Citation
Modality related performances in educationally handicapped children

Material Information

Title:
Modality related performances in educationally handicapped children : an operant analysis
Creator:
Koorland, Mark Allan, 1947- ( Dissertant )
Wolking, William D. ( Thesis advisor )
Forgnone, Charles ( Reviewer )
Mercer, Cecil D. ( Reviewer )
Pennypacker, Henry S. ( Reviewer )
Place of Publication:
Gainesville, Fla.
Publisher:
University of Florida
Publication Date:
Copyright Date:
1977
Language:
English
Physical Description:
xi, 135 leaves : ; 28 cm.

Subjects

Subjects / Keywords:
Aural learning ( jstor )
Child psychology ( jstor )
Educational research ( jstor )
Experimentation ( jstor )
Learning ( jstor )
Learning disabilities ( jstor )
Learning modalities ( jstor )
Short term memory ( jstor )
Special education ( jstor )
Visual learning ( jstor )
Children with mental disabilities -- Education ( lcsh )
Dissertations, Academic -- Special Education -- UF
Learning, Psychology of ( lcsh )
Reinforcement (Psychology) ( lcsh )
Special Education thesis Ph. D
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Abstract:
There has been a great deal of interest in modality preference as a learner characteristic and its relation to learner performance. To the extent that it is difficult to determine consistent auditory or visual modality preferences, some investigators, nevertheless, claim that aptitude treatment interactions between modality preference and presentation mode indicate preferences which are stable across tasks, and that these preferences control the subjects' performances on a short term memory task. If a positive reinforcer proves to control performances allegedly dependent on modality preferences, then there will be a greater understanding of the variables that contribute to the consistency of modality preference, and subsequently an explanation of the reasons for a subject's performance could be offered at a functional level, rather than at a nominal level. For these reasons the study was designed to use a set of tasks previously found to be successful in demonstrating an aptitude treatment interrelation between modality preference and presentation mode. The study employed operant techniques to further clarify the variables that might control alleged modality dependent performances with educationally handicapped children. Subjects were classified as to modality preference by their performance on a bisensory digit span task. This task consisted of simultaneous presentation of four-digit spans in five criterion trials. The number of visual responses were subtracted from the number of auditory responses and the magnitude and direction of the differences were used to identify the subjects' preference. Subjects with a preference then performed on a bisensory missing units short term memory task. This task consisted of simultaneous auditory and visual presentations of two different sets of three words. Following the bisensory presentation the subject was shown two of the three words from each set and asked to recall the missing word. To meet the criteria for an aptitude treatment interrelation, the subject had to recall, in any ten trials, six missing words from the same presentation mode as their preference classification. After screening 18 elementary aged children, one male with an auditory preference and treatment interrelation and one female with a visual preference and treatment interrelation were identified for the study. A single subject design employing a multielement baseline was used to evaluate the effects of contingencies of reinforcement on the subjects' bisensory missing units task performance, allegedly under the control of modality preference. Data were evaluated by visual inspection of the subjects' cumulative response records and distributions of the subjects' response latencies. Contingencies of reinforcement were found to control both subjects' missing units task performances. Performances in nonpreferred modalities could be made to approach or exceed performances in preferred modalities by the systematic use of contingencies of positive reinforcement. These results imply that the source of control for modality related performances and subsequently the source of stability or consistency of a subject's alleged preference controlled performances may depend on controllable variables in the immediate environment, such as reinforcement. One implication of these findings is that a learner may perform better in a supposedly nonpreferred learning modality than expected, if the consequences for that performance are reinforcing, and if the consequences are arranged systematically.
Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 73-78.
General Note:
Typescript.
General Note:
Vita.

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:
022249508 ( ALEPH )
04290082 ( OCLC )
AAG9926 ( NOTIS )

Downloads

This item has the following downloads:


Full Text












MODALITY RELATED PERFORMANCES IN EDUCATIONALLY
HANDICAPPED CHILDREN: AN OPERANT ANALYSIS











By

MARK A. KOORLAND


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































To the late Henry Koorland,

a very gentle man












ACKNOWLEDGMENTS


Undoubtedly, one of the most difficult tasks is to write

acknowledgments because the completion of a dissertation is just

one step in a series of accomplishments that is contributed to

by many people. Every person's involvement, even if not immedi-

ately apparent, leaves a subtle mark.

My acknowledgments must first extend back to those who

helped to set my career direction. No doubt long forgotten by

them, conversations with Ms. Coramae Paganini and Ms. Esther

Morgan set the starting point. Shortly after a discussion with

Dr. Myron Cunningham nudged me a bit further. Advice and direc-

tion from Dr. Wayne C. Richard generated on my part an interest

in exceptional behavior. Dr. Lyndal Bullock, Ms. Laverne Graves,

Ms. Geraldine Getzen, and Dr. Barbara Reid provided opportunities

to learn and to continue my education.

Throughout this process, support from close personal friends

has often been the mainstay of any accomplishment along the way.

It is difficult to fully describe each contribution but, in my

eyes, each was and continues to be valued. The people whose

names follow have contributed just a small part of their spirit

and purpose to the extent that I am different than before:

Bill and Kathy Steen, Daniel Hobby, Ryan Beaty, Richard Hersh,

Dan Sokol, Robert Alan Tate, Paul and Steve Rozynes, Mary Walsh,










Hugh M. Gramling, Terry Cronis, Marilyn and Gary Warrington,

William C. T. Graham, David Westling, Jim Altman, Elliott

Lessen, Val and Laura Valentine, and Chari Campbell.

A special note of thanks is extended to Grace Hodgson,

Carol Springer, and Leila Cantara for their patience, friendship,

and capable assistance in the completion of this dissertation.

Terry Rose has contributed to the present work and to my personal

growth during the course of our 13 year friendship. The number

of years reflects the opportunities I have had for his contribu-

tions. His tremendous academic and personal capability has been

invaluable in the completion of this dissertation. Mitchell

Martin has been a long time invaluable friend. He is a top notch

professional whose standards have greatly influenced me. Without

his cooperation and effort this dissertation could not have been

completed.

My supervisory committee is composed of individuals who I

regard as true teachers in the greatest sense of the word. Dr. Bill

Wolking, my chairman, has always been enthusiastic about my work

and supportive of me to the extent that I have felt no other direc-

tion to move in except forward. Every question Bill has asked has

resulted in an answer that made this dissertation more complete.

Dr. Cecil flercer has contributed consistently to my professional

growth for a number of years. His input has never been less than

essential, especially for this dissertation. He has sharpened my

ability to think critically. Dr. Charles Forgnone has permitted










me to gain a multidimensional view of not only special education

but also of special education's interface with the culture. I

consider him as a valued friend and advisor. From our earliest

association, Dr. Bill Reid has consistently been supportive. He

has always been positive about my work. I especially value Dr.

H. S. Pennypacker's friendship and his readiness to answer all my

questions, assist me in my career, and in general arrange the

conditions necessary for my academic growth. He has, over the

years, taught me no less than the fundamentals of human behavior

and the skills necessary to ask questions of nature.

Finally, I wish to acknowledge my parents. They have been

totally supportive throughout every phase of my education. It

made no difference if times were good or bad. They never waivered

in their faith in me. For my mother and father, who may only see

this work indirectly, I can only say a simple thank you.












TABLE OF CONTENTS


ACKNOWLEDGEMENTS . . . . . . .

LIST OF TABLES . . . . . .

ABSTRACT . . . . . . . . . . .

CHAPTER I

INTRODUCTION . . . . .

Statement of the Problem ......
Questions Under Investigation ......
Rationale . . . . . .
Variables Affecting Performances ....
Determination of Stable Modality Preference .
Definition of Terms . . . ..
Delimitations of the Study ......
Sumnary . . . . . .

CHAPTER II

REVIEW OF RELATED LITERATURE . . . . . .

Modality Preference . . . . . . .
Studies of Significant Interaction . . . . .
Studies of Nonsignificant Interaction . . . .
Presentation Mode . . . . . . . . .


Studies of Handicapped . .
Studies of Non-Handicapped .
Comparison of Handicapped and
Operant Technology . . .

CHAPTER III


Non-Handicapped


METHOD . . . . .

Subjects ..
Materials and Aoparatus ......
Apparatus and Stimuli .....
Reinforcing Stimuli .....
Settings . . . . .
Procedure . . .. .
Preliminary Procedure .....
Experimental Procedure .....


iii

viii

ix


12

12
13
17
22
22
23
25
27


j I I











Experimental Design . . . . . .... . 40
Dependent Variables . . . . . . 41
Data Collection .... ..... . . . . 41
Data Analysis . . . . . 42
Interobserver Reliability . . . . . .... 43

CHAPTER IV

RESULTS . . . . . . . . . . . 45

Reliability . . . . . . 46
Analysis of Cumulative Responses . . . . . .. 47
Analysis of Baseline 1 . .. . . . . 48
Analysis of Reinforcement Phase . . . . .. 49
Analysis of Incorrect Responding . . . . . 52
Analysis of Baseline 2 . ........ . . 53
Analysis of Response Latencies . . .... .... 55

CHAPTER V

DISCUSSION . . ... .............. . 60

Findings ... . ... ............ . 61
Interpretation of the Findings . . . . . .... 63
Problems and Limitations of the Study . . . ... 66
Practical Implications . . . . . . . 67
Suggestions for Future Research . . . . .... 70

REFERENCES .................. ..... 73

APPENDICES

A. CVC TRIGRAMS . ... ............. . 80

B. DIGIT SPANS . . .. . . . . . . 81

C. MISSING UNITS TASK STIMULUS SHEET . ... . 82

D. DESCRIPTION OF TRIAL PROGRAMMING . . . ... 84

E. RAW DATA ........ . . . ..... 86

F. VERBATIM DIRECTIONS FOR SUBJECTS .. ...... 111

G. RECORD SHEETS . . . . . . .. . 114

H. CUMULATIVE GRAPHS ... . . .... .. .118

I. GRAPHIC DISPLAYS OF RESPONSE LATENCIES . . .. .127

BIOGRAPHICAL SKETCH .. . . . . . . .134












LIST OF TABLES


Table Pag

1 Demographic and Academic Data . . .... 33

2 Percentage Interobserver Agreement for Dependent
Variables Across Experimental Conditions . .. 47

3 Percent of Responses Across Trials by Response
Type and Experimental Phase . . . ... 50

4 Latency Means by Response Type and Experimental
Condition .. .... . . . . . 56













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



MODALITY RELATED PERFORMANCES IN EDUCATIONALLY
HANDICAPPED CHILDREN: AN OPERANT ANALYSIS

By

Mark A. Koorland

December, 1977

Chiarman: William D. Wolking
Major Department: Special Education

There has been a great deal of interest in modality preference

as a learner characteristic and its relation to learner performance.

To the extent that it is difficult to determine consistent auditory

or visual modality preferences, some investigators, nevertheless,

claim that aptitude treatment interactions between modality

preference and presentation mode indicate preferences which are

stable across tasks, and that these preferences control the

subjects' performances on a short term memory task.

If a positive reinforcer proves to control performances

allegedly dependent on modality preferences, then there will be

a greater understanding of the variables that contribute to the

consistency of modality preference, and subsequently an explana-

tion of the reasons for a subject's performance could be offered

at a functional level, rather than at a nominal level. For these

reasons the study was designed to use a set of tasks previously











found to be successful in demonstrating an aptitude treatment inter-

relation between modality preference and presentation mode. The

study employed operant techniques to further clarify the variables

that might control alleged modality dependent performances with

educationally handicapped children.

Subjects were classified as to modality preference by their

performance on a bisensory digit span task. This task consisted

of simultaneous presentation of four-digit spans in five criterion

trials. The number of visual responses were subtracted from the

number of auditory responses and the magnitude and direction of

the differences were used to identify the subjects' preference.

Subjects with a preference then performed on a bisensory

missing units short term memory task. This task consisted of

simultaneous auditory and visual presentations of two different

sets of three words. Following the bisensory presentation the

subject was shown two of the three words from each set and asked

to recall the missing word. To meet the criteria for an aptitude

treatment interrelation, the subject had to recall, in any ten

trials, six missing words from the same presentation mode as their

preference classification. After screening 18 elementary aged

children, one male with an auditory preference and treatment inter-

relation and one female with a visual preference and treatment

interrelation were identified for the study.

A single subject design employing a multielement baseline

was used to evaluate the effects of contingencies of reinforcement

on the subjects' bisensory missing units task performance, allegedly

under the control of modality preference. Data were evaluated by










visual inspection of the subjects' cumulative response records

and distributions of the subjects' response latencies.

Contingencies of reinforcement were found to control both

subjects' missing units task performances. Performances in non-

preferred modalities could be made to approach or exceed per-

formances in preferred modalities by the systematic use of

contingencies of positive reinforcement.

These results imply that the source of control for modality

related performances and subsequently the source of stability or

consistency of a subject's alleged preference controlled per-

formances may depend on controllable variables in the immediate

environment, such as reinforcement. One implication of these

findings is that a learner may perform better in a supposedly non-

preferred learning modality than expected, if the consequences

for that performance are reinforcing, and if the consequences are

arranged systematically.













CHAPTER I

INTRODUCTION



Operant conditioning and its associated technologies have

demonstrated, in various investigations, control over measures

of abilities or traits that have been considered relatively

stable or immutable characteristics of the person, e.g., IQ

scores (Clingman & Fowler, 1976), and sensory thresholds

(Goldiamond, 1962). Operant technology provides a viable set

of methods and procedures for demonstrating the dynamic nature

of learning and associated performances (Edinger, 1969). More

specifically, modality preference, a type of individual charac-

teristic (DiVesta, 1975; Ingersoll & DiVesta, 1972), may be

under the control of contingencies of reinforcement, rather than

under the control of a pervasive trait, aptitude or innate charac-

teristic. Therefore, the purpose of this research was to investi-

gate the stability or consistency across tasks of modality

preferences or aptitudes by using operant procedures to demonstrate

the function of dynamic environmental variables in the control of

performances alleged to be dependent on these preferences or

aptitudes.

Modality preferences have been considered a type of aptitude

that would greatly enhance the study of presentation modalities

and their effects on learning and recall (Ingersoll & DiVesta,

1972; Sabatino & Dorfman, 1974). Lilly and Kelleher (1973) used





2




the terms modality strength, modality preference, and learning

style interchangeably in their investigation of modality preference

and learning. In all subsequent sections, because of the frequent

interchangeable use of the above terms in the modality literature,

aptitudes, modality strengths, and preferred learning modality are

used interchangeably when referring to modality preference.

Special educators have been proposing that learning disabled

children have discrepancies in visual and auditory processes that

result in poor achievement (Waugh, 1973). Attempts to determine

accurately the visual or auditory modality preferences of handi-

capped children have not often met with success (Mann, Proger, &

Cross, 1973; Sabatino, Ysseldyke, & Woolston, 1972). Consequently,

matching these children to appropriate instructional programs also

remains difficult. Some investigators have suggested that in-

adequate instrumentation and measurement procedures may contribute

to the difficulty in identifying a learner's modality preference

(Estes & Stewart, 1975; Foster, Reese, Schmidt, & Ohrtman, 1976;

Mann, Proger, & Cross, 1973; Sabatino, Ysseldyke, & Woolston,

1972; Wolpert, 1971). Because the primary means for identifying

modality preference involves the use of norm measures such as the

ITPA (Kirk, McCarthy, & Kirk, 1961) and other psychometric modality

tasks of many forms, e.g., Wepman, Frostig, Ladd, Mills, and

Valett tests (Meeham, 1974), other types of tasks, may warrant

investigation. Smith (1971) suggests the use of techniques which

are more specifically auditory or visual in nature, i.e., eliminat-

ing motor responses or instructions using other modalities.










Inqersoll and DiVesta (1972) used a laboratory type task to

identify visual and auditory attenders among a group of college age

students. The task was a bisensory digit span type, consisting of

simultaneous presentation of different auditory and visual four-

digit spans. After presentation of the stimuli, the subject

was to recall as many digits as possible. The modality in which

the subject recalled the most digits correctly was used to classify

each subject's preference as either auditory or visual. This task

was followed by a type of short term memory exercise consisting of

a bisensory missing units task that required the subject to respond

with missing words from two independent sets of words. Each

independent set was presented simultaneously in a separate modality.

The results indicated an aptitude (modality preference) by treatment

(presentation mode) interaction implying that the identified

modality preferences were stable across these tasks (Ingersoll &

DiVesta, 1972). This series of tasks was replicated with a number

of educationally handicapped children and indicated what appeared

to be stable modality preferences for these learners, while avoiding

difficulties found in the use of psychometric modality tasks.



Statement of the Problem


This study was designed to determine if modality related per-

formances in elementary aged educationally handicapped children

were subject to control by contingencies of reinforcement rather

than control by a pervasive trait, aptitude, or innate characteristic.










Consequently, existence of a functional relationship between

contingencies of reinforcement and the type of response,

allegedly controlled by modality preference, on bisensory short

term memory trials was investigated.



Questions Under Investigation


This study was designed to answer the following questions

regarding the influence of contingencies of reinforcement on

performances allegedly controlled by modality preference in

elementary school aqed educationally handicapped children.

1. Is there a functional relationship between contin-

gencies of reinforcement and short term memory performance

allegedly under control of a visual modality preference?

2. Is there a functional relationship between contin-

gencies of reinforcement and short term memory performance

allegedly under control of an auditory modality preference?



Rationale


At the core of special education is the search for strategies

leading to effective individualization. The determination of a

student's preferred learning modality in order to insure that

appropriate treatments are provided, is one way in which indivi-

dualization is carried out. Because the modality concept has been

promoted as well as used as an explanation for achievement or

lack thereof (Wepman, 1968), it has resulted in a proliferation











of assessment devices and even modality specific commercial

teaching materials. Unfortunately the emphasis on the modality

concept has resulted not only in commercialism, but also in wide

acceptance of the general strategy of determining modality prefer-

ence and subsequent selection of materials by sensory modality.

Many involved in the modality enterprise have overlooked the

complexity of the learning processes involved (Landers, 1976)

and the inconclusive nature of modality research. The rationale

for this study is derived from two substantive areas related to

modality research. They consist of the variables affecting per-

formances allegedly under the control of modality preference,

and the detenrination of stable modal preferences.


Variables Affecting Performances

When determining the variables that affect modality dependent

performances, most often demographic characteristics such as age,

IQ, socioeconomic status (Blanton, 1971; Williams, Williams, &

Blumberg, 1973), developmental processes (Wepman, 1971), and sex

(Dwyer, 1971; May & Hutt, 1974) are investigated.

Infrequently, but nevertheless important, motivational charac-

teristics of the immediate learning environment have been

suggested as factors that may affect the student's performance

in learning modality studies (Dunn & Dunn, 1974; Williams, Williams,

& Blumberg, 1973). In contrast, Wepman (1971), in reference to

modality preference, states, "the choice appears not to be condi-

tioned by the environment but seems more likely to be an innate

characteristic" (p. 6). In order to clarify the relationship










of variables in the immediate environment to modality dependent

performance, the use of operant technology would seem warranted.

Operant conditioning procedures have been shown to modify

behaviors thought to be controlled by pervasive organismic

variables, relatively resistant to intervention. These variables

are, for example, autonomic responses that control human gastric

acid secretion (Whitehead, Renault, & Goldiamond, 1975), neuro-

logical impairments that control written letter reversals (Smith

& Lovitt, 1973) and learning disabilities and general mental

abilities that control aptitude and intelligence test scores

(Clingman & Fowler, 1976; Kubany & Sloggett, 1971). These studies

demonstrate the influence of environmental variables and subsequent

changes in behaviors thought to represent pervasive characteristics.

There has been no systematic investigation, using operant technology,

of performances allegedly dependent on modality preferences.


Determination of Stable Modality Preference

There appears to be wide agreement that the detection of

modality preference is impeded by the lack of reliable and valid

instrumentation (Ysseldyke, 1973). In order to move toward tasks

that effectively divide attention between visual and auditory

channels and thus indicate the primary modality of attending, the

bisensory simultaneous stimulation paradigm has been employed

(Dornbush, 1968). Ingersoll and DiVesta (1972) mention that

Broadbent, while employing this type of recall task noted that

subjects had their own consistent preference for giving information










received in the auditory or visual channel first. For this reason,

Ingersoll and DiVesta used the bisensory digit span task to

identify visual and auditory preferences of their subjects.

After identifying the preferences, the subjects were given a

second task which was a bisensory modification of the missing

unit paradigm. This procedure consisted of presenting five

words to the subject, then repeating four of the five words and

asking the subject to write on a record sheet the missing word

or unit. The bisensory modification involves the simultaneous

auditory and visual presentation of different sets of words.

The authors found that those subjects identified as having auditory

preferences recalled more auditory materials and those identified

as having visual preferences recalled more visual materials.

DiVesta (1975) notes that to identify the missing word from the

originally presented sets is an impossible task unless one assumes

that people develop consistent, selective, and automatic information

processing strategies.

Consequently, these laboratory-type tasks might be more

directly responsive to auditory or visual preferences and thus

avoid some of the problems encountered with other methods of

assessing modality preference, e.g., psychometric instruments.

Additionally, Ingersoll and DiVesta (1972) state that replica-

tion with different subjects is necessary. They point out that

the use of elementary school children, a more heterogeneous group

than the original population of college students, would allow

selection of more extreme scores relative to modality preference.










Extreme scores should produce clear response patterns on the bi-

sensory missing units task. If the results can be replicated, then

the authors believe that the generality of the construct of

modality preference can be extended.

Elementary age educationally handicapped children are often

the focus of individualized learning treatments and modality

related materials. Their use as subjects seemed warranted,

because of age and relevance to the modality enterprise. This

set of tasks reliably indicated whether modality preferences

existed in the population under study and also provided results

similar to the aptitude treatment interaction (ATI) between

modality preference (aptitude) and presentation mode (treatment)

obtained by Ingersoll and DiVesta (1972). Results similar to the

original ATI implied that these modality preferences were stable

across tasks. Once results similar to Ingersoll and DiVesta's

(1972) findings were obtained, the effect of contingencies of

reinforcement on task performances allegedly under the control

of modality preference were investigated.



Definition of Terms


1. Modality Preference Preferences are "monitoring and

filtering systems which control the flow of information within

a processing system" (Ingersoll & DiVesta, 1972, p. 391). This

is determined by the magnitude and direction of correct recall

on a bisensory digit span task. After five sets of trials, the

number of visual responses are subtracted from the number of










auditory responses. The previously set criteria of 3

(Ingersoll & DiVesta, 1972) was used to classify subjects.

Those who did not meet the established criteria were dismissed

from the study.

2. Bisensory Digit-Span Task Simultaneous presentation

of different auditory and visual four-digit spans. The subject

is asked to recall the set in whatever order he prefers.

3. Bisensory Missing Units Task Simultaneous presentation

of different three word sets. One set is presented auditorily

and the other visually. The words are drawn from A Teacher's

Word Book of Twenty Thousand Words (Thorndike, 1931). The subject

responds with the missing word, when two words from each set are

presented visually following the original presentation.

4. Aptitude Treatment Interaction (ATI) An ATI is a signi-

ficant interaction between personological variables and alterna-

tive treatments. This study does not use statistical tests of

the subjects' data to operationally define an ATI. Rather, an

ATI is said to exist for those subjects who demonstrate a modality

preference and also remember 60% of the words presented on a

sensory channel corresponding to their indicated modality preference.

5. Educationally Handicapped Child Children who are

classified as non-retarded and who are at least one year below

age appropriate grade placement in one or more academic areas.

The child has no physiologically based sensory handicaps.











Delimitations of the Study


The use of the bisensory digit-span tasks have not been

validated as an indicator of modality preference with any task

other than the missing units task. The generality of any

modality classification derived from this procedure to tasks

of varying content and complexity levels is not yet established.

As such, a modality preference determined with the laboratory

tasks may not be generalized beyond short term memory functions.

The use of educationally handicapped students as defined

earlier may impose restrictions on the generality of the findings.

Any operational definition for educationally handicapped students

is bound by the limits of the instruments used to determine subject

characteristics.



Summary


The determination of modality preferences that are stable

across tasks in exceptional students has met with limited success.

The use of psychometric instrumentation has been cited as a

factor contributing to these results. Laboratory tasks, however,

were used by Ingersoll and DiVesta (1972) to demonstrate stable

modality preferences in adults. Using these similar tasks

evidence of stable modality preferences were sought with educa-

tionally handicapped children and, when found, further investi-

gated by the use of operant technology. Operant techniques





11




were used to investigate the responsiveness of alleged preference

controlled behavior to environmental influence.












CHAPTER II

REVIEW OF RELATED LITERATURE



The following review is divided into three sections. In the

first two sections salient studies related to modality preference

and mode of presentation are reviewed. Within these sections,

studies are further divided by type of experimental population.

In the third section, operant technology and its value for modality

research are discussed.



Modality Preference


A mode or modality is a sensory channel, such as visual,

auditory, tactile or muscular movement, through which sensations

are transmitted and received (Jones, 1972). General interest in

the area began in the latter part of the nineteenth century. The

purpose was to compare listening and reading as input channels for

comprehension and analysis of various materials (Jones, 1972).

Ysseldyke (1973) points out the influence of Cronbach and

Reynolds in urging psychologists and special educators to research

how individual's aptitudes could be joined with instructional

systems or treatments in order to identify and use aptitude-

treatment-interactions.

Bracht (1970) states the goal of aptitude treatment interaction

research is identification of disordinal interactions between person-

ological variables (IQ, specific abilities, personality variables,










and others). In light of the above, an interest developed among

special educators in determining the most efficient channels of

learning. Evaluation instruments and complimentary instructional

programs have been devised, based on the premise that the identifi-

cation of preferred modalities in instruction would lead to more

effective learning (Waugh, 1973). Additionally, it was felt that

weak modalities could be identified and instructional activities

could be designed to remediate those weak modalities or processes.

There remains much emphasis on the determination of modality

strength for purposes of matching specific instructional procedure

to a given learner's aptitude (Dunn & Dunn, 1974; Sabatino & Dorfman,

1974). Despite often contradictory results, some investigators feel

that it is premature to discard research on modality preference

(Foster, Reese, Schmidt, & Ohrtman, 1976).

Most often problems in modality research are related to accurate

assessment of modality preference (Jones, 1971; Jones, 1972; Mann et al.,

1973; Smith, 1971) and the use of treatments that are uniquely auditory

or visual in nature (Foster, Reese, Schmidt, & Ohrtman, 1976; Lilly &

Kelleher, 1973; Smith, 1971). The following review compares and

contrasts successful and unsuccessful modality treatment interaction

studies with handicapped and non-handicapped populations.


Studies of Significant Interaction

The published studies indicating modality preference and treat-

ment interaction are relatively few. Indeed, the number of studies

indicating any aptitude by treatment interaction is small. Bracht










(1970) reviewed 90 aptitude-treatment-interaction studies and only

found five that met the conditions for a disordinal interaction.

It must be kept in mind that interaction can be of two types:

with or without reaching significance. Bracht's requirements were

of the former type and as such fairly rigorous. Researchers after

1970 have attempted to improve on methodological features.

Studies of non-handicapped. An investigation (Ingersoll &

DiVesta, 1972) with a college age population obtained an interaction

of modality preference and presentation mode for a short term memory

task. The value of this work to the present discussion is the nature

of the task. The subjects were engaged in a bisensory digit span task

that required them to recall numbers presented simultaneously in an

auditory and visual presentation. Those identified as auditory or

visual attenders were then given a bisensory missing units task con-

sisting of five words simultaneously presented in the auditory and

visual mode. The subject was then presented four words from each

modality and asked to recall the missing word. Modality preference

and its interaction with the treatment were found. The authors state

that the results indicate the stability of modality preferences

across the two tasks. This procedure meets many of the criteria for

tasks that have been found to produce accurate indications of modality

preference, i.e., stability across tasks or interaction with mode of

presentation, a task involving short term memory, and relative face

validity of the task for a particular modality.

Kalin and McAvoy (1973) demonstrated that college age students pro-

vided valid data when asked to rank their modality preference. The










validity was apparent, because the data were used to provide accurate

matches to instructional methods. Those students choosing visual

presentation performed better in that mode than in the auditory mode

and vice versa. Of course, this self disclosure of preference was

obtained from a group keenly aware of study procedures and

associated difficulties. Consequently, the utility of this

approach, for children, would be limited.

Wepman and Morency (1975) obtained a successful modality-treat-

ment interaction in normal first grade children. They used tests

with high content validity in order to measure the perceptual

parameters of interest. Some of the tasks involved auditory and

visual presentation of discrimination tasks (requires same or

different response) and memory tasks (repeat numbers in order and

words in any order). A standardized reading achievement test was

given after phonic and sight word approaches were used for treat-

ments. An interaction between modality preference and presentation

mode of treatment was obtained. Wepman and Morency (1975) noted

that success was due to the use of auditory memory and sequential

memory tasks that others have failed to use. Also, there were no

motor components involved in the tasks and it is felt that motor

components may have confounded previous attempts at modality

preference identification.

In the most recent and successful study, Foster, Reese, Schmidt,

and Ohrtman (1976) used a small number of non-motoric tasks, related

directly to reading, e.g., phoneme discrimination and word recognition.

These tasks were used for modality classification. The subjects suc-

cessfully were taught unknown words by the use of pictures and by the










use of phoneme-grapheme relationships. Parenthetically, the authors

note that many children were screened in order to form the small

experimental groups for this study. The effort required to identify

such a small sample prompted the authors to point out that teaching

through the preferred modality may be unimportant for the majority

of children. Conversely, research on teaching through a preferred

modality might be all the more important for the minority of learners

classified as educationally handicapped.

Studies of handicapped. To date there is only one study using

a population identified as handicapped that found modality preference

to interact with a treatment. In 1973, Lilly and Kelleher attempted

to correct some problems typically found in modality studies. They

cited difficulty with dependent variable measurement and noted that

tasks with at least face validity in relation to the reading process

were necessary. In order to identify modality preference in subjects,

they used a visual or auditory memory task for words. Points were

assigned for words remembered and preference was determined by sub-

tracting auditory scores from visual scores and using a certain number

of points in one direction as criterion for a particular preference.

The dependent measure involved reading or listening to a passage

and recalling salient facts. Those subjects originally identified

as preferring the auditory modality performed better on the audio

passage and those identified as preferring the visual modality

performed better on material read.

In order to perform this study, however, the students had to

read at least at a 2.5 grade level. Consequently, the determination










of modality preference for younger students would present diffi-

culties. Of value, though, is the demonstration of high face

validity in tasks and more or less direct measures of performance,

i.e., number of facts remembered as a dependent variable.


Studies of Nonsignificant Interaction

This category is the largest of modality preference studies.

The dependent measures are primarily concerned with some aspect of

reading. Jones (1971), in a review, points out that various researchers

in studies employing psychometric tests such as the ITPA (Kirk,

McCarthy, & Kirk, 1961) or combined batteries of tests (deHirsh,

Jansky, & Langford, 1966) have difficulty supporting the theory

that modal preference of an individual should be considered in

teaching him to read. Jones concludes that it is difficult to

isolate the auditory or visual modality, but recommends more experi-

mentation to detennine the relationship of modal preference to

learning.

Studies of non-handicapped. Bruininks (1963) used the instru-

mentation approach typical of many unsuccessful investigations, i.e.,

psychometric instruments of a normative nature. Using a battery of

12 tests, he found that groups of auditory and visual learners did

not profit from reading instructions based on perceptual strengths

or weaknesses. Because the subjects, low socioeconomic status (SES)

second and third graders, had been in school up to three years, the

authors concluded that the students' reading habits from previous

instruction might be more influential than perceptual characteristics

in determining their performance. This is an atypical conclusion in










modality studies in that experience is noted as a possible factor

affecting the results. The author also states that the tests used

may not measure the most significant factors involved in reading

instruction.

Cullinan, Ringler, and Smith (1969) using a relatively concise

test, the New York Modality Test, composed of visual, auditory, and

kinesthetic subtests, identified three groups of first grade urban

learners corresponding to three modal preferences. The learners

received visual, auditory, and kinesthetic word recognition training

emphasizing configuration, specific sounds, and words formed by

using pipe cleaners as letters fixed to a card. Using the Metropolitan

Achievement Test as a dependent variable, they found no differences

between the experimental groups. The authors conclude that the use

of a standardized reading achievement test with inner city subjects,

insufficient teaching time (71 hours), and no coordination between

the regular class activities and the various experimental treatments

were possible problems. This study is one of the earliest to mention

difficulty with using standardized tests as the dependent measure.

Smith (1971) in a study of low SES primary subjects found no

relation between modality strength, classified using ITPA subtests,

and visual treatments (use of the Initial Teaching Alphabet and Words

in Color) or auditory treatments (reading with phonics). The

Metropolitan Achievement Test served as a dependent measure. Smith

concludes that if differences in children, relative to auditory and

visual functioning exist, then the ITPA may be inappropriate for

diagnosis. This finding is in agreement with earlier work (Bateman,










1968; Cripe, 1966). The dependent measure, however, was not mentioned

as possible limitation for Smith's (1971) study.

Newcomer and Goodman (1975) used the ITPA as part of a battery

of 18 tests. They found no interaction with an associative learning

task or a meaningful learning task. Additionally, they found the

visual condition of the treatment superior for the majority of the

4th grade normal subjects in the study. This result is in keeping

with Bruininks (1968) earlier work.

Studies of handicapped. Waugh (1973) using the ITPA and an

additional test attempted to correct for inadequacies of tasks used

for dependent measures in earlier studies. Simplified recall along

with word recognition tasks, similar to tasks used in actual instruc-

tion, were used. Waugh found that there was no relation of prefer-

ence to treatment. In an ad hoc analysis it was noted that the

auditory subjects did better on the auditory and visual recall.

task. Thus, the ITPA might serve to identify auditory perceptual

functioning related to tasks at the recall level of complexity.

For the reading task, however, no statements could be made for

either the auditory or visual groups. Possibly reflecting the

results of all modality preference research up to that point in

time, Waugh (1973) states, "special educators at all levels must

continue to ask an essential question is this really a useful

activity?" (p. 469). Waugh's population was handicapped only to

the extent that they were enrolled in Title I reading. To charge

special educators with such an essential question was probably pre-

mature in view of the heterogeneity to be found in most special










education populations. At best, Waugh's population, as a whole,

would be considered borderline exceptional by most special educators.

Sabatino, Ysseldyke, and Woolston (1972) gave a battery of

perceptual tests to mentally retarded subjects. The treatment

involved unisensory perceptual training matched to the indicated

strength of the subject. Gains were found on perceptual measures

but were not specific to the kinds of training received. Group

reading achievement tests were given as an additional dependent

measure and no significant gains were detected. The authors con-

clude that present modality classification instruments have sizable

error and may classify children in specific modality areas by

chance.

Using 116 mentally retarded children as subjects, Sabatino and

Dorfman (1974) arranged three groups consisting of auditory, visual,

and non-preferred categories. They used a normative battery of

tests; some derived from the ITPA. These groups were matched to

a commercial visual reading curriculum (Sullivan Series) and an

auditory reading program (Distar). Pre and posttest scores, obtained

using an individual achievement test, produced no interaction. The

authors conclude that possible lack of reliability in the instruments

and use of measures that were not factorily simple may have contri-

buted to the nonsignficant findings.

Tyler (1974) using a similar population and ins-trumentation

received similar results. Parenthetically, the author mentions

that all responses of the subjects in the present study were rein-

forced. This is the first statement about environmental contingencies










to be found in any of the modality preference-treatment studies to

date.

In summary there are some statements that apply to most of

these studies in modality preference:

1. Every test of modality preference employing the ITPA or

selected subtests provided non-significant modality-treatment

interactions.

2. The use of large comprehensive batteries to determine

modality preference essentially did not prove to be more effective

than the use of a smaller number of tests.

3. The use of normative or psychometric instrumentation

appeared, in general, to lack the sensitivity to detect modality

preferences or treatment gains.

4. Little or no information is provided about experimenter

application of contingent reinforcement to the performance of

subjects.

Unless there is an interaction between the preferred modality

and the specialized treatment, it is very difficult to determine

where the basic weakness in the study lies. It could be in the

modality identification procedures, purity of the treatment, or the

method of analysis. There is a possibility that, because of the

present state of the art, few reliable generalities can be made

about modality preference.










Presentation Mode


Many studies related to the learning modality area are con-

cerned with presentation mode in particular. A unifying feature

of many of the studies included in this area is the type of depen-

dent measure used and the data analysis model. For the most part,

tasks used for dependent measures are paired associate, short term

memory, or other specific skills (Silverston & Deichmann, 1975).

In every study, group research designs and confidence level statis-

tics are used for analysis.

For these studies, researchers may employ handicapped subjects

or handicapped and non-handicapped comparisons, but the majority

employ elementary level low SES or normal populations. The

studies are reviewed by population and various distinguishing

features are mentioned.


Studies of Handicapped

In studies of learning disabled children, Estes and Huizinga

(1974) and Estes and Stewart (1975), using paired associate tasks

under visual and auditory presentation nodes, noted visual tasks

produced a greater number of correct responses than auditory. It

is suggested (Estes & Stewart, 1975) that instructional materials

for these subjects should be oriented to visual processing, yet in

the same discussion the authors state, "in learning disabilities .

heterogenity of psycholinguistic functioning and learning styles

is the rule rather than the exception" (p. 5). This seems incon-

sistent with the determination that learning disabled children in

their studies performed best under the visual presentation mode.










In a study of retarded adults, McConkey and Green (1975) inves-

tigated presentation method and its effect on free recall of

categorically related items. They used auditory, visual sequen-

tial, and visual simultaneous presentation modes. Although primarily

interested in the use of categorial strategies during recall, they

noted that both groups receiving visual input recalled more items

than the group receiving auditory presentation. This result is

congruent with many of the other studies (Bruininks, 1968; Estes

& Huizinga, 1974; Estes & Stewart, 1975; Newcomer & Goodman, 1975).


Studies of Hon-Handicapped

Galfo (1970) in a study based on Broadbent's limited capacity

processing model, which states that redundant presentation of program

material will result in no increase in learning, obtained results at

variance with the limited capacity theory. He found that simultaneous

sight sound presentation with normal subjects proved superior to

sight and sound separation. The task, however, involved academic

material, and as such might involve level of familiarity factors or

complexity exceeding that of tasks used in other studies in this

area.

A study more typical of the mode of presentation and specific

skills type under review was conducted by Filmer and Linder (1970).

Using low SES subjects, they employed a short term memory task

requiring recall of colors, digits, and pictures. The task was

presented in the auditory, visual, and auditory-visual combined

modes. The combined mode yielded the highest scores on each task;

the auditory yielded the lowest. These results are in agreement










with Galfo's (1970) study. Of special importance, however, is that

the researcher states that immediate feedback on correct and in-

correct responses was given. This is the earliest and one of the

few indications of the use of feedback or reinforcement. It must be

kept in mind that the feedback was not a variable under investigation.

Reynolds, Bickley, Champion, and Deckle (1971) noted that very

few of the studies employing word association tasks take mode of

presentation of stimulus words into consideration. Consequently,

they obtained results with low SES subjects that indicate responses

to word association tasks are a function of stimulus modality. They

also noted that as children get older the difference in type of

response to oral and graphic presentation of words becomes less

pronounced. Silverston and Deichmann (1975) noted age dependent

relationships between modality skills and various reading indices.

It would appear, then, that experience and other environmental

variables may interact with performance on tasks related to various

modalities.

In a study of encoding processes, in second grade children's

short term memory, Corsale (1974) used mode of presentation as an

independent variable. The subjects were required to recall cate-

gorically related words after a brief interference period. Although

mode of presentation for this type of task plays a role in adults,

it was not found to have an effect on the children's encoding.

Although Corsale's procedures differed because of the use of an

interference procedure, this result is interesting in relation to

McConkey and Green's (1973) finding that mode of presentation for

retarded adults differentially affected recall of categorically










related items. It might be of value to determine if there is any

parallel performance for handicapped children as there is for both

handicapped and non-handicapped adults. Mode of presentation

occasionally appears to have generality across very different

populations.

iilliams, Williams, and Blumberg (1973) found that middle class

subjects perform better with visual presentation in a paired associate

task while lower class subjects learned better under an auditory

condition. In relation to social class and performance on paired

associate learning tasks, they state the following:

because of related motivational differences it is
difficult (unfortunately) to find a situation
entirely free of these possible confounding factors.
(p. 358)

It is interesting to note that some researchers of modality

choose either to ignore motivational variables or to treat motiva-

tional variables as a nuisance rather than subject matter worthy of

investigation in its own right.


Comparison of Handicapped and Non-Handicapped

Budoff and Quinlan (1964) compared auditory and visual presenta-

tion modes in normal and retarded readers. They used a meaningful

paired associate learning task for second graders and found that the

auditory presentation was superior in both groups. This finding

does not hold true across other studies (Bruininks & Clark, 1972;

Katz, 1967).

Katz (1967) employed a discrimination task presented in the

auditory and visual mode at different levels of familiarity using










normal and retarded readers. Familiarity was defined as frequency

of occurrence in the culture. The authors employed the Hebrew

alphabet and associated letter names as the unfamiliar stimuli and

the English alphabet and associated letter names as the familiar

stimuli. Katz (1967) found that stimulus modality interacted with

level of familiarity. He concluded that effects of stimulus modality

are differentially related to the response required of the subject.

The best performance was obtained with the visually familiar

stimuli. These results raise the issue of stimulus meaning as

a variable which will influence results. In many of the presenta-

tion mode studies this variable is not taken into consideration.

Bruininks and Clark (1972), using first grade disadvantaged

educable mentally retarded children, disadvantaged non-retarded,

and advantaged non-retarded, studied the efficacy of auditory, visual,

and combined presentation modes. The results failed to show any

difference on paired associate tasks between disadvantaged and non-

disadvantaged of similar intellectual ability. Overall the total

group responded best under the visual condition. The authors con-

cluded that young retarded children exposed to auditory activities

ought to have visual picture opportunities at the same time. To the

extent that this study is typical of the others, it is also unusual

because the authors point out that the subject was not informed of

his correct or incorrect responses. This is opposite of the tactics

that Filmer and Linder (1970) employed. Both studies used audio,

visual and combined presentation modes, both studies employed low

SES subjects and recall tasks, yet different results were obtained.










The differential use of feedback or reinforcement may have contri-

buted to the difference in outcomes.

In summary, from the studies reviewed some conclusions may be

offered.

1. Rarely is information concerning consequences of the

subjects' responding made available.

2. The majority of the studies appear to demonstrate favorable

results under visual presentation modes, but because of the prevalent

group research strategies the performance of individuals are not

reported.

3. Studies employing a non-complex task, i.e., short term memory,

provided more consistent results than studies employing complex tasks,

i.e., study of academic materials.

From these conclusions and conclusions of the Modality Preference

section, the need for alternative research strategies, continued use

of non-complex tasks, and attention to environmental variables becomes

apparent. In the following section on Operant Technology alternative

research strategies and environmental variables are considered.



Operant Technology


The growth and development of operant technology has been

steady and has had impact on diverse areas of research (Kazdin,

1975). More specifically, the techniques of applied behavior

analysis, which embrace many of the traditions of present laboratory

practices, have been directed toward the needs of special educators,

who, by definition, are charged with bringing unique procedures,










materials, and modifications of, or additions to school practice

for the handicapped (Kirk, 1962).

Applied behavior analysis attempts to demonstrate quantitative

characteristics of the behavior, the experimental manipulations,

and the technologically exact description of the methods used to

bring about change in behavior. Additionally, the change must be

effective enough to have value and also possess generality within

limits (Baer, Wolf, & Risley, 1968).

To the extent that science is a rigorous exercise in measurement,

applied behavior analysis (ABA) requires certain standards of

measurement and design. The use of intra-subject designs permit-

ting replication (Sidman, 1960) is considered basic. Additionally,

measurement requirements for the assessment of reliability of

behaviors under observation are an integral part of ABA practice.

To appreciate that behavior of importance in applied settings

is the focus of ABA, one only needs to consider the diversity of

settings in which ABA techniques have taken place: day care

centers, early education, Head Start, elementary schools, secondary

education, college settings, special education, community settings,

and clinics (Ulrich, Stachnik, & Mabry, 1974).

Operant conditioning has been applied not only in a wide range

of settings, but also for the purpose of investigating a wide range

of behaviors in those settings. Repeated demonstration of rein-

forcement as the controlling variable for certain classes of behavior

have established the generality and usefulness of the reinforcement

principle for those classes.











Recently, operant conditioning strategies have been applied

to relatively traditional research areas (Clingman & Fowler, 1976;

Goldiamond, 1962; Lovitt & Smith, 1973). Goldiamond (1962), in a

discussion of perception, points out the importance of questioning

the locus of response change in sensory threshold, hypnotic percep-

tion and related perceptual research. In relation to controlling

variables, Goldiamond states, "where it is possible to choose between

sensory and response loci, the evidence has supported a response

locus" (p. 315).

Just as the application of new procedures to perceptual research

(Goldiamond, 1962) may have evoked comments of doubt by those who had

a vested interest in the area, i.e., psychophysics researchers and

traditional psychologists, so has the initial application of operant

procedures to newer areas.

Psychometric testing has in the recent past received some

attention from individuals working with operant technology. Because

of the standard procedures that must be adhered to (Cronback, 1970)

in order to appropriately administer a psychometric instrument such

as an intelligence test, any variation from that procedure is seen

as a source of invalidity for the results. Commenting on the use of

contingent reinforcement for correct answers on such instruments,

Conner and Weiss (1974) state that all scores will shift upward and

each subject's position in the distribution will remain the same.

The problem with Conner and Weiss's conceptual analysis is that it

was not based on experimental data; it was based, no doubt, on their

measurement training and theoretical background. When the effect of











reinforcement was determined, it was found that the scores in

different ranges were differentially affected by the experimental

manipulation (Clingman & Fowler, 1976). Of more important value

than the effect obtained was the process uncovered and the additional

understanding of the interaction between psychometrics and con-

tingencies of reinforcement.

Examples of the use of operant technology to examine areas

previously thought to be confounded by or impervious to the in-

fluence of immediate environmental variables are increasing.

Whitehead, Renault, and Goldiamond (1975) studied human gastric

acid secretion with operant conditioning procedures. Ayllon,

Layman, and Kandel (1975) investigated the comparative use of

medically prescribed drugs and reinforcement to control hyperactive

behavior, often thought to be under the relatively exclusive control

of physiological variables. Smith and Lovitt (1973) have demon-

strated the remediation of number reversals that were previously

thought to be under the control of neurological impairments or

immature physiology. In the main, operant technology has been

helpful in providing understanding of the extra-organismic processes

that interact with many behaviors thought to be under the primary

control or organismic variables.

Skinner (1974) points out the lack of utility in discussing

the causes of behavior, using instinct related explanations. To

say that the duckling follows the mother because of imprinting

is not a thorough enough explanation. "Imprinting" as the

explanation of "following behavior" eventually becomes accepted










as the cause of the behavior, when, in fact, it is simply explana-

tion at a nominal level. The more subtle effects of the environ-

ment are overlooked in this type of analysis.

In much the same way, to say that a student learns visual

material best because he has a preference for the visual mode is

to explain cause at a nominal level, rather than at a functional

level. The further discovery of controlling variables may be

undermined by the use of current terminology and explanatory

constructs.

In summary, the area of modality preference in children has

not been subject to investigation using operant technology. Con-

sequently, the purpose of this study is to investigate the sensi-

tivity of modality controlled performances to environmental

variables, thus leading to a fuller understanding of their effect

on this phenomenon. Important considerations for this investiga-

tion, emerging from the previous review of literature, are the

inclusion of non-complex tasks and employment of alternative research

strategies.













CHAPTER III

METHOD



In Chapter III the methods and procedures of the study are

presented. This includes a description of the subjects, materials

and apparatus, settings, procedures, and experimental design.

The study was composed of three phases:

1. Identification of eligible children.

2. Identification, using non-complex tasks, of those

eligible children meeting criteria for having a modality preference.

3. Application of reinforcement procedures to alleged modality

dependent performance.



Subjects


The two subjects were educationally handicapped (EH) elementary

students. The subjects were non-retarded and at least one year

behind age-appropriate grade level in reading or in arithmetic skills

as assessed on the following achievement tests:

a. Slosson Oral Reading Test (Slosson, 1963).

b. Wide Range Achievement Test Arithmetic Subtest

Level I (Jastak & Jastak, 1965).

In order to qualify for the study, the subjects had to know:

a. Names of numbers 1, 2, 3, 4, 5, 6, 7, 8, 9, 0.

(Integers 1-0).










b. Consonant vowel consonant (CVC) trigram words.

(See Appendix A).

Each child was shown a list of the words and numbers used

in the study. They were required to read the material with 100%

accuracy. Demographic and academic data are reported on the

children qualifying for the study (see Table 1).



Table 1

Demographic and Academic Data



Father's Appropriate Reading Arithmetic
Subject Age Sex Occupation Grade* Grade Grade


1 8 yrs.ll mos. M Professional 4.0 4.9 3.0

2 9 yrs. 8 mos. F Semi-skilled 5.0 6.2 2.8



*Because academic testing and demographic data were obtained in late
summer, appropriate fall grade placement is given as age-appropriate.



Materials and Apparatus


Apparatus and Stimuli

Bisensory digit span task. Integers (1, 2, 3, 4, 5, 6, 7,

8, 9, 0) were randomly assigned to series of digit spans which were

four digits in size ( see Appendix B). For each series there was a

simultaneous audio and visual presentation. The same number could

not occur twice within a bisensory presentation.










Bisensory missing units task. CVC trigrams were selected

from a list of words (Thorndike, 1931) within the typical subject's

reading capability (see Appendix A). Ambiguous words or homophones

(son and sun), as well as words of related meaning were eliminated

from any bisensory trial. Numbers were also eliminated (ten and

six).

Media. Visual stimuli were projected onto a suitable surface

and the audio stimuli were provided by a cassette tape recorder

(Buhl Programma, Model 616-150) built to operate the projection

equipment (Kodak Carousel Slide Projector, Model E-2). The equip-

ment was programmed so that a sound-relay synchronized the auditory

and visual presentation in order to deliver them simultaneously.

Digits were presented at a rate of one every two seconds on each

channel. Because of limitations of the equipment, the visual dura-

tion of exposure was .9 second. The auditory duration averaged

.6 seconds in length. These presentation times approximate the

specifications provided in Ingersoll and DiVesta's (1972) procedure

description.

Stimulus sheet. After a bisensory missing units trial, a

stimulus sheet (see Appendix C), containing two of the three words

from the auditory and two of the three words from the visual pre-

sentation, was placed in front of the subject. A cover sheet with

a 175 mm. x 12 mm. opening was placed on top of the stimulus sheet.

In this fashion, after each trial, the experimenter would slide the

cover down the stimulus sheet in order to expose the proper set of










stimulus words, one set at a time. A complete description of the

arrangement used in programming the trials is given in Appendix D.


Reinforcing Stimuli

Nickels were chosen as the reinforcer. This particular

currency was chosen because of its generalized value and ease of

handling. Nickels were given to the subjects for emitting the

appropriate response upon completion of certain missing units

trials. As they were earned, nickels were placed in a clear

plastic container in direct view of the subject.



Settings


Subject 1 was enrolled in a private school for learning dis-

abled children. An unused classroom was available for carrying

out the experiment.

The classroom was well lighted and air conditioned. The

subject was seated at a 106 cm. x 106 cm. table. The projection

equipment was immediately to the subject's left and the audio equip-

ment was at the subject's right where it could be controlled by the

experimenter. The experimenter was seated to the right of the

subject.

The experimenter placed the paper stimulus materials necessary

for the missing units task directly in front of the subject. As

the subject proceeded from one trial to the next, the experimenter

controlled the paper stimulus materials with his left hand while

he operated the equipment and recorded responses with his right










hand. The slide portions of the bisensory tasks were projected

on a surface approximately 106 cm. in front of the subject.

Subject 2 was enrolled in a special summer program for mildly

handicapped children. This program was operated at a laboratory

school and a spare classroom was used. The classroom was

similar to the one used with Subject 1, except that it was not

air conditioned. The experimental arrangement was similar to

that of Subject 1.



Procedure


Preliminary Procedure

Bisensory digit span task reliability. Because the bisensory

digit span task had only been used with college freshmen, it was

necessary to determine its reliability with a younger population.

Internal consistency was determined by giving a group of 11 subjects

(x C.A. = 10 yrs. 5 mos.) ten criterion trials and then forming a

ratio of the number of auditory correct responses to the number of

visual correct responses for the first five criterion trials and

then for the second five criterion trials (see Appendix E). A

Pearson product moment correlation resulted in r = .69 for these

two sets of ratios.

Pilot study. In order to develop and refine the procedures

for the actual experiment, a pilot study was performed. Two

important outcomes resulted from the pilot effort.

First, the multielement baseline that was used incorporated

a verbal command as a discriminative stimulus for the subject. In










other words, the subject was told beforehand whether to attend

to the auditory or to the visual stimuli and was reinforced for

a correct response. After using this procedure with the pilot

subject, it became apparent that it was difficult, if not impossible,

to determine whether the instructions or the reinforcement schedule

controlled the subject's responding. Consequently, in the actual

experiment the instructions were eliminated and only the reinforce-

ment schedule was used to influence responding.

The second outcome was the attainment of a recoverable baseline

condition that required a relatively large number of trials before

the experimenter observed stability. This was in contrast to the

smaller number of trials necessary to observe stability in the

original baseline phase. This outcome alerted the experimenter

to the possibility of using a larger number of trials than was

anticipated originally. In fact, many trials were necessary in

order to gain stability in a number of phases during the actual

experiment.


Experimental Procedure

Reinforcement. After the modality classification and missing

units task baseline were obtained, each subject was asked, prior

to the actual contingent use, if he or she would like to earn

nickels. Both subjects responded emphatically that they would.

Bisensory Tasks. In the digit span task, the subject received

ten trials during which he or she was told beforehand from which

modality to recall first. These responses were not recorded; they

served to familiarize the subject with recall by modality set.










Five additional trials were given and the subject was told

to recall whichever set first that he or she wanted. Performance

on these five trials served as the modality preference criteria.

The subject dictated the numbers recalled and the experimenter

recorded the responses. The number of visual correct responses

were subtracted from the number of auditory correct responses

and a 3 value served to classify the child as having either a

visual or auditory preference.

This session was followed by the missing units trials. The

different words (CVC trigrams) were presented simultaneously in

the auditory and visual modes. The subject received trials until

he achieved six correct auditory or six correct visual responses

in any group of ten trials. If the subject achieved the six

correct responses in the mode corresponding to his classification

on the digit span task, then he was said to demonstrate a signifi-

cant aptitude treatment relationship. The next series of trials

was used for reinforcement and the final series of trials comprised

the return to baseline condition. The subject saw both sets of words,

less one, on a stimulus sheet after each trial (see Appendix C).

In order to avoid problems with motor variables, the subject was

required to respond orally with the missing word. The experimenter

recorded the responses.

A detailed summary of the experimental procedure is as follows:

1. The student was screened in order to determine if he or

she met the criteria for word and number recognition and for educa-

tionally handicapped classification.

2. Subjects meeting the requirements for the experimental










population were given ten bisensory digit span trials. The

subject was told from which modality to recall. Verbatim direc-

tions are provided in Appendix F.

3. The subject was given the five criterion trials (see

Appendix F). If he met the criterion for having a modality

preference, he proceeded to the next task.

4. The subject engaged in bisensory missing units trials

(see Appendix F). If he did not reach criterion on this task he

was dismissed. If criterion was reached, the subject proceeded

to the next series of trials. In these trials the experimenter

announced that reinforcement would be given for correct recall of

the missing units. The criteria for reinforcement was shifted

from the non-preferred to the preferred modality until the experi-

menter could demonstrate control over responding in both. The

subject was not told when the criterion shifted, thereby permitting

an assessment of the contingencies alone as an independent variable.

5. After completion of the reinforcement trials, the subject

received the third and final series of missing units trials. For

this series the experimenter announced that reinforcement would no

longer be given for correct recall of missing units.

In summary, seven children, in addition to the 11 children

necessary for the digit span reliability procedure, were screened.

The pilot subject came from the group of 11 reliability subjects.

Three other subjects, who were qualified for the study, were found

in the additional group of seven. One subject was dismissed

because of a procedural error related to verbal commands given by










the experimenter. The two subjects who remained were employed

in the present study. Therefore, 18 children were screened, four

were found to be qualified, and two were used for the actual

study.



Experimental Design


The design is a single subject reversal type (Baer, Wolf, &

Risley, 1968). The first phase was the baseline of bisensory

missing units trials. The second phase was the series of missing

units trials on a multielement baseline (Sidman, 1960) that con-

sisted of two concurrent discriminated operants, each on a contin-

uous reinforcement (CRF) and extinction multiple schedule. The

experimenter systematically applied the reinforcement contingency

to responses requiring either auditory or visual attending. This

resulted in the visually discriminated operant coming under CRF,

while the auditorily discriminated operant came under extinction

and vice versa. If the subject was classified as having an auditory

preference then the first condition in the second phase consisted

of reinforcement for visual responses. If the subject was classified

as having a visual preference then the first condition in the second

phase consisted of reinforcement for auditory responses. The

conditions in the second phase were alternated until the return

to baseline phase.

The multielement baseline controls for intertrial difficulty

and practice effects by distributing the requirement for responses

to any particular modality across a large number of trials. The










multielement baseline also provides the opportunity for repeated

demonstrations of experimental control.


Dependent Variables

Two classes of dependent variable were necessary for this

study. The first class was the type and number of responses,

i.e., number of correct words recalled from each channel on the

missing units task and number of incorrect responses. Because

an incorrect response could not be assigned to either a visual

or auditory mode, incorrect responses were not used to determine

accuracy of responding. They served, however, as an indicator of

difficulty.

A second class of variable, which was necessary for a fine-

grain analysis of the experimental conditions was response latency

(Alba, 1975; Ferster & Skinner, 1957). These latencies were

recorded for visual, auditory, and incorrect responses. Response

latency was defined as the time from offset of the missing units

task presentation to the onset of the subject's response.


Data Collection

The data were collected during experimental periods approxi-

mately 30 to 45 minutes in length. The subjects' responses were

recorded on the experimenter's record sheet for the digit span

tasks and the missing units tasks (see Appendix G). Additionally,

each session was recorded on audio tape in order to permit

measurement of response latencies and checks on reliability.










The latencies, while somewhat difficult to record during

the actual experimental sessions, were obtained from the tapes

not only by the reliability observer, but also by the experimenter.

The latencies were obtained in this fashion, because the time

required to time and record the latencies would have slowed down

the experimental sessions and limited the number of trials that

could be given during any particular session. Additionally, the

overt timing may have produced reactivity on the part of the

subjects.

The experimenter and reliability observer listened to the

tapes and obtained response latencies by starting and stopping

a 1/10 second stopwatch at the offset of the bisensory presentation

equipment and at the onset of a response. The point of offset was

indicated on the tape by the distinct noise occurring when the

experimenter operated controls that stopped the equipment.

Listeners stopped timing when the subject made an audible response.

Latencies were recorded on the experimenter's record sheet.


Data Analysis

The number of auditory, visual, and incorrect responses was

plotted using, for each subject, a cumulative graphic display.

The cumulative records were subjected to visual inspection as the

primary analytic method. The trends in the data from phase to

phase and from condition to condition within a phase were used to

evaluate the reinforcement effect. The percent of visual, auditory,

and incorrect responses in any particular phase or condition was










computed in order to yield a simple quantification of the subjects'

response. This information is used to supplement the data in the

cumulative records.

The latency data were used to further compare and contrast

differences between correct and incorrect responses across the

various phases of the experiment. These data are displayed as

distributions on a logarithmic scale. The logarithmic scale

aids in interpretation because it assists in normalizing the

display of the data.


Interobserver Reliability

The reliability procedure chosen for this study was the

relatively stringent point-by-point agreement method (Kelly, 1977).

It is possible to use the point-by-point method with trial and

latency data by recording latency values and codes for various

responses within the experimental conditions.

In the present study any trial could be coded as having

produced a visual (V), auditory (A), incorrect response (I),

experimenter error (EE), or equipment malfunction (EM) (see

Appendix E). In order to permit proper coding, the reliability

observer was given a copy of the correct auditory and visual

answers for any set of trials to be evaluated. The observer

listened to audio tapes of the various experimental conditions

and recorded response codes accordingly.

After the response codes of the reliability observer were

recorded, they could be compared to the corresponding experimenter






44



records. The point-by-point agreement was computed by dividing

the number of specific notations of the observers' records

agreed on by the total number they both recorded plus the number

of notations only one or the other recorded (Kelly, 1977).

The latency data were checked for reliability in the same

manner as the trial notation data. The point-by-point computa-

tional procedure was the same. Rather than computing the number

of notations agreed upon, the number of latency values agreed

upon becomes the unit for reliability evaluation. When determin-

ing agreement between two observed latency values,a .5 second

observer error was permitted.












CHAPTER IV

RESULTS



The measures used for analysis of the experimental results

are the cumulative number of responses and response latencies.

The three types of responses were auditory (A), visual (V), and

incorrect or error (I). For each type of response, the durations

of the response latencies were measured.

In the following sections, both the number of responses and

response latencies are analyzed across phases, beginning with the

baseline phase (Bl) and proceeding through subsequent phases in

which reinforcement was contingent on auditory or visual responses.

The subscript indicates the particular order and position a base-

line or reinforcement phase held within the experiment. For

example, phase V2 indicates that reinforcement was provided for

visual responses and that this was the second such reinforcement

phase for the subject. The missing units task raw data, indicat-

ing trial numbers, phase changes, and day changes are found in

Appendix E.

The data were collected according to the following schedule:










Trials, Phases, and Day Sequence Per Child

Subject 1 Subject 2

Trial Number Phase Day Number Trial Number Phase Day Number

1-39 B1 1, 2 1-68 B1 1, 2

40-151 V1 2, 3 69-106 A1 3, 4

152-174 Al 3, 4 107-226 V1 4, 5, 6

175-201 V2 4 227-274 A2 6

202-213 A2 4 275-295 V2 6, 7

214-238 V3 4, 5 296-330 A3 7

239-264 B2 5 331-498 B2 8, 9, 10



Reliability


Table 2 provides percentage agreements between the experimenter

and the reliability observer for each experimental condition.

Agreement information is shown for each dependent variable, i.e.,

type of response and response latency. The minimum acceptable

total agreement percentage for type of or latency of response was

80%.

For Subject 1, certain conditions, notably BI, V1, A1, and

V3, produced low agreement percentages for either response type or

response latency. These low percent values are due, in part, to

background noise on the tapes. These noises were present because

of air conditioning sounds and occasional interruptions by

custodial staff. The raw data taken from the trials that were

rated for reliability are provided in Appendix E.










Table 2

Percentage Interobserver Agreement for Dependent
Variables Across Experimental Conditions


Phase

Response Type

Response Latency


Phase

Response Type

Response Latency

*Not available


BI

72.7

81.8


BI

92.5

92.5


V1

76.4

82.3


Al

100.0

76.4


Subject 1

A1 V2

100.0 91.6

78.2 83.3


A2 V3 B2 Total

NA* 83.3 87.5 87.3

NA* 75.0 81.2 80.5


Subject 2

V1 A2 V2 A3 B2 Total

93.7 96.2 100.0 100.0 100.0 98.3

81.2 92.3 95.4 89.4 100.0 90.0


Analysis of Cumulative Responses


The responses are displayed by the use of cumulative records

(Ferster & Skinner, 1957; Skinner, 1968). The cumulative record

was chosen as the display method, because it offers a potent

technique for establishing functional relationships between environ-

mental variables and behavior (Alba, 1975). Consequently, for each

experimental condition, the type of response is graphed cumulatively

by trials. The actual calendar day is noted on the upper part of

the auditory grid. Broken vertical lines indicate the trial pre-

ceding a day change and the solid vertical lines indicate the trial

preceding a phase change. The cumulative graphs are found in

Appendix H.


~~










During the course of the experiment, there were occasional

experimenter errors and equipment malfunctions. In order to

clearly represent every trial outcome, these errors are noted

by a break in the cumulative record line at the point that the

error or malfunction occurred.

It should be noted at the outset that the steepest slopes

possible for the cumulative displays of this experiment are approxi-

mately 45 degrees. A 45 degree slope would represent the occurrence

of one response per trial for any given response type. Because of

the design of the experimental trials, only one response of any

particular type is possible. Slopes of less than 45 degrees indi-

cate a lower number of responses across trials and horizontal

cumulative record lines indicate no responding across trials.

Additionally, after 20 cumulative responses occur, the cumulative

record line resets and starts again on the zero line. The con-

tinuous record from trial to trial clearly illustrates steady

state responding, transition states, and any local fluctuations

in the general trend of the data.


Analysis of Baseline 1

An analysis of the performance of Subjects 1 and 2 during

the B1 phase (see Appendix H-1 and H-4) yields a major difference

in type of response. Subject 1 strongly responds with auditory

responses, while Subject 2 responds with visual responses. These

responses reflect their original modality preference classification.

There appears to be a steeper cumulative response curve for Subject

l's auditory responding than for Subject 2's visual responding.










While responses to non-preferred presentation modalities also

differ as to total number between subjects, the differences in

incorrect responding for each subject can likewise be appreciated.

Subject 2 makes incorrect responses approximately two times as often

as Subject 1. This high error rate probably contributed to Subject

2's extremely low number of responses to the auditory presentation,

her non-preferred presentation modality. Furthermore, because of

the high number of errors, almost twice as many trials were necessary

in order to gain a relatively stable baseline before beginning the

reinforcement phases.


Analysis of Reinforcement Phase

The first reinforcement phase (V1) for Subject 1 consisted of

continuous reinforcement for any response that was under the

stimulus control of the visual portion of the bisensory presenta-

tion. This was Subject l's non-preferred modality. The first

reinforcement phase for Subject 2 (A1) was arranged in a similar

fashion, i.e., reinforcement for the non-preferred modality.

Subject 1, possibly because of a high rate of auditory responding

during baseline (B1), persisted in high auditory responding for

approximately 50 trials before the contingency began to exert

control over the visual responses. In contrast, the reinforcement

contingency in Phase A1 appeared to influence Subject 2's responding

after approximately 14 trials.

In order to further clarify the subjects' response patterns,

percentage measures of each subject's responses are provided in

Table 3. The number of responses in a phase for each of the three











Table 3

Percent of Responses Across Trials
by Response Type and Experimental Phase


Subject 1


B1 V1 A1 V2 A2 V3 B2

61.0 40.5 68.1 11.5 66.6 16.6 64.0

17.9 35.1 18.2 61.5 16.6 54.2 20.0

35.8 29.7 22.7 30.7 16.6 33.3 20.0


Subject 2


B1 Al V1 A2 V2 A3 B2

07.0 45.9 43.5 57.1 33.3 58.8 47.8

29.5 08.0 24.2 33.3 61.9 23.5 26.6

60.6 37.8 27.1 19.0 09.5 26.4 23.6


*Subjects occasionally would give both a correct auditory and
visual response, and subjects would periodically give no scorable
response. Because of these responses, sums of auditory, visual,
and incorrect responses, if computed, give values somewhat above
or below 100% depending on the particular phase.


Phases


Auditory*


Visual*


Incorrect*


Phases


Auditory*

Visual*

Incorrect*










response types was divided by the number of trials in that phase.

This measure yields an index of response proportion.

When inspecting incorrect response percentages, it becomes

clear that Subject 2's A1 responses took place at 37.8%, about

half of the B1 amount of 60.6%. This reduction of errors is

related to the relatively large, sixfold, increase in auditory

responses. The increases in non-preferred responding from the B1

to V1 phase for Subject 1 and from the B1 to A1 phase for Subject

2 argue convincingly for the existence of a functional relationship

between the reinforcement contingency and bisensory missing units

task performance. It must be kept in mind that the auditory V1

value for Subject 1 is relatively high because of the large number

of trials that elapsed before the contingency had an effect on

visual responding. From trial 100 to 150, however, the rapid

changeover can be appreciated easily.

Returning to the cumulative displays, it is interesting to

note, during the V1 phase, Subject 2 persisted for a large number

of trials in auditory responding. Around trial 228 the visual

responding began to increase. For Subject 2 the large number of

trials necessary to gain control of the visual responding and to

note a reduction in auditory responding contributed to the high

V1 auditory value found in Table 3.

Weiner (1970) discusses the existence of behavioral persistence

in human subjects. He has attributed this behavior to either

organisms with limited response repertoires or to uncontrolled

variables affecting performances. Because Subject 2 had both










auditory and visual responses in her repertoire, the possibility

exists that during the V1 phase an uncontrolled variable was

operating. In either event, responding did come under control

dependably from trial 228 through trial 252 and, in those 24

trials, 75% of the responses were visual.

Except for a slow changeover in condition V1, experimental

control was demonstrated consistently for Subject 1 in phases V2,

A2, and V3. The trend in responding was quickly brought under

control by the reinforcement contingency, and consequently fewer

trials were needed to obtain stability in these phases.

Subject 2 performed in a similar manner but with some varia-

tion. Upon close inspection of phase A2 and V2 it can be seen that

the effects from a previous phase carried over as much as 14 trials

in the following phase, as illustrated by continued visual responding

from phase VI to A2. For Subject 1, there were virtually no

occurrences of behavioral persistence in the latter phases (Al, V2,

A2, and V3) for either auditory or visual responses.


Analysis of Incorrect Responding

Incorrect responding served as an indicator of difficulty of

the material in any particular phase. As previously mentioned,

the number of incorrect responses cannot be computed with the

number of correct responses to provide measures of accuracy, because

of the difficulty in reliably determining whether any particular

incorrect response was visual or auditory in nature. An interesting

pattern, however, appears from phases A2 through B2 in Subject 2,

and from phase V1 through B2 in Subject 1. For the phases noted,










when reinforcement was applied to performances in non-preferred

learning modalities, the incorrect responses increased (see Table

3). When reinforcement was applied to performances in preferred

learning modalities, then incorrect responses decreased. It must

be kept in mind that this pattern is related to difficulty but not

accuracy of responding. Furthermore, there were no contingencies

placed on incorrect responses in any phase of the experiment.


Analysis of Baseline 2

The final return to baseline (B2) phase for each subject yielded

contrasting results. Subject 2 initially gave a brief run of visual

responses from approximately trial 65 to trial 80, while auditory

responding decreased during the same period. After this set of

trials, however, Subject 2 engaged in continued auditory responding.

As noted in Table 3, error percentages during B2 responding were

not especially high. In fact, errors occurred two thirds less

often than during the B1 phase.

Subject 1 quickly regained the original response curves

observed during the B1 phase. Briefly returning to the pilot

study, the pilot subject, classified like Subject 1 as having an

auditory preference, demonstrated a similar recovery of baseline

responding. During the pilot study, however, the recovery of the

auditory preference occurred after 50 trials of baseline. In

other words, the data would have appeared to recover the original

B1 visual stability; when in actuality, a transition or local

fluctuation was in effect (see Appendix H-7, H-8).










Regarding the B2 phase, it is necessary to keep in mind that

the multielement baseline design, by permitting the repeated

demonstration of experimental control, usually diminishes the

need for returning to a lengthy baseline condition. Because

only one more affirmation of the consequent (Sidman, 1960),

showing that the independent variable was responsible for the

change in responding, is gained, the additional expenditure of

time may not always be justifiable.

When, after numerous experimental manipulations, irrevers-

ability is encountered, Sidman (1960) points out:

If non recoverability is indeed a fact, intrasubject
replication is not possible. I have brought up the
consideration, however, that the irreversability may
be elsewhere than in the organism or in his behavior.
.The extinction operation, as it is normally
carried out, only destroys some of the relations
that were established during original acquisition.
(pp. 101-102)

Because of numerous experimental manipulations of Subject 2's

baseline performance, the possibility exists that the irrevers-

ability encountered is a function of environmental variables and

not organismic ones. If this is true, then the baseline condition

could, no doubt, be recovered had further experimental manipulation

been attempted.

In the main, the three replications (V1, V2, V3) with Subject

1, and the three replications (Al, A2, A3) with Subject 2 argue

convincingly for the existence of a functional relationship between

the contingencies of reinforcement and both subjects' performances

on the bisensory missing units task. With few exceptions, experi-

mental control was demonstrated across both subjects in a relatively

rapid and predictable fashion.










Analysis of Response Latencies


Response latencies were graphed as distributions on semi-

logarithmic scales. Latencies are displayed for each type of

response (A, V, and I) by phase. These displays are provided in

Appendix I. Means of the latency distributions are provided in

Table 4.

Response latencies did not prove to be greatly responsive to

the experimental manipulations. Consequently, few reliable

generalizations can be made from the latency data. There are,

however, a number of trends that became apparent upon close inspec-

tion of the data.

For most conditions, the duration of latencies associated

with incorrect or error responses was uniformly high across both

subjects. The only example of an auditory or visual latency

higher than incorrect latencies occurred once during Subject 2's

visual B1 performance (see Appendix 1-4). With that exception,

Subject 2's incorrect latencies were consistently higher than any

latencies obtained for auditory or visual responses. High latencies

associated with incorrect responses have also been noted by Alba

(1976) in work involving the use of time parameters for investiga-

tion of academic responding.

In the majority of measures a relatively high auditory or

visual response latency occurred during reinforcement conditions.

Overall means for baselines and reinforcement conditions are noted

in Table 4. The visual overall mean for Subject 1 provides the

only overall measure wherein the baseline mean is greater than the










Table 4

Latency Means by Response Type and Experimental Condition


Overall Mean
Subject 1 Bj V1 Al V2 A2 V3 82 B1,B2 V1,A1,V2,A2,V3

Auditory 1.68 3.00 2.12 4.90 NA* .00 2.59 2.12 3.34

Visual 2.12 1.60 1.15 4.00 NA* 2.87 2.78 2.45 2.40

Incorrect 3.00 3.90 2.98 8.50 NA* 6.80 2.59 3.25 5.54



Overall Mean
Subject 2 81 Al V1 A2 V2 A3 B2 B ,B2 A1,V1,A2,V2,A3

Auditory 1.50 2.39 3.52 4.40 4.83 3.17 3.32 2.41 3.66

Visual 1.59 .80 3.75 3.10 3.70 2.25 2.58 2.08 2.72

Incorrect .99 1.89 5.20 5.18 9.10 5.17 3.49 2.24 5.30



*Not available










reinforcement mean. It should be noted, however, that these two

measures are very similar in value. Relative to means for indivi-

dual conditions, those means representing latency measures during

reinforcement conditions exceeded baseline measures in 18 out of

26 conditions (70%). The shift in latency duration also carried

over into the B2 condition for both subjects. The carry over was

more noticeable for Subject 2 than for Subject 1. For Subject 2,

the clarity of the greater latencies in the B2 phase may also be

a reflection of larger amounts of data gathered in this condition.

Regarding individual latency values, the exception to higher

latencies during reinforcement was again the one outlying latency

in the visual Bi condition for Subject 2 (see Appendix 1-4).

Relative to individual incorrect latency values, Subject 2's

B2 latency durations exceeded those of the B1 condition (see

Appendix 1-4 and 1-7). The longer individual latencies found under

reinforcement were not found for incorrect responses in Subject 1

(see Appendix I-1 and 1-3). The data might be better understood

when one considers a possible source of short latencies in general.

It was noted during the baseline trials that Subject 2 engaged

in frequent guessing, to the extent that this particular subject

might give a response before the stimulus sheet cover was moved to

expose the appropriate stimulus words for the trial. As the experi-

ment proceeded this behavior decreased, no doubt, because of its

limited reinforcement. The final product, however, of guessing

behavior was incorrect responses and short latencies. As Subject

2 was exposed to various reinforcement conditions, guessing

stopped and the duration of latencies increased.










In attempting to understand greater latencies, when careful

inspection of the latency data during reinforcement is made, it

is possible to note a trend for latency durations to cluster

around greater time values. The clustering is noticeable for

the type of response being reinforced. In other words, with two

exceptions, visual latencies were at least as long if not longer

than auditory latencies under visual reinforcement conditions.

The V3 phase for Subject 1 resulted in no auditory responses

thereby eliminating this phase from the above comparisons. The

trend to longer latencies under various reinforcement conditions

might be indicative of the more careful attempt on the part of

the subject to respond correctly.

The experimenter would occasionally hear both subjects employ

a rehearsal strategy. This strategy consisted of audible stimulus

item repetition taking place while the subject looked at the stimulus

sheet. The process of stimulus item repetition, while determining

the missing unit, probably required somewhat more time than other

non audible strategies. In either case, it became apparent that

both subjects at some points in the experiment would use audible

verbal rehearsal strategies requiring more time between the offset

of the bisensory presentation and the onset of their responses.

An additional subtle relationship was noted in Subject l's

data. At one point, the experimenter ran out of reinforcers for

Subject 1. The problem occurred during the V2 phase. At this

time the experimenter substituted (as the medium for reinforcement)

check marks on a card rather than nickels. The card was to be

traded in for cash during the following session. Although it is










difficult to make an experimentally verifiable statement, this

phase produced, as a whole, somewhat longer latency durations

for all three categories of responding (see Appendix 1-2).

Further replications, however, would be needed to clarify any

reliable functional relation between latency and change in the

reinforcing stimulus.

In sum, the most convincing data that a functional relation-

ship exists between the subject's performance and the reinforcement

contingency are found by visual inspection of the cumulative response

data. The latency data, while serving to clarify subtle character-

istics of the responses, are secondary to the large effects and

numerous direct replications (Sidman, 1960) of experimental control

over the subjects short term memory performance on the bisensory

missing units tasks.















CHAPTER V

DISCUSSION



There has been a great deal of interest in modality preference

as a learner characteristic and its relation to learner performance.

To the extent that it is difficult to determine consistent auditory

or visual modality preferences, some investigators (Ingersoll &

DiVesta, 1972), nevertheless, claim that aptitude treatment inter-

actions between modality preference and presentation mode indicate

preferences which are stable across tasks. Additionally, they

claim that these preferences control the subject's performances

on a short term memory task.

The present study employed operant techniques to further

clarify the variables that act to control alleged modality depen-

dent performances with mildly educationally handicapped children.

A positive reinforcer proved to control the modality related

performances of these children. This demonstration provided an

illustration of the environmental variables contributing to the

consistency of modality preference, and subsequently provides an

explanation for a subject's performance at a functional level,

rather than at a nominal level.











Findings


The findings of this study are discussed relative to the

experimental questions. The questions are listed and a brief

discussion follows.

1. Is there a functional relationship between contingencies

of reinforcement and short term memory performance, allegedly

under control of a visual modality preference?

The data drawn from Subject 2, who was classified as having a

visual preference,indicate that the reinforcement effect, while

not always similar in pattern from phase to phase, was present.

Visual inspection of the cumulative response curves demonstrates,

especially at phases Al, A2, V2, and A3, consistent control over

the type of responding.

Data from Table 3 indicate that contingencies, when placed on

auditory responses, increased percent response values over those

obtained during both the visual baselines. Except for phase V2,

auditory values, obtained when reinforcement for visual responding

was in effect, exceeded those values obtained for visual responding.

In other words, the application of reinforcement increased perform-

ance in the non-preferred modality to the extent that it exceeded

performances in the preferred modality with one exception. In-

creased performance is visible for auditory responses in phase

Al and A3, when compared to visual responses in phases Bl, B2, and

VI.

Visual latencies for Subject 2, who exhibited a visual

preference, were not greatly sensitive to reinforcement when











applied to specific classes of response. In other words, when

auditory responses were reinforced there were no easily dis-

cernible shifts in the auditory latency distributions. Thus,

for the present study these measures are equivocal in there sensi-

tivity to the independent variable.

2. Is there a functional relationship between contingencies

of reinforcement and short term memory performance allegedly under

control of an auditory modality preference?

Visual inspection of the trends in the cumulative records

indicates relatively reliable control over the direction of the

response curves.

The data gathered on Subject 1 indicated that he exhibited

an auditory preference. When reinforcement contingencies were

placed on visual responding, the percent value not only doubled

its B1 value (see Table 3), but also, in two out of three visual

reinforcement phases (V2 and V3), exceeded the B2 auditory condi-

tion. Further, the visual responding under visual reinforcement

exceeded or approached, within 6.8%, the B1 auditory responding,

in phases V2 and V3.

As with Subject 2, Subject 1 when under reinforcement for

the non-preferred class of response, did not demonstrate changes

of a noticeable magnitude on the latency measures for the non-

preferred response. Therefore, the usefulness of these measures

for addressing the second experimental question is limited.

Supplementary information, not directly related to the experi-

mental questions, bears some discussion. First, although the










latency measures did not prove especially sensitive to the specific

experimental manipulations per se, the latency data did seem

responsive to reinforcement conditions in general. Second,

latency measures did provide data useful for a topographical

analysis of the responses. As mentioned earlier, subjects'

responses that were characterized best as guesses exhibited ex-

tremely short latency durations. Longer response latencies, for

correct responding, appeared to take place on occasions character-

ized by a subject's use of a vocal or almost subvocal rehearsal

strategy. Incorrect responses were characterized by latencies

frequently approaching or exceeding a ten second value. For

correct responses, time values exceeding ten seconds only occurred

twice out of 13 possible phases in which the latency datum was

kept. Third, that the incorrect responding was differentially

sensitive to experimental phases is suggested by the data in Table

3. Because incorrect responding cannot be clearly identified with

any response type, it is of little use in discussing the subjects'

accuracy. It does, however, provide an indicator of differential

difficulty for the subjects.



Interpretation of the Findings


That a functional relationship between the type of responses

on the missing units task and contingencies of reinforcement exists

is clear. Interpretation of the data at a molecular level was

provided in Chapter IV. At a molar level, interpretations of the

experimental findings, relative to Ingersoll and DiVesta's (1972)









suppositions of modality preference control and stability, are

necessary.

Ingersoll and DiVesta (1972) observed their subjects' per-

formances on a bisensory task and then observed similar perform-

ances on another set of bisensory tasks. From this set of

observations they proceeded to suggest a causal relationship between

preference and performance. Realistically, the relationship

observed could be categorized as correlational. To say one set

of behaviors indicates control over another is to make a statement

needing a strong experimental foundation.

First and foremost, the repeated demonstrations of a relatively

weak reinforcer as an independent variable capable of controlling

which of two discriminated operants a subject would use most fre-

quently, certainly places the question of control in a functional

perspective, rather than a correlational one. The numerous experi-

mental manipulations using an independent variable under control

of the experimenter would appear to be a more viable method to

approach the questions as to what, in fact, might control a subject's

bisensory short term memory performance.

Bearing on modality preference stability, the inference made

by Ingersoll and DiVesta (1972) that the observed aptitude treatment

interaction in their study implied stability or consistency across

tasks is certainly reasonable. Stability from one task to another,

however, implies stability through time. The point is of moment,

because Ingersoll and DiVesta performed their tasks in sequence.

The modality preference identification task was performed first

and the missing units task second. The data in the present experi-










ment give support for relative flexibility of response direction

through time. Additionally, Subject 2 did not recover the original

baseline responding she demonstrated earlier on each modality, but

showed a pattern of alternating between both.

Whether or not responding could be recovered is not as important

as the fact that it did change and remained so after experimental

manipulations ceased. This is even more interesting in light of

Subject 2's performance on the bisensory digit span task. Subject

2 scored the greatest value (-12) in the visual direction of any child

screened for the study. If the visual modality preference was truly

stable or consistent across these tasks, then it seems likely that

the baseline condition would have been recovered, at least partially.

It must be kept in mind that Ingersoll and DiVesta (1972) concluded

that extreme digit span scores would be associated with clear

response patterns on the bisensory missing units task. The validity

of this proposition may be restricted when one considers the data

of the present experiment.

Smaller incorrect response magnitudes did appear in phases

under reinforcement for responses similar to the original modality

preference classification; nonetheless, there is little reason to

believe that experimental control of incorrect responding in any

phase could not be achieved given the appropriate experimental

controls.

In sum, regarding control of alleged modality dependent per-

formances and their stability across tasks, it appears that the

source of control for a subject's performance is more likely to










be found in specifiable, measurable, and for the most part,

controllable variables such as positive reinforcement, in the

immediate learning environment. Resorting to explanations based

on data essentially correlational in nature has very limited

utility in uncovering the variables responsible for behavior.

Stability across tasks is very likely to be a function of these

same variables acting in concert with time parameters.



Problems and Limitations of the Stud


The foremost limitation of any study performed in the area of

modality preference is the instrumentation and criteria used for

modality preference identification. The second pressing problem

is the provision of a treatment that actually possesses the compon-

ents for classification as exclusively auditory or visual in nature.

Relatively non-complex tasks were chosen to minimize the threats

of both foregoing limitations. The use of non-complex tasks may

have limited generality to the domain of tasks that are primarily

the focus of modality based treatment, i.e., reading. Short term

memory, on the other hand, has been suggested as a skill necessary

for reading (Senf, 1969). Consequently, the tasks used for the

modality treatment in this study may be components of more complex

behavior.

A problem that deserves some discussion, because it is notice-

able with Subject 2, concerns equipment malfunctions. The precision

of the equipment was limited. Changes in available line voltage

and even changes in temperature would periodically effect the










performance of the audio-cassette sound relay that controlled the

visual portion of the bisensory presentation. The effect of the

occasional interruptions is not known. Fortunately, the greatest

proportion of trials took place without interruption. This problem

could be avoided in future studies by the use of more sophisticated

programming equipment.

The second problem was the behavioral persistence encountered

with Subject 2. The possibility exists that an uncontrolled

variable (Weiner, 1970), such as subtle cues by the experimenter,

accidental reinforcement, or a practice effect, contributed to

the observed persistence. Additionally, in any study there is a

chance that the subject reaches a state of reinforcer satiation or

that the reinforcer itself may have been identified incorrectly.

In any operant study of reinforcement the only control for these

possibilities is inspection of the data. Control was achieved

eventually in all experimental conditions, but the possibility

that a loss of reinforcer effectiveness or, on the other hand, a

competing reinforcer during certain trials remains. In either

event, it is worthy to note that the persistence took place in the

subject's auditory response, her non-preferred modality. This fact

subsequently limits the value of an auditory preference as a

plausible alternative explanation.



Practical Implications


A number of outcomes in this study have practical implica-

tions for those involved in research or applied work. First,










methodologically, the value of a dynamic or continuous measure-

ment strategy, an integral component of operant technology,

becomes evident. The continuous measurement through time employed

in this study permitted the investigator to not only note changes

in the data that were necessary for making experimental decisions,

but also note changes in responding that occasionally might be at

variance with the subject's original preferred modality classifica-

tion. The continuous measurement process could be crucial in

programming the most effective modality treatment, had the study

been concerned with making an educational intervention based on

the subject's modality use.

Second, the direct measurement strategy employed in operant

research deserves comment. The data in the present experiment did

not undergo a statistical averaging in preparation for tests of

significance nor did the primary measurement units evolve from any

statistical summarization of the data. For the most part the data

was interpreted in units that were compatible with the continuous

measurement strategy. In other words, direct measurement procedures

are interpreted in a display that reflects each trial outcome and

the impact of time as a parameter. For example, had B1 cumulative

data been averaged with B2 cumulative data in Subject 2, conclusions

about her modality performance across tasks would have been entirely

different from those originally presented.

Third, for the practitioner of modality based teaching inter-

ventions, the data of this study tentatively indicate that modality

use in various presentation modes is not undivided. That by using










the proper environmental arrangements performances in an alleged

non-preferred modality can approach or in some cases exceed the

performance in a preferred modality. This is not to say that

performance in non-preferred modes can consistently be equated

with performances in the preferred, but the possibility for

approximate performances exists.

The error patterns in both subjects showed differential

sensitivity to the experimental conditions. In order to increase

performance in a non-preferred modality there may be some temporary

trade off in terms of error rates, but not necessarily in terms of

accuracy. In the response latency measures there were no examples

of differential error sensitivity.

There is convincing evidence that to exclude a learner from

material presented in a non-preferred mode would be a great

injustice and possibly even an inefficient teaching tactic. It

may, in fact, be a service to the learner to be switched to a non-

preferred modality. Subject 2 used her non-preferred modality

frequently in the B2 condition after exposure to the experimental

conditions. Professionals have assumed that the preferred modality

is in the service of the learner while at the same time attributing

the preference to sensory and other constitutional traits.

It is entirely possible that an individual's observed modality

use is a function of his or her past reinforcement history. In

this sense a modality preference could be superstitious in origin.

In other words, the modality use frequency that one observes in an

individual could have come about by accidental reinforcement of










learning in that particular modality. It must be kept in mind

that reinforcer control of modality use was demonstrated repeatedly

in the present experiment. And, if one considers the term

preference in a functional perspective it actually connotes a

high probability or frequency of an activity on the part of the

individual who exhibits the preference. Thus, modality use

frequency is the defining characteristic of a modality preference

and reinforcement influences the frequency of modality use. In

sum, a learner may perform better in a supposedly non-preferred

learning modality than expected, if the consequences for that

performance are reinforcing and if the consequences are arranged

systematically.



Suggestions for Future Research


The foregoing research, in many ways, stimulates an unlimited

variety of questions related to modality preference research. Much

research is needed before one can fully understand the relationship

of reinforcement to modality related performances.

The first requirement for any science is that of replication,

for it is the soundest empirical test of reliability (Sidman, 1960).

Operant procedures lend themselves well to replication. Because

of the reliance on experimental control, the great number of

observations on relatively few subjects, the strategy of holding

many variables constant, while manipulating only one, and because

of the technologically exact descriptions of the procedure, replica-

tion is possible.










Systematic replication (Sidman, 1960) of the present experi-

ment should give a fuller understanding of modality use and

related performances. First, replications with children presenting

various behavioral repertoires, such as different baseline

modality use and error response patterns, is of the utmost

importance for understanding the generality of the present results.

There is a need to determine, with other children, not only the

generality of the functional relations noted in this study, but

also the general usefulness of the methods employed.

Second, the complexity level of the task may interact with the

subjects' modality use. As noted with Subject 2, incorrect respond-

ing took place at a high rate in phase B1 when compared to Subject

I's responding. Whether or not Subject 2's incorrect responding

during baseline contributed to her behavior in the experimental

condition is certainly a valid question that can be addressed by

investigating various levels of complexity in task arrangements

or in stimulus materials.

Third, it is assumed in modality investigations that subjects

naturally provide responses associated with that presentation

modality that is the most effective for them. In the present

study contingencies of reinforcement proved to influence the type

of response that a subject gave. It may be of value to investigate

variables such as various verbal directions to the subject. It is

possible that directions can effectively control performance to

the extent that the subjects' optimum performance in a modality

can be obtained by instructing them to attend to that modality.










In the same manner, instructing a subject in various rehearsal

techniques may also optimize their learning in an infrequently

used modality while also increasing the chances for increased

use of that modality.

Fourth, the schedule of reinforcement in the present study

was continuous. The area of schedule effects on modality use

should be systematically investigated. Not only the schedule

effects of reinforcing stimuli, but also the schedule effects of

aversive stimuli could prove valuable as a research area.

Schedule research with both types of stimuli could illuminate

behavioral processes contributing to particular modality use

patterns.

Fifth, in the present study accuracy of responding was

difficult to evaluate. Further studies of a similar nature should

concentrate on the accuracy of a subject's response under varying

reinforcement conditions. In order to carry out such research,

techniques need to be developed to determine if an incorrect

response is associated with auditory or with visual stimuli.

By performing replications of the present experiment and by

pursuing the suggested research, the knowledge base associated

with modality preference can expand and ultimately become

consumable by the population in greatest need of viable teaching

strategies handicapped children.












REFERENCES


Alba, E. The experimental analysis of academic behavior. In
L. E. Fraley & E. A. Vargas (Eds.), Proceedinqs of the
Third National Conference on Behavior Research and Technolog
in Higher Education. Champaign, Ill.: Research Press, 1975.

Allyon, T., Layman, D., & Kandel, H. J. A behavioral educational
alternative to drug control of hyperactive children. Journal
of Applied Behavior Analysis, 1975, 8, 137-146.

Baer, D. M., iolf, M. M., & Risley, T. R. Some current dimensions
of applied behavior analysis. Journal of Applied Behavior
Analysis, 1968, 1, 91-97.

Bateman, B. The efficacy of an auditory and visual method of
first grade reading instruction with auditory and visual
learners. In H. K. Smith (Ed.), Perception and Reading.
Newark, Del.: International Reading Association, 1968.

Blanton, B. Modalities and reading. Reading Teacher, 1971, 25,
210-212.

Bracht, G. H. Experimental factors related to aptitude-treatment-
interactions. Review of Educational Research, 1970, 40,
627-645.

Bruininks, R. H. The relationship of auditory and visual per-
ceptual strengths to methods of teaching word recognition
among disadvantaged Negro bys. Nashville, Tenn.: George
Peabody College for Teachers, 1968. (ERIC Document
Reproduction Service No. ED 043 721)

Bruininks, R. H., & Clark, C. B. Auditory and visual paired
associate learning in first grade retarded and non-retarded
children. American Journal of Mental Deficiency, 1972, 76,
561-567.

Budoff, 11., & Quinlan, D. Reading progress as related to efficiency
of visual and aural learning in the primary grades. Journal
of Educational Psychology, 1964, 55, 247-252.









Clingman, J., & Fowler, R. L. The effects of primary reward on
the I..1 performance of grade-school children as a function
of initial I.Q. level. Journal of Applied Behavior Analysis,
1976, 9, 19-23.

Conner, J. J., & Weiss, F. L. A brief discussion of the efficacy
of raising standardized test scores by contingent reinforcement.
Journal of Applied Behavior Analysis, 1974, 7, 351-352.

Corsale, K. The effects of mode of presentation on encoding
processes in children's short term memory New Orleans, La.:
American Psychological Association, 1974. (ERIC Document
Reproduction Service No. ED 114 763)

Cripe, A. G. Auditory and visual learning related to ITPA sensory
channels (Doctoral dissertation, Purdue University, 1966).
Dissertation Abstracts International, 1966, 27, 635B. (Univer-
sity Microfilms No. 66-7407)

Cronbach, L. J. Essentials of psychological testing. New York:
Harper and Row, 1970.

Cullinan, B. E., Ringler, T., & Smith, J. Preferred learning
modalities and differentiated presentation on reading tasks.
New York: [Jew York University, 1969. (ERIC Document Repro-
duction Service No. ED 042 589)

deHirsh, K., Jansky, J. J., & Langford, W. S. Predictingreading
failure. New York: Harper, 1966.

DiVesta, F. J. Trait-treatment interactions, cognitive processes,
and research on communication media. Audio Visual Communica-
tions Review, 1975, 23, 185-196.

Dornbush, R. L. Input variables in bisensory memory. Perception
and Psychophysics, 1968, 4, 41-44.

Dunn, R., & Dunn, K. Learning style as a criterion for placement
in alternative programs. Phi Delta ppan, 1974, 56, 275-278.

Dwyer, F. 1I. Visual learning: An analysis by sex and grade level.
California Journal of Educational Research, 1971, 22, 170-176.

Estes, R. E., & Huizinga, R. J. A comparison of visual and auditory
presentations of a paired associate learning task with learning
disabled children. Journal of Learning Disabilities, 1974,
10, 33-42.

Estes, R. E., & Stewart, J. C. A comparison of visual and auditory
channels in learning disabled and control children. Washington,
D.C.: American Educational Research Association, 1975.
(ERIC Document Reproduction Service No. ED 116 420)











Edinger, D. L. A free operant analysis of programmed instruction
performance with reading disabled children (Doctoral disserta-
tion, University of Florida, 1969). Dissertation Abstracts
International, 1970, 31(5), 2215-A. University Microfilms
No. 70-20, 749)

Ferster, C. B., & Skinner, B. F. Schedules of reinforcement. New
York: Appleton-Century-Crofts, 1957.

Filmer, H. T., & Linder, R. Comparison of auditory and visual
modalities. Education, 1970, 91, 110-113.

Foster, G. G., Reese, J. H., Schmidt, C. R., & Ohrtman, W. F.
Modality preference and the learning of sight words. Journal
of Special Education, 1976, 10, 253-258.

Galfo, A. J. Effects of certain audio and visual presentation
sequences on pupil information acquisition. Journal of
Educational Research, 1970, 69, 172-176.

Goldiamond, I. Perception. In A. J. Bachrach (Ed.), Experimental
foundations of clinical psychology. New York: Basic Books,
1962.

Ingersoll, G. M., & DiVesta, F. J. Effects of modality preferences
on performance on a bi-sensory missing units task. Journal of
Experimental Psychology, 1972, 93, 386-391.

Jastak, J. F., & Jastak, S. R. The wide range achievement test.
Wilmington, Del.: Guidance Associates of Delaware, 1965.

Jones, J. P. Learning modalities should they be considered?
Atlanta: International Reading Association, 1972. (ERIC
Document Reproduction Service No. ED 052 899)

Jones, J. P. Intersensory transfer, perceptual shifting, modal
preference and reading information series: Where do we go?
Newark, Del.: International Reading Association, 1972.
(ERIC Document Reproduction Service No. ED 052 899)

Kalin, M., & McAvoy, R. The influence of choice on the acquisition
and retention of learning materials in different modes of
instruction. New Orleans, La.: American Educational Research
Association, 1973. (ERIC Document Reproduction Service No.
ED 094 381)

Katz, P. A. Visual discrimination performance of disadvantaged
children: Stimuli and subject variables. Child Development,
1967, 38, 233-242.

Kazdin, A. E. The impact of applied behavior analysis on diverse
areas of research. Journal of Applied Behavior Analysis, 1975,
8, 213-229.










Kelly, M. B. A review of the observational data collection and
reliability procedures reported in The Journal of Applied
Behavior Analysis. Journal of Applied Behavior Analysis,
1977, 10, 97-101.

Kirk, S. A. Educating exceptional children. Boston: Houghton
Mifflin, 1962.

Kirk, S. A., McCarthy, J. J., & Kirk, iW. D. Illinois test of
psycholinguistic abilities (experimental ed.J. Urbana:
University of Illinois Press, 1961.

Kubany, E. S., & Sloggett, B. B. The role of motivation in test
performance and remediation. Journal of Learning Disabilities,
1971, 4, 426-428.

Landers, M. F. The modality concept goes commercial. DCLD
Newsletter, 1976, 2, 53-56.

Lilly, S. M., & Kelleher, J. Modality strengths and aptitude-
treatment interaction. Journal of Special Education, 1973,
7, 5-13.

Mann, L. E., Proger, B. B., & Cross, L. H. Aptitude-treatment
interaction with handicapped children: A focus on the
measurement of the aptitude component. New Orleans, La.:
American Educational Research Association, 1973. (ERIC
Document Reproduction Service No. ED 075 510)

May, R. B., & Hutt, C. Modality and sex differences in recall and
recognition memory. Child Development, 1974, 45, 228-231.

McConkey, R., & Green, J. M. Presentation method and the free
recall performance of retarded adults. American Journal of
Mental Deficiency, 1973, 78, 95-97.

Meeham, T. An informal modality inventory. Elementary English,
1974, 51, 901-904.

Newcomer, P. L., & Goodman, L. Effects of modality of instruction
on the learning of meaningful and non-meaningful material by
auditory and visual learners. Journal of Special Education,
1975, 9, 261-268.

Reynolds, R. J., Bickley, A. C., Champion, S., & Deckle, 0.
Effects of mode of presentation of stimulus materials in
word association tasks. Psychological Reports, 1971, 28,
211-215.

Sabatino, D. A., & Dorfman, N. Matching learning aptitude to two
commercial reading programs. Exceptional Children, 1974, 41,
85-90.










Sabatino, D. A., Ysseldyke, J. E., & Woolston, J. Diagnostic-
prescriptive teaching utilizing perceptual strengths of
mentally retarded children. American Journal of Mental
Deficiency, 1972, 78, 7-14.

Senf, G. Development of immediate memory for bisensory stimuli
in normal children and in children with learning disorders.
Developmental Psychology onograph, 1969, 1, 6 (pt. 2).

Sidman, M. Tactics of scientific research. New York: Basic
Books, 1960.

Silverston, R. A., & Deichmann, J. W. Sense modality research and
the acquisition of reading skills. Review of Educational
Research, 1975, 45, 149-171.

Skinner, B. F. A case history in scientific method. In A. C.
Catania (Ed.), Contemporary research in operant behavior.
Glenview, CA.: Scott, Foresman, and Company, 1968.

Skinner, B. F. About behaviorism. New York: Vintage Books, 1974.

Slosson, R. L. Slosson oral reading test. New York: Slosson
Educational Publications, 1963.

Smith, C. M. The relationship of reading method and reading
achievement to ITPA sensory modalities. Journal of Special
Education, 1971, 5, 143-149.

Smith, D. D., & Lovitt, T. C. The educational diagnosis and remedia-
tion of b and d written reversal problems: A case study.
Journal of Learning Disabilities, 1973, 6, 356-363.

Thorndike, E. L. A teacher's word book of twenty thousand words.
New York: Columbia Teachers College, 1931.

Tyler, J. L. Modality preference and reading task performance among
the mildly retarded. Training School Bulletin, 1974, 70, 208-
214.

Ulrich, R., Stachnik, T., & Mabry, J. Control of human behavior.
Glenview, Ill.: Scott, Foresman, and Company, 1974.

Waugh, R. P. Relationship between modality preference and performance.
Exceptional Children, 1973, 39, 465-469.

Weiner, H. Human behavioral persistence. The Psychological Record,
1970, 20, 445-456.

Wepman, J. M. The modality concept. In H. K. Smith (Ed.),
Perception and reading. Newark, Del.: International Reading
Association, 1968.










Wepman, J. M. The modality concept: Background and research.
Atlanta: International Reading Association, 1971. ERIC
Document Reproduction Service No. ED 053 864)

Wepman, J. M., & Morency, A. S. Perceptual development and learning:
An experimental study on modality and reading instruction,
Section II, final report. Washington, D.C.: National Institute
of Education (DHEW), 1975. (ERIC Document Reproduction Serivce
No. ED 013 371)

Whitehead, W. E., Renault, P. F., & Goldiamond, I. Modification of
human gastric acid secretion with operant conditioning procedures.
Journal of Applied Behavior Analysis, 1975, 8, 147-156.

Williams, J., Williams, D. V., & Blumberg, E. L. Visual and aural
learning in urban children. Journal of Educational Psychology,
1973, 64, 353-359.

Wolpert, E. B. Modality and reading: A perspective. Reading
Teacher, 1971, 24, 640-643.

Ysseldyke, J. E. Diagnostic-prescriptive teaching: The search
for aptitude-treatment interactions. In L. Mann & D. A.
Sabatino (Eds.), The first review of special education.
Philadelphia: JSE Press, 1973.

































APPENDICES












APPENDIX A
CVC TRIGRAMS

V I
Trial 1 rib cad dig fig den dim
2 bet mar car pax hap rib
3 bid cab beg fop but dam
4 cam cud den fan gap jet
5 box peg cog lap hew dip
6 tug new his fob mop bib
7 gem cob bug dog bag dot
8 jag ham lot cap keg put
9 rid bus fed bat gag pun
10 gat dad mix fog sin hod
11 bed map jug don pug cut
12 din lid lex sot run sad
13 bog gas sap big get fit
14 mat rov hot pip mad hut
15 bad fad ran gin hop hub
16 pet sod rug sac wig pit
17 jaw mot lip cat fix kid
18 pen kin bed job can hip
19 bas con sup cub nob she
20 way ton log pot sop pap
21 rut hat rip mum hap sis
22 dub wed had gut leg pin
23 mug wan jut top pew set
24 hen jig fun hog hid let
25 dab lit pod him cot kit
26 mid hit rot rim tab bun
27 men wag pop met jet lax
28 yet tot hip day lag tap











APPENDIX B

DIGIT SPANS



1. 2789 4035

2. 4267 1859

3. 1354 0672

4. 2604 1738

5. 7369 4218

6. 9034 2681

7. 6349 1052

8. 2753 4198

9. 1382 9705

10. 5927 8064

11. 5306 1472

12. 6398 2174

13. 2496 0351

14. 8513 2076

15. 0231 6485











APPENDIX C


MISSING UNITS TASK STIMULUS SHEET

TRIALS 1-28

V


Trial 1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28


DEN FIG
PAX HAP


BUT
FAN JET
HEW
MOB BIB
BAG DOT


DAM


CAD DIG
CAR MAR
BEG CAB
DEN CAM
COG PEG
NEW HIS
BUG GEM
LOT JAG
FED RID
DAD MIX
MAP JUG
LEX DIN
GAS SAP
HOT MAT
RAN FAD
RUG SOD
JAW LIP
KIN BED
BAS CON
LOG TON
RIP RUT
HAD WED
WAN MUG
HEN FUN
LIT DAB
ROT MID
POP MEN
HIP IOT


CAP KEG
BAT PUN
FOG HOD
PUG DON
RUN SOT
GET FIT
MAD HUT
GIN HUB
WIG SAC
FIX CAT
CAN JOB
NOB CUB
POT SOP
SIS MUM
PIN LEG
PEW TOP
LET HOG
KIT IIM _
RIM BUN
IMET JET
TAP LAG












MISSING UNITS TASK STIMULUS SHEET

TRIALS 29-56


Trial 29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56


DIG JIB
MAR BET
CAB BID
DEN CAM
COG BOX
HIS TUG
COB GEM
HAM JAG
RID BUS
DAD GAT
MAP BED
LID DIN
GAS BOG
HOT ROB
BAD RAN
SOD PET
MOT JAW
KIN PEN
SUP CON
TON WAY
RIP HAT
WED DUB
WAN JUT
FUN JIG
POD LIT
HIT MID
WAG POP
HIP YET


KIT COT
BUN


LAG


TAB


MET LAX
DAY


DEN DIM
RIB PAX
FOP DAM
JET GAP
HEW LAP
FOB BIB
BAG DOG
PUT CAP
GAG PUN
HOD SIN
CUT DON
RUN SAD
GET PIG
HUT PIP
HUB HOP
SAC PIT
KID CAT
CAN HIP
SAG CUB
PAP SOT
HAP SIS
LEG GUT
SET TOP
HID LET













APPENDIX D

DESCRIPTION OF TRIAL PROGRAMMING



Missing Units Task


There were basically 28 unique sets of trials, each containing

three auditory and three visual stimuli. The slides necessary for

the visual presentation were photographed on reverse black and white

film in order to produce a white word with black background when

projected.

A 140 slide capacity Kodak slide tray was used. The slides

were loaded into the tray in groups of three and a blank space

was left between each group. This provision enabled the experi-

menter to see when the end of any particular trial occurred,

because a large white light would appear on the projection surface.

This light, resulting from the blank slot in the slide tray,

served as a signal for the experimenter to stop the audio cassette

which controlled the slide projector. At this point, the experi-

menter would move the cover sheet in order to expose the stimulus

words associated with the auditory and visual stimuli that were

just presented.

The two stimulus sheets provided 56 sets of stimulus words.

This number was accomplished by varying the missing words from one

sheet to the next. Thus, the experimenter could perform 28 trials,

with one stimulus sheet (trials 1-28), then reset the audio and

visual equipment to start over and use a second stimulus sheet










(trials 29-56) containing different missing words in each trial.

Consequently, after 56 trials, the experimenter would start over

with the first sheet (trials 1-28).

By using 56 trials, each with a different missing unit, it

became very difficult for the subject to memorize answers, even

after repeated exposure to the basic 28 sets of auditory and

visual stimuli.



Digit Span Task


Fifteen unique sets of digits were photographed on reverse

film and placed in an 80 slide capacity Kodak slide tray. These

were programmed simultaneously with 15 unique sets of four digits

on audio tape. The first ten sets were used for the familiariza-

tion trials and the last five sets were used for the criterion

trials. As in the bisensory missing units task, a space was

left after each set of four slides in order to permit the experi-

menter time to stop the equipment and record the subject's

response. The numbers contained in each set appear on the digit

span record sheet.
































APPENDIX E

RAW DATA







BISENSORY DIGIT SPAN RELIABILITY RAW DATA


Reliability
Subject No.

1

2

3

4

5

6

c- 7

8

9

10

11


Age N


10 yrs 1 mo

8 y-s

8 yrs 10 mo

12 yrs 1 mo

12 yrs 10 mo

11 yrs 1 mo

9 yrs 1 mo

9 yrs 11 mo

10 yrs 11 mo

14 yrs 5 no

8 yrs


First Five Trials


No. Correct Responses A/V
Auditory Visual Ratio

16 5 3.20

9 13 .69

14 18 .78

17 12 1.42

14 18 .78

15 13 1.15

16 11 1.45

10 17 .59

17 13 1.31

14 16 .88

20 11 1.82


A-V
Difference

+11

4

4

+ 5

4

+2

+ 5

7

+4

-2

+ 9


Second Five Trials
No. Correct Responses A/V


Auditory


Visual Ratio Difference








RAW DATA SUBJECT 1


Trial No. Response Latency Trial No. Response Latency Trial No. Response Latency
(in seconds) (in seconds) (in seconds)


---- Start Baseline ----

---- Start Day 1 ----

1 I NA

2 I NA

3 I NA

4 A NA


15 A 1

16 V 2,

17 A 1

18 I 2

19 I 6

20 A 2

21 I 1

22 A 1

23 V 3

24 A 1

25 A

26 A 1

27 V 2

28 A 2

----- Start Day 2 -----


30 I

31 I

32 A

33 I

34 A

35 I

36 A

37 V

38 A

39 V

- Visual Reinforce.

40 A

41 A

42 A

43 V,A

44 A


1.0

4.0

4.0

5.0

1.0

5.0

3.7

1.5

1.0

1.5

Starts

1.0

.7

2.0

1.8

2.5










Trial No. Response Latency
(in seconds)


Trial No. Response Latency
(in seconds)


_L~


Trial No. Response Latency
(in seconds)

79 A 1.7

80 A 2.5

81 A 8.0

82 A 2.2

83 A 4.5

84 I 7.2

85 V 3.1

86 A 4.4

87 A 2.7

88 V 1.6

89 I 6.7

90 I 7.0

91 A 5.5

92 A 4.4

93 V .9

94 A 5.0

95 I 4.6




Full Text

PAGE 1

MODALITY RELATED PERFORMANCES IN EDUCATIONALLY HANDICAPPED CHILDREN: AN OPERANT ANALYSIS By MARK A. KOORLAND 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 1977

PAGE 2

To the late Henry Koorland, a very gentle man

PAGE 3

ACKNOWLEDGMENTS Undoubtedly, one of the most difficult tasks is to write acknowledgments because the completion of a dissertation is just one step in a series of accomplishments that is contributed to by many people. Every person's involvement, even if not immediately apparent, leaves a subtle mark. My acknowledgments must first extend back to those who helped to set my career direction. No doubt long forgotten by them, conversations with Ms. Coramae Paganini and Ms. Esther Morgan set the starting point. Shortly after a discussion with Dr. Myron Cunningham nudged me a bit further. Advice and direction from Dr. Wayne C. Richard generated on my part an interest in exceptional behavior. Dr. Lyndal Bullock, Ms. Laverne Graves, Ms. Geraldine Getzen, and Dr. Barbara Reid provided opportunities to learn and to continue my education. Throughout this process, support from close personal friends has often been the mainstay of any accomplishment along the way. It is difficult to fully describe each contribution but, in my eyes, each was and continues to be valued. The people whose names follow have contributed just a small part of their spirit and purpose to the extent that I am different than before: Bill and Kathy Steen, Daniel Hobby, Ryan Beaty, Richard Hersh, Dan Sokol , Robert Alan Tate, Paul and Steve Rozynes, Mary Walsh, l 1 1

PAGE 4

Hugh M. Gramling, Terry Cronis, Marilyn and Gary Warrington, William C. T. Graham, David Westling, Jim Altman, Elliott Lessen, Val and Laura Valentine, and Chari Campbell. A special note of thanks is extended to Grace Hodgson, Carol Springer, and Leila Cantara for their patience, friendship, and capable assistance in the completion of this dissertation. Terry Rose has contributed to the present work and to my personal growth during the course of our 13 year friendship. The number of years reflects the opportunities I have had for his contributions. His tremendous academic and personal capability has been invaluable in the completion of this dissertation. Mitchell Martin has been a long time invaluable friend. He is a top notch professional whose standards have greatly influenced me. Without his cooperation and effort this dissertation could not have been completed. My supervisory committee is composed of individuals who I regard as true teachers in the greatest sense of the word. Dr. Bill Wolking, my chairman, has always been enthusiastic about my work and supportive of me to the extent that I have felt no other direction to move in except forward. Every question Bill has asked has resulted in an answer that made this dissertation more complete. Dr. Cecil fiercer has contributed consistently to my professional growth for a number of years. His input has never been less than essential, especially for this dissertation. He has sharpened my ability to think critically. Dr. Charles Forgnone has permitted IV

PAGE 5

me to gain a multidimensional view of not only special education but also of special education's interface with the culture. I consider him as a valued friend and advisor. From our earliest association, Dr. Bill Reid has consistently been supportive. He has always been positive about my work. I especially value Dr. H. S. Pennypacker's friendship and his readiness to answer all my questions, assist me in my career, and in general arrange the conditions necessary for my academic growth. He has, over the years, taught me no less than the fundamentals of human behavior and the skills necessary to ask questions of nature. Finally, I wish to acknowledge my parents. They have been totally supportive throughout every phase of my education. It made no difference if times were good or bad. They never waivered in their faith in me. For my mother and father, who may only see this work indirectly, I can only say a simple thank you.

PAGE 6

TABLE OF CONTENTS ACKNOWLEDGEMENTS iii LIST OF TABLES viii ABSTRACT ix CHAPTER I INTRODUCTION 1 Statement of the Problem 3 Questions Under Investigation 4 Rationale 4 Variables Affecting Performances 5 Determination of Stable Modality Preference .... 6 Definition of Terms 8 Delimitations of the Study 10 Summary 10 CHAPTER II REVIEW OF RELATED LITERATURE 12 Modality Preference 12 Studies of Significant Interaction 13 Studies of Nonsignificant Interaction 17 Presentation Mode 22 Studies of Handicapped 22 Studies of Non-Handicapped 23 Comparison of Handicapped and Non-Handicapped ... 25 Operant Technology 27 CHAPTER III METHOD 32 Subjects 32 Materials and Apparatus 33 Apparatus and Stimuli 33 Reinforcing Stimuli 35 Settings 35 Procedure 36 Preliminary Procedure 36 Experimental Procedure 37 VI

PAGE 7

Experimental Design 40 Dependent Variables 41 Data Collection 41 Data Analysis 42 Interobserver Reliability 43 CHAPTER IV RESULTS 45 Reliability 46 Analysis of Cumulative Responses 47 Analysis of Baseline 1 48 Analysis of Reinforcement Phase 49 Analysis of Incorrect Responding 52 Analysis of Baseline 2 53 Analysis of Response Latencies 55 CHAPTER V DISCUSSION 60 Findings 61 Interpretation of the Findings 63 Problems and Limitations of the Study 66 Practical Implications 67 Suggestions for Future Research 70 REFERENCES 73 APPENDICES A. CVC TRIGRAMS . . .' 80 B. DIGIT SPANS 81 C. MISSING UNITS TASK STIMULUS SHEET 82 D. DESCRIPTION OF TRIAL PROGRAMMING 84 E. RAW DATA 86 F. VERBATIM DIRECTIONS FOR SUBJECTS Ill G. RECORD SHEETS 114 H. CUMULATIVE GRAPHS 118 I. GRAPHIC DISPLAYS OF RESPONSE LATENCIES 127 BIOGRAPHICAL SKETCH 134 VII

PAGE 8

LIST OF TABLES Table Pacje 1 Demographic and Academic Data 33 2 Percentage Interobserver Agreement for Dependent Variables Across Experimental Conditions .... 47 3 Percent of Responses Across Trials by Response Type and Experimental Phase 50 4 Latency Means by Response Type and Experimental Condition 56 vi 1 1

PAGE 9

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 MODALITY RELATED PERFORMANCES IN EDUCATIONALLY HANDICAPPED CHILDREN: AN OPERANT ANALYSIS By Mark A. Koorland December, 1977 Chiarman: William D. Wol king Major Department: Special Education There has been a great deal of interest in modality preference as a learner characteristic and its relation to learner performance. To the extent that it is difficult to determine consistent auditory or visual modality preferences, some investigators, nevertheless, claim that aptitude treatment interactions between modality preference and presentation mode indicate preferences which are stable across tasks, and that these preferences control the subjects' performances on a short term memory task. If a positive reinforcer proves to control performances allegedly dependent on modality preferences, then there will be a greater understanding of the variables that contribute to the consistency of modality preference, and subsequently an explanation of the reasons for a subject's performance could be offered at a functional level, rather than at a nominal level. For these reasons the study was designed to use a set of tasks previously

PAGE 10

found to be successful in demonstrating an aptitude treatment interrelation between modality preference and presentation mode. The study employed operant techniques to further clarify the variables that might control alleged modality dependent performances with educationally handicapped children. Subjects were classified as to modality preference by their performance on a bisensory digit span task. This task consisted of simultaneous presentation of four-digit spans in five criterion trials. The number of visual responses were subtracted from the number of auditory responses and the magnitude and direction of the differences were used to identify the subjects' preference. Subjects with a preference then performed on a bisensory missing units short term memory task. This task consisted of simultaneous auditory and visual presentations of two different sets of three words. Following the bisensory presentation the subject was shown two of the three words from each set and asked to recall the missing word. To meet the criteria for an aptitude treatment interrelation, the subject had to recall, in any ten trials, six missing words from the same presentation mode as their preference classification. After screening 18 elementary aged children, one male with an auditory preference and treatment interrelation and one female with a visual preference and treatment interrelation were identified for the study. A single subject design employing a multielement baseline was used to evaluate the effects of contingencies of reinforcement on the subjects' bisensory missing units task performance, allegedly under the control of modality preference. Data were evaluated by

PAGE 11

visual inspection of the subjects' cumulative response records and distributions of the subjects' response latencies. Contingencies of reinforcement were found to control both subjects' missing units task performances. Performances in nonpreferred modalities could be made to approach or exceed performances in preferred modalities by the systematic use of contingencies of positive reinforcement. These results imply that the source of control for modality related performances and subsequently the source of stability or consistency of a subject's alleged preference controlled performances may depend on controllable variables in the immediate environment, such as reinforcement. One implication of these findings is that a learner may perform better in a supposedly nonpreferred learning modality than expected, if the consequences for that performance are reinforcing, and if the consequences are arranged systematically. XI

PAGE 12

CHAPTER I INTRODUCTION Operant conditioning and its associated technologies have demonstrated, in various investigations, control over measures of abilities or traits that have been considered relatively stable or immutable characteristics of the person, e.g., IQ scores (Clingman & Fowler, 1976), and sensory thresholds (Goldiamond, 1962). Operant technology provides a viable set of methods and procedures for demonstrating the dynamic nature of learning and associated performances (Edinger, 1969). More specifically, modality preference, a type of individual characteristic (DiVesta, 1975; Ingersoll & DiVesta, 1972), may be under the control of contingencies of reinforcement, rather than under the control of a pervasive trait, aptitude or innate characteristic. Therefore, the purpose of this research was to investigate the stability or consistency across tasks of modality preferences or aptitudes by using operant procedures to demonstrate the function of dynamic environmental variables in the control of performances alleged to be dependent on these preferences or aptitudes. Modality preferences have been considered a type of aptitude that would greatly enhance the study of presentation modalities and their effects on learning and recall (Ingersoll & DiVesta, 1972; Sabatino & Dorfman, 1974). Lilly and Kelleher (1973) used

PAGE 13

the terms modality strength, modality preference, and learning style interchangeably in their investigation of modality preference and learning. In all subsequent sections, because of the frequent interchangeable use of the above terms in the modality literature, aptitudes, modality strengths, and preferred learning modality are used interchangeably when referring to modality preference. Special educators have been proposing that learning disabled children have discrepancies in visual and auditory processes that result in poor achievement (Waugh, 1973). Attempts to determine accurately the visual or auditory modality preferences of handicapped children have not often met with success (Mann, Proger, & Cross, 1973; Sabatino, Ysseldyke, SWoolston, 1972). Consequently, matching these children to appropriate instructional programs also remains difficult. Some investigators have suggested that inadequate instrumentation and measurement procedures may contribute to the difficulty in identifying a learner's modality preference (Estes & Stewart, 1975; Foster, Reese, Schmidt, & Ohrtman, 1976; Mann, Proger, & Cross, 1973; Sabatino, Ysseldyke, & Woolston, 1972; Wolpert, 1971). Because the primary means for identifying modality preference involves the use of norm measures such as the ITPA (Kirk, McCarthy, & Kirk, 1961) and other psychometric modality tasks of many forms, e.g., Wepman, Frostig, Ladd, Mills, and Valett tests (Keeham, 1974), other types of tasks, may warrant investigation. Smith (1971) suggests the use of techniques which are more specifically auditory or visual in nature, i.e., eliminating motor responses or instructions using other modalities.

PAGE 14

Inqersoll and DiVesta (1972) used a laboratory type task to identify visual and auditory attenders among a group of college age students. The task was a bisensory digit span type, consisting of simultaneous presentation of different auditory and visual fourdigit spans. After presentation of the stimuli, the subject was to recall as many digits as possible. The modality in which the subject recalled the most digits correctly was used to classify each subject's preference as either auditory or visual. This task was followed by a type of short term memory exercise consisting of a bisensory missing units task that required the subject to respond with missing words from two independent sets of words. Each independent set was presented simultaneously in a separate modality. The results indicated an aptitude (modality preference) by treatment (presentation mode) interaction implying that the identified modality preferences were stable across these tasks (Inqersoll & DiVesta, 1972). This series of tasks was replicated with a number of educationally handicapped children and indicated what appeared to be stable modality preferences for these learners, while avoiding difficulties found in the use of psychometric modality tasks. Statement of the Problem This study was designed to determine if modality related performances in elementary aged educationally handicapped children were subject to control by contingencies of reinforcement rather than control by a pervasive trait, aptitude, or innate characteristic,

PAGE 15

Consequently, existence of a functional relationship between contingencies of reinforcement and the type of response, allegedly controlled by modality preference, on bi sensory short term memory trials was investigated. Questions Under Investigation This study was designed to answer the following questions regarding the influence of contingencies of reinforcement on performances allegedly controlled by modality preference in elementary school aqed educationally handicapped children. 1. Is there a functional relationship between contingencies of reinforcement and short term memory performance allegedly under control of a visual modality preference? 2. Is there a functional relationship between contingencies of reinforcement and short term memory performance allegedly under control of an auditory modality preference? Rationale At the core of special education is the search for strategies leading to effective individualization. The determination of a student's preferred learning modality in order to insure that appropriate treatments are provided, is one way in which individualization is carried out. Because the modality concept has been promoted as well as used as an explanation for achievement or lack thereof (Wepman, 1968), it has resulted in a proliferation

PAGE 16

of assessment devices and even modality specific commercial teaching materials. Unfortunately the emphasis on the modality concept has resulted not only in commercialism, but also in wide acceptance of the general strategy of determining modality preference and subsequent selection of materials by sensory modality. Many involved in the modality enterprise have overlooked the complexity of the learning processes involved (Landers, 1976) and the inconclusive nature of modality research. The rationale for this study is derived from two substantive areas related to modality research. They consist of the variables affecting performances allegedly under the control of modality preference, and the determination of stable modal preferences. Variables Affecting Performa nces When determining the variables that affect modality dependent performances, most often demographic characteristics such as age, IQ, socioeconomic status (Blanton, 1971; Williams, Williams, & Blumberg, 1973), developmental processes (Wepman, 1971), and sex (Dwyer, 1971; May & Hutt, 1974) are investigated. Infrequently, but nevertheless important, motivational characteristics of the immediate learning environment have been suggested as factors that may affect the student's performance in learning modality studies (Dunn & Dunn, 1974; Williams, Williams, & Blumberg, 1973). In contrast, Wepman (1971), in reference to modality preference, states, "the choice appears not to be conditioned by the environment but seems more likely to be an innate characteristic" (p. 6). In order to clarify the relationship

PAGE 17

of variables in the immediate environment to modality dependent performance, the use of operant technology would seem warranted. Operant conditioning procedures have been shown to modify behaviors thought to be controlled by pervasive organismic variables, relatively resistant to intervention. These variables are, for example, autonomic responses that control human gastric acid secretion (Whitehead, Renault, & Gol diamond, 1975), neurological impairments that control written letter reversals (Smith & Lovitt, 1973) and learning disabilities and general mental abilities that control aptitude and intelligence test scores (Clingman & Fowler, 1976; Kubany & Sloggett, 1971). These studies demonstrate the influence of environmental variables and subsequent changes in behaviors thought to represent pervasive characteristics. There has been no systematic investigation, using operant technology, of performances allegedly dependent on modality preferences. Determination of Stable Modality Preference There appears to be wide agreement that the detection of modality preference is impeded by the lack of reliable and valid instrumentation (Ysseldyke, 1973). In order to move toward tasks that effectively divide attention between visual and auditory channels and thus indicate the primary modality of attending, the bi sensory simultaneous stimulation paradigm has been employed (Dornbush, 1968). Ingersoll and DiVesta (1972) mention that Broadbent, while employing this type of recall task noted that subjects had their own consistent preference for giving information

PAGE 18

received in the auditory or visual channel first. For this reason, Ingersoll and DiVesta used the bisensory digit span task to identify visual and auditory preferences of their subjects. After identifying the preferences, the subjects were given a second task which was a bisensory modification of the missing unit paradigm. This procedure consisted of presenting five words to the subject, then repeating four of the five words and asking the subject to write on a record sheet the missing word or unit. The bisensory modification involves the simultaneous auditory and visual presentation of different sets of words. The authors found that those subjects identified as having auditory preferences recalled more auditory materials and those identified as having visual preferences recalled more visual materials. DiVesta (1975) notes that to identify the missing word from the originally presented sets is an impossible task unless one assumes that people develop consistent, selective, and automatic information processing strategies. Consequently, these laboratory-type tasks might be more directly responsive to auditory or visual preferences and thus avoid some of the problems encountered with other methods of assessing modality preference, e.g., psychometric instruments. Additionally, Ingersoll and DiVesta (1972) state that replication with different subjects is necessary. They point out that the use of elementary school children, a more heterogeneous group than the original population of college students, would allow selection of more extreme scores relative to modality preference.

PAGE 19

Extreme scores should produce clear response patterns on the bisensory missing units task. If the results can be replicated, then the authors believe that the generality of the construct of modality preference can be extended. Elementary age educationally handicapped children are often the focus of individualized learning treatments and modality related materials. Their use as subjects seemed warranted, because of age and relevance to the modality enterprise. This set of tasks reliably indicated whether modality preferences existed in the population under study and also provided results similar to the aptitude treatment interaction (ATI) between modality preference (aptitude) and presentation mode (treatment) obtained by Ingersoll and DiVesta (1972). Results similar to the original ATI implied that these modality preferences were stable across tasks. Once results similar to Ingersoll and DiVesta's (1972) findings were obtained, the effect of contingencies of reinforcement on task performances allegedly under the control of modality preference were investigated. Definitio n of T erms 1. Modality Preference Preferences are "monitoring and filtering systems which control the flow of information within a processing system" (Ingersoll & DiVesta, 1972, p. 391). This is determined by the magnitude and direction of correct recall on a bisensory digit span task. After five sets of trials, the number of visual responses are subtracted from the number of

PAGE 20

auditory responses. The previously set criteria of + 3 (Ingersoll & DiVesta, 1972) was used to classify subjects. Those who did not meet the established criteria were dismissed from the study. 2. Bi sensory Digit-Span Task Simultaneous presentation of different auditory and visual four-digit spans. The subject is asked to recall the set in whatever order he prefers. 3. Bisensory Missing Units Task Simultaneous presentation of different three word sets. One set is presented auditorily and the other visually. The words are drawn from A Teacher's Word Book of Twenty Thousand Wo rds (Thorndike, 1931). The subject responds with the missing word, when two words from each set are presented visually following the original presentation. 4. Aptitude Treatment Interaction (ATI) An ATI is a significant interaction between personological variables and alternative treatments. This study does not use statistical tests of the subjects' data to operationally define an ATI. Rather, an ATI is said to exist for those subjects who demonstrate a modality preference and also remember 60% of the words presented on a sensory channel corresponding to their indicated modality preference. 5. Educationally Handicapped Child Children who are classified as non-retarded and who are at least one year below age appropriate grade placement in one or more academic areas. The child has no physiologically based sensory handicaps.

PAGE 21

10 Delimitations of th e Study The use of the bisensory digit-span tasks have not been validated as an indicator of modality preference with any task other than the missing units task. The generality of any modality classification derived from this procedure to tasks of varying content and complexity levels is not yet established. As such, a modality preference determined with the laboratory tasks may not be generalized beyond short term memory functions. The use of educationally handicapped students as defined earlier may impose restrictions on the generality of the findings. Any operational definition for educationally handicapped students is bound by the limits of the instruments used to determine subject characteristics. Summary The determination of modality preferences that are stable across tasks in exceptional students has met with limited success. The use of psychometric instrumentation has been cited as a factor contributing to these results. Laboratory tasks, however, were used by Ingersoll and DiVesta (1972) to demonstrate stable modality preferences in adults. Using these similar tasks evidence of stable modality preferences were sought with educationally handicapped children and, when found, further investigated by the use of operant technology. Operant techniques

PAGE 22

11 were used to investigate the responsiveness of alleged preference controlled behavior to environmental influence.

PAGE 23

CHAPTER II REVIEW OF RELATED LITERATURE The following review is divided into three sections. In the first two sections salient studies related to modality preference and mode of presentation are reviewed. Within these sections, studies are further divided by type of experimental population. In the third section, operant technology and its value for modality research are discussed. Modality Preference A mode or modality is a sensory channel, such as visual, auditory, tactile or muscular movement, through which sensations are transmitted and received (Jones, 1972). General interest in the area began in the latter part of the nineteenth century. The purpose was to compare listening and reading as input channels for comprehension and analysis of various materials (Jones, 1972). Ysseldyke (1973) points out the influence of Cronbach and Reynolds in urging psychologists and special educators to research how individual's aptitudes could be joined with instructional systems or treatments in order to identify and use aptitudetreatmentinteract ions. Bracht (1970) states the goal of aptitude treatment interaction research is identification of disordinal interactions between personological variables (IQ, specific abilities, personality variables, 12

PAGE 24

13 and others). In light of the above, an interest developed among special educators in determining the most efficient channels of learning. Evaluation instruments and complimentary instructional programs have been devised, based on the premise that the identification of preferred modalities in instruction would lead to more effective learning (Waugh, 1973). Additionally, it was felt that weak modalities could be identified and instructional activities could be designed to remediate those weak modalities or processes. There remains much emphasis on the determination of modality strength for purposes of matching specific instructional procedure to a given learner's aptitude (Dunn & Dunn, 1974; Sabatino & Dorfman, 1974). Despite often contradictory results, some investigators feel that it is premature to discard research on modality preference (Foster, Reese, Schmidt, & Ohrtman, 1976). Most often problems in modality research are related to accurate assessment of modality preference (Jones, 1971; Jones, 1972; Mann et al. 1973; Smith, 1971) and the use of treatments that are uniquely auditory or visual in nature (Foster, Reese, Schmidt, & Ohrtman, 1976; Lilly & Kelleher, 1973; Smith, 1971). The following review compares and contrasts successful and unsuccessful modality treatment interaction studies with handicapped and non-handicapped populations. Studies of Significant Interaction The published studies indicating modality preference and treatment interaction are relatively few. Indeed, the number of studies indicating any aptitude by treatment interaction is small. Bracht

PAGE 25

14 (1970) reviewed 90 aptitude-treatment-interaction studies and only found five that met the conditions for a disordinal interaction. It must be kept in mind that interaction can be of two types: with or without reaching significance. Bracht's requirements were of the former type and as such fairly rigorous. Researchers after 1970 have attempted to improve on methodological features. Studies of non-handicappe d. An investigation (Ingersoll & DiVesta, 1972) with a college age population obtained an interaction of modality preference and presentation mode for a short term memory task. The value of this work to the present discussion is the nature of the task. The subjects were engaged in a bisensory digit span task that required them to recall numbers presented simultaneously in an auditory and visual presentation. Those identified as auditory or visual attenders were then given a bisensory missing units task consisting of five words simultaneously presented in the auditory and visual mode. The subject was then presented four words from each modality and asked to recall the missing word. Modality preference and its interaction with the treatment were found. The authors state that the results indicate the stability of modality preferences across the two tasks. This procedure meets many of the criteria for tasks that have been found to produce accurate indications of modality preference, i.e., stability across tasks or interaction with mode of presentation, a task involving short term memory, and relative face validity of the task for a particular modality. Kalin and McAvoy (1973) demonstrated that college age students provided valid data when asked to rank their modality preference. The

PAGE 26

15 validity was apparent, because the data were used to provide accurate matches to instructional methods. Those students choosing visual presentation performed better in that mode than in the auditory mode and vice versa. Of course, this self disclosure of preference was obtained from a group keenly aware of study procedures and associated difficulties. Consequently, the utility of this approach, for children, would be limited. Wepman and Morency (1975) obtained a successful modality-treatment interaction in normal first grade children. They used tests with high content validity in order to measure the perceptual parameters of interest. Some of the tasks involved auditory and visual presentation of discrimination tasks (requires same or different response) and memory tasks (repeat numbers in order and words in any order). A standardized reading achievement test was given after phonic and sight word approaches were used for treatments. An interaction between modality preference and presentation mode of treatment was obtained. Wepman and Morency (1975) noted that success was due to the use of auditory memory and sequential memory tasks that others have failed to use. Also, there were no motor components involved in the tasks and it is felt that motor components may have confounded previous attempts at modality preference ^identification. In the most recent and successful study, Foster, Reese, Schmidt, and Ohrtman (1976) used a small number of non-motoric tasks, related directly to reading, e.g., phoneme discrimination and word recognition. These tasks were used for modality classification. The subjects successfully were taught unknown words by the use of pictures and by the

PAGE 27

16 use of phoneme-grapheme relationships. Parenthetically, the authors note that many children were screened in order to form the small experimental groups for this study. The effort required to identify such a small sample prompted the authors to point out that teaching through the preferred modality may be unimportant for the majority of children. Conversely, research on teaching through a preferred modality might be all the more important for the minority of learners classified as educationally handicapped. Stud ies of handicap ped. To date there is only one study using a population identified as handicapped that found modality preference to interact with a treatment. In 1973, Lilly and Kelleher attempted to correct some problems typically found in modality studies. They cited difficulty with dependent variable measurement and noted that tasks with at least face validity in relation to the reading process were necessary. In order to identify modality preference in subjects, they used a visual or auditory memory task for words. Points were assigned for words remembered and preference was determined by subtracting auditory scores from visual scores and using a certain number of points in one direction as criterion for a particular preference. The dependent measure involved reading or listening to a passage and recalling salient facts. Those subjects originally identified as preferring the auditory modality performed better on the audio passage and those identified as preferring the visual modality performed better on material read. In order to perform this study, however, the students had to read at least at a 2.5 grade level. Consequently, the determination

PAGE 28

17 of modality preference for younger students would present difficulties. Of value, though, is the demonstration of high face validity in tasks and more or less direct measures of performance, i.e., number of facts remembered as a dependent variable. St udies of Nonsignificant Interaction This category is the largest of modality preference studies. The dependent measures are primarily concerned with some aspect of reading. Jones (1971), in a review, points out that various researchers in studies employing psychometric tests such as the ITPA (Kirk, McCarthy, & Kirk, 1961) or combined batteries of tests (deHirsh, Jansky, & Langford, 1966) have difficulty supporting the theory that modal preference of an individual should be considered in teaching him to read. Jones concludes that it is difficult to isolate the auditory or visual modality, but recommends more experimentation to determine the relationship of modal preference to learning. Studies of non-handicap ped. Bruininks (1963) used the instrumentation approach typical of many unsuccessful investigations, i.e., psychometric instruments of a normative nature. Using a battery of 12 tests, he found that groups of auditory and visual learners did not profit from reading instructions based on perceptual strengths or weaknesses. Because the subjects, low socioeconomic status (SES) second and third graders, had been in school up to three years, the authors concluded that the students' reading habits from previous instruction might be more influential than perceptual characteristics in determining their performance. This is an atypical conclusion in

PAGE 29

13 modality studies in that experience is noted as a possible factor affecting the results. The author also states that the tests used may not measure the most significant factors involved in reading instruction. Cullinan, Ringler, and Smith (1969) using a relatively concise test, the New York Modality Test, composed of visual, auditory, and kinesthetic subtests, identified three groups of first grade urban learners corresponding to three modal preferences. The learners received visual, auditory, and kinesthetic word recognition training emphasizing configuration, specific sounds, and words formed by using pipe cleaners as letters fixed to a card. Using the Metropolitan Achievement Test as a dependent variable, they found no differences between the experimental groups. The authors conclude that the use of a standardized reading achievement test with inner city subjects, insufficient teaching time (7h hours), and no coordination between the regular class activities and the various experimental treatments were possible problems. This study is one of the earliest to mention difficulty with using standardized tests as the dependent measure. Smith (1971) in a study of low SES primary subjects found no relation between modality strength, classified using ITPA subtests, and visual treatments (use of the Initial Teaching Alphabet and Words in Color) or auditory treatments (reading with phonics). The Metropolitan Achievement Test served as a dependent measure. Smith concludes that if differences in children, relative to auditory and visual functioning exist, then the ITPA may be inappropriate for diagnosis. This finding is in agreement with earlier work (Bateman,

PAGE 30

19 1963; Cripe, 1966). The dependent measure, however, was not mentioned as possible limitation for Smith's (1971) study. Newcomer and Goodman (1975) used the ITPA as part of a battery of 18 tests. They found no interaction with an associative learning task or a meaningful learning task. Additionally, they found the visual condition of the treatment superior for the majority of the 4th grade normal subjects in the study. This result is in keeping with Bruininks (1968) earlier work. Studies of handicapped . Waugh (1973) using the ITPA and an additional test attempted to correct for inadequacies of tasks used for dependent measures in earlier studies. Simplified recall along with word recognition tasks, similar to tasks used in actual instruction, were used. Waugh found that there was no relation of preference to treatment. In an ad hoc analysis it was noted that the auditory subjects did better on the auditory and visual recall . task. Thus, the ITPA might serve to identify auditory perceptual functioning related to tasks at the recall level of complexity. For the reading task, however, no statements could be made for either the auditory or visual groups. Possibly reflecting the results of all modality preference research up to that point in time, Waugh (1973) states, "special educators at all levels must continue to ask an essential question is this really a useful activity?" (p. 469). Waugh' s population was handicapped only to the extent that they were enrolled in Title I reading. To charge special educators with such an essential question was probably premature in view of the heterogeneity to be found in most special

PAGE 31

20 education populations. At best, Waugh's population, as a whole, would be considered borderline exceptional by most special educators. Sabatino, Ysseldyke, and Woolston (1972) gave a battery of perceptual tests to mentally retarded subjects. The treatment involved unisensory perceptual training matched to the indicated strength of the subject. Gains were found on perceptual measures but were not specific to the kinds of training received. Group reading achievement tests were given as an additional dependent measure and no significant gains were detected. The authors conclude that present modality classification instruments have sizable error and may classify children in specific modality areas by chance. Using 116 mentally retarded children as subjects, Sabatino and Dor f man (1974) arranged three groups consisting of auditory, visual, and non-preferred categories. They used a normative battery of tests; some derived from the ITPA. These groups were matched to a commercial visual reading curriculum (Sullivan Series) and an auditory reading program (Distar). Pre and posttest scores, obtained using an individual achievement test, produced no interaction. The authors conclude that possible lack of reliability in the instruments and use of measures that were not factorily simple may have contributed to the nonsignficant findings. Tyler (1974) using a similar population and instrumentation received similar results. Parenthetically, the author mentions that all responses of the subjects in the present study were reinforced. This is the first statement about environmental contingencies

PAGE 32

21 to be found in any of the modality preference-treatment studies to date. In summary there are some statements that apply to most of these studies in modality preference: 1. Every test of modality preference employing the ITPA or selected subtests provided non-significant modality-treatment interactions. 2. The use of large comprehensive batteries to determine modality preference essentially did not prove to be more effective than the use of a smaller number of tests. 3. The use of normative or psychometric instrumentation appeared, in general, to lack the sensitivity to detect modality preferences or treatment gains. 4. Little or no information is provided about experimenter application of contingent reinforcement to the performance of subjects. Unless there is an interaction between the preferred modality and the specialized treatment, it is \iery difficult to determine where the basic weakness in the study lies. It could be in the modality identification procedures, purity of the treatment, or the method of analysis. There is a possibility that, because of the present state of the art, few reliable generalities can be made about modality preference.

PAGE 33

22. Presentation Mode Many studies related to the learning modality area are concerned with presentation mode in particular. A unifying feature of many of the studies included in this area is the type of dependent measure used and the data analysis model. For the most part, tasks used for dependent measures are paired associate, short term memory, or other specific skills (Silverston & Deichmann, 1975). In every study, group research designs and confidence level statistics are used for analysis. For these studies, researchers may employ handicapped subjects or handicapped and non-handicapped comparisons, but the majority employ elementary level low SES or normal populations. The studies are reviewed by population and various distinguishing features are mentioned. Studi es of H andicapped In studies of learning disabled children, Estes and Huizinga (1974) and Estes and Stewart (1975), using paired associate tasks under visual and auditory presentation modes, noted visual tasks produced a greater number of correct responses than auditory. It is suggested (Estes & Stewart, 1975) that instructional materials for these subjects should be oriented to visual processing, yet in the same discussion the authors state, "in learning disabilities . . . heterogenity of psycholinguistic functioning and learning styles is the rule rather than the exception" (p. 5). This seems inconsistent with the determination that learning disabled children in their studies performed best under the visual presentation mode.

PAGE 34

23 In a study of retarded adults, McConkey and Green (1975) investigated presentation method and its effect on free recall of categorically related items. They used auditory, visual sequential, and visual simultaneous presentation modes. Although primarily interested in the use of categorial strategies during recall, they noted that both groups receiving visual input recalled more items than the group receiving auditory presentation. This result is congruent with many of the other studies (Bruininks, 1963; Estes & Huizinga, 1974; Estes & Stewart, 1975; Newcomer & Goodman, 1975). Stu dies of Non-Handica pped Galfo (1970) in a study based on Broadbent's limited capacity processing model, which states that redundant presentation of program material will result in no increase in learning, obtained results at variance with the limited capacity theory. He found that simultaneous sight sound presentation with normal subjects proved superior to sight and sound separation. The task, however, involved academic material, and as such might involve level of familiarity factors or complexity exceeding that of tasks used in other studies in this area. A study more typical of the mode of presentation and specific skills type under review was conducted by Filmer and Linder (1970). Using low SES subjects, they employed a short term memory task requiring recall of colors, digits, and pictures. The task was presented in the auditory, visual, and auditory-visual combined modes. The combined mode yielded the highest scores on each task; the auditory yielded the lowest. These results are in agreement

PAGE 35

24 with Galfo's (1970) study. Of special importance, however, is that the researcher states that immediate feedback on correct and incorrect responses was given. This is the earliest and one of the few indications of the use of feedback or reinforcement. It must be kept in mind that the feedback was not a variable under investigation. Reynolds, Bickley, Champion, and Deckle (1971) noted that very few of the studies employing word association tasks take mode of presentation of stimulus words into consideration. Consequently, they obtained results with low SES subjects that indicate responses to word association tasks are a function of stimulus modality. They also noted that as children get older the difference in type of response to oral and graphic presentation of words becomes less pronounced. Silverston and Deichmann (1975) noted age dependent relationships between modality skills and various reading indices. It would appear, then, that experience and other environmental variables may interact with performance on tasks related to various modalities. In a study of encoding processes, in second grade children's short term memory, Corsale (1974) used mode of presentation as an independent variable. The subjects were required to recall categorically related words after a brief interference period. Although mode of presentation for this type of task plays a role in adults, it was not found to have an effect on the children's encoding. Although Corsale's procedures differed because of the use of an interference procedure, this result is interesting in relation to McConkey and Green's (1973) finding that mode of presentation for retarded adults differentially affected recall of categorically

PAGE 36

25 related items. It might be of value to determine if there is any parallel performance for handicapped children as there is for both handicapped and non-handicapped adults. Mode of presentation occasionally appears to have generality across wery different populations. Williams, Williams, and Blumberg (1973) found that middle class subjects perform better with visual presentation in a paired associate task while lower class subjects learned better under an auditory condition. In relation to social class and performance on paired associate learning tasks, they state the following: because of related motivational differences it is difficult (unfortunately) to find a situation entirely free of these possible confounding factors, (p. 358) It is interesting to note that some researchers of modality choose either to ignore motivational variables or to treat motivational variables as a nuisance rather than subject matter worthy of investigation in its own right. Comparison of Handicapped and N o n-Handicapped Budoff and Quinlan (1964) compared auditory and visual presentation modes in normal and retarded readers. They used a meaningful paired associate learning task for second graders and found that the auditory presentation was superior in both groups. This finding does not hold true across other studies (Bruininks & Clark, 1972; Katz, 1967). Katz (1967) employed a discrimination task presented in the auditory and visual mode at different levels of familiarity using

PAGE 37

26 normal and retarded readers. Familiarity was defined as frequency of occurrence in the culture. The authors employed the Hebrew alphabet and associated letter names as the unfamiliar stimuli and the English alphabet and associated letter names as the familiar stimuli. Katz (1967) found that stimulus modality interacted with level of familiarity. He concluded that effects of stimulus modality are differentially related to the response required of the subject. The best performance was obtained with the visually familiar stimuli. These results raise the issue of stimulus meaning as a variable which will influence results. In many of the presentation mode studies this variable is not taken into consideration. Bruininks and Clark (1972), using first grade disadvantaged educable mentally retarded children, disadvantaged non-retarded, and advantaged non-retarded, studied the efficacy of auditory, visual, and combined presentation modes. The results failed to show any difference on paired associate tasks between disadvantaged and nondisadvantaged of similar intellectual ability. Overall the total group responded best under the visual condition. The authors concluded that young retarded children exposed to auditory activities ought to have visual picture opportunities at the same time. To the extent that this study is typical of the others, it is also unusual because the authors point out that the subject was not informed of his correct or incorrect responses. This is opposite of the tactics that Filmer and Linder (1970) employed. Both studies used audio, visual and combined presentation modes, both studies employed low SES subjects and recall tasks, yet different results were obtained.

PAGE 38

27 The differential use of feedback or reinforcement may have contributed to the difference in outcomes. In summary, from the studies reviewed some conclusions may be offered. 1. Rarely is information concerning consequences of the subjects' responding made available. 2. The majority of the studies appear to demonstrate favorable results under visual presentation modes, but because of the prevalent group research strategies the performance of individuals are not reported. 3. Studies employing a non-complex task, i.e., short term memory, provided more consistent results than studies employing complex tasks, i.e., study of academic materials. From these conclusions and conclusions of the nodal ity Preference section, the need for alternative research strategies, continued use of non-complex tasks, and attention to environmental variables becomes apparent. In the following section on Operant Technology alternative research strategies and environmental variables are considered. O perant Technology The growth and development of operant technology has been steady and has had impact on diverse areas of research (Kazdin, 1975). More specifically, the techniques of applied behavior analysis, which embrace many of the traditions of present laboratory practices, have been directed toward the needs of special educators, who, by definition, are charged with bringing unique procedures,

PAGE 39

28 materials, and modifications of, or additions to school practice for the handicapped (Kirk, 1962). Applied behavior analysis attempts to demonstrate quantitative characteristics of the behavior, the experimental manipulations, and the technologically exact description of the methods used to bring about change in behavior. Additionally, the change must be effective enough to have value and also possess generality within limits (Baer, Wolf, & Risley, 1968). To the extent that science is a rigorous exercise in measurement, applied behavior analysis (ABA) requires certain standards of measurement and design. The use of intra-subject designs permitting replication (Sidman, 1960) is considered basic. Additionally, measurement requirements for the assessment of reliability of behaviors under observation are an integral part of ABA practice. To appreciate that behavior of importance in applied settings is the focus of ABA, one only needs to consider the diversity of settings in which ABA techniques have taken place: day care centers, early education, Head Start, elementary schools, secondary education, college settings, special education, community settings, and clinics (Ulrich, Stachnik, & Mabry, 1974). Operant conditioning has been applied not only in a wide range of settings, but also for the purpose of investigating a wide range of behaviors in those settings. Repeated demonstration of reinforcement as the controlling variable for certain classes of behavior have established the generality and usefulness of the reinforcement principle for those classes.

PAGE 40

29 Recently, operant conditioning strategies have been applied to relatively traditional research areas (Clingman & Fowler, 1976; Goldiamond, 1962; Lovitt & Smith, 1973). Goldiamond (1962), in a discussion of perception, points out the importance of questioning the locus of response change in sensory threshold, hypnotic perception and related perceptual research. In relation to controlling variables, Goldiamond states, "where it is possible to choose between sensory and response loci, the evidence has supported a response locus" (p. 315). Just as the application of new procedures to perceptual research (Goldiamond, 1962) may have evoked comments of doubt by those who had a vested interest in the area, i.e., psychophysics researchers and traditional psychologists, so has the initial application of operant procedures to newer areas. Psychometric testing has in the recent past received some attention from individuals working with operant technology. Because of the standard procedures that must be adhered to (Cronback, 1970) in order to appropriately administer a psychometric instrument such as an intelligence test, any variation from that procedure is seen as a source of invalidity for the results. Commenting on the use of contingent reinforcement for correct answers on such instruments, Conner and Weiss (1974) state that all scores will shift upward and each subject's position in the distribution will remain the same. The problem with Conner and Weiss's conceptual analysis is that it was not based on experimental data; it was based, no doubt, on their measurement training and theoretical background. When the effect of

PAGE 41

30 reinforcement was determined, it was found that the scores in different ranges were differentially affected by the experimental manipulation (Clingman & Fowler, 1976). Of more important value than the effect obtained was the process uncovered and the additional understanding of the interaction between psychornetrics and contingencies of reinforcement. Examples of the use of operant technology to examine areas previously thought to be confounded by or impervious to the influence of immediate environmental variables are increasing. Whitehead, Renault, and Goldiamond (1975) studied human gastric acid secretion with operant conditioning procedures. Ayllon, Layman, and Kandel (1975) investigated the comparative use of medically prescribed drugs and reinforcement to control hyperactive behavior, often thought to be under the relatively exclusive control of physiological variables. Smith and Lovitt (1973) have demonstrated the remediation of number reversals that were previously thought to be under the control of neurological impairments or immature physiology. In the main, operant technology has been helpful in providing understanding of the extra-organismic processes that interact with many behaviors thought to be under the primary control or organismic variables. Skinner (1974) points out the lack of utility in discussing the causes of behavior, using instinct related explanations. To say that the duckling follows the mother because of imprinting is not a thorough enough explanation. "Imprinting" as the explanation of "following behavior" eventually becomes accepted

PAGE 42

31 as the cause of the behavior, when, in fact, it is simply explanation at a nominal level. The more subtle effects of the environment are overlooked in this type of analysis. In much the same way, to say that a student learns visual material best because he has a preference for the visual mode is to explain cause at a nominal level, rather than at a functional level. The further discovery of controlling variables may be undermined by the use of current terminology and explanatory constructs. In summary, the area of modality preference in children has not been subject to investigation using operant technology. Consequently, the purpose of this study is to investigate the sensitivity of modality controlled performances to environmental variables, thus leading to a fuller understanding of their effect on this phenomenon. Important considerations for this investigation, emerging from the previous review of literature, are the inclusion of non-complex tasks and employment of alternative research strategies.

PAGE 43

CHAPTER III METHOD In Chapter III the methods and procedures of the study are presented. This includes a description of the subjects, materials and apparatus, settings, procedures, and experimental design. The study was composed of three phases: 1. Identification of eligible children. 2. Identification, using non-complex tasks, of those eligible children meeting criteria for having a modality preference. 3. Application of reinforcement procedures to alleged modality dependent performance. Subjec ts The two subjects were educationally handicapped (EH) elementary students. The subjects were non-retarded and at least one year behind age-appropriate grade level in reading or in arithmetic skills as assessed on the following achievement tests: a. Slosson Oral Reading Test (Slosson, 1963). b. Wide Range Achievement Test Arithmetic Subtest Level I (Jastak & Jastak, 1965). In order to qualify for the study, the subjects had to know: a. Names of numbers 1, 2, 3, 4, 5, 6, 7, 8, 9, 0. (Integers 1-0). 32

PAGE 44

33 b. Consonant vowel consonant (CVC) trigram words. (See Appendix A) . Each child was shown a list of the words and numbers used in the study. They were required to read the material with 100% accuracy. Demographic and academic data are reported on the children qualifying for the study (see Table 1). Table 1 Demographic and Academic Data Father's Appropriate Reading Arithmetic Subject Age Sex Occupation Grade* Grade Grade 1 8 yrs.ll mos. M Professional 4.0 4.9 3.0 2 9 yrs. 8 mos. F Semi-skilled 5.0 6.2 2.8 ^Because academic testing and demographic data were obtained in late summer, appropriate fall grade placement is given as age-appropriate Materials and Apparatus Apparatus and Stimuli Bisensory digit span task . Integers (1, 2, 3, 4, 5, 6, 7, 8, 9, 0) were randomly assigned to series of digit spans which were four digits in size ( see Appendix B). For each series there was a simultaneous audio and visual presentation. The same number could not occur twice within a bisensory presentation.

PAGE 45

34 Bisensory missing units task . CVC tri grams were selected from a list of words (Thorndike, 1931) within the typical subject's reading capability (see Appendix A). Ambiguous words or homophones (son and sun), as well as words of related meaning were eliminated from any bisensory trial. Numbers were also eliminated (ten and six). Media . Visual stimuli were projected onto a suitable surface and the audio stimuli were provided by a cassette tape recorder (Buhl Programma, Model 616-150) built to operate the projection equipment (Kodak Carousel Slide Projector, Model E-2). The equipment was programmed so that a sound-relay synchronized the auditory and visual presentation in order to deliver them simultaneously. Digits were presented at a rate of one every two seconds on each channel. Because of limitations of the equipment, the visual duration of exposure was .9 second. The auditory duration averaged .6 seconds in length. These presentation times approximate the specifications provided in Ingersoll and DiVesta's (1972) procedure description. Stimulus sheet . After a bisensory missing units trial, a stimulus sheet (see Appendix C), containing two of the three words from the auditory and two of the three words from the visual presentation, was placed in front of the subject. A cover sheet with a 175 mm. x 12 nm. opening was placed on top of the stimulus sheet. In this fashion, after each trial, the experimenter would slide the cover down the stimulus sheet in order to expose the proper set of

PAGE 46

35 stimulus words, one set at a time. A complete description of the arrangement used in programming the trials is given in Appendix D. Reinforcing Stimuli Nickels were chosen as the reinforcer. This particular currency was chosen because of its generalized value and ease of handling. Nickels were given to the subjects for emitting the appropriate response upon completion of certain missing units trials. As they were earned, nickels were placed in a clear plastic container in direct view of the subject. Settings Subject 1 was enrolled in a private school for learning disabled children. An unused classroom was available for carrying out the experiment. The classroom was well lighted and air conditioned. The subject was seated at a 106 cm. x 106 cm. table. The projection equipment was immediately to the subject's left and the audio equipment was at the subject's right where it could be controlled by the experimenter. The experimenter was seated to the right of the subject. The experimenter placed the paper stimulus materials necessary for the missing units task directly in front of the subject. As the subject proceeded from one trial to the next, the experimenter controlled the paper stimulus materials with his left hand while he operated the equipment and recorded responses with his right

PAGE 47

36 hand. The slide portions of the bisensory tasks were projected on a surface approximately 106 cm. in front of the subject. Subject 2 was enrolled in a special summer program for mildly handicapped children. This program was operated at a laboratory school and a spare classroom was used. The classroom was similar to the one used with Subject 1, except that it was not air conditioned. The experimental arrangement was similar to that of Subject 1 . Procedure Preli minary Pr ocedure^ Bisensory digit span task reliab ility. Because the bisensory digit span task had only been used with college freshmen, it was necessary to determine its reliability with a younger population. Internal consistency was determined by giving a group of 11 subjects (x C.A. = 10 yrs. 5 mos.) ten criterion trials and then forming a ratio of the number of auditory correct responses to the number of visual correct responses for the first five criterion trials and then for the second five criterion trials (see Appendix E). A Pearson product moment correlation resulted in r = .69 for these two sets of ratios. Pilot study . In order to develop and refine the procedures for the actual experiment, a pilot study was performed. Two important outcomes resulted from the pilot effort. First, the multielement baseline that was used incorporated a verbal command as a discriminative stimulus for the subject. In

PAGE 48

37 other words, the subject was told beforehand whether to attend to the auditory or to the visual stimuli and was reinforced for a correct response. After using this procedure with the pilot subject, it became apparent that it was difficult, if not impossible, to determine whether the instructions or the reinforcement schedule controlled the subject's responding. Consequently, in the actual experiment the instructions were eliminated and only the reinforcement schedule was used to influence responding. The second outcome was the attainment of a recoverable baseline condition that required a relatively large number of trials before the experimenter observed stability. This was in contrast to the smaller number of trials necessary to observe stability in the original baseline phase. This outcome alerted the experimenter to the possibility of using a larger number of trials than was anticipated originally. In fact, many trials were necessary in order to gain stability in a number of phases during the actual experiment. Experimental Pr oced ure Re inforcement . After the modality classification and missing units task baseline were obtained, each subject was asked, prior to the actual contingent use, if he or she would like to earn nickels. Both subjects responded emphatically that they would. Bisensory Tas ks. In the digit span task, the subject received ten trials during which he or she was told beforehand from which modality to recall first. These responses were not recorded; they served to familiarize the subject with recall by modality set.

PAGE 49

38 Five additional trials were given and the subject was told to recall whichever set first that he or she wanted. Performance on these five trials served as the modality preference criteria. The subject dictated the numbers recalled and the experimenter recorded the responses. The number of visual correct responses were subtracted from the number of auditory correct responses and a + 3 value served to classify the child as having either a visual or auditory preference. This session was followed by the missing units trials. The different words (CVC trigrams) were presented simultaneously in the auditory and visual modes. The subject received trials until he achieved six correct auditory or six correct visual responses in any group of ten trials. If the subject achieved the six correct responses in the mode corresponding to his classification on the digit span task, then he was said to demonstrate a significant aptitude treatment relationship. The next series of trials was used for reinforcement and the final series of trials comprised the return to baseline condition. The subject saw both sets of words less one, on a stimulus sheet after each trial (see Appendix C). In order to avoid problems with motor variables, the subject was required to respond orally with the missing word. The experimenter recorded the responses. A detailed summary of the experimental procedure is as follows: 1. The student was screened in order to determine if he or she met the criteria for word and number recognition and for educationally handicapped classification. 2. Subjects meeting the requirements for the experimental

PAGE 50

39 population were given ten bisensory digit span trials. The subject was told from which modality to recall. Verbatim directions are provided in Appendix F. 3. The subject was given the five criterion trials (see Appendix F). If he met the criterion for having a modality preference, he proceeded to the next task. 4. The subject engaged in bisensory missing units trials (see Appendix F). If he did not reach criterion on this task he was dismissed. If criterion was reached, the subject proceeded to the next series of trials. In these trials the experimenter announced that reinforcement would be given for correct recall of the missing units. The criteria for reinforcement was shifted from the non-preferred to the preferred modality until the experimenter could demonstrate control over responding in both. The subject was not told when the criterion shifted, thereby permitting an assessment of the contingencies alone as an independent variable, 5. After completion of the reinforcement trials, the subject received the third and final series of missing units trials. For this series the experimenter announced that reinforcement would no longer be given for correct recall of missing units. In summary, seven children, in addition to the 11 children necessary for the digit span reliability procedure, were screened. The pilot subject came from the group of 11 reliability subjects. Three other subjects, who were qualified for the study, were found in the additional group of seven. One subject was dismissed because of a procedural error related to verbal commands given by

PAGE 51

40 the experimenter. The two subjects who remained were employed in the present study. Therefore, 18 children were screened, four were found to be qualified, and two were used for the actual study. Experimental Desi gn The design is a single subject reversal type (Baer, Wolf, & Risley, 1968). The first phase was the baseline of bisensory missing units trials. The second phase was the series of missing units trials on a multielement baseline (Sidman, 1960) that consisted of two concurrent discriminated operants, each on a continuous reinforcement (CRF) and extinction multiple schedule. The experimenter systematically applied the reinforcement contingency to responses requiring either auditory or visual attending. This resulted in the visually discriminated operant corning under CRF, while the auditorily discriminated operant came under extinction and vice versa. If the subject was classified as having an auditory preference then the first condition in the second phase consisted of reinforcement for visual responses. If the subject was classified as having a visual preference then the first condition in the second phase consisted of reinforcement for auditory responses. The conditions in the second phase were alternated until the return to baseline phase. The multielement baseline controls for intertrial difficulty and practice effects by distributing the requirement for responses to any particular modality across a large number of trials. The

PAGE 52

multielement baseline also provides the opportunity for repeated demonstrations of experimental control. Depe ndent Variabl es Two classes of dependent variable were necessary for this study. The first class was the type and number of responses, i.e., number of correct words recalled from each channel on the missing units task and number of incorrect responses. Because an incorrect response could not be assigned to either a visual or auditory mode, incorrect responses were not used to determine accuracy of responding. They served, however, as an indicator of difficulty. A second class of variable, which was necessary for a finegrain analysis of the experimental conditions was response latency (Alba, 1975; Ferster & Skinner, 1957). These latencies were recorded for visual, auditory, and incorrect responses. Response latency was defined as the time from offset of the missing units task presentation to the onset of the subject's response. Data Collection The data were collected during experimental periods approximately 30 to 45 minutes in length. The subjects' responses were recorded on the experimenter's record sheet for the digit span tasks and the missing units tasks (see Appendix G). Additionally, each session was recorded on audio tape in order to permit measurement of response latencies and checks on reliability.

PAGE 53

4? The latencies, while somewhat difficult to record during the actual experimental sessions, were obtained from the tapes not only by the reliability observer, but also by the experimenter. The latencies were obtained in this fashion, because the time required to time and record the latencies would have slowed down the experimental sessions and limited the number of trials that could be given during any particular session. Additionally, the overt timing may have produced reactivity on the part of the subjects. The experimenter and reliability observer listened to the tapes and obtained response latencies by starting and stopping a 1/10 second stopwatch at the offset of the bisensory presentation equipment and at the onset of a response. The point of offset was indicated on the tape by the distinct noise occurring when the experimenter operated controls that stopped the equipment. Listeners stopped timing when the subject made an audible response. Latencies were recorded on the experimenter's record sheet. Data Analysis The number of auditory, visual, and incorrect responses was plotted using, for each subject, a cumulative graphic display. The cumulative records were subjected to visual inspection as the primary analytic method. The trends in the data from phase to phase and from condition to condition within a phase were used to evaluate the reinforcement effect. The percent of visual, auditory, and incorrect responses in any particular phase or condition was

PAGE 54

43 computed in order to yield a simple quantification of the subjects response. This information is used to supplement the data in the cumulative records. The latency data were used to further compare and contrast differences between correct and incorrect responses across the various phases of the experiment. These data are displayed as distributions on a logarithmic scale. The logarithmic scale aids in interpretation because it assists in normalizing the display of the data. Interobserver Reliability The reliability procedure chosen for this study was the relatively stringent point-by-point agreement method (Kelly, 1977). It is possible to use the point-by-point method with trial and latency data by recording latency values and codes for various responses within the experimental conditions. In the present study any trial could be coded as having produced a visual (V), auditory (A), incorrect response (I), experimenter error (EE), or equipment malfunction (EM) (see Appendix E). In order to permit proper coding, the reliability observer was given a copy of the correct auditory and visual answers for any set of trials to be evaluated. The observer listened to audio tapes of the various experimental conditions and recorded response codes accordingly. After the response codes of the reliability observer were recorded, they could be compared to the corresponding experimenter

PAGE 55

44 records. The point-by-point agreement was computed by dividing the number of specific notations of the observers' records agreed on by the total number they both recorded plus the number of notations only one or the other recorded (Kelly, 1977). The latency data were checked for reliability in the same manner as the trial notation data. The point-by-point computational procedure was the same. Rather than computing the number of notations agreed upon, the number of latency values agreed upon becomes the unit for reliability evaluation. When determining agreement between two observed latency values, a t .5 second observer error was permitted.

PAGE 56

CHAPTER IV RESULTS The measures used for analysis of the experimental results are the cumulative number of responses and response latencies. The three types of responses were auditory (A), visual (V), and incorrect or error (I). For each type of response, the durations of the response latencies were measured. In the following sections, both the number of responses and response latencies are analyzed across phases, beginning with the baseline phase (B]) and proceeding through subsequent phases in which reinforcement was contingent on auditory or visual responses. The subscript indicates the particular order and position a baseline or reinforcement phase held within the experiment. For example, phase V 2 indicates that reinforcement was provided for visual responses and that this was the second such reinforcement phase for the subject. The missing units task raw data, indicating trial numbers, phase changes, and day changes are found in Appendix E. The data were collected according to the following schedule: 45

PAGE 57

46 Trials, Phases, and Day Sequence Per Child Subject 1 Trial Number Phase Day Number 1

PAGE 58

47 Table 2 Percentage Interobserver Agreement for Dependent Variables Across Experimental Conditions Subject 1 Phase B] V] A] V 2 A 2 V 3 B 2 Total Response Type 72.7 76.4 100.0 91.6 NA* 83.3 87.5 87.3 Response Latency 81.8 82.3 78.2 83.3 NA* 75.0 81.2 80.5 Subject 2 Phase B-| A] V ] A 2 V 2 A 3 B 2 Total Response Type 92.5 100.0 93.7 96.2 100.0 100.0 100.0 98.3 Response Latency 92.5 76.4 81.2 92.3 95.4 89.4 100.0 90.0 *Not available Analysis of Cumu lativ e Responses The responses are displayed by the use of cumulative records (Ferster & Skinner, 1957; Skinner, 1968). The cumulative record was chosen as the display method, because it offers a potent technique for establishing functional relationships between environmental variables and behavior (Alba, 1975). Consequently, for each experimental condition, the type of response is graphed cumulatively by trials. The actual calendar day is noted on the upper part of the auditory grid. Broken vertical lines indicate the trial preceding a day change and the solid vertical lines indicate the trial preceding a phase change. The cumulative graphs are found in Appendix H.

PAGE 59

During the course of the experiment, there were occasional experimenter errors and equipment malfunctions. In order to clearly represent every trial outcome, these errors are noted by a break in the cumulative record line at the point that the error or malfunction occurred. It should be noted at the outset that the steepest slopes possible for the cumulative displays of this experiment are approximately 45 degrees. A 45 degree slope would represent the occurrence of one response per trial for any given response type. Because of the design of the experimental trials, only one response of any particular type is possible. Slopes of less than 45 degrees indicate a lower number of responses across trials and horizontal cumulative record lines indicate no responding across trials. Additionally, after 20 cumulative responses occur, the cumulative record line resets and starts again on the zero line. The continuous record from trial to trial clearly illustrates steady state responding, transition states, and any local fluctuations in the general trend of the data. A nalysis of Baseline 1 An analysis of the performance of Subjects 1 and 2 during the B-] phase (see Appendix H-l and H-4) yields a major difference in type of response. Subject 1 strongly responds with auditory responses, while Subject 2 responds with visual responses. These responses reflect their original modality preference classification, There appears to be a steeper cumulative response curve for Subject l's auditory responding than for Subject 2's visual responding.

PAGE 60

49 While responses to non-preferred presentation modalities also differ as to total number between subjects, the differences in incorrect responding for each subject can likewise be appreciated. Subject 2 makes incorrect responses approximately two times as often as Subject 1. This high error rate probably contributed to Subject 2's extremely low number of responses to the auditory presentation, her non-preferred presentation modality. Furthermore, because of the high number of errors, almost twice as many trials were necessary in order to gain a relatively stable baseline before beginning the reinforcement phases. Analysis of Reinf o rcement Phase The first reinforcement phase (V]) for Subject 1 consisted of continuous reinforcement for any response that was under the stimulus control of the visual portion of the bisensory presentation. This was Subject l's non-preferred modality. The first reinforcement phase for Subject 2 (A]) was arranged in a similar fashion, i.e., reinforcement for the non-preferred modality. Subject 1, possibly because of a high rate of auditory responding during baseline (B]), persisted in high auditory responding for approximately 50 trials before the contingency began to exert control over the visual responses. In contrast, the reinforcement contingency in Phase A-j appeared to influence Subject 2's responding after approximately 14 trials. In order to further clarify the subjects' response patterns, percentage measures of each subject's responses are provided in Table 3. The number of responses in a phase for each of the three

PAGE 61

50 Table 3 Percent of Responses Across Trials by Response Type and Experimental Phase Subject 1 Phases Q ] y } A] V 2 A 2 V 3 B 2 Auditory* 61.0 40.5 68.1 11.5 66.6 16.6 64.0 Visual* 17.9 35.1 18.2 61.5 16.6 54.2 20.0 Incorrect* 35.8 29.7 22.7 30.7 16.6 33.3 20.0 Subject 2 Phases B ] A-| Vi A 2 V 2 A 3 B 2 Auditory* 07.0 45.9 43.5 57.1 33.3 58.8 47.8 Visual* 29.5 08.0 24.2 33.3 61.9 23.5 26.6 Incorrect* 60.6 37.8 27.1 19.0 09.5 26.4 23.6 ^Subjects occasionally would give both a correct auditory and visual response, and subjects would periodically give no scorable response. Because of these responses, sums of auditory, visual, and incorrect responses, if computed, give values somewhat above or below 100% depending on the particular phase.

PAGE 62

51 response types was divided by the number of trials in that phase. This measure yields an index of response proportion. When inspecting incorrect response percentages, it becomes clear that Subject 2's A-j responses took place at 37.8%, about half of the B^ amount of 60.6%. This reduction of errors is related to the relatively large, sixfold, increase in auditory responses. The increases in non-preferred responding from the B] to V] phase for Subject 1 and from the B-j to A-| phase for Subject 2 argue convincingly for the existence of a functional relationship between the reinforcement contingency and bisensory missing units task performance. It must be kept in mind that the auditory V-j value for Subject 1 is relatively high because of the large number of trials that elapsed before the contingency had an effect on visual responding. From trial 100 to 150, however, the rapid changeover can be appreciated easily. Returning to the cumulative displays, it is interesting to note, during the V] phase, Subject 2 persisted for a large number of trials in auditory responding. Around trial 228 the visual responding began to increase. For Subject 2 the large number of trials necessary to gain control of the visual responding and to note a reduction in auditory responding contributed to the high V] auditory value found in Table 3. Weiner (1970) discusses the existence of behavioral persistance in human subjects. He has attributed this behavior to either organisms with limited response repertoires or to uncontrolled variables affecting performances. Because Subject 2 had both

PAGE 63

52 auditory and visual responses in her repertoire, the possibility exists that during the V-j phase an uncontrolled variable was operating. In either event, responding did come under control dependably from trial 228 through trial 252 and, in those 24 trials, 75% of the responses were visual. Except for a slow changeover in condition V-j , experimental control was demonstrated consistently for Subject 1 in phases V 2 , A 2 , and V3. The trend in responding was quickly brought under control by the reinforcement contingency, and consequently fewer trials were needed to obtain stability in these phases. Subject 2 performed in a similar manner but with some variation. Upon close inspection of phase A 2 and V 2 it can be seen that the effects from a previous phase carried over as much as 14 trials in the following phase, as illustrated by continued visual responding from phase V] to A 2 . For Subject 1, there were virtually no occurrences of behavioral persistence in the latter phases (A], V 2 , A 2 , and V3) for either auditory or visual responses. • Analysis of Incorr e ct Respondin g Incorrect responding served as an indicator of difficulty of the material in any particular phase. As previously mentioned, the number of incorrect responses cannot be computed with the number of correct responses to provide measures of accuracy, because of the difficulty in reliably determining whether any particular incorrect response was visual or auditory in nature. An interesting pattern, however, appears from phases A 2 through B 2 in Subject 2, and from phase V-| through B 2 in Subject 1. For the phases noted,

PAGE 64

53 when reinforcement was applied to performances in non-preferred learning modalities, the incorrect responses increased (see Table 3). When reinforcement was applied to performances in preferred learning modalities, then incorrect responses decreased. It must be kept in mind that this pattern is related to difficulty but not accuracy of responding. Furthermore, there were no contingencies placed on incorrect responses in any phase of the experiment. Anal ysis of Ba seline 2 The final return to baseline (B 2 ) phase for each subject yielded contrasting results. Subject 2 initially gave a brief run of visual responses from approximately trial 65 to trial 80, while auditory responding decreased during the same period. After this set of trials, however. Subject 2 engaged in continued auditory responding. As noted in Table 3, error percentages during B 2 responding were not especially high. In fact, errors occurred two thirds less often than during the B-, phase. Subject 1 quickly regained the original response curves observed during the B] phase. Briefly returning to the pilot study, the pilot subject, classified like Subject 1 as having an auditory preference, demonstrated a similar recovery of baseline responding. During the pilot study, however, the recovery of the auditory preference occurred after 50 trials of baseline. In other words, the data would have appeared to recover the original B-) visual stability; when in actuality, a transition or local fluctuation was in effect (see Appendix H-7, H-8).

PAGE 65

54 Regarding the B 2 phase, it is necessary to keep in mind that the multielement baseline design, by permitting the repeated demonstration of experimental control, usually diminishes the need for returning to a lengthy baseline condition. Because only one more affirmation of the consequent (Sidman, 1960), showing that the independent variable was responsible for the change in responding, is gained, the additional expenditure of time may not always be justifiable. When, after numerous experimental manipulations, irreversability is encountered, Sidman (1960) points out: If non recoverability is indeed a fact, intrasubject replication is not possible. I have brought up the consideration, however, that the irreversabi li ty may be elsewhere than in the organism or in his behavior. ... The extinction operation, as it is normally carried out, only destroys some of the relations that were established during original acquisition, (pp. 101-102) Because of numerous experimental manipulations of Subject 2's baseline performance, the possibility exists that the irreversabi lity encountered is a function of environmental variables and not organismic ones. If this is true, then the baseline condition could, no doubt, be recovered had further experimental manipulation been attempted. In the main, the three replications (V-,, V2, V3) with Subject 1, and the three replications (A-| , A 2 , A 3 ) with Subject 2 argue convincingly for the existence of a functional relationship between the contingencies of reinforcement and both subjects' performances on the bisensory missing units task. With few exceptions, experimental control was demonstrated across both subjects in a relatively rapid and predictable fashion.

PAGE 66

55 Analysis of Resp onse L atencies Response latencies were graphed as distributions on semilogarithmic scales. Latencies are displayed for each type of response (A, V, and I) by phase. These displays are provided in Appendix I. Means of the latency distributions are provided in Table 4. Response latencies did not prove to be greatly responsive to the experimental manipulations. Consequently, few reliable generalizations can be made from the latency data. There are, however, a number of trends that became apparent upon close inspection of the data. For most conditions, the duration of latencies associated with incorrect or error responses was uniformly high across both subjects. The only example of an auditory or visual latency higher than incorrect latencies occurred once during Subject 2's visual B] performance (see Appendix 1-4). With that exception, Subject 2's incorrect latencies were consistently higher than any latencies obtained for auditory or visual responses. High latencies associated with incorrect responses have also been noted by Alba (1976) in work involving the use of time parameters for investigation of academic responding. In the majority of measures a relatively high auditory or visual response latency occurred during reinforcement conditions. Overall means for baselines and reinforcement conditions are noted in Table 4. The visual overall mean for Subject 1 provides the only overall measure wherein the baseline mean is greater than the

PAGE 67

56

PAGE 68

57 reinforcement mean. It should be noted, however, that these two measures are very similar in value. Relative to means for individual conditions, those means representing latency measures during reinforcement conditions exceeded baseline measures in 18 out of 26 conditions (70%). The shift in latency duration also carried over into the B 2 condition for both subjects. The carry over was more noticeable for Subject 2 than for Subject 1. For Subject 2, the clarity of the greater latencies in the B 2 phase may also be a reflection of larger amounts of data gathered in this condition. Regarding individual latency values, the exception to higher latencies during reinforcement was again the one outlying latency in the visual B ] condition for Subject 2 (see Appendix 1-4). Relative to individual incorrect latency values, Subject 2's B 2 latency durations exceeded those of the B-| condition (see Appendix 1-4 and 1-7). The longer individual latencies found under reinforcement were not found for incorrect responses in Subject 1 (see Appendix 1-1 and 1-3). The data might be better understood when one considers a possible source of short latencies in general. It was noted during the baseline trials that Subject 2 engaged in frequent guessing, to the extent that this particular subject might give a response before the stimulus sheet cover was moved to expose the appropriate stimulus words for the trial. As the experiment proceeded this behavior decreased, no doubt, because of its limited reinforcement. The final product, however, of guessing behavior was incorrect responses and short latencies. As Subject 2 was exposed to various reinforcement conditions, guessing stopped and the duration of latencies increased.

PAGE 69

58 In attempting to understand greater latencies, when careful inspection of the latency data during reinforcement is made, it is possible to note a trend for latency durations to cluster around greater time values. The clustering is noticeable for the type of response being reinforced. In other words, with two exceptions, visual latencies were at least as long if not longer than auditory latencies under visual reinforcement conditions. The V 3 phase for Subject 1 resulted in no auditory responses thereby eliminating this phase from the above comparisons. The trend to longer latencies under various reinforcement conditions might be indicative of the more careful attempt on the part of the subject to respond correctly. The experimenter would occasionally hear both subjects employ a rehearsal strategy. This strategy consisted of audible stimulus item repetition taking place while the subject looked at the stimulus sheet. The process of stimulus item repetition, while determining the missing unit, probably required somewhat more time than other non audible strategies. In either case, it became apparent that both subjects at some points in the experiment would use audible verbal rehearsal strategies requiring more time between the offset of the bisensory presentation and the onset of their responses. An additional subtle relationship was noted in Subject Ts data. At one point, the experimenter ran out of reinforcers for Subject 1. The problem occurred during the V 2 phase. At this time the experimenter substituted (as the medium for reinforcement) check marks on a card rather than nickels. The card was to be traded in for cash during the following session. Although it is

PAGE 70

59 difficult to make an experimentally verifiable statement, this phase produced, as a whole, somewhat longer latency durations for all three categories of responding (see Appendix 1-2). Further replications, however, would be needed to clarify any reliable functional relation between latency and change in the reinforcing stimulus. In sum, the most convincing data that a functional relationship exists between the subject's performance and the reinforcement contingency are found by visual inspection of the cumulative response data. The latency data, while serving to clarify subtle characteristics of the responses, are secondary to the large effects and numerous direct replications (Sidman, 1960) of experimental control over the subjects short term memory performance on the bi sensory missing units tasks.

PAGE 71

CHAPTER V DISCUSSION There has been a great deal of interest in modality preference as a learner characteristic and its relation to learner performance. To the extent that it is difficult to determine consistent auditory or visual modality preferences, some investigators (Ingersoll & DiVesta, 1972), nevertheless, claim that aptitude treatment interactions between modality preference and presentation mode indicate preferences which are stable across tasks. Additionally, they claim that these preferences control the subject's performances on a short term memory task. The present study employed operant techniques to further clarify the variables that act to control alleged modality dependent performances with mildly educationally handicapped children. A positive reinforcer proved to control the modality related performances of these children. This demonstration provided an illustration of the environmental variables contributing to the consistency of modality preference, and subsequently provides an explanation for a subject's performance at a functional level, rather than at a nominal level. 60

PAGE 72

61 F indings The findings of this study are discussed relative to the experimental questions. The questions are listed and a brief discussion follows. 1. Is there a functional relationship between contingencies of reinforcement and short term memory performance, allegedly under control of a visual modality preference? The data drawn from Subject 2, who was classified as having a visual preference, indicate that the reinforcement effect, while not always similar in pattern from phase to phase, was present. Visual inspection of the cumulative response curves demonstrates, especially at phases A], A 2 , V 2 , and A 3 , consistent control over the type of responding. Data from Table 3 indicate that contingencies, when placed on auditory responses, increased percent response values over those obtained during both the visual baselines. Except for phase V 2 , auditory values, obtained when reinforcement for visual responding was in effect, exceeded those values obtained for visual responding. In other words, the application of reinforcement increased performance in the non-preferred modality to the extent that it exceeded performances in the preferred modality with one exception. Increased performance is visible for auditory responses in phase A-| and A3, when compared to visual responses in phases B] , B 2 , and VlVisual latencies for Subject 2, who exhibited a visual preference, were not greatly sensitive to reinforcement when

PAGE 73

62 applied to specific classes of response. In other words, when auditory responses were reinforced there were no easily discernible shifts in the auditory latency distributions. Thus, for the present study these measures are equivocal in there sensitivity to the independent variable. 2. Is there a functional relationship between contingencies of reinforcement and short term memory performance allegedly under control of an auditory modality preference? Visual inspection of the trends in the cumulative records indicates relatively reliable control over the direction of the response curves. The data gathered on Subject 1 indicated that he exhibited an auditory preference. When reinforcement contingencies were placed on visual responding, the percent value not only doubled its B] value (see Table 3), but also, in two out of three visual reinforcement phases (V2 and V 3 ), exceeded the B 2 auditory condition. Further, the visual responding under visual reinforcement exceeded or approached, within 6.8%, the B-j auditory responding, in phases V 2 and V3. As with Subject 2, Subject 1 when under reinforcement for the non-preferred class of response, did not demonstrate changes of a noticeable magnitude on the latency measures for the nonpreferred response. Therefore, the usefulness of these measures for addressing the second experimental question is limited. Supplementary information, not directly related to the experimental questions, bears some discussion. First, although the

PAGE 74

63 latency measures did not prove especially sensitive to the specific experimental manipulations per se, the latency data did seem responsive to reinforcement conditions in general. Second, latency measures did provide data useful for a topographical analysis of the responses. As mentioned earlier, subjects' responses that were characterized best as guesses exhibited extremely short latency durations. Longer response latencies, for correct responding, appeared to take place on occasions characterized by a subject's use of a vocal or almost subvocal rehearsal strategy. Incorrect responses were characterized by latencies frequently approaching or exceeding a ten second value. For correct responses, time values exceeding ten seconds only occurred twice out of 13 possible phases in which the latency datum was kept. Third, that the incorrect responding was differentially sensitive to experimental phases is suggested by the data in Table 3. Because incorrect responding cannot be clearly identified with any response type, it is of little use in discussing the subjects' accuracy. It does, however, provide an indicator of differential difficulty for the subjects. Interpretation of the Findings That a functional relationship between the type of responses on the missing units task and contingencies of reinforcement exists is clear. Interpretation of the data at a molecular level was provided in Chapter IV. At a molar level, interpretations of the experimental findings, relative to Ingersoll and DiVesta's (1972)

PAGE 75

64 suppositions of modality preference control and stability, are necessary. Ingersoll and DiVesta (1972) observed their subjects' performances on a bisensory task and then observed similar performances on another set of bisensory tasks. From this set of observations they proceeded to suggest a causal relationship between preference and performance. Realistically, the relationship observed could be categorized as correlational. To say one set of behaviors indicates control over another is to make a statement needing a strong experimental foundation. First and foremost, the repeated demonstrations of a relatively weak reinforcer as an independent variable capable of controlling which of two discriminated operants a subject would use most frequently, certainly places the question of control in a functional perspective, rather than a correlational one. The numerous experimental manipulations using an independent variable under control of the experimenter would appear to be a more viable method to approach the questions as to what, in fact, might control a subject's bisensory short term memory performance. Bearing on modality preference stability, the inference made by Ingersoll and DiVesta (1972) that the observed aptitude treatment interaction in their study implied stability or consistency across tasks is certainly reasonable. Stability from one task to another, however, implies stability through time. The point is of moment, because Ingersoll and DiVesta performed their tasks in sequence. The modality preference identification task was performed first and the missing units task second. The data in the present experi-

PAGE 76

65 merit give support for relative flexibility of response direction through time. Additionally, Subject 2 did not recover the original baseline responding she demonstrated earlier on each modality, but showed a pattern of alternating between both. Whether or not responding could be recovered is not as important as the fact that it did change and remained so after experimental manipulations ceased. This is even more interesting in light of Subject 2's performance on the bisensory digit span task. Subject 2 scored the greatest value (-12) in the visual direction of any child screened for the study. If the visual modality preference was truly stable or consistent across these tasks, then it seems likely that the baseline condition would have been recovered, at least partially. It must be kept in mind that Ingersoll and DiVesta (1972) concluded that extreme digit span scores would be associated with clear response patterns on the bisensory missing units task. The validity of this proposition may be restricted when one considers the data of the present experiment. Smaller incorrect response magnitudes did appear in phases under reinforcement for responses similar to the original modality preference classification; nonetheless, there is little reason to believe that experimental control of incorrect responding in any phase could not be achieved given the appropriate experimental controls. In sum, regarding control of alleged modality dependent performances and their stability across tasks, it appears that the source of control for a subject's performance is more likely to

PAGE 77

66 be found in specifiable, measurable, and for the most part, controllable variables such as positive reinforcement, in the immediate learning environment. Resorting to explanations based on data essentially correlational in nature has very limited utility in uncovering the variables responsible for behavior. Stability across tasks is ^ery likely to be a function of these same variables acting in concert with time parameters. Problems and Limita tions of the Study The foremost limitation of any study performed in the area of modality preference is the instrumentation and criteria used for modality preference identification. The second pressing problem is the provision of a treatment that actually possesses the components for classification as exclusively auditory or visual in nature. Relatively non-complex tasks were chosen to minimize the threats of both foregoing limitations. The use of non-complex tasks may have limited generality to the domain of tasks that are primarily the focus of modality based treatment, i.e., reading. Short term memory, on the other hand, has been suggested as a skill necessary for reading (Senf, 1969). Consequently, the tasks used for the modality treatment in this study may be components of more complex behavior. A problem that deserves some discussion, because it is noticeable with Subject 2, concerns equipment malfunctions. The precision of the equipment was limited. Changes in available line voltage and even changes in temperature would periodically effect the

PAGE 78

67 performance of the audio-cassette sound relay that controlled the visual portion of the bisensory presentation. The effect of the occasional interruptions is not known. Fortunately, the greatest proportion of trials took place without interruption. This problem could be avoided in future studies by the use of more sophisticated programming equipment. The second problem was the behavioral persistence encountered with Subject 2. The possibility exists that an uncontrolled variable (Weiner, 1970), such as subtle cues by the experimenter, accidental reinforcement, or a practice effect, contributed to the observed persistence. Additionally, in any study there is a chance that the subject reaches a state of reinforcer satiation or that the reinforcer itself may have been identified incorrectly. In any operant study of reinforcement the only control for these possibilities is inspection of the data. Control was achieved eventually in all experimental conditions, but the possibility that a loss of reinforcer effectiveness or, on the other hand, a competing reinforcer during certain trials remains. In either event, it is worthy to note that the persistence took place in the subject's auditory response, her non-preferred modality. This fact subsequently limits the value of an auditory preference as a plausible alternative explanation. Practical Implication s A number of outcomes in this study have practical implications for those involved in research or applied work. First,

PAGE 79

methodologically, the value of a dynamic or continuous measurement strategy, an integral component of operant technology, becomes evident. The continuous measurement through time employed in this study permitted the investigator to not only note changes in the data that were necessary for making experimental decisions, but also note changes in responding that occasionally might be at variance with the subject's original preferred modality classification. The continuous measurement process could be crucial in programming the most effective modality treatment, had the study been concerned with making an educational intervention based on the subject's modality use. Second, the direct measurement strategy employed in operant research deserves comment. The data in the present experiment did not undergo a statistical averaging in preparation for tests of significance nor did the primary measurement units evolve from any statistical summarization of the data. For the most part the data was interpreted in units that were compatible with the continuous measurement strategy. In other words, direct measurement procedures are interpreted in a display that reflects each trial outcome and the impact of time as a parameter. For example, had B-] cumulative data been everaged with B 2 cumulative data in Subject 2, conclusions about her modality performance across tasks would have been entirely different from those originally presented. Third, for the practitioner of modality based teaching interventions, the data of this study tentatively indicate that modality use in various presentation modes is not undivided. That by using

PAGE 80

69 the proper environmental arrangements performances in an alleged non-preferred modality can approach or in some cases exceed the performance in a preferred modality. This is not to say that performance in non-preferred modes can consistently be equated with performances in the preferred, but the possibility for approximate performances exists. The error patterns in both subjects showed differential sensitivity to the experimental conditions. In order to increase performance in a non-preferred modality there may be some temporary trade off in terms of error rates, but not necessarily in terms of accuracy. In the response latency measures there were no examples of differential error sensitivity. There is convincing evidence that to exclude a learner from material presented in a non-preferred mode would be a great injustice and possibly even an inefficient teaching tactic. It may, in fact, be a service to the learner to be switched to a nonpreferred modality. Subject 2 used her non-preferred modality frequently in the B 2 condition after exposure to the experimental conditions. Professionals have assumed that the preferred modality is in the service of the learner while at the same time attributing the preference to sensory and other constitutional traits. It is entirely possible that an individual's observed modality use is a function of his or her past reinforcement history. In this sense a modality preference could be superstitious in origin. In other words, the modality use frequency that one observes in an individual could have come about by accidental reinforcement of

PAGE 81

70 learning in that particular modality. It must be kept in mind that reinforcer control of modality use was demonstrated repeatedly in the present experiment. And, if one considers the term preference in a functional perspective it actually connotes a high probability or frequency of an activity on the part of the individual who exhibits the preference. Thus, modality use frequency is the defining characteristic of a modality preference and reinforcement influences the frequency of modality use. In sum, a learner may perform better in a supposedly non-preferred learning modality than expected, if the consequences for that performance are reinforcing and if the consequences are arranged systematical ly. Suggestions for Future Resea rch The foregoing research, in many ways, stimulates an unlimited variety of questions related to modality preference research. Much research is needed before one can fully understand the relationship of reinforcement to modality related performances. The first requirement for any science is that of replication, for it is the soundest empirical test of reliability (Sidman, 1960). Operant procedures lend themselves well to replication. Because of the reliance on experimental control, the great number of observations on relatively few subjects, the strategy of holding many variables constant, while manipulating only one, and because of the technologically exact descriptions of the procedure, replication is possible.

PAGE 82

71 Systematic replication (Sidman, I960) of the present experiment should give a fuller understanding of modality use and related performances. First, replications with children presenting various behavioral repertoires, such as different baseline modality use and error response patterns, is of the utmost importance for understanding the generality of the present results. There is a need to determine, with other children, not only the generality of the functional relations noted in this study, but also the general usefulness of the methods employed. Second, the complexity level of the task may interact with the subjects' modality use. As noted with Subject 2, incorrect responding took place at a high rate in phase B 1 when compared to Subject l's responding. Whether or not Subject 2's incorrect responding during baseline contributed to her behavior in the experimental condition is certainly a valid question that can be addressed by investigating various levels of complexity in task arrangements or in stimulus materials. Third, it is assumed in modality investigations that subjects naturally provide responses associated with that presentation modality that is the most effective for them. In the present study contingencies of reinforcement proved to influence the type of response that a subject gave. It may be of value to investigate variables such as various verbal directions to the subject. It is possible that directions can effectively control performance to the extent that the subjects' optimum performance in a modality can be obtained by instructing them to attend to that modality.

PAGE 83

72 In the same manner, instructing a subject in various rehearsal techniques may also optimize their learning in an infrequently used modality while also increasing the chances for increased use of that modality. Fourth, the schedule of reinforcement in the present study was continuous. The area of schedule effects on modality use should be systematically investigated. Not only the schedule effects of reinforcing stimuli, but also the schedule effects of aversive stimuli could prove valuable as a research area. Schedule research with both types of stimuli could illuminate behavioral processes contributing to particular modality use patterns. Fifth, in the present study accuracy of responding was difficult to evaluate. Further studies of a similar nature should concentrate on the accuracy of a subject's response under varying reinforcement conditions. In order to carry out such research, techniques need to be developed to determine if an incorrect response is associated with auditory or with visual stimuli. By performing replications of the present experiment and by pursuing the suggested research, the knowledge base associated with modality preference can expand and ultimately become consumable by the population in greatest need of viable teaching strategies handicapped children.

PAGE 84

REFERENCES Alba, E. The experimental analysis of academic behavior. In L. E. Fraley & E. A. Vargas (Eds.), Proceedings of the Third Nation al Conference o n Behavi or Research an d Technology in Higher Education . Champaign, 111.: Research Press, 1975." Allyon, T., Layman, D. , & Kandel , II. J. A behavioral educational alternative to drug control of hyperactive children. Journal of Applied Beh avior An alysis, 1975, 8, 137-146. Baer, D. M. , Wolf, M. M., & Risley, T. R. Some current dimensions of applied behavior analysis. Journal of Applied Behavior Analysis, 1968, 1, 91-97. Bateman, B. The efficacy of an auditory and visual method of first grade reading instruction with auditory and visual learners. In H. K. Smith (Ed.), Perception and Reading. Newark, Del.: International Reading Association', 1968. Blanton, B. Modalities and reading. 210-212. Reading Teacher, 1971, 25, Bracht, G. H. Experimental factors related to aptitude-treatmentinteractions. Review of Educ ational Research, 1970, 40, 627-645. " "" " Bruin inks, R. H. The relationsh ip of audi tory and visual_p_erceptual stre ngths to me j^jo_d_s^f_tej^_nc j word re cognition among disad va ntaged Negro boys. Nashville, Tenn.: George Peabody College for Teachers, 1968. (ERIC Document Reproduction Service No. ED 043 721) Bruininks, R. H., & Clark, C. B. Auditory and visual paired associate learning in first grade retarded and non-retarded children. American Jou rnal of Mental Deficiency, 1972, 76, 561-567. " "" — Budoff, M., & Quinlan, D. Reading progress as related to efficiency of visual and aural learning in the primary grades. Journal of Edu cational Ps ychology, 1964, 55, 247-252.

PAGE 85

74 Clingman, J., & Fowler, R. L. The effects of primary reward on the I.Q. performance of grade-school children as a function of initial I.Q. level. Journal of Applied Behavior Analysis, 1976, 9, 19-23. J Conner, J. J., & Weiss, F. L. A brief discussion of the efficacy of raising standardized test scores by contingent reinforcement. Journal of App l ied Behavior A nalysis, 1974, 7, 351-352. Corsale, K. The effects of mode of presentation on encoding processes in children's short term memory . New Orleans, La.: American Psychological Association, "1974. (ERIC Document Reproduction Service No. ED 114 763) Cripe, A. G. Auditory and visual learning related to ITPA sensory channels (Doctoral dissertation, Purdue University, 1966). Dissert ation A bstracts InternationaJ_, 1966, 27, 635B. (University Microfilms No. 66-7407) Cronbach, L. J. .E^sjntj_al_s_of __p_s_y_ chol ogi cal testi n g . New York: Harper and Row, 1970. Cullinan, B. E., Ringler, T., & Smith, J. Pr eferr ed l earni ng m odalities and differentia ted presentation on rea di nq ta s_k_s . New York: New York University, 1969. (ERIC Document RepTo'duction Service No. ED 042 589) deHirsh, K. , Jansky, J. J., & Langford, W. S. Predi cti ng readi ng fail ure. New York: Harper, 1966. DiVesta, F. J. Trait-treatment interactions, cognitive processes, and research on communication media. Audio Visual Communic ations Review , 1975, 23, 185-196. Dornbush, R. L. Input variables in bisensory memory. Percep tion and Psychophy sics, 1968, 4, 41-44. Dunn, R. , & Dunn, K. Learning style as a criterion for placement in alternative programs. Phi Del ta Kap pan, 1974, 56, 275-278. Dwyer, F. 11. Visual learning: An analysis by sex and grade level. California Jo urnal of Educ ational Research, 1971, 22, 170-176. Estes, R. E., & Huizinga, R. J. A comparison of visual and auditory presentations of a paired associate learning task with learning disabled children. Jo urnal of Le arning Disabilities, 1974, 10, 33-42. ~ Estes, R. E., & Stewart, J. C. A^jcompjrjson of_visu al and auditory channels in l earning disabled an d control children . Washington, D.C.: American Educational Research Association, 1975. (ERIC Document Reproduction Service No. ED 116 420)

PAGE 86

75 Edinger, D. L. A free operant analysis of programmed instruction performance with reading disabled children (Doctoral dissertation, University of Florida, 1969). Di sser tation Abstracts Internationa l , 1970, 31(5), 2215-A. (University Microfilms No. 70-20, 749) Ferster, C. B., & Skinner, B. F. Sc hedules of rein forcement. New York: Appleton-Century-Crofts , 1957. Filmer, H. T., & Under, R. Comparison of auditory and visual modalities. Education, 1970, 91_, 110-113. Foster, G. G. , Reese, J. H., Schmidt, C. R. , & Ohrtman, W. F. Modality preference and the learning of sight words. Journal of Special Education, 1976, 10., 253-258. Galfo, A. J. Effects of certain audio and visual presentation sequences on pupil information acquisition. Journal of Educational Research , 1970, 69, 172-176. Goldiamond, I. Perception. In A. J. Bachrach (Ed.), Experimenta l foundations of clinical psycholo gy. New York: Basic Books," 1962. Ingersoll, G. M., & Di Vesta, F. J. Effects of modality preferences on performance on a bi -sensory missing units task. Jo urnal of Experiment al Psychol ogy, 1972, 93, 386-391. Jastak, J. F., & Jastak, S. R. The wide range ac hie vement test. Wilmington, Del.: Guidance Associates of Delaware, 1965. Jones, J. P. Lear ning modaliti es sh ould they be c onside red ? Atlanta: International Reading Association, 1972. ("ERIC Document Reproduction Service No. ED 052 899) Jones, J . P . Int ersenso ry trans fer, perceptual shi fting, modal preference and reading information series: Wh ere do we go? Newark, Del.: International Reading Association, 1972. (ERIC Document Reproduction Service No. ED 052 899) Kalin, M., & McAvoy, R. The influenc e of c hoice o n the ac quis ition a nd retention of learn ing materials in diff e rent modes of inst r uction . New Orleans, La.: American Educational Research Association, 1973. (ERIC Document Reproduction Service No. ED 094 381) Katz, P. A. Visual discrimination performance of disadvantaged children: Stimuli and subject variables. Child Dev elopment, 1967, 38, 233-242. Kazdin, A. E. The impact of applied behavior analysis on diverse areas of research. Jo urnal of Applied B eha vi o r Analys is, 1975, 8, 213-229.

PAGE 87

76 Kelly, M. B. A review of the observational data collection and reliability procedures reported in The Journal of Applied Behavior Analysis. Journal of Applied Behavior Analysis, 1977, j_0, 97-101. " " ~ Kirk, S. A. Educating exce ptional children. Boston: Houqhton Mifflin, 1962. Kirk, S. A., McCarthy, J. J., & Kirk, H. D. Ill inois_Jest^if psycholin g uistic abilities (experimental~ed~TUrbana: University of Illinois Press, 1961. Kubany, E. S., & Sloggett, B. B. The role of motivation in test performance and remediation. Jo urnal of Learning Disabilities 1971, 4, 426-428. Landers, M. F. The modality concept goes commercial Newslette r, 1976, 2, 53-56. DCLD Lilly, S. M., & Kelleher, J. Modality strengths and aptitudetreatment interaction. Journal of S pecial Education, 1973, 7, 5-13. Mann, L. E., Proger, B. B. , & Cross, L. H. Aptitude -trea tment i nteracti on with handicapped chi ldren: A focus_on_the measurement of the aptitude component . New Orleans, La.: American Educational Research Association, 1973. (ERIC Document Reproduction Service No. ED 075 510) flay, R. B., & Hutt, C. Modality and sex differences in recall and recognition memory. Child Develop ment, 1974, 45, 228-231. McConkey, R. , & Green, J. M. Presentation method and the free recall performance of retarded adults. American Journal of Mental Deficiency , 1973, 78, 95-97. —--..-Meeham, T. An informal modality inventory. Elementary English, 1974, 51_, 901-904. ' Newcomer, P. L., & Goodman, L. Effects of modality of instruction on the learning of meaningful and non-meaningful material by auditory and visual learners. Journal of Speci al Education, 1975, 9, 261-268. ~ Reynolds, R. J., Bickley, A. C. , Champion, S. , & Deckle, 0. Effects of mode of presentation of stimulus materials in word association tasks. P sycholog i cal Re ports, 1971, 28, 211-215. " " ' ~~ Sabatino, D. A., & Dorfman, N. Matching learning aptitude to two commercial reading programs. Exceptional Children, 1974, 41. 85-90. -. —

PAGE 88

77 Sabatino, D. A., Ysseldyke, J. E., & Woolston, J. Diagnosticprescriptive teaching utilizing perceptual strengths of mentally retarded children. American Jo urnal of Mental Deficiency, 1972, 78, 7-14. " ~~ Senf, G. Development of immediate memory for bi sensory stimuli in normal children and in children with learning disorders. Developmental Psych ology Monograph, 1969, t, 6 (pt. 2). Sidman, M. Tactics of scientific research . New York: Basic Books, 1960. Silverston, R. A., & Deichmann, J. W. Sense modality research and the acquisition of reading skills. Review o f Educational Research, 1975, 45, 149-171. ~~ ~~ "" Skinner, B. F. A case history in scientific method. In A. C. Catania (Ed.), Contemporary research in operant be havior. Glenview, CA.: Scott, Foresman, and Company, 1968. Skinner, B. F. Abo ut behaviorism . New York: Vintage Books, 1974. Slosson, R. L. Slo sson ora l re adin g test . New York: Slosson Educational Publications, 1963. Smith, C. M. The relationship of reading method and reading achievement to ITPA sensory modalities. Journal o f Special Education , 1971, 5, 143-149. ""'""" " ' Smith, D. D. , & Lovitt, T. C. The educational diagnosis and remediation of b and d written reversal problems: A case study. Journal of Learning Di sabilities , 1973, 6, 356-363. Thorndike, E. L. A teacher's word bo ok o f twenty thousand w ords. New York: Columbia Teachers College, 1931. Tyler, J. L. Modality preference and reading task performance among the mildly retarded. Training S chool Bulletin, 1974, 70, 208214. Ul rich, R., Stachnik, T., & Mabry, J. Control of human behavior. Glenview, 111.: Scott, Foresman, and Company, 1974. Waugh, R. P. Relationship between modality preference and performance. Exceptional Childre n, 1973, 39, 465-469. Weiner, H. Human behavioral persistance. The Psychologi cal Record, 1970, 20, 445-456. ' ' — Wepman, J. M. The modality concept. In H. K. Smith (Ed.), Perception and reading . Newark, Del.: International Reading Association, 1968.

PAGE 89

78 Wepman, J. M. The modality con ceptj Atlanta: l ckground and r esearch. International Reading Association, 197T ("ERIC Document Reproduction Service No. ED 053 864) Wepman, J. M., & Morency, A. S. Perceptu al development and learning : An expe rimental study on modality and reading instruction" Section II, f inal repo rt. Washington, D.C.: National Institute of Education (DHEW)7l975. (ERIC Document Reproduction Serivce No. ED 013 371) Whitehead, W. E., Renault, P. F., & Goldiamond, I. Modification of human gastric acid secretion with operant conditioning procedures. Journal of A ppl ied Behavior Analysi s, 1975, 8, 147-156. Williams, J., Williams, D. V., & Blumberg, E. L. Visual and aural learning in urban children. Journ al of Educational Psychology, 1973, 64, 353-359. Wolpert, E. B. Modality and reading: A perspective. Reading Teacher , 1971, 24, 640-643. Ysseldyke, J. E. Diagnostic-prescriptive teaching: The search for aptitude-treatment interactions. In L. Mann & D. A. Sabati no ( Eds . ) , The first review of special educa tion. Philadelphia: JSE Press, 1973.

PAGE 90

APPENDICES

PAGE 94

83 MISSING UNITS TASK STIMULUS SHEET TRIALS 29-56 Tria

PAGE 95

APPENDIX D DESCRIPTION OF TRIAL PROGRAMMING Missing Units Task There were basically 28 unique sets of trials, each containing three auditory and three visual stimuli. The slides necessary for the visual presentation were photographed on reverse black and white film in order to produce a white word with black background when projected. A 140 slide capacity Kodak slide tray was used. The slides were loaded into the tray in groups of three and a blank space was left between each group. This provision enabled the experimenter to see when the end of any particular trial occurred, because a large white light would appear on the projection surface. This light, resulting from the blank slot in the slide tray, served as a signal for the experimenter to stop the audio cassette which controlled the slide projector. At this point, the experimenter would move the cover sheet in order to expose the stimulus words associated with the auditory and visual stimuli that were just presented. The two stimulus sheets provided 56 sets of stimulus words. This number was accomplished by varying the missing words from one sheet to the next. Thus, the experimenter could perform 28 trials, with one stimulus sheet (trials 1-28), then reset the audio and visual equipment to start over and use a second stimulus sheet 84

PAGE 96

85 (trials 29-56) containing different missing words in each trial. Consequently, after 56 trials, the experimenter would start over with the first sheet (trials 1-28). By using 56 trials, each with a different missing unit, it became very difficult for the subject to memorize answers, even after repeated exposure to the basic 28 sets of auditory and visual stimuli. Digit Span Task Fifteen unique sets of digits were photographed on reverse film and placed in an 80 slide capacity Kodak slide tray. These were programmed simultaneously with 15 unique sets of four digits on audio tape. The first ten sets were used for the familiarization trials and the last five sets were used for the criterion trials. As in the bisensory missing units task, a space was left after each set of four slides in order to permit the experimenter time to stop the equipment and record the subject's response. The numbers contained in each set appear on the digit span record sheet.

PAGE 97

APPENDIX E RAW DATA

PAGE 98

I CD cC 4C_J> -t— rJD U T3 CD r— •<-) 87

PAGE 99

w >>

PAGE 100

CXJ , —

PAGE 101

90

PAGE 102

91 ID , — c

PAGE 103

9?

PAGE 104

93 >> -r(O 4-) E
PAGE 105

94

PAGE 106

95 >1

PAGE 107

96 CD in c o CL

PAGE 108

97 >^

PAGE 109

98 i — C\J

PAGE 110

99 i— co >^

PAGE 111

100 C\J i— ro , — — in

PAGE 112

10 i >>

PAGE 113

02

PAGE 114

103 en ,_ >> c r— **

PAGE 115

104

PAGE 116

105 ir> i —
PAGE 117

106

PAGE 118

107 LT> i —

PAGE 119

08 I— Lf)

PAGE 120

109

PAGE 121

10

PAGE 122

APPENDIX F VERBATIM DIRECTIONS FOR SUBJECTS Bisensory Digit Span Task Familiarization Tr ials Today we're going to do some things to see how you remember. I'm going to let you see a group of numbers and hear a different group at the same time. Here, I'll show you. (Experimenter runs one trial.) That was called a trial and we will do ten more. I want you to try to remember as many of the numbers that you see and hear as you can. I wi 11 ask you at the end of the trial to tell me all you remember. Now, we will do the trials, but I will tell you befo re each trial what group of numbers to tell me first when it comes time to answer. For example, I will say, "(child' s name), tell me the numbers that you saw first." Sometimes I will say, "tell me the numbers that you heard first." Remember to tell me as many numbers as you can, but always tell me the group I asked for first. Then you can tell me any other numbers that you can remember. Let's begin. Cri terion Trials Now, we will do some more trials, but this time I'm not going to tell you which group to say first. At the end of the trial, tell me as many numbers as you can remember in any order you want. Ill

PAGE 123

112 Missing U nits Task Now we are going to do another memory task. I'm going to show you some words on the screen and you will also hear some words at the same time. It will be like this. (Experimenter runs one missing units trial). When the trial is over, I will show you, on this sheet (experimenter places stimulus sheet in front of the child), two of the words you have just seen and two of the words you have just heard. One of the words from the group you saw and one of the words from the group you heard will be missing. So you have to try very hard to tell me the missing words. Try to remember both, if you can. Let's begin.

PAGE 124

APPENDIX G RECORD SHEETS

PAGE 125

MISSING UNITS RECORD SHEET TRIALS 1-28 V A Trial 1 jib dim 2 bet rib 3 bid fop 4 cud gap 5 box lap 6 tug fob 7 cob dog 8 ham put 9 bus gag 10 gat sin 11 bed cut 12 lid sad 13 bog pig 14 rob pip 15 bad hop 16 pet pit 1 7 mo t kid 18 pen hip 19 sup sag 20 way pap 21 hat hap 22 dub gut 23 jut set 24 jig hid 25 pod cot 26 hit tab 27 wag lax 28 yet day 114

PAGE 126

115 Trial 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 56 SSING UNITS RECORD

PAGE 127

116 DIGIT SPAN RECORD SHEET Name

PAGE 128

APPENDIX H CUMULATIVE GRAPHS

PAGE 129

rnrmrrrnr ill Hi n o 1 IM|iii:)iiirr-rrTTr ;l[!|iJ|M[jj|tf{i|'

PAGE 130

119 AUoiipny LBnsLA :pa.woDui S3SN0dS3y BAiivnnwno

PAGE 131

20 w

PAGE 132

121 r IVI CM s3SNOdS3cj HAiivinwno

PAGE 133

122 tftitt t • I PTR 1 1 , !:|jt|t Fj iiiiiiji ' -J-444-4. ! t J ill: tim m : t it .-.'44-14.— ,-*4-+— it Tit! jTrrni lilt i! I| J: !:!|i:tf!;!lj!r!:L n || I liltffTtp Ti i , i ; t nl ] ; :: ! I F ti I hr Li I.Li iihim;:. It! !iirriH:t tit hi fj rttttrtiij I : I , l I I tL i .-. 1 1 1 1, f-t-; ihthH j — o A"jo 4 Lpn\/ Lens la ^ +oa-uoDui S3SN0dS3y 3AiivinwnD

PAGE 134

123 mmm n T "Hirrrr r PF 7 "?T;ii;l; i! i !! :}!rr: t tiiil Hit! !]i! 1 i }TTtt fi i llTntTrr: t i in ! {' I i lilMiiLL itnTuMl i iii 1 1 mi rrtjr: (tiii iiii . 1 1 1 ,. m rrtTtnxlti^t.. i ii i lu f ,1 1 i.i ri. i_.j_. ';;;! |. i.'it -hrtt I I : 1 I I ] | I , I r t I | I J 1. 1 •' < j r!~rf n rj i ii *"f jnijijiifte: lMiiiiilil£tiJ = sBSNOdsny 3Aiivinwno

PAGE 135

124 i 'I itiTt nrffitHi I if I Iff L ill lirtm.'lttliy ihxr iiflftSS tf m-rmtt ntittf* /Uo^ipny LBnsiA 139.U0DUI S3SNOdS3y 3 A I 1 v i n W fl 3

PAGE 136

125 [rnirnniTniiii'i i t M • " H I t I 1 1 1 H . I Mr 1 ' ... . .... i h lit lit H-:i tnttr III I Pi "Wl iin III] hi lhJtr;.t f i 4— l_J L_ tfH ii ij \\vi\hu p... t,.,;ut„.. ;.;.: ItUtiiiiliifc . I 1 1 i s 1 1 i I ( , i . I , <3I ._ — litlill LensLA loajjooui S 3 S N d S 3 d 3 i\ I 1 V 1 fl W fl 3

PAGE 137

APPENDIX I GRAPHIC DISPLAYS OF RESPONSE LATENCIES

PAGE 138

. . h -44i I nttn M jili ! I 1 it -44 ij U44llili trt IS H~ 111 tt j tit til 111! u Xouanbajj 127

PAGE 139

123 III

PAGE 140

129 'I! "' H["| ii'in

PAGE 141

130 — t-o m m t |!|,|j ti'i rt"fl i iiiiii liiiiii miiiiiiiiriiin

PAGE 142

131 I "11 J ~ 1 1° CM _ O ill] iUKU 1 ' 1 Will T+T (ill vr u l : . 1 ! ! ftt!" IUJ4-4 t+— : : IU111J PIIII llll'l 1

PAGE 143

132 II 'illl'li'l Hill

PAGE 144

133

PAGE 145

BIOGRAPHICAL SKETCH Mark Alan Koorland was born at Cincinnati, Ohio, on F'larch 22, 1947. When he was 15, he and his family moved to St. Petersburg, Florida. He was graduated from Boca Ciega High School in 1965 and St. Petersburg Junior College in 1967. After graduating in special education from the University of Florida in 1970, he taught middle school educable mentally retarded children in Marion County, Florida. In 1972, upon completion of a Master of Education degree in special education with an emphasis in emotional disturbance, he worked as a consultant in the area of paraprofessional training for an NIMH drug abuse project operated at the University of Florida. In the Fall of 1972, he returned to the Marion County School system and took a position as a diagnostic-prescriptologist. The next year he taught in a resource room for emotionally disturbed elementary school children in Alachua County, Florida. Mr. Koorland entered the doctoral program in special education in the area of learning disabilities at the University of Florida in 1974. At the same time he took a position as an instructor with the Career Associate in Special Education (CASE) program at Santa Fe Community College. 134

PAGE 146

135 In 1975, Mr. Koorland held a summer appointment in special education at the University of South Alabama in Mobile. He returned to Santa Fe Community College in the fall and also resumed his doctoral studies. In 1976, he became coordinator of the CASE program. Mr. Koorland will complete the requirements for the degree of Doctor of Philosophy in December, 1977. He has accepted a position as assistant professor of special education in the Department of Psychoeducational Services at Florida International University in Miami, Florida.

PAGE 147

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. nriiam D. Wolking, Chairperson Professor of Special Education 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. /./'' Charles Forgnone Professor of Special Education 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. ^Lti £i~ Cecil D.' Mercer Associate Professor of Special Education 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. ) /$,? :',<£'.' ,'i, Henry S. Penhypacker Professor of Psychology

PAGE 148

UNIVERSITY OF FLORIDA 3 1262 08552 9492