Citation
The effect of delaying consequences upon the learning of emotionally disturbed adolescents

Material Information

Title:
The effect of delaying consequences upon the learning of emotionally disturbed adolescents
Creator:
Brown, Robert Keith, 1946-
Publication Date:
Copyright Date:
1973
Language:
English
Physical Description:
ix, 82 leaves. : illus. ; 28 cm.

Subjects

Subjects / Keywords:
Adolescents ( jstor )
Analysis of variance ( jstor )
Celebrations ( jstor )
Delay lines ( jstor )
Emotional disturbance ( jstor )
Emotional expression ( jstor )
Null hypothesis ( jstor )
Standard deviation ( jstor )
Students ( jstor )
Test ranges ( jstor )
Adolescence ( lcsh )
Dissertations, Academic -- Special Education -- UF
Learning, Psychology of ( lcsh )
Mentally ill ( lcsh )
Special Education thesis Ph. D
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 78-80.
Additional Physical Form:
Also available on World Wide Web
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:
022915861 ( AlephBibNum )
14171923 ( OCLC )
ADB3647 ( NOTIS )

Downloads

This item has the following downloads:


Full Text








THE EFFECT OF DELAYING
CONSEQUENCES UPON THE LEARNING
OF EMOTIONALLY DISTURBED ADOLESCENTS















By

ROBERT KEITH BROWN


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




UNIV RSITY OF FLORIDA


JUNE, 1973












DEDICATION




To Barbara and Kelly













ACKNOWLEDGEMENTS


The writer wishes to express his thanks to the chairman and

co-chairman of his committee, Dr. William Reid and Dr. Lyndal Bullock

respectively. Sincere appreciation is also expressed to the other

committee members, Dr. Henry Boudin and Dr. Hank Pennypacker.

The encouragement and guidance of three other individuals greatly

contributed to my being at this point in my career. Dr. Wayne Richard,

Dr. Robert Soar and Dr. Wilson Guertin deserve special thanks in

appreciation for their influence on my professional growth from the

beginning of my graduate program.

The study could not have occurred without the willing consent of

Mrs. Julia Wickersham and Mr. Jack Barile. I thank them for allowing

me the time to undertake the study. The cooperation of the teachers

and students of the program for emotionally disturbed adolescents in

Duval County is also greatly appreciated.

A very special thanks goes to Mr. William Geiger, Miss Catherine

Baltzell and Dr. Charles Forgnone for their comments and suggestions

on the initial drafts of the dissertation.













TABLE OF CONTENTS


ACKNOWLEDGEMENTS iii

LIST OF TABLES v

LIST OF FIGURES vii

ABSTRACT viii

CHAPTER

I. INTRODUCTION .... . . . . 1

II. REVIEW OF LITERATURE . . . . . 4

III. PROCEDURES . . . . . . . . 10

Statement of the Problem . . .. 10
Hypotheses . . . . . . 11
Delimitations. . . . . .. 12
Definitions. . . . . . ... 12
Subjects ............. 13
Learning Task . . . . .. 13
Administration Procedures. .. ... 15

IV. RESULTS. . . . . . . . ... 21

Analysis of Number of Correct
Responses . . . . . .. 22
Analysis of Response Time. . ... 29
Analysis of Log of Correct
Responding . . . . . . 39
Analysis of Celeration Ratios . 50
Su!~mary of Results . . . ... 56

V. COICLSIfO S. . . . . . . ... 59

APPENDIX A: SAFP-LE LEARNING TASK SHARES . . ... 62

APPENDIX B: CHARTS OF INDIVIDUAL STUDENT PERFORMANCE. 67

BIBLIOGRAPHY . . . . . . . . . . . 78

BIOGRAPHICAL SKETCH. . . . . . . . . ... 81













LIST OF TABLES


TABLE

1. ORDER OF PRESENTATION OF
DELAY/CUE CONDITIONS. . . . . .. 18

2. ANALYSIS OF VARIANCE OF
NUMBER OF CORRECT RESPONSES ...... 23

3. DUNCAN'S NEW MULTIPLE RANGE TEST
ON NUMBER OF CORRECT RESPONSES IN
EACH TRIAL/RESPONSE PRODUCED CUE
CONDITION . . . . . . ... .25

4. DUNCAN'S NEW MULTIPLE RANGE TEST
ON NUMBER OF CORRECT RESPONSES IN
EACH DELAY/RESPONSE PRODUCED CUES
CONDITION . . . . . . ... .27

5. ANALYSIS OF VARIANCE OF RESPONSE
TIME. . . . . . . . . ... 30

6. DUNCAN'S NEW MULTIPLE RANGE TEST
ON RESPONSE TIME FOR EACH TRIAL ..... 31

7. MEANS AND STANDARD DEVIATIONS ON
RESPONSE TIME FOR EACH TRIAL/DELAY
CONDITION . . . . . . ... .34

8. DJUCAN'S NEW MULTIPLE RANGE TEST
ON RESPONSE TIME FOR EACH TRIAL/
DELAY CONDITION . . . . .... .35

9. DUNCAN'S NEW MULTIPLE RANGE TEST
ON RESPONSE TIME IN EACH DELAY/
REPRODUCED CUE CONDITION. ....... .. 37

10. ANALYSIS OF VARIANCE OF LOG OF
RATE OF CORRECT RESPONDING. . . .. 40

11. DUNCAN'S NEW MULTIPLE RANGE TEST
OF LOG OF RATE OF CORRECT RESPONDING
FOP EACH TRIAL. . . . . . ... 41








12. DUNCAN'S NEW MULTIPLE RANGE TEST
OF LOG OF RATE OF CORRECT RESPONDING
FOR EACH DELAY LENGTH. . . . . .. 44

13. DUNCAN'S NEW MULTIPLE RANGE TEST
OF LOG OF RATE OF CORRECT RESPONDING
FOR EACH TRIAL AND CUE CONDITION .... . 46

14. DUNCAN'S NEW MULTIPLE RANGE TEST
ON LOG OF RATE OF CORRECT RESPONDING
IN EACH DELAY/RESPONSE PRODUCED CONDITION. 48

15. CELEBRATION COEFFICIENTS FOR CORRECT
RESPONDING . . . .... . . 51

16. CELEBRATION COEFFICIENTS FOR
INCORRECT RESPONDING . . . .... .52

17. ANALYSIS OF VARIANCE OF
CELERATION RATIOS. . . . . . ... 54

18. CELERATION RATIOS. . . . . . ... 55

19. SUMMARY OF SIGNIFICANT EFFECTS . . .. 58













LIST OF FIGURES

FIGURE

1. MEAN NUMBER OF CORRECT RESPONSES
IN EACH TRIAL/RESPONSE PRODUCED
CUE CONDITION. . . . . . . ... 26

2. MEAN NUMBER OF CORRECT RESPONSES IN
EACH CUE/DELAY CONDITION . . . ... 28

3. MEAN RESPONSE TIME FOR EACH TRIAL. ... 32

4. MEAN RESPONSE TIME FOR EACH TRIAL
UNDER EACH DELAY CONDITION . . ... 36

5. MEAN RESPONSE TIME IN EACH CUE/
DELAY CONDITION. . . . . . . 38

6. MEAN LOG OF RATE OF CORRECT RESPONDING
ON EACH TRIAL. . . . . . . ... 42

7. MEAN LOG OF CORRECT RESPONDING IN
EACH DELAY CONDITION . . . .... .45

8. MEAN LOG OF RATE OF CORRECT
RESPONDING ON EACH TRIAL UNDER
CUES ABSENT/CUES PRESENT
CONDITIONS . . . . . . . . 47

9. MEAN LOG OF RATE OF CORRECT
RESPONDING IN EACH CUE/DELAY
CONDITION. . . . . . . . ... 49













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

THE EFFECT OF DELAYING
CONSEQUENCES UPON THE
LEARNING OF EMOTIONALLY
DISTURBED ADOLESCENTS

By

Robert Keith Brown

1973

Chairman: Dr. William R. Reid
Co-Chairman: Dr. Lyndal M. Bullock
Major Department: Special Education

This study investigated the effects of delayed consequences on the

learning of emotionally disturbed adolescent students. The effects of

three lengths of delays under two conditions of feedback were studied.

The three delay lengths were: 1) a minimum delay of 15 to 45 seconds;

2) a 15 minute delay; and 3) a 3-hour delay of consequence. Each delay

length was investigated under two types of feedback: 1) the stimulus

and response produced cues (work sheets of the student) were returned

when the consequence was delivered; and 2) the stimulus and response

produced cues were not returned to the students (stimulus and response

produced cues absent).

The study investigated the effects of the six conditions on the

number of correct responses, response time, rate of correct responding

and the celebration ratio. As the core of the learning task 48 nonsense

geometric shapes were developed for this study. These 48 shapes were


viii








paired so that each student received a different set of 24 pairs of

these shapes. The 24 pairs were divided so that four pairs of shapes

were presented in each of the six experimental conditions. The

student's task was to learn the correct shape of each pair. The

learning task was presented to each student a total of seven times.

The analysis of the results of the effects of the six conditions

on the four dependent measures indicated that,

1) Delays of 15 minutes and 3 hours had a detrimental
effect on the rate of correct responding when no
response produced cues were present.

2) A delay of 15 minutes had a detrimental effect on
the response time when no response produced cues
were present.

3) Delays of 15 minutes or 3 hours had no detrimental
effect on number of correct responses or celebration
ratios.

4) Delays of 15 minutes or 3 hours had no detrimental
effect on any of the four dependent measures when
response produced cues were present. In fact
performance was superior under the conditions of
3 hour delay as measured by number of correct
responses.

It therefore appears that the detrimental effects of delays of reinforce-

ment can be overcome by the presentation of stimulus and response

produced cues at the time the consequence is delivered. This indicated

that modification techniques used with emotionally disturbed students

nay employ techniques not centered around immediate delivery of

consequences and expect some success. This should result in the

design and adoption of techniques which are not as tedious as the

present ones and which may be utilized more readily without tremendous

outside support. These techniques however should be limited to behaviors

in which stimulus and response produced cues can be presented during

consErcation.













CHAPTER I


INTRODUCTION


The effectiveness of behavior modification techniques with

emotionally disturbed students has been demonstrated by numerous

studies (Allen, Hart, Buell, Harris and Wolk, 1964; Quay, Werry,

McQueen and Sprague, 1966; Graubard, 1969). However, little evidence

of the effective application of such techniques is present in most

nonexperimental public school programs serving disturbed students.

Many differences could be pointed out between the nonexperimental

public school programs and the experimental or university supported

settings in which the published studies have taken place with the more

significant ones being, 1) the lack of professional support in public

school programs as opposed to the university or experimental settings;

and 2) teachers highly trained and experienced in behavior modification

techniques have not systematically applied the techniques when the type

of support supplied in a university or experimental setting is not

present.

Several alternative solutions to this dilemma are posed. The

teacher trainers could attempt to develop a schedule of support for

the students while they are undergoing training which would more

readily withstand extinction. That is they could give high levels of

support during the first part of the teacher's training and gradually





2

fade into an intermittent schedule of support on a very infrequent

basis. Thus the support during the last phases of training would

approximate support in a public school system. This may result in

the teacher's maintaining the effective use of the techniques.

A second alternative is that the public school system may attempt

to increase the professional support. It is doubtful however that

they could duplicate the quality or quantity of support present in

the university or experimental setting.

A third alternative is to devise modified techniques which could

more easily and practically be applied in the public school program.

If effective variants in the usual procedures can be found which are

less tedious and more practical these measures may be maintained with

the present schedule of support. The focal point of this study is

one aspect of the third alternative.

The variation in usual behavior modification procedures to be

investigated in this study is the delay in delivery of consequences

for speciFied behaviors. Most studies demonstrating effective modifi-

cation of emotionally disturbed students' academic or social behavior

have involved contingencies delivered during or immediately after

completion of the behavior undergoing modification. In classroom

situations where the teacher is attempting to modify the behavior of

the disturbed student, it may be impractical to deliver consequences

immediately, especially where large volumes of written academic work

are produced.

Is the effectiveness of the modification techniques dependent on

the immediate application of consequences. or are there ways of

systematically applying effective consequences even when they may be







delayed minutes, hours, or even days? Research by Brackbill and

Kappy (1962), an' Goldstein and Seigel (1971) suggest that the retarding

effect of delay sf consequences may be overcome by presentation of the

stimulus and response produced cues when the consequences are delivered.

Successful modification of behavior has been demonstrated in two

recent studies (Schwartz and Hawkins, 1970; Sluyter and Hawkins, 1972)

where the contingencies were delayed for several hours. These studies

indicate that the usual procedure of immediate presentation of the

consequence may, under certain circumstances, be replaced by systemati-

cally delaying consequences, and still maintain the effectiveness of

the consequence Fn modifying behavior.













CHAPTER II


REVIEW OF LITERATURE


Public school education has steadily been increasing its role in

the remediation of the social and academic behavior of emotionally

disturbed students in recent years. A combination of factors has

contributed to this increase: 1) the failure of the medical profession

to solve all of these students' problems; 2) the shortage of clinically

trained personnel; 3) the demands of parents of disturbed students for

services; and 4) the availability of federal funds for teacher training,

demonstration programs, research (PL 88-164) and support of local

programs (Titles I and III of Elementary and Secondary Education Act

PL 89-10).

The necessity of developing effective and practical techniques

for the remediation of disturbed students' behavior has been brought

to the forefront because of this increasing role. It has become of

critical importance to determine the effect of variations in remediation

techniques on the academic ana social behavior of these students. This

should enable the implementation of more practical and effective

programs.

One of the most effective techniques has been the use of behavior

modification procedures (Graubard, 1969; Hewitt, Taylor and Artuso,

1969; Hanley, 1970; Lovitt, 1970; Glavin, Quay, Annesley and

Werry, 1971). There are a multitude of variations of such techniques







which have been employed with emotionally disturbed students. Most

variations have one procedure in common. This is the immediate

delivery of consequences of some form.

The type of reinforcers used have varied widely. Primary rein-

forcers such as food have found frequent use (Carlson, Arnold, Becker,

and Madsen, 1968). Tokens which lead to a variety of back-up rein-

forcers have been employed in many projects (O'Leary and Becker, 1967;

Glavin, Quay, Annesley, and Werry, 1971). Tokens are utilized due to

the inconvenience of using primary reinforcers and the possible

satiation if only one type of reinforcer is used. Social consequences

such as praise contingent upon the occurrence of specified behaviors

has also proved effective (Allen, Hart, Buell, Harris, and Wolk, 1964;

Becker, Madsen, Arnold, and Thomas, 1967).

The immediate delivery of consequences such as those cited above

may be inconvenient and impractical in many classroom situations for the

teacher of the emotionally disturbed. This is especially true where a

large amount of written academic work is produced which must be

corrected by the teacher before she can distribute the consequences

for correct responses. The delays under these circumstances may be

measured by hours or even days.

Delay of consequence has been studied extensively in populations

other than the emotionally disturbed. A review of these studies adds

light to the major variables to be considered and the possible effects

on learning. Historically, delay of reinforcement was first investigated

with animals. These studies led to research with normal human subjects.

Interest has also been present for the last decade and a half in the

effects delays have on the learning of retarded students.








The studies with animals are important not only because of their

historical value, but because much more control of the significant

variables surrounding the behavior to be acquired and the reinforcement

is possible. This allows for a more precise investigation of the

effect of the delays.

Renner (1964) has complied an excellent review (Watson, 1917;

Mischel and Metzner, 1962) of the effects of delay of reinforcement.

The results of the studies with animals suggest the following concerning

the effects on the acquisition of a response:

1) A constant delay of consequence will retard acquisition.

2) The detrimental effects become more so as indirect
rewards (such as are present in the goal box) are removed.

When shifting focus from animal studies to studies with human

subjects two important differences should be noted: 1) the compar-

ability of the animal data to human data is only rough since many

human studies used delay of feedback of knowledge as opposed to the

primary reinforcement used in animal studies; and 2) there is much

evidence (Brackbill and Kappy, 1962; Erickson and Lipsitt, 1960) that

verbal cues may play an important role in human studies, creating

differential effects above and beyond that of the delay (Renner, 1964).

Renner concluded in his review that the data on humans agreed

essentially with the animal data. In his review, however, two factors

come into more prominent view when consideration is given to the human

studies: 1) the noticable effect of response produced cues (corres-

ponding to the indirect rewards in animal studies) in reducing the

detrimental effect of the delay; and 2) the effect of the intertrial

interval in retarding acquisition.







Brackbill and Kappy (1962) investigated the effect of response

produced cues, finding that they reduce the detrimental effect of

delays of reinforcement and may even enhance resistance to extinction

of a response. This reduction of the detrimental effects has also been

produced by using goal orientation instructions (Erickson and Lipsitt,

1960), a motor response (Biloudeau and Biloudeau, 1958), and stimulus

exposure during delay (Goldstein and Siegel, 1971). Brackbill and

Kappy (1962) theorized that when a subject links the consequence to

the response through the use of mediating cues the delay of reinforce-

ment has no detrimental effect.

The possibility that the major retarding variable in delay of

reinforcement was the intertrial interval rather than response-

consequence interval was suggested by studies by Biloudeau and

Biloudeau (1958) and Denny, Allard, Hall, and Rokeach (1960). These

studies demonstrated that when long intertrial intervals are used

acquisition is retarded. Thus it seems evident from studies with

animals and normal human subjects that there are variables surrounding

delays of consequence which may be manipulated to decrease the negative

effects.

Several studies have been undertaken to investigate the effects of

delays of consequences with retarded children. Ross, Hetherington, and

Wray (1964) found the effects of delay of reinforcement to be similar

for both retarded and normals. Delays of six seconds or more retarded

learning for both populations. Various delay lengths (3,5,6,12, and 18

seconds) were found to differentially affect speed of acquisition of a

response by retarded students (Schoelkopf and Orlando, 1965, 1966).

Presentation of the stimulus during the delay has also been proved to







be an important variable with retarded students (Ward and Baumeister,

1971).

Most of the aforementioned studies with animals and humans have

been done in laboratory conditions with delays measured in seconds. In

a recent study with retarded students in the classroom Piper (1971)

found that a delay of 15 minutes in the delivery of consequence had a

detrimental effect on learning only when response produced cues were

absent. Thus the studies with animals, normal human subjects and

retarded subjects point to the reduction of the detrimental effects of

delay of reinforcement when response produced cues are present during

the delay or when the consequence is delivered.

Mower and Ullman (1945) made a speculation regarding emotional

disturbance and delay of reinforcement. They note that neurotics

frequently persist in responding in ways that lead to punishment.

They contend that the particular behavior may result in some immediate

reinforcer before the delayed punishment. An interrelationship between

ineffectiveness of delayed consequences on behavior, delay of grati-

fication and abnormal behavior has also been postulated through the

work of Mischel (1958, 1961). These studies point to the possibility

that emotionally disturbed students may acquire responses less readily

under delay consequence conditions than normals.

However other research raises questions about the correctness of

this conclusion. Morris (1969) found that a 10-second delay of reward

had no effect on the number of trials emotionally disturbed students

took to reach the criterion of correct responses. In applying the

techniques of delayed consequences to the modification of a maladjusted








student's behavior, Schwartz and Hawkins (1970) found that the desired

behavior could be modified. The behaviors undergoing modification were

presented on videotape several hours after their occurrence and the

consequences were systematically applied resulting in modification

of the behavior in a short period of time. Sluyter and Hawkins (1972)

demonstrated the modification of classroom behavior by the application

of specific consequence by parents based upon a note dispensed after

school. The last two studies demonstrate the possibility of effective

modification of disturbed students' behavior even when consequences

are applied several hours after the behavior occurs.













CHAPTER III


PROCEDURES


Statement of the Problem


This study investigated the effects of delayed consequences on the

learning of emotionally disturbed adolescent students. The effects of

three lengths of delays under two conditions of feedback were studied.

The three 6dlay lengths were, 1) a minimum delay of 15 to 45

seconds; 2) a 15-minute delay; and 3) a 3-hour delay of consequence.

Each delay length was investigated under two types of feedback, 1) the

stimulus and response produced cues (work sheets of the student) were

returned when the consequence was delivered; and 2) the stimulus and

response produced cues were not returned to the students (stimulus and

response produced cues absent).

The study investigated the effects of the six conditions or' the

number of correct responses, response time, rate of correct responding

and the celebration pair. The following are the six conditions

investigated,

A) Minimum delay of consequence, stimulus and response
produced cues absent.

B) 15-minute delay of consequence, stimulus and response
produced cues absent.

C) 3-hour delay of consequence, stimulus and response
produced cues absent.







D) Minimum delay of consequence, stimulus and response
produced cues present.

E) 15-minute delay of consequence, stimulus and response
produced cues present.

F) 3-hour delay of consequence, stimulus and response
produced cues absent.

The following questions were investigated:

1) Do delays in the delivery of consequence of 15 minutes
or 3 hours have detrimental effects on the learning of
emotionally disturbed adolescents when no stimulus or
response produced cues are present?

2) Do delays in the delivery of consequences of 15 minutes
or 3 hours have detrimental effects on the learning of
emotionally disturbed adolescents when stimulus and
response produced cues are present?


Hypothesis

Specifically the following null hypotheses were tested.

1) There are no significant differences among the number
of correct responses in the three delay conditions
(minimum, 15 minutes, 3 hours).

2) There are no significant differences among the response
times under the three delay conditions.

3) There are no significant differences among the rate of
correct responding under the three delay conditions.

4) There are no significant differences among the
celebration ratios under the three delay conditions.

5) There are no significant differences between the number
of correct responses in the two types of consequence
presentations (stimulus and response produced cues
absent, or stimulus and response produced cues present).

6) There are no significant differences between the response
times in the two types of consequence presentations.

7) There are no significant differences between the rate
of correct responding in the two types of consequence
presentations.

8) There are no significant differences between the
celebration ratios under the two types of consequence
presentations.








9) There are no significant interactions among the delay
conditions and consequence types on the number of correct
responses.

10) There are no significant interactions among the delay
conditions and consequence types on the response time.

11) There are no significant interactions among the delay
conditions and consequence types on the rate of correct
responding.

12) There are no significant interactions among the delay
'conditions and consequence types on the celebration ratios.


Delimitations


The population of concern was restricted to emotionally disturbed

adolescent students enrolled in the Duval County Exceptional Child

Program. Caution should be taken against generalizing these results

to other populations.

This study investigated the effect of delays of consequences only

under the conditions specified. This study did not investigate

behavioral characteristics of the students other than in the learning

task.


Definitions


Emotionally disturbed students: refers to students
identified by a school psychologist as emotion-
ally disturbed and enrolled in a program for
emotionally disturbed students.

Celeration Coefficient: refers to how much one must
multiply the predicted rate of responding on
day n to obtain the predicted rate or responding
on day n + 7.

Celeration Ratio: refers to how much one must multiply
the celebration coefficient of rate of incorrect
responding to obtain the celebration coefficient
of rate of correct responding.








Subjects


Ten adolescent emotionally disturbed children enrolled in the

Duval County Exceptional Child Program during May of 1972 constituted

the sample for this study. All of these students had been identified

by a school psychologist as emotionally disturbed prior to enrollment

in the Exceptional Child Program.

The students ranged in age from 12 to 16 years. There were nine

boys and one girl included in the sample. The students were selected

on the basis of low absenteeism from a total of 16 students assigned

to the program at that time.


Learning Task


As the core of the learning task 48 nonsense geometric shapes were

developed for this study (see Appendix A for examples). These 48 shapes

were paired so that each student received a different set of 24 pairs of

these snapes. The 24 pairs were divided so that four pairs of shapes

were presented in each of the six experimental conditions.

One shape from each of the pairs was designated as correct by the

experimenter before the start of the study. The selection of the correct

shape was made using a random number table.

The student's task was to learn the correct shape of each pair

under conditions of:

(A) minimum delay of consequence-stimulus and response produced
cues absent.

(D) 15 minutes delay of consequence-stimulus and response
produced cues absent.

(C) 3 hours delay of consequence-stimulus and response produced
cues absent.







(D) minimum delay of consequence-stimulus and response produced
cues present.

(E) 15 minutes delay of consequence-stimulus and response and
produced cues present.

(F) 3 hours delay of consequence-stimulus and response produced
cues present.


Utilization of this two-choice discrimination task allows for

comparison with similar studies and controls for differential exposure

to the task, which would have been present if material from the academic

program were utilized. Each of the 24 unique, nonsense geometric pairs

was presented on a separate 8i by 11 sheet. Each of the student's

sheets were numbered for identification purposes, 1 to 24. There were

two sheets for each unique, nonsense geometric pair; one sheet with the

correct geometric shape on the right and one with the correct geometric

shape on the left.

The correct shape for each pair was randomly assigned by the

experimenter before the first trial. During the first trial the students

had no information on which shape of each pair was correct but had to

guess. After the first trial the students had information on the

correctness or incorrectness of his first guess. The specific condition

under which each pair was presented determined the types of feedback

and the length o-f time the feedback was delayed.

The student indicated his selection of the correct shape of each

pair by circling that shape. The dependent variables were the number

of correct choices of the four pairs under each condition, the response

time, the rate of correct responding, and the celebration ratio (the

celebration of rate correct compared to the celebration of rate

incorrect).







Administration Procedures

On the Friday previous to the beginning of the week of data

collection, the students were oriented to the study and the learning

task. The study was described as an attempt to discover how the

students learn 'ulder various conditions. It was explained that the

experimenter needed the cooperation of the students by their regular

attendance and appropriate behavior during the learning situations. In

return the students were told that they would earn five cents for each

correct choice they made.

The experimenter then told them that, as an example of what they

would be doing the next week, he would like to have each of them do a

sample of the type of problems they would be working with. The

experimenter then showed the students four pairs of nonsense geometric

shapes (not used in the actual trials) saying,

These are four pairs of oddly shaped figures like the ones
yo0'u ;,ill be dealing with next week. One of the figures from
each of these pairs has been decided upon as being correct
by a friendd of mine. When I hand out your sheet it's each
of your jobs to guess (without help from anyone else) which
is the correct shape of each figure. Do you have any questions?

After answering all questions, most of which dealt with how they would

be paid, the experimenter continued.

OK, where I hand you your papers leave them turned over until
I give you your instructions to start. Indicate which figure
is correct oii your sheet by circling the shape you think is
correct, whe~ you finish your paper please raise your hand
until I come over to your desk. I'll take up your papers and
hand them back to you, telling you which ones you got correct
and give you a nickle for each one you get correct.

After all the papers had been corrected and the money was given out

the experimenter said,








OK, you all did very well. Next week each of you will have
24 pairs of the odd shapes like the ones today. Each of you
however will have pairs that are different from anyone elses.

I will give you 4 papers at a time exactly like I did today
except next week I won't always return your papers or tell
you which papers were correct right away. Sometimes I'll
tell you right away and sometimes I won't be able to tell
you until in the afternoon. Are there any questions?

After answering all questions, again most of which concerned the amount

of money, the experimenter continued.

Each day next week you will be working with the same pairs,
and you will have a chance to improve the number you get
correct by remembering which shapes you get correct from
day to day.

Now remember we will do this learning game every day next
week and you will be earning the money we discussed. So
be sure and be here all next week.


Week of Testing

The students were administered the learning task during art class

in small groups of 2 to 5. Before the beginning of the session the

experimenter reminded them of the task by saying,

Today we're going to get into the learning for real. Remember,
I said last week that each of you would be presented pairs of
odd shapes. Your job is to learn which shape of each pair is
the correct shape. Indicate which pair is correct by circling
it like we did last week. Remember each of you has 24 pairs
of shapes made up just for you which you will be working with
each day this week, your pairs are different than anyone else's.
Again I want to remind you that the most important thing is how
many of the pairs you get correct. Your art teacher will,
however, be timing how long it takes you, take as much time as
you need but raise your hand as soon as you are finished with
each four pairs.

Sometimes I will give your paper back right away and tell you
right away which ones you got correct (by telling you the
identification number of the ones you got correct) and give
you your reward, and sometimes I'll tell you and won't give
you back your paper, and sometimes I won't tell you or return
your paper until later.








As some will still be working when you finish I would like no
talking whiinle you are in this room during the learning games.

Are there any questions?

After all questions were answered the experimenter continued.

"OK, don't turn the paper over until I tell you to begin."

The experimenter handed out the papers and told the students to begin.

The art teacher Imegan the stop watch when the experimenter told the

students to begin and recorded the time each student took to complete

the described task. The order of presentation of the conditions was

counter balanced to prevent any condition from occurring predominately

in any presentation position. The order is presented in Table 1.

The experimenter responded to each student in the following way

dependent upon the condition.

Condition A

Experimenter collected each student's papers and within 15 to 45

seconds told the student which number he correctly circled (i.e. number

1 and 3 were corr-ect), and gave the student the appropriate amount of

money.

Condition B

Experimenter collected each student's papers and in 15 minutes

(during which time the student was working on other tasks) told the

student which numbers he circled correctly and gave the student the

appropriate amount of money.

Condition C

Experimenter collected each student's papers and in 3 hours told

the student the numbers he circled correctly and gave the student the

appropriate amount of money.








TABLE 1


ORDER OF PRESENTATION OF
DELAY/CUE CONDITIONS


ORDER FIRST SECOND THIRD FOURTH FIFTH SIXTH

DAY CONDITIONS

One B C F D E A

Two F B C E A D

Three C F A B D E

Four E A D F C B

Five A D E C B F

Six D E F B A C

Seven B F D A C E







Condition D

Experimenter collected each student's papers and within 15 to 45

seconds gave the student the paper back saying "Look at the pairs you

circled number and correctly. Here is your money

cents." The experimenter then collected the student's papers

again.

Condition E

Experimenter collected each student's papers and in 15 minutes

gave the student the paper back saying "Look at the pairs, you circled

number and correctly. Here is your money cents."

The experimenter then collected the student's papers again.

Condition F

Experimenter collected each student's papers and in three hours

gave the student the paper back saying "Look at the pairs, you circled

number and correctly. Here is your money cents."

The experimenter then collected the student's papers again.

Second, third, fourth, and fifth day

Experimenter said,

Today we are going to do the same things as yesterday. You
each will have the same pairs and the same shape of each pair
will be correct as it was yesterday. OK?

Now remember don't turn your paper over until I say begin
and remember to raise your hand when you have completed each
set of four pairs.

The experimenter then handed out the papers and said,

"Now we're reedy to begin."

The experimenter response was the same as during the first day

depending on the conditions.





20

Eighth and ninth day

The experimenter told the students before they began the eighth day

that they would now be able to earn $.25 for three correct responses on

a set and $.50 for four correct responses a set. No money would be

earned for one or two correct responses. The experimenter response

was similar to the previous week but the amount of money earned was

changed. After the final consequence was delivered on the ninth day

the purpose of the study was discussed with the students.













CHAPTER IV


RESULTS


Three measures of the students' behavior were collected as raw

data to be used in the analyses to determine the effects of the six

conditions on the learning of the emotionally disturbed adolescents.

Tnese variables were the number of correct responses, the number of

incorrect responses, and the time (in seconds) required for all

responses in each condition. From these measures the rate of correct

responding and the rate of incorrect responding were calculated. These

new variables were then plotted on six-cycle semilog paper (Lindsley,

Behavior Research Co.) which provides a display of relative rather than

absolute change.

The log of correct and incorrect responding provided the data for

the calculation of the celebration coefficients for rate of correct and

incorrect responding. The coefficients provide a measure of change of

behavior over a week's period of time. A ratio of the celebration

coefficients of the log of the rate of correct responding to the log

of the rate of incorrect responding was then calculated. This ratio

provides a measure of the rate of diversion of the two rates and is

referred to as the celebration ratio (for a detailed description of

celebration coefficients and ratios see Geneviene Skypek's dissertation,

1971 pages 24-27).







To provide answers to the two major questions and the nine null

hypotheses (see page 11) the following analyses of the data were

performed:

1) a four way analysis of variance repeated measures design
on the number of correct responses;

2) a four way analysis of variance repeated measures design
on the response time;

3) a four way analysis of variance repeated measures design
on the log of the rate of correct responding;

4) a three way analysis of variance repeated measures design
on the celebration ratios.


The analyses are presented in the above order. A summary comparison of

the results of the various analyses is presented at the end of the

chapter.


Analysis of Number of Correct Responses


The number of correct responses has been utilized as the major

dependent variable in several studies on the effects of delayed

consequences. The analysis of variance of the number of correct

responses (Table 2) reveals several significant effects. There was a

significantly higher (P(.01) number of correct responses when the

response produced cues were returned as the consequence was delivered

than when the response produced cues were not returned. The mean

number of responses when the cues were present was 2.433 as opposed to

a mean of 2.152 when the cues were absent. This difference is

complicated by two significant interactions. The first is an interaction

between the trials and the presence or absence of response produced cues

(P(.05). The utilization of a Duncan's New Multiple Range Test (Li,








TABLE 2


ANALYSIS W. VARIANCE OF NUMBER OF CORRECT RESPONSES


SOURCE OF DEGREES OF SUM OF MEAN F
VARIATION FREEDOM SQUARES SQUARES RATIO


Trials 6 8.362 1.394 1.722

Cues 1 8.288 8.288 10.240**

Delay 2 2.871 1.436 1.774

Subjects 9 33.002 3.667

Trials X Cues 6 13.229 2.205 2.724*

Trials X Delay 12 7.295 0.608 .751

Trials X Subjects 54 46.448 0.860

Cues X Delay 2 8.233 4.117 5.087**

Cues X Subjects 9 33.689 3.743

Delay X Subjects 18 40.748 2.264

Trials X Cues X 12 12.200 1.017 1.256
Delay
Trials X Cues X 54 51.295 0.950
Subjects
Trials X Delay X 108 97.752 0.905
Subjects
Cues X Delay X 18 24.147 1.342
Subjects
Residual 108 87.405 0.809

TOTAL 419 474.964


*Significant at .05
**Significant at .01








1964, pages 270-273) presented in Table 3 indicated the following:

1) That the number of correct responses on trial six when cues were

present was significantly different from trial one when cues were

present and trials three and four when cues were absent; and 2) That

the number of correct responses on trial seven with cues present was

significantly higher than on either trials three or four when cues

were absent.

A plot (Figure 1) of the mean number of correct responses on each

trial under the two-cue conditions further illustrates the interaction.

Under conditions in which the response produced cues were present a

steady increase in the number of correct responses was made. However

in conditions where response produced cues were absent no steady

increase in number of correct responses is shown.

The second interaction which complicates the interpretation and

the effect due to cue presentation is between cues and delay conditions

(P<.01). A Duncan's New Multiple Range Test (Table 4) of the six

conditions reveals the source of the interaction. The three-hour

delay of consequence in which response produced cues were present along

with the consequence differed significantly from every other condition

except immediate presentation of consequence with response produced

cues absent.

The mean number of correct responses are plotted (Figure 2) to

illustrate the interaction. From this it can be seen that the mean

number of correct responses was very similar in both cues present and

cues absent conditions when the consequence was delivered immediately.

As the length of the delay was increased to three hours, the presen-

tation of the response produced cues when the consequence was delivered








TABLE 3

DUNCAN'S NEW MULTIPLE RANGE TEST ON NUMBER
OF CORRECT RESPONSES IN EACH TRIAL/
RESPONSE PRODUCED CUE CONDITION


LABEL TRIAL / RESPONSE MEANS STANDARD
PRODUCED CUES DEVIATIONS


INC

1C
IC

2NC

2C

3NC

3C

4NC

4C

5NC

5C

6NC

6C

7NC

7C


No Cues

Cues

No Cues

Cues

No Cues

Cues

No Cues

Cues

No Cues

Cues

No Cues

Cues

No Cues

Cues


2.4000

2.000

2.267

2.167

1.833

2.400

1.900

2.433

2.267

2.567

2.233

2.767

2.167

2.7000


1.102

1.174

0.827

0.950

0.950

1.102

0.950

1.006

1.172

1.040

1.006

0.935

1.085

1.208


LABEL 3NC 4NC 1C 7NC 2C 6NC 5NC 2NC 3C 1NC 4C 5C 7C 6C

ORDERED 1.8 1.9 2.0 2.2 2.2 2.2 2.3 2.3 2.4 2.4 2.4 2.6 2.7 2.8
MEANS
SUBSETS







MEAN
NUMBER
OF CORRECT
RESPONSES



3.0

2.9

2.8

2.7

2.6

2.5

2.4

2.3

2.2

2.1

2.0

1.9

1.8

1.7

1.6

1.5

TRIAL


SNNo cues present

X,--- Cues present

FIGURE 1

MEAN NUMBER OF CORRECT RESPONSES IN
EACH TRIAL/RESPONSE PRODUCED
CUE CONDITION








TABLE 4


DUNCAN'S NEW MULTIPLE RANGE TEST ON NUMBER OF
CORRECT RESPONSES IN EACH DELAY/RESPONSE
PRODUCED CUES CONDITION


DELAY


A Minimum

B 15 Minutes

C 3 Hours

D Minimum

E 15 Minutes

F 3 Hours


/ CUES


/ No Cues


No Cues


/ No Cues

/ Cues


Cues


/ Cues


MEANS


2.357

2.043

2.057

2.300

2.314

2.686


STANDARD
DEVIATIONS

.964

1.056

1.062

1.095

1 .001

1.110


LABEL I 8 C D E A F
ORDERED 2.043 2.057 2.300 2.314 2.357 2.686
MEAN S_
SUBSETS








MEAN NUMBER
OF CORRECT
RESPONSES





3.0

2.9

2.8

2.7

2.6

2.5

2.4

2.3

2.2

2.1

2.0

1.9

1.8

1.7

1.6

1.5

DELAY


MINIMUM


15 MINUTES


---- No cues present

*----*Cues present


FIGURE 2

MEAN NUMBER OF CORRECT RESPONSES
IN EACH CUE/DELAY CONDITION


3 HOURS







resulted in a significantly higher number of correct responses.

In summary the analysis of correct responses indicates that,

1) The presentation of response produced cues resulted in a
greater number of correct responses;

2) A 3-hour delay of consequence with response produced
cues presented at the time of consequentation (condition
F) resulted in a higher number of correct responses
than in any other condition, except immediate presen-
tation of consequence when no response produced cues
were presented during consequentation (condition A);

3) A greater number of correct responses were obtained on
trial six when cues were present than on trial one when
cues were present and trials three and four when cues
were absent, and a greater number of correct responses
were obtained on trial seven when cues were present than
on trials three and four when cues were absent.


Analysis of Response Time

The mean response time was the next variable to be investigated.

The analysis provides information on the differential effect of the

various conditions on the speed of response regardless of the accuracy.

The analysis is parallel to analyzing the overall rate of responding due

to the fact that the number of responses under each condition was the

same.

The four way analysis of variance (Table 5) depicts a significant

effect (P.01O) due to trials on the response time. Further investiga-

tion into this effect was conducted using a Duncan's New Multiple range

test (Table 6). Tnis analysis shows that trials two and five result

in significantly lower response times (higher rates of behavior) than

trial one. Figure 3 portrays the relationship between the number of

the trial and mean response time.








TABLE 5


ANALYSIS OF VARIANCE OF RESPONSE TIME


SOURCE OF DEGREES OF SUM OF MEAN F
VARIATION FREEDOM SQUARES SQUARES RATIO


Trials

Cues

Delay

Subjects

Trials X Cues

Trials X Delay

Trials X Subjects

Cues X Delay

Cues X Subjects

Delay X Subjects

Trials X Cues X
Delay
Trials X Cues X
Subjects
Trials X Delay X
Subjects
Cues X Delay X
Subjects
Residual

TOTAL


6

1

2

9

6

12

54

2

9

18

12

54

108

18

108

419


2855.385

132.606

736.455

35293.667

513.924

4406.064

4394.438

1450.216

1626.365

1104.413

3020.469

4394.438

11980.733

2485.853

15280.875

96856.313


475.897

132.606

368.228

3921.518

85.654

367.172

81.378

725.108

180.707

61.356

251.706

81.378

110.933

138.103

141.490


3.363**

0.937

2.603


0.605

2.595**



5.125**





1.779


*Significant at .05
**Significant at .01









TABLE 6


DUNCAN'S NEW MULTIPLE RANGE TEST
ON RESPONSE TIME FOR EACH TRIAL


TRIAL MEANS STANDARD
DEVIATIONS


One 27.950 17.280

Two 21.100 12.722

Three 26.433 14.252

Four 22.333 9.2583

Five 21.750 14.050

Six 22.933 17.391

Seven 27.000 18.552


TRIAL 2 5 4 6 3 7 1


ORDERED 21.1 21.8 22.3 22.9 26.4 27.0 28.0
MEANS
SUBSETS








MEAN
NUMBER OF
SECONDS





35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15

TRIALS


FIGURE 3


MEAN RESPONSE TIME
FOR EACH TRIAL








The effect of the number of trials on the response time cannot be

adequately explained without taking into account the significant

interaction (P<.01) between number of trials and length of delay of

consequences. This interaction shows that on trial one the 15 minute

delay condition resulted in a significantly higher mean response time

than in a three hour delay condition on trial five. This difference

is shown more emphatically in figure 4.

As was present in the analysis of correct responses, a significant

interaction (P<.01) was found between the length of delay and the

presence or absence of response produced cues. The analysis of the six

conditions using Duncan's New Multiple Range Test (Table 9) reveals

the differences. The condition of minimum delay of consequence with

response produced cues absent (condition A) and three hours delay of

consequence with response produced cues present (condition F) resulted

in significantly lower response times than the condition of 15 minutes

delay of consequence with response produced cues absent (condition B).

This interaction is graphically portrayed (figure 5) for further

clarification.

The analysis of response time then points to the following,

1) That a 15 minute delay of consequence with response
Produced cues absent (condition B) results in signi-
ficantly higher response time than when consequences
are presented immediately with response produced cues
absent (condition A) or when consequences are delayed
3 hours and response produced cues are present
(condition F);

2) That trials two and five resulted in significantly
lower response times than trial one.

3) That trial one in the 15 minute delay condition
resulted in a significantly higher mean response
time than trial five in the three hour delay condition.







TABLE 7


MEANS AND STANDARD DEVIATIONS
ON RESPONSE TIME FOR EACH
TRIAL/DELAY CONDITION


LABEL TRIAL / DELAY MEAN NUMBER STANDARD
OF SECONDS DEVIATIONS

1/0 one / Minimum 20.850 14.229

1/15 one / 15 minutes 38.750 20.805

1/3 one / 3 hours 24.250 10.083

2/0 two / Minimum 19.550 12.198

2/15 two / 15 minutes 19.050 10.812

2/3 two / 3 hours 24.700 14.708

3/0 three / Minimum 26.600 15.329

3/15 three / 15 minutes 25.100 12.506

3/3 three / 3 hours 27.600 15.364

4/0 four / Minimum 20.550 7.522

4/15 four / 15 minutes 23.100 9.904

4/3 four / 3 hours 23.350 10.317

5/0 five / Minimum 23.100 10.736

5/15 five / 15 minutes 26.250 19.676

5/3 five / 3 hours 15.900 7.166

6/0 six / Minimum 24.450 16.694

6/15 six / 15 minutes 23.300 21.263

6/3 six / 3 hours 21.050 14.2170

7/0 seven / Minimum 25.950 18.345

7/15 seven / 15 minutes 27.150 18.180

7/3 seven / 3 hours 27.900 19.989








TABLE 8


DUNCAN'S NEW MULTIPLE RANGE TEST
ON RESPONSE TIME FOR EACH
TRIAL/DELAY CONDITION


TRIAL/DELAY
CONDITION


5/3 2/15 2/0 4/0 1/0 6/3 5/0 4/14 6/15 4/3 1/3 6/0 2/3 3/15 7/0 5/15 3/0 7/15 3/3 7/3 1/15


ORDERED
MEANS 16 19 20 21 21 21 23 23 23 23 24 24 25 25 26 26 27 27 28 29 39

SUBSETS








MEAN
NUMBER
OF
SECONDS



38
37
36
35
34
33
32
31
30
29
28
27
26
25
24
23
22
21
20
19
18
17
16
15

TRIALS


-- Minimum delay


U--.


15 minute delay


-& 3 hour delay


FIGURE 4

MEAN RESPONSE TIME FOR
EACH TRIAL UNDER EACH
DELAY CONDITION


1 2 3


I I I 1 I I I








TABLE 9


DUNCAN'S NEW MULTIPLE RANGE TEST ON
RESPONSE TIME IN EACH DELAY/
REPRODUCED CUE CONDITION


CONDITION MEANS STANDARD
DEVIATIONS

(A) 0 min. delay / No Cues 21.100 12.554

(B) 15 min. delay / No Cues 28.714 20.383

(C) 3 hr. delay / No Cues 24.514 14.860

(D) 0 min. delay / Cues 24.800 15.144

(E) 15 min. delay / Cues 23.371 13.386

(F) 3 hr. delay / Cues 22.786 12.949

CONDITIONS A F E C D B
ORDERED
MEANS 21.100 22.736 23.371 24.514 24.800 28.714

SUBSETS








MEAN NUMBER
OF SECONDS





30

29

28

27

26

25

24

23

22

21

20

19

18

17

16

15


MINIMUM 15 MINUTES

No Cues present

----- Cues present


FIGURE 5

MEAN RESPONSE TIME IN EACH
CUE/DELAY CONDITION


3 HOURS








Analysis of Log of Rate of Correct Responding


From the number of correct responses and the response time the

rate of correct responding was calculated. The log 10 of these rates

was then calculated. This provided the variable which is used in this

analysis. By combining both time and frequency of responding this

variable provides a more precise measure of the effect of the various

conditions on the learning of the emotionally disturbed adolescents.

The analysis of variance (Table 10) provides evidence of the

increased precision of the measure. This is indicated by the fact

that five significant differences were present in this analysis as

opposed to three significant differences in each of the previous

analyses.

The number of trials had a significant effect (P<.O1) on the rate

of correct responding. The rate of correct responding on trial six

is significantly higher than the rate of correct responding on trials

one and three. The log of the rate of correct responding on trial

three is in addition lower than on all other trials with the exception

of trials one and four. Plotting of the data in figure 6 provides

visual display of the trend of increasing rates of correct responding as

the number of trials increases.

The presence of response produced cues when the consequence was

delivered resulted in a significantly higher (P(.01) mean rate than

when the response produced cues were absent. The mean log of correct

responses when response produced cues were present was .815 as opposed

to a mean of .749 when they were absent. The length of the delay had a

differentially significant effect (P(.05) on the rate of correct








TABLE 10


ANALYSIS OF VARIANCE OF LOG OF RATE OF CORRECT RESPONDING


SOURCE OF DEGREE OF SUM OF MEAN F
VARIATION FREEDOM SQUARES SQUARES RATIO


Trials 6 1.909 0.318 4.863**

Cues 1 0.462 0.462 7.057**

Delay 2 0.477 0.238 3.542*

Subjects 9 8.197 0.912

Trials X Cues 6 0.925 0.154 2.356*

Trials X Delay 12 1.162 0.097 1.492

Trials X Subjects 54 5.839 0.108

Cues X Delay 2 1.117 0.558 8.535**

Cues X Subjects 9 1.712 0.190

Delay X Subjects 18 2.085 0.116

Trials X Cues X 12 1.280 0.107 1.631
Delay
Trials X Cues X 54 2.825 0.052
Subjects
Trials X Delay X 108 6.390 0.059
Subjects
Cues X Delay X 18 1.398 0.078
Subjects
Residual 108 7.065 0.065

TOTAL 419 42.842


*Significant at .05
**Significant at .01








TABLE 11


DUNCAN'S NEW
OF RATE


MULTIPLE RANGE TEST OF LOG
OF CORRECT RESPONDING
FOR EACH TRIAL


TRIAL MEANS STANDARD
DEVIATIONS

One 0.731 0.3167

Two 0.822 0.2763

Three 0.666 0.3741

Four 0.754 0.2433

Five 0.850 0.3310

Six 0.862 0.2295

Seven 0.790 0.3640


TRIAL 3 1 4 7 2 5 6
ORDERED
MEANS .67 .73 .75 .79 .82 .85 .86

SUBSETS









LOG OF
RATE OF
CORRECT
RESPONDING



.86
.85
.84
.83
.82
.81
.80
.79
.78
.77
.76
.75
.74
.73
.72
.71
.70
.69
.68
.67
.66
.65
.64
.63
.62 1 11

TRIALS 1 2 3 4 5 6 7



FIGURE 6


MEAN LOG OF RATE OF CORRECT RESPONDING ON EACH TRIAL








responding. The Duncan's New Multiple Range Test (Table 12) pointed

out that a minimum delay in consequation produced a significantly

higher log of rate of correct responding than the 15 minute delay of

consequence. Figure 7 provides a visual display of the differences.

In addition to the three significant main effects two significant

interactions were found. Trials by cues (P .05) and cues by delay

(P .01) proved significant. Analysis of the trials by cues interactions

(Table 13) generated the conclusions that, 1) trial six with cues

present resulted in a higher rate than trials three or four with cues

absent; and 2) trial three with cues absent resulted in a lower rate

than all other trial/cue conditions except trial four with cues absent.

The graphic portrayal (Figure 8) of the means provides a very good

picture of the interaction between number of trials and presence or

absence of cues. In the plot it is readily apparent that rate of

correct responding tends to steadily increase as the number of trials

increase when response produced cues are present. However no definite

trend either in increase or decrease is observed when response produced

cues are absent.

The analysis of the cues by delay interaction resulted in several

significant differences (Table 14): Immediate delivery of consequence

with response produced cues absent (condition A) resulted in a higher

correct rate of responding than under conditions of a 15 minute or three

hour delay cues absent (conditions B and C). A higher rate of correct

responding was also present under conditions of 15 minute and 3 hour

delays cues present (conditions E and F) than under the condition of

a 15 minute delay of consequence cues absent. To point out clearly

these differences the mean values are plotted in figure 9.








TABLE 12


DUNCAN'S NEW MULTIPLE RANGE TEST OF
LOG OF RATE OF CORRECT RESPONDING
FOR EACH DELAY LENGTH


DELAY LENGTH MEANS STANDARD
DEVIATIONS

Minimum 0.817 0.287

15 Minutes 0.730 0.312

3 Hours 0.798 0.353



DELAY MINIMUM 3 HOURS 15 MINUTES

ORDERED .82 .80 .73
MEANS

SUBSETS








LOG OF RATE
OF CORRECT
RESPONDING





.86
.85
.84
.83
.82
.81
.80
.79
.78
.77
.76
.75
.74
.73
.72
.71
.79
.69
.68
.67
.66
.65
.64
.63
.62

DELAY


MINIMUM


15 MINUTES


FIGURE 7


MEN: LOG OF RATE OF CORRECT RESPONDING
IN EACH DELAY CONDITION


3 HOURS








TABLE 13


DUNCAN'S NEW MULTIPLE RANGE TEST OF
THE LOG OF RATE OF CORRECT
RESPONDING FOR EACH TRIAL
AND CUE CONDITION


LABEL TRIAL RESPONSE MEANS STANDARD
PRODUCED CUES DEVIATIONS


1NC

1C

2NC

2C

3NC

3C

4NC

4C

5NC

5C

6NC

6C

7NC

7C


No Cues

Cues

No Cues

Cues

No Cues

Cues

No Cues

Cues

No Cues

Cues

No Cues

Cues

No Cues

Cues


0.742

0.719

0.845

0.798

0.550

0.781

0.693

0.815

0.842

0.858

0.832

0.893

0.747

0.833


0.353

0.281

0.288

0.266

0.422

0.282

0.221

0.252

0.328

0.339

0.228

0.283

0.375

0.354


LABEL 3NC 4NC 1C 1NC 7NC 3C 2C 4C 6NC 7C 5NC 2NC 5C 6C
ORDERED
MEANS .55 .69 .72 .74 .75 .78 .80 .82 .83 .83 .84 .85 .86 .89
SUBSETS








LOG OF
CORRECT
RESPONDING

.89
.88
.87
.86
.85
.84
.83
.82
.81
.80
.79
.78
.77
.76
.75
.74
.73
.72
.71
.70
.69
.68
.67
.66
.65
.64
.63
.62
.61
.60
.59
.58
.57
.56
.55

TRIALS


1 2 3 4 5 6 7

---* Cues absent

C--a Cues present


FIGURE 8

MEAN LOG OF RATE OF CORRECT
RESPONDING ON EACH TRIAL
UNDER CUES ABSENT/CUES
PRESENT CONDITIONS








TABLE 14


*UUNCAN'S NEW MULTIPLE RANGE TEST ON
LOG OF RATE OF CORRECT RESPONDING
IN EACH DELAY/RESPONSE
PRODUCED CUE CONDITION


CONDITION MEANS STANDARD
DEVIATIONS

A 0.877 0.272

B 0.675 0.333

C 0.72 0.384

D 0.774 0.236

E 0.795 0.282

F 0.852 0.302



CONDITIONS B C D E F A
ORDERED
MEANS 0.675 0.720 0.774 0.795 0.852 0.877

SUBSETS







LOG OF
CORRECT
RESPONDING





.88
.87
.86
.85
.84
.83
.82
.81
.80
.79
.78
.77
.76
.75
.74
.73
.72
.71
.70
.69
.68
.67
.66
.65

DELAY


MINIMUM


15 MINUTES


--- No cues present

&-a Cues present


FIGURE 9

MEA~ LOG OF RATE OF CORRECT RESPONDING
IN EACH CUE/DELAY CONDITION


3 HOURS








In summary the analysis of the log of rate of correct responding

provides the following:

1) That the presentation of response produced cues results
in higher rates of correct responding;

2) That immediate consequentation resulted in higher rates
of correct responding than a 15 minute delay in consequence;

3) That immediate delivery of consequence cues absent (condition
A) resulted in a higher rate of correct responding than under
conditions of a 15 minute or three hour delay cues absent
(conditions B and C).

4) That a higher rate of correct responding resulted from
a 15 minute or 3 hour delay cues present (conditions
E and F) than under the condition of a 15 minute delay
of consequence cues absent.

5) That the rate of correct responding was lower on trial
three than on all trials except one and four.

6) That trial six resulted in a higher rate of correct
responding than on trials one and three.

7) That trial three with cues absent resulted in a lower
rate than all other trial cue conditions except trial
four with cues absent; and

8) That trial six with cues present resulted in a higher
rate than trials three or four with cues absent.


Analysis of Celeration Ratios

A line of best fit was calculated for the log 10 of correct

responding using linear regression. From this the predicted log 10 of

the rate of correct and incorrect responding for each day was calculated.

From this information the celebration coefficient for correct and

incorrect was calculated. This is the number one must multiply the

predicted rate of responding on day n to obtain the predicted rate of

responding on day n + 7. The coefficients are presented in Table 15

and 16. Celeration ratios were then calculated by dividing the

celebration coefficient of rate of correct responding for each subject








TABLE 15


CELEBRATION COEFFICIENTS FOR CORRECT RESPONDING




CONDITIONS


Subject A B C D E F

1 1.354 1.201 0.767 1.460 0.520 2.785

2 0.682 1.248 0.250 0.297 0.139 0.716

3 2.257 0.778 0.458 0.633 1.831 1.482

4 0.540 2.121 0.360 2.187 0.454 2.515

5 1.045 1.959 1.877 1.064 1.646 3.065

6 0.955 2.079 0.943 1.090 2.392 1.429

7 1.050 1.094 0.979 1.153 2.347 1.760

8 1.0S2 2.078 1.364 0.693 0.688 3.261

9 0.889 1.744 8.459 2.389 5.293 1.922

10 0.667 0.481 0.841 1.821 1.076 1.355





52


TABLE 16


CELEBRATION COEFFICIENTS FOR INCORRECT RESPONDING








in each condition by the respective celebration coefficient for rate of

incorrect responding.

The three way analysis of variance (Table 17) reveals one signifi-

cant effect. The absence or presence of cues during consequation has a

significant effect on the celebration ratio. The mean celebration ratio

was significantly higher in the cues present condition than in the cues

absent condition. This is in line with the analysis of number correct

and rate of correct responding.

Looking at the data on an individual student basis provides some

additional information. A celebration ratio of over one indicates

higher increase in rates of correct responding than incorrect responding,

which is an indication that learning in the desired direction has

occurred for that subject in that condition.

Analyzing the celebration ratios in this fashion indicates that

learning took place for,

1) six subjects under condition A.

2) seven subjects under condition B.

3) two subjects under condition C.

4) eight subjects under condition D.

5) five subjects under condition E.

6) ten subjects under condition F.

This indicates more subjects learned under condition F than under

any other condition.








TABLE 17


ANALYSIS OF VARIANCE OF CELEBRATION RATIOS


SOURCE OF DEGREES OF SUM OF MEAN F
VARIATION FREEDOM SQUARES SQUARES RATIO


Cues 1 1.09595 1.09595 6.0316*

Delay 2 0.31030 0.15515 .8539

Subjects 9 1.38916 0.15435

Cues X Delay 2 0.70913 0.35457 1.9514

Cues X Subjects 9 1.13922 0.12658

Delay X Subjects 18 2.80871 0.15604

Residual 18 3.27060 0.1870

TOTAL 59 10.72307


*Significant at .05








TABLE 18


CELEBRATION RATIOS


CONDITIONS


Subject A B C D E F

1 2.033 0.712 0.752 1.989 0.947 2.237

2 0.942 1.373 0.415 0.577 0.409 1.178

3 1.028 1.911 0.382 2.237 1.068 1.602

4 0.567 2.088 0.206 3.713 0.239 2.246

5 1.131 2.132 0.847 0.433 1.749 5.377

6 1.226 1.022 0.851 1.069 6.056 5.878

7 3.088 1.602 0.183 7.297 1.392 1.811

8 2.415 1.387 2.312 1.707 0.766 2.353

9 0.692 0.311 17.513 4.465 2.735 2.324

10 0.930 0.099 0.305 1.654 0.866 6.985








Summary of Results


This section summarizes the foregoing analysis in terms of the

null hypotheses stated on page 11.

1) Analysis of correct responding indicated no
significant effect due to length of delay
taken by itself. So null hypotheses one
cannot be rejected.

2) Analysis of response time resulted in no
significant effect due to length of delay.
Therefore null hypotheses two cannot be
rejected.

-3) Analysis of rate of correct responding resulted
in a significant effect due to length of delay.
Therefore null hypotheses three is rejected.
There is an effect on rate of correct responding
due to the length of the delay. The minimum
delay of consequence resulted in a higher correct
rate of responding than a 15-minute delay of
consequence.

4) Analysis of the celebration ratios resulted in no
significant effects due to length of delay.
Therefore null hypotheses four cannot be
rejected.

'5) Analysis of correct responses resulted in a
significant effect due to presence or absence
of response produced cues. Thus null hypotheses
five is rejected. The number of correct responses
is effected by presence or absence of cues. The
presence of response produced cues during conse-
quation results in higher numbers of correct
responses.

6) Analysis of response time resulted no significant
effects due to cue presence or absence. Therefore
null hypotheses six cannot be rejected.

/ 7) Analysis of rate of correct responding resulted in
a significant effect due to cue presence or absence.
Thus null hypothesis seven is rejected. The rate
of correct responding is effected by presence or
absence of cues. The presence or response produced
cues during consequention results in higher rates
of correct responding.








8) Analysis of the celebration ratios resulted in a
significant effect due to presence or absence of
response produced cues. Therefore hypotheses eight
is rejected. The presence of response produced cues
during consequation results in higher celebration
ratios.

9) Analysis of the number of correct responses resulted
in a significant effect due to the interaction of
delay length and cue condition. Thus null hypotheses
nine is rejected. The interaction is due to the fact
that the number of correct responses with a three
hour delay-cues present (condition F) was higher than
every other combination except a minimum delay of
consequence-cues absent (condition A).

10) Analysis of the response time resulted in a significant
effect due to the interaction of delay length and cue
condition. Thus null hypotheses ten is rejected. The
interaction is due to a lower response time when
consequences are presented immediately and response
produced cues are absent (condition A) and when
consequences are delayed three hours and response
produced cues are present (condition F) than when
consequences are delayed 15 minutes and response
produced cues are absent. (Condition B).

11i) Analysis of rate of correct responding resulted in a
significant effect due to the interaction of delay
length and cue condition. Thus null hypotheses 11 is
rejected. This interaction is due to, 1) a higher
rate of correct responding with immediate delivery of
consequence-cues absent (condition A) than under
conditions of a 15 minute or three hour delay cues
absent (conditions B and C); and 2) a higher rate of
correct responding with a 15 minute or 3 hour delay
of consequence-cues present (conditions E and F) than
with a 15 minute delay of consequence-cues absent.

12) Analysis of celebration ratios resulted in no significant
interaction due to delay length and cue condition combi-
nation. Therefore hypotheses 12 is not rejected.


Table 19 presents a summary of significance. This indicates that

7 of the null hypothesis were rejected, 5 of the 7 at the .01 level and


2 at the .05 level.








TABLE 19

SUrMARY OF SIGNIFICANT EFFECTS


VARIABLES NUMBER RESPONSE RATE OF CELEBRATION
CORRECT TIME CORRECT RATIOS
RESPONDING

Delay Length Not Signi- Not Signi- Significant Not Signi-
ficant ficant (P(.05) ficant

Cues Significant Not Signi- Significant Significant
(P.01O) ficant (P(.01) (P(.05)

Delay Length Significant Significant Significant Not Signi-
X Cues (P(.O1) (P<.01) (P(.01) ficant













CHAPTER V


CONCLUSIONS


This study was undertaken to investigate the effects of delays of

consequence on the learning of emotionally disturbed adolescents. The

effects of the delays were investigated in situations where response

produced cues were presented at the time the consequence was delivered

and where response produced cues were absent during consequation.

The major questions were whether:

1) delays in the delivery of consequences of 15
minutes or 3 hours have detrimental effects
on the learning of emotionally disturbed adoles-
cents when no stimulus or response produced cues
are present;

2) delays in the delivery of consequences of 15 minutes
or 3 hours have detrimental effects on the learning
of emotionally disturbed adolescents when stimulus
and response produced cues are present.

The analysis of the results of the effects on four dependent

measures indicates that:

1) Delays of 15 minutes and 3 hours have a detrimental
effect on the rate of correct responding when no
response produced cues are present.

2) A delay of 15 minutes has a detrimental effect on the
response time when no response produced cues are present.

3) Delays of 15 minutes or 3 hours have no detrimental effect
on number of correct responses or celebration ratios.

4) Delays of 15 minutes and 3 hours have no detrimental effect
on any of the four dependent measures when response produced
cues are present. In fact performance was superior under
conditions of 3 hour delay as measured by number of correct
responses.







It therefore appears that the detrimental effects of delays of

reinforcement can be overcome by the presentation of stimulus and

response produced cues at the time the consequence is delivered. This

means that modification techniques used with emotionally disturbed

students may successfully employ techniques not centered around

immediate delivery of consequences, resulting in the design and

adoption of techniques which are not as tedious as the present ones and

which may be utilized more readily without tremendous outside support.

These techniques however should be limited to behaviors in which

stimulus and response produced cues can be presented during consequation.

Several factors are deserving of further investigation. The most

unexpected result was the higher mean number of correct responses

under a three hour delay than a delay of 15 minutes or the immediate

presentation of the consequence when cues were present. This effect

may be due to the fact that a smaller time period passed between the

time the cues were presented and the next trial (approximately 20 hours

as opposed to 23 hours). It may also be due to the fact that due to

the school's schedule academic work followed the presentation of cues

in all the conditions except the three hour delay. Thus the academic

work may have interfered with remembering the correct responses. Both

these possibilities should be investigated.

It would be desirable to investigate the effect of delays in other

populations of emotionally disturbed adolescents, with different

learning tasks, stronger consequences and different time lengths. There

is enough evidence from this study and two applied studies (Schwartz

and Hawkins, 1970; Sluyter and Hawkins, 1971) to show teachers of





61


emotionally disturbed adolescents that less tedious modification

procedures centered around delayed consequences can and should be

employed effectively.































APPENDIX

A





63






7 7






































Z___/





64




65






























i1





66

































A P P E N D I X







1000
500

100
50


C


W- Rate Incorrect


r)


A






fA


E


B


10
5

1
.5


.1
.05

.01
.005

.001
DAYS


2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8 SU
SUBJECT 1
Rate Correct PROTEGE







1000
500

100
50


2 4 8 2 4 8 4 8


B






F4r


E


F


S4 6 8 2 4 6 8 2 4 6 8


SRate Correct SUBJECT 2
Rate Incorrect PROTEGE


D





1i,


.1
.05

.01
.005

.001
DAYS


-~i-?-~---~ -- -. -_








1000

500


100

50


A


F


2468 2468 2468 2468 2468 2468


--- Rate Correct

--' Rate Incorrect


SUBJECT 3
PROTEGE


A


.1

.05


.01

.005


.001
DAYS






























__ __ _ JI -_ -I -


SUBJECT 4
PROTEGE


A-0 Rate Incorrect


'1000
500

100
50


E


C


B





(T


.05

.01
.005

.001
DAYS


2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8
Rate Correct








































SUBJECT 5
PROTEGE


--- Rate Correct

x-x Rate Incorrect


1000

500


100

50


10

5


1

.5


.1


.05


.01

.005


.001
DAYS





























SUBJECT 6
PROTEGE


-- Rate Incorrect


1000
500

100
50


D



r~""I


A




X1


.05

.01
.005

.001
DAYS


2 4 6 8 2 4 6 8 2468 2 4 6 8 2 4 6 8 2 4 6 8
,---- Rate Correct


_ _ _ _ _ _ _ _ _ _ _







1000

500


100

50


SUBJECT 7
PROTEGE


- Rate Incorrect


A


B


C


E





Si^


F


5


1

.5


.1


.05


.01

.005


.001

DAYS


2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8
S-- Rate Correct


I I I I I I






1000
500
100
50


I


2468 2468 2468 2468 2468 2 4 6 8


c--- Rate Correct
--T< Rate Incorrect


SUBJECT 8
PROTEGE


B


I-


D


0N


.05
.01
.005
.001
DAYS


L~



































SUBJECT 9
PROTEGE


x--- Rate Incorrect


1000
500

100
50


I


.05

.01
.005

.001
DAYS


{


2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8 2 4 6 8
S-- Rate Correct


____ ____ ____ _ _ ____


E







1000
500

100


SUBJECT 10
PROTEGE


-- Rate Correct
--A-- Rate Incorrect


E



$


.05

.01
.005

.001
DAYS













BIBLIOGRAPHY


Allen, E., Hart, B., Buell, J., Harris, F., and Wolk, M. "Effects of
Social Behavior on Isolate Behavior of a Nursery School Child,"
Child Development, 1964, 35, 511-518.

Becker, W., Madsen, C., Arnold, C. and Thomas, D. "The Contingent Use
of Teacher Attention and Praise in Reducing Classroom Behavior
Problems," Journal of Special Education, 1967, 1, 287-307.

Bilodeau, E. A. and Bilodeau, I. M. "Variation of Temporal Intervals
Among Critical Events in Five Studies of Knowledge of Results,"
Journal of Experimental Psychology, 1958, 55, 603-612.

Brackbill, Y., and Kappy, M. S. "Delay of Reinforcement and Retention,"
Journal of Comparative Physiological Psychology, 1962, 20, 433-434.

Carlson, C., Arnold, C., Becker, W. and rladsen, C. "The Elimination of
Tantrum Behavior of a Child in an Elementary Classroom," Behavior
Research and Therapy, 1968, 6, 117-119.

Denny, M. R., Allard, M., Hall, E. and Rokeach, M. "Supplementary
Report: Delay of Knowledge of Results, Knowledge of the Task
and Intertrial Interval," Journal of Experimental Psychology,
1960, 60, 327.

Erickson, M. and Lipsitt, L. P. "Effect of Delayed Reward on Simultaneous
and Successive Discrimination Learning in Children," Journal of
Comparative Physiology, 1960, 53, 256-260.

Glavin, J. P., Quay, H. C., Annesley, F. R., and Werry, J. S. "An
Experimental Resource Room for Behavior Problem Children,"
Exceptional Children, 1971, 38, 131-137.

Goldstein, S. B., and Siegel, A. W. "Observing Behavior and Children's
Learning," Child Development, 1971, 42, 1608-1613.

Grubard, P. S. "Utilizing the Group in Teaching Disturbed Delinquents
to Learn," Exceptional Child, 1969, 36, 267-72.

Hanley, E. M. "Review of Research Involving Applied Behavioral Analysis
in the Classroom," Review of Educational Research, 1970, 40, 597-
627.







Hewett, F. M., Taylor, F. D., and Artuso, A. A. "The Santa Monia
Project: Evaluation of an Engineered Classroom Design with
Emotional/ Disturbed Children," Exceptional Child, 1969, 35
523-529.

Li, J. C. Statistical Inference I, Edward Brothers, Inc., Ann Arbor
Michigan, l1964.

Lovitt, T. "Behavior Modification: The Current Scene," Exceptional
Children, 1970, 37, 85-93.

Mischel, W. "Preference for Delayed Reinforcement: An Experimental
Study of a Cultural Observation," Journal of Abnormal and Social
Psychology, 1958, 56, 57-61.

Mischel, W. "Preference for Delayed Reinforcement and Social Respon-
sibility," Journal of Abnormal and Social Psychology, 1961, 62,
1-7.

Mischel, W. and Netzner, R. "Preference for Delayed Reward as a
Function of Aae Intelligence and Length of Delay Interval,"
Journal of Abnormal and Social Psychology, 1962, 64, 425-431.

Morris, S. M. The Effects of Reward Value and Delay of Reward on the
Acquisition and Retention of a Simultaneous Discrimination Task K
with Emotionally Disturbed Children. Dissertation Abstracts,
70-1299, lUniversity of Connecticut, 1969.

Mower, 0. H. and Ullman, A. D. "Time as a Determinant in Integrative
Learning," Psychological Review, 1945, 52, 61-90.

O'Leary, K. and Becker, W. "Behavior Modification of an Adjustment
Class: A Token Reinforcement Program," Exceptional Children,
1967, 37, 637-642.

Piper, T. J. "Effects of Delay of Reinforcement on Retarded Children's
Learning," Exceptional Children, 1971, 38, 139-145.

Quay, H. Sprague, R., Werry, J. and McQueen, M. "Conditioning Visual
Orientation of Conduct Problem Children in the Classroom,"
Journal of Experimental Child Psychology, 1966, 5, 512-517.

Renner, K. E. "Delay of Reinforcement: A Historical Review,"
Psychological Bulletin, 1964, 61, 341-361.

Ross, L. E., Hetherington, M., and Wray, N. P. "Delay of Reward and the
Learning of a Size Problem by Normal and Retarded Children," Child
Development, 1965, 36, 509-517.

Schoelkopf, A. M. and Orlando, R. "Delayed Versus Immediate Reinforce-
ment in Simultaneous Discrimination Problems with Mentally Retarded
Children," Psychological Record, 1965, 16, 15-23.








Schoelkopf, A. M. and Orlando, R. "Reinforcement Delay Gradients of
Retardates with a Current Discrimination Task Procedure,"
Psychological Record, 1966, 16, 113-128.

Schwartz, M. L. and Hawkins, R. P. "Application of Delayed Reinforce-
ment Procedures to the Behavior of an Elementary School Child,"
Journal of Applied Behavior Analysis, 1970, 3, 85-96.

Skypeck, G. A Descriptive Study of Three Methods of Programming
Flash Cards. Doctoral Dissertation, University of Florida,
1971.

Sluyter, D. J. and Hawkins, R. P. "Delayed Reinforcement of Classroom
Behavior by Parents," Journal of Learning Disabilities, 1972, 5
20-21.

Ward, L. C. and Baumeister, A. A. "Effects of Temporal Variables,
Monetary Incentive, and Method of Information Feedback on the
Paired-Associate Learning of Retardates," American Journal of
Mental Deficiency, 1971, 75, 712-718.

Watson, J. B. "The Effect of Delayed Feeding Upon Reaction,"
Psychobiology, 1917, 1, 51-60.













BIOGRAPHICAL SKETCH


Robert Keith Brown was born October 22, 1946, at Winter Haven,

Florida and is the second of six children of Mr. and Mrs. R. McDonald

3rown. He attended elementary school at Winter Haven, Florida and

was a member of the last graduating class of St. Leo College

Preparatory School, St. Leo, Florida in 1964.

In September, 1964 he entered the University of Florida and in

Au;ust, 1968 received the Bachelor of Arts Degree, majoring in

;;ychology. He then enrolled in the graduate program in special

education at the University of Florida. Mr. Brown received his Master

cf Education Degree, majoring in teaching emotionally disturbed

c';i-den, in 1970.

He worked full time for Dr. Robert S. Soar at the Institute for

the Development of Human Resources as a research assistant for 6 months

before ccntinuinc his work toward a Doctorate of Philosophy in Special

Education. Since June, 1971 he has been employed by the Duval County

School Board in the position of Coordinator of Research and Evaluation

for Exceptional Child Education.

Mr. Brown has made a number of presentations on topics ranging

from accountability to sensitivity training at state and National

Conventions. He has three major publications in the area of

emotionally disturbed children and is presently engaged in developing

evaluative technique for the trainable mentally retarded.





82


Mr. Brown holds memberships in the Council for Exceptional

Children, the American Educational Research Association and Phi

Kappa Phi.

Mr. Brown is divorced and has one lovely daughter Kelly Christine

Brown.













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.




William R. Rei Chairman
Professor of 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.




-Lyl M Bullock, Co-Chairman
Assistant Professor of Educatio


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.




Henry .,d din
Assistant Professor of 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.



Henry S. Pennypacker
Professor of Psychology


This dissertation was submitted to the Dean of the College of Education
and to the Graduate Council, and was accepted as partial fulfillment of
the requirements for the degree of Doctor of Philosophy.

June, 1973


Dean, College f education


Dean, Graduate School




Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E26A5FRFJ_CPGTPK INGEST_TIME 2017-07-14T23:24:12Z PACKAGE UF00097566_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES



PAGE 1

THE EFFECT OF DELAYING CONSEQUENCES UPON THE LEARNING OF EflOTIONALLY DISTURBED ADOLESCENTS By ROBERT KEITH BROWN A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVr:RSITY OF FLORIDA JUNE, 1973

PAGE 2

DEDICATION To Barbara and Kelly

PAGE 3

ACKNOWLEDGEMENTS The writer wishes to express his thanks to the chairman and co-chairman of his committee, Dr. William Reid and Dr. Lyndal Bullock respectively. Sincere appreciation is also expressed to the other committee members, Dr. Henry Boudin and Dr. Hank Pennypacker. The encouragement and guidance of three other individuals greatly contributed to my being at this point in my career. Dr. Wayne Richard, Dr. Robert Soar and Dr. Wilson Guertin deserve special thanks in appreciation for their influence on my professional growth from the beginning of my graduate program. Trie study could not have occurred without the willing consent of Mrs. Julia Wickersham and Mr. Jack Barile. I thank them for allowing Fiie ths time to undertake the study. The cooperation of the teachers and students of the program for emotionally disturbed adolescents in Duval County is also greatly appreciated. A very special thanks goes to Mr. William Geiger, Miss Catherine Ealtzell and Dr. Charles Forgnone for their comments and suggestions on the initial drafts of the dissertation. m

PAGE 4

TABLE OF CONTENTS ACKNOWLEDGEMENTS iii LIST OF TABLES v LIST OF FIGURES vii ABSTRACT viii CHAPTER I. INTRODUCTION 1 II. REVIEW OF LITERATURE 4 III. PROCEDURES 10 Statement of the Problem 10 Hypotheses 11 Del 'imitations 12 Definitions 12 Subjects 13 Learning Task. 13 .Adnjini strati on Procedures 15 IV. RESULTS 21 AnaTysis of Number of Correct Responses 22 Ar-alysis of Response Time 29 Analysis ct Log of Correct Responding 39 Analysis Df Celeration Ratios. ... 50 Siiramary of Results 56 V. COf^atJSIG'IS 59 APPENDIX A: SAM-LE LEARNING TASK SHAPES 62 APPENDIX B: CHARTS GF INDIVIDUAL STUDENT PERFORMANCE. . 67 BIBLIOGRAPHY •. . 78 BIOGRAPHICAL SKETCH 81

PAGE 5

TABLE LIST OF TABLES 1. ORDER OF PRESENTATION OF DELAY/CUE CONDITIONS 18 2. ANALYSIS OF VARIANCE OF NUMBER OF CORRECT RESPONSES 23 3. DUNCAN'S NEW MULTIPLE RANGE TEST ON NUMBER OF CORRECT RESPONSES IN EACH TRIAL/RESPONSE PRODUCED CUE COMDITION 25 4. DUNCAN'S NE'W MU'.TIPLE RANGE TEST OM NUMBER OF CORRECT RESPONSES IN EACH DELAY/RESPONSE PRODUCED CUES CONDITION 27 5. AiiALYSIS OF VARIANCE OF RESPONSE TIME 30 6. DU^'JCAN'S NEW MULTIPLE RANGE TEST ON RESPONSE TIME FOR EACH TRIAL 31 7. MEANS AND STANDARD DEVIATIONS ON RESPONSE TIME FOR EACH TRIAL/DELAY Ca?IDITION 34 8. DL^NCAN'S NEW MULTIPLE RANGE TEST ON RESPONSE TIME FOR EACH TRIAL/ DELAY CONDITION 35 9. DUIICAN'S NEW MULTIPLE RANGE TEST ON RESPONSE TIME IN EACH DELAY/ REPRODUCED CUE CONDITION 37 10. ANALYSIS OF VARIANCE OF LOG OF RATE OF CORRECT RESPONDING 40 n. D'J'iCAN'S NE'^' MULTIPLE P.ANGE TEST OF LOG OF RATE OF CORRECT RESPONDING FOR EACH TRIAL 41

PAGE 6

12. DUNCAN'S NEW MULTIPLE RANGE TEST OF LOG OF RATE OF CORRECT RESPONDING FOR EACH DELAY LENGTH 44 13. DUNCAN'S NEW MULTIPLE RANGE TEST OF LOG OF RATE OF CORRECT RESPONDING FOR EACH TRIAL AND CUE CONDITION 46 14. DUNCAN'S NEW MULTIPLE RANGE TEST ON LOG OF RATE OF CORRECT RESPONDING IN EACH DELAY/RESPONSE PRODUCED CONDITION. 48 15. CELERATION COEFFICIENTS FOR CORRECT RESPONDING 51 16. CELERATION COEFFICIENTS FOR INCORRECT RESPONDING 52 17. ANALYSIS OF VARIANCE OF CELERATION RATIOS 54 18. CELERATION RATIOS 55 19. SUMMARY OF SIGNIFICANT EFFECTS 58 VI

PAGE 7

FIGURE LIST OF FIGURES 1 . MEAN NUMBER OF CORRECT RESPONSES IN EACH TRIAL/RESPONSE PRODUCED CUE CONDITION 26 2. MEAN NUMBER OF CORRECT RESPONSES IN EACH CUE/DELAY CONDITION 28 3. MEAN RESPONSE TIME FOR EACH TRIAL 32 4. MEAN RESPONSE TIME FOR EACH TRIAL UNDER EACH DELAY CONDITION 36 5. MEAN RESPONSE TIME IN EACH CUE/ DELAY CONDITION. . 38 6. MEAN LOG OF RATE OF CORRECT RESPONDING ON EACH TRIAL 42 7. MEAN LOG OF CORRECT RESPONDING IN EACH DELAY CONDITION 45 8. MEAN LOG OF PJ\TE OF CORRECT RESPONDING ON EACH TRIAL UNDER CUES ABSENT/CUES PRESENT CONDITIONS 47 9. MEAN LOG OF RATE OF CORRECT RESPONDING IN EACH CUE/DELAY CONDITION 49 vn

PAGE 8

Abstract of Dissertation Presented to the Graduate Council of the University of rlorida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy THE EFFECT OF DELAYING CONSEQUENCES UPON THE LEARNING OF EMOTIONALLY DISTURBED ADOLESCENTS By Robert Keith Brown 1973 Chairman: Dr. Wtlliam R. Reid Co-Chairman: Dr. Lyndal M. Bullock Major Department: Special Education This study i'fivesti gated the effects of delayed consequences on the learning of emotionally disturbed adolescent students. The effects of three lengths of delays under twc conditions of feedback were studied. The three delay l^engths were: 1) a minimum delay of 15 to 45 seconds; 2) a 15 minute de-lay; and 3) a 3-hour delay of consequence. Each delay length was investigated under two types of feedback: 1) the stimulus and response prodiuced cues (work sheets of the student) were returned when the consequence was delivered; and 2) the stimulus and response produced cues were not returned to the students (stimulus and response produced cues abstrit). The stuoy investigated the effects of the six conditions on the number of correct responses, response time, rate of correct responding and the celeration ratio. As the core of the learning task 48 nonsense geometric shapes were developed for this study. These 48 shapes were Vll 1

PAGE 9

paired so that each student received a different set of 24 pairs of these shapes. The 24 pairs were divided so that four pairs of shapes were presented in each of the six experimental conditions. The student's task was to learn the correct shape of each pair. The learning task was presented to each student a total of seven times. The analysis of the results of the effects of the six conditions on the four dependent measures indicated that^ 1) Delays of 15 minutes and 3 hours had a detrimental effect on the rate of correct responding when no response produced cues were present. 2) A delay of 15 minutes had a detrimental effect on the response time when no response produced cues v/ere present. 3) Delays of 15 minutes or 3 hours had no detrimental effect on number of correct responses or celeration ratios. 4) Delays of 15 minutes or 3 hours had no detrimental effect on any of the four dependent measures when response produced cues were present. In fact performance was superior under the conditions of 3 hour delay as measured by number of correct responses. It therefore appears that the detrimental effects of delays of reinforcement can be overcome by the presentation of stimulus and response p>-cduceri cues at the time the consequence is delivered. This indicated that modification techniques used with emotionally disturbed students may employ techniques not centered around immediate delivery o^ consequences and expect some success. This should result in the design and adoption of techniques which are not as tedious as the present ones and which may be utilized more readily without tremendous outside support. These techniques however should be limited to behaviors in whicn stimulus and response produced cue;i can be presented during consequation. IX

PAGE 10

CHAPTER I INTRODUCTION The effectiveness of behavior modification techniques with emotionally disturbed students has been demonstrated by numerous studies (Allen, Hart, Buell, Harris and Wolk, 1964; Quay, Werry, McQueen and Snrague, 1966; Graubard, 1969). However, little evidence of the effective application of such techniques is present in most nonexperimental public school programs serving disturbed students. Many differences could be pointed out between the nonexperimental public school programs and the experimental or university supported settings in which the published studies have taken place with the more significant ones being, 1) the lack of professional support in public school programs as opposed to the university or experimental settings; and 2) teachers highly tr-ained and experienced in behavior modification techniques have not systematically applied the techniques when the type of support supplied in a university or experimental setting is not present. Several alternative solution:^ to this dilemma are posed. The teacher trainers could attempt to develop a schedule of support for the students while they are undergoing training which would more readily withstand extinction. That is they could give high levels of support during the first part of the teacher's training and gradually

PAGE 11

fade into an intermittent schedule of support on a very infrequent basis. Thus the support during the last phases of training would approximate support in a public school system. This may result in the teacher's maintaining the effective use of the techniques. A second alternative is that the public school system may attempt to increase the professional support. It is doubtful however that they could duplicate the quality or quantity of support present in the university or experimental setting. A third alternative is to devise modified techniques which could more easily and practically be applied in the public school program. If effective variants in the usual procedures can be found which are less tedious and more practical these measures may be maintained with the present schedule of support. The focal point of this study is one aspect of the third alternative. The variation in usual behavior modification procedures to be invest iqated in this study is the delay in delivery of consequences for specified behaviors. Most studies demonstrating effective modification of emotionally disturbed students' academic or social behavior hir/e involved contingencies delivered during or immediately after completion of the behavior undergoing modification. In classroom situations whore the teacher is attempting to modify the behavior of the disturbed student, it may be impractical to deliver consequences immediately, especially where large volumes of v^ritten academic work are produced. Is the effectiveness of the modificaticr; techniques dependent on the immediate application of consequences, or 6re there ways of systematically applying effective consequences even when they may be

PAGE 12

delayed minutes, hours, or even days? Research by Brackbill and Kappy (1962), ant Goldstein and Seigel (1971) suggest that the retarding effect of delay sf consequences may be overcome by presentation of the stimulus and respinse produced cues when the consequences are delivered. Successful nadification of behavior has been demonstrated in two recent studies (Schwartz and Hawkins, 1970; Sluyter and Hawkins, 1972) where the contincencies were delayed for several hours. These studies indicate that the usual procedure of immediate presentation of the consequence may, under certain circumstances, be replaced by systematically delaying ccnsequences, and still maintain the effectiveness of the consequence rn modifying behavior.

PAGE 13

CHAPTER II REVIEW OF LITERATURE Public school education has steadily been increasing its role in the remediation of the social and academic behavior of emotionally disturbed students in recent years. A combination of factors has contributed to this increase: 1) the failure of the medical profession to solve all of these studetits' problems; 2) the shortage of clinically trained personnel; 3) the demands of parents of disturbed students for services; and 4) the availability of federal funds for teacher training, demonstration programs, research (PL 88-164) and support of local programs (Titles I and III of Elementary and Secondary Education Act PL 89-10). The necessity of developing effective and practical techniques for the remediotion of disturbed students' behavior has been brought to the forefront because of this increasing role. It has become of critical importance to determine the effect of variations in remediation techniques on me academic ana social behavior of these students. This should enable the implementation o^ mere practical and effective programs. One of the most effective techniques has been the use of behavior modification procedures (Graubard, 1969; Hewitt, Taylor and Artuso, 1969; Hanley. 1970, Lovitt, 1970; Glavin, Quay, Annesley and Werry, 1971). There are a multitude of variations of such techniques

PAGE 14

which have been employed with emotionally disturbed students. Most variations have one procedure in common. This is the immediate delivery of consequences of some form. The type of reinforcers used have varied widely. Primary reinforcers such as food have found frequent use (Carlson, Arnold, Becker, and Madsen, 1958). Tokens which lead to a variety of back-up reinforcers have been employed in many projects (O'Leary and Becker, 1967; Glavin, Quay, Annesley, and Werry, 1971). Tokens are utilized due to the inconvenience of using primary reinforcers and the possible satiation if only one type of reinfcrcer is used. Social consequences such as praise contingent upon the occurrence of specified behaviors has also proved effective (Allen, Hart, Buell, Harris, and Wolk, 1964; Becker, Madsen, Arnold, and Thomas, 1967). The immediate delivery of consequences such as those cited above may be inconvenient and impractical in many classroom situations for the teacher of the emotionally disturbed. This is especially true where a large amount of written academic work is produced which must be corrected by the teacher before she can distribute the consequences for correct responses. The delays under these circumstances may be measured by hours or even days. Delay of consequence has been studied extensively in populations other than the emotionally disturbed. A review of these studies adds light to the major variables to be considered and the possible effects on learning. Historically, delay of reinforcement was first investigated with animals. These studies led to research with normal human subjects. Interest has also been present for the last decade and a half in the effects delays have on the learning of retarded students.

PAGE 15

The studies with animals are important not only because of their historical value, but because much more control of the significant variables surrounding the behavior to be acquired and the reinforcement is possible. This allows for a more precise investigation of the effect of the delays. Renner (1964) has complied an excellent review (Watson, 1917; 'lischel and Metzner, 1962) of the effects of delay of reinforcement. The results of the studies with animals suggest the following concerning the effects on the acquisition of a response: 1) A constant delay of consequence will retard acquisition. 2) The detrimental effects become more so as indirect rewards (such as are present in the goal box) are removed. When shifting focus from animal studies to studies with human subjects two important differences should be noted: 1) the comparability of the animal data to human data is only rough since many human studies used delay of feedback of knowledge as opposed to the priirary reinforcement used in animal studies; and 2) there is much evidence (Brackbill and Kappy, 1962; Erickson and Lipsitt, 1960) that verbal cues may play an important role in human studies, creating differential effects above and beyond that of the delay (Renner, 1964), Renner concluded in his review that the data on humans agreed essentially with the animal data. In his review, however, two factors come into more prominent view when consideration is given to the human studies: 1) the noti cable effect of response produced cues (corresponding to the indirect rewards in animal studies) in reducing the detrimental effect of the delay; and 2) the effect of tne intertrial interval in retarding acquisition.

PAGE 16

Brackbill and Kappy (1962) investigated the effect of response produced cues, finding thdt they red'jce the detrimental effect of delays of reinforcement and may even enhance resistance to extinction of a response. This reduction of the detrimental effects has also been produced by using goal orientation instructions (Erickson and Lipsitt, 1960), a motor ^-esponse (Biloudeau and Biloudeau, 1958), and stimulus exposure during delay (Goldstein and Siegel , 1971). Brackbill and Kappy (1962) theorized that when a subject links the consequence to the response through the use of mediating cues the delay of reinforcement has no detrimental effect. The possibility that the major retarding variable in delay of reinforcement was the intertrial interval rather than responseconsequence interval was suggested by studies by Biloudeau and Biloudeau (1958) and Denny, Allard, Hall, and Rokeach (1960). These studies demonstrated that when long intertrial intervals are used acquisition is retarded. Thus it seems evident from studies with animals and normal human subjects that there are variables surrounding delays of consequence which may be manipulated to decrease the negative effects. Several studies have been undertaken to investigate the effects of delays of consequences with retarded children. Ross, Hetherington, and Wray (1964) found the effects of delay of reinforcement to be similar for both retarded and normals. Delays of six seconds or more retarded learnino for both poDulations. Various delay lengths (3,5,6,12, and 18 seconds) were found to differentially affect speed of acquisition of a resDonse by retarded students (Schoelkopf and Orlando, 1965, 1966). Presentation of the stimulus during the delay has also been proved to

PAGE 17

8 be an important variable with retarded students (Ward and Baumeister, 1971). Most of the aforementioned studies with animals and humans have been done in laboratory conditions with delays measured in seconds. In a recent study with retarded students in the classroom Piper (1971) found that a deley of 15 minutes in the delivery of consequence had a detrimental effect on learning only when response produced cues were absent. Thus the studies with animals, normal human subjects and retarded subjects point to the reduction of the detrimental effects of delay of reinforcement when response produced cues are present during the delay or whem the consequence is delivered. Mower and Ullman (1945) made a speculation regarding emotional disturbance and delay of reinforcement. They note that neurotics frequently persist in responding in ways that lead to punishment. They contend that the particular behavior may result in some immediate reinforcer before the delayed punishment. An interrelationship between ineffectiveness of delayed consequences on behavior, delay of gratification and abnormal behavior has also been postulated through the v;ork of Mischel (li958, 1961). These studies point to the possibility that emotionally disturbed students may acquire responses less readily under delay consequence conditions than normals. However other research raises questions about the correctness of this conclusion. Morris (1969) found that a 10-second delay of reward had no effect on the number of trials emotionally disturbed students took to reach the criterion of correct responses. In applying the techniques of delayed consequences to the modification of a maladjusted

PAGE 18

student's behavior, Schwartz and Hawkins (1970) found that the desired behavior could be modified. The behaviors undergoing modification were presented on videotape several hours after their occurrence and the consequences v;ere systematically applied resulting in modification of the behavior in a short period of time. Sluyter and Hawkins (1972) demonstrated the modification of classroom behavior by the application of specific consequence by parents based upon a note dispensed after school. The last two studies demonstrate the possibility of effective modification of disturbed students' behavior even when consequences are applied several hours after the behavior occurs.

PAGE 19

CHAPTER III PROCEDURES Statement of the Problem This study investigated the effects of delayed conseauences on the learning of emotionally disturbed adolescent students. The effects of three lengths of delays under two conditions of feedback v^ere studied. The three d-'lay lengths were, 1) a minimum delay of 15 to 45 seconds; 2) a 15-minute delay; and 3) a 3-hour delay of conseauence. Each delay length was investigated under two types of feedback, 1) the stimulus and response produced cues (work sheets of the student) were returned v^hen the consequence was delivered; and 2) the stiiTiulus ard response produced cues were not returned to the students (stimulus and response produced cues absent). The study investigated the effects of the six condition-^ on the number of correct responses, response time, rate of correct responding and the celeration pair. The following are the six conditions investigated, A) Minimum delay of consequence, stimulus and response produced cues absent. B) 15-minute delay of consequence, stimulus and response produced cues absent. C) 3-hour delay of consequence, stimulus and response produced cues absent. 10

PAGE 20

n D) Minimum delay of consequence, stimulus and response produced cues present. E) 15-minute delay of consequence, stimulus and response produced cues present. F) 3-hour delay of consequence, stimulus and response produced cues absent. The following questions were investigated: 1) Do delays in the delivery of consequence of 15 minutes or 3 hours have detrimental effects on the learning of emotionally disturbed adolescents when no stimulus or response produced cues are present? 2) Do delays in the delivery of consequences of 15 minutes or 3 hours have detrimental effects on the learning of emotionally disturbed adolescents when stimulus and response produced cues are present? Hypothesis Specifically the following null hypotheses were tested. 1) There are no significant differences among the number of correct responses in the three delay conditions (minimum, 15 minutes, 3 hours). 2) There are no significant differences among the response times under the three delay conditions, 3) There are no significant differences among the rate of correct responding under the three delay conditions. 4) There are no significant differences among the celeration ratios under the three delay conditions. 5) There are no significant differences between the number of correct responses in the two types of consequence presentations (stimulus and response produced cues absent, or stimulus and response produced cues present). 6) There are no significant differences between the response times in the two types of consequence presentations. 7) There are no significant differences between the rate of correct responding in the two types of consequence presentations. 3) There are no significant differences between the celeration ratios under the two types of consequence presentations.

PAGE 21

12 9) There are no significant interactions among the delay conditions and consequence types on the number of correct responses. 10) Tivsre are no significant interactions among the delay conditions and consequence types on the response time. 11) Thei^e are no significant interactions among the delay conditions and consequence types on the rate of correct responding. 12) There are no significant interactions among the delay conditions and consequence types on the celeration ratios. Delimitations The population of concern was restricted to emotionally disturbed adolescent students enrolled in the Duval County Exceptional Child Program. Caution should be taken against generalizing these results to other populations. This study investigated the effect of delays of consequences only under the conditions specified. This study did not investigate behavioral characteristics of the students other than in the learning task. Definitions Emotionally disturbed students: refers to students identified by a school psychologist as emotionally disturbed and enrolled in a program for emotionally disturbed students. Celeration Coefficient: refers to how much one must multiply the predicted rate of responding on day n to obtain the predicted rate or responding on day n + 7. Celeration Ratio: refers to how much one must multiply the celeration coefficient of rate of incorrect responding to obtain the celeration coefficient of rate of correct responding.

PAGE 22

13 Subjects Ten adolescent emotionally disturbed children enrolled in the Duval County Exceptional Child Program during May of 1972 constituted the saTiDle for this study. All of these students had been identified by a school psychologist as emotionally disturbed prior to enrollment in the Exceptional Child Program. The students ranged in age from 12 to 16 years. There were nine boys and one girl included in the sample. The students were selected on the basis of low absenteeism from a total of 16 students assigned to the program at that time. Learning Task As the core of the learning task 48 nonsense geometric shapes were developed for this study (see Appendix A for examples). These 48 shapes v/ere paired so that each student received a different set of 24 pairs of these snapes. The 24 pairs were divided so that four pairs of shapes were presented in each of the six experimental conditions. Gne shape from each of the pairs was designated as correct by the experimenter before the start of the study. The selection of the correct shape was made using a random number table. The student's task was to learn the correct shape of each pair under conditions of: (A) minimum delay of consequence-stimulus and response produced cues absent. (C) 15 minutes delay of consequence-stimulus and response produced cues absent. (C) 3 hours delay of consequence-stimulus and response produced cues absent.

PAGE 23

14 (D) minlmuTi) delay of consequence-stimulus and response produced cues present. (E) 15 mirutes delay of consequence-stimulus and response and produced cues present. (F) 3 hours delay of consequence-stimulus and response produced cues present. Utilization of this two-choice discrimination task allows for comparison with similar studies and controls for differential exposure to the task, which would have been present if material from the academic program were utilized. Each of the 24 unique, nonsense geometric pairs was presented on a separate ^ by 11 sheet. Each of the student's sheets were numb-ered for identification purposes, 1 to 24. There were two sheets for eAach unique, nonsense geometric pair; one sheet with the correct geometri'c shape on the right and one with the correct geometric shape on the left. The correct shape for each pair was randomly assigned by the experimenter bef.ore the first trial. During the first trial the students had no information on which shape of each pair was correct but had to guess. After the first trial the students had information on the correciriess or incorrectness of his first guess. The specific condition under which eachi pair was presented determined the types of feedback and the length af time the feedback was delayed. The student indicated his selection of the correct shape of each pair by circling! that shape. The dependent variables were the number of correct choices of the four pairs under each condition, the response time, the rate of correct responding, and the celeration ratio (the celeration of rate correct compared to the celeration of rate incorrect).

PAGE 24

15 Administration Procedures On the Friday previous to the beginning of the week of data collection, the students were oriented to the study and the learning task. The study -was described as an attempt to discover how the students lea'Ti i-nder various conditions. It was explained that the experimenter needed the cooperation of the students by their regular attendance and appropriate behavior during the learning situations. In return the studer/ts were told that they would earn five cents for each correct choice they made. The experin^enter then told them that, as an example of what they would be doing the next week, he would like to have each of them do a sample of the type of problems they would be working with. The experi:Tient£r then showed the students four pairs of nonsense geometric shapes (not used in the actual trials) saying, These are four pairs of oddly shaped figures like the ones yc'j will be dealing with next week. One of the figures from each of these pairs has been decided upon as being correct by a friend of mine. When I hand out your sheet it's each of your jobs to guess (without help from anyone else) which iS the corrffct shape of each figure. Do you have any questions? After answering all questions, must of which dealt with how they would be paid, the experimenter continued. OK, when I hand you your papers leave them turned over until 1 give you your instruciions to siart. Indicate which figure is correct 'or, your sheet by circling the shape you think is correct, vvht'i you finish your paper please raise your hand until I coiTTt^ over to your desk. I'll take up your papers and hand them back to you, telling you which ones you got correct and give vol; a nickle for each one you get correct. After all the papers had been corrected and the money was given out the experitTienter Tiaid,

PAGE 25

16 OK, you all did "^ery well. Next week each of you will have 24 pairs of the odd shapes like the ones today. Each of you however will have pairs that are different from anyone elses. I will give you 4 papers at a time exactly like I did toaay except next week I won't always return your papers or tell you which papers were correct right away. Sometimes I'll tell you right away and sometimes I won't be able to tell you until in the afternoon. Are there any questions? After answering all questions, again most of which concerned the amount of money, the experimenter continued. Each day next week you will be working with the same pairs, and you will have a chance to improve the number you get correct by remembering which shapes you get correct from day to day. Now remember we will do this learning game every day next week and you will be earning the money we discussed. So be sure and be here all next week. Week of Testing The students were administered the learning task dur-ing art class in small groups of 2 to 5. Before the beginning of the session the experimenter reminded them of the task by saying, Today we're going to get into the learning for real. Remember, I said last week that each of you would be presented pairs of odd shapes. Your job is to learn which shape of each pair is the correct shape. Indicate which pair is correct by ci>"cling it like we did last week. Remember each of you has 24 pairs of shapes made up just for you which you will be working with each day this week, your pairs are different than anyone else's. Again I want to remind you that the most important thing is how many of the pairs you get correct, Your art teacher will, however, be timing how long it takes you, take as much time as you need but raise your hand as soon as you are finished with each four pairs. Sometimes I will give your paper back right away and tell you right away which ones you got correct (by telling you the identification nurfber of the ones you got correct) and give you your reward, and sometimes I'll tell you and won't give you back your paper, and sometimes I won't tell you or return your paper until later.

PAGE 26

17 As some will still be working when you finish I would like no telkinq wnile you are in this room during the learning games. Are there any questions? After all questions were answered the experimenter continued. "OK, don't t.urr) the paper over until I tell you to begin." The experimenter handed out the papers and told the students to begin. The art teacher h-egan the stop watch when the experimenter told the students to begin and recorded the time each student took to complete the described ta?k. The order of presentation of the conditions was counter balanced to prevent any condition from occurring predominately in any presentation position. The order is presented in Table 1. The experimenter responded to each student in the following way dependent upon the condition. Condition A Expernnenter collected each student's papers and within 15 to 45 seconds told the student which number he correctly circled (i.e. number 1 and 3 were corvect), and gave the student the appropriate amount of money. Condition 3 Experimenter collected each student's papers and in 15 minutes (during which tiriG the student was working on other tasks) told the student which nup.bers he circled correctly and gave the student the appropriate amount of money. Condition C Experimenter collected each student's papers and in 3 hours told the student the rfumbers he circled correctly and gave the student the appropriate amount of money.

PAGE 27

TABLE 1 ORDER OF PRESENTATION OF DELAY/CUE CONDITIONS 18 ORDER DAY

PAGE 28

19 Condition D Experimenter collected each student's papers and within 15 to 45 • seconds gave the student the paper back saying "Look at the pairs you circled number , and correctly. Here is your money cents." The experimenter then collected the student's papers again. Condition E Experimenter collected each student's papers and in 15 minutes gave the student the paper back saying "Look at the pairs, you circled number and correctly. Here is your money cents." The experimente«^ then collected the student's papers again. Condition F Experimenter collected each student's papers and in three hours gave the student the paper back saying "Look at the pairs, you circled number and correctly. Here is your money cents." The experimenter then collected the student's papers again. Second, third, fourth, and fifth day Experimenter said. Today we are going to do the same things as yesterday. You each will have the same pairs and the same shape of each pair vnll be correct as it was yesterday. OK? Now remember don't turn your paper over until I say begin and remember to raise your hand when you have completed each set of four pairs. The experimenter then handed out the papers and said, "Now vie' re reedy to begin." The experimenter response was the same as during the first day depending on the conditions.

PAGE 29

20 Eighth and ninth day The experimenter told the students before they began the eighth day that they would now be able to earn $.25 for three correct responses on a set and $.50 for four correct responses a set. No money would be earned for one or two correct responses. The experimenter response was similar' to tne previous week but the amount of money earned was changed. After the final consequence was delivered on the ninth day the purpose of the study was discussed with the students.

PAGE 30

CHAPTER IV RESULTS Three measures of the students' behavior were collected as raw data to be used in the analyses to determine the effects of the six conditions on the learning of the emotionally disturbed adolescents. Tnese variables were the number of correct responses, the number of incorrect responses, and the time (in seconds) required for all responses in each condition. From these measures the rate of correct responding and the rate of incorrect responding were calculated. These new variables were then plotted on six-cycle semilog paper (Lindsley, Behavior Research Co.) which provides a display of relative rather than absolute change. The log of correct and incorrect responding provided the data for the calculation of the celeration coefficients for rate of correct and incorrect responding. The coefficients provide a measure of change of behavior over a week's period of time. A ratio of the celeration coefficients of the log of the rate of correct responding to the log of the rate of incorrect responding was then calculated. This ratio provides a measure of the rate of diversion of the two rates and is referred to as the celeration ratio (for a detailed description of celeration coefficients and ratios see Geneviene Skypek's dissertation, 1971 pages 24-27). 21

PAGE 31

22 To provida answers to the two major questions and the nine null hypotheses (see page 11) the following analyses of the data were performed: 1) a four way analysis of variance repeated measures design on the number of correct responses; 2) a four way analysis of variance repeated measures design on the response time; 3) a four way analysis of variance repeated measures design on the log of the rate of correct responding; 4) a three way analysis of variance repeated measures design on the celeration ratios. The analyses are presented in the above order. A summary comparison of the results of the various analyses is presented at the end of the chapter. Analysis of Number of Correct Responses The number of correct responses has been utilized as the major dependent variable in several studies on the effects of delayed consequences. The analysis of variance of the number of correct responses (Table 2) reveals several significant effects. There was a significantly higher (P<.01) number of correct responses when the response produced cues were returned as the consequence was delivered than when the response produced cues were not returned. The mean number of responses when the cues were present was 2.433 as opposed to a mean of 2.152 when the cues were absent. This difference is complicated by two significant interactions. The first is an interaction between the trials and the presence or absence of response produced cues (P<.05). The utilization cf a Duncan's New Multiple Range Test (Li,

PAGE 32

23 TABLE 2 ANALYSIS .QF VARIANCE OF NUMBER OF CORRECT RESPONSES SOURCE OF VARIATION

PAGE 33

24 1964, pages 270-273) presented in Table 3 indicated the following: 1) That the number of correct responses on trial six when cues were present was significantly different from trial one when cues were present and trials three and four when cues were absent; and 2) That the number of correct responses on trial seven with cues present was significantly higher than on either trials three or four when cues were absent. A plot (Figure 1) of the mean number of correct responses on each trial under the two-cue conditions further illustrates the interaction. Under conditions in which the response produced cues were present a steady increase in the number of correct responses was made. However in conditions where response produced cues were absent no steady increase in number of correct responses is shown. The second interaction which complicates the interpretation and the effect due to cue presentation is between cues and delay conditions (P^.Ol). A Duncan's New Multiple Range Test (Table 4) of the six conditions reveals the source of the interaction. The three-hour delay of consequence in which response produced cues were present along with the consequence differed significantly from eyery other condition except immediate presentation of consequence with response produced cues absent. The mean number of correct responses are plotted (Figure 2) to illustrate the interaction. From this it can be seen that the mean number of correct responses was ^ery similar in both cues present and cues absent conditions when the consequence was delivered immediately. As the length of the delay was increased to three hours, the presentation of the response produced cues when the consequence was delivered

PAGE 34

TABLE 3 DUNCAN'S NEW MULTIPLE RANGE TEST ON NUMBER OF CORRECT RESPONSES IN EACH TRIAL/ RESPONSE PRODUCED CUE CONDITION 25 LABEL

PAGE 35

MEATJ NUMBER OF CORRECT RESPONSES 26 3.0 2.9 2.8 2.7 2.6 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 TRIAL 2 3 4 No cues present Cues present FIGURE 1 MEAN NUMBER OF CORRECT RESPONSES IN EACH TRIAL/RESPONSE PRODUCED CUE CONDITION

PAGE 36

TABLE 4 DUNCAN'S NEW MULTIPLE RANGE TEST ON NUr-IBER OF CORfiECT RESPONSES IN EACH DELAY/RESPONSE PRODUCED CUES CONDITION 27 DELAY

PAGE 37

28 MEAN NUMBER OF CORRECT RESPONSES 3.0 2.9 2.8 2.7 2.6 2.5 2.4 2.3 2.2 2.1 2.0 1.9 1.8 1.7 1.6 1.5 DELAY MINIMUM (5 MINUTES -» No cues present -Cues present 3 HOURS FIGURE 2 MEAN NUMBER OF CORRECT RESPONSES IN EACH CUE/DELAY CONDITION

PAGE 38

29 resulted in a significantly higher number of correct responses. In summary the analysis of correct responses indicates that, 1) The presentation of response produced cues resulted in a greater number of correct responses; 2) A 3-hour delay of consequence with response produced cues presented at the time of consequentation (condition F) resulted in a higher number of correct responses than in any other condition; except immediate presentation of consequence when no response produced cues v;ere presented during consequentation (condition A); 3) A greater number of correct responses were obtained on trial six when cues were present than on trial one when cues were present and trials three and four when cues were absent, and a greater number of correct responses were obtained on trial seven when cues were present than on triaJs three and four when cues were absent. Analysis of Response Time The mean response time was the next variable to be investigated. The analysis provides information on the differential effect of the various conditions on the speed of response regardless of the accuracy. The analysis is parallel to analyzing the overall rate of responding due to the fact that the number of responses under each condition was the same. The four way analysis of variance (Table 5) depicts a significant effect (pCOI) due to trials on the response time. Further investigation into this effect was conducted using a Duncan's New Multiple range test (Table 6). Tnis analysis shows that trials two and five result in significantly lower response times (higher rates of behavior) than trial one. Figure 3 portrays the relationship between the number of the trial and mean "^esconse time.

PAGE 39

30 TABLE 5 ANALYSIS OF VARIANCE OF RESPONSE TIME SOURCE OF VARIATION

PAGE 40

TABLE 6 DUNCAN'S NEW MULTIPLE RANGE TEST ON RESPONSE TIME FOR EACH TRIAL 31 TRIAL

PAGE 41

32 MEAN NUMBER OF SECONDS 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 TRIALS FIGURE 3 MEAN RESPONSE TIME FOR EACH TRIAL

PAGE 42

33 The effect of the number of trials on the response time cannot be adequately explained v/ithout taking into account the significant interaction (P^.Ol) between number of trials and length of delay of consequences. This interaction shows that on trial one the 15 minute delay condition resulted in a significantly higher mean response time tfian in a three hour delay condition on trial five. This difference is shown more emphatically in figure 4. As was present in the analysis of correct responses, a significant interaction (P<.01) was found between the length of delay and the presence or absence of response produced cues. The analysis of the six conditions using Duncan's New Multiple Range Test (Table 9) reveals the differences. The condition of minimum delay of consequence with response produced cues absent (condition A) and three hours delay of consequence with response produced cues present (condition F) resulted in significantly lower response times than the condition of 15 minutes delay of consequence with response produced cues absent (condition B) . This interaction is graphically portrayed (figure 5) for further clarification. The analysis of response time then points to the following, 1) That a 15 minute delay of consequence with response Droduced cues absent (condition B) results in significantly higher response time than when consequences are presented immediately with response produced cues absent (condition A) or when consequences are delayed 3 hours and response produced cues are present (condition F) ; 2) That trials two and five resulted in significantly lower response times than trial one. 3) That trial one in the 15 minute delay condition resulted in a significantly higher mean response time than trial five in the three hour delay condition.

PAGE 43

TABLE 7 MEANS AND STANDARD DEVIATIONS ON RESPONSE TIME FOR EACH TRIAL/DELAY CONDITION 34 LABEL

PAGE 44

35

PAGE 45

MEAN NUMBER OF SECONDS 36 38 37 36 35 34 33 32 31 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 TRIALS h -k 2 3 4 Minimum delay 15 minute delay 3 hour delay FIGURE 4 MEAN RESPONSE TIME FOR EACH TRIAL UNDER EACH DELAY CONDITION

PAGE 46

TABLE 9 DUNCAN'S NEW MULTIPLE RANGE TEST ON RESPONSE Tire IN EACH DELAY/ REPRODUCED CUE CONDITION 37 CONDITION

PAGE 47

MEAN NUMBER OF SECONDS 38 30 29 28 27 26 25 24 23 22 21 20 19 18 17 16 15 MINIMUM 15 MINUTES No Cues present -4i Cues present 3 HOURS FIGURE 5 MEAN RESPONSE TIME IN EACH CUE/DELAY CONDITION

PAGE 48

39 Analysis of Log of Rate of Correct Responding From the number of correct responses and the response time the rate of correct responding was calculated. The log 10 of these rates was then calculated. This provided the variable which is used in this analysis. By combining both time and frequency of responding this variable provides a more precise measure of the effect of the various conditions on the learning of the emotionally disturbed adolescents. The analysis of variance (Table 10) provides evidence of the increased precision of the measure. This is indicated by the fact that five significant differences were present in this analysis as opposed to three significant differences in eacii of the previous analyses. The number of trials had a significant effect (P^.Ol) on the rate of correct responding. The rate of correct responding on trial six is significantly higher than the rate of correct responding on trials one and three. The log of the rate of correct responding on trial three is in addition lower than on all other trials with the exception of trials one and four. Plotting of the data in figure 6 provides visual display of the trend of increasing rates of correct responding as the number of trials increases. The presence of response produced cues when the consequence was delivered resulted in a significantly higher (P<.01) mean rate than when the response produced cues were absent. The mean log of correct responses when response produced cues were present was .815 as opposed to a mean of .749 when they were absent. The length of the delay had a differentially significant effect (P<.05) on the rate of correct

PAGE 49

TABLE 10 ANALYSIS OF VARIANCE OF LOG OF RATE OF CORRECT RESPONDING 40 SOURCE OF

PAGE 50

TABLE n DUNCAN'S NEW MULTIPLE RANGE TEST OF LOG OF RATE OF CORRECT RESPONDING FOR EACH TRIAL 41 TRIAL

PAGE 51

42 LOG OF RATE OF CORRECT RESPONDING .86 .85 .84 .83 .82 .81 .80 .79 .78 .77 .76 .75 .74 .73 .72 .71 .70 .69 .68 .67 .66 .65 .64 .63 .62 TRIALS FIGURE 6 MEAN LOG OF RATE OF CORRECT RESPONDING ON EACH TRIAL

PAGE 52

43 responding. The Duncan's flew Multiple Range Test (Table 12) pointed out that a minimum delay in consequation produced a significantly higher log of rate of correct responding than the 15 minute delay of consequence. Figure 7 provides a visual display of the differences. In addition to the three significant main effects two significant interactions were found. Trials by cues (P .05) and cues by delay (F .01) proved significant. Analysis of the trials by cues interactions (Table 13) generated the conclusions that, 1) trial six with cues present resulted in a higher rate than trials three or four with cues absent; and 2) trial three with cues absent resulted in a lower rate than all other trial/cue conditions except trial four with cues absent. The graphic portrayal (Figure 8) of the means provides a very good picture of the interaction between number of trials and presence or absence of cues. In the plot it is readily apparent that rate of correct responding tends to steadily increase as the number of trials increase when response produced cues are present. However no definite trend either in increase or decrease is observed when response produced cues are absent. The analysis of the cues by delay interaction resulted in several significant differences (Table 14): Immediate delivery of consequence with response produced cues absent (condition A) resulted in a higher correct rate of responding than under conditions of a 15 minute or three hour delay cues absent (conditions B and C). A higher rate of correct responding was also present under conditions of 15 minute and 3 hour delays cues present (conditions E and F) than under the condition of a 15 minute delay of consequence cues absent. To point out clearly these differences the mean values are plotted in figure 9.

PAGE 53

TABLE 12 DUNCAN'S NEW MULTIPLE RANGE TEST OF LOG OF RATE OF CORRECT RESPONDING FOR EACH DELAY LENGTH 44 DELAY LENGTH

PAGE 54

45 LOG OF RATE OF CORRECT RESPONDING .86 .85 .84 .83 .82 .81 .80 .79 .78 .77 .76 .75 .74 .73 .72 .71 .79 .69 .68 .67 .66 .65 .64 .63 .62 DELAY MINIMUM 15 MINUTES 3 HOURS FIGURE 7 ME ^i LOG OF RATE OF CORRECT RESPONDING IN EACH DELAY CONDITION

PAGE 55

TABLE 13 DUNCAN'S NEW MULTIPLE RANGE TEST OF THE LOG OF RATE OF CORRECT RESPONDING FOR EACH TRIAL . AND CUE CONDITION 46 LABEL

PAGE 56

47 LOG OF CORRECT RESPONDING .89 .88 .87 .86 .85 .84 .83 .82 .81 .80 .79 .78 .77 .76 .75 .74 .73 .72 .71 .70 .69 .68 .67 .66 .55 .64 .63 .62 .61 .60 .59 .58 .57 . 56 .55 TRIALS FIGURE 8 MEAN LOG OF RATE OF CORRECT RESPONDING ON EACH TRIAL UNDER CUES ABSENT/CUES PRESENT CONDITIONS

PAGE 57

TABLE 14 QUNCAN'S NEW MULTIPLE RANGE TEST ON LOG OF RATE OF CORRECT RESPONDING IN EACH DELAY/RESPONSE PRODUCED CUE CONDITION 48 CONDITION

PAGE 58

LOG OF CORRECT RESPONDING 49 .88 .87 .86 .85 .84 .83 .82 .81 .80 .79 .78 .77 .76 .75 .74 .73 .72 .71 .70 .69 .68 .67 .56 .65 DELAY MINIMUM 15 MINUTES 3 HOURS -• No cues present Cues present FIGURE 9 MEAT^ LOG OF RATE OF CORRECT RESPONDING IN EACH CUE/DELAY CONDITION

PAGE 59

50 In summary the analysis of ttie log of rate of correct responding provides the following: 1) That the presentation of response produced cues results in higher rates of correct responding; , 2) That immediate consequentation resulted in higher rates of correct responding than a 15 minute delay in consequence; 3) That im.niediate delivery of consequence cues absent (condition A) resulted in a higher rate of correct responding than under conditions of a 15 minute or three hour delay cues absent (conditions B and C) . 4) That a higher rate of correct responding resulted from a 15 minute or 3 hour delay cues present (conditions E and F) than under the condition of a 15 minute delay of consequence cues absent. 5) That the rate of correct responding was lower on trial three than on all trials except one and four. 5) That trial six resulted in a higher rate of correct responding than on trials one and three. 7) That trial three with cues absent resulted in a lower rate than all other trial cue conditions except trial four with cues absent; and 8) That trial six with cues present resulted in a higher rate than trials three or four with cues absent. Analysis of Celeration Ratios A line of best fit was calculated for the log 10 of correct responding using linear regression. From this the predicted log 10 of the rate of correct and incorrect responding for each day was calculated. From this information the celeration coefficient for correct and incorrect was calculated. This is the number one must multiply the predicted rate of responding on day n to obtain the predicted rate of responding on day n + 7. The coefficients are presented in Table 15 and 16. Celeration ratios were then calculated by dividing the celeration coefficient of rate of correct responding for each subject

PAGE 60

TABLE 15 CELERAVION COEFFICIENTS FOR CORRECT RESPONDING 51

PAGE 61

TABLE 16 CELERATION COEFFICIENTS FOR INCORRECT RESPONDING 52

PAGE 62

53 in each condition by the respective celeration coefficient for rate of incorrect responding. The three way analysis of variance (Table 17) reveals one significant effect. The absence or presence of cues during consequation has a significant effect on the celeration ratio. The mean celeration ratio was significantly higher in the cues present condition than in the cues absent condition. This is in line with the analysis of number correct and rate of correct responding. Looking at the data on an individual student basis provides some additional information. A celeration ratio of over one indicates higher increase in rates of correct responding than incorrect responding, which is an indication that learning in the desired direction has occurred for thai subject in that condition. Analyzing the celeration ratios in this fashion indicates that learning took place for, 1) six subjects under condition A. 2) seven subjects under condition B. 3) two subjects under condition C. 4) eight subjects under condition D. 5) five subjects under condition E. 6) ten subjects under condition F. This indicates more subjects learned under condition F than under any other condition.

PAGE 63

54 TABLE 17 ANALYSIS OF VARIANCE OF CELERATION RATIOS SOURCE OF VARIATION

PAGE 64

TABLE 18 CELERATION RATIOS 55

PAGE 65

56 Summary of Results This section summarizes the foregoing analysis in terms of the null hypotheses stated on page 11. 1) Analysis of correct responding indicated no significant effect due to length of delay taken by itself. So null hypotheses one cannot be rejected. 2) Analysis of response time resulted in no significant effect due to length of delay. Therefore null hypotheses two cannot be rejected. 3) Analysis of rate of correct responding resulted in a significant effect due to length of delay. Therefore null hypotheses three is rejected. There is an effect on rate of correct responding due to the length of the delay. The minimum delay of consequence resulted in a higher correct rate of responding than a 15-minute delay of consequence. 4) Analysis of the celeration ratios resulted in no significant effects due to length of delay. Therefore null hypotheses four cannot be rejected. 5) Analysis of correct responses resulted in a significant effect due to presence or absence of response produced cues. Thus null hypotheses five is rejected. The number of correct responses is effected by presence or absence of cues. The presence of response produced cues during consequation results in higher numbers of correct responses. 5) Analysis of response time resulted no significant effects due to cue presence or absence. Therefore null hypotheses six cannot be rejected. /^ 7) Analysis of rate of correct responding resulted in ' a significant effect due to cue presence or absence. Thusnull hypothesis seven is rejected. The rate of correct responding is effected by presence or absence of cues. The presence or response produced cues during consequention results in higher rates of correct responding.

PAGE 66

57 8) Analysis of the celeration ratios resulted in a significant effect due to presence or absence of response produced cues. Therefore hypotheses eight is rejected. The presence of response produced cues during consequation results in higher celeration ratios. 9) Analysis of the number of correct responses resulted in a significant effect due to the interaction of delay length and cue condition. Thus null hypotheses nine is rejected. The interaction is due to the fact that the number of correct responses with a three hour delay-cues present (condition F) was higher than every other combination except a minimum delay of consequence-cues absent (condition A). 10) Analysis of the response time resulted in a significant effect due to the interaction of delay length and cue condition. Thus null hypotheses ten is rejected. The interaction is due to a lower response time when consequences are presented immediately and response produced cues are absent (condition A) and when consequences are delayed three hours and response produced cues are present (condition F) than when consequences are delayed 15 minutes and response produced cues are absent. (Condition B). j^l) Analysis of rate of correct responding resulted in a significant effect due to the interaction of delay length and cue condition. Thus null hypotheses 11 is rejected. This interaction is due to, 1) a higher rate of correct responding with immediate delivery of consequence-cues absent (condition A) than under conditions of a 1 5 minute or three hour delay cues absent (conditions B and C); and 2) a higher rate of correct responding with a 15 minute or 3 hour delay of consequence-cues present (conditions E and F) than with a 15 minute delay of consequence-cues absent. 12) Analysis of celeration ratios resulted in no significant interaction due to delay length and cue condition combination. Therefore hypotheses 12 is not rejected. Table 19 presents a summary of significances. This indicates that 7 of the null hypothesis were rejected, 5 of the 7 at the .01 level and 2 at the .05 level .

PAGE 67

TABLE 19 SUmARY OF SIGNIFICANT EFFECTS 58 VARIABLES

PAGE 68

CHAPTER V CONCLUSIONS This study was undertaken to investigate the effects of delays of consequence on the learning of emotionally disturbed adolescents. The effects of the delays were investigated in situations where response produced cues were presented at the time the consequence was delivered and where response product^d cues were absent during consequation. The major questions were whether: 1) delays in the delivery of consequences of 15 minutes or 3 hours have detrimental effects on the learning of emotionally disturbed adolescents v.hen no stimulus or response produced cues are present; 2) delays in the delivery of consequences of 15 minutes or 3 hours have detrimental effects on the learning of emotionally disturbed adolescents when stimulus and response produced cues are present. The analysis of the results of the effects on four dependent measures indicates that: 1) Delays of 15 minutes and 3 hours have a detrimental effect on the rate of correct responding when no response produced cues are present. 2) A delay of 15 minutes has a detrimental effect on the response time when no response produced cues are present. 3) Delays of 15 minutes or 3 hours have no detrimental effect on number of correct responses or celeration ratios. 4) Delays of 15 minutes and 3 hours have no detrimental effect on any of the four dependent measures when response produced cuss are present. In fact performance was superior under conditions of 3 hour delay as measured by number of correct responses . 59

PAGE 69

60 It therefore appears that the detrimental effects of delays of reinforcement can be overcome by the presentation of stimulus and response produced cues at the time the consequence is delivered. This means that modification techniques used with emotionally disturbed students may successfully employ techniques not centered around immediate delivery of consequences, resulting in the design and adoption of techniques which are not as tedious as the present ones and which may be utilized more readily without tremendous outside support. These techniques however should be limited to behaviors in which stimulus and response produced cues can be presented during consequation. Several factors are deserving of further investigation. The most unexpected result was the higher mean number of correct responses under a three hour delay than a delay of 15 minutes or the immediate presentation of the consequence when cues were present. This effect may be due to the fact that a smaller time period passed between the time the cues were presented and the next trial (approximately 20 hours as opposed to 23 hours). It may also be due to the fact that due to the school's schedule academic work followed the presentation of cues in all the conditions except the three hour delay. Thus the academic work may have interfered with remembering the correct responses. Both these possibilities should be investigated. It would be desirable to investigate the effect of delays in other populations of emotionally disturbed adolescents, with different learning tasks, stronger consequences and different time lengths. There is enough evidence from this study and two applied studies (Schwartz and Hawkins, 1970; Sluyter and Hawkins, 1971) to show teachers of

PAGE 70

61 emotionally disturbed adolescents that less tedious modification procedures centered around delayed consequences can and should be employed effecti'*ely.

PAGE 71

APPENDIX 62

PAGE 72

63 ^ / / y / X

PAGE 73

64

PAGE 74

65 K A

PAGE 75

66

PAGE 76

APPENDIX 67

PAGE 77

68 V. in S.^ CQ ^^ o UJ CQ CJ3 O co CO «^ c\j CO U3 «^ CO CO UD <* CVJ 00 CM 00 (£) CO +-> O CLi Ss_ o CD ro -M U O) So o 4-> 1 1 o o o o O Lf) o o o tn o o o 3inNIW y3d S1N3W3A0W

PAGE 78

69 <^ <-i ^U CO CO CSJ 00 CO W3 CO <^ cv CO CM 00 CO C3 CO o ex oo

PAGE 79

70 u. LU < e t ^ J^^ 00 CVJ 00 evi CO vo oo CO CNJ CO CM 00 CVJ CO o -3 CO o OO o o o o o tn o o O LO O LO r— IT) LT) O rLO O O o CD O OO • >-

PAGE 80

71 CO CM LU 00 CVJ V ^ CQ C_3 C3 CO O 00 00 <^ CM 00 CM GO CM o o o o o o O Ln O LD I— LO o r — LO o o o o o tn • ><: o

PAGE 81

72 'W <.=^ (-> CQ >*-* % •" "*^ t_) C3 -3 CQ o 00 00 CM 00 lo Od 00 'aCM 00 CM 00 ID «=f CM 00 CM

PAGE 82

73 S X CD ^ .^=-^ VO t_) o CO CM CO CM 00 lO C\J OD CM CO
PAGE 83

74 J r ''^^ (_> OQ O C_) LiJ o Q. CO >=300 +-> u a; s_ so u E OJ (13 O O o o O Lf) o o o uia Ln o 1 — LD o o o o O 1/1 >o HinNIl-J y3d S1N3W3A0W

PAGE 84

75 St ^^23v> CO t \ ^ CO o CO o 00 00 kO CM 00 CM CX3 \D CM CO CM CO •acsj 00 lO CVJ

PAGE 85

76 CQ ai o CQ o a. 00 lo CO IX) C\J 00 U3 ^^ CM CO CM CO CVI 00 CM (J q; ss_ o <_>
PAGE 86

77 <^ ^ o cn o co CVJ 00 00 00 CO CM CO CM o o o o O Lf) O D o o o 1 — Ln o o o o O oo • >-

PAGE 87

BIBLIOGRAPHY Allen, E., Hart, B., Buell, J., Harris, F., and Wolk, M. "Effects of Social Behavior on Isolate Behavior of a Nursery School Child," Child Development , 1964, 35, 511-518. Becker, W., Madsen, C, Arnold, C. and Thomas, D. "The Contingent Use of Teacher Attention and Praise in Reducing Classroom Behavior Problems," Journal of Special Education , 1967, 1, 287-307. Bilodeau, E. A. and Bilodeau, I. M. "Variation of Temporal Intervals Among Critical Events in Five Studies of Knowledge of Results," Journal of Experimental Psychology , 1958, 55, 603-612. Brackbill, Y., and Kappy, M. S. "Delay of Reinforcement and Retention," Journal of Comparative Physiological Psychology , 1962, 20, 433-434. Carlson, C, Arnold, C, Becker, W. and fladsen, C. "The Elimination of Tantrum Behavior of a Child in an Elementary Classroom," Behavior Research and Therapy , 1968, 6, 117-119. Denny, M. R., Allard, M., Hall, E. and Rokeach, M. "Supplementary Report: Delay of Knowledge of Results, Knowledge of the Task and Intertrial Interval," Journal of Experimental Psychology , 1950, 60, 327. Erickson, M. and Lipsitt, L. P. "Effect of Delayed Reward on Simultaneous and Successive Discrimination Learning in Children," Journal of Comparative Physiology , 1960, 53, 256-260. Glavin, J. P., Quay, H. C, Annesley, F. R., and Werry, J. S. "An Experimental Resource Room for Behavior Problem Children," Exceptional Children , 1971, 38, 131-137. Goldstein, S. B., and Siegel, A. W. "Observing Behavior and Children's Learning," Child Development , 1971, 42, 1608-1613. Grubard, P. S. "Utilizing the Group in Teaching Disturbed Delinquents to Learn," Exceptional Child , 1969, 36, 267-72. Hanley, E. M. "Review of Research Involving Applied Behavioral Analysis in the Classroom," Review of Educational Research , 1970, 40, 597627. 78

PAGE 88

79 Hewett, F. M. , Taylor, F. D., and Artuso, A. A. "The Santa Monia Project: Evaluation of an Engineered Classroom Design with Emotionally Disturbed Children," Exceptional Child , 1969, 35 523-529. Li, J. C. Statistical Inference I , Edward Brothers, Inc., Ann Arbor Michigan, li964. Lovitt, T. "Behavior Modification: The Current Scene," Exceptional Children , T970, 37, 85-93. Mischel, W. "Preference for Delayed Reinforcement: An Experimental Study of a Cultural Observation," Journal of Abnormal and Social Psychology , 1958, 56, 57-51. Mischel, W. "Preference for Delayed Reinforcement and Social Responsibility," Journal of Abnormal and Social Psychology , 1961, 62, 1-7. Mischel, W. and Tletzner, R. "Preference for Delayed Reward as a Function of Age Intelligence and Length of Delay Interval," Journal of Abnormal and Social Psychology , 1962, 64, 425-431. Morris, S. M. The Effects of Reward Value and Delay of Reward on the Acquisition and Retention of a Simultaneous Discrimination Task \ with Emotionally Disturbed Children. Dissertation Abstracts, 70-1299, U'r.iversity of Connecticut, 1969. Mower, 0. H. and Ullman, A. D. "Time as a Determinant in Integrative Learning," Psychological Review, 1945, 52, 61-90. O'Leary, K. and Becker, W. "Behavior Modification of an Adjustment Class: A Token Reinforcement Program," Exceptional Children , ^ 1967, 37, 537-642. Piper, T. J. "Effects of Delay of Reinforcement on Retarded Children's Learning," Excep tional Children , 1971, 38, 139-145. Quay, H. Sprague, R., Werry, J. and McQueen, M. "Conditioning Visual Orientation of Conduct Problem Children in the Classroom," Journal of Experimental Child Psychology . 1966, 5, 512-517. Renner, K. E. "Delay of Reinforcement: A Historical Review," Psychological Bulletin , 1964, 61, 341-361. Ross, L. E., Hetherington, M. , and Wray, N. P. "Delay of Reward and the Learning of a Size Problem by Normal and Retarded Children," Child Development . 1965, 36. 509-517. Schoelkopf, A. M. and Orlando, R. "Delayed Versus Immediate Reinforcement in Simultaneous Discrimination Problems with Mentally Retarded Children," Psychological Record , 1965, 16, 15-23.

PAGE 89

80 Schoelkopf, A. M. and Orlando, R. "Reinforcement Delay Gradients of Retardates with a Current Discrimination Task Procedure," Psychological Record , 1966, 16, 113-128. Schwartz, M. L. and Hawkins, R. P. "Application of Delayed Reinforcement Procedures to the Behavior of an Elementary School Child," J ournal of Applied Behavior Analysis , 1970, 3, 85-96. Skypeck, G, A Descriptive Study of Three Methods of Programming Flash Cards . Doctoral Dissertation , University of Florida, 1971. Sluyter, D. J. and Hawkins, R. P. "Delayed Reinforcement of Classroom Behavior by Parents," Journal of Learning Disabilities , 1972, 5 20-21. Ward, L. C. and Baumeister, A. A. "Effects of Temporal Variables, Monetary Incentive, and Method of Information Feedback on the Paired-Associate Learning of Retardates," American Journal of Mental Deficiency , 1971, 75, 712-718. Watson, J. B. "The Effect of Delayed Feeding Upon Reaction," Psychobiology , 1917, 1, 51-60.

PAGE 90

BIOGRAPHICAL SKETCH Robert Keith Brown was born October 22, 1946, at Winter Haven, Florida and is the second of six children of Mr. and Mrs. R. McDonald Brown. He attended elementary school at Winter Haven, Florida and •yjas a member of the last graduating class of St. Leo College rreparatory School, St. Leo, Florida in 1964. In September, 1964 he entered the University of Florida and in Aiicust, 1968 received the Bachelor of Arts Degree, majoring in ;}'=;ychology. He then enrolled in the graduate program in special education at the University of Florida. Mr. Brov'jn received his Master c*-' Education Degree, majoring in teaching emotionally disturbed chi'^d-'en, in 1970. He v/orked full time for Dr. Robert S. Soar at the Institute for the Development of Human Resources as a research assistant for 6 months before ccntinuinc his work toward a Doctorate of Philosophy in Special Education. Since June, 1971 he has been employed by the Duval County School Board in the position of Coordinator of Research and Evaluation for Exceptional Child Education. Mr. Brown has made a number of presentations on topics ranging froii". accountability to sensitivity training at state and National Conventions. He has three major publications in the area of emotionally disturbed children and is presently engaged in developing evaluative technique for the trainable mentally retarded. 81

PAGE 91

82 Mr. Brown holds memberships in the Council for Exceptional Children, the American Educational Research Association and Phi Kappa Phi . Mr. Brown is divorced and has one lovely daughter Kelly Christine Brown.

PAGE 92

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. William R.'Reid, Chairman Professor of 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. Assistant Professor of Educatio 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. "^ Henry Ji?ud in Assistant Professor of Education

PAGE 93

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. Henry S. Pennypacker Professor of Psychology This dissertation was submitted to the Dean of the College of Education and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. June, 1973 < ?^;c'/d./^u^ Dean, College f Education Dean, Graduate School

PAGE 94

^7 S M2. 2. 213 4 5