGH RI WKH VXEMHFWVn VHOIHIILFDF\ ZDV DVVHVVHG E\ JLYLQJ HDFK VXEMHFW D OLVW RI WKH WDVNV XVHG RQ D EHKDYLRUDO DYRLGDQFH SUHWHVW UHODWHG WR DSSURDFKLQJ D VQDNHf 7KH VXEMHFWV ZHUH DVNHG WR UDWH WKH WDVNV WKH\ H[SHFWHG WR SHUIRUP VXFFHVVIXOO\ 7KH VWUHQJWK RI VHOIHIILFDF\ ZDV DVVHVVHG E\ DVNLQJ HDFK VXEMHFW WR UDWH H[SHFWDQF\ RI VXFFHVV IRU HDFK WDVN LQ WHUPV RI D SRLQW VFDOH GLYLGHG LQWR LQWHUYDOV 7KH JHQHUDOLW\ RI VHOIHIILFDF\ ZDV DVVHVVHG E\ DVNLQJ VXEMHFWV WR UDWH WKH OHYHO PDJQLWXGHf DQG VWUHQJWK RI WKHLU H[SHFWDWLRQV ZLWK UHJDUG WR VXFFHVVIXOO\ FDUU\LQJ RXW WKH VDPH EHKDYLRUDO WDVN ZLWK DQ XQIDPLOLDU VQDNH 7KH UHVXOWV VXSSRUWHG WKH K\SRWKHVHV WKDW WKH SDUWLFLSDQW PRGHOLQJ WHFKQLTXH ZLWK LWV SHUIRUPDQFH PDVWHU\ FRPSRQHQW LQVWLOOHG WKH KLJKHVW VWURQJHVW DQG PRVW JHQHUDOL]HG HIILFDF\ H[SHFWDWLRQV IROORZHG E\ WKH YLFDULRXV WUHDWPHQW ,Q WKH IROORZXS DVVHVVPHQW LW ZDV IRXQG WKDW SDVW EHKDYLRU GLG QRW FRQVLVWHQWO\ RU DFFXUDWHO\ SUHGLFW IXWXUH EHKDYLRU

PAGE 46

EXW WKDW VHOIHIILFDF\ GLG DFFXUDWHO\ SUHGLFW IXWXUH EHKDYLRU $V D IXUWKHU H[WHQVLRQ RI WKH WKHRU\ %DQGXUD DQG $GDPV f H[DPLQHG WKH UROH RI UHGXFLQJ HPRWLRQDO DURXVDO WKURXJK GHVHQVLWL]DWLRQ RQ VHOIHIILFDF\ 7KH UHVXOWV VXSSRUWHG WKH K\SRWKHVHV WKDW HOLPLQDWLQJ HPRWLRQDO DURXVDO WKURXJK GHVHQVLWL]DWLRQ ZRXOG HQKDQFH VXEMHFWVn OHYHOV RI VHOIHIILFDF\ ,W ZDV DOVR IRXQG WKDW WKH OHYHO RI VHOIHIILFDF\ YDULHG IRU HDFK VXEMHFW DQG WKDW WKH KLJKHU DQG VWURQJHU WKH OHYHO RI VHOIHIILFDF\ LQGXFHG E\ WKH WUHDWPHQW WKH JUHDWHU WKH UHGXFWLRQ LQ DYRLGDQFH EHKDYLRU 7KH OHYHO RI VHOIHIILFDF\ FORVHO\ FRUUHVSRQGHG WR DSSURDFK EHKDYLRU DQG DFFXUDWHO\ SUHGLFWHG VXEVHTXHQW EHKDYLRU WR VLPLODU DQG GLVVLPLODU WDVNV ,Q VXPPDU\ %DQGXUD DQG $GDPV f DQG %DQGXUD $GDPV DQG %H\HU f GHPRQn VWUDWHG WKDW WKH VXFFHVVHV UHSRUWHG E\ WKH YDU\LQJ WHFKQLTXHV RI HQDFWLYH LH SHUIRUPDQFHf PDVWHU\ YLFDULRXV H[SHULHQFH DQG GHVHQVLWL]DWLRQ FDQ EH UHDGLO\ DQG DGHTXDWHO\ H[SODLQHG WKURXJK VHOIHIILFDF\ WKHRU\ 7KH\ KDYH DOVR GHPRQVWUDWHG WKDW WKH OHYHO RI HIILFDF\ HVWDEn OLVKHG E\ WKH LQWHUYHQWLRQ VLJQLILFDQWO\ DIIHFWV WKHUDSHXWLF VXFFHVV LQ RYHUFRPLQJ IHDUIXO DQG DYRLGDQW EHKDYLRU 7KH UHVXOWV VXJJHVW WKDW HIILFDF\ LV DFWXDOO\ D EHWWHU SUHGLFWRU RI IXWXUH EHKDYLRU WKDQ LV SDVW EHKDYLRU ,W LV PRUH FRQVLVWHQW DQG PRUH DFFXUDWH DV D SUHGLFWRU %DQGXUD DQG KLV FROOHDJXHV KDYH ZRUNHG H[FOXVLYHO\ ZLWK SKRELF FOLHQWV ERWK VQDNH SKRELFV DQG DJRUDSKRELFV %DQGXUD

PAGE 47

%DQGXUD $GDPV +DUG\ t +RZHOOV LQ SUHVVf ,W VHHPV OLNHO\ WKDW WKH OHYHO RI VHOIHIILFDF\ PLJKW SURYH WR EH DQ H[FHOOHQW SUHGLFWRU RI DFDGHPLF VXFFHVV DQG FDUHHU FKRLFH ([WHQGLQJ %DQGXUDnV UHVXOWV WR WKH DFKLHYHPHQW VLWXDWLRQ RQH ZRXOG DVVXPH WKDW VWXGHQWV ZLWK ORZ OHYHOV RI DFDGHPLF VHOIHIILFDF\ ZRXOG IHDU DQG DYRLG VXEMHFWV DQG H[WUDFXUULFXODU DFWLYLWLHV ZKLFK WKH\ EHOLHYHG OD\ RXWVLGH WKHLU UHSHUWRLUH RI FRSLQJ VNLOOV 7KHVH VWXGHQWV PLJKW UHVWULFW WKHPVHOYHV WR DFDGHPLF SURJUDPV SHUFHLYHG DV UHTXLULQJ OHVVHU DELOLWLHV 7KH\ PLJKW DOVR FKRRVH FDUHHUV ZKLFK DUH ORZHU LQ VRFLRHFRQRPLF FDUHHU VWDWXV 7KH\ ZRXOG DOVR EH SUHGLFWHG WR KDYH SRRUHU DFDGHPLF SHUn IRUPDQFH LQ VFKRRO /DORQGH f H[WHQGHG %DQGXUDnV f WKHRU\ E\ LQYHVWLJDWLQJ WKH UHODWLRQVKLS EHWZHHQ VHOIHIILFDF\ DQG DFKLHYHPHQWRULHQWHG EHKDYLRU 6KH K\SRWKHVL]HG WKDW ORZ H[SHFWDWLRQV RI VHOIHIILFDF\ ZRXOG SOD\ D PDMRU UROH LQ SRRU JUDGHV UHVWULFWHG IXWXUH DFDGHPLF H[SHFWDWLRQV DQG UHVWULFWHG FDUHHU FKRLFH ORZHU VRFLRHFRQRPLF FDUHHU FKRLFHf 7R WHVW WKH K\SRWKHVHV OLQNLQJ DFDGHPLF SHUIRUn PDQFH DV ZHOO DV DFDGHPLF DQG FDUHHU H[SHFWDWLRQV WR HIILFDF\ H[SHFWDWLRQV D PHDVXUH RI VHOIHIILFDF\ SHUWDLQLQJ WR DFDGHPLF DQG DFKLHYHPHQWRULHQWHG VLWXDWLRQV ZDV GHYHORSHG 7KLV PHDVXUH ZDV QHFHVVDU\ VLQFH WKH RQO\ SUHYLRXV PHDVXUH RI VHOIHIILFDF\ KDG EHHQ WKH UDWLQJ VFDOH XVHG LQ WKH %DQGXUD DQG $GDPV f DQG %DQGXUD $GDPV

PAGE 48

DQG %H\HU f UHVHDUFK 7KLV VLPSOLILHG PHWKRG RI DVNLQJ VXEMHFWV WR UDWH WKHLU HIILFDF\ H[SHFWDWLRQV IRU VXFFHVVIXOO\ FRPSOHWLQJ FHUWDLQ WDVNV ZDV HIIHFWLYH IRU VQDNH SKRELFV +RZHYHU LW ZDV QRW VR HDV\ WR GHWHUPLQH HIILFDF\ H[SHFWDWLRQV LQ DFKLHYHPHQW VLWXDWLRQV VLQFH VXEMHFWV GR QRW KDYH D SDUWLFXODU LGHQWLILDEOH VHW RI WDVNV ZKLFK LQGLFDWH WKH OHYHO RI WKHLU DYRLGDQFH EHKDYLRU /DORQGH f WKHUHIRUH LQYHVWLJDWHG VHOIHIILFDF\ DV UHODWHG WR DFKLHYHPHQW EHKDYLRU E\ WKUHH VWDJHV RI UHVHDUFK f WKH GHYHORSPHQW RI D SDSHU DQG SHQFLO VHOIUHSRUW PHDVXUH RI DFDGHPLF VHOIHIILFDF\ f WKH HVWDEOLVKPHQW RI GLVFULPLQDQW DQG FRQYHUJHQW YDOLGLWLHV f WKH XQGHUWDNLQJ RI D YDOLGDWLRQDO H[SHULPHQWDO VWXG\ ,Q RUGHU WR GHYHORS WKH LWHPV IRU WKH 0HDVXUH RI $FDGHPLF 6HOI(IILFDF\ 0$6(f %DQGXUDnV VHOIHIILFDF\ WKHRU\ DQG UHVHDUFK ZHUH FDUHIXOO\ H[DPLQHG DV EDFNJURXQG GDWD LQ GHILQLQJ WKH EHKDYLRUDO FKDUDFWHULVWLFV RI VWURQJO\ DQG ZHDNO\ HIILFDFLRXV SHRSOH 7R GHILQH DV FOHDUO\ DV SRVVLEOH WKH EHKDYLRUDO FRUUHODWHV RI VHOIHIILFDF\ WZHQW\ MXGJHV JUDGXDWH SV\FKRORJ\ VWXGHQWV DQG WHDFKLQJ VWDIIf ZHUH JLYHQ D ZULWWHQ GHILQLWLRQ RI HIILFDF\ DV ZHOO DV WKH WKHRUHWLFDO VWDWHPHQWV DQG HPSLULFDO UHVXOWV RI %DQGXUD DQG KLV FROOHDJXHV 7KH MXGJHV ZHUH DVNHG WR GHILQH WKH EHKDYLRUDO FKDUDFWHULVWLFV RI ZHDNO\ DQG VWURQJO\ HIILFDFLRXV SHRSOH 7KH WHVW LWHPV JHQHUDWHG E\ WKH MXGJHV ZHUH WR LQFOXGH D ZLGH VHOHFWLRQ RI H[WUDFXUULFXODU DQG

PAGE 49

FODVVURRP EHKDYLRU WR HQVXUH WKDW WKH JHQHUDOLW\ RI VWXGHQWVn DFDGHPLF VHOIHIILFDF\ ZDV DVVHVVHG WKH H[WHQW WR ZKLFK IHHOLQJV RI HIILFDF\ JHQHUDOL]H WR YDULRXV DFDGHPLF VLWXDWLRQVf %DQGXUD f VWDWHG WKDW HIILFDF\ H[SHFn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f 0$6(f KDG D +R\W f HVWLPDWH RI LQWHUQDO UHOLDELOLW\ VFDOH KRPRJHQHLW\f RI DQG D VWDQGDUG HUURU RI PHDVXUHPHQW RI 7KH WHVWUHWHVW UHOLDELOLW\ IRU WKH 0$6( ZDV RYHU D RQH PRQWK SHULRG &RQYHUJHQW DQG GLVFULPLQDQW YDOLGLW\ RI WKH 0$6( ZDV DVVHVVHG ZLWK D QXPEHU RI PHDVXUHV ,W ZDV SUHGLFWHG WKDW WKH 0$6( ZRXOG FRUUHODWH SRVLWLYHO\ ZLWK JUDGH SRLQW DYHUDJH VFKRODVWLF DSWLWXGH 2WLV/HQQRQ 0HQWDO $ELOLW\ 7HVWf ZLOOLQJQHVV WR DFFHSW UHVSRQVLELOLW\ IRU SRVLWLYH

PAGE 50

LQWHOOHFWXDO DFKLHYHPHQW ,QWHOOHFWXDO $FKLHYHPHQW 5HVSRQ VLELOLW\ 4XHVWLRQQDLUHf &UDQGDOO .DWNRYVN\ t &UDQGDOO f FDUHHU DVSLUDWLRQV DQG H[SHFWDWLRQV FRGHG IRU VRFLRHFRQRPLF VWDWXVf SRVW KLJK VFKRRO FDUHHU SODQV DQG VHOIFRQFHSW 3LHUV+DUULV &KLOGUHQnV 6HOI&RQFHSW 6FDOHf 3LHUV f ,W ZDV DOVR K\SRWKHVL]HG WKDW VHOIHIILFDF\ ZRXOG FRUUHODWH QHJDWLYHO\ ZLWK VRFLDO DQ[LHW\ 6RFLDO $YRLGDQFH DQG 'LVWUHVV 6FDOH :DWVRQ t )ULHQG f DQG WHVW DQ[LHW\ 7HVW $Q[LHW\ 6FDOH 6DUDVRQ f )LQDOO\ DQ H[SHULPHQW ZDV GHVLJQHG WR EHKDYLRUDOO\ YDOLGDWH WKH 0$6( E\ WHVWLQJ K\SRWKHVHV GHULYHG IURP VHOI HIILFDF\ WKHRU\ $Q DQDJUDPV WDVN ZDV GHVLJQHG WR WHVW WKH K\SRWKHVHV WKDW f KLJKO\HIILFDFLRXV KLJK VFKRRO VWXGHQWV ZRXOG SHUVLVW ORQJHU DW DQ DQDJUDPV WDVN WKDQ ZRXOG OHVV HIILFDFLRXV VWXGHQWV f KLJKO\HIILFDFLRXV VWXGHQWV ZRXOG SHUVLVW ORQJHU LQ WKH IDFH RI IDLOXUH WKDQ ZRXOG OHVV HIILFDFLRXV VWXGHQWV f IRUFHG IDLOXUH H[SHULHQFHG HDUOLHU LQ WKH DQDJUDPV WDVN ZRXOG KDYH D JUHDWHU HIIHFW RQ OHVV HIILFDFLRXV VWXGHQWV WKDQ ZRXOG IDLOXUH H[SHULHQFHG ODWHU 7KHVH K\SRWKHVHV ZHUH EDVHG RQ K\SRWKHVHV IURP VHOIHIILFDF\ WKHRU\ ZKLFK VWDWH WKDW KLJKO\ VHOIHIILFDFLRXV SHRSOH ZLOO SHUVLVW ORQJHU DW RYHUFRPLQJ GLIILFXOWLHV DQG DUH OHVV DIIHFWHG E\ IDLOXUH ZKHUHDV OHVV VHOIHIILFDFLRXV SHUVRQV DUH PRUH DIIHFWHG E\ HDUO\ H[SHULHQFHV RI IDLOXUH 3HUVLVn WHQFH DW WKH DQDJUDPV WDVN ZDV GHILQHG LQ WHUPV RI QLQH YDULDEOHV WLPH VSHQW RQ VROYHG DQDJUDPV VNLSSHG DQDJUDPV

PAGE 51

WRWDO DQDJUDPV QXPEHU RI DQDJUDPV UHWXUQHG WR SURSRUWLRQ RI DQDJUDPV VNLSSHG ZRUNLQJ XQWLO IRUFHG WR VWRS ZLOOLQJn QHVV WR UHWXUQ WR VNLSSHG DQDJUDPV DQG ZLOOLQJQHVV WR NHHS ZRUNLQJ LI IRUFHG WR ILQLVK SUHPDWXUHO\f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f YLHZV DQ[LHW\ DQG HIILFDF\ DV DQWLWKHWLFDO 2QH FDQQRW EH DQ[LRXV DERXW RQHnV DELOLW\ DQG HIILFDFLRXV DW WKH VDPH WLPH 7KH VWURQJHU WKH HIILFDF\ H[SHFWDWLRQV WKH JUHDWHU WKH UHGXFWLRQ LQ DQ[LHW\ DQG DYRLGDQFH EHKDYLRU 7KH K\SRWKHVHV FRQFHUQLQJ WKH SURSRVHG UHODWLRQVKLS EHWZHHQ DFDGHPLF VHOIHIILFDF\ DQG FDUHHU H[SHFWDWLRQV DQG SRVWVHFRQGDU\ VFKRRO SODQV ZHUH VXSSRUWHG +LJKHU HIILFDF\ VFRUHV ZHUH DVVRFLDWHG ZLWK KLJKHU FDUHHU H[SHFWDn WLRQV DQG SODQV WR DWWHQG D GHJUHHJUDQWLQJ LQVWLWXWLRQ

PAGE 52

,Q DGGLWLRQ ZLWK WKH HIIHFWV RI YHUEDO LQWHOOLJHQFH UHPRYHG DFDGHPLF HIILFDF\ VFRUHV UHPDLQHG ORZ EXW SRVLWLYH SUHGLFWRUV RI JUDGH SRLQW DYHUDJH DQG SRVWn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n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

PAGE 53

VHOIHIILFDF\ WDVN GLIILFXOW\ DQG SHUVLVWHQFH $ QXPEHU RI H[SODQDWLRQV ZHUH SURYLGHG SRVWKRF IRU WKHVH UHVXOWV 7KH PRVW FRJHQW H[SODQDWLRQ DSSHDUV WR EH WKDW WKH GXUDWLRQ RI WLPH SURYLGHG IRU WKH WDVN ZDV QRW VXIILFLHQW WR DGHn TXDWHO\ PHDVXUH SHUVLVWHQFH $ ODUJH SURSRUWLRQ bf RI WKH VXEMHFWV ZHUH IRUFHG WR ILQLVK SUHPDWXUHO\ DW WKH PLQXWH WLPH OLPLW 7KLV VXJJHVWV WKDW PLQXWHV PD\ QRW KDYH EHHQ VXIILFLHQW WLPH WR GLVFULPLQDWH EHWZHHQ KLJKDQG ORZHIILFDF\ VWXGHQWV LQ WHUPV RI WDVN SHUVLVWHQFH +RZHYHU WKH VWXG\ GLG ILQG WKDW KLJKO\HIILFDFLRXV VXEMHFWV SHUIRUPHG EHWWHU RQ WKH KDUG DQDJUDPV WKDQ GLG WKH ORZ HIILFDFLRXV VXEMHFWV DQG H[KLELWHG PRUH ZLOOLQJQHVV WR UHWXUQ WR XQILQLVKHG VNLSSHGf DQDJUDPV 2YHUDOO WKH UHVXOWV RI WKH VWXG\ SURYLGH WHQWDWLYH VXSSRUW IRU DVSHFWV RI %DQGXUDnV f WKHRU\ RI VHOI HIILFDF\ &RUUHODWLRQV EHWZHHQ 0$6( VFRUHV RWKHU VHOI UHSRUW LQGLFHV DQG WHDFKHU UDWLQJV LQGLFDWHG WKDW DFDGHPLFDOO\ KLJKO\HIILFDFLRXV VWXGHQWV DUH PRUH ZLOOLQJ WR DFFHSW UHVSRQVLELOLW\ IRU WKHLU DFDGHPLF VXFFHVVHV WKDQ IDLOXUHV %DQGXUD f KDG VWDWHG WKDW VHOIHIILFDF\ ZDV HQKDQFHG RQO\ ZKHQ VXEMHFWV YLHZHG WKHLU VXFFHVVHV DV UHVXOWLQJ IURP WKHLU RZQ DELOLW\ UDWKHU WKDQ DV D IXQFWLRQ RI FRQWH[WXDO IDFWRUV VXFK DV FKDQFH OXFN RU HDVH RI WKH WDVN )ROORZLQJ WKLV UHDVRQLQJ LW ZDV DOVR K\SRWKHVL]HG WKDW WKH 0$6( ZRXOG FRUUHODWH SRVLWLYHO\ ZLWK D VHOIFRQFHSW VFDOH DQG WKLV K\SRWKHVLV ZDV YDOLGDWHG

PAGE 54

7KH 0$6( DOVR FRUUHODWHG SRVLWLYHO\ ZLWK JUDGH SRLQW DYHUDJH IRUPDO SRVWVHFRQGDU\ HGXFDWLRQ SODQV SODQV WR HQUROO LQ D GHJUHHJUDQWLQJ LQVWLWXWLRQ DQG FDUHHU H[SHFWDWLRQV ZLWK KLJKHU VRFLRHFRQRPLF LQGLFHV DFFRUGLQJ WR D UDQNLQJ RI FDUHHU SRVVLELOLWLHV SUHVHQWHG WR WKH VXEMHFWVf )LQDOO\ ZLWK WKH YHUEDO LQWHOOLJHQFH IDFWRU UHPRYHG WKH 0$6( FRUUHODWHG DW D ORZ EXW SRVLWLYH OHYHO ZLWK JUDGH SRLQW DYHUDJH DQG SRVW VHFRQGDU\ VFKRRO SODQV 7KHVH UHVXOWV VXJJHVW WKDW WKH PDVWHU\ H[SHFWDQF\ YDULDEOH RI VHOIHIILFDF\ LV UHOHYDQW WR SUHGLFWLRQ RI DFDGHPLF VXFFHVV DQG SODQQLQJ 7KH UHVXOWV DOVR UHODWH WR WKH ZRUN RI 0F&OHOODQG f 7DLW f DQG 7D\ORU f ZKR IRXQG WKDW QRQLQWHOOHFn WXDO IDFWRUV SOD\ D UROH LQ DFDGHPLF DQG YRFDWLRQDO VXFFHVV 6LQFH DFDGHPLF VHOIHIILFDF\ LV DQ H[SHFWDQF\ YDULDEOH LW LV LPSRUWDQW WR GLVWLQJXLVK LW IURP RWKHU H[SHFWDQF\ YDULDEOHV 5RWWHUnV f ORFXV RI FRQWURO WKHRU\ DQG 0DLHU DQG 6HOLJPDQnV f OHDUQHG KHOSOHVVQHVV WKHRU\ PD\ EH VHHQ DV VRPHZKDW DQDORJRXV WR %DQGXUDnV f VHOI HIILFDF\ WKHRU\ DQG WKH 0HDVXUH RI $FDGHPLF 6HOI(IILFDF\ /RFXV RI FRQWURO DQG OHDUQHG KHOSOHVVQHVV KRZHYHU ERWK PHDVXUH RXWFRPH H[SHFWDQFLHV UDWKHU WKDQ PDVWHU\ H[SHFn WDQFLHV 5RWWHUnV WKHRU\ LV FRQFHUQHG ZLWK D SHUVRQnV EHOLHI WKDW RXWFRPHV DUH GHWHUPLQHG HLWKHU E\ RQHnV RZQ DFWLRQV LQWHUQDO FRQWUROf RU E\ IRUFHV RXWVLGH RQHnV FRQWURO H[WHUQDO FRQWUROf 6HOIHIILFDF\ LV GLIIHUHQW IURP ORFXV RI FRQWURO EHFDXVH SHRSOH PD\ EHOLHYH WKDW

PAGE 55

RXWFRPHV DUH LQWHUQDOO\ FRQWUROOHG EXW PD\ IHHO D VHQVH RI ORZ HIILFDF\ LI WKH\ EHOLHYH WKDW WKH\ ODFN WKH VNLOOV UHTXLUHG WR SHUIRUP WKH DFWLYLW\ 0DLHU DQG 6HOLJPDQnV f WKHRU\ RI OHDUQHG KHOSOHVVQHVV DQG WKH +RSHOHVVQHVV 6FDOH %HFN HW DO f UHVHPEOH WKH FRQFHSW RI VHOIHIILFDF\ \HW DUH GLIIHUn HQW IURP VHOIHIILFDF\ DQG WKH 0$6( /HDUQHG KHOSOHVVQHVV LV DOVR DQ RXWFRPH H[SHFWDQF\ YDULDEOH ,W VWDWHV WKDW ZKHQ SHRSOH FDQQRW DIIHFW ZKDW KDSSHQV WR WKHP WKH\ EHFRPH DSDWKHWLF RU OHDUQHG KHOSOHVV DQG VWRS WU\LQJ WR FRSH /DORQGH f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f 7KH FRQFHSW RI DFKLHYHPHQW PRWLYDWLRQ GHYHORSHG E\ 0F&OHOODQG HW DO f FDQ DOVR EH FRQVLGHUHG WR EH GLIIHUHQW IURP WKH FRQFHSW RI DFDGHPLF VHOIHIILFDF\ ,Q WKH GHILQLWLRQ RI KLJK DFKLHYHPHQW PRWLYDWLRQ WKH NH\

PAGE 56

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n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t /RZHOO f 7KH LQGLYLGXDO UHVSRQGV WR SLFWXUHV ZKLFK VXJJHVW DFKLHYHPHQW

PAGE 57

WKHPHV 7KLV LV D UDGLFDOO\ GLIIHUHQW PHWKRGRORJ\ IURP WKDW SURSRVHG E\ /DORQGH f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nV VWXG\ f KDV EHJXQ WR GHOLQHDWH WKH UROH RI DFDGHPLF VHOIHIILFDF\ DV D UHOHYDQW YDULDEOH LQ WKH SUHGLFWLRQ RI DFDGHPLF VXFFHVV DQG FDUHHU SODQQLQJ 7R WKH H[WHQW WKDW VWXGHQWV EHOLHYH WKDW WKH\ KDYH WKH FRSLQJ VNLOOV WR PDVWHU DFDGHPLF VLWXDWLRQV WKH\ WHQG WR KDYH KLJKHU JUDGH SRLQW DYHUDJHV WDNH PRUH UHVSRQVLELOLW\ IRU

PAGE 58

WKHLU DFDGHPLF VXFFHVVHV DQG KDYH FDUHHU H[SHFWDWLRQV DQG IXWXUH SODQV DVVRFLDWHG ZLWK JUHDWHU VRFLRHFRQRPLF DQG DFDGHPLF DFKLHYHPHQW ,Q KHU VWXG\ /DORQGH DOVR IRXQG WKDW WHVW DQ[LHW\ DV PHDVXUHG RQ WKH 7HVW $Q[LHW\ 6FDOH FRUUHODWHG ZLWK DFDGHPLF VHOIHIILFDF\ 7KLV FRUUHODWLRQ LQGLFDWHG VRPH RYHUODSSLQJ YDULDQFH DFURVV WKHVH WZR YDULDEOHV EXW QRW LGHQWLW\ 7KLV GLVVHUWDWLRQ VWXG\ H[WHQGV WKH ZRUN RI /DORQGH LQ H[DPLQLQJ WKH UHODn WLRQVKLS RI WHVW DQ[LHW\ DQG DFDGHPLF VHOIHIILFDF\ DV SUHGLFWRUV RI WDVN SHUIRUPDQFH XQGHU YDULRXV FRQGLWLRQV ,W LV SURSRVHG WKDW VHOIHIILFDF\ LV D UHOHYDQW FRJQLWLYH FRQFRPLWDQW RI WHVW DQ[LHW\ ,W DSSHDUV WR UHIOHFW VRPH RI WKH FRUH SUREOHPV RI WKH WHVWDQ[LRXV LQGLYLGXDO LQ WKH DFDGHPLFDOO\HYDOXDWLYH VLWXDWLRQ $V :LQH f DQG RWKHUV 6DUDVRQ 0HLFKHQEDXP f KDYH K\SRWKHVL]HG WKH WHVWDQ[LRXV SHUVRQ GHPRQVWUDWHV D SHUIRUPDQFH GHFUHPHQW LQ WKH HYDOXDWLYH VLWXDWLRQ EHFDXVH KH IDLOV WR IRFXV RQ WKH UHOHYDQW SDUWV RI WKH WDVN ,QVWHDG KH UXPLQDWHV RQ KLV SHUFHLYHG LQFRPSHWHQFH LQ WKH WHVW VLWXDWLRQ 6LQFH WKH FRQFHSW RI DFDGHPLF VHOIHIILFDF\ GHDOV ZLWK D SHUVRQnV EHOLHI WKDW KH KDV RU GRHV QRW KDYHf WKH FRSLQJ VNLOOV WR PDVWHU DFDGHPLF VLWXDWLRQV LW DSSHDUV WKDW DFDGHPLF VHOIHIILFDF\ LV FHUWDLQO\ UHODWHG DW D FRQFHSWXDO OHYHO WR WKH FRJQLWLYH DVSHFWV RI WHVW DQ[LHW\ GLVFXVVHG HDUOLHU LQ WKLV UHYLHZ $FDGHPLF VHOIHIILFDF\

PAGE 59

PD\ WKXV SURYLGH VRPH IXUWKHU LQIRUPDWLRQ DERXW WKH PRGXV RSHUDQGL RI WKH WHVWDQ[LRXV LQGLYLGXDO 7KLV GLVVHUWDWLRQ VWXG\ LV DQ DWWHPSW WR GHOLQHDWH WKH UROH RI DFDGHPLF VHOIHIILFDF\ DV LW UHODWHV WR WKH PXOWLGLPHQVLRQDO FRQVWUXFW RI WHVW DQ[LHW\ 7KH VWXG\ GHWHUPLQHV YLD D PXOWLYDULDWH SUHGLFWRU PRGHO WKH H[WHQW WR ZKLFK D PHDVXUH RI DFDGHPLF VHOIHIILFDF\ RSHUn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

PAGE 60

KDV UHODWHG WR SHUIRUPDQFH GLIIHUHQFHV LQ KLJK DQG ORZ WHVWDQ[LRXV VXEMHFWV 0DQ\ RI WKH VWXGLHV PDQLSXODWHG LQVWUXFWLRQDO FRQGLWLRQV ZKLOH KROGLQJ WDVN GHPDQGV FRQVWDQW 6DUDVRQ f 7KHVH VWXGLHV KDYH JHQHUDOO\ LQYROYHG FRPSDULVRQV RI Df HJRLQYROYLQJ LQVWUXFWLRQVf§ WKHVH LQVWUXFWLRQV LQIRUP WKH VXEMHFW WKDW KLV WDVN SHUIRUPDQFH OHYHO LV UHIOHFWLYH RI KLV LQWHOOHFWXDO FDSDELOLWLHV Ef PLQLPDO WDVN LQVWUXFWLRQVf§LQVWUXFWLRQV ZKLFK VSHFLILFDOO\ DGGUHVV WKH FRJQLWLYH WDVN HJ GLUHFWLRQV IRU VROYLQJ DQ DQDJUDPV WDVNf DQG Ff UHDVVXU DQHH LQVWUXFWLRQVf§LQVWUXFWLRQV ZKLFK UHDVVXUH WKH VXEMHFW WKDW KLV WDVN SHUIRUPDQFH ZLOO QRW EH XVHG WR HYDOXDWH KLP )UHTXHQWO\ XQGHU WKLV ODVW FRQGLWLRQ WKH VXEMHFW LV WROG WKDW KLV SHUIRUPDQFH LV DQRQ\PRXV DQG RQO\ LPSRUWDQW VR WKDW WKH H[DPLQHU FDQ H[DPLQH DYHUDJH JURXS SHUIRUPDQFH GDWD :LQH 6DUDVRQ 6DUDVRQ t *DQ]HU 5XVVHOO t 6DUDVRQ *DQ]HU f +LJKO\ WHVWDQ[LRXV VXEMHFWV WHQG WR SHUIRUP PRUH SRRUO\ IROORZLQJ KLJKO\ HYDOXDWLYH HJRLQYROYLQJ LQVWUXFWLRQV WKDQ IROORZLQJ UHDVVXULQJ DQRQ\PRXV LQVWUXFWLRQV 7KH UHYHUVH WHQGV WR EH WUXH IRU ORZ WHVWDQ[LRXV VXEMHFWV )ROORZLQJ KLJKO\ HYDOXDWLYH LQVWUXFWLRQV ORZ WHVWDQ[LRXV VXEMHFWV SHUIRUP EHWWHU WKDQ KLJK WHVWDQ[LRXV VXEMHFWV +RZHYHU IROORZLQJ WKH UHDVVXULQJ DQRQ\PRXV LQVWUXFWLRQV KLJK WHVWDQ[LRXV VXEMHFWV SHUIRUP EHWWHU WKDQ ORZ WHVWDQ[LRXV VXEMHFWV 8QGHU WKH WKLUG W\SH RI FRQGLWLRQ WKH QHXWUDO

PAGE 61

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f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

PAGE 62

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n VLRQ WKLV VWXG\ VXJJHVWV WKDW WDVN GLIILFXOW\ PLJKW GLIIHUn HQWLDOO\ DIIHFW WKH SHUIRUPDQFH RI ORZ DQG KLJK VHOI HIILFDFLRXV VWXGHQWV EXW D IRUFHG IDLOXUH H[SHULHQFH SURGXFHG E\ WKH SUHVHQWDWLRQ RI KDUG DQDJUDPV HDUO\ LQ WKH WDVN ZLOO QRW GLIIHUHQWLDOO\ DIIHFW WKH SHUIRUPDQFH RI WKH WZR JURXSV RI VXEMHFWV 7KH VWUHQJWK RI WKH UHVXOWV RI WKLV VWXG\ DUH VRPHZKDW GLPLQLVKHG E\ WKH IDFW WKDW WKH HDV\ DQG PHGLXP DQDJUDP VHWV ZHUH H[WUHPHO\ HDV\ IRU ERWK JURXSV WKXV UHGXFLQJ WKH GLVFULPLQDWLYH SRZHU RI WKH DQDJUDP YDULDEOH ,Q WHUPV RI WKH SURSRUWLRQV RI DQDJUDPV VROYHG WR b RI DOO RI WKH HDV\ DQG PHGLXP DQDJUDPV ZHUH VROYHG E\ ERWK JURXSV 8QIRUWXQDWHO\ WKLV VWXG\ LV QRW FRPSDUDEOH WR WKH WHVW DQ[LHW\ VWXGLHV ZKLFK PDQLSXODWHG WDVN LQVWUXFWLRQV RI DQ HYDOXDWLYH YV QHXWUDO DQGRU UHDVVXULQJ QDWXUH &RQVHTXHQWO\ K\SRWKHVHV FDQ QRW EH GHULYHG IURP WKH

PAGE 63

/DORQGH VWXG\ UHJDUGLQJ WKH GLIIHUHQWLDO SHUIRUPDQFH RI KLJKO\HIILFDFLRXV DQG OHVVHIILFDFLRXV VWXGHQWV XQGHU YDU\LQJ HYDOXDWLYH FRQGLWLRQV 7KH GDWD GR VXJJHVW KRZHYHU WKDW WKHUH PD\ EH RYHUDOO FRJQLWLYH SHUIRUPDQFH GLIIHUHQFHV EHWZHHQ ORZDQG KLJKHIILFDF\ VWXGHQWV EXW VSHFLILF IDLOXUH H[SHULHQFHV ZLOO QRW GLIIHUHQWLDOO\ DIIHFW WKH SHUIRUPDQFH RI WKH WZR JURXSV +\SRWKHVHV UHJDUGLQJ WKH GLIIHUHQWLDO SHUIRUPDQFH RI ORZ YV KLJKHIILFDFLRXV VWXGHQWV FDQ EH JHQHUDWHG IURP %DQGXUDnV WKHRU\ )URP %DQGXUDnV WKHRU\ LW LV H[SHFWHG WKDW WKHUH ZRXOG EH DQ RYHUDOO SHUIRUPDQFH GLIIHUHQFH RQ DQ DFDGHPLFDOO\UHODWHG FRJQLWLYH WDVN KRZHYHU WKH PDQLSXODWLRQ RI HYDOXDWLYH YV QRQHYDOXDWLYH LQVWUXFWLRQV ZRXOG QRW SURGXFH GLIIHUHQWLDO SHUIRUPDQFH UHVXOWV VXFK WKDW WKH HIIHFWV RI WKH HIILFDF\ YDULDEOH ZRXOG RQO\ EH VLJQLILFDQW XQGHU RQH RU WKH RWKHU FRQGLWLRQ 7KHUH DUH VHYHUDO DVSHFWV RI %DQGXUDnV WKHRU\ ZKLFK ZRXOG VXJJHVW WKDW DQ H[SHULPHQWDO PDQLSXODWLRQ ZRXOG QRW SURGXFH GLIIHUHQWLDO UHVXOWV )LUVW RI DOO DFFRUGLQJ WR %DQGXUDnV WKHRU\ H[SHFWDWLRQV RI SHUVRQDO HIILFDF\ DUH GHULYHG IURP IRXU VRXUFHV RI LQIRUPDWLRQ SHUIRUPDQFH DFFRPSOLVKPHQWV YLFDULRXV H[SHULHQFH YHUEDO SHUVXDVLRQ DQG SK\VLRORJLFDO GDWD %DQGXUD f 7KH PRUH GHSHQGDEOH WKH H[SHULHQWLDO VRXUFH RI HIILFDF\ LQIRUPDWLRQ WKH JUHDWHU WKH OLNHOLKRRG RI FKDQJH LQ SHUFHLYHG VHOIHIILFDF\ 7KH WDVN LQVWUXFWLRQV SURYLGHG E\ WKH HYDOXDWLYH FRQGLWLRQ

PAGE 64

FRXOG EH FRQVLGHUHG WR IDOO LQWR WKH FDWHJRU\ RI YHUEDO SHUVXDVLRQ 7KH VXEMHFW LV OHG YLD VXJJHVWLRQ WR EHOLHYH WKDW WKH SUREOHPVROYLQJ WDVN IRU H[DPSOH WKH DQDJUDPV WDVN XVHG LQ WKH WHVW DQ[LHW\ VWXGLHV LV D PHDVXUH RI LQWHOOLJHQFH RU DELOLW\ 7KXV WKH VXEMHFW SUHVXPDEO\ EHFRPHV HJRLQYROYHG $FFRUGLQJ WR %DQGXUD f YHUEDO SHUVXDVLRQ LV D UHODWLYHO\ SRRU VRXUFH IRU SURYLGLQJ FKDQJH LQ SHUFHLYHG VHOIHIILFDF\ (IILFDF\ H[SHFWDWLRQV LQGXFHG E\ YHUEDO SHUVXDVLRQ DUH OLNHO\ WR EH ZHDNHU WKDQ WKRVH DULVLQJ IURP RQHnV RZQ DFFRPSOLVKPHQWV 7KLV LV GXH WR WKH IDFW WKDW YHUEDO SHUVXDVLRQV GR QRW SURYLGH DQ DXWKHQWLF H[SHULHQWLDO EDVH IRU WKH EHOLHIV %DQGXUD S f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n WHQFH ZLOO EH VLJQLILFDQWO\ UHODWHG WR SHUIRUPDQFH %DQGXUD VWDWHV LQ DGGLWLRQ WKDW WKH HIIHFWV RI IDLOXUH RQ SHUVRQDO HIILFDF\ GHSHQG SDUWO\ RQ WKH WLPLQJ DQG WKH

PAGE 65

WRWDO SDWWHUQ RI H[SHULHQFHV LQ ZKLFK WKH IDLOXUHV RFFXU ,Q WKH VWURQJO\ HIILFDFLRXV SHUVRQ DIWHU HIILFDF\ H[SHFn WDWLRQV DUH GHYHORSHG WKURXJK UHSHDWHG VXFFHVV WKH QHJDWLYH LPSDFW RI RFFDVLRQDO IDLOXUHV LV OLNHO\ WR EH UHGXFHG %DQGXUD f %\ WKH VDPH WRNHQ UHDVVXULQJ LQVWUXFWLRQV DUH XQOLNHO\ WR LQGXFH IHHOLQJV RI FRSLQJ PDVWHU\ LQ D SHUVRQ ZLWK SRRU VHOIHIILFDF\ RI ORQJ VWDQGLQJ LQ DFDGHPLF VLWXDWLRQV 7KXV IRU WKH YDULDEOH RI DFDGHPLF VHOIHIILFDF\ WDVN LQVWUXFWLRQV DUH QRW OLNHO\ WR VLJQLILn FDQWO\ UHGXFH RU LQFUHDVH WKH UHODWLRQVKLS EHWZHHQ VHOI HIILFDF\ DQG FRJQLWLYH WDVN SHUIRUPDQFH ,Q WKH SUHVHQW VWXG\ WHVW DQ[LHW\ DQG VHOIHIILFDF\ DUH ERWK XVHG WR SUHGLFW SHUIRUPDQFH XQGHU HYDOXDWLYH QHXWUDO DQG UHDVVXULQJ WDVN LQVWUXFWLRQV $FFRUGLQJ WR WKH UHVHDUFK ILQGLQJV GLVFXVVHG DERYH LQ UHJDUG WR WHVW DQ[LHW\ DQG WKH WKHRUHWLFDO LVVXHV SUHVHQWHG LQ WHUPV RI VHOIHIILFDF\ WHVW DQ[LHW\ ZLOO LQWHUDFW ZLWK WKH LQVWUXFn WLRQDO FRQGLWLRQ ZKHQ SUHGLFWLQJ SHUIRUPDQFH RQ D FRJQLWLYH WDVN ,W LV H[SHFWHG WKDW VHOIHIILFDF\ ZLOO QRW LQWHUDFW ZLWK WKH LQVWUXFWLRQDO FRQGLWLRQ ZKHQ SUHGLFWLQJ SHUIRUPDQFH RQ D FRJQLWLYH WDVN ,W LV H[SHFWHG WKDW WHVW DQ[LHW\ ZLOO DFFRXQW IRU D VLJQLILFDQW DPRXQW RI WKH YDULDQFH LQ SHUIRUPDQFH XQGHU HYDOXDWLYH DQG UHDVVXULQJ WDVN LQVWUXFWLRQV DQG LW ZLOO DFFRXQW IRU OHVV RI WKH YDULDQFH XQGHU QHXWUDO LQVWUXFWLRQV ,W LV H[SHFWHG WKDW VHOIHIILFDF\ ZLOO VLJQLILFDQWO\ DFFRXQW IRU SHUIRUPDQFH

PAGE 66

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t 6WRRSV 6DUDVRQ 5XVVHOO t 6DUDVRQ f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

PAGE 67

ILQDO JUDGH LQ *HQHUDO 3V\FKRORJ\ ZDV UHDGLO\ DYDLODEOH IRU WKH XVH RI WKH LQYHVWLJDWRU 7KH K\SRWKHVHV SURSRVHG IRU WKLV VWXG\ JLYHQ WKH H[LVWLQJ OLWHUDWXUH DQG WKHRU\ ZHUH WKH IROORZLQJ f $FDGHPLF VHOIHIILFDF\ ZRXOG EH PRGHUDWHO\ QHJDWLYHO\ FRUUHODWHG ZLWK WHVW DQ[LHW\ f 7HVW DQ[LHW\ ZRXOG SUHGLFW DQDJUDP SHUIRUPDQFH XQGHU HYDOXDWLYH DQG UHDVVXULQJ WDVN LQVWUXFWLRQV EXW QRW DV HIIHFWLYHO\ XQGHU QHXWUDO LQVWUXFWLRQV LQ DFFRUGDQFH ZLWK WKH WHVW DQ[LHW\ OLWHUDWXUH f 6HOIHIILFDF\ ZRXOG SUHGLFW DQDJUDP SHUIRUPDQFH XQGHU HYDOXDWLYH QHXWUDO DQG UHDVVXULQJ WDVN LQVWUXFWLRQV LQ DFFRUGDQFH ZLWK %DQGXUDnV WKHRU\ f &RJQLWLYH DELOLW\ OHYHO DV PHDVXUHG E\ DQ DSWLWXGH WHVW ZRXOG DFFRXQW IRU D VLJQLILFDQW DPRXQW RI YDULDQFH LQ SUHGLFWLQJ WR FRJQLWLYH WDVN SHUIRUPDQFH SURYLGHG WKHUH ZHUH VXIILFLHQW YDULDWLRQ LQ DSWLWXGH VFRUHV ,I WKH UDQJH RI DSWLWXGH VFRUHV ZDV UHGXFHG WKHQ WKH SUHGLFWLYH SRZHU RI FRJQLWLYH DELOLW\ ZRXOG EH UHGXFHG DV ZHOO f 7HVW DQ[LHW\ VHOIHIILFDF\ DQG DSWLWXGH VFRUHV ZRXOG DOO EH SUHGLFWLYH RI SHUIRUPDQFH LQ *HQHUDO 3V\FKRORJ\ DV PHDVXUHG E\ WKH VXEMHFWnV ILQDO JUDGH LQ WKH FRXUVH

PAGE 68

&+$37(5 7:2 0(7+2' 6XEMHFWV 7KH VXEMHFWV ZHUH XQGHUJUDGXDWH VWXGHQWV HQUROOHG LQ WKH LQWURGXFWRU\ SV\FKRORJ\ FRXUVH VHFWLRQV 36< f DW WKH 8QLYHUVLW\ RI )ORULGD GXULQJ WKH )DOO 6HPHVWHU 7KH\ ZHUH UHTXLUHG WR SDUWLFLSDWH LQ H[SHULPHQWDO VWXGLHV LQ RUGHU WR FRPSOHWH WKHLU FRXUVH UHTXLUHPHQW 7KH VXEMHFWV VLJQHG XS IRU WZR H[SHULPHQWDO VHVVLRQV ZKLFK ZHUH KHOG RQH ZHHN DSDUW 7KH\ ZHUH UHTXLUHG WR FRPSOHWH ERWK VHVVLRQV LQ RUGHU WR UHFHLYH FUHGLW IRU SDUWLFLSDWLRQ LQ WKH VWXG\ 7KRVH VXEMHFWV ZKR FRPSOHWHG ERWK VHVVLRQV ZHUH LQFOXGHG LQ WKH VWXG\ Q f 7KH LQYHVWLJDWRU ZDV DVVLVWHG LQ WKH GDWD FROOHFWLRQ VHVVLRQV E\ IRXU XQGHUJUDGXDWH SV\FKRORJ\ PDMRUV ZKR DGPLQLVWHUHG TXHVWLRQQDLUHV GXULQJ WKH ILUVW VHVVLRQV DQG FRQGXFWHG WKH H[SHULPHQWDO SDUW RI WKH VWXG\ GXULQJ WKH VHFRQG VHVVLRQV 7KH\ ZHUH EOLQG WR WKH H[SHULPHQWDO K\SRWKHVHV

PAGE 69

'XULQJ WKH LQLWLDO WHVWLQJ VHVVLRQ DOO VXEMHFWV UHFHLYHG WKH 7HVW $Q[LHW\ 6FDOH DQG WKH 0HDVXUH RI $FDGHPLF 6HOI(IILFDF\ 7KH 7HVW $Q[LHW\ 6FDOH GHYHORSHG E\ 6DUDVRQ f LV D LWHP VFDOH IUHTXHQWO\ XVHG WR DVVHVV WHVW DQ[LHW\ ,W LV D WUXHIDOVH WHVW DQG LW \LHOGV RQH WRWDO VFRUH 7HVWUHWHVW UHOLDELOLWLHV RYHU KDYH EHHQ UHSRUWHG IRU LQWHUYDOV RI VHYHUDO ZHHNV 6DUDVRQ f RQ D VDPSOH RI PDOHV DQG IHPDOH XQGHUJUDGXDWHV :DJDPDQ &RUPLHU DQG &RUPLHU f UHSRUWHG D WHVWUHWHVW UHOLDELOLW\ FRHIILFLHQW RI 7KLV WHVW KDV EHHQ XWLOL]HG WR LGHQWLI\ KLJK DQG ORZ WHVWDQ[LRXV LQGLYLGXDOV LQ PRVW RI WKH UHVHDUFK FRQGXFWHG E\ 6DUDVRQ DQG KLV FROOHDJXHV 6DUDVRQ D 6DUDVRQ 3HGHUVRQ t 1\PDQ 6DUDVRQ t 6WRRSV f ,W KDV DOVR EHHQ XVHG LQ PDQ\ RWKHU UHVHDUFK VWXGLHV LQ WKH DUHD RI WHVW DQ[LHW\ :LQH 6SLHOEHUJHU $QWRQ t %HGHOO :LQH f $OO VXEMHFWV ZHUH DOVR DGPLQLVWHUHG WKH 0HDVXUH RI $FDGHPLF 6HOI(IILFDF\ 0$6(f GHYHORSHG E\ /DORQGH f 7KLV LV D LWHP WHVW GHVLJQHG WR DVVHVV VHOIHIILFDF\ LQ DFDGHPLF VLWXDWLRQV 7KH +R\W HVWLPDWH RI LQWHUQDO FRQVLVWHQF\ UHOLDELOLW\ IRU WKH WHVW ZDV IRXQG WR EH ZKLOH WHVWUHWHVW UHOLDELOLW\ RYHU D RQH PRQWK SHULRG ZDV 7KH GLVWULEXWLRQ RI WKH 0$6( RQ D VDPSOH RI PDOH DQG IHPDOH KLJK VFKRRO VWXGHQWVf UHDVRQDEO\ DSSUR[LPDWHG D QRUPDO FXUYH 3VHXGRUDQGRP UHVSRQGLQJ DQG DFTXLHVFHQW

PAGE 70

UHVSRQVH VHW ZHUH FRQWUROOHG E\ KDYLQJ HTXDO QXPEHUV RI QHJDWLYHO\ DQG SRVLWLYHO\ ZRUGHG LWHPV DQG WKH LQFOXVLRQ RI LQIUHTXHQF\ LWHPV LQ WKH VFDOH 6RFLDO GHVLUDELOLW\ DV D UHVSRQVH VHW ZDV FRQWUROOHG E\ HPSOR\LQJ -DFNVRQnV f VWULQJHQW LWHP VHOHFWLRQ SURFHGXUH 2QO\ WKRVH LWHPV ZKLFK FRUUHODWHG PRUH KLJKO\ ZLWK WKH HIILFDF\ VFDOH WKDQ ZLWK WKH VRFLDO GHVLUDELOLW\ VFDOH ZHUH LQFOXGHG LQ WKH ILQDO VFDOH 7KH 0$6( FRQWDLQV D /LNHUWW\SH IRUPDW 7KH VXEMHFWV ZHUH UHTXLUHG WR DVVHVV WKHLU UHVSRQVH WR HDFK LWHP RQ D ILYH SRLQW VFDOH UDQJLQJ IURP QHYHU RU UDUHO\ WUXH WR DOPRVW DOZD\V RU DOZD\V WUXH 7KLV WHVW WRRN DSSUR[Ln PDWHO\ PLQXWHV WR FRPSOHWH VR WKH HQWLUH LQLWLDO WHVWLQJ VHVVLRQ WRRN DSSUR[LPDWHO\ PLQXWHV $W WKH WLPH RI WKH LQLWLDO WHVWLQJ VHVVLRQ VXEMHFWV DOVR JDYH WKHLU ZULWWHQ SHUPLVVLRQ WR UHOHDVH WR WKH LQYHVWLJDWRU WKHLU YHUEDO DQG TXDQWLWDWLYH 6FKRODVWLF $SWLWXGH 7HVW VFRUHV DQG ILQDO JUDGHV LQ 36< ,QWURn GXFWRU\ 3V\FKRORJ\f 7KH 6FKRODVWLF $SWLWXGH 7HVW 6$7f VFRUHV ZHUH REWDLQHG IURP WKH 5HJLVWUDUnV RIILFH DQG WKH JUDGHV ZHUH REWDLQHG IURP WKH LQGLYLGXDO LQVWUXFWRUV $W WKH HQG RI WKH LQLWLDO WHVWLQJ VHVVLRQ WKH VXEMHFWV ZHUH WROG WR UHWXUQ IRU DQ DGGLWLRQDO WHVWLQJ VHVVLRQ RQH ZHHN ODWHU 7KH ILQDO WHVWLQJ VHVVLRQ LQYROYHG WKH DGPLQLVWUDWLRQ RI D VHULHV RI DQDJUDPV VHOHFWHG IURP D OLVW RI DQDJUDPV GHYHORSHG E\ 6DUJHQW f 7KHVH DQDJUDPV KDYH EHHQ XVHG

PAGE 71

LQ VHYHUDO WHVW DQ[LHW\ VWXGLHV 6DUDVRQ t 6WRRSV 6DUDVRQ D E 5XVVHOO t 6DUDVRQ &UDJHU f 7KH 6DUJHQW VHULHV RI DQDJUDPV FRQVLVWV RI DQDJUDPV ZKLFK DUH GLYLGHG LQWR WKUHH GLIILFXOW\ OHYHOV HDV\ PHGLXP KDUG 3ULRU WR WKH EHJLQQLQJ RI WKH DFWXDO GLVVHUWDWLRQ VWXG\ D SLORW VWXG\ KDG EHHQ FRQGXFWHG WR VHOHFW DQDJUDPV ZKLFK ZRXOG EH GLIILFXOW HQRXJK WR SUHVHQW D FKDOOHQJH \HW QRW VR GLIILFXOW DV WR SURYLGH QR UDQJH RI SHUIRUPDQFH ,Q RUGHU WR DVVHVV WKH OHYHO RI GLIILFXOW\ RI WKH 6DUJHQW f DQDJUDPV LQ WKLV FROOHJH SRSXODWLRQ VXEMHFWV ZHUH DGPLQLVWHUHG DOO RI 6DUJHQWnV DQDJUDPV 7KH\ ZHUH JLYHQ XQOLPLWHG WLPH WR FRPSOHWH WKH DQDJUDPV KRZHYHU WKH\ ZHUH DVVHVVHG IRU WKH DPRXQW RI WLPH LW WRRN WR FRPSOHWH WKH PRVW GLIILFXOW DQDJUDPV 7KLV ZDV GRQH LQ RUGHU WR SURYLGH DQ HVWLPDWH RI DQ DSSURSULDWH WLPH OLPLW IRU SUHVHQWDWLRQ RI WKH DQDJUDPV HYHQWXDOO\ VHOHFWHG 7KH SHUFHQWDJH RI VXEMHFWV VROYLQJ HDFK DQDJUDP FRUUHFWO\ ZDV FRPSXWHG ,W ZDV IRXQG WKDW IRU WKLV SLORW VDPSOH QRW DOO RI 6DUJHQWnV PRVW GLIILFXOW DQDJUDPV ZHUH WKH PRVW GLIILFXOW 7KH PRVW GLIILFXOW DQDJUDPV IRU WKLV VDPSOH ZHUH VROYHG E\ b WR b RI WKH VDPSOH (LJKW RI WKHVH DQDJUDPV ZHUH LQ 6DUJHQWnV HDV\ JURXS DQG ILYH ZHUH LQ 6DUJHQWnV PHGLXP GLIILFXOW\ JURXS $V PHQWLRQHG DERYH WKH VXEMHFWV ZHUH DOVR DVVHVVHG IRU WLPH WR

PAGE 72

FRPSOHWLRQ RI 6DUJHQWnV KDUG DQDJUDPV ,W ZDV IRXQG WKDW WKH\ WRRN D PHDQ WLPH RI PLQXWHV WR FRPSOHWH WKHVH DQDJUDPV $IWHU WKLV LQLWLDO SLORW VWXG\ ZDV FRPSOHWHG DQ DGGLWLRQDO SLORW VWXG\ ZDV FRQGXFWHG )RU WKLV VWXG\ VXEMHFWV ZHUH JLYHQ DOO DVSHFWV RI WKH IXOO GLVVHUWDWLRQ VWXG\ LQFOXGLQJ ERWK WKH TXHVWLRQQDLUH VHVVLRQ ILUVW VHVVLRQf DQG WKH H[SHULPHQWDO VHVVLRQ VHFRQG VHVVLRQf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n WDQWV &RQGLWLRQ RQH WKH QHXWUDO LQVWUXFWLRQDO FRQGLWLRQ LQYROYHG WKH IROORZLQJ LQVWUXFWLRQV DGDSWHG IURP 5XVVHOO t 6DUDVRQ f 7KHVH LQVWUXFWLRQV ZHUH UHDG DORXG WR WKH VXEMHFWV DV D JURXS

PAGE 73

7KLV SDUW RI WKH H[SHULPHQW LQYROYHV D SUREOHPVROYLQJ WDVN 7KH SUREOHPV ZLOO JLYH \RX WR VROYH DUH FDOOHG DQDJUDPV $Q DQDJUDP LV D ZRUG ZLWK WKH OHWWHUV DOO PL[HG XS DQG \RXU WDVN LV WR PDNH D PHDQn LQJIXO ZRUG 2QO\ RQH ZRUG FDQ EH PDGH IURP HDFK DQDJUDP DQG WKH\ DUH DOO FRPPRQ ZRUGV 6ROYH WKH DQDJUDPV DQG ZULWH WKH DQVZHUV RQ WKH SUREOHP VKHHWV LQ WKH DQVZHU FROXPQ ,I \RX QHHG VFUDWFK SDSHU XVH WKH EDFNV RI WKH DQVZHU VKHHWV
PAGE 74

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nW ZRUU\ DERXW LW 1R RQH ZLOO ILQG WKH DQDJUDPV HDV\
PAGE 75

6WDWLVWLFDO $QDO\VLV $OO DQDO\VHV ZHUH FRQGXFWHG XVLQJ WKH 6WDWLVWLFDO $QDO\VLV 6\VWHP 6$6f +HOZLJ t &RXQFLO f SDFNDJHG SURJUDPV IRU PXOWLSOH UHJUHVVLRQ DQDO\VHV 7KH WZR SULPDU\ SURJUDPV XWLOL]HG ZHUH */0 JHQHUDO OLQHDU PRGHOVf DQG 6WHSZLVH VWHSZLVH UHJUHVVLRQf 3ULRU WR DQ\ UHJUHVVLRQ DQDO\VHV EHLQJ FRQGXFWHG LW ZDV QHFHVVDU\ WR UDQGRPO\ VHOHFW VXEMHFWV IRU D PDLQ VDPSOH DQG D YDOLGDWLRQ VDPSOH 2XW RI WKH VXEMHFWV WKHUH ZHUH VXEMHFWV ZKR ZHUH PLVVLQJ GDWD IRU LQGHSHQGHQW YDULDEOHV HLWKHU DSWLWXGH VFRUHV WHVW DQ[LHW\ VFRUHV RU VHOIHIILFDF\ VFRUHV 0RVW RI WKHVH VXEMHFWV ZHUH PLVVLQJ 6$7 VFRUHV IRU YDULRXV UHDVRQV IRUHLJQ VWXGHQWV MXQLRU FROOHJH WUDQVIHUV HWFf )RXU VXEMHFWV ZHUH PLVVLQJ WHVW DQ[LHW\ RU VHOI HIILFDF\ VFRUHV GXH WR ODFN RI FRPSOHWLRQ RI WKH TXHVWLRQQDLUHV 7KHVH VXEMHFWV ZHUH JHQHUDOO\ H[FOXGHG IURP UHJUHVVLRQ DQDO\VHV DOWKRXJK WKHLU VFRUHV ZHUH LQFOXGHG LQ PHDQV DQDO\VHV DQG 3HDUVRQnV 5 FRUUHODWLRQV $Q DGGLWLRQDO WKUHH VXEMHFWV ZHUH LQFOXGHG LQ WKH DQDJUDP DQDO\VHV EXW H[FOXGHG LQ WKH JUDGH DQDO\VHV EHFDXVH WKH\ GLG QRW KDYH JUDGHV IRU WKH 36< ,QWURn GXFWRU\ 3V\FKRORJ\f FRXUVH 7KH UHPDLQLQJ VXEMHFWV ZHUH UDQGRPO\ VHOHFWHG WR EH HLWKHU LQ WKH PDLQ VDPSOH FRQVLVWLQJ RI VXEMHFWV RU D YDOLGDWLRQ VDPSOH FRQVLVWLQJ RI VXEMHFWV 7KH

PAGE 76

UDQGRPL]DWLRQ ZDV DFFRPSOLVKHG E\ FRPSXWHU XVLQJ D UDQGRP VHHG SURFHGXUH IURP WKH 6$6 V\VWHP +HOZLJ t &RXQFLO f 7KH LQLWLDO DQDO\VHV LQYROYHG WHVWLQJ WKH RULJLQDO SURSRVHG PRGHOV RQH IRU DQDJUDP SHUIRUPDQFH DV WKH GHSHQGHQW YDULDEOH DQG RQH IRU JUDGH SHUIRUPDQFH DV WKH GHSHQGHQW YDULDEOH $IWHU WKHVH PRGHOV ZHUH WHVWHG D VWHSZLVH UHJUHVVLRQ ZDV FRQGXFWHG WR HVWDEOLVK WKH EHVW PRGHOV IRU ERWK JUDGH DQG DQDJUDP 7KHVH DQDO\VHV LQFOXGHG WKH YDULDEOHV IURP WKH RULJLQDO SURSRVHG PRGHOV DV ZHOO DV RQH DGGLWLRQDO LQWHUDFWLRQDO YDULDEOH WHVW DQ[LHW\ LQWHUn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

PAGE 77

&+$37(5 7+5(( 5(68/76 'HVFULSWLYH VWDWLVWLFV ZHUH FDOFXODWHG IRU DOO RI WKH FRQWLQXRXV YDULDEOHV LQ WKH VWXG\ WHVW DQ[LHW\ VHOI HIILFDF\ 6FKRODVWLF $SWLWXGH 7HVW 6$7f YHUEDO DQG TXDQWLWDWLYH FRPELQHG 6$7 YHUEDO DORQH 6$7 TXDQWLWDWLYH DORQH JUDGH DQG DQDJUDP 7DEOH f ,Q DGGLWLRQ 3HDUVRQ U FRUUHODWLRQ PDWULFHV ZHUH FDOFXODWHG RQ DOO RI WKH TXDQWLWDWLYH YDULDEOHV IRU WKH FRPELQHG PDLQ DQG YDOLGDWLRQ VDPSOH RI Q 7DEOH f $V K\SRWKHVL]HG WHVW DQ[LHW\ FRUUHODWHG PRGHUDWHO\ QHJDWLYHO\ ZLWK VHOIHIILFDF\ U S f ,QLWLDOO\ PXOWLSOH UHJUHVVLRQ DQDO\VHV ZHUH FRQGXFWHG IRU ERWK JUDGH DQG DQDJUDP DV WKH GHSHQGHQW YDULDEOHV 7KH LQLWLDO PRGHO LQFOXGHG JURXS LH LQVWUXFWLRQDO FRQGLWLRQf WHVW DQ[LHW\ VHOIHIILFDF\ 6$7 YHUEDO 6$79f 6$7 TXDQWLn WDWLYH 6$74f DQG LQWHUDFWLRQV EHWZHHQ JURXS DQG WHVW DQ[LHW\ JURXS DQG VHOIHIILFDF\ JURXS DQG 6$79 DQG JURXS DQG 6$74 7KH LQLWLDO DQDO\VLV XVLQJ DQDJUDP DV WKH GHSHQGHQW YDULDEOH LQGLFDWHG WKDW WKH IROORZLQJ YDULDEOHV ZHUH DOO VLJQLILFDQW LQGLYLGXDO SUHGLFWRUV RI DQDJUDP SHUIRUPDQFH DW WKH S OHYHO RU EHWWHU WHVW DQ[LHW\ S f

PAGE 78

7DEOH 'HVFULSWLYH 6WDWLVWLFV IRU ,QGHSHQGHQW DQG 'HSHQGHQW 9DULDEOHV 9DULDEOH 1 0HDQ 6WDQGDUG 'HYLDWLRQ 0LQLPXP 9DOXH 0D[LPXP 9DOXH 7HVW $Q[LHW\ 6HOI(IILFDF\ 6$79HUEDO 6$74XDQWLWDWLYH $QDJUDP *UDGH 6$7 6FKRODVWLF $SWLWXGH 7HVWf

PAGE 79

7DEOH ,QWHUFRUUHODWLRQ 0DWUL[ RQ WKH ,QGHSHQGHQW DQG 'HSHQGHQW 9DULDEOHV Q f 7HVWDQ[ 6HOIHI 6$79 6$74 *UDGH $QDJUDP 7HVWDQ[ rr rr rr r 6HOIHI 6$79 rr r 6$74 r *UDGH $QDJUDP 1RWH 7HVWDQ[ 7HVW $Q[LHW\ 6HOIHI 6HOI(IILFDF\ 6$79 6FKRODVWLF $SWLWXGH 7HVW9HUEDO 6$74 6FKRODVWLF $SWLWXGH 7HVW4XDQWLWDWLYH r S rr S J $OO FRUUHODWLRQV RI JUDGH ZLWK WKH RWKHU YDULDEOHV ZHUH GRQH RQ D VDPSOH RI Q EHFDXVH WKUHH VXEMHFWV KDG PLVVLQJ JUDGH GDWD 1-

PAGE 80

VHOIHIILFDF\ S f 6$79 S f 6$74 S f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f OHYHO RU EHWWHU 7KH U IRU WKLV PRGHO ZDV 9DOLGDWLRQ DQDO\VHV ZHUH FRQGXFWHG IRU ERWK WKH DQDJUDP PRGHO DQG WKH JUDGH PRGHO )RU WKH DQDJUDP PRGHO WKH REVHUYHG VFRUHV IRU WKH YDOLGDWLRQ VDPSOH Q f ZHUH FRUUHODWHG ZLWK WKHLU H[SHFWHG VFRUHV EDVHG RQ WKH UHJUHVVLRQ HTXDWLRQ JHQHUDWHG IURP WKH PDLQ VDPSOH 7KH 3HDUVRQ U FRUUHODWLRQ ZDV ZKLFK ZDV QRW VLJQLILFDQW LQGLFDWLQJ WKDW WKH UHVXOWV ZLWK WKLV VDPSOH RI Q PD\ QRW EH VWDEOH ZKHQ WKLV PRGHO LV UHSOLFDWHG ZLWK RWKHU VDPSOHV RI HTXLYDOHQW VL]H 9DOLGDWLRQ DQDO\VHV ZHUH DOVR FRQGXFWHG IRU WKH JUDGH PRGHO 7KH REVHUYHG VFRUHV IRU WKH YDOLGDWLRQ VDPSOH Q f ZHUH FRUUHODWHG ZLWK WKHLU H[SHFWHG VFRUHV EDVHG RQ WKH UHJUHVVLRQ HTXDWLRQ IRU JUDGH LQ WKH PDLQ VDPSOH

PAGE 81

7KH 3HDUVRQ U FRUUHODWLRQ ZDV ZKLFK ZDV VLJQLILFDQW DW WKH S OHYHO RI SUREDELOLW\ 7KLV PRGHO GLG WR WKH YDOLGDWLRQ VDPSOH $GGLWLRQDO DQDO\VHV ZHUH DOVR FRQGXFWHG ZLWK UHGXFHG PRGHOV IRU ERWK JUDGH DQG DQDJUDP LQFOXGLQJ PDLQ HIIHFWV RQO\f ,Q ERWK DQDJUDP DQG JUDGH PRGHOV PDLQ HIIHFWV IRU JURXS DQG WKH LQWHUDFWLRQ RI JURXS ZLWK WKH RWKHU YDULDEOHV ZHUH HOLPLQDWHG IURP WKH PRGHOV ) WHVWV ZHUH FRQGXFWHG FRPSDULQJ WKH UHGXFHG PRGHOV WR WKH IXOO PRGHOV )RU ERWK JUDGH DQG DQDJUDP WKH UHVXOWV ZHUH QRW VLJQLILFDQW LQGLn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

PAGE 82

EHWZHHQ WKH YDULDEOHV 6LQFH WH[W DQ[LHW\ DQG VHOIHIILFDF\ ZHUH K\SRWKHVL]HG WR MRLQWO\ SUHGLFW FRJQLWLYH WDVN SHUn IRUPDQFH LW ZDV DGGLWLRQDOO\ K\SRWKHVL]HG WKDW WKH LQWHUDFWLRQ RI WHVW DQ[LHW\ DQG VHOIHIILFDF\ PLJKW SUHGLFW FRJQLWLYH SHUIRUPDQFH PRUH HIIHFWLYHO\ WKDQ WHVW DQ[LHW\ DQG VHOIHIILFDF\ RSHUDWLQJ ZLWKLQ WKH PRGHO LQ D OLQHDU IDVKLRQ ,Q WKLV VWHSZLVH SURFHGXUH DOO WKH PDLQ HIIHFW YDULDEOHV ZHUH LQFOXGHG LH WHVW DQ[LHW\ VHOIHIILFDF\ 6$79 DQG 6$74 ,Q DGGLWLRQ WKH LQWHUDFWLRQ RI WHVW DQ[LHW\ DQG VHOIHIILFDF\ 7(676(f ZDV LQFOXGHG LQ WKH PRGHO DV GLVFXVVHG DERYH ,Q RUGHU WR HVWLPDWH WKH EHVW PRGHO LW ZDV QHFHVVDU\ WR XVH DOO DYDLODEOH GDWD +HQFH WKH PDLQ VDPSOH DQG YDOLGDWLRQ VDPSOH ZHUH PHUJHG DQG D VWHSZLVH UHJUHVVLRQ ZDV UXQ IRU WKH FRPSOHWH VDPSOH ,Q WKLV FDVH VXEMHFWV ZHUH LQFOXGHG LQ WKH VWHSZLVH DQDO\VLV IRU JUDGH DQG VXEMHFWV ZHUH LQFOXGHG LQ WKH VWHSZLVH DQDO\VLV IRU DQDJUDP $ VWHSZLVH UHJUHVVLRQ DQDO\VLV RQ WKH FRPSOHWH VDPSOH ZDV FRQGXFWHG IRU DQDJUDP DV WKH GHSHQGHQW YDULDEOH Q f 7KH VWHSZLVH DQDO\VLV \LHOGHG WZR YDULDEOHV DV VLJQLILFDQW SUHGLFWRUV RI DQDJUDP WDVN SHUIRUPDQFH 7KHVH YDULDEOHV ZHUH WHVW DQ[LHW\ DQG WKH LQWHUDFWLRQ RI WHVW DQ[LHW\ ZLWK VHOIHIILFDF\ 7HVW DQ[LHW\ SUHGLFWHG SHUIRUPDQFH DW D SUREDELOLW\ RI S 7HVW DQ[LHW\ LQ LQWHUDFWLRQ ZLWK VHOIHIILFDF\ SUHGLFWHG DQDJUDP SHUIRUPDQFH DW S

PAGE 83

7KH U IRU WKLV PRGHO ZDV 7KH RYHUDOO ) YDOXH ZDV ) f S ,Q WKH VWHSZLVH UHJUHVVLRQ DQDO\VLV IRU JUDGH RQ D VDPSOH RI Q WKUHH VXEMHFWV LQ WKH PDLQ VDPSOH GLG QRW KDYH JUDGH GDWDf WZR YDULDEOHV ZHUH VLJQLILFDQW SUHGLFWRUV RI JUDGH SHUIRUPDQFH 7KHVH YDULDEOHV ZHUH WHVW DQ[LHW\ DQG 6$74 7KH U IRU WKLV PRGHO ZDV U 7KH RYHUDOO ) YDOXH IRU WKLV PRGHO ZDV ) f S 7HVW DQ[LHW\ ZDV D VLJQLILFDQW SUHGLFWRU YDULDEOH DW S DQG 6$74 ZDV VLJQLILFDQW DW S 9DOLGDWLRQ DQDO\VHV ZHUH FRQGXFWHG IRU WKH WZR VWHSZLVH UHJUHVVLRQ PRGHOV GHYHORSHG IURP WKH FRPSOHWH VDPSOH RQH IRU JUDGHV DQG RQH IRU DQDJUDPf )LUVW WKH PRGHOV IRU DQDJUDP DQG JUDGH GHYHORSHG LQ WKH IXOO VWHSZLVH UHJUHVVLRQ ZHUH XVHG LQ D VWDQGDUG PXOWLSOH UHJUHVVLRQ SHUIRUPHG RQ WKH PDLQ VDPSOHV 7KHQ WKHVH PRGHOV ZHUH YDOLGDWHG RQ WKH VPDOOHU YDOLGDWLRQ VDPSOHV 7KH VDPSOHV ZHUH PHUJHG DQG WKH ILQDO UHJUHVVLRQ UHVXOWV DUH UHSRUWHG )RU WKH DQDJUDP PRGHO WHVW DQ[LHW\ DQG 7(676( DV SUHGLFWRUVf WKH PDLQ VDPSOH ZDV YDOLGDWHG RQ WKH YDOLGDWLRQ VDPSOH 7KH FRUUHODWLRQ EHWZHHQ WKH H[SHFWHG VFRUH DQG WKH REVHUYHG VFRUH ZDV U OO ZKLFK ZDV QRQVLJQLILFDQW 7KH YDOLGDWLRQ VDPSOH ZDV FRPELQHG ZLWK WKH PDLQ VDPSOH DQG D UHJUHVVLRQ DQDO\VLV ZDV GRQH 7KLV DQDO\VLV \LHOGHG DQ U RI 7HVW DQ[LHW\ SUHGLFWHG DQDJUDP SHUIRUPDQFH DW WKH S OHYHO DQG 7(676( SUHGLFWHG DW WKH S OHYHO

PAGE 84

7KH VDPH DQDO\VHV GHVFULEHG DERYH IRU DQDJUDP ZHUH DOVR GRQH ZLWK JUDGH DV WKH GHSHQGHQW YDULDEOH 7KH VWHSZLVH UHJUHVVLRQ KDG IRXQG WKDW WKH EHVW PRGHO SUHGLFWLQJ JUDGH SHUIRUPDQFH ZDV WHVW DQ[LHW\ DQG 6$74 7KH VWDQGDUG UHJUHVVLRQ SURFHGXUH */0f ZDV XVHG WR DQDO\VH WKLV PRGHO RQ WKH PDLQ VDPSOH 7KLV PRGHO ZDV WKHQ YDOLGDWHG RQ WKH FURVVYDOLGDWLRQ VDPSOH 7KH H[SHFWHG YDOXH IRU JUDGH ZDV FRUUHODWHG ZLWK WKH REVHUYHG YDOXH RQ WKH YDOLGDWLRQ VDPSOH DQG WKH 3HDUVRQ U ZDV IRXQG WR EH U ZKLFK ZDV VLJQLILFDQW DW S +HQFH WKLV PRGHO ZDV YDOLGDWHG 7KH YDOLGDWLRQ VDPSOH ZDV WKHQ FRPELQHG ZLWK WKH PDLQ VDPSOH DQG D PXOWLSOH UHJUHVVLRQ DQDO\VLV ZDV SHUIRUPHG RQ WKH FRPELQHG VDPSOH Q f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

PAGE 85

FRQGLWLRQ ,Q WKH WHVW DQ[LHW\ FODVVLILFDWLRQ WKH KLJK DQG ORZ WHVWDQ[LRXV VXEMHFWV ZHUH GUDZQ IURP WKH XSSHU DQG ORZHU b RI WKH VFRUH GLVWULEXWLRQ 7KH VFRUH UDQJHV IRU WKH WKUHH 7HVW $Q[LHW\ 6FDOH JURXSV ZHUH KLJK WHVW DQ[LHW\ PLGGOH WHVW DQ[LHW\ DQG ORZ WHVW DQ[LHW\ 7KHVH FXWRII VFRUHV FRUUHVSRQGHG DOPRVW LGHQWLFDOO\ WR WKRVH XVHG E\ 6DUDVRQ f ,Q WKLV DQDO\VLV RI YDULDQFH ZLWK JURXS DQG WHVW DQ[LHW\ DV WKH WZR LQGHSHQGHQW YDULDEOHV DQG DQDJUDP SHUIRUPDQFH DV WKH GHSHQGHQW YDULDEOH WKH RYHUDOO ) YDOXH ZDV QRW VLJQLILFDQW DQG WKHUH ZHUH QR VLJQLILFDQW PDLQ HIIHFWV RU LQWHUDFWLRQV DOWKRXJK WKH JURXS PDLQ HIIHFW DSSURDFKHG VLJQLILFDQFH S f 7KH VDPH DQDO\VLV RI YDULDQFH ZDV DOVR FRQGXFWHG IRU JUDGH DV WKH GHSHQGHQW YDULDEOH ,Q WKLV DQDO\VLV WKH RYHUDOO ) YDOXH ZDV VLJQLILFDQW ) f S f 7KHUH ZDV D VLJQLILn FDQW PDLQ HIIHFW IRU WHVW DQ[LHW\ S f )XUWKHU DQDO\VHV ZHUH FRQGXFWHG ZLWK 'XQFDQnV 0XOWLSOH 5DQJH 7HVW WR FRPSDUH WKH LQGLYLGXDO WHVW DQ[LHW\ PHDQV 7KH UHVXOWV LQGLFDWHG WKDW WKH KLJK WHVW DQ[LHW\ JURXS UHFHLYHG VLJQLILn FDQWO\ ORZHU JUDGHV S f WKDQ WKH ORZ DQG PLGGOH WHVW DQ[LHW\ JURXSV EXW WKHVH WZR JURXSV GLG QRW VLJQLILFDQWO\ GLIIHU IURP HDFK RWKHU $Q DGGLWLRQDO DQDO\VLV RI YDULDQFH ZDV FRQGXFWHG LQFRUSRUDWLQJ WHVW DQ[LHW\ VHOIHIILFDF\ DQG JURXS LQVWUXFWLRQV 7KH VXEMHFWV ZHUH FODVVLILHG LQWR WKH WHVW

PAGE 86

DQ[LHW\ JURXSV ZLWK WKH VDPH FXWRII VFRUHV DV DERYH )RU WKH VHOIHIILFDF\ PHDVXUH WKH VXEMHFWV ZKR ZHUH FODVVLILHG LQWR WKH KLJK DQG ORZ VHOIHIILFDF\ JURXSV ZHUH LQ WKH XSSHU DQG ORZHU b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f 'XQFDQnV 0XOWLSOH 5DQJH 7HVW LQGLFDWHG DJDLQ WKDW WKH KLJK WHVWDQ[LRXV JURXS UHFHLYHG VLJQLILFDQWO\ ORZHU JUDGHV WKDQ WKH PLGGOH DQG ORZ WHVWDQ[LRXV JURXS S f

PAGE 87

&+$37(5 )285 ',6&866,21 7HVW DQ[LHW\ DQG VHOIHIILFDF\ FRUUHODWHG VLJQLILFDQWO\ LQ D PRGHUDWH QHJDWLYH UHODWLRQVKLS 7KH FRUUHODWLRQ RI ZDV YHU\ VLPLODU WR WKH FRUUHODWLRQ FRHIILFLHQW IRXQG E\ /DORQGH LQ KHU GLVVHUWDWLRQ VWXG\ U /DORQGH f 7KHVH UHVXOWV LQGLFDWH WKDW WHVW DQ[LHW\ DQG VHOIHIILFDF\ DUH WZR PRGHUDWHO\ UHODWHG EXW QRW LGHQWLFDO FRQVWUXFWV ,Q WKH LQLWLDO PRGHO ZLWK DQDJUDP DV WKH GHSHQGHQW YDULDEOH WHVW DQ[LHW\ VHOIHIILFDF\ 6$7 YHUEDO DQG 6$7 TXDQWLWDWLYH ZHUH DOO VLJQLILFDQW SUHGLFWRUV RI SHUIRUPDQFH +RZHYHU WKH LQVWUXFWLRQDO FRQGLWLRQ HYDOXDWLYH UHDVVXULQJ QHXWUDOf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

PAGE 88

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f +RZHYHU H[DPLQDWLRQ RI SDVW VWXGLHV XVLQJ

PAGE 89

WKH LQVWUXFWLRQDO PDQLSXODWLRQ LQGLFDWHG LQFRQFOXVLYH UHVXOWV 6RPH VWXGLHV KDYH IRXQG D PDLQ HIIHFW IRU WHVW DQ[LHW\ ZLWK KLJK WHVWDQ[LRXV VXEMHFWV GRLQJ PRUH SRRUO\ WKDQ ORZ WHVWDQ[LRXV VXEMHFWV EXW QR WHVW DQ[LHW\ E\ LQVWUXFWLRQ LQWHUDFWLRQ 6DUDVRQ 0DQGOHU t &UDLJKLOO 0DQGOHU t 6DUDVRQ &UDJHU 5XVVHOO t 6DUDVRQ f $OWKRXJK WKH WHVW DQ[LHW\ E\ LQVWUXFWLRQ LQWHUDFWLRQ KDV EHHQ IRXQG E\ VRPH UHVHDUFKHUV LW LV QRW DQ XQHTXLYRFDO ILQGLQJ ,W FRXOG EH VSHFXODWHG WKDW WKH FXUUHQW VWXG\ GLG QRW ILQG D WHVW DQ[LHW\ E\ LQVWUXFWLRQ LQWHUDFWLRQ EHFDXVH RI WKH PHWKRG RI GDWD DQDO\VLV $OO RI WKH SUHYLRXV VWXGLHV XWLOL]HG DQ DQDO\VLV RI YDULDQFH $129$f PRGHO LQ VWDWLVn WLFDO DQDO\VLV FODVVLI\LQJ WKH VXEMHFWV LQWR KLJK DQG ORZ WHVW DQ[LHW\ JURXSV PHGLDQ VSOLWf RU WKUHH WHVW DQ[LHW\ JURXSV ORZ PLGGOH KLJKf 7KH FXUUHQW VWXG\ WUHDWHG WKH GDWD LQ FRQWLQXRXV IRUP DQG DQDO\]HG WKH GDWD ZLWKLQ D PXOWLSOH UHJUHVVLRQ IUDPHZRUN ,Q RUGHU WR DQVZHU WKLV TXHVWLRQ WKH GDWD IURP WKLV VWXG\ ZHUH DQDO\]HG YLD WKH $129$ PRGHO WR PDNH WKH DQDO\VHV PRUH FRPSDUDEOH WR SUHYLRXV VWXGLHV 7KH DQDO\VLV RI YDULDQFH IRU WHVW DQ[LHW\ DQG LQVWUXFn WLRQV ZLWK DQDJUDP DV WKH GHSHQGHQW YDULDEOH IRXQG QR VLJQLILFDQW UHVXOWV 7KDW LV WKHUH ZDV QR VLJQLILFDQW UHODWLRQVKLS EHWZHHQ WHVW DQ[LHW\ DQG DQDJUDP SHUIRUPDQFH QRU DQ\ VLJQLILFDQW LQWHUDFWLRQ EHWZHHQ WHVW DQ[LHW\ DQG

PAGE 90

W\SH RI LQVWUXFWLRQ DV UHODWHG WR DQDJUDP SHUIRUPDQFH 2Q WKH RWKHU KDQG WKHUH ZDV D VLJQLILFDQW UHODWLRQVKLS EHWZHHQ WHVW DQ[LHW\ DQG JUDGH SHUIRUPDQFH ZLWK KLJK WHVWDQ[LRXV VXEMHFWV GRLQJ VLJQLILFDQWO\ PRUH SRRUO\ WKDQ PLGGOH RU ORZ WHVWDQ[LRXV VXEMHFWV LQ WKHLU JHQHUDO SV\FKROn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f :LWK DQDJUDP DV WKH GHSHQGHQW YDULDEOH KRZHYHU QR VLJQLILFDQW UHVXOWV ZHUH IRXQG 7KLV VWXG\ GLG QRW UHSOLFDWH WKRVH SUHYLRXV WHVW DQ[LHW\ VWXGLHV ZKLFK IRXQG D WHVW DQ[LHW\ E\ LQVWUXFWLRQ LQWHUDFWLRQ 6DUDVRQ 6DUDVRQ Ef KRZHYHU LW GLG UHSOLFDWH WKRVH VWXGLHV ZKLFK IRXQG D JHQHUDO QHJDWLYH UHODWLRQVKLS EHWZHHQ WHVW DQ[LHW\ DQG SHUIRUPDQFH 5XVVHOO t 6DUDVRQ &UDJHU f $SSDUHQWO\ WKH LQVWUXFWLRQV VXEMHFWV RU H[SHULn PHQWDO FRQGLWLRQV LQ WKLV VWXG\ GLIIHUHG LQ VRPH ZD\ IURP SUHYLRXV VWXGLHV ZKLFK HOLFLWHG D WHVW DQ[LHW\ E\ LQVWUXFWLRQ

PAGE 91

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n VKLS &RKHQ f KDV SURSRVHG WKDW OLQHDU UHJUHVVLRQ DQDO\VHV LQFUHDVH WKH VL]H RI VWDWLVWLFDO HIIHFWV RYHU $129$ PRGHOV E\ XS WR b $OWKRXJK WKHUH DUH PDMRU OLPLWDWLRQV RQ WKH LQWHUSUHn WDWLRQV WR EH PDGH IURP WKH UHJUHVVLRQ DQDO\VHV IRU DQDJUDP SHUIRUPDQFH JLYHQ WKH DPRXQW RI YDULDQFH DFFRXQWHG IRU E\ WKH PRGHO VRPH VSHFXODWLYH DQG WHQWDWLYH LQWHUSUHWDWLRQV FDQ EH PDGH IURP WKH ILQGLQJ RI DQ LQWHUDFWLRQ EHWZHHQ WHVW DQ[LHW\ DQG VHOIHIILFDF\ $OWKRXJK LW LV RQO\ D WHQWDWLYH ILQGLQJ WKLV LQWHUDFWLRQ LV UDWKHU LQWHUHVWLQJ DQG ZRUWK GLVFXVVLQJ 7KH LQWHUDFWLRQ EHWZHHQ WHVW DQ[LHW\ DQG VHOI HIILFDF\ FDQ EH DQDO\]HG E\ EUHDNLQJ LW XS LQWR WZR VXEVHFn WLRQV DOJHEUDLFDOO\ )LUVW RQH FDQ H[DPLQH WKH HIIHFW RI WHVWDQ[LHW\ RQ DQDJUDP SHUIRUPDQFH WDNLQJ LQWR DFFRXQW WKH HIIHFW RI VHOIHIILFDF\ 7KHQ RQH FDQ H[DPLQH WKH HIIHFW RI VHOIHIILFDF\ RQ SHUIRUPDQFH WDNLQJ LQWR DFFRXQW WKH

PAGE 92

HIIHFW RI WHVWDQ[LHW\ ,Q RUGHU WR LQWHUSUHW WKHVH HIIHFWV RQH WDNHV WKH RULJLQDO UHJUHVVLRQ HTXDWLRQ IRU WKH WRWDO VDPSOH Q f DQG VHWWLQJ WKH HTXDWLRQ WR ]HUR RQH FDQ FRPSDUH WKH HIIHFW RI D RQH XQLW FKDQJH LQ WHVWDQ[LHW\ ZLWK QR FKDQJH LQ WHVWDQ[LHW\ WR DVVHVV WKH HIIHFW RI WHVW DQ[LHW\ RQ SHUIRUPDQFH 6LPLODUO\ RQH FDQ FRPSDUH WKH HIIHFW RI D RQH XQLW FKDQJH LQ VHOIHIILFDF\ ZLWK QR FKDQJH LQ VHOIHIILFDF\ WR DVVHVV WKH HIIHFW RI VHOI HIILFDF\ RQ SHUIRUPDQFH 7KH WHVW DQ[LHW\ HIIHFW ZDV IRXQG DV IROORZV $ $QO 7$f7$r6(f $Q 7$f7$fr6(ff $ $ $QO$Q 6(f 6( 6( 7$ WHVWDQ[LHW\ 6( VHOIHIILFDF\ AQ H[SHFWHG DQDJUDP VFRUHf $ RQH XQLW FKDQJH LQ WHVW DQ[LHW\ VFRUH LV H[SHFWHG WR FRUUHVSRQG WR D FKDQJH LQ DQDJUDP VFRUH VHOIHIILFDF\f ,I RQH VROYHV WKH HTXDWLRQ VHWWLQJ WKH HTXDWLRQ WR ]HURf LW LV IRXQG WKDW DW D VHOIHIILFDF\ VFRUH RI WHVW DQ[LHW\ KDV QR HIIHFW :KHQ VHOIHIILFDF\ VFRUHV H[FHHG WKHQ WKH HIIHFW RI WHVW DQ[LHW\ RQ DQDJUDP SHUIRUPDQFH LV SRVLWLYH :KHQ VHOIHIILFDF\ VFRUHV IDOO EHORZ WKHQ WHVW DQ[LHW\ KDV D QHJDWLYH HIIHFW RQ SHUIRUPDQFH

PAGE 93

7KHVH UHVXOWV VXJJHVW WKDW IRU WZR SHUVRQV ZLWK KLJK VHOIHIILFDF\ LI RQH KDV UHODWLYHO\ KLJKHU WHVW DQ[LHW\ WKHQ WKDW SHUVRQ ZLOO KDYH EHWWHU DQDJUDP SHUIRUPDQFH &RQYHUVHO\ IRU WZR VXEMHFWV ZLWK ORZ VHOIHIILFDF\ LI RQH KDV KLJKHU WHVW DQ[LHW\ UHODWLYH WR WKH RWKHU WKHQ KLV DQDJUDP SHUIRUPDQFH LV H[SHFWHG WR EH ORZHU WKDQ WKH RWKHU SHUVRQn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

PAGE 94

GHSHQGLQJ XSRQ WKH SHUVRQnV JHQHUDO H[SHFWDQFLHV UHJDUGLQJ WKHLU DELOLW\ WR VXFFHVVIXOO\ FRSH ZLWK DFDGHPLF WDVNV $OWKRXJK WKH JHQHUDO WUHQG LQ WKH OLWHUDWXUH KDV EHHQ WR SHUFHLYH RI WHVW DQ[LHW\ DV GHELOLWDWLYH WKHRULVWV LQ UHODWHG DUHDV KDYH VXJJHVWHG WKDW ZRUU\ FDQ KDYH SRVLWLYH DVSHFWV -DQLV f KDV GLVFXVVHG WKH ZRUN RI ZRUU\LQJ DV D PHDQV RI GHDOLQJ HIIHFWLYHO\ ZLWK D FKDOOHQJLQJ UHDOLW\ VLWXDWLRQ :RUU\ FDQ EH XWLOL]HG DV D PHDQV RI PRELOL]LQJ VWUDWHJLHV DKHDG RI WLPH WR KHOS VROYH DQ DSSURDFKLQJ GLIILFXOW VLWXDWLRQ -DQLV KDV WKHRUL]HG WKDW ZKHQ ZRUU\ LV H[SHULHQFHG DGDSWLYHO\ WKH SHUVRQ EHFRPHV LQYROYHG LQ PHQWDO UHKHDUVDO ZKLFK SURGXFHV UHDOLW\EDVHG FRJQLWLRQV DQG H[SHFWDWLRQV ZKLFK IXQFWLRQ DV D VRXUFH RI KRSH DQG UHDVVXUDQFH 2I FRXUVH ZRUU\ FDQ DOVR EHFRPH GHELOLWDWLYH 0DUPRU f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

PAGE 95

ZKLFK LV LGHQWLFDO ZLWK ZKDW LV XVXDOO\ FOLQLFDOO\ ODEHOHG DV REVHVVLRQDO UXPLQn DWLRQff§RU HOVH WKH HIIRUW WR GHDO ZLWK WKH SUREOHP DW DQ LQWHOOHFWXDO OHYHO LV JLYHQ XS HQWLUHO\ DQG UHJUHVVLRQ WDNHV SODFH WR WKH HPRWLRQDO OHYHO LQ ZKLFK DQ[LHW\ ERXQG RU XQERXQG LV WKH GRPLQDQW IHDWXUH 7KH FOLQLFDO SLFWXUH PD\ WKHQ EH WKDW RI DQ DQ[LHW\ VWDWH SKRELD RU D FRQYHUVLRQ K\VWHULD 0DUPRU LQ -DQLV S f $OWKRXJK WKH FOLQLFDO VWDWH 0DUPRU LV GHVFULELQJ PLJKW EH PRUH H[WUHPH WKDQ WKDW RI D WHVWDQ[LRXV FROOHJH VWXGHQW KLV GHVFULSWLRQ RI UHDOLVWLF DQG QHXURWLF ZRUU\ DSSHDUV WR EH VRPHZKDW DQDORJRXV WR WKH ZD\ WHVW DQ[LHW\ PLJKW RSHUDWH LH LQ HLWKHU DQ DGDSWLYH RU PDODGDSWLYH IDVKLRQ ,Q $OSHUW DQG +DEHU FRQFHLYHG RI WKH LGHD WKDW WHVW DQ[LHW\ FRXOG KDYH ERWK GHELOLWDWLYH DQG IDFLOLWDWLYH FRPSRQHQWV 7KH\ GHYHORSHG D VFDOH WKH $FKLHYHPHQW $Q[LHW\ 7HVW $$7f ZKLFK DWWHPSWHG WR PHDVXUH IDFLOLWDWLYH DQG GHELOLWDWLYH WHVW DQ[LHW\ 7KH WHVW FRQVLVWHG RI WKH )DFLOLWDWLQJ $Q[LHW\ 6FDOH DQG WKH 'HELOLWDWLQJ $Q[LHW\ 6FDOH 7KH )DFLOLWDWLQJ $Q[LHW\ 6FDOH FRUUHODWHG SRVLWLYHO\ r ZLWK DFDGHPLF SHUIRUPDQFH JUDGH SRLQW DYHUDJHf DQG WKH 'HELOLWDWLQJ $Q[LHW\ 6FDOH FRUUHODWHG QHJDWLYHO\ ZLWK DFDGHPLF SHUIRUPDQFH 8VLQJ PXOWLSOH FRUUHODWLRQV WR SUHn GLFW SHUIRUPDQFH $OSHUW DQG +DEHU f IRXQG WKDW XVLQJ ERWK WKH SOXV DQG PLQXV VFDOHV WRJHWKHU HQKDQFHG SHUIRUPDQFH VLJQLILFDQWO\ EH\RQG XVH RI WKH LQGLYLGXDO VFDOHV 7KHVH VFDOHV KDYH QRW EHHQ XVHG DV RIWHQ DV WKH PRUH FRQYHQWLRQDO GHELOLWDWLQJ DQ[LHW\ VFDOHV EXW WKLV PD\ EH GXH SULPDULO\

PAGE 96

WR WKH LQYHVWLJDWRUn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

PAGE 97

FKDQJH LQ VHOIHIILFDF\ RQ DQDJUDP SHUIRUPDQFH WKH DOJHEUD LV DEEUHYLDWHG LQ WKLV FDVHf $ $QO 7$f7$f6(f $ $ $QO$Q 7$f 6HOI(IILFDF\ HIIHFW 7$f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nW PDWWHU ZKDW WKH\ EHOLHYH WKH\ FDQ GR RQ DFDGHPLF WDVNV 7KHUH PD\ EH PRUH RI D GLVFUHSDQF\ EHWZHHQ WKHLU JHQHUDO EHOLHIV DERXW WKHPVHOYHV LQ DFDGHPLF VLWXDWLRQV DQG WKHLU SHUIRUPDQFH 6LWXDWLRQDO RU WDVN

PAGE 98

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f DQG +DUULV f UHVHDUFK FLWHG HDUOLHU VXJJHVWV WKDW PHQWDO LPDJHU\ LV DQ LPSRUWDQW IDFWRU LQ WKH HIIHFWV RI WHVW DQ[LHW\ RQ SHUIRUPDQFH +DUULV f IRXQG WKDW XQGHU DQ[LHW\HYRNLQJ VLWXDWLRQV KLJK WHVWDQ[LRXV LQGLYLGXDOV WHQG WR KDYH QHJDWLYH LPDJHU\ ZKLFK LV YLYLG DQG SHUVLVWHQW :KHQ +DUULV f WUDLQHG WKH VXEMHFWV WR YLVXDOL]H LPDJHV RI FRPSHWHQFH EDVHG RQ SUHYLRXV SHUVRQDO VXFFHVV H[SHULHQFHV WKH LQGLYLGXDOV PDQLIHVWHG D UHGXFWLRQ LQ WHVW DQ[LHW\ DQG LPSURYHG DFDGHPLF SHUIRUPDQFH 7KHUHIRUH LI RQH ZHUH WR DVVHVV WKH H[WHQW RI QHJDWLYH LPDJHU\ LQ VXEMHFWV SHUKDSV WKLV ZRXOG SURYH WR EH D VWURQJHU SUHGLFWRU WKDQ WHVW DQ[LHW\

PAGE 99

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f 7KH UHODWLRQVKLS RI WHVW DQ[LHW\ WR JUDGH SHUIRUPDQFH LV QHJDWLYH 7KH PRUH WHVW DQ[LHW\ D SHUVRQ KDV WKH PRUH OLNHO\ LW ZLOO KDYH D GHELOLWDWLQJ HIIHFW RQ SHUIRUPDQFH 7KXV ZLWK JUDGH DV WKH PHDVXUH RI SHUIRUPDQFH WHVW DQ[LHW\ DSSHDUV WR KDYH WKH OLQHDU HIIHFW H[SHFWHG IURP PXFK RI WKH OLWHUDWXUH 7KLV LV QRW VXUSULVLQJ JLYHQ WKDW WKH PHDQV IRU REWDLQLQJ FODVVURRP JUDGHV DUH JHQHUDOO\ WHVWV DQG WKH WHVWLQJ

PAGE 100

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n ORJLFDO SHUVSHFWLYH ERWK WKH JUDGH PHDVXUH DQG DQDJUDP PHDVXUH KDG VRPH OLPLWDWLRQV 7KH DQDJUDP PHDVXUH RQO\ XVHG DQDJUDPV WKH W\SLFDO QXPEHU RI DQDJUDPV XVHG LQ SUHYLRXV VWXGLHVf ZKLFK PD\ KDYH OLPLWHG WKH VHQVLWLYLW\ RI WKH PHDVXUH ,Q DGGLWLRQ WKH JUDGH PHDVXUH ILQDO JUDGH LQ *HQHUDO 3V\FKRORJ\ RQO\ KDG D SRWHQWLDO IRU HLJKW GDWD SRLQWV DOVR SRVVLEO\ OLPLWLQJ WKH VHQVLWLYLW\ RI WKLV PHDVXUH $QRWKHU OLPLWDWLRQ WR WKH LQWHUSUHWDWLRQV WR EH PDGH IURP WKH DERYH PRGHOV LV EDVHG RQ D EULHI DQDO\VLV RI WKH

PAGE 101

SV\FKRPHWULF SURSHUWLHV RI WKH VHOIHIILFDF\ VFDOH 0HDVXUH RI $FDGHPLF 6HOI(IILFDF\f $OWKRXJK H[WHQVLYH DQDO\VHV ZHUH QRW SHUIRUPHG RQ WKLV LQVWUXPHQW LW DSSHDUHG WKDW WKH GLVWULEXWLRQ RI VFRUHV ZDV VRPHZKDW VNHZHG WR WKH XSSHU HQG RI SRVVLEOH VFRUHV 7KH UDQJH RI VFRUHV ZDV IURP WR ZLWK D PHDQ RI DQG D VWDQGDUG GHYLDWLRQ RI ,Q WHUPV RI WKH SRWHQWLDO UDQJH RQ WKLV VFDOH D VXEMHFW FRXOG FRQFHLYDEO\ VFRUH EHWZHHQ DQG ,Q DFWXDOLW\ PRVW VXEMHFWV VFRUHG DERYH RQ WKH VFDOH ,I RQH UHDOL]HV WKDW WKHUH LV D SRLQW SRWHQWLDO UDQJH RQ WKLV VFDOH RQH ZRXOG H[SHFW WKH PHDQ RI D QRUPDO GLVWULn EXWLRQ WR IDOO VRPHZKHUH DURXQG KDOIZD\ EHWZHHQ DQG f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f LQ KHU VWXG\ )RU KHU VDPSOH RI Q WKH GLVWULEXWLRQ \LHOGHG D PHDQ VFRUH RI D VWDQGDUG GHYLDWLRQ RI D

PAGE 102

PLQLPXP YDOXH RI DQG D PD[LPXP YDOXH RI ,Q KHU UHWHVW VDPSOH Q f KHU GHVFULSWLYH VWDWLVWLFV ZHUH VLPLODU KRZHYHUVKH GLG KDYH D EURDGHU UDQJH f ,W DSSHDUV SUREDEOH IURP TXDOLWDWLYH H[DPLQDWLRQ RI WKH 0$6( LWHPV WKDW WKH UHDVRQ WKH GLVWULEXWLRQ LV VR VNHZHG LV EHFDXVH RI WKH JOREDO QDWXUH RI WKH LQGLYLGXDO LWHPV $OO RI WKH LWHPV GHDO ZLWK WKH QRWLRQ RI SHUVLVWHQFH DW GLIILFXOW WDVNV DQG FRQWLQXLQJ WR SHUVHYHUH LQ VSLWH RI HDUO\ IDLOXUH RU JLYLQJ XS HDVLO\ +RZHYHU WKH LWHPV GR QRW DGGUHVV VSHFLILF VXEMHFW DUHDV QRU GR WKH\ SDUWLFXODUO\ DGGUHVV DFDGHPLF LVVXHV +HQFH WKHUH PD\ EH D WHQGHQF\ IRU PRVW VXEMHFWV WR DQVZHU PDQ\ RI WKH LWHPV LQ D VLPLODU IDVKLRQ ZLWKRXW PXFK GLVFULPLQDWLRQ ,I D SHUVRQ IHHOV LQ JHQHUDO WKDW WKH\ SHUVLVW DW GLIILFXOW WDVNV WKHQ WKHUH LV QR UHDVRQ IRU WKHP WR DQVZHU VRPH RI WKH LQGLYLGXDO LWHPV GLIIHUHQWO\ 7KHUH LVWKHUHIRUH D EXLOWLQ ELDV LQ WKLV WHVW WRZDUG D UHVSRQVH VHW ,I WKH LWHPV ZHUH PRUH GLYHUVLILHG SHUKDSV WKLV WHQGHQF\ WRZDUG D VNHZHG GLVWULEXWLRQ RI KLJK VHOIHIILFDF\ VFRUHV ZRXOG EH DYRLGHG 7KH SUHVHQW VWXG\ IRXQG WKDW ZKLOH DQ DSWLWXGH PHDVXUH ZDV D VLJQLILFDQW SUHGLFWRU RI FRJQLWLYH WDVN SHUIRUPDQFH RWKHU YDULDEOHV ZHUH DOVR VLJQLILFDQW LQ SUHGLFWLQJ FRJQLWLYH SHUIRUPDQFH :LWK DQ DQDJUDP WDVN DV WKH FRJQLWLYH SHUIRUPDQFH PHDVXUH WHVW DQ[LHW\ DQG VHOIHIILFDF\ ZHUH VLJQLILFDQW SUHGLFWRUV ZKHUHDV DSWLWXGH PHDVXUHV ZHUH QRW 7HVW DQ[LHW\ DQG VHOIHIILFDF\ DOVR DSSHDUHG WR SUHGLFW

PAGE 103

FRJQLWLYH SHUIRUPDQFH EHVW LQ LQWHUDFWLRQ ZLWK HDFK RWKHU UDWKHU WKDQ DV PDLQ HIIHFWV 7KHVH UHVXOWV VXSSRUW 7RELDVn UHYLHZ RI WKH WHVW DQ[LHW\ OLWHUDWXUH 7RELDV f ZKLFK VXJJHVWV WKDW WHVW DQ[LHW\ PHDVXUHV DUH OLNHO\ WR LQFUHDVH LQ WKHLU SUHGLFWLYH SRZHUV ZKHQ XVHG LQ FRPELQDWLRQ ZLWK PHDVXUHPHQW GHYLFHV ZKLFK DVVHVV SRVLWLYH RULHQWDWLRQV WR HYDOXDWLRQ )RU H[DPSOH &UDZIRUG f IRXQG WKDW WHVW DQ[LHW\ VLJQLILFDQWO\ LQWHUDFWHG ZLWK QHHG IRU DFKLHYHPHQW LQ SUHGLFWLQJ SHUIRUPDQFH XQGHU GLIIHUHQW LQVWUXFWLRQDO FRQGLWLRQV XVLQJ SURJUDPPHG LQVWUXFWLRQV 7KH WHVW DQ[LHW\VHOIHIILFDF\ LQWHUDFWLRQ DOWKRXJK D VPDOO HIIHFW ZDUUDQWV IXUWKHU LQYHVWLJDWLRQ +RZHYHU EHIRUH DGGLWLRQDO VWXGLHV DUH FRQGXFWHG LW DSSHDUV QHFHVVDU\ WR GR VRPH IXUWKHU WHVW GHYHORSPHQW ZLWK WKH FRQVWUXFW RI VHOIHIILFDF\ 7KLV FRXOG HLWKHU EH GRQH WKURXJK PRGLILFDn WLRQ RI /DORQGHnV 0HDVXUH RI $FDGHPLF 6HOI(IILFDF\ f RU WKURXJK WKH GHYHORSPHQW RI DQRWKHU PHDVXUH RI VHOI HIILFDF\ 3HUKDSV D KLJKO\ VSHFLILF PHDVXUH RI DFDGHPLF VHOIHIILFDF\ VLPLODU WR WKRVH HPSOR\HG LQ %DQGXUDnV f UHVHDUFK ZRXOG EH D PRUH SRZHUIXO LQVWUXPHQW ,I DQ DFDGHPLF VHOIHIILFDF\ PHDVXUH ZHUH GHYHORSHG VLPLODU WR WKRVH XVHG LQ WKH %DQGXUD VWXGLHV LW ZRXOG KDYH WR EH KLJKO\ VSHFLILF DQG SHUWDLQ WR WKH WDVN RU WDVNV UHTXLUHG RI VXEMHFWV GXULQJ WKH VWXG\ %DQGXUD f EHOLHYHV WKDW VHOIHIILFDF\ LV QRW D JOREDO FRQVWUXFW DQG WKHUHIRUH DVVHVVPHQW RI H[SHFWDQFLHV LQ VSHFLILF VNLOO DUHDV LV WKH

PAGE 104

RQO\ ZD\ WR PHDVXUH VHOIHIILFDF\ ,W PLJKW EH LQWHUHVWLQJ WR FRPSDUH D JOREDO PHDVXUH RI DFDGHPLF VHOIHIILFDF\ WR D VSHFLILF VNLOOEDVHG PHDVXUH RI DFDGHPLF VHOIHIILFDF\ DV SUHGLFWRUV RI SHUIRUPDQFH 7KLV VWXG\ DOVR IRXQG WKDW XQGHU FHUWDLQ FRQGLWLRQV WHVW DQ[LHW\ FDQ KDYH D SRVLWLYH HIIHFW RQ SHUIRUPDQFH VSHFLILFDOO\ ZKHQ VXEMHFWV KDYH UHODWLYHO\ KLJKHU VHOI HIILFDF\ ,W ZRXOG EH LQWHUHVWLQJ WR LQYHVWLJDWH WKH FRQWHQWV RI FRQVFLRXVQHVV :LQH f RI KLJK WHVW DQ[LRXV VXEMHFWV WR GHWHUPLQH KRZ VRPH SHUVRQV XWLOL]H WKHLU ZRUU\ RYHU WHVWV LQ D FRQVWUXFWLYH IDVKLRQ DQG KRZ RWKHUV PD\ DOORZ WKHLU ZRUU\ WR EHFRPH SDUDO\]LQJ $ SHUVRQnV OHYHO RI VHOIHIILFDF\ PD\ VRPHKRZ FRXQWHUDFW WKH ZRUU\ WKDW WKH\ H[SHULHQFH RYHU WHVW VLWXDWLRQV HQDEOLQJ WKH ZRUU\ WR EHFRPH FRQVWUXFWLYH UDWKHU WKDQ GHVWUXFWLYH 7KH +DUULV f DQG +DUULV DQG -RKQVRQ f VWXGLHV FHUWDLQO\ VXJJHVW WKDW LPDJHV RI FRPSHWHQFH FDQ UHGXFH WHVW DQ[LHW\ DQG LPSURYH SHUIRUPDQFH 7KH VHOIHIILFDF\ HIIHFW PD\ DOVR KDYH VRPHWKLQJ WR GR ZLWK WKH DWWULEXWLRQV WKDW GLIIHUHQW LQGLYLGXDOV PDNH UHJDUGLQJ WKHLU RZQ HIIHFW RQ WKHLU SHUIRUPDQFH ,I WKH\ DWWULEXWH VXFFHVVIXO SHUIRUPDQFH WR OXFN RU RWKHU VLWXDWLRQDO YDULn DEOHV LW ZRXOG DSSHDU WKDW WKH\ ZRXOG EH OHVV OLNHO\ WR PRELOL]H WKHLU RZQ UHVRXUFHV WR PDNH DQ DWWHPSW WR EH VXFFHVVIXO ,I WKH\ DWWULEXWH VXFFHVVIXO SHUIRUPDQFH WR HIIRUW WKH\ ZLOO EH PRUH OLNHO\ WR LQYHVW HIIRUW WR VXFFHHG ,W PD\ EH WKDW KLJKO\ VHOIHIILFDFLRXV LQGLYLGXDOV DUH PRUH

PAGE 105

OLNHO\ WR PDNH HIIRUW DWWULEXWLRQV :HLQHU f DQG RWKHUV KDYH EHJXQ WR ORRN DW WKH UHODWLRQVKLS EHWZHHQ DQ[LHW\ DQG DWWULEXWLRQ 6FKXQN f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f 7KH DQVZHU PD\ DOVR EH IRXQG LQ FDUHIXO DVVHVVPHQW RI WKH VXEMHFWnV LQGLYLGXDO LQWHUSUHWDWLRQ RI WKH VLWXDWLRQ WKDW LV WKHLU LQGLYLGXDO FRJQLWLYH DSSUDLVDO RI WDVN GLIILFXOW\ WKHLU H[SHFWDQF\ DV WR KRZ ZHOO WKH\ ZLOO SHUIRUP RQ WKDW WDVN WKH NLQGV RI QHJDWLYH DQGRU SRVLWLYH VHOIVWDWHPHQWV WKH\ PDNH WKHLU SHUVRQDO DWWULEXWLRQV RI FDXVDOLW\ IRU VXFFHVV RU IDLOXUH RU WKHLU PHQWDO LPDJHU\ UHJDUGLQJ DFDGHPLF VLWXDWLRQV 7KH FXUUHQW GLUHFWLRQ LQ ERWK WKH WHVW DQ[LHW\ DQG

PAGE 106

VHOIHIILFDF\ UHVHDUFK DUHDV LV WRZDUG JUHDWHU XQGHUn VWDQGLQJ RI WKH VHOIPRWLYDWLRQDO SURFHVVHV XQGHUO\LQJ SHUIRUPDQFH ,W ZRXOG EH PRVW GHVLUDEOH IRU WKHVH WZR UHVHDUFK DUHDV WR SRRO IRUFHV LQ GHYHORSLQJ DQVZHUV WR WKHVH TXHVWLRQV

PAGE 107

$33(1',&(6

PAGE 108

$33(1',; $ $1$*5$06 7$6. $QDJUDP $QVZHU 10*2,15 63588( 687&%,, 6&/,$2 (968251 56$1(2 ,07&(/$ (/6$8; *,15()2 &(3$57 (21635 (,9$55 7+:*,(

PAGE 109

$33(1',; % ,17(5&255(/$7,21 0$75,; 21 7+( ,1'(3(1'(17 $1' '(3(1'(17 9$5,$%/(6 )25 7+( 0$,1 6$03/( Q f 7HVWDQ[ 6HOIHI 6$79 6$74 *UDGH $QDJUDP 7HVWDQ[ rrr rrr rrr rr 6HOIHI 6$79 rrr rr r 6$74 r *UDGH $QDJUDP 7HVWDQ[ 7HVW $Q[LHW\ 6HOIHI 6HOI(IILFDF\ 6$79 6FKRODVWLF $SWLWXGH 7HVW9HUEDO 6$74 6FKRODVWLF $SWLWXGH 7HVW4XDQWLWDWLYH r S rr S rrr S D$OO FRUUHODWLRQV RI JUDGH ZLWK WKH RWKHU YDULDEOHV ZHUH GRQH RQ D VDPSOH RI Q EHFDXVH WZR VXEMHFWV KDG PLVVLQJ JUDGH GDWD 1RWH

PAGE 110

$33(1',; & ,17(5&255(/$7,21 0$75,; 21 7+( ,1'(3(1'(17 $1' '(3(1'(17 9$5,$%/(6 )25 727$/ 6$03/( 7HVWDQ[ 6HOIHI 6$79 6$74 *UDGH $QDJUDP 7HVWDQ[ rr Q f rr Q f rr Q f r Q f Q f 6HOIHI Q f Q f Q f Q f 6$79 rr Q f r Q f Q f 6$74 r Q f Q f *UDGH Q f $QDJUDP 1RWH 7HVWDQ[ 7HVW $Q[LHW\ 6HOIHI 6HOI (IILFDF\ 6$79 6FKRODVWLF $SWLWXGH 7HVW9HUEDO 6$74 6FKRODVWLF $SWLWXGH 7HVW4XDQWLWDWLYH r S rr S ,QFOXGHV DOO VXEMHFWV LQ WKH PDLQ VDPSOH WKH YDOLGDWLRQ VDPSOH DQG WKRVH ZLWK PLVVLQJ GDWD

PAGE 111

5()(5(1&(6 $OOHQ *7KH EHKDYLRUDO WUHDWPHQW RI WHVW DQ[LHW\ 5HFHQW UHVHDUFK DQG IXWXUH WUHQGV %HKDYLRU 7KHUDS\ $OSHUW 5 t +DEHU 5 $Q[LHW\ LQ DFDGHPLF DFKLHYHPHQW VLWXDWLRQV -RXUQDO RI $EQRUPDO DQG 6RFLDO 3V\FKRORJ\ $SRQWH t $SRQWH & *URXS SUHSURJUDPPHG V\VWHPDWLF GHVHQVLWL]DWLRQ ZLWKRXW WKH VLPXOWDQHRXV SUHVHQWDWLRQ RI DYHUVLYH VFHQHV ZLWK UHOD[DWLRQ WUDLQLQJ %HKDYLRU 5HVHDUFK DQG 7KHUDS\ B %DQGXUD $ 6HOIHIILFDF\ 7RZDUG D XQLI\LQJ WKHRU\ RI EHKDYLRUDO FKDQJH 3V\FKRORJLFDO 5HYLHZ %DQGXUD $ 6HOIHIILFDF\ PHFKDQLVP LQ KXPDQ DJHQF\ $PHULFDQ 3V\FKRORJLVW B %DQGXUD $ t $GDPV 1 $QDO\VLV RI VHOIHIILFDF\ WKHRU\ RI EHKDYLRU FKDQJH &RJQLWLYH 7KHUDS\ %DQGXUD $ $GDPV 1 t %H\HU &RJQLWLYH SURFHVVHV PHGLDWLQJ EHKDYLRU FKDQJHV -RXUQDO RI 3HUVRQDOLW\ DQG 6RFLDO 3V\FKRORJ\ B %DQGXUD $ $GDPV 1 +DUG\ $ t +RZHOOV 7HVWV RI WKH JHQHUDOLW\ RI VHOIHIILFDF\ WKHRU\ &RJQLWLYH 7KHUDS\ DQG 5HVHDUFK %DUWOHWW 5 3URJUDP IRU H[SHULPHQWV RQ WKLQNLQJ 4XDUWHUO\ -RXUQDO RI ([SHULPHQWDO 3V\FKRORJ\ %HFN $7 :HLVVPDQ $ /HVWHU t 7UH[OHU / 7KH PHDVXUHPHQW RI SHVVLPLVP 7KH KRSHOHVVQHVV VFDOH -RXUQDO RI &RQVXOWLQJ DQG &OLQLFDO 3V\FKRORJ\

PAGE 112

%URG\ (% t %URG\ 1 ,QWHOOLJHQFH 1DWXUH GHWHUPLQDQWV DQG FRQVHTXHQFHV 1HZ
PAGE 113

'HIIHQEDFKHU -/ 0DWKLV + t 0LFKDHOV $ & 6HOIFRQWURO SURFHGXUHV LQ WKH UHGXFWLRQ RI WDUJHWHG DQG QRQWDUJHWHG DQ[LHWLHV -RXUQDO RI &RXQVHOLQJ 3V\FKRORJ\ 'HIIHQEDFKHU -/ t 3DUNV '+ $ FRPSDULVRQ RI WUDGLn WLRQDO DQG VHOIFRQWURO V\VWHPDWLF GHVHQVLWL]DWLRQ -RXUQDO RI &RXQVHOLQJ 3V\FKRORJ\ 'HQQH\ '5 6HOIFRQWURO DSSURDFKHV WR WKH WUHDWPHQW RI WHVW DQ[LHW\ ,Q ,* 6DUDVRQ (Gf 7HVW DQ[LHW\ 7KHRU\ UHVHDUFK DQG DSSOLFDWLRQV +LOOVGDOH 1/DZUHQFH (UOEDXP $VVRFLDWHV 'HQQH\ '5 t 5XSHUW 3$ 'HVHQVLWL]DWLRQ DQG VHOI FRQWURO LQ WKH WUHDWPHQW RI WHVW DQ[LHW\ -RXUQDO RI &RXQVHOLQJ 3V\FKRORJ\ B 'RFWRU 50 t $OWPDQ ) :RUU\ DQG HPRWLRQDOLW\ DV FRPSRQHQWV RI WHVW DQ[LHW\ 5HSOLFDWLRQ DQG IXUWKHU GDWD 3V\FKRORJLFDO 5HSRUWV B 'RQQHU / t *XHUQH\ % $XWRPDWHG JURXS GHVHQVLWL]DWLRQ IRU WHVW DQ[LHW\ %HKDYLRU 7KHUDS\ DQG 5HVHDUFK 'XGHN 6$ *ROGEHUJ -6 /HVWHU (3 t +DUULV %5 7KH YDOLGLW\ RI FRJQLWLYH SHUFHSWXDOPRWRU DQG SHUVRQDOLW\ YDULDEOHV IRU SUHGLFWLRQ RI DFKLHYHPHQW LQ JUDGH DQG JUDGH ,, -RXUQDO RI &OLQLFDO 3V\FKRORJ\ 'XVHN -% 7KH GHYHORSPHQW RI WHVW DQ[LHW\ LQ FKLOGUHQ ,Q ,* 6DUDVRQ (Gf 7HVW DQ[LHW\ 7KHRU\ UHVHDUFK DQG DSSOLFDWLRQV@ +LOOVGDOH 1 /DZUHQFH (UOEDXP $VVRFLDWHV 'XVHN -% 0HUJOHU 1/ t .HUPLV 0' $WWHQWLRQ HQFRGLQJ DQG LQIRUPDWLRQSURFHVVLQJ LQ ORZ DQG KLJK WHVWDQ[LRXV FKLOGUHQ &KLOG 'HYHORSPHQW e (DVWHUEURRN -$ 7KH HIIHFW RI HPRWLRQ RQ FXH XWLOL]DWLRQ DQG WKH RUJDQL]DWLRQ RI EHKDYLRU 3V\FKRORJLFDO 5HYLHZ (PHU\ t .UXPEROW] 6WDQGDUG YHUVXV LQGLYLGXDOL]HG KLHUDUFKLHV LQ GHVHQVLWL]DWLRQ WR UHGXFH WHVW DQ[LHW\ -RXUQDO RI &RXQVHOLQJ 3V\FKRORJ\

PAGE 114

)LVKPDQ -$ t 3DVDQHOOD $. &ROOHJH DGPLVVLRQ VHOHFWLRQ VWXGLHV 5HYLHZ RI (GXFDWLRQDO 5HVHDUFK )UHHOLQJ 1: t 6KHPEHUJ 7KH DOOHYLDWLRQ RI WHVW DQ[LHW\ E\ V\VWHPDWLF GHVHQVLWL]DWLRQ %HKDYLRU 5HVHDUFK DQG 7KHUDS\ *DQ]HU 9(IIHFWV RI DXGLHQFH SUHVHQFH DQG WHVW DQ[LHW\ RQ OHDUQLQJ DQG UHWHQWLRQ LQ D VHULDO OHDUQLQJ VLWXDWLRQ -RXUQDO RI 3HUVRQDOLW\ DQG 6RFLDO 3V\FKRORJ\ *DUOLQJWRQ :. t &RWLHU 6 6\VWHPDWLF GHVHQVLWL]DWLRQ RI WHVW DQ[LHW\ %HKDYLRU 5HVHDUFK DQG 7KHUDS\ *HHQ 5* 7HVW DQ[LHW\ REVHUYDWLRQ DQG UDQJH RI FXH XWLOL]DWLRQ %ULWLVK -RXUQDO RI 6RFLDO DQG &OLQLFDO 3V\FKRORJ\ 7A *HHQ 5* 7HVW DQ[LHW\ DQG FXH XWLOL]DWLRQ ,Q ,* 6DUDVRQ (Gf 7HVW DQ[LHW\ 7KHRU\ UHVHDUFK DQG DSSOLFDWLRQV +LOOVGDOH 1/DZUHQFH (DUOEDXP $VVRFLDWHV *ROGIULHG 05 6\VWHPDWLF GHVHQVLWL]DWLRQ DV WUDLQLQJ LQ VHOIFRQWURO -RXUQDO RI &RQVXOWLQJ DQG &OLQLFDO 3V\FKRORJ\ *ROGIULHG 05 /LQHKDQ 00 DQG 6PLWK -/ 5HGXFWLRQ RI WHVW DQ[LHW\ WKURXJK FRJQLWLYH UHVWUXFWXULQJ -RXUQDO RI &RQVXOWLQJ DQG &OLQLFDO 3V\FKRORJ\ f +DPLOWRQ 9 6RFLDOL]DWLRQ DQ[LHW\ DQG LQIRUPDWLRQ SURFHVVLQJ $ FDSDFLW\ PRGHO RI DQ[LHW\LQGXFHG SHUIRUPDQFH ,Q ,* 6DUDVRQ t &' 6SLHOEHUJHU (GVf 6WUHVV DQG DQ[LHW\ 9RO :DVKLQJWRQ '& +HPLVSKHUH 7" +DUULV *0 $ FRPSDULVRQ RI VHYHUDO GLIIHUHQW LPDJHU\ LQVWUXFWLRQ SURFHGXUHV IRU WKH WUHDWPHQW RI WHVW DQ[LHW\ 'RFWRUDO GLVVHUWDWLRQ 8QLYHUVLW\ RI )ORULGD f 'LVVHUWDWLRQ $EVWUDFWV ,QWHUQDWLRQDO % 8QLYHUVLW\ 0LFURILOPV 1R f +DUULV *0 t -RKQVRQ 6 &RPSDULVRQ RI LQGLYLGXDOL]HG FRYHUW PRGHOLQJ VHOIFRQWURO GHVHQVLWL]DWLRQ DQG VWXG\ VNLOOV WUDLQLQJ IRU DOOHYLDWLRQ RI WHVW DQ[LHW\ -RXUQDO RI &RQVXOWLQJ DQG &OLQLFDO 3V\FKRORJ\ +HOZLJ -/ t &RXQFLO .$ 6$6 XVHUnV JXLGH HGLWLRQ &DU\ 1& 6$6 ,QVWLWXWH ,QF

PAGE 115

+ROUR\G &RJQLWLRQ DQG GHVHQVLWL]DWLRQ LQ WKH JURXS WUHDWPHQW RI WHVW DQ[LHW\ -RXUQDO RI &RQVXOWLQJ DQG &OLQLFDO 3V\FKRORJ\ +ROUR\G :HVWEURRN 7 :ROI 0 t %DGKRUQ ( 3HUIRUPDQFH FRJQLWLRQ DQG SK\VLRORJLFDO UHVSRQGLQJ LQ WHVW DQ[LHW\ -RXUQDO RI $EQRUPDO 3V\FKRORJ\ +R\W &/ 7HVW UHOLDELOLW\ HVWLPDWHG E\ DQDO\VLV RI YDULDQFH 3V\FKRPHWULND -DFNVRQ '1 3HUVRQDOLW\ UHVHDUFK IRUP *RVKHQ 1< 5HVHDUFK 3V\FKRORJLVWV 3UHVV ,QF -DFREVHQ ( 3URJUHVVLYH UHOD[DWLRQ &KLFDJR 8QLYHUVLW\ RI &KLFDJR -DQLV ,/ 3V\FKRORJLFDO VWUHVV 1HZ
PAGE 116

0DLHU 65 t 6HOLJPDQ 0( /HDUQHG KHOSOHVVQHVV 7KHRU\ DQG HYLGHQFH -RXUQDO RI ([SHULPHQWDO 3V\n FKRORJ\ 0DQGOHU t 6DUDVRQ 6% $ VWXG\ RI DQ[LHW\ DQG OHDUQLQJ -RXUQDO RI $EQRUPDO DQG 6RFLDO 3V\FKRORJ\ 0DQGOHU t :DWVRQ '/ $Q[LHW\ DQG WKH LQWHUUXSWLRQ RI EHKDYLRU ,Q &' 6SLHOEHUJHU (Gf $Q[LHW\ DQG EHKDYLRU 1HZ
PAGE 117

0RUULV /: t 3HUH] 7/ (IIHFWV RI WHVWLQWHUUXSWLRQ RQ HPRWLRQDO DURXVDO DQG SHUIRUPDQFH 3V\FKRORJLFDO 5HSRUWV 1HDOH ,0 t .DWDKQ 0 $Q[LHW\ FKRLFH DQG VWLPXOXV XQFHUWDLQW\ -RXUQDO RI 3HUVRQDOLW\ 2VWHUKRXVH 5$ 'HVHQVLWL]DWLRQ DQG VWXG\VNLOOV WUDLQLQJ DV WUHDWPHQW IRU WZR W\SHV RI WHVWDQ[LRXV VWXGHQWV -RXUQDO RI &RXQVHOLQJ 3V\FKRORJ\ 3DXO */ 0RGLILFDWLRQV RI V\VWHPDWLF GHVHQVLWL]DWLRQ EDVHG RQ FDVH VWXG\ 3DSHU SUHVHQWHG DW WKH PHHWLQJ RI WKH :HVWHUQ 3V\FKRORJLFDO $VVRFLDWLRQ 3RUWODQG 2UHJRQ $SULO 3DXO */ ,QVLJKW YHUVXV GHVHQVLWL]DWLRQ LQ SV\FKRWKHUDS\ 6WDQIRUG &DOLI 6WDQIRUG 8QLYHUVLW\ 3UHVV 3DXO */ t (ULNVHQ &: (IIHFWV RI WHVW DQ[LHW\ RQ UHDO OLIH H[DPLQDWLRQV -RXUQDO RI 3HUVRQDOLW\ 3LHUV (% 0DQXDO IRU WKH 3LHUV+DUULV FKLOGUHQnV VHOI FRQFHSW VFDOH 7KH ZD\ IHHO DERXW P\VHOIf 1DVKYLOOH 7HQQ &RXQVHORU 5HFRUGLQJV DQG 7HVWV 3ODQW :t 5LFKDUGVRQ + 7KH ,4 RI WKH DYHUDJH FROOHJH VWXGHQW -RXUQDO RI &RXQVHOLQJ 3V\FKRORJ\ 5LPP '& t 0DVWHUV -& %HKDYLRU WKHUDS\ 7HFKQLTXHV DQG HPSLULFDO ILQGLQJV 1HZ
PAGE 118

,OO 5XVVHOO 5. 0LOOHU '( t -XQH /0 $ FRPSDULVRQ EHWZHHQ JURXS V\VWHPDWLF GHVHQVLWL]DWLRQ DQG FXH FRQWUROOHG UHOD[DWLRQ LQ WKH WUHDWPHQW RI WHVW DQ[LHW\ %HKDYLRU 7KHUDS\ B 5XVVHOO 5. t 6LSLFK -) &XHFRQWUROOHG UHOD[DWLRQ LQ WKH WUHDWPHQW RI WHVW DQ[LHW\ -RXUQDO RI %HKDYLRU 7KHUDS\ DQG ([SHULPHQWDO 3V\FKLDWU\A A 5XVVHOO 5. :LVH 5 t 6WUDWRXGDNLV -3 7UHDWPHQW RI WHVW DQ[LHW\ E\ FXHFRQWUROOHG UHOD[DWLRQ DQG V\VWHPDWLF GHVHQVLWL]DWLRQ -RXUQDO RI &RXQVHOLQJ 3V\FKRORJ\ 6DPXHOV 0 t 6DPXHOV 1 6HHLQJ ZLWK WKH PLQGnV H\H 7KH KLVWRU\ WHFKQLTXHV DQG XVHV RI YLVXDOL]DWLRQ 1HZ
PAGE 119

6DUDVRQ ,* t *DQ]HU 9$Q[LHW\ UHLQIRUFHPHQW DQG H[SHULPHQWDO LQVWUXFWLRQV LQ D IUHH YHUEDOL]DWLRQ VLWXDWLRQ -RXUQDO RI $EQRUPDO DQG 6RFLDO 3V\FKRORJ\ 6DUDVRQ ,* t *DQ]HU 9(IIHFWV RI WHVW DQ[LHW\ DQG UHLQIRUFHPHQW KLVWRU\ RQ YHUEDO EHKDYLRU -RXUQDO RI $EQRUPDO DQG 6RFLDO 3V\FKRORJ\ BL 6DUDVRQ ,* 3HGHUVRQ $ t 1\PDQ %$ 7HVW DQ[LHW\ DQG WKH REVHUYDWLRQ RI PRGHOV -RXUQDO RI 3HUVRQDOLW\ 6DUDVRQ ,* t 6WRRSV 5 7HVW DQ[LHW\ DQG WKH SDVVDJH RI WLPH -RXUQDO RI &RQVXOWLQJ DQG &OLQLFDO 3V\FKRORJ\ f 6DUDVRQ 6% 'DYLGVRQ /LJKWKDOO ) t :DLWH 5 &ODVVURRP REVHUYDWLRQ RI KLJK DQG ORZDQ[LRXV FKLOGUHQ &KLOG 'HYHORSPHQW B 6DUDVRQ 6% 'DYLGVRQ /LJKWKDOO ) :DLWH 5 t 5XHEXVK %. $Q[LHW\ LQ HOHPHQWDU\ VFKRRO FKLOGUHQ 1HZ
PAGE 120

7D\ORU 5 3HUVRQDOLW\ WUDLWV DQG GLVFUHSDQW DFKLHYHPHQW $ UHYLHZ -RXUQDO RI &RXQVHOLQJ 3V\FKRORJ\ f§ 7RELDV 6 $Q[LHW\ DQG LQVWUXFWLRQ ,Q ,* 6DUDVRQ (Gf 7HVW DQ[LHW\ 7KHRU\ UHVHDUFK DQG DSSOLFDn WLRQV +LOOVGDOH 1 /DZUHQFH (DUO3DXP $VVRFLDWHV 7ROPDQ (& &RJQLWLYH PDSV LQ UDWV DQG PHQ 3V\FKRORJLFDO 5HYLHZ :DJDPDQ */ &RUPLHU :+ t &RUPLHU /6 &RJQLWLYH PRGLILFDWLRQ RI WHVWDQ[LRXV VWXGHQWV 3DSHU SUHVHQWHG DW WKH PHHWLQJ RI WKH $PHULFDQ (GXFDWLRQDO 5HVHDUFK $VVRFLDWLRQ :DVKLQJWRQ '& :DWVRQ t )ULHQG 5 0HDVXUHPHQW RI VRFLDOHYDOXDWLYH DQ[LHW\ -RXUQDO RI &RQVXOWLQJ DQG &OLQLFDO 3V\FKRORJ\ B n :HLQHU % $Q DWWULEXWLRQ WKHRU\ RI PRWLYDWLRQ DQG HPRWLRQ ,Q + .URKQH t / /DX[ (GVf $FKLHYHPHQW VWUHVV DQG DQ[LHW\ 1HZ
PAGE 121

%,2*5$3+,&$/ 6.(7&+ 5XWK /DXULH /D[ ZDV ERUQ LQ 6\UDFXVH 1HZ
PAGE 122

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nU AA Y YV a 6X]DQQH %HQQHWW -RKQVRQ $VVRFLDWH 3URIHVVRU RI &OLQLFDO 3V\FKRORJ\ 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 GLVVHUDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\ :DOWHU 5 &XQQLQJKDP RU RI 3V\FKRORJ\

PAGE 123

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

PAGE 124

81,9(56,7< 2) )/25,'$


TEST ANXIETY AND ACADEMIC SELF-EFFICACY
AS PREDICTORS OF COGNITIVE TASK PERFORMANCE
BY
RUTH LAURIE LAX
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
1982

This dissertation
is dedicated to my parents,
Judith and Melvin Lax.

ACKNOWLEDGEMENTS
I would like to thank my dissertation committee chairman,
Dr. Nathan W. Perry, Jr., for providing support and guidance
this year, and for graciously taking over as chairman, on
short notice, when my former chairman, Dr. Wiley C. Rasbury,
obtained a new position in Michigan. I also greatly appre¬
ciate the time that Dr. Rasbury put into the formulation
and planning stages of this dissertation. His assistance
was invaluable. My dissertation committee, Dr. Walter
Cunningham, Dr. Suzanne Bennett Johnson, Dr. Hugh C. Davis,
Jr., and Dr. Sandra Damico, provided many instructive
comments, advice, and support.
I am especially grateful for the many hours that Dr.
Randy Carter spent with me, reviewing statistical analyses,
and providing psychological support when the computer
programs began to grow rapidly in number. He always made
himself available, even when his schedule was extremely
tight, which it almost always was. I am also deeply grate¬
ful to the four undergraduate students who helped me with
data collection and coding. It was a mammoth task; and
Patty Bell, Erin Boyer, Vivian Montes, and Adam Rosenberg,
were hard working and enthusiastic throughout the entire
process. I also greatly appreciate the time and energy
iii

that the Registrar, Robert Fullington, and his staff, put
into obtaining aptitude scores for me.
I would also like to thank Gary Rushakoff, a special
friend, who taught me the ins and outs of text-editing on
the Apple Microcomputer. His technical assistance and
psychological support have been invaluable over the past
two and a half years.
Dr. Warren Rice was a great source of support. He
allowed me to use his Apple Microcomputer to write this
dissertation. He also helped make the work environment
at the Gainesville V.A. Medical Center very conducive to
speedy completion of my dissertation.
Finally, I would like to express my gratitude to my
family, who have persevered with me, and provided support
throughout my entire graduate career. It would have been
difficult without them.
IV

TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS iii
ABSTRACT vi
CHAPTER
ONE INTRODUCTION 1
Cognitive and Emotional Concomitants of
Test Anxiety 4
Treatment Studies 20
Self-efficacy Theory 36
TWO METHOD 61
Subjects 61
Procedure 61
Statistical Analysis 68
THREE RESULTS 70
FOUR DISCUSSION 80
APPENDICES
A ANAGRAMS TASK 101
B INTERCORRELATION MATRIX ON THE INDEPENDENT
AND DEPENDENT VARIABLES FOR THE MAIN
SAMPLE (n=16 2) 102
C INTERCORRELATION MATRIX ON THE INDEPENDENT
AND DEPENDENT VARIABLES FOR TOTAL SAMPLE 103
REFERENCES 104
BIOGRAPHICAL SKETCH 114
V

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
TEST ANXIETY AND ACADEMIC SELF-EFFICACY
AS PREDICTORS OF COGNITIVE TASK PERFORMANCE
By
Ruth Laurie Lax
August 1982
Chairman: Nathan W. Perry, Jr.
Major Department: Clinical Psychology
This study assessed the effectiveness of test anxiety,
academic self-efficacy, and aptitude as predictors of
cognitive task performance under neutral, evaluative, or
reassuring task instructions. It was hypothesized that
although test anxiety has been found to predict cognitive
task performance under reassuring and neutral task condi¬
tions, the variables of academic self-efficacy and aptitude
would enhance prediction. It was theorized that a person's
academic self-efficacy, or their belief in their ability
to cope with or master academic situations, would be
related to cognitive aspects of test anxiety (negative
self-statements, negative imagery). Academic self-efficacy
was, therefore, hypothesized to enhance prediction of
vx

cognitive task performance beyond knowledge of a person's
level of test anxiety, alone. In addition, it was hypothe¬
sized that aptitude would account for a significant amount
of variance in cognitive task performance.
It was found that test anxiety did predict cognitive
task performance, as measured by an anagrams task, and
final grade in General Psychology. However, test anxiety
predicted performance under all task instructional conditions,
rather than differentially, as predicted. Academic self-
efficacy predicted cognitive task performance, in inter¬
action with test anxiety, on the anagrams task, alone.
Aptitude scores predicted grade performance but not
anagram task performance. The results were considered as
speculative, because the predictor variables only accounted
for a small percentage of the variance in task performance.
The findings were discussed in terms of methodological
improvements in the academic self-efficacy measure, and in
terms of other possible relevant predictor variables.
vii

CHAPTER ONE
INTRODUCTION
It has been generally found that there is a correla¬
tion between aptitude measures (e.g. I.Q. measures,
Scholastic Aptitude Test) and academic performance in
elementary, secondary, and college populations (Plant &
Richardson, 1958; Fishman & Pasanella, 1960; Lavin, 1965;
Conry & Plant, 1965; Dudek, Goldberg, Lester, & Harris,
1969; Cronbach, 1970; McClelland, 1973). The correlations
tend to be somewhat higher in the elementary and high
school years, than in college or older populations (Brody &
Brody, 1976). The lower correlations for the relationship
between aptitude and grades in college and other select
populations, are most likely due to the restricted range
of aptitude test scores found in these populations.
While aptitude measures account for a large proportion
of the variance in cognitive performance, other factors
are considered important, especially for populations
exhibiting a narrow range of aptitude scores. These factors
generally fall into the domains of "personality" or
"cognitive style" or motivational factors. One of the most
frequently cited of these factors is test anxiety. Experi¬
mental research has demonstrated that highly test-anxious
1

2
individuals perform most poorly on cognitive tasks under
conditions of high evaluative stress, and optimally when
evaluation is minimal. The reverse tends to be true for
low test-anxious individuals (Mandler & Sarason, 1952;
Sarason, Mandler, & Craighill, 1952; Ganzer, 1968; Paul &
Eriksen, 1964; Russell & Sarason, 1965; Sarason, 1958,
1959, 1961). Whereas test anxiety has been found to be
related to a cognitive performance decrement, the historically
popular explanation for the cause of this decrement in
performance has been autonomic reactivity, or what Liebert
and Morris (1967) have termed "emotionality." Recent
research suggested that cognitive components of test
anxiety, rather than autonomic reactivity, may have greater
predictive validity in regard to cognitive performance
under stressful evaluative circumstances (Neale & Katahn,
1968; Sarason & Stoops, 1978; Deffenbacher, 1978; Dusek,
Mergler, & Kermis, 1976; Denney, 1980). These cognitive
components include
a) worry over performance, and comparison with
how well others are doing.
b) rumination over alternatives.
c) being preoccupied with such things as
feelings of inadequacy, anticipation of
punishment, and loss of status and esteem
(Meichenbaum, 1972; Mandler & Watson, 1966;
Marlett & Watson, 1968).

3
While earlier definitions of test anxiety had empha¬
sized a common denominator of "anxiety," or emotional or
autonomic reactivity (Wine, 1971), this common denominator
does not appear to capture the most outstanding differences
between persons who score at extremes on test anxiety
measures. These differences appear to lie in the nature
of cognitive structures (Wine, 1980). There is a change
in the thrust of current test anxiety research toward
cognitive interpretations of test anxiety. This disserta¬
tion study examines the relationship of test anxiety to
other cognitive process variables, in an effort to delineate
more specifically, the cognitive processes underlying
performance decrement in test anxiety.
Test anxiety research initially focused (Sarason, 1961)
on emotional reactivity as the mediating variable in the
relationship of test anxiety to academic performance. More
recent research has suggested that test anxiety has a
strong cognitive component, as well. It has been proposed
that this cognitive component accounts for much of the
variance in task performance (Wine, 1980). This literature
review examines the contribution of cognitive and emotional
factors in test anxiety, as both relate to cognitive task
performance. Experimental studies will be discussed from
several diverse areas of research, which bear upon this
issue. In addition, the treatment literature in test
anxiety will be examined. The treatment literature is

4
relevant to an understanding of the relative contribution
of cognitive and emotional components, to test anxiety.
Treatment studies have utilized either cognitive, or
emotionality reduction approaches, or combinations of
these approaches, to attempt to alleviate test anxiety.
An examination of the effectiveness of these various
treatment modalities will provide information as to the
underlying mediating variables in test anxiety.
Cognitive and Emotional
Concomitants of Test Anxiety
"Direction of Attention" Hypothesis
Wine (1980, 1971) has developed a cognitive-attentional
theory of test or "evaluational" anxiety, the "direction
of attention" hypothesis. She has proposed that the
performance differences of low and high test-anxious
individuals reflect basic differences in their task atten¬
tion during evaluative tasks. The high test-anxious
individual is hypothesized to divide his attention between
self-preoccupied worry and task cues. The low test-anxious
individual is presumably focused more fully on the task¬
relevant cues. The high test-anxious person, thus, has a
tendency to direct his attention internally rather than to
the task. He focuses on self-evaluative and self-deprecatory

5
thinking. There is a significant amount of evidence from
the experimental and treatment literature which suggests
that a cognitive-attentional approach to test anxiety is
important in understanding the construct, perhaps more .so
than a traditional autonomic arousal interpretation of
test anxiety (Wine, 1971; Sarason, 1975; Denney, 1980).
Worry-Emotionality Distinction
A closely related theoretical development to Wine's
(1971) "direction of attention" hypothesis, is the Worry-
Emotionality distinction developed by Liebert and Morris
(1967). These authors have analyzed test anxiety into the
two components of 1) Worry, which is the focusing of
attention on concerns about performance, consequences,
negative self-evaluation, evaluation of one's ability
relative to others, and, 2) Emotionality, which refers
to the affective-physiological experience generated from
increased autonomic arousal. Interfering test anxiety,
therefore, takes the form of heightened worry and/or
emotionality.
Two separate self-report scales have been developed
to measure the Worry and Emotionality components of test
anxiety. One is the ten item, Worry-Emotionality Inventory
(Liebert & Morris, 1967). The other scale is the 16 item

6
Inventory of Test Anxiety (Osterhouse, 1972). Worry items
include a) not feeling confident about performance,
b) worrying a great deal, c) thinking how much brighter
others are, d) thinking about the consequences of failure,
e) feeling not as prepared as possible. Emotionality
items include a) so nervous cannot remember facts,
b) heart beating faster, c) upset stomach, d) uneasy,
upset feeling, e) feeling anxiety.
Worry and Emotionality scores on the Worry-Emotionality
Inventory have been related to performance expectations
regarding examinations, and actual examination performance.
Several studies have shown that Worry was inversely related
to performance expectations of college and high school
students taking classroom exams (Doctor & Altman, 1969;
Liebert & Morris, 1967; Morris & Liebert, 1970). Perfor¬
mance expectancy was measured by having subjects assess
how well they expected to perform on a particular test.
Whereas Worry was consistently related to performance
expectations, Emotionality was inversely related to
performance expectations in some samples (Doctor & Altman,
1969; Morris & Liebert, 1970; Spiegler, Morris, & Liebert,
1968), but unrelated in others (Liebert & Morris, 1967;
Morris & Liebert, 1970; Spiegler, Morris, & Liebert,
1968). In all but one of the samples in the Morris and
Liebert (1970) studies, Worry correlated more negatively
with performance expectations than Emotionality.

7
A similar pattern exists for the relation of Worry
and Emotionality to actual performance on exams. Morris
and Liebert (1970) reported that in a college sample, Worry,
but not Emotionality, formed a significant inverse relation¬
ship with test performance, and the Worry-performance
relationship was stronger than the Emotionality-performance
relationship. In the high school sample, however, both
Worry and Emotionality were inversely related to performance,
and the strengths of the relationships did not differ. In
both age samples, Worry and Emotionality were significantly
correlated with each other, and consequently shared
considerable variance. Partial correlation analyses
revealed, however, that for both samples, when the common
variance between Worry and Emotionality was controlled,
Worry was negatively correlated with test grades, whereas
Emotionality was no longer significantly related to test
performance.
In five unpublished studies involving college students,
Deffenbacher (1980) found results similar to those of
the Morris and Liebert (1970) college sample. Worry was
consistently negatively related to exam performance, while
Emotionality related on some occasions and not on others.
Again, partial correlations demonstrated that when the
effects of Emotionality were partialled out, Worry continued
to form a significant negative correlation with performance.
However, when Worry was partialled out, Emotionality was

8
not significantly correlated with performance. This
suggests that when the common variance is removed, Worry,
but not Emotionality, is significantly and negatively
correlated with performance on college exams.
Deffenbacher (1977) also investigated the relation¬
ship of Worry and Emotionality to performance on the Miller
Analogies Test (MAT), a test which is important because it
is used for graduate admissions, but for which little
preparation is possible (i.e. there is no specific knowledge
that some people may have access to over others, in regard
to performance on the task). Deffenbacher found that both
Worry and Emotionality were negatively related to performance
on the MAT. When partial correlations were calculated, as
in the previous studies, only Worry continued to form a
significant relationship with performance on the MAT.
Further analysis revealed a complex relationship among
Worry, Emotionality, and performance. Worry was broadly
related to performance. Subjects high on Worry solved fewer
analogies than subjects low on Worry. The effects of
Emotionality varied with Worry level. At low levels of
Worry, Emotionality was unrelated to performance, but at
high levels of Worry, high Emotionality was debilitative.
That is, the negative effects of Emotionality were nested
within the upper range of Worry, while Worry contributed
more pervasively to the relationship of test anxiety and
performance.

9
Morris and Liebert (1969) employed Worry and Emotionality
in a design which manipulated level of stress through obvious
and non-obvious timing and difficulty of items for the timed
subtests of the Wechsler Adult Intelligence Scale. Worry
level was found to interact with timing and difficulty level.
Subjects who were high on Worry performed better in the
untimed than the timed condition. The reverse was true
for subjects low on Worry. They performed better on the
timed condition. Emotionality level did not interact with
difficulty or timing conditions and was unrelated to per¬
formance in this study.
Finally, Morris and Perez (1972) found that Worry, but
not Emotionality, was significantly negatively related to
reading test performance. Interestingly, although the
correlation was not significant, Emotionality was positively
correlated with performance.
It appears that a fairly consistent pattern of results
emerges from the research of the Liebert and Morris group.
Worry and Emotionality are significantly and consistently
correlated (correlations range from .55 to .76). However,
it is the Worry component which consistently forms a nega¬
tive relationship with performance expectations and actual
test performance, not Emotionality. The findings for
Emotionality are much more inconsistent and mixed. Emotion¬
ality was unrelated, related only within certain strata of
Worry, or negatively related to performance measures.

10
Therefore, for both performance measures and performance
expectancies, Emotionality is not as consistently related
to outcome measures, as Worry.
As Morris and Liebert (1970) point out, the failure
to find consistent differences for Emotionality, among
performance expectancy groups, may reflect a generally high
stress level for all groups. These studies used actual
examinations as the tasks to rate expectancy against. This
situation may have provided subjects with greater stress
than less familiar tasks of lesser importance. The high
stress level of the tasks in these studies may have obscured
differences among expectancy levels, which might have been
seen with less immediately relevant tasks.
It is also conceivable that autonomic symptoms are
only relevant to performance to the extent that the subject
attends to them. That is, if the Worry level is high, the
subject attends to his affective reaction. If Worry is
not present, the affective reaction is not attended to.
Thus Emotionality would only be related to performance
under high Worry levels.
The Emotionality variable in the Morris and Liebert
(1969, 1970) studies refers to self-report of affective reac¬
tivity. It is important to examine the relationship between
self-report of emotionality, and actual physiological indices
of emotional reactivity. Holroyd, Westbrook, Wolf, and
Badhorn (1978) attempted to assess the relationship between

11
self-report of anxious arousal, physiological indices of
emotional arousal, and performance measures, in high and
low test-anxious subjects. The Holroyd et al. (1978)
study found that high test-anxious individuals performed
more poorly and reported higher levels of anxious arousal
and worry in an analog testing situation, than did low
test-anxious individuals. This is consistent with previous
findings. During the testing situation, actual autonomic
arousal was monitored continuously. Interestingly, results
from these measures indicated that obtained differences in
performance were not accompanied by corresponding difference
in autonomic reactivity. High and low test-anxious
individuals manifested virtually identical tonic and phasic
electrodermal activity and heart rate responses. Only heart
rate variability, which appeared to reflect differences
in the cognitive and attentional responses of the test-
anxious groups, successfully differentiated high and low
test-anxious individuals. The results lend further support
to the theory that the cognitive performance deficits of
high test-anxious subjects are not a result of maladaptive
levels of autonomic arousal. Rather, it appears that test
anxiety may be more accurately conceptualized as a cognitive-
attentional phenomenon, as has been proposed by Wine (1971).

12
Attention to Task Cues
In an effort to identify cognitive-attentional factors
related to test anxiety, a group of researchers is examining
the utilization of task cues by low and high test-anxious
subjects. The purpose of this research is to assess whether
high and low test-anxious subjects differentially attend to
task-relevant stimuli.
Quality of performance on a task depends on the extent
to which task-relevant stimuli are attended to and utilized.
If relevant stimuli are not attended to, then performance
will be impaired.
Current research in the area of test anxiety and task
cue utilization has been generated from the earlier motiva¬
tional or drive theory of Tolman (1948). The influence of
motivation on cue utilization was recognized by Tolman
(1948) . He listed "strong motivational conditions" as one
of several causes of the formation of narrow cognitive maps
in animals and humans. Bartlett (1950) introduced the
concept of "range of cues utilized" to examine the effects
of motivation on task performance. The early research
culminated in an influential paper by Easterbrook (1959).
The main point of the Easterbrook paper (1959) was that
"when direction of behavior is constant, increase in drive
(in this case anxiety drive) is associated with a reduction
in the range of cue use." One implication of Easterbrook's

13
analysis is that drive may either facilitate or inhibit
performance, depending on the importance of peripheral cues
for successful performance. Although there has not been much
research generated in this area, the initial studies seem
to indicate that test anxiety is related to narrowed cue
utilization. Interestingly, the low test-anxious subjects
appear to be more affected by the presence of added task cues
than high test-anxious subjects. The performance of the low
test-anxious person is facilitated more by the addition of
relevant cues, and debilitated more by the addition of irrele¬
vant cues. That is, the performance of low test-anxious
subjects reflects a wider range of task cue utilization,
in accordance with the Easterbrook hypothesis (Wine, 1980).
West, Lee, and Anderson (1969) tested children who had scored
high or low on the Test Anxiety Scale for Children (Sarason,
Davidson, Lighthall, Waite, & Ruebush, 1960) with forty
arithmetic "story" problems constructed in such a way that
some contained additional irrelevant information. Although
the highly test-anxious subjects showed a slightly poorer
level of performance overall than the less test-anxious
group, their performance was not hindered by the addition
of irrelevant information. The low test-anxious subjects
performed more efficiently on the unembellished items than
on the ones containing irrelevant cues. Geen (1976) found
further evidence of test anxiety as a source of restricted
cue utilization in an experiment involving serial learning.

14
The task required the subject to remember which of a set of
several stimuli was the one designated correct on each of
15 trials. In two conditions, additional cues were added
to the task. In one case, the cues aided solution of the
task. In the other case, the cues were irrelevant and
potentially distracting to task solution. High test-anxious
subjects were helped less by the addition of the relevant
information than the low test-anxious subjects. They were
also less hindered in their recall, by the insertion of
irrelevant information. In both cases, subjects who scored
high in test anxiety, behaved as if they were less affected
by the additional cues than those low in test anxiety. This
finding is consistent with an assumption of a restricted
range of cue utilization. Whatever the underlying mechanism
of the narrowed cue utilization, this restriction in cue
use appears to represent a shift in attentional bias or
priorities for the high test-anxious subjects. Wine (1980)
and Geen (1980) propose that the "primary task" of the high
test-anxious subject is not the assigned task. Wine describes
the situation in terms of diversion of attention away from the
task and toward the self. Concentration of attention inwardly
toward the self draws attention away from task-related
stimuli. Hamilton (1975) has proposed an interesting model
of attention related to Wine's hypothesis:
Whenever a person's priorities (from the avail¬
able limited store) give precedence to the
demands of "enduring dispositions," less spare

15
capacity is left over for demands of the
task. The person may respond by paying less
attention to external, task-related inputs,
either by narrowing the area focussed for
selective attention or by restricting atten¬
tion to a small number of foci. . . .
(Hamilton, 1975, pp. 54-55)
These cue utilization studies support the proposition
that the high test-anxious person does not attend to relevant
task cues, as much as the low test-anxious person. Further
research will be necessary to determine whether the attention
of the high test-anxious person is diverted to internal
concerns. The cue utilization research does support the view
that cognitive factors are relevant to an understanding
of the construct of test anxiety. This should prove to be
a fruitful area of research in furthering understanding of
cognitive processing in test-anxious populations.
Social-Evaluative Variables
While the high test-anxious subject may not be entirely
attentive to task-relevant cues, he may be more attentive to
social-evaluative cues than the low test-anxious subject
(Wine, 1980; Dusek, 1980). There is some evidence to support
this notion in experimental modeling research. Most of this
research has been conducted by Sarason and his colleagues
(Sarason, 1968, 1972, 1973a, 1975; and Sarason, Pederson, &
Nyman, 1968). Sarason, Pederson, and Nyman (1968) reasoned
that

16
test anxiety may contribute to increased
vigilance with regard to possibly helpful
cues in the environment . . . because high
test-anxious individuals are more insecure
than other persons, they may be relatively
more interested and active in attending to
cues in strange situations and in "borrowing"
behaviors and attitudes from models.
(Sarason et al., 1968, pp. 496-497)
In an initial study, Sarason et al. (1968) studied the
effects of modeling and test anxiety on learning paired
associates. There was a significant anxiety by conditions
interaction. The high and middle test anxiety groups demon¬
strated superior performance on the task, in the situations
where the subjects could observe models learning the lists.
Subsequent studies explored the effects of character¬
istics of models' problem-solving behaviors, and feedback
to models, on the cognitive performance of high and low
test-anxious subjects. Sarason (1973a) had subjects solve
difficult anagrams under one of several experimental modeling
conditions. One group observed a model silently solve sample
problems. Another observed a model describing the manipula¬
tion of letters while solving the anagrams. A third group
observed a model comment on principles underlying the success¬
ful solution of anagrams while working. There was also a
no-model control condition. While overall, the low test-
anxious subjects performed better than the high test-anxious
subjects, there was an anxiety by experimental condition
interaction. The high test-anxious subjects were superior
in performance to the low test-anxious subjects under the

17
condition in which the model stated problem-solving principles.
Although it is not clear why the task performance of the
high test-anxious subject is generally inferior to the low
test-anxious subject , it may be, in part, because of an
actual or assumed paucity of task-relevant responses. The
modeling situation which provided cognitive strategies may
have given the high test-anxious subjects enough information
so that they were confident enough to become less self-
preoccupied, and more able to guide their own task-relevant
behavior.
While these studies suggest that the observation of
a model may be beneficial to the test-anxious subject, other
studies which have examined the effects of exposure to
models who succeed, fail, or are self-derogatory are some¬
what more difficult to interpret. Sarason (1972) investigated
whether persons differing in test anxiety would respond
differentially to a neutral model vs. a model who is self¬
derogatory regarding her performance. It was expected that
the opportunity to observe a model would have a facilitative
effect on all subjects, especially those high in test anxiety.
However, it was expected that the high test-anxious group
would show this facilitative effect to a lesser extent under
the condition involving the self-derogatory model. While
there was, overall, a facilitative effect for modeling,
exposure to the self-derogatory model did not differentially
affect the different anxiety groups. Post hoc review of the

18
self-derogatory condition led to the observation that while
the model had derogated herself, her performance had not been
evaluated by an authority figure to be a failure. There
might be a difference between observing someone who is self-
debasing and thinks he is failing, vs. someone who is told
by an authority that he has failed. Consequently, another
experiment was conducted including success and failure
conditions determined by an authority figure. There was a
test anxiety by treatment interaction. For the condition in
which the model did not fail (given success feedback by an
authority), anxious subjects' performance was superior to
that of a middle and low test anxiety group. For the failure
condition (failed by an authority figure), the high test-
anxious group performed at a significantly lower level than
did the high test anxiety group that was exposed to the
nonfailing model. The low test anxiety group performed
better under the failed model condition than the non-failed
model condition. The results of the two experiments suggest
that observing someone derogating himself vs. observing
someone who is declared by an authority figure to be a
failure can have very different effects on performance for
the high test-anxious subject. The self-derogatory condition
and the authority-failure condition were both apparently
motivating conditions for the low test-anxious subjects,
who performed best under both of these conditions. Performance
was quite different for the high test-anxious subjects. The

19
observation of failure feedback had a much more negative
effect, on this group, than the observation of a self¬
derogatory model. Sarason suggests that observing someone
who is dissatisfied with himself (self-derogatory condition),
may not be convincing to an observer, since there may be
no objective basis for the self-debasing reaction. However,
the observation of a negative authority figure is much more
convincing, and certainly would affect the evaluational
anxiety of the high test-anxious subject. Another explana¬
tion for the lack of detrimental impact of the self-derogatory
model lies in the construct of the "coping model," described
by Meichenbaum (1972). The self-debasing model may have been
seen by the high test-anxious subject as someone who exper¬
iences problems similar to those of the subject, but who
actively copes with and overcomes those difficulties.
Failure feedback, however, is clearcut evidence of a failure
in coping efforts (Wine, 1980). Sarason (1975) examined
whether a coping model was, in fact, facilitative of cogni¬
tive task performance. Subjects were exposed to a model
who admitted experiencing anxiety and described ways of
coping with it, or to one who only admitted to the anxiety,
versus a model who did not admit to any anxiety. Exposure
to the self-disclosing coping model was highly facilitative
of the test-anxious subjects' cognitive task performance.
The results of these studies suggest, in general, that
exposure to models who are task-oriented, and who provide

20
attention-directing cognitive strategies, is beneficial to
the performance of test-anxious persons. Performance is
also enhanced if there is evidence that the model is
successfully coping with the worry and tension associated
with test anxiety. The observation of models performing
tasks does seem to have powerful effects on the subsequent
task performance of high test-anxious subjects. The social-
evaluative and problem-solving information can have either
positive or negative effects depending on the characteristics
of the modeling situation.
These studies suggest that the high test-anxious
subject may attend to cues in the task situation other than
the immediate task cues. Social-evaluative feedback appears
to influence the cognitive task performance of the high test-
anxious subject. This is further evidence for a cognitive-
attentional component to test anxiety.
Treatment Studies
In an effort to further examine the contribution of
cognitive and emotional factors to test anxiety, the treatment
research in test anxiety will be examined. Treatment approaches
have focused primarily on either emotionality reduction,
cognitive coping strategies, or a combination of these
approaches. It is, thus, possible to examine the relative

21
effects of these approaches on both cognitive task perfor¬
mance and self-report of test anxiety.
Systematic Desensitization
Most of the early treatment literature assumed
emotional reactivity to be the major characteristic of test
anxiety. As such, much of the early treatment literature
focused on systematic desensitization, progressing from
simple demonstrations of the efficacy of systematic desensi¬
tization (Katahn, Strenger, & Cherry, 1966; Garlington &
Cotier, 1968; Paul, 1964) through systematic manipulations
of the desensitization process (e.g. Freeling & Shemberg,
1970). Systematic desensitization treatment for test
anxiety basically involves the pairing of deep muscle relax¬
ation with the instructed imagination of aversive scenes,
to inhibit autonomic arousal and provide a new response,
to be conditioned to previously anxiety-arousing stimuli.
The systematic desensitization procedure followed the general
outline used by Paul (1966) in one of the earliest demon¬
strations of this treatment for text anxiety. Studies
compared individually-designed vs. standardized desensiti¬
zation hierarchies (Emery & Krumboltz, 1967), systematic
desensitization vs. relaxation therapy (Johnson & Secrest,
1968), automated vs. live-therapist desensitization (Donner

22
St Guerney, 1969) , and other variations of presenting the
systematic desensitization treatment (Aponte & Aponte, 1971;
Freeling & Shemberg, 1970; Suinn, 1968). In general, these
studies have found that systematic desensitization is effec¬
tive in alleviating test anxiety when self-report measures
of anxiety level are used as outcome criteria. However, the
effects of desensitization on academic performance are not
quite as consistent. Some studies demonstrate an increase
in grade point average for treated groups compared to
control groups (Aponte & Aponte, 1971; Donner & Guerney,
1969; Johnson & Secrest, 1968; and Katahn, Strenger, &
Cherry, 1966). Others showed no increment in academic
performance (Garlington & Cotier, 1968; Emery & Krumboltz,
1967) or did not use performance measures as outcome
criteria (Freeling & Shemberg, 1970; Suinn, 1968). Allen
(1972) states the fact that the desensitization studies
did not control for the effects of non-specific therapeutic
factors, which might enhance treatment benefits in some
studies and not others. These non-specific factors include
heightened subject expectancy, the use of only one therapist,
the use of first semester freshmen (have not yet adapted to
college routine and may interpret their stress as test
anxiety), and the lack of appropriate control groups
(e.g. waiting list control group).
The systematic desensitization research does demonstrate
that an emotionality reduction technique can effect changes

23
in self-report of anxiety level. However, an emotionality-
oriented treatment approach may not be sufficient to con¬
sistently alter poor performance under academically-evaluative
conditions. More extensive, methodologically-sound research
on the effects of systematic desensitization treatments on
cognitive performance, is required, to make more definitive
statements.
Self-Control Treatment Techniques
Denney (1980) has reviewed a series of self-control
treatment studies. He orders the studies on a continuum
on which one extreme represents approaches focused solely
on emotionality reduction, and the other extreme represents
approaches emphasizing a totally cognitive approach. The
labels for the three basic types of studies reviewed are
the following: applied relaxation, self-control training
techniques, and cognitive coping techniques. As one moves
from treatment approaches relying heavily on emotionality
reduction (applied relaxation) to more cognitively oriented
treatments (self-control training and cognitive coping),
there is increasing emphasis on in-vivo application training
(training in the actual anxiety-provoking setting), guided
rehearsal, and cognitive restructuring. All of these
approaches emphasize the attainment of coping skills which

24
the client can effectively apply toward the management and
reduction of test anxiety when it arises in real life
situations. The more cognitively-oriented treatment approaches
attend to thoughts, worries, and negative self-statements,
rather than alleviation of emotional arousal.
Applied Relaxation Techniques
Applied relaxation techniques constitute the simplest
examples of self-control procedures for the reduction of
anxiety. The procedures classified as applied relaxation
techniques share three common features: 1) These procedures
are introduced with a self-control rationale similar to the
one advocated by Goldfried (1971). Clients are usually
informed that the purpose of the treatment is to provide
them with skills for actively coping with anxiety. They
are told that the relaxation training they receive is a
method for helping them to bring the response of relaxation
under voluntary control. Further, they are told, that with
greater proficiency at inducing relaxation they will be
able to apply relaxation to reduce feelings of anxiety in
stressful situations encountered outside of treatment.
2) The second feature of the applied relaxation techniques
is training in the induction of relaxation. Progressive
muscle relaxation exercises (Rimm & Masters, 1974) adapted
from Jacobsen's (1938) work, are the most common method

25
for inducing relaxation, although numerous other approaches
have been developed, supplementing relaxation with biofeed¬
back-assistance (Romano & Cabianca, 1978), cue-controlled
relaxation (Russell & Sipich, 1973), and imagery exercises
(Samuels & Samuels, 1975) , and other supplementary exercises.
3) The third feature of applied relaxation techniques involves
training in the application of relaxation within stressful
settings encountered outside of treatment. Typically, in
these studies, the clients are merely instructed to begin
applying relaxation outside of treatment.
Self-control Training Techniques
Self-control training techniques include the three
main features found in applied relaxation techniques.
However, in addition, all of the self-control training
techniques include guided rehearsal, which is introduced
after relaxation induction training. During guided rehearsal,
clients are confronted with some type of stressful stimulus
presented within the consultative setting. Usually the
stressful stimuli are evoked by the use of imagery. Thus,
there is greater personal involvement in cognitive
approaches.

26
Cognitive Coping Techniques
Cognitive coping techniques, which are the most
cognitively-oriented approaches, attempt to provide clients
with an additional repertoire of coping skills beyond, or
in some cases instead of, relaxation skills. The distinc¬
tive feature of the cognitive coping techniques is some
form of cognitive restructuring. Cognitive restructuring
has three basic objectives: 1) Clients are persuaded that
the beliefs that they entertain while confronting certain
situations affect the emotional reactions they have in these
situations, 2) They are, next, encouraged to identify the
particular negative self-statements that they make when
confronting a stressful situation, 3) Clients are helped
to formulate positive and rational self-statements which
will replace the negative self-statements and thereby
attenuate their stress in the evaluative situation.
As discussed earlier in regard to the systematic
desensitization studies, most of the applied relaxation
studies (Russell, Miller, & June, 1975; Russell, Wise, &
Stratoudakis, 1976; Marchetti, McGlynn, & Patterson, 1977;
and Chiang-Liang, & Denney, 1976) found that this treatment
approach demonstrated changes on self-report measures of
test-anxiety. The same generalization could be applied
to the results of the self-control training techniques
(Deffenbacher, Mathis, & Michaels, 1979; Zemore, 1975;

27
Deffenbacher & Parks, 1979; Denney & Rupert, 1977), and the
cognitive coping techniques (Meichenbaum, 1972; Holroyd,
1976; Lavigne, 1974; Wine, 1970; Goldfried, Linehan, &
Smith, 1978). Analyses of the comparative impact of these
treatment strategies on cognitive performance demonstrates
quite different findings. Only one out of three of the
applied relaxation studies resulted in cognitive performance
improvement. It was found that four of the eight self-control
training studies reviewed, revealed improvements on perfor¬
mance measures, whereas five of the seven cognitive-coping
technique studies demonstrated improvements in these types
of measures. Thus, the cognitive coping techniques, which
incorporate some form of cognitive restructuring, are
observed to have a higher rate of success than self-control
training or applied relaxation techniques, in terms of
performance improvement.
Both Meichenbaum (1972) and Holroyd (1976) compared
cognitive coping treatment approaches directly with systematic
desensitization, which is an arousal reduction technique.
Meichenbaum's "cognitive modification" treatment procedure
combined an insight-oriented therapy which was designed to
make test-anxious subjects aware of their anxiety-engendering
thoughts, with a modified desensitization procedure which
employed a) coping imagery on how to handle anxiety, and
b) self-instructional training to attend to the task and
not ruminate about oneself. Results of the study indicated

28
that the cognitive modification group was most effective in
significantly reducing test anxiety as assessed by a) test
performance obtained in an analogue test situation, b) self-
report of anxiety immediately after treatment and at a
one-month follow-up, and c) grade point average. There was
less consistent, but general improvement by the desensitiza¬
tion treatment group, who appeared significantly more improved
than the waiting list controls.
The Holroyd (1976) study is actually a follow-up study
on the Meichenbaum (1972) study. Meichenbaum's results are
somewhat difficult to interpret because his cognitive
modification treatment is actually a combination of a
cognitive treatment (the use of self-statements and cognitive
"coping" imagery) and an emotional arousal treatment (the
use of relaxation). It is not clear whether the positive
results of the cognitive modification treatment are primarily
due to the cognitive aspects, the arousal reduction aspects,
or the combination of the two treatment forms. Holroyd
sought to separate out these individual treatment components.
He compared the effectiveness of cognitive, arousal reduction,
and combined arousal-cognitive treatments for test anxiety.
The cognitive therapy focused exclusively on eliminating the
test-anxious client's task-irrelevant ruminations and
attentional focus. The clients were taught to emit incom¬
patible self-statements designed to facilitate task-attention
in stressful situations. The emotional arousal treatment

29
was the traditional group systematic desensitization
paradigm (Paul, 1966). The combined treatment focused on
both eliminating the test-anxious client's task-irrelevant
ruminations and task focus, and reducing their emotional
arousal. There was also a pseudotherapy control group
included in the study, to assess the degree of improvement
attributed to nonspecific treatment effects.
The results demonstrated that the cognitive therapy
approach was significantly more effective than the other
treatment approaches. The treatment gains were seen in both
laboratory measures of test anxiety (self-report), test
performance in an analog test situation, and grade point
average. Only the cognitive therapy approach demonstrated
significantly better results than the pseudotherapy control
procedure. The systematic desensitization group, the
combined treatment group, and the pseudotherapy groups
produced similar outcomes, except for the analog test
situation (digit symbol test performance), where the
combined treatment approach was superior to the systematic
desensitization and the pseudotherapy approach. The results
suggest that while systematic desensitization is an effective
treatment for test anxiety, it may be no more effective than
an equally convincing treatment ritual of any other sort.
The results for the combined treatment approach are somewhat
puzzling. Other studies using combined treatments have
yielded conflicting results, with the combined treatment

30
being less effective than cognitive or systematic desensi¬
tization treatments alone (Meichenbaum, Gilmore, & Fedoravicius,
1971), or more effective than systematic desensitization
alone (Meichenbaum, 1972). Holroyd (1976) suggests that the
results obtained with combined treatments may be influenced
by procedural variables such as length of treatments, order
of treatments, and the specific treatments which are combined.
There have been other cognitive coping approaches
which have proven to be successful in improving academic
performance and reducing self-report of test anxiety.
Recently, Harris and Johnson (1980) have compared a unique
covert modeling approach, to a self-control desensitization
treatment, and a study-skills training group (with the
covert modeling group and the desensitization group receiving
some study skills training, as well). Harris and Johnson
(1980) hypothesized that an individual would be likely
to benefit more from a modeling procedure, if the model
performing the behavior is the individual himself. There¬
fore they developed a treatment in which the individual
utilizes his own highly personalized images of self-competency
in stressful situations to cope with his test anxiety. These
images are taken from life situations other than test-taking,
in which the individual operates successfully, for example,
playing in a music recital or running a radio show. After
the individual develops his own unique "success" imagery,
these images are paired with anxiety-eliciting images taken

31
from the test-taking situation. This procedure allows the
individual to utilize his own resources, rather than having
a standardized "coping skills" package imposed upon him.
Harris and Johnson found, similar to previous studies comparing
cognitive approaches to desensitization, that the covert
modeling procedure was superior to both the self-control
desensitization paradigm and the study-skills training group.
The individualized covert modeling procedure combined with
study skills training significantly reduced test-anxiety,
as assessed by several self-report measures. It also
significantly increased academic performance, as assessed
by grade point average. The self-control desensitization
treatment was also effective in reducing self-reported test
anxiety, but was not successful in improving academic per¬
formance. Subjects in the study skills training group had
a small, but nonsignificant increase in grade point average,
but did not show reduction in test-anxiety on all of the
anxiety measures. Although the study skills training
procedure presumably prevented a decrement in academic
performance, it was not effective in either reducing test-
anxiety or significantly increasing grade point average.
The Harris and Johnson (1980) study clearly indicates that
a cognitive coping approach to treatment in which positive
imagery supplants negative anxiety-provoking imagery, is
highly effective in altering both self-report of test
anxiety and, importantly, academic performance. This

32
treatment also serves to highlight to the individual his own
areas of competence. Reinforcement of a sense of mastery
appears to be a more effective treatment than muscular
relaxation.
Harris (1980) investigated the concept of individualized
coping imagery in more detail in a follow-up study. She was
interested in the effectiveness of different imagery instruc¬
tion procedures in the treatment of test anxiety. Specifically
Harris compared individualized coping imagery treatment
based on non-academic experiences of competence and success
to individualized coping imagery treatment based on academic
experiences of competence and success. Since the images of
competence utilized in the Harris and Johnson (1980) study
were non-academic in nature, Harris hypothesized that
instruction in coping images based on previous academic
success experiences might prove even more effective in
reducing test anxiety and improving academic performance.
In addition, Harris wanted to assess the extent to which
effective relaxation training would enhance the overall
treatment effectiveness of the individualized coping imagery
techniques. Consequently, she compared four treatment
techniques: 1) non-academic individualized coping imagery
with relaxation training, 2) non-academic individualized
coping imagery without relaxation training, 3) academic
coping imagery with relaxation training, and 4) academic
coping imagery without relaxation training. The study found
that only the academic coping imagery group combined v/ith

33
relaxation and the individualized coping imagery group
without relaxation were significantly different from the
control group in academic performance (grade point average)
at post-test. Pre-test to post-test treatment changes in
grade point average were significant for both non-academic
imagery groups (p < .05) and approached significance for
the academic coping imagery group with relaxation (p < .089)
The academic coping imagery treatment without relaxation
did not show a significant pre-treatment to post-treatment
improvement in grade point average. All of the variations
of individualized coping imagery manifested significant
decreases in test anxiety on self-report measures. It
is apparent from the results of this study, that the type
of imagery instruction used in treatment has a bearing on
treatment effectiveness, particularly in regard to academic
performance improvement. The academic imagery condition
without relaxation proved not to enhance treatment effective
ness. Harris (1980) hypothesized that the academic coping
imagery treatment without relaxation was a less effective
procedure because academic images were likely to elicit
fear and avoidance rather than images and feelings of
competence in these subjects, since academic experiences
are so closely linked to the test-taking situation. Con¬
sequently, they had more difficulty evoking images of
competence regarding academic experiences. The addition
of relaxation to the academic coping imagery paradigm may

34
have allowed the subjects to be more comfortable about
academic imagery, since relaxation training has been
hypothesized to reduce distracting thoughts and physiolog¬
ical responses (Lang, 1979)• Hence, the academic coping
imagery with relaxation treatment was more efficacious
than the non-relaxation group, and the desensitization
hierarchy was completed more efficiently. The relaxation
component did not significantly add to the effectiveness
of the non-academic individualized coping imagery paradigm,
presumably because the subjects were initially more able
to visualize non-academic imagery without fear or avoidance.
The Harris (1980) study and the Harris and Johnson
(1980) study highlight the findings of previous research,
that while both emotionality and cognitively-oriented
treatments are effective in improving self-report of test
anxiety, it is the cognitive treatment approaches and
multimodal approaches which alleviate the cognitive
performance deficits of the test-anxious person. Further¬
more, the Harris (1980) study indicates that multimodal
approaches may not be necessary, if a powerful cognitive
coping model is applied, such as the non-academic
individualized coping imagery paradigm utilized in that
s tudy.

35
General Conclusions
While most of the early research in the area of test
anxiety was focused upon the debilitating effects of test
anxiety (Wine, 1971), current research has begun to investi¬
gate the contribution of cognitive and affective variables
in understanding the construct of test anxiety. Research
in the areas of task cue utilization, social-evaluative
cues, worry-emotionality distinction, and treatment approaches
to test anxiety, has demonstrated that cognitive factors
appear to be operative in the relationship between test
anxiety and cognitive task performance in evaluative
situations. High test-anxious subjects appear to be less
attentive to task cues and more attentive to social-evalua¬
tive cues. Cognitive treatment approaches which address
the negative self-statements, task-irrelevant attentional
patterns, or negative imagery of the test-anxious individual
appear to enhance task performance. While these studies
have suggested that the cognitive concomitants of test
anxiety are related to cognitive task performance, the
description of specific cognitive concomitants remains
at a very global level. That is, the high test-anxious
individual is considered to be negatively self-preoccupied,
and the low test-anxious person is presumably task-oriented
during the evaluatively-stressful test situation (Sarason,
1972; Wine, 1971; Wine, 1980). In order to more effectively
propose cognitively-oriented treatment strategies, as well

36
as more fully understand test anxiety as a cognitive
construct, we need to investigate the extent to which test
anxiety is related to specific cognitive-mediational
variables.
Self-Efficacy Theory
One cognitive-mediational variable which has been
examined primarily in relation to phobic behavior, is the
variable of "self-efficacy." The term "self-efficacy" was
originally conceptualized by Bandura (1977) as a mastery
expectancy variable which could be a powerful predictor of
current level of behavior, as well as a predictor of behavior
change following therapeutic intervention. In fact, the
basic premise of the theory is that the client's efficacy
expectations are central to behavior change. The theory pro¬
poses that expectations of personal efficacy "determine whether
coping behavior will be initiated, how much effort will be
expanded, and how long it will be sustained in the face of
obstacles and aversive circumstances" (Bandura, 1977, p. 191).
Through experiences of mastery provided via various thera¬
peutic modes, there is enhancement of self-efficacy and
corresponding reduction in task or situational avoidance
behavior. Those who persist in subjectively threatening
activities, that turn out to be relatively safe,

37
gain corrective experiences that reinforce their sense of
self-efficacy (Bandura, 1977). When individuals cease their
coping efforts prematurely, they retain their debilitating
expectations and fears indefinitely. Perceived self-efficacy
is, thus, presumed to affect coping efforts, as well as
choice of behavioral settings and activities. Bandura
(1977) emphasizes that expectation is not the sole deter¬
minant of behavior. Expectation of mastery, alone, will not
produce desired performance if the component capabilities
are lacking. In addition, there are many tasks people can
do successfully, which are not performed, because adequate
incentives are not present. However, given appropriate
skills and adequate incentives, efficacy expectations
are seen by Bandura and his colleagues (Bandura & Adams,
1977; Bandura, Adams, & Beyer, 1977), as a major deter¬
minant of people's choice of activities, the employment of
their coping efforts, and their persistence at the coping
effort when dealing with stressful situations. In Bandura's
theoretical analysis, expectations of self-efficacy are
based on four sources of information: 1) performance
accomplishments, 2) vicarious experience, 3) verbal
persuasion, 4) emotional arousal. Bandura (1977) predicted
that therapeutic interventions employing enactive (i.e.
performance accomplishments) techniques would be more
effective in raising subjects' self-efficacy expectations
than would be the interventions employing vicarious
experience (modeling), verbal persuasion (self-instruction,

38
interpretive treatments), or emotional arousal (relaxation,
desensitization). Bandura, Adams, and Beyer (1977)
designed an experiment to test the hypothesized relation¬
ship between self-efficacy and behavioral change, and to
compare the effectiveness of enactive and vicarious exper¬
ience in enhancing efficacy expectations. Severely snake-
phobic subjects were assigned to a participant modeling
condition, a traditional modeling condition, or a no¬
treatment control group. The magnitude of the subjects'
self-efficacy was assessed by giving each subject a list
of the tasks used on a behavioral avoidance pretest
(related to approaching a snake). The subjects were asked
to rate the tasks they expected to perform successfully.
The strength of self-efficacy was assessed by asking each
subject to rate expectancy of success for each task in terms
of a 100 point scale divided into 10 intervals. The
generality of self-efficacy was assessed by asking subjects
to rate the level (magnitude) and strength of their
expectations with regard to successfully carrying out the
same behavioral task with an unfamiliar snake. The results
supported the hypotheses that the participant modeling
technique, with its performance mastery component,
instilled the highest, strongest, and most generalized
efficacy expectations, followed by the vicarious treatment.
In the follow-up assessment, it was found that past behavior
did not consistently or accurately predict future behavior,

39
but that self-efficacy did accurately predict future
behavior. As a further extension of the theory, Bandura
and Adams (1977) examined the role of reducing emotional
arousal through desensitization, on self-efficacy. The
results supported the hypotheses that eliminating emotional
arousal through desensitization would enhance subjects'
levels of self-efficacy. It was also found that the level
of self-efficacy varied for each subject, and that the higher
and stronger the level of self-efficacy induced by the
treatment, the greater the reduction in avoidance behavior.
The level of self-efficacy closely corresponded to approach
behavior, and accurately predicted subsequent behavior to
similar and dissimilar tasks. In summary, Bandura and
Adams (1977) and Bandura, Adams, and Beyer (1977) demon¬
strated that the successes reported by the varying techniques
of enactive (i.e. performance) mastery, vicarious
experience, and desensitization can be readily and
adequately explained through self-efficacy theory. They
have also demonstrated that the level of efficacy estab¬
lished by the intervention significantly affects therapeutic
success in overcoming fearful and avoidant behavior. The
results suggest that efficacy is actually a better predictor
of future behavior, than is past behavior. It is more
consistent and more accurate as a predictor.
Bandura and his colleagues have worked exclusively with
phobic clients, both snake phobics and agoraphobics (Bandura,

40
1977, Bandura, Adams, Hardy, & Howells, in press). It
seems likely that the level of self-efficacy might prove to
be an excellent predictor of academic success and career
choice. Extending Bandura's results to the achievement
situation, one would assume that students with low levels
of academic self-efficacy would fear and avoid subjects
and extracurricular activities which they believed lay
outside their repertoire of coping skills. These students
might restrict themselves to academic programs perceived
as requiring lesser abilities. They might also choose
careers which are lower in socio-economic career status.
They would also be predicted to have poorer academic per¬
formance in school.
Lalonde (1979) extended Bandura's (1977) theory by
investigating the relationship between self-efficacy and
achievement-oriented behavior. She hypothesized that low
expectations of self-efficacy would play a major role in
poor grades, restricted future academic expectations, and
restricted career choice (lower socio-economic career
choice). To test the hypotheses linking academic perfor¬
mance, as well as academic and career expectations, to
efficacy expectations, a measure of self-efficacy pertaining
to academic and achievement-oriented situations was
developed. This measure was necessary, since the only
previous measure of self-efficacy had been the rating scale
used in the Bandura and Adams (1977) and Bandura, Adams,

41
and Beyer (1977) research. This simplified method of
asking subjects to rate their efficacy expectations for
successfully completing certain tasks was effective for
snake phobics. However, it was not so easy to determine
efficacy expectations in achievement situations, since
subjects do not have a particular identifiable set of tasks
which indicate the level of their avoidance behavior.
Lalonde (1979), therefore, investigated self-efficacy, as
related to achievement behavior, by three stages of research
1) the development of a paper and pencil self-report
measure of academic self-efficacy, 2) the establishment of
discriminant and convergent validities, 3) the undertaking
of a validational experimental study.
In order to develop the items for the Measure of
Academic Self-Efficacy (MASE), Bandura's self-efficacy
theory and research were carefully examined as background
data in defining the behavioral characteristics of strongly
and weakly efficacious people. To define as clearly as
possible the behavioral correlates of self-efficacy,
twenty judges (graduate psychology students and teaching
staff) were given a written definition of efficacy as well
as the theoretical statements and empirical results of
Bandura and his colleagues. The judges were asked to
define the behavioral characteristics of weakly and strongly
efficacious people. The test items generated by the judges
were to include a wide selection of extra-curricular and

42
classroom behavior to ensure that the generality of
students' academic self-efficacy was assessed (the extent
to which feelings of efficacy generalize to various academic
situations). Bandura (1977) stated that efficacy expec¬
tancies varied along the three dimensions of magnitude,
generality, and strength, and that an adequate analysis
of expectancy must include all three dimensions. Efficacy
strength and generality were assessed by employing a five-
point response scale, and by tapping a wide variety of
academic activities and behaviors. Magnitude of efficacy
was more difficult to assess for achievement-oriented
situations, since achievement situations do not as easily
lend themselves to rank ordering by difficulty, as do
performance tasks related to snakes. Thus, a magnitude
component was not included in this scale. A reliable scale
was developed from an initial item pool of 350 items. The
final form of the Measure of Academic Self-Efficacy (60
items) (MASE) had a Hoyt (1941) estimate of internal
reliability (scale homogeneity) of .95, and a standard
error of measurement of 5.87. The test-retest reliability
for the MASE was .85, over a one month period.
Convergent and discriminant validity of the MASE was
assessed with a number of measures. It was predicted that
the MASE would correlate positively with grade point
average, scholastic aptitude (Otis-Lennon Mental Ability
Test), willingness to accept responsibility for positive

43
intellectual achievement (Intellectual Achievement Respon-
sibility Questionnaire) (Crandall, Katkovsky, & Crandall,
1965), career aspirations and expectations (coded for
socioeconomic status), post high school career plans, and
self-concept (Piers-Harris Children's Self-Concept Scale)
(Piers, 1969). It was also hypothesized that self-efficacy
would correlate negatively with social anxiety (Social
Avoidance and Distress Scale, Watson & Friend, 1969) and
test anxiety (Test Anxiety Scale, Sarason, 1958).
Finally, an experiment was designed to behaviorally
validate the MASE by testing hypotheses derived from self-
efficacy theory. An anagrams task was designed to test the
hypotheses that 1) highly-efficacious high school students
would persist longer at an anagrams task than would less-
efficacious students 2) highly-efficacious students would
persist longer in the face of failure than would less-
efficacious students 3) forced failure experienced earlier
in the anagrams task would have a greater effect on less-
efficacious students, than would failure experienced later.
These hypotheses were based on hypotheses from self-efficacy
theory which state that highly self-efficacious people will
persist longer at overcoming difficulties and are less
affected by failure, whereas less self-efficacious persons
are more affected by early experiences of failure. Persis¬
tence at the anagrams task was defined in terms of nine
variables: time spent on solved anagrams, skipped anagrams,

44
total anagrams, number of anagrams returned to, proportion
of anagrams skipped, working until forced to stop, willing¬
ness to return to skipped anagrams, and willingness to keep
working (if forced to finish prematurely). Failure was
defined in terms of the early vs. late presentation of hard
anagrams. That is, an early failure experience had hard
anagrams presented before medium and easy anagrams. A
late failure experience had hard anagrams presented following
the presentation of medium and easy anagrams.
In terms of the convergent and discriminant validity
analyses, the MASE correlated moderately and positively, as
predicted, with grade point average, scholastic aptitude,
willingness to accept responsibility for positive intellectual
achievement, and self-concept. In terms of discriminant
validity, the MASE correlated moderately negatively with
social and test anxiety. This is predicted from self-
efficacy theory in that Bandura (1977) views anxiety and
efficacy as antithetical. One cannot be anxious about one's
ability and efficacious at the same time. The stronger the
efficacy expectations, the greater the reduction in anxiety
and avoidance behavior.
The hypotheses concerning the proposed relationship
between academic self-efficacy and career expectations,
and post-secondary school plans were supported. Higher
efficacy scores were associated with higher career expecta¬
tions, and plans to attend a degree-granting institution.

45
In addition, with the effects of verbal intelligence
removed, academic efficacy scores remained low but
positive predictors of grade point average and post¬
secondary school plans.
Analyses of the experimental task results failed to
support the hypothesis that academic self-efficacy would
correlate with persistence at an anagrams task, except for
one instance. This was the variable of willingness, on the
part of subjects forced to finish early, to return to skipped
anagrams. More highly-efficacious subjects indicated this
willingness. In terms of overall performance, the highly-
efficacious subjects correctly solved more of the hard
anagrams than did the les3-efficacious subjects. The hypoth¬
esis concerning the effects of encountering failure earlier
on in the task was not supported. Also, no interaction
effects of efficacy and experimental condition were found,
indicating that less-efficacious subjects were not, in
general, more affected by failure than high-efficacious
subjects, as hypothesized.
In summary, the data suggest that, in terms of internal
consistency and test-retest reliability, the MASE is a
reliable instrument. The convergent and discriminant
validity correlations with other self-report instruments,
provide support for the construct validity of the MASE.
The behavioral validation study did not entirely support
the hypothesized relationships between level of academic

46
self-efficacy, task difficulty, and persistence. A number
of explanations were provided post-hoc for these results.
The most cogent explanation appears to be that the duration
of time provided for the task was not sufficient to ade¬
quately measure persistence. A large proportion (71%) of
the subjects were forced to finish prematurely at the 40
minute time limit. This suggests that 40 minutes may not
have been sufficient time to discriminate between high-and
low-efficacy students, in terms of task persistence.
However, the study did find that highly-efficacious
subjects performed better on the hard anagrams than did
the low efficacious subjects, and exhibited more willingness
to return to unfinished (skipped) anagrams.
Overall, the results of the study provide tentative
support for aspects of Bandura's (1977) theory of self-
efficacy. Correlations between MASE scores, other self-
report indices, and teacher ratings, indicated that
academically highly-efficacious students are more willing
to accept responsibility for their academic successes
than failures. Bandura (1977) had stated that self-efficacy
was enhanced only when subjects viewed their successes
as resulting from their own ability, rather than as a
function of contextual factors such as chance, luck, or
ease of the task. Following this reasoning, it was also
hypothesized that the MASE would correlate positively with
a self-concept scale, and this hypothesis was validated.

47
The MASE also correlated positively with grade point average,
formal post-secondary education plans, plans to enroll in
a degree-granting institution, and career expectations with
higher socio-economic indices (according to a ranking of
career possibilities presented to the subjects). Finally,
with the verbal intelligence factor removed, the MASE
correlated at a low, but positive level with grade point
average and post secondary school plans. These results
suggest that the mastery expectancy variable of self-efficacy
is relevant to prediction of academic success and planning.
The results also relate to the work of McClelland (1973),
Tait (1973), and Taylor (1964), who found that non-intellec¬
tual factors play a role in academic and vocational success.
Since academic self-efficacy is an expectancy variable,
it is important to distinguish it from other expectancy
variables. Rotter's (1966) locus of control theory and
Maier and Seligman's (1976) learned helplessness theory
may be seen as somewhat analogous to Bandura's (1977) self-
efficacy theory and the Measure of Academic Self-Efficacy.
Locus of control and learned helplessness, however, both
measure outcome expectancies, rather than mastery expec¬
tancies. Rotter's theory is concerned with a person's
belief that outcomes are determined either by one's own
actions (internal control), or by forces outside one's
control (external control). Self-efficacy is different
from locus of control, because people may believe that

48
outcomes are internally controlled, but may feel a sense of
low efficacy, if they believe that they lack the skills
required to perform the activity.
Maier and Seligman's (1976) theory of learned
helplessness, and the Hopelessness Scale (Beck et al.,
1974) resemble the concept of self-efficacy, yet are differ¬
ent from self-efficacy and the MASE. Learned helplessness
is also an outcome expectancy variable. It states that
when people cannot affect what happens to them, they become
apathetic or "learned helpless" and stop trying to cope
(Lalonde, 1979). People may give up trying because they
lack a sense of efficacy in achieving required performances,
or they may give up trying because they expect their efforts
to produce no results in an unresponsive or punishing
environment. To change efficacy-based futility requires
development of a sense of personal effectiveness. To
change outcome-based futility requires changing the social
environment so that people are rewarded for using the
competencies they already possess. Bandura suggests
that considering both efficacy and outcome expectancies
would produce the best prediction of behavior (Lalonde,
1979).
The concept of achievement motivation developed by
McClelland et al. (1958) can also be considered to be
different from the concept of academic self-efficacy. In
the definition of high achievement motivation, the key

49
issue is concern over competition with a standard of
excellence, and concern with success or failure in an
academic task. A subject may describe either positive or
negative feelings toward achievement in the achievement
motive index, and still score high in need for achievement.
Thus, high need achievement refers to any concerns regarding
achievement, including fear of failure, or disappointment
at lack of success. The absence of achievement concerns
indicates low achievement motivation.
In the conceptualization of academic self-efficacy,
high self-efficacy is seen in terms of a positive orienta¬
tion toward achievement situations. That is, the person
has a mastery expectancy in achievement situations. Low
self-efficacy is concerned with achievement failure, in
that the person perceives a lack of personal coping skills
in the achievement situation. Concern over competitive
situations and academic failure would, therefore, fall
into the domain of low self-efficacy. Those who are
indifferent to achievement situations would not fall into
the category of low self-efficacy. These two concepts of
achievement motivation, and academic self-efficacy, thus,
appear to differ at a theoretical level. They also differ
methodologically, in that achievement motivation is
measured by assessing achievement imagery via a projective
technique similar to the Thematic Apperception Test
(McClelland, Atkinson, Clark, & Lowell, 1958). The
individual responds to pictures which suggest achievement

50
themes. This is a radically different methodology from that
proposed by Lalonde (1979) for assessing academic self-
efficacy.
The research and treatment literature in test anxiety
has strongly suggested that cognitive concomitants of test
anxiety are important in predicting performance in evaluative
situations. Although emotional arousal factors do appear
to be related to self-report of test anxiety, these factors
apparently are less predictive of performance differences
in high and low test-anxious individuals. At this point,
the cognitive factors which underlie test anxiety have
only been generally identified. In order to more adequately
understand the cognitive concomitants of test anxiety, it
is necessary to identify specific cognitive variables which
mediate the observed performance differences. Bandura has
proposed a cognitive-mediating variable, self-efficacy,
which appears to predict the extent to which individuals
approach or avoid task situations. This variable may also
have relevance for understanding the cognitive factors
underlying test anxiety.
Lalonde's study (1979) has begun to delineate the role
of academic self-efficacy as a relevant variable in the
prediction of academic success and career planning. To
the extent that students believe that they have the coping
skills to master academic situations, they tend to have
higher grade point averages, take more responsibility for

51
their academic successes, and have career expectations and
future plans associated with greater socio-economic and
academic achievement. In her study, Lalonde also found
that test anxiety, as measured on the Test Anxiety
Scale, correlated -.36 with academic self-efficacy. This
correlation indicated some overlapping variance across
these two variables, but not identity. This dissertation
study extends the work of Lalonde, in examining the rela¬
tionship of test anxiety and academic self-efficacy as
predictors of task performance, under various conditions.
It is proposed that self-efficacy is a relevant
cognitive concomitant of test anxiety. It appears to
reflect some of the core problems of the test-anxious
individual in the academically-evaluative situation. As
Wine (1971) and others (Sarason, 1975; Meichenbaum, 1972)
have hypothesized, the test-anxious person demonstrates
a performance decrement in the evaluative situation,
because he fails to focus on the relevant parts of the task.
Instead he ruminates on his perceived incompetence in the
test situation. Since the concept of academic self-efficacy
deals with a person's belief that he has (or does not have)
the coping skills to master academic situations, it appears
that academic self-efficacy is certainly related, at a
conceptual level, to the cognitive aspects of test anxiety
discussed earlier in this review. Academic self-efficacy

52
may, thus, provide some further information about the
modus operandi of the test-anxious individual.
This dissertation study is an attempt to delineate
the role of academic self-efficacy as it relates to the
multidimensional construct of test anxiety. The study
determines, via a multivariate predictor model, the
extent to which a measure of academic self-efficacy, oper¬
ationalized as a performance mastery expectancy variable
in academic situations, is related to a measure of test
anxiety. Furthermore, it determines whether these two
variables jointly and singly can predict task performance
under evaluative, reassuring, and neutral task conditions.
This study is, thus, an attempt to establish whether self-
efficacy is a cognitive concomitant of test anxiety, and
the extent to which there is non-overlapping variance
across these two variables, in predicting to cognitive
performance. It also determines the extent to which both
variables predict performance under evaluatively-stressful
conditions, and the extent to which this prediction generalizes
to non-stressful conditions. That is, it addresses the
question of whether these variables are generalized predictors
of academic performance, or only relevant under evalúatively-
stressful academic task conditions.
Some previous test anxiety studies, which have
manipulated task conditions, have found that manipulation
of the level of evaluative stress presented to subjects,

53
has related to performance differences in high and low
test-anxious subjects. Many of the studies manipulated
instructional conditions, while holding task demands
constant (Sarason, 1972). These studies have generally
involved comparisons of a) "ego-involving" instructions—
these instructions inform the subject that his task
performance level is reflective of his intellectual
capabilities; b) minimal task instructions—instructions
which specifically address the cognitive task (e.g.
directions for solving an anagrams task); and c) reassur¬
ance instructions—instructions which reassure the subject
that his task performance will not be used to evaluate him.
Frequently, under this last condition, the subject is
told that his performance is anonymous, and only important
so that the examiner can examine average group performance
data (Wine, 1971; Sarason, 1958; Sarason & Ganzer, 1962,
1963; Russell & Sarason, 1965; Ganzer, 1968). Highly
test-anxious subjects tend to perform more poorly following
highly evaluative "ego-involving" instructions, than
following reassuring "anonymous" instructions. The reverse
tends to be true for low test-anxious subjects. Following
highly evaluative instructions, low test-anxious subjects
perform better than high test-anxious subjects. However,
following the reassuring "anonymous" instructions, high
test-anxious subjects perform better than low test-anxious
subjects. Under the third type of condition, the neutral

54
or minimal task instruction condition, the effects of
test anxiety on task performance, are wiped out. High and
low test-anxious subjects perform at equivalent levels,
intermediate between their performances in the highly-
evaluative and reassuring conditions. Thus, these particular
studies indicate that test anxiety appears to be a relevant
variable in cognitive task performance, if task instructions
create an evaluative situation, or, the opposite of an
evaluative situation, that is, a reassuring situation,
where individual performances do not count. However, if
task instructions are neutral, minus any evaluative
component, the high and low test-anxious subjects perform
at equivalent levels on a cognitive task.
There is only one study which has examined self-efficacy
in relation to task conditions. This is the behavioral
validation study, from the Lalonde (1979) dissertation.
This study did not manipulate task instructions in the
manner described for the test anxiety studies described
above. In other words, the subjects were not presented
with instructions varying in the amount of evaluative
emphasis on performance. The study did, however, attempt
to develop a failure experience, by presenting subjects
with anagrams of varying degrees of difficulty. Anagrams
were presented to subjects in various orders: hard,
medium, easy. It was hypothesized that highly-efficacious
high school students would persist longer in the face of

55
failure than less-efficacious students. Presumably, the
condition with the hard anagrams presented first would
produce the most difficult task situation by providing
early failure. The results indicated that, in terms of
persistence, the less-efficacious students were not more
affected by the failure experiences. However, over all
the experimental conditions, the highly-efficacious students
solved more of the hard anagrams. Thus, the creation of
a "forced" failure experience did not differentially
affect the low-and high-efficacy students. In conclu¬
sion, this study suggests that task difficulty might differ¬
entially affect the performance of low and high self-
efficacious students, but a forced failure experience
produced by the presentation of hard anagrams early in
the task will not differentially affect the performance
of the two groups of subjects. The strength of the
results of this study are somewhat diminished by the
fact that the easy and medium anagram sets were extremely
easy for both groups, thus reducing the discriminative
power of the anagram variable. In terms of the proportions
of anagrams solved, 91 to 100% of all of the easy and
medium anagrams were solved by both groups.
Unfortunately, this study is not comparable to the
test anxiety studies, which manipulated task instructions of
an evaluative vs. neutral and/or reassuring nature.
Consequently, hypotheses can not be derived from the

56
Lalonde study regarding the differential performance of
highly-efficacious and less-efficacious students under
varying evaluative conditions. The data do suggest,
however, that there may be overall cognitive performance
differences between low-and high-efficacy students,
but specific failure experiences will not differentially
affect the performance of the two groups.
Hypotheses regarding the differential performance
of low vs. high-efficacious students can be generated
from Bandura's theory. From Bandura's theory, it is
expected that there would be an overall performance
difference on an academically-related cognitive task,
however, the manipulation of evaluative vs. non-evaluative
instructions would not produce differential performance
results, such that the effects of the efficacy variable
would only be significant under one or the other condition.
There are several aspects of Bandura's theory which
would suggest that an experimental manipulation would
not produce differential results. First of all, according
to Bandura's theory, expectations of personal efficacy
are derived from four sources of information: performance
accomplishments, vicarious experience, verbal persuasion,
and physiological data (Bandura, 1977). The more dependable
the experiential source of efficacy information, the greater
the likelihood of change in perceived self-efficacy. The
task instructions provided by the evaluative condition

57
could be considered to fall into the category of verbal
persuasion. The subject is led, via suggestion, to believe
that the problem-solving task, for example the anagrams
task used in the test anxiety studies, is a measure of
intelligence or ability. Thus, the subject presumably
becomes ego-involved. According to Bandura (197 7) , verbal
persuasion is a relatively poor source for providing change
in perceived self-efficacy. Efficacy expectations induced
by verbal persuasion are likely to be weaker than those
arising from one's own accomplishments. This is due to
the fact that verbal persuasions do not provide an "authentic
experiential base" for the beliefs (Bandura, 1977, p. 198).
If subjects already have a long history of failure to cope
with academically-evaluative situations, expectations
induced by suggestion are not likely to dramatically alter
performance in a positive or negative direction. Thus,
provided the individual has a long history of failure or
success at academically-related tasks, the imposition of
a brief laboratory manipulation is unlikely to reduce or
greatly increase the relationship of self-efficacy to
performance. If task conditions suggest an academically-
difficult task, regardless of whether instructions are
neutral, reassuring, or evaluative, perceived self compe¬
tence will be significantly related to performance.
Bandura states, in addition, that the effects of failure
on personal efficacy depend partly on the timing and the

58
total pattern of experiences in which the failures occur.
In the strongly efficacious person, after efficacy expec¬
tations are developed through repeated success, the
negative impact of occasional failures is likely to be
reduced (Bandura, 1977). By the same token, "reassuring"
instructions are unlikely to induce feelings of coping
mastery in a person with poor self-efficacy of long standing,
in academic situations. Thus, for the variable of academic
self-efficacy, task instructions are not likely to signifi¬
cantly reduce or increase the relationship between self-
efficacy and cognitive task performance.
In the present study, test anxiety and self-efficacy
are both used to predict performance under evaluative,
neutral, and reassuring task instructions. According to
the research findings discussed above in regard to test
anxiety and the theoretical issues presented in terms of
self-efficacy, test anxiety will interact with the instruc¬
tional condition, when predicting performance on a
cognitive task. It is expected that self-efficacy will not
interact with the instructional condition when predicting
performance on a cognitive task. It is expected that test
anxiety will account for a significant amount of the
variance in performance under evaluative and reassuring
task instructions, and it will account for less of the
variance under neutral instructions. It is expected that
self-efficacy will significantly account for performance

59
on a cognitive task regardless of the task instructions
presented to the subject. If this differential prediction
for test anxiety and academic self-efficacy is supported,
this suggests that these are two different, but overlapping
dimensions, which facilitate the prediction of cognitive
performance, beyond knowledge of ability level.
The cognitive task which was presented to the subjects
was a series of anagrams. This laboratory analog task was
chosen because it has been shown to be a sensitive measure
of cognitive task performance differences related to test
anxiety under evaluative and reassuring instructional
conditions (Sarason & Stoops, 1978; Sarason, 1972; Russell
& Sarason, 1965).
In addition, final grades for each of the subjects
in their General Psychology class were obtained, to see
whether test anxiety and self-efficacy would also predict
to a cognitive performance measure which has meaning outside
of the laboratory setting. The final grade in General
Psychology was chosen as the external cognitive performance
measure because it was one of the few performance measures
which all subjects would have in common. Even though the
subjects were not all in the same sections, there was some
similarity in the type of materials and level of difficulty
presented to all subjects. Other course grades or grade
point average would present a problem in terms of lack of
generality across all of the subjects. In addition, the

60
final grade in General Psychology was readily available for
the use of the investigator.
The hypotheses proposed for this study, given the
existing literature and theory, were the following:
1) Academic self-efficacy would be moderately
negatively correlated with test anxiety.
2) Test anxiety would predict anagram performance
under evaluative and reassuring task instructions,
but not as effectively under neutral instructions,
in accordance with the test anxiety literature.
3) Self-efficacy would predict anagram performance
under evaluative, neutral, and reassuring task
instructions, in accordance with Bandura's theory.
4) Cognitive ability level, as measured by an aptitude
test, would account for a significant amount of
variance in predicting to cognitive task performance,
provided there were sufficient variation in aptitude
scores. If the range of aptitude scores was reduced,
then the predictive power of cognitive ability
would be reduced, as well.
5) Test anxiety, self-efficacy, and aptitude scores,
would all be predictive of performance in General
Psychology, as measured by the subject's final
grade in the course.

CHAPTER TWO
METHOD
Subjects
The subjects were 234 undergraduate students enrolled
in the introductory psychology course sections (PSY 2013)
at the University of Florida during the Fall Semester,
1981. They were required to participate in experimental
studies in order to complete their course requirement.
Procedure
The subjects signed up for two experimental sessions,
which were held one week apart. They were required to
complete both sessions in order to receive credit for
participation in the study. Those subjects who completed
both sessions were included in the study (n=234). The
investigator was assisted in the data collection sessions
by four undergraduate psychology majors, who administered
questionnaires during the first sessions, and conducted
the experimental part of the study, during the second
sessions. They were blind to the experimental hypotheses.
61

62
During the initial testing session, all subjects
received the Test Anxiety Scale and the Measure of
Academic Self-Efficacy. The Test Anxiety Scale, developed
by Sarason (1972), is a 37 item scale frequently used to
assess test anxiety. It is a true/false test, and it
yields one total score. Test-retest reliabilities over
.80 have been reported for intervals of several weeks
(Sarason, 1978) on a sample of 283 males and 237 female
undergraduates. Wagaman, Cormier, and Cormier (1975)
reported a test-retest reliability coefficient of .87. This
test has been utilized to identify high and low test-anxious
individuals in most of the research conducted by Sarason and
his colleagues (Sarason, 1968, 1972, 1973a, 1975; Sarason,
Pederson, & Nyman, 1968; Sarason & Stoops, 1978). It has
also been used in many other research studies in the area
of test anxiety (Wine, 1971; Spielberger, Anton, & Bedell,
1976; Wine, 1980).
All subjects were also administered the Measure of
Academic Self-Efficacy (MASE) developed by Lalonde (1979).
This is a 64 item test designed to assess self-efficacy
in academic situations. The Hoyt estimate of internal
consistency reliability for the test was found to be .95,
while test-retest reliability, over a one month period,
was .85. The distribution of the MASE (on a sample of 885
male and female high school students) reasonably approximated
a normal curve. Pseudo-random responding and acquiescent

63
response set were controlled by having equal numbers of
negatively and positively worded items, and the inclusion
of infrequency items in the scale. Social desirability as
a response set was controlled by employing Jackson's (1966)
stringent item selection procedure. Only those items which
correlated more highly with the efficacy scale than with the
social desirability scale were included in the final scale.
The MASE contains a Likert-type format. The subjects
were required to assess their response to each item on a
five point scale, ranging from "never or rarely true"
to "almost always or always true." This test took approxi¬
mately 10-15 minutes to complete, so the entire initial
testing session took approximately 30 minutes.
At the time of the initial testing session, subjects
also gave their written permission to release to the
investigator, their verbal and quantitative Scholastic
Aptitude Test scores and final grades in PSY 2013 (Intro¬
ductory Psychology). The Scholastic Aptitude Test (SAT)
scores were obtained from the Registrar's office, and the
grades were obtained from the individual instructors. At
the end of the initial testing session, the subjects were
told to return for an additional testing session, one week
later.
The final testing session involved the administration
of a series of 13 anagrams selected from a list of anagrams
developed by Sargent (1940). These anagrams have been used

64
in several test anxiety studies (Sarason, & Stoops, 1978;
Sarason, 1973a, 1973b; Russell, & Sarason, 1965; Crager,
1959). The Sargent series of anagrams consists of 40
anagrams which are divided into three difficulty levels;
easy, medium, hard.
Prior to the beginning of the actual dissertation
study, a pilot study had been conducted, to select 13
anagrams which would be difficult enough to present a
challenge, yet not so difficult as to provide no range
of performance. In order to assess the level of difficulty
of the Sargent (1940) anagrams in this college population,
52 subjects were administered all 40 of Sargent's anagrams.
They were given unlimited time to complete the anagrams;
however, they were assessed for the amount of time it took
to complete the 13 most difficult anagrams. This was done
in order to provide an estimate of an appropriate time
limit for presentation of the 13 anagrams eventually
selected.
The percentage of subjects solving each anagram correctly
was computed. It was found that, for this pilot sample,
not all of Sargent's 13 "most difficult" anagrams were the
most difficult. The 13 most difficult anagrams for this
sample were solved by 42% to 69% of the sample. Eight of
these anagrams were in Sargent's "easy" group, and five
were in Sargent's "medium" difficulty group. As mentioned
above, the subjects were also assessed for time to

65
completion of Sargent's "hard" anagrams. It was found that
they took a mean time of 43 minutes to complete these
anagrams.
After this initial pilot study was completed, an
additional pilot study was conducted. For this study,
31 subjects were given all aspects of the full dissertation
study, including both the questionnaire session (first
session) and the experimental session (second session).
The purpose of the second pilot study was to essentially
try out the experimental materials, instructions, and time
limits to be used in the full dissertation study.
In the actual dissertation study, the second session
was set up in the following manner. Prior to the second
session, all subjects were randomly assigned to one of three
experimental conditions: neutral, reassuring, or evaluative
instructions. When they attended the second session they
were taken to separate rooms and given the anagrams
task under one of the three instructional conditions. The
instructions were administered by the undergraduate assis¬
tants .
Condition one, the "neutral" instructional condition,
involved the following instructions (adapted from Russell &
Sarason, 1965). These instructions were read aloud to the
subjects as a group:

66
This part of the experiment involves a
problem-solving task. The problems I will
give you to solve are called anagrams. An
anagram is a word with the letters all
mixed up, and your task is to make a mean¬
ingful word. Only one word can be made
from each anagram, and they are all common
words. Solve the anagrams, and write the
answers on the problem sheets in the answer
column. If you need scratch paper, use
the backs of the answer sheets. You will
have 45 minutes to solve the anagram task.
The subjects in the "evaluative" condition received
the same basic instructions given to the subjects in the
"neutral" condition, plus an added evaluative component:
This part of the experiment involves a
problem-solving task. The problems I will
give you to solve are called anagrams. An
anagram is a word with the letters all
mixed up, and your task is to make a mean¬
ingful word. Only one word can be made
from each anagram, and they are all common
words. Solve the anagrams, and write the
answers on the problem sheets in the answer
column. If you need scratch paper, use the
backs of the answer sheets. These anagrams
are a measure of the ability to think in
abstract terms. High school students of
above average intelligence and most college
students have been found to solve these
particular anagrams quickly and with little
difficulty, so you should have little trouble.
You will have 45 minutes to solve the anagram
task.
The subjects in the "reassuring" condition were also
given the same basic instructions as subjects in the "neutral"
condition; however, they were given additional reassuring
comments:

67
This part of the experiment involves a
problem-solving task. The problems I will
give you to solve are called anagrams. An
anagram is a word with the letters all mixed
up, and your task is to make a meaningful
word. Only one word can be made from each
anagram, and they are all common words.
Solve the anagrams, and write the answers on
the problem sheets in the answer column. If
you need scratch paper, use the backs of the
answer sheets. Most of you have probably seen
anagrams before, or have worked on them.
These anagrams work the same way, however,
they are harder than most you have seen in
books or magazines. Consequently, you may
not finish all of them, and you may find
some of the anagrams very difficult. If
this happens, don't worry about it. No one
will find the anagrams easy. You will have
45 minutes to solve the anagrams.
The anagrams task was scored for total number of
anagrams successfully completed. As shown above, the
subjects were given 45 minutes to solve the anagrams task.
This time limit was selected from the second pilot study
discussed previously. In that study, the subjects had been
given 40 minutes to solve the anagrams, and although
there was an acceptable range of performance, some anagrams
were solved by very few subjects. "ELSAUX" was solved by
only 29% of the subjects, and "EIVARR" was solved by only
35% of the subjects. In general, these thirteen words
appeared slightly more difficult for these subjects than
for the original pilot group. Therefore, for the actual
dissertation study, it was decided that these same words
should be used, but that subjects should be given slightly
more time to solve them.

68
Statistical Analysis
All analyses were conducted using the Statistical
Analysis System (SAS) (Helwig, & Council, 1979) packaged
programs for multiple regression analyses. The two
primary programs utilized were GLM (general linear models)
and Stepwise (stepwise regression).
Prior to any regression analyses being conducted,
it was necessary to randomly select subjects for a main
sample and a validation sample. Out of the 234 subjects,
there were 21 subjects who were missing data for independent
variables, either aptitude scores, test anxiety scores,
or self-efficacy scores. Most of these subjects were
missing SAT scores for various reasons (foreign students,
junior college transfers, etc.). Four subjects were
missing test anxiety or self efficacy scores, due to lack
of completion of the questionnaires. These 21 subjects
were generally excluded from regression analyses, although
their scores were included in means analyses and Pearson's
R correlations. An additional three subjects were included
in the anagram analyses, but excluded in the grade analyses,
because they did not have grades for the PSY 2013 (Intro¬
ductory Psychology) course.
The remaining 213 subjects were randomly selected to
be either in the main sample consisting of 162 subjects, or
a validation sample, consisting of 51 subjects. The

69
randomization was accomplished by computer, using a random
seed procedure from the SAS system (Helwig & Council, 1979).
The initial analyses involved testing the original
proposed models, one for anagram performance as the
dependent variable, and one for grade performance as the
dependent variable. After these models were tested, a
stepwise regression was conducted to establish the best
models for both grade and anagram. These analyses included
the variables from the original proposed models as well as
one additional interactional variable, test anxiety inter¬
acting with self-efficacy. This variable was included
because it was hypothesized that test anxiety might
interact with self-efficacy in predicting performance on
the dependent measures, and it might enhance the model
beyond the original proposed linear variables.
Finally, analyses of variance were conducted to
establish whether this study would replicate the findings
of previous studies which employed similar manipulations
of test anxiety and instructions, yet utilized an analysis
of variance model rather than a linear regression model in
data analyses.

CHAPTER THREE
RESULTS
Descriptive statistics were calculated for all of the
continuous variables in the study: test anxiety, self-
efficacy, Scholastic Aptitude Test (SAT) verbal and
quantitative combined, SAT verbal alone, SAT quantitative
alone, grade, and anagram (Table 1). In addition, Pearson
r correlation matrices were calculated on all of the
quantitative variables, for the combined main and validation
sample of n=213 (Table 2). As hypothesized, test anxiety
correlated moderately negatively with self-efficacy, r=-.42
(p < .0001) .
Initially, multiple regression analyses were conducted
for both grade and anagram as the dependent variables. The
initial model included group (i.e. instructional condition),
test anxiety, self-efficacy, SAT verbal (SATV), SAT quanti¬
tative (SATQ), and interactions between group and test
anxiety, group and self-efficacy, group and SATV, and
group and SATQ.
The initial analysis, using anagram as the dependent
variable, indicated that the following variables were all
significant individual predictors of anagram performance
at the p < .05 level or better: test anxiety (p < .03),
70

Table 1
Descriptive Statistics for Independent and Dependent Variables
Variable
N
Mean
Standard
Deviation
Minimum
Value
Maximum
Value
Test Anxiety
232
16.14
7.60
2.00
34.00
Self-Efficacy
232
240.34
27.34
164.00
298.00
SAT-Verbal
216
476.71
82.22
230.00
700.00
SAT-Quantitative
216
520.88
87.31
240.00
770.00
Anagram
234
8.29
3.28
1.00
13.00
Grade
228
3.02
.90
0.00
4.00
(SAT = Scholastic Aptitude Test)

Table 2
Intercorrelation Matrix on the Independent and Dependent Variables (n=213)
Testanx Selfef
SATV
SATQ
Grade3
Anagr,
Testanx
-.42**
-.34**
-.28**
-.21*
00
o
•
1
Selfef
00
o
•
in
o
•
.12
. 09
SATV
.46**
. 20*
1
•
o
SATQ
. 22*
.05
Grade3
.03
Anagram
Note: Testanx = Test Anxiety; Selfef = Self-Efficacy; SATV ^Scholastic
Aptitude Test-Verbal; SATQ = Scholastic Aptitude Test-Quantitative
* p < .005
** p < .0001
All correlations of grade with the other variables were done on
a sample of n=210, because three subjects had missing grade data.

73
self-efficacy (p < .05), SATV (p < .02), SATQ (p < .03).
The main effect for group instructional condition was not
significant, and neither were the interaction effects for
test anxiety by group, self-efficacy by group, SATV by
group, and SATQ by group. The r2 for this model was .14.
The same process of testing out the original proposed
model was also carried out for the grade variable. The
model tested included the five main variables of group,
test anxiety, self-efficacy, SATV, and SATQ, as well as
the interaction of group with the other four variables.
In this initial model, none of the variables reached
significance at the (p < .05) level or better. The r2 for
this model was .13.
Validation analyses were conducted for both the
anagram model and the grade model. For the anagram model,
the observed scores for the validation sample (n=51) were
correlated with their expected scores, based on the
regression equation generated from the main sample. The
Pearson r correlation was .12, which was not significant,
indicating that the results with this sample of n=162 may
not be stable when this model is replicated with other
samples of equivalent size.
Validation analyses were also conducted for the grade
model. The observed scores for the validation sample
(n=49) were correlated with their expected scores, based
on the regression equation for grade in the main sample.

74
The Pearson r correlation was .34, which was significant
at the p < .02 level of probability. This model did
replicate to the validation sample.
Additional analyses were also conducted with reduced
models for both grade and anagram (including main effects
only). In both anagram and grade models, main effects for
group and the interaction of group with the other variables
were eliminated from the models. F tests were conducted,
comparing the reduced models to the full models. For both
grade and anagram, the results were not significant, indi¬
cating that the presence of the group main effects and group
interactions did not significantly enhance the models.
Hence, they were removed from additional analyses.
Since the r2 values for both the anagram model and the
grade model were relatively low, .14 and.13, respectively,
it appeared that the initial proposed variables accounted
for only a relatively small proportion of the variance in
performance. It appeared likely that other variables might
account for more of the variance and provide better models.
Other variables not currently in the model might correlate
more highly with anagram and grade performance. It
appeared appropriate, after testing the initial models, to
build a better model through the stepwise regression
procedure. Examination of the plots of expected vs.
observed scores suggested that perhaps there was a pattern
to the residuals which might be explained by interactions

75
between the variables. Since text anxiety and self-efficacy
were hypothesized to jointly predict cognitive task per¬
formance, it was additionally hypothesized that the
interaction of test anxiety and self-efficacy might
predict cognitive performance more effectively than test
anxiety and self-efficacy operating within the model in a
linear fashion.
In this stepwise procedure all the main effect
variables were included, i.e., test anxiety, self-efficacy,
SATV, and SATQ. In addition, the interaction of test
anxiety and self-efficacy (TESTSE) was included in the
model, as discussed above. In order to estimate the "best"
model, it was necessary to use all available data. Hence,
the main sample and validation sample were merged, and a
stepwise regression was run for the complete sample. In
this case, 210 subjects were included in the stepwise
analysis for grade, and 213 subjects were included in the
stepwise analysis for anagram.
A stepwise regression analysis on the complete sample
was conducted for anagram as the dependent variable (n=213).
The stepwise analysis yielded two variables as significant
predictors of anagram task performance. These variables
were test anxiety, and the interaction of test anxiety with
self-efficacy. Test anxiety predicted performance at a
probability of p < .02. Test anxiety in interaction with
self-efficacy predicted anagram performance at p < .006.

76
The r2 for this model was .04. The overall F value was
F (2,210)=4.54, p < .01.
In the stepwise regression analysis for grade on a
sample of n=210 (three subjects in the main sample did not
have grade data), two variables were significant predictors
of grade performance. These variables were test anxiety
and SATQ. The r2 for this model was r =.07. The overall F
value for this model was F (2, 207)=7.74, p < .0006. Test
anxiety was a significant predictor variable at p < .02,
and SATQ was significant at p < .02.
Validation analyses were conducted for the two stepwise
regression models developed from the complete sample (one
for grades, and one for anagram). First, the models for
anagram and grade developed in the full stepwise regression,
were used in a standard multiple regression performed on
the main samples. Then these models were validated on the
smaller validation samples. The samples were merged and
the final regression results are reported.
For the anagram model (test anxiety and TESTSE as
predictors) the main sample was validated on the validation
sample. The correlation between the expected score and
the observed score was r=.ll, which was non-significant.
The validation sample was combined with the main sample,
and a regression analysis was done. This analysis yielded
an r2 of .03. Test anxiety predicted anagram performance
at the p < .02 level, and TESTSE predicted at the p < .01
level.

77
The same analyses described above for anagram were
also done with grade as the dependent variable. The
stepwise regression had found that the best model predicting
grade performance was test anxiety and SATQ. The standard
regression procedure (GLM) was used to analyse this
model on the main sample. This model was then validated
on the cross-validation sample. The expected value for
grade was correlated with the observed value, on the
validation sample, and the Pearson r was found to be r=.30,
which was significant at p < .03. Hence, this model was
validated. The validation sample was then combined with
the main sample, and a multiple regression analysis was
performed on the combined sample (n=211). The results
yielded an r2 of .07. Test anxiety contributed significantly
to the model, with a probability of p < .03. SATQ also
contributed significantly to the model, with a probability
of p < .02.
In addition, several analyses of variance were conducted
on smaller models to determine whether the results of this
study would replicate previous test anxiety studies which
had employed an analysis of variance model in statistical
analyses. The first analysis of variance examined the
instruction variable, group, and test anxiety, in a 3 x 3
model. The subjects were classified into high, middle,
and low test anxiety, and they had previously been assigned
to the reassuring, neutral, or evaluative instructional

78
condition. In the test anxiety classification, the high
and low test-anxious subjects were drawn from the upper
and lower 25% of the score distribution. The score ranges
for the three Test Anxiety Scale groups were high test
anxiety, 21-37, middle test anxiety, 11-20, and low test
anxiety, 0-10. These cutoff scores corresponded almost
identically to those used by Sarason (1972).
In this analysis of variance with group and test
anxiety as the two independent variables, and anagram
performance as the dependent variable, the overall F value
was not significant, and there were no significant main
effects or interactions, although the group main effect
approached significance (p < .07). The same analysis of
variance was also conducted for grade as the dependent
variable. In this analysis, the overall F value was
significant, F (8, 218)=2.14 (p < .03). There was a signifi¬
cant main effect for test anxiety (p < .01). Further
analyses were conducted with Duncan's Multiple Range Test
to compare the individual test anxiety means. The results
indicated that the high test anxiety group received signifi¬
cantly lower grades (p < .05) than the low and middle test
anxiety groups, but these two groups did not significantly
differ from each other.
An additional analysis of variance was conducted
incorporating test anxiety, self-efficacy, and group
instructions. The subjects were classified into the test

79
anxiety groups with the same cutoff scores as above. For
the self-efficacy measure, the subjects who were classified
into the high and low self-efficacy groups, were in the
upper and lower 25% of the score distribution. The remaining
subjects were classified into the middle self-efficacy
groups. The score ranges for the three self-efficacy
groups were as follows: high self-efficacy, 262 or greater,
middle self-efficacy, 222-261, low self-efficacy, 211 or
lower. Again, as in the previous analysis of variance
with anagram as the dependent variable, the overall F
value was not significant, and there were no significant
main effects or interactions. When grade was used as the
dependent variable, the overall F value was not significant;
however, there was a significant main effect for test anxiety
(p < .01). Duncan's Multiple Range Test indicated again
that the high test-anxious group received significantly
lower grades than the middle and low test-anxious group
(p < .05) .

CHAPTER FOUR
DISCUSSION
Test anxiety and self-efficacy correlated significantly
in a moderate negative relationship. The correlation of
-.42 was very similar to the correlation coefficient found
by Lalonde in her dissertation study, r=-.35 (Lalonde, 1979)
These results indicate that test anxiety and self-efficacy
are two moderately related, but not identical constructs.
In the initial model, with anagram as the dependent
variable, test anxiety, self-efficacy, SAT verbal, and SAT
quantitative, were all significant predictors of performance
However, the instructional condition (evaluative, reassuring
neutral) was neither significant as a main effect, nor in
interaction with the other variables. Therefore, the
hypothesis that test anxiety would interact with task
instructions, was not supported by the data. It was
predicted that self-efficacy would not interact with
instructions, and this was supported by these results.
It was also hypothesized that test anxiety, self-
efficacy, SAT verbal, and SAT quantitative, would predict
grade performance in the introductory psychology class.
However, each of these variables, individually, did not
significantly predict grade performance.
80

81
Validation analyses conducted on both the original
anagram and grade models yielded different results for
both. The original anagram model, based on 162 subjects,
did not cross-validate, indicating that the results for this
model may not be stable when replicated with a comparably
sized sample. However, if the data from the main sample
are merged with the data from the validation sample, the
resultant regression model, based on 213 subjects, may
prove to be stable.
Validation analyses conducted on the original grade
model indicated that this original model did validate to the
smaller validation sample. This result indicates that this
model would be stable if replicated in another sample.
The main results of the study, based on the stepwise
regression with anagram as the dependent variable, indicated
that test anxiety, and the interaction of test anxiety with
self-efficacy, were both significant predictors of anagram
performance. Test anxiety, as a main effect, related
negatively to performance. In the stepwise regression analysis
for grades, test anxiety and SATQ were found to be significant
predictors of grades received in the introductory psychology
course.
The fact that test anxiety did not interact with
instructional condition is contrary to several of the
previous research studies in the area which have found a
test anxiety by instruction interaction (Sarason, 1961;
Sarason, 1973b). However, examination of past studies using

82
the instructional manipulation indicated inconclusive
results. Some studies have found a main effect for test
anxiety, with high test-anxious subjects doing more poorly
than low test-anxious subjects, but no test anxiety by
instruction interaction (Sarason, Mandler, & Craighill,
1952; Mandler & Sarason, 1952; Crager, 1959; Russell &
Sarason, 1965). Although the test anxiety by instruction
interaction has been found by some researchers, it is not
an unequivocal finding.
It could be speculated that the current study did not
find a test anxiety by instruction interaction because of
the method of data analysis. All of the previous studies
utilized an analysis of variance (ANOVA) model in statis¬
tical analysis, classifying the subjects into high and low
test anxiety groups (median split) or three test anxiety
groups (low, middle, high). The current study treated the
data in continuous form and analyzed the data within a
multiple regression framework. In order to answer this
question, the data from this study were analyzed via
the ANOVA model to make the analyses more comparable to
previous studies.
The analysis of variance for test anxiety and instruc¬
tions, with anagram as the dependent variable, found no
significant results. That is, there was no significant
relationship between test anxiety and anagram performance,
nor any significant interaction between test anxiety and

83
type of instruction, as related to anagram performance.
On the other hand, there was a significant relationship
between test anxiety and grade performance, with high
test-anxious subjects doing significantly more poorly than
middle or low test-anxious subjects in their general psychol¬
ogy course. When analyses were conducted using test anxiety,
self-efficacy, and instructions as independent variables
in the ANOVA model, similar results were obtained. Although
neither the anagram analysis nor the grade analysis were
significant overall, the grade analysis found that test
anxiety was significantly related to grade performance.
The grade analyses replicated the findings in the regression
analyses that test anxiety has a negative relationship to
school performance. These analyses also confirmed previous
test anxiety studies which have found that test anxiety has
a negative effect on cognitive task performance (Wine,
1971). With anagram as the dependent variable, however,
no significant results were found. This study did not
replicate those previous test anxiety studies which found
a test anxiety by instruction interaction (Sarason, 1961;
Sarason, 19 73b); however, it did replicate those studies
which found a general negative relationship between test
anxiety and performance (Russell & Sarason, 1965; Crager,
1959). Apparently, the instructions, subjects, or experi¬
mental conditions in this study differed, in some way, from
previous studies which elicited a test anxiety by instruction

84
interaction. It also appears from these ANOVA analyses,
that the grade measure is more sensitive to the effects
of test anxiety. This is fortunately the case, because
the anagram task is a contrived laboratory measure, and
grade is a meaningful measure of performance in school.
The ANOVA analyses with the anagram task did not
replicate the findings from the multiple regression model,
that test anxiety and self-efficacy interact to predict
performance. It is possible that the results did not
replicate because the effect is small, and the ANOVA
analysis is not sensitive enough to pick up this relation¬
ship. Cohen (1968) has proposed that linear regression
analyses increase the size of statistical effects over
ANOVA models by up to 36%.
Although there are major limitations on the interpre¬
tations to be made from the regression analyses for anagram
performance, given the amount of variance accounted for
by the model, some speculative and tentative interpretations
can be made from the finding of an interaction between test
anxiety and self-efficacy. Although it is only a tentative
finding, this interaction is rather interesting and worth
discussing. The interaction between test anxiety and self-
efficacy can be analyzed by breaking it up into two subsec¬
tions algebraically. First one can examine the effect of
test-anxiety on anagram performance, taking into account the
effect of self-efficacy. Then one can examine the effect
of self-efficacy on performance, taking into account the

85
effect of test-anxiety. In order to interpret these
effects, one takes the original regression equation for the
total sample (n=213), and, setting the equation to zero, one
can compare the effect of a one unit change in test-anxiety
with no change in test-anxiety, to assess the effect of test
anxiety on performance. Similarly, one can compare the
effect of a one unit change in self-efficacy with no
change in self-efficacy, to assess the effect of self-
efficacy on performance. The test anxiety effect was
found, as follows:
A
Anl=7.6518-.2375(TA)+.0012(TA*SE)
An2=7.6518-.2375(TA+1)+.0012((TA+1)*(SE))
A A
Anl-An2=-.2375+.0012(SE)
SE=.2375/.0012
SE=197.9
(TA= test-anxiety, SE=self-efficacy, ^n=expected
anagram score)
A one unit change in test anxiety score is expected to
correspond to a change in anagram score = -.2375 + .0012
(self-efficacy). If one solves the equation (setting the
equation to zero), it is found that at a self-efficacy score
of 197.9, test anxiety has no effect. When self-efficacy
scores exceed 197.9, then the effect of test anxiety on
anagram performance is positive. When self-efficacy scores
fall below 197.9, then test anxiety has a negative effect
on performance.

86
These results suggest that for two persons with high
self-efficacy, if one has relatively higher test anxiety,
then that person will have better anagram performance.
Conversely, for two subjects with low self-efficacy, if
one has higher test anxiety relative to the other, then his
anagram performance is expected to be lower than the other
person's. The magnitude of the expected difference will
necessarily depend on the actual placement of the self-
efficacy scores within the distribution. Test anxiety can,
therefore, have either a relatively positive or negative
effect depending on the level of self-efficacy. At higher
levels of self-efficacy, test anxiety has a positive effect,
and at lower levels of self-efficacy, test anxiety has a
negative effect.
These results suggest that when a person has high
self-efficacy, i.e. , a strong belief in their ability
persist at, and to master academic situations, then the
worry they experience regarding tests and other evaluative
situations is facilitative of performance. If, on the other
hand, a person has low self-efficacy, or believes that they
lack the coping skills to master academic situations, then
their worry and preoccupation regarding testing situations
is likely to be debilitating to performance. For less
self-efficacious persons, as test anxiety increases,
performance is likely to decrease. Therefore, test
anxiety can be either facilitative or debilitative,

87
depending upon the person's general expectancies regarding
their ability to successfully cope with academic tasks.
Although the general trend in the literature has been
to perceive of test anxiety as debilitative, theorists
in related areas have suggested that "worry" can have
positive aspects. Janis (1958) has discussed the "work of
worrying" as a means of dealing effectively with a challenging
reality situation. Worry can be utilized as a means of
mobilizing strategies ahead of time, to help solve an
approaching difficult situation. Janis has theorized that
when worry is experienced adaptively, the person becomes
involved in mental rehearsal which produces reality-based
cognitions and expectations, which function as a source of
hope and reassurance. Of course, worry can also become
debilitative. Marmor (1958) makes a distinction between
"realistic" and "neurotic" worrying. He describes realistic
worry as "a defensive function of the ego, the purpose of
which is either to ward off an anticipated real trauma or
to deal with the painful consequences of one already
experienced." He theorizes that when this realistic worry
is successful, the mental activity involved leads to action
and mastery of the threat situation. However, some people
engage in "neurotic" worry, and these people are unable to
achieve "inner mastery":
If the ego fails in its integrative task
. . . decompensation takes place in the
form of ineffective or circular worrying

88
(which is identical with what is usually
clinically labeled as obsessional rumin¬
ation)—or else the effort to deal with
the problem at an intellectual level is
given up entirely, and regression takes
place to the emotional level, in which
anxiety, bound or unbound, is the dominant
feature. The clinical picture may then be
that of an anxiety state, phobia, or a
conversion hysteria. (Marmor, in Janis,
1958, p. 377)
Although the clinical state Marmor is describing might
be more extreme than that of a test-anxious college student,
his description of "realistic" and "neurotic" worry appears
to be somewhat analogous to the way test anxiety might
operate, i.e. in either an adaptive or maladaptive fashion.
In 1960, Alpert and Haber conceived of the idea that test
anxiety could have both debilitative and facilitative
components. They developed a scale, the Achievement
Anxiety Test (AAT), which attempted to measure facilitative
and debilitative test anxiety. The test consisted of the
Facilitating Anxiety Scale, and the Debilitating Anxiety
Scale. The Facilitating Anxiety Scale correlated positively
with academic performance (grade point average), and the
Debilitating Anxiety Scale correlated negatively with
academic performance. Using multiple correlations to pre¬
dict performance, Alpert and Haber (1960) found that using
both the plus and minus scales together enhanced performance
significantly beyond use of the individual scales. These
scales have not been used as often as the more conventional
debilitating anxiety scales, but this may be due primarily

89
to the investigator's a priori conceptualization of test
anxiety as a linear dimension, affecting performance solely
in a negative fashion. The results of the current study
suggest that test anxiety may be a more complex dimension,
interacting with another cognitive-mediational variable,
self-efficacy, to affect performance in either positive or
negative ways.
These results suggest that in a highly self-efficacious
person, some cognitive processes may occur which counteract
the maladaptive effects of test anxiety, and channel the
person into productive, positive, self-referential thinking.
This positive thinking may then serve as a motivator for
adaptive performance. In individuals with low self-efficacy,
as the degree of test anxiety increases, the negative effect
of test anxiety on performance becomes greater. Therefore,
the combined effects of low self-efficacy and low test
anxiety on cognitive processes are presumed to be maladaptive
in nature, leading to decreased performance.
Although several interpretations have been generated
regarding the interaction of test anxiety and self-efficacy,
it must be remembered that the above interpretations relate
to only 1/2 of the interactional effect. It is also necessary
to interpret the effect of self-efficacy on performance,
taking into account the effect of test-anxiety. This
interpretation is generated in the same manner. One compares
the effect of a one unit change in self-efficacy to no

90
change in self-efficacy, on anagram performance (the algebra
is abbreviated in this case):
A
Anl=7.6518-0.2375(TA)+0.0012(TA)(SE)
A A
Anl-An2=0.0012(TA)
Self-Efficacy effect=0.0012(TA)
This result can be interpreted as saying that there is a
positive effect of self-efficacy on anagram performance,
and this effect increases with increasing test anxiety.
Self-efficacy has a greater effect on performance at higher
levels of test anxiety, and a lesser effect at lower levels
of test anxiety. For some reason, self-efficacy appears to
interact more strongly with test anxiety in predicting
performance, at higher levels of test anxiety. Perhaps,
at high levels of test anxiety, a person is more self-
focused and concerned with how well they are going to
perform. Self-referential thought may, therefore, have a
greater bearing on performance for these individuals.
Consequently, it is more crucial to account for their
level of self-efficacy. At lower levels of test anxiety,
the person may not be as invested in how well they will do
on the task, so it doesn't matter what they believe they
can do on academic tasks. There may be more of a discrepancy
between their general beliefs about themselves in academic
situations, and their performance. Situational or task

91
variables may be more significant predictors of performance,
in the case of low test-anxious persons.
Although these results are quite interesting, they
must remain highly speculative due to the small amount of
variance accounted for by this model. The pattern to the
residuals in the earlier models had suggested that other
variables might be systematically accounting for the
results. Perhaps other cognitive factors, not currently
in this model, are important in predicting task performance.
Other variables, such as persistence at the task, or how
much the subject cared about his performance, or the
quantity of negative self-statements made by the subjects
may have had a stronger effect on performance than those
variables systematically measured by this study. The
Harris and Johnson (1980) and Harris (1980) research cited
earlier suggests that mental imagery is an important factor
in the effects of test anxiety on performance. Harris
(19 80) found that under anxiety-evoking situations, high
test-anxious individuals tend to have negative imagery,
which is vivid and persistent. When Harris (1980) trained
the subjects to visualize images of competence based on
previous personal success experiences, the individuals
manifested a reduction in test anxiety and improved
academic performance. Therefore, if one were to assess
the extent of negative imagery in subjects, perhaps this
would prove to be a stronger predictor than test anxiety.

92
An additional limiting factor to the test anxiety/
self-efficacy predictor model is the finding that the
relationship of these variables to anagram performance did
not replicate in the grade analyses. The final regression
model for grade yielded test anxiety and SATQ as the
significant predictors. The rationale for using grade as
an additional dependent variable was to see if the variables
selected to predict anagram performance, a laboratory analog
task, would generalize to a meaningful measure in an
academic setting. From these results, it appears that
self-efficacy does not predict grade performance, although
test anxiety does. This suggests that the particular model
generated to predict anagram performance does not entirely
generalize to this external measure of achievement. Test
anxiety is related to performance in both circumstances?
however, the test anxiety effect is different under each
circumstance. The relationship between test anxiety and
anagram performance is nonlinear (positive under some
circumstances and negative under others). The relationship
of test anxiety to grade performance is negative. The more
test anxiety a person has, the more likely it will have a
debilitating effect on performance. Thus, with grade as
the measure of performance, test anxiety appears to have
the linear effect expected from much of the literature.
This is not surprising, given that the means for obtaining
classroom grades are generally tests, and the testing

93
situation, creating the "evaluative" situational cues,
elicits those effects predicted for test anxiety under
"evaluative" conditions. In addition, as might be
expected, aptitude scores appear to have a strong effect
in the prediction of grades. In this case, SATQ significantly
predicts grades in General Psychology. It is unclear why
self-efficacy is predictive of anagram performance but is
not predictive of grades. Additional studies will be
necessary to answer this question.
The finding that the "best" predictor model for anagram
performance is not identical to the "best" model for grade
performance, does not totally negate the results found for
the anagram model. It suggests that either the model has
limited generalizability, or that the external measure,
itself, may not have been adequate to bring out the potential
effects of test anxiety and self-efficacy. From a methodo¬
logical perspective, both the grade measure and anagram
measure had some limitations. The anagram measure only
used 13 anagrams (the typical number of anagrams used in
previous studies), which may have limited the sensitivity
of the measure. In addition, the grade measure, final grade
in General Psychology only had a potential for eight data
points, also possibly limiting the sensitivity of this
measure.
Another limitation to the interpretations to be made
from the above models is based on a brief analysis of the

94
psychometric properties of the self-efficacy scale (Measure
of Academic Self-Efficacy). Although extensive analyses
were not performed on this instrument, it appeared that the
distribution of scores was somewhat skewed to the upper
end of possible scores. The range of scores was from 164
to 298, with a mean of 240.34, and a standard deviation of
27.34. In terms of the potential range on this scale, a
subject could conceivably score between 60 and 300. In
actuality, most subjects scored above 200, on the scale.
If one realizes that there is a 240 point potential range
on this scale, one would expect the mean of a normal distri¬
bution to fall somewhere around 180 (halfway between 60 and
300). Thus, although self-efficacy was found to be a
significant predictor of anagram performance, in interaction
with test anxiety, the results must be interpreted with
caution, because the range of scores on the MASE was
relatively narrow. There were relatively few subjects
scoring at the low self-efficacy end of the scale. This
makes interpretation of these results somewhat difficult,
because no one in this sample had truly low self-efficacy.
Most subjects appeared to have relatively high, to high
self-efficacy. From a reliability standpoint, however,
the descriptive statistics on the MASE for this sample were
strikingly similar to those found by Lalonde (1979) in her
study. For her sample of n=344, the distribution yielded
a mean score of 235, a standard deviation of 27.85, a

95
minimum value of 149, and a maximum value of 298. In her
re-test sample (n=303), her descriptive statistics were
similar; however,she did have a broader range (126-300).
It appears probable, from qualitative examination of
the MASE items, that the reason the distribution is so
skewed is because of the global nature of the individual
items. All of the items deal with the notion of persistence
at difficult tasks, and continuing to persevere in spite
of early failure, or giving up easily. However, the items
do not address specific subject areas, nor do they particularly
address academic issues. Hence, there may be a tendency
for most subjects to answer many of the items in a similar
fashion, without much discrimination. If a person feels,
in general, that they persist at difficult tasks, then there
is no reason for them to answer some of the individual
items differently. There is,therefore, a built-in bias
in this test toward a response set. If the items were
more diversified, perhaps this tendency toward a skewed
distribution of high self-efficacy scores would be avoided.
The present study found that while an aptitude measure
was a significant predictor of cognitive task performance,
other variables were also significant in predicting
cognitive performance. With an anagram task as the cognitive
performance measure, test anxiety and self-efficacy were
significant predictors, whereas aptitude measures were not.
Test anxiety and self-efficacy also appeared to predict

96
cognitive performance best in interaction with each other,
rather than as main effects. These results support Tobias'
review of the test anxiety literature (Tobias, 1980), which
suggests that test anxiety measures are likely to increase
in their predictive powers when used in combination with
measurement devices which assess positive orientations to
evaluation. For example, Crawford (1976) found that test
anxiety significantly interacted with "need for achievement"
in predicting performance under different instructional
conditions, using programmed instructions.
The test anxiety/self-efficacy interaction, although
a small effect, warrants further investigation. However,
before additional studies are conducted, it appears necessary
to do some further test development with the construct of
self-efficacy. This could either be done through modifica¬
tion of Lalonde's Measure of Academic Self-Efficacy (1979),
or through the development of another measure of self-
efficacy. Perhaps a highly specific measure of academic
self-efficacy, similar to those employed in Bandura's
(1982) research would be a more powerful instrument. If
an academic self-efficacy measure were developed similar
to those used in the Bandura studies, it would have to be
highly specific and pertain to the task or tasks required
of subjects during the study. Bandura (1982) believes
that self-efficacy is not a global construct, and therefore
assessment of expectancies in specific skill areas is the

97
only way to measure self-efficacy. It might be interesting
to compare a global measure of academic self-efficacy to a
specific skill-based measure of academic self-efficacy as
predictors of performance.
This study also found that under certain conditions,
test anxiety can have a positive effect on performance,
specifically, when subjects have relatively higher self-
efficacy. It would be interesting to investigate the
"contents of consciousness" (Wine, 1980) of high test-
anxious subjects, to determine how some persons utilize
their worry over tests in a constructive fashion, and how
others may allow their worry to become paralyzing. A
person's level of self-efficacy may somehow counteract
the worry that they experience over test situations,
enabling the worry to become constructive rather than
destructive. The Harris (1980) and Harris and Johnson
(1980) studies certainly suggest that images of competence
can reduce test anxiety and improve performance. The
self-efficacy effect may also have something to do with
the attributions that different individuals make regarding
their own effect on their performance. If they attribute
successful performance to luck or other situational vari¬
ables, it would appear that they would be less likely to
mobilize their own resources to make an attempt to be
successful. If they attribute successful performance to
effort, they will be more likely to invest effort to succeed.
It may be that highly self-efficacious individuals are more

98
likely to make "effort" attributions. Weiner (1982) and
others have begun to look at the relationship between
anxiety and attribution. Schunk (1981) has examined the
effects of effort attribution training and modeling on
self-efficacy judgments and performance. Effort attribution
training was helpful in producing the highest congruence
between efficacy judgment and performance. These studies
suggest that attributions of causality are highly related
to the self-evaluative process.
Finally, since the test anxiety/self-efficacy model
only accounted for a small percentage of the variance in
performance, it is necessary to investigate what other
variables are accounting for performance differences. The
answer to this question may lie in careful assessment of
situational variables, such as the degree of difficulty of
the task, or the environment presented by the experimenter
(e.g. stressful, reassuring). The answer may also be
found in careful assessment of the subject's individual
interpretation of the situation, that is, their individual
cognitive appraisal of task difficulty, their expectancy
as to how well they will perform on that task, the kinds
of negative and/or positive self-statements they make, their
personal attributions of causality for success or failure,
or their mental imagery regarding academic situations.
The current direction, in both the test anxiety and

99
self-efficacy research areas, is toward greater under¬
standing of the self-motivational processes underlying
performance. It would be most desirable for these two
research areas to pool forces in developing answers to
these questions.

APPENDICES

APPENDIX A
ANAGRAMS TASK
Anagram
NMGOINR
SPRUUE
SUTCBII
SCLIAO
EVSUORN
RSANEO
IMTCELA
ELSAUX
GINREFO
CEPART
EONSPR
EIVARR
THWGIE
Answer
101

102
APPENDIX B
INTERCORRELATION MATRIX ON THE INDEPENDENT AND DEPENDENT
VARIABLES FOR THE MAIN SAMPLE (n=162)
Testanx
Selfef
SATV
SATQ
Grade3
Anagra
Testanx
-.38***
-.33***
-.33***
-.22**
. 13
Selfef
. 10
. 05
.11
.05
SATV
.45***
. 22**
-.16*
SATQ
.17*
. 04
Grade3
-.009
Anagram
Note: Testanx=Test Anxiety; Selfef=Self-Efficacy; SATV=Scholastic Aptitude
Test-Verbal; SATQ=Scholastic Aptitude Test-Quantitative
* p < .05
** p < .01
*** p < .0001
aAll correlations of grade with the other variables were done on a
sample of n=160, because two subjects had missing grade data.

103
APPENDIX C
INTERCORRELATION MATRIX ON THE INDEPENDENT
DEPENDENT VARIABLES FOR TOTAL SAMPLE3
AND
Testanx Selfef
SATV
SATQ
Grade
Anagram
Testanx
-.37**
(n=230)
-.34**
(n=214)
-.28**
(n=214)
-.19*
(n=227)
.03
(n=232)
Selfef
.08
(n=215)
.05
(n=215)
.11
(n=226)
.09
(n=232)
SATV
. 46**
(n=216)
. 20*
(n=212)
-.05
(n=216)
SATQ
. 21*
(n=212)
.05
(n=216)
Grade
.04
(n=228)
Anagram
Note: Testanx = Test Anxiety; Selfef = Self Efficacy; SATV = Scholastic
Aptitude Test-Verbal; SATQ = Scholastic Aptitude Test-Quantitative
* p < .005
** p < .0001
Includes all subjects in the main sample, the validation sample,
and those with missing data.

REFERENCES
Allen, G.J. The behavioral treatment of test anxiety:
Recent research and future trends. Behavior Therapy,
1972, 3, 253-262.
Alpert, R., & Haber, R. Anxiety in academic achievement
situations. Journal of Abnormal and Social Psychology,
1960, 61, 207-215.
Aponte, J., & Aponte, C. Group preprogrammed systematic
desensitization without the simultaneous presentation
of aversive scenes with relaxation training. Behavior
Research and Therapy, 1971, 9_, 337-346.
Bandura, A. Self-efficacy: Toward a unifying theory of
behavioral change. Psychological Review, 1977, 84,
191-215.
Bandura, A. Self-efficacy mechanism in human agency.
American Psychologist, 1982 , 37_, 122-147.
Bandura, A., & Adams, N. Analysis of self-efficacy theory
of behavior change. Cognitive Therapy, 1977, 1, 287-
310.
Bandura, A., Adams, N., & Beyer, J. Cognitive processes
mediating behavior changes. Journal of Personality
and Social Psychology, 1977, 3_5' 125-139.
Bandura, A., Adams, N., Hardy, A., & Howells, G. Tests
of the generality of self-efficacy theory. Cognitive
Therapy and Research, 1980, 4, 39-66.
Bartlett, R. Program for experiments on thinking.
Quarterly Journal of Experimental Psychology, 1950,
2, 145-152.
Beck, A.T., Weissman, A., Lester, D., & Trexler, L. The
measurement of pessimism: The hopelessness scale.
Journal of Consulting and Clinical Psychology, 1974,
42, 861-865.
104

105
Brody, E.B., & Brody, N. Intelligence: Nature, determinants,
and consequences. New York: Academic Press, 1976.
Chiang-Liang, R., & Denney, D.R. Applied relaxation as
training in self-control. Journal of Counseling
Psychology, 1976, 2_3 , 183-189".
Cohen, J. Multiple regression as a general data analytic
system. Psychological Bulletin, 1968, 70, 426-443.
Conry, R., & Plant, W.J. WAIS and group test predictions
of an academic success criterion: High school and
college. Educational and Psychological Measurement,
1965, 25_, 493-500.
Crager, R.L. The relation of anxiety, sex, and instructions
to performance and verbal behavior during anagram
solution (Doctoral dissertation, University of Washing¬
ton, 1959) . Dissertation Abstracts International,
1960, 20, 2900-2901.
Crandall, V.C., Katkovsky, W., & Crandall, V.J. Children's
belief in their own control of reinforcements in
intellectual-academic achievement situations. Child
Development, 1965, 3_6, 91-109.
Crawford, J. An investigation of internally and externally
operative variables in an interactive learning model.
Unpublished manuscript (Report No. 76-8), Austin,
Texas: Research and Development Center for Teacher
Education, 1976.
Cronbach, L.J. Essentials of psychological testing. 3rd
Edition. New York: Harper & Row, 1970.
Deffenbacher, J.L. Relationship of worry and emotionality
to performance on the Miller Analogies Test. Journal
of Educational Psychology, 1977, 6J?, 191-195.
Deffenbacher, J.L. Worry, emotionality, and task-generated
interference in test anxiety: An empirical test of
attentional theory. Journal of Educational Psychology,
1978, 70(2), 248-254.
Deffenbacher, J.L. Worry and emotionality in test anxiety.
In: I.G. Sarason (Ed.), Test anxiety: Theory,
research, and applications. Hillsdale, N.J.: Lawrence
Erlbaum Associates, 1980.

106
Deffenbacher, J.L., Mathis, H., & Michaels , A. C. Self-control
procedures in the reduction of targeted and nontargeted
anxieties. Journal of Counseling Psychology, 1979,
26_, 120-127.
Deffenbacher, J.L., & Parks, D.H. A comparison of tradi¬
tional and self-control systematic desensitization.
Journal of Counseling Psychology, 1979, £6, 93-97.
Denney, D.R. Self-control approaches to the treatment of
test anxiety. In: I.G. Sarason (Ed.), Test anxiety:
Theory, research, and applications. Hillsdale, N.J.:
Lawrence Erlbaum Associates, 1980.
Denney, D.R., & Rupert, P.A. Desensitization and self-
control in the treatment of test anxiety. Journal of
Counseling Psychology, 1977, 4_5, 272-280.
Doctor, R.M., & Altman, F. Worry and emotionality as
components of test anxiety: Replication and further
data. Psychological Reports, 1969 , 24_, 563-568.
Donner, L., & Guerney, B. Automated group desensitization
for test anxiety. Behavior Therapy and Research, 1969,
7, 1-13.
Dudek, S.A., Goldberg, J.S., Lester, E.P., & Harris, B.R.
The validity of cognitive, perceptual-motor, and
personality variables for prediction of achievement
in grade I and grade II. Journal of Clinical Psychology,
1969, 25, 165-170.
Dusek, J.B. The development of test anxiety in children.
In: I.G. Sarason (Ed.), Test anxiety: Theory,
research, and applications"! Hillsdale, N.J.: Lawrence
Erlbaum Associates, 1980.
Dusek, J.B., Mergler, N.L., & Kermis, M.D. Attention,
encoding, and information-processing in low and
high test-anxious children. Child Development, 1976,
£7, 201-207.
Easterbrook, J.A. The effect of emotion on cue utilization
and the organization of behavior. Psychological Review
1959, £6, 183-201.
Emery, J., & Krumboltz, J. Standard versus individualized
hierarchies in desensitization to reduce test anxiety.
Journal of Counseling Psychology, 1967, 14, 204-209.

107
Fishman, J.A., & Pasanella, A.K. College admission-
selection studies. Review of Educational Research,
1960 , 30_, 198-310.
Freeling, N.W., & Shemberg, K. The alleviation of test
anxiety by systematic desensitization. Behavior
Research and Therapy, 1970, 8, 293-300.
Ganzer, V.J. Effects of audience presence and test
anxiety on learning and retention in a serial
learning situation. Journal of Personality and
Social Psychology, 1968, 194-199.
Garlington, W.K., & Cotier, S. Systematic desensitization
of test anxiety. Behavior Research and Therapy, 1968,
6, 247-256.
Geen, R.G. Test anxiety, observation, and range of cue
utilization. British Journal of Social and Clinical
Psychology, 1976, 15, 253-259.
Geen, R.G. Test anxiety and cue utilization. In: I.G.
Sarason (Ed.), Test anxiety: Theory, research, and
applications. Hillsdale, N.J.: Lawrence Earlbaum
Associates, 1980.
Goldfried, M.R. Systematic desensitization as training in
self-control. Journal of Consulting and Clinical
Psychology, 19 71, TT~, 228-234.
Goldfried, M.R., Linehan, M.M., and Smith, J.L. Reduction
of test anxiety through cognitive restructuring.
Journal of Consulting and Clinical Psychology, 1978,
¥(D7'Í2'-T^
Hamilton, V. Socialization anxiety and information processing
A capacity model of anxiety-induced performance. In:
I.G. Sarason & C.D. Spielberger (Eds.), Stress and
anxiety, Vol. 2, Washington, D.C.: Hemisphere, IU75,
45-68.
Harris, G.M. A comparison of several different imagery
instruction procedures for the treatment of test
anxiety (Doctoral dissertation, University of Florida,
1980). Dissertation Abstracts International, 1981, 42,
770B (University Microfilms No. 8115639).
Harris, G.M., & Johnson, S. Comparison of individualized
covert modeling, self-control desensitization, and study
skills training for alleviation of test anxiety.
Journal of Consulting and Clinical Psychology, 1980,
48, 186-194.
Helwig, J.L., & Council, K.A. SAS user's guide, 1979
edition. Cary, N.C.: SAS Institute, Inc., 1979.

108
Holroyd, K. Cognition and desensitization in the group
treatment of test anxiety. Journal of Consulting and
Clinical Psychology, 1976, 4V, 991-1001.
Holroyd, K., Westbrook, T., Wolf, M., & Badhorn, E.
Performance, cognition, and physiological responding
in test anxiety. Journal of Abnormal Psychology,
1978, 87_, 442-451.
Hoyt, C.L. Test reliability estimated by analysis of
variance. Psychometrika, 1941, 6, 153-160.
Jackson, D.N. Personality research form. Goshen, N.Y.:
Research Psychologists Press, Inc., 1966.
Jacobsen, E. Progressive relaxation. Chicago: University
of Chicago, 1938.
Janis, I.L. Psychological stress. New York: Wiley, 1958.
Johnson, S.M., & Secrest, L. Comparison of desensitization
and progressive relaxation in treating test anxiety.
Journal of Consulting and Clinical Psychology, 1968,
T27~28'0-28'6~
Katahn, M., Strenger, S., Cherry, N. Group counseling and
behavior therapy with test-anxious college students.
Journal of Consulting Psychology, 1966, 30, 544-549.
Lalonde, B. The construction and validation of a measure of
academic self-efficacy (Doctoral dissertation, Univer¬
sity of Toronto, 1979). Dissertation Abstracts
International, 1980, £0_, 4510A.
Lang, P. A bio-informational theory of emotional imagery.
Psychophysiology, 1979 , 16., 495-512.
Lavigne, J.V. The relative efficacy of cognitive-behavior
rehearsal and systematic desensitization in the treat¬
ment of test anxiety (Doctoral dissertation, University
of Texas at Austin, 1974). Dissertation Abstracts
International, 1975, 35, 2437B-2438B (University
Microfilms No. 742489TT.
Lavin, D.E. The prediction of academic performance: A
theoretical analysis and review of research. New
York: Russell Sage Foundation, 1965.
Liebert, R.M., & Morris, L.W. Cognitive and emotional
components of test anxiety: A distinction and some
initial data. Psychological Reports, 1967, 20, 975-
978.

109
Maier, S.R., & Seligman, M.E. Learned helplessness:
Theory and evidence. Journal of Experimental Psy¬
chology, 1976, 105, 3-46.
Mandler, G., & Sarason, S.B. A study of anxiety and learning.
Journal of Abnormal and Social Psychology, 1952, 47,
166-173. —
Mandler, G. , & Watson, D.L. Anxiety and the interruption
of behavior. In: C.D. Spielberger (Ed.), Anxiety
and behavior. New York: Academic Press, 1966.
Marchetti, A., McGlynn, F.D., & Patterson, A.S. Effects
of cue-controlled relaxation, a placebo treatment, and
no treatment on changes in self-reported and psycho-
physiological indices of test anxiety among college
students. Behavior Modification, 1977, 1_, 47-72.
Marlett, N.J., & Watson, D. Test anxiety and immediate or
delayed feedback in a test-like avoidance task.
Journal of Personality and Social Psychology, 1968,
8, 200-203.
Marmor, J. The psychodynamics of realistic worry.
Unpub1ished manuscript, 195Ó.
McClelland, D.C. Testing for competence rather than
intelligence. American Psychologist, 1973, 2_8, 1-13.
McClelland, D.C., Atkinson, J.W., Clark, R.A., & Lowell,
E.J. A scoring manual for the achievement motive.
In J.W. Atkinson (Ed.), Motives in fantasy, action,
and society. Princeton: Van Nostrand, 1958.
Meichenbaum, D.H. Cognitive modification of test anxious
college students. Journal of Consulting and Clinical
Psychology, 1972, _39 , 370-380.
Meichenbaum, D.H., Gilmore, J., & Fedoravicius, A.
Group insight vs. group desensitization in treating
speech anxiety. Journal of Consulting and Clinical
Psychology, 1971 410-421.
Morris, L.W., & Liebert, R.M. Effects of anxiety on timed
and untimed intelligence tests. Journal of Consulting
and Clinical Psychology, 1969 , 33~j 240-244.
Morris, L.W., & Liebert, R.M. Relationship of cognitive
and emotional components of test anxiety to physiologi¬
cal arousal and academic performance. Journal of
Consulting and Clinical Psychology, 1970, 35, 332-
3IT.

110
Morris, L.W., & Perez, T.L. Effects of test-interruption
on emotional arousal and performance. Psychological
Reports, 1972, 31, 559-564.
Neale, I.M., & Katahn, M. Anxiety, choice, and stimulus
uncertainty. Journal of Personality, 1968, 36, 235-
245.
Osterhouse, R.A. Desensitization and study-skills training
as treatment for two types of test-anxious students.
Journal of Counseling Psychology, 1972, 19., 301-307.
Paul, G.L. Modifications of systematic desensitization based
on case study. Paper presented at the meeting of the
Western Psychological Association, Portland, Oregon,
April, 1964.
Paul, G.L. Insight versus desensitization in psychotherapy.
Stanford, Calif.: Stanford University Press, 1966.
Paul, G.L., & Eriksen, C.W. Effects of test anxiety on
"real life" examinations. Journal of Personality, 1964,
32, 480-494.
Piers, E.B. Manual for the Piers-Harris children's self-
concept scale (The way I feel about myself). Nashville,
Tenn.: Counselor Recordings and Tests, 1969.
Plant, W.J., & Richardson, H. The IQ of the average college
student. Journal of Counseling Psychology, 1958, 5,
229-231.
Rimm, D.C., & Masters, J.C. Behavior therapy: Techniques
and empirical findings. New York: Academic Press,
1974.
Romano, J.L., & Cabianca, W.A. EMG biofeedback training
versus systematic desensitization for test anxiety
reduction. Journal of Counseling Psychology, 1978,
25, 8-13.
Rotter, J.B. Generalized expectancies for internal vs.
external control of reinforcement. Psychological
Monographs, 1966, J30 (1, Whole No. 609) .
Russell, D.G., & Sarason, I.G. Test anxiety, sex, and
experimental conditions in relation to anagram solution.
Journal of Personality and Social Psychology, 1965,
1, 493-496.

Ill
Russell, R.K., Miller, D.E., & June, L.M. A comparison
between group systematic desensitization and cue-
controlled relaxation in the treatment of test anxiety.
Behavior Therapy, 1975, 6_, 172-177.
Russell, R.K., & Sipich, J.F. Cue-controlled relaxation
in the treatment of test anxiety. Journal of Behavior
Therapy and Experimental Psychiatry-^ 1973, 47-49.
Russell, R.K., Wise, R., & Stratoudakis, J.P. Treatment
of test anxiety by cue-controlled relaxation and
systematic desensitization. Journal of Counseling
Psychology, 1976, 2_3, 563-566.
Samuels, M., & Samuels, N. Seeing with the mind's eye:
The history, techniques, and uses of visualization.
New York: Random House, 1975.
Sarason, I.G. Interrelationships among individual differ¬
ence variables, behavior in psychotherapy, and verbal
conditioning. Journal of Abnormal and Social
Psychology, 1958, 56] 339-344.
Sarason, I.G. Relationships of measures of anxiety and
experimental instructions to word association test
performance. Journal of Abnormal and Social
Psychology, 1959, _59, 37-42.
Sarason, I.G. The effects of anxiety and threat on solution
of a difficult task. Journal of Abnormal and Social
Psychology, 1961, 6_2, 165-168.
Sarason, I.G. Verbal learning, modeling, and juvenile
delinquency. American Psychologist, 1968, 23, 254-
266.
Sarason, I.G. Test anxiety and the model who fails. Journal
of Personality and Social Psychology, 1972, 2_2, 410-413.
Sarason, I.G. Test anxiety and cognitive modeling. Journal
of Personality and Social Psychology, 1973a, 2_8, 58-61.
Sarason, I.G. Test anxiety and social influence. Journal
of Personality, 1973b, .41, 261-271.
Sarason, I.G. Anxiety and self-preoccupation. In: I.G.
Sarason & C.D. Spielberger (Eds.), Stress and anxiety,
(Vol. 2), Washington, D.C.: Hemisphere, 1975, 27-44.
Sarason, I.G. The Test Anxiety Scale: Concept and research.
In: C.D. Spielberger & I.G. Sarason (Eds.), Stress and
anxiety (Vol. 5), Washington, D.C.: Hemisphere, 1978.

112
Sarason, I.G., & Ganzer, V.J. Anxiety, reinforcement,
and experimental instructions in a free verbalization
situation. Journal of Abnormal and Social Psychology,
1962, 65, 300-307.
Sarason, I.G., & Ganzer, V.J. Effects of test anxiety and
reinforcement history on verbal behavior. Journal
of Abnormal and Social Psychology, 1963, 677 513-519.
Sarason, I.G., Pederson, A., & Nyman, B.A. Test anxiety
and the observation of models. Journal of Personality,
1968, 36, 493-511.
Sarason, I.G., & Stoops, R. Test anxiety and the passage
of time. Journal of Consulting and Clinical Psychology,
1977, 46(1) , 102-109.
Sarason, S.B., Davidson, K., Lighthall, F., & Waite, R.
Classroom observation of high and low-anxious children.
Child Development, 1958, 29_, 287-295.
Sarason, S.B., Davidson, K., Lighthall, F., Waite, R., &
Ruebush, B.K. Anxiety in elementary school children.
New York: Wiley, 1960.
Sarason, S.B., Mandler, G., & Craighill, P.G. The effects
of differential instructions on anxiety and learning.
Journal of Abnormal and Social Psychology, 1952,
£7, 561-565.
Sargent, S.S. Thinking processes at various levels of
difficulty. Archives of Psychology, New York, 1940,
No. 249.
Schunk, D. Modeling and attributional effects on children's
achievement: A self-efficacy analysis. Journal of
Educational Psychology, 1981, 1_3 , 93-105.
Speigler, M.D., Morris, L.W., & Liebert, R.M. Cognitive
and emotional components of test anxiety: Temporal
factors. Psychological Reports, 1968, 22_, 451-456.
Spielberger, C.D., Anton, W.D., & Bedell, J. The nature
and treatment of test anxiety. In: M. Zuckerman,
& C.D. Spielberger (Eds.), Emotions and anxiety:
New concepts, methods, and applications. New York:
Lawrence Earlbaum Associates, 1976, 317-345.
Suinn, R.M. The desensitization of test anxiety by group
and individual treatment. Behavior Research and
Therapy, 1968, 6_, 385-387.
Tait, A.M. Competence in fifth form students. Unpublished
paper, Monash University, 1973.

113
Taylor, R. Personality traits and discrepant achievement:
A review. Journal of Counseling Psychology, 1964, 11,
76-82. —
Tobias, S. Anxiety and instruction. In: I.G. Sarason
(Ed.), Test anxiety: Theory, research, and applica¬
tions. Hillsdale, N.J.: Lawrence EarlPaum Associates,
imr.
Tolman, E.C. Cognitive maps in rats and men. Psychological
Review, 1948, 55_, 189-208.
Wagaman, G.L., Cormier, W.H., & Cormier, L.S. Cognitive
modification of test-anxious students. Paper presented
at the meeting of the American Educational Research
Association, Washington, D.C., 1975.
Watson, D., & Friend, R. Measurement of social-evaluative
anxiety. Journal of Consulting and Clinical Psychology,
1969 , 33_, 448-457.
Weiner, B. An attribution theory of motivation and emotion.
In: H. Krohne & L. Laux (Eds.), Achievement, stress,
and anxiety. New York: Hemisphere Publishing Corpor-
ation, 1982.
West, C.K., Lee, J.F., & Anderson, T.H. The influence of
test anxiety on the selection of relevant from
irrelevant information. The Journal of Educational
Research, 1969, 63_, 51-52.
Wine, J.D. Investigation of an attentional interpretation
of test anxiety (Doctoral dissertation, University of
Waterloo, 1970). Dissertation Abstracts International,
1971, 32^ 1199B-1200B"!
Wine, J.D. Test anxiety and direction of attention.
Psychological Bulletin, 1971, 76_, 92-104.
Wine, J.D. Cognitive-attentional theory of test anxiety.
In: I.G. Sarason (Ed.), Test anxiety: Theory,
research, and application" Hillsdale, N.J.: Lawrence
Earlbaum Associates, 19 80.
Wine, J.D. Evaluation anxiety: A cognitive-attentional
construct. In H.W. Krohne & L. Laux (Eds.), Achieve-
ment, stress, and anxiety. Washington D.C.:
Hemisphere Publishing Corporation, 1982.
Zemore, R. Systematic desensitization as a method of teaching
a general anxiety-reducing skill. Journal of Consulting
and Clinical Psychology, 1975, 43, 157-161.

BIOGRAPHICAL SKETCH
I, Ruth Laurie Lax, was born in Syracuse, New York,
on April 2, 1952. I grew up in northern New Jersey, and
attended public schools there, with the exception of one
year in Oxford, England (1960-61). I graduated from
Summit High School in Summit, New Jersey, and then attended
the University of Rochester, in Rochester, New York. It
was there that my interest in psychology, and in particular,
clinical psychology was fostered. I received a Bachelor
of Arts degree with honors in psychology from the University
of Rochester in May, 1974, and was inducted into Phi Beta
Kappa. I obtained my master's degree from the University
of Florida in 1977, in the area of cognitive style research
with children. I have continued my interest in applied
cognitive psychology with this dissertation in the area of
test anxiety. I completed my clinical internship at
Milwaukee County Mental Health Complex in Milwaukee,
Wisconsin, with a specialization in child-clinical psychology.
114

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.
Nathan W. Perry, Jr., Chairman
Professor of Clinical Psychology
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.
Hugh C. .Davis, Jr.
Professor of Clinical Psychology
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.
J; tí\y via, y.V
Suzanne Bennett Johnson
Associate Professor of Clinical
Psychology
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 disseration for the degree of Doctor of Philosophy.
.
Walter R. Cunningham
Associate Professor of
Psychology

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.
Sandra B. Damico
Professor of Foundations of
Education
This dissertation was submitted to the Graduate Faculty of
the College of Health Related Professions and to the
Graduate Council, and was accepted as partial fulfillment
of the requirements for the degree of Doctor of Philosophy.
Dean, College of Health Related
Professions
Dean for Graduate Studies and
Research
August 1982

UNIVERSITY OF FLORIDA
3 1262 08554 1174



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 EKPLVQ1GB_0WUYOU INGEST_TIME 2017-07-12T21:11:15Z PACKAGE AA00002183_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES