ABSTRACT

CONSTRUCT VALIDATION OF THE MOTIVE FOR SUCCESS AS
INTERPRETED BY GSCI TEST SCORES AND AS EXPRESSED
IN BEHAVIOR AND ACHIEVEMENT WITHIN THE CLASSROOM

by Wayne H. Chubb

The study was an investigation of the construct (nomo—
logical) validity of the Motive for Success as, (a) repre-
sented within Atkinson's theory of achievement motivation,
(b) interpreted by scores on the Generalized Situational.
Choice Inventory (GSCI) and, (c) eXpressed in behavior and
achievement in the classroom situation.

Atkinson’s statements concerning the empirical problem
of motivation and Farquhar's definition of the role of
motivation within the academic situation were combined to
select behavioral criteria for the expression of motivation

in the classroom——the initiating, directing and sustaining

 

 

 

of achievement—related activity.

Students in ninth grade English classes were presented
with both an achievement task (Theme Writing) and a non-
achievement task (Joke Rating). Instructions emphasized
the importance of the theme (grade counted toward final
course grade) and the "spare-time" value of rating the Jokes.
A time limit of 20 minutes was set and all students were

required to read and rate "some" of the Jokes before starting

Wayne H. Chubb

to write and, if they chose, to re-write (revise) a theme

concerning "the importance of doing well in English class."
Support was found for the nomological validity of

the Motive for Success through relationships (academic

ability held constant by partial correlation) between

GSCI scores and both the directing criterion (Expected

 

Grade), r = .21U (p < .01), and the sustaining criterion

 

(time spent on the achievement task), r = .170 (p < .05),
of motivation for 252 males.
The relationship between academic ability (DAT-V

scores) and the initiating criterion (time spent on the non-

 

achievement task before starting the achievement task),
r = —.19U (p < .01), suggested that providing an alterna-
tive activity for males allowed expression of tendencies to
"avoid failure."
Support was also found for the nomological validity
‘of the Motive for Success through confirmation of a pre-
dicted stronger relationship between GSCI test scores and
teacher-assigned grades for a subgroup of males assumed to
be performing school tasks of an "intermediate" level of
difficulty, partial r = .417 (p < .01), than for subgroups
of males performing easier, partial r = .100 (p > .05),
or more difficult, partial r = .151 (p > .05), school tasks.
Findings with females indicated that achievement-
related behaviors may differ for males and females and
require future studies of motivation in the classroom to

analyze results separately for males and females.

CONSTRUCT VALIDATION OF THE MOTIVE FOR SUCCESS AS
INTERPRETED BY GSCI TEST SCORES AND AS EXPRESSED

IN BEHAVIOR AND ACHIEVEMENT WITHIN THE CLASSROOM

By

Wayne H. Chubb

A THESIS

Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Counseling, Personnel Services,
and Educational Psychology

College of Education

1968

ACKNOWLEDGMENTS

Throughout the preparation of the study, the assis—
tance and encouragement of the writer's major advisor,
Dr. William W. Farquhar, was most helpful. In addition
the members of the committee, Dr. Walter F. Johnson, Dr.
Bill L. Kell and Dr. Norman Abeles were always available
for consultation.

Appreciation is also extended to my colleagues at
Wayne State University and particular gratitude is due
Dr. Stanford Glazer, Director, Counseling and Testing
Bureau, Wayne State University.

Finally, a sincere expression of gratitude is
extended to my wife, Marie, and my daughter, Jo Ann, for

their understanding and patience.

ii

TABLE OF CONTENTS

ACKNOWLEDGMENTS . . . . . . . . . .
LIST OF TABLES . . . . . . . . . . . .
LIST OF APPENDICES . . . . . . . . . .
Chapter

I INTRODUCTION TO THE PROBLEM . . . . .

II

Two Problems of Motive Assessment .

The Problem of Motivation . . . . .

Need for the Present Study . . . . .

Construct Validation . . . .

Construct (Trait) Validity of the M-Scales

Construct (Nomological) Validity of the
Generalized Situational Choice

Inventory (GSCI) o o o ' o o o o .
Purpose of the Present Study . . . .
Theory . . . . . . . . . . .

Prediction of Behavior in the Classroom
Situation . .

Hypotheses for Stage I--Nomological Validity
Hypotheses for Stage II--Criterion Validity

Hypotheses for Stage III--Predictive
and Nomological Validity . . ., .
Overview of the Study . . . . . .

REVIEW OF THE LITERATURE . . . .

Achievement- Related Motives and the
Initiation of Behavior . . .
Achievement Related Motives and the
Direction of Behavior . . . . ..
Persistence as a Personality Trait . .
Persistence as a Motivational Phenomenon
in the Laboratory . .> . . . . .

iii

Page

ii
viii

xii

29

35
A2,

A7

Chapter

Strengths and Weaknesses of Laboratory
Studies of Persistence . . . .

Persistence as a Motivational Phenomenon

in the Classroom . . ._ . . .
Strengths and Weaknesses of Studies of

Persistence in the Classroom . .
Studies of Attrition in the

Academic Situation . . .
Similarities and Differences in the

Study of Persistence . . . .

The Measurement of Individual Differences

in Achievement Related Motives .
Subjects . . . . . . . . . .
The Situation . . . . . . . .

Apparent Difficulty. . . . . .
Group and Individual Performance .
Novelty of the Task . . .' . .
Knowledge of Results . . .' . .
The Alternate Activity . .
Moderator Variables and the Prediction

of Achievement . . . . . . .

Summary . . . . .' . . . . .

III DESIGN OF THE STUDY . . . o . . .

sample 0 o o o o o o o o ' 0
Test Instruments .

Differential Aptitude Test--Verbal Scale
Generalized Situational Choice Inventory

Research Instruments . . . . . .
Non-Achievement Task . . .
Achievement Task . . . .
Achievement Questionnaire .

Instructions . . . . .~ .

Time Measurement Procedure . .
Pilot Study Modifications .

Motivation Criteria . .
Latency (Initiating Behavior)
Persistence (Sustaining Behavior) .
Expected Grades (Directing Behavior)

Indices of Academic Achievement Level
Level of Achievement on the Academic

Level of Achievement in the

Academic Situation . .

0 0 O 9 O 0 O

0 0 O O O 0 9 0

Two Indices of "Apparent Task Difficulty"
The Formation of Groups Assumed to Vary in
Level of "Apparent Task Difficulty:

iv

0

O
O
0
O
O
0
O
O
O
6
O
T

x. o o o o o o o o o o o

0 O O 0 O 0 O O O 0 O 0 O O O O 0

Page

57
58
63
64
67

69
72
7M
75
76
77

80

82
89

92
93

94
9H
95
95
96
97

98
100
100
100
101
101
101
101

102
103

103

Chapter Page

Reliability of the Generalized Situational
Choice Inventory (GSCI) . . . . . . lON
Reliability Estimates for Three Indices

for Academic Achievement . . . . . . 106
Theme Grade . . . . . . . . . . . 107
Semester English Grade . . . . . . . 107
Cumulative GPA . . . . . . . . . . 107

Analysis . . . . . . . . . . . . . 109
Research Hypotheses . . . . . . . 112
Stage I--Nomologica1 Validity . . . . . 112
Stage II--Criterion Validity . . . . . 113
Stage III--Predictive and
Nomological Validity . . . . . . . 115
Summary . . . . . . . . . . . . . 116
IV ANALYSIS OF RESULTS . . . . . . . .- . . 120

Stage I--Nomologica1 Validity . . . . . . 120

The Strength and Direction of Relation-

ships Stated in Hypotheses 1, 2 and 3 . . 121
The Relative Strength and Direction of

Relationships Stated in Hypotheses

AA, 5A and 6A . .. . . . . . . 126
The Relative Strength and Direction

of Relationships Stated in

Hypotheses AB, 5B and 6B . . . . . . 133

Stage II--Criterion Validity . . . . . 141
The Strength and Direction of Relation—

ships Stated in Hypotheses 7, 8 and 9 . . 1A1

The Relative Strength and Direction

of Relationships Stated in

Hypotheses 10A, 11A and 12A . . . . . 1A8
The Relative Strength and Direction

of Relationships Stated in

Hypotheses 10B, 11B and 12B . . . . . 155
Stage III--Predictive and
Nomological Validity . . . . . 162

Strength and Direction of the Relation-

ship Stated in Hypothesis l3 . . . . . 162
The Relative Strength and Direction

of Relationships Stated in Hypothesis 1“ . 166
The Strength and Direction of Relation-

ships Stated in Hypothesis IAB . . . . 173
Summary of the Analysis of Results . . . 179
Results within Stage I Analyses . . . . 179
Results within Stage II Analyses . . . 182
Results within Stage III Analyses . . . . 185

Chapter
V DISCUSSION OF RESULTS . . . . . . . .

Stage I--Nomologica1 Validity . . . . .
Discussion of Relationships Between GSCI
Scores and Three Criteria of
Motivation for Males . . .- . . .
Discussion of Relationships Between GSCI
Scores and Three Criteria of
Motivation for Females . .

Discussion of Relationships of GSCI Scores

with Three Criteria of Motivation for

Subgroups of Males or Females Varying in

Objective or Subjective Estimates of
the Difficulty of School Success . .
Stage II--Criterion Validity . . . . .
Discussion of Relationships for the Three
Behavioral Criteria of Motivation with
Indices of Academic Achievement
for Males . . . . . .
Discussion of Relationships for the Three
Behavioral Criteria of Motivation with
Indices of Academic-Achievement
for Females . . . . .
Discussion of Relationships for the Three
Behavioral Criteria of Motivation with
Indices of Academic Achievement for
Subgroups of Males or Females Varying
in Objective or Subjective Estimates of
the Difficulty of School Success . .
Stage III-—Predictive and Nomological
Validity . . . . .

Discussion of Relationships for GSCI Scores

for Males, and GSCI Scores for Females,
with Indices of Academic Achievement .

Discussion of Relationships for GSCI Scores
with Indices of Academic Achievement for

Subgroups of Males Varying in Objective

or Subjective Estimates of the Difficulty

of School Success . . .

Discussion of Relationships for GSCI Scores
with Indices of Academic Achievement for

Subgroups of Females Varying in
Objective or Subjective Estimates of
the Difficulty of School Success . .

vi

Page
188
188

189

197

199
205

207

213

215

218

220

222

227

Chapter Page

VI SUMMARY, CONCLUSIONS AND IMPLICATIONS . . . . 231

Summary . . . . . . . . . . . . . 231

The Problem . . . . . . . . . . . 231

Theory and Instrumentation . . . . . . 232

Sample and Procedure . . . . . . 232

Design and Analysis . . . . . o 234

Findings for Males . . . o . . . . o 236

Findings for Females . . . . . . . . 2A1
Conclusions . . . . . . . . . . . .‘ 2AA
Implications . . . . . . . . . . . 2A9
BIBLIOGRAPHY . o o . o o o . . . . . . o 252

APPENDICES O 0 0 0 0 0 O 0 0 0 0 O 0 0 0 258

vii

LIST OF TABLES

Table Page

1.1 Motivation (eXpressed in behavior) as a joint
function of the Motive for Success (Ms),
EXpectancy of Success (Ps) and Incentive
Value of Success (Is) for Individuals in
Whom Ms = l and Ms = 2 . . . . . . . . 20

3.1 Kuder—Richardson estimates (Formula #20)of
reliability for male and female forms of
the Generalized Situational Choice Inventory
(GSCI) for the total samples of subjects and
for each of the subgroups formed by tricho-
tomizing the distributions of DAT-V scores
and Difficulty Ratings of school success . . 105

3.2 Reliability estimates for the "Theme Grade,"
"Semester English Grade" and "Cumulative
GPA" indices of level of academic achievement
obtained by correlating Theme Grade with
Current Semester English Grade, Current
Semester English Grade with previous semester
English grade and current year grade averages
with previous year grade averages within
total samples of male and female students
and within each of the subgroups formed by
trichotomizing the separate distributions
of DAT- V scores and Rated Difficulty of
school succeess. . . . . . . . . . 108

“.1 Simple and partial correlation (DAT-V scores
controlled) of the Generalized Situational
Choice Inventory Test scores with each of
three theory-selected criteria of motiva-
tion for 252 junior high school males . . . 123

“.2 Simple and partial coefficients of correlation
(DAT-V scores controlled) of the Generalized
Situational Choice Inventory with each of
three theory-selected criteria of motivation
for 259 Junior high school females . . . . 125

4.3 Simple and partial coefficients of correlation
(DAT-V scores controlled) of the Generalized
Situational Choice Inventory Test scores with
each of three theory-selected criteria of
motivation at each of three levels of ability
(DAT- V scores) for 252 junior high school
males . . . . . . . 130

viii

Table
4.4

4.5

4.6

4.7

4.8

4.9

4.10

Simple and partial coefficients of correlation
(DAT-V scores controlled) of the Generalized
Situational Choice Inventory Test scores
with each of three theory-selected criteria
of motivation at each of three levels of
ability (DAT- V scores) for 259 junior high
school females . .

Simple and partial correlation coefficients
(DAT-V scores controlled) of the Generalized
Situational Choice Inventory Test scores with
each of three theory-selected criteria of
motivation at each of three levels of the
rated difficulty of school achievement for
252 junior high school males . . .

Simple and partial correlation coefficients
(DAT-V scores controlled) of the Generalized
Situational Choice Inventory Test scores with
each of three theory-selected criteria of
motivation at each of three levels of the
rated difficulty of school achievement for
259 junior high school females. .

Simple and partial correlation coefficients
(DAT-V scores controlled) of three theory-
selected criteria of motivation with two
indices of academic achievement for 252
junior high school males.

Simple and partial correlation coefficients
(DAT-V scores controlled) of three theory-
selected criteria of motivation with each of
two indices of academic achievement for
259 junior high school females.

Simple and partial correlation coefficients
(DAT-V scores controlled) of three theory-
selected criteria of motivation with each of
two indices of academic achievement within
each of three levels of ability (DAT- V scores)
for 252 junior high school males .

Simple and partial correlation coefficients
(DAT-V scores controlled) of three theory-
selected criteria of motivation with each of
two indices of academic achievement within
each of three levels of ability (DAT-V scores)
for 259 junior high school females . .

ix

Page

132

137

139

144

147

151

154

Table Page

4.11 Simple and partial correlation coefficients
(DAT-V scores controlled) of three theory-
selected criteria of motivation with each
of two indices of academic achievement
within each of three levels of ability
(DAT—V scores) for 252 junior high school

males 159

4.12 Simple and partial correlation coefficients
(DAT—V scores controlled) of three theory-
selected criteria of motivation with each of
two indices of academic achievement within
each of three levels of rated difficulty of
school success for 259 junior high school 6
. . . . . . . . . . . l 1

females

4.13 Simple and partial correlation coefficients
(DAT-V scores controlled) of the Generalized
Situational Choice Inventory (GSCI) and the
Differential Aptitude Test--Verba1 Scale
with each of two indices of academic achieve-
ment for 252 junior high school males . 165

4.14 Simple and partial correlation coefficients
(DAT-V scores controlled) of the Generalized
Situational Choice Inventory (GSCI) and the
Differential Aptitude Test--Verbal Scale with
each of two indices of academic achievement

for 259 junior high school females 167

4.15 Simple and partial correlation coefficients
(DAT‘V scores controlled) of the Generalized
Situational Choice Inventory (GSCI) and the
Differential Aptitude Test--Verbal Scale
with each of two indices of academic achieve-
ment within each of three levels of ability
(DAT- V scores) for 254 junior high school

males 170

4.16 Simple and partial correlation coefficients
(DAT-V scores controlled) of the Generalized
Situational Choice Inventory (GSCI) and the
Differential Aptitude Test-~Verbal Scale with
each of two indices of academic achievement
within each of three levels of ability (DAT—V
scores) for 259 Junior high school females. . 172

4.17 Simple and partial correlation coefficients
(DAT-V scores controlled) of the Generalized
Situational Choice Inventory (GSCI) and the

Table

Page

Differential Aptitude Test-—Verbal Scale

(DAT-V) with each of two indices of academic
achievement within each of three levels of

the rated difficulty of school success for

254 junior high school males . . . . . . 176

4.18 Simple and partial correlation coefficients

(DAT-V scores controlled) of the Generalized
Situational Choice Inventory (GSCI) and the
Differential Aptitude Test-—Verba1 Scale

(DAT-V) with each of two indices of academic
achievement within each of three levels of

the rated difficulty of school success for

259 junior high school females. . . . . . 178

xi

Appendix
A.

B
C
D.
E

LIST OF APPENDICES

Jokes (Sample).

Paragraph — Essay (Male)

Achievement Questionnaire

Instructions

Means and Standard Deviations of the
Variables for 252 Junior High School
Males. . . . . . . .

Means and Standard Deviations of the

Variables for 259 Junior High School
Females . . . .

xii

Page
259
260
261

262

268

269

CHAPTER I
INTRODUCTION TO THE PROBLEM

The problem of motivation within the academic situation
has been studied by Farquhar.l His basic assumption was that
theory and instrumentation developed without specific regard
for the academic situation were inadequate to the task of
prediction and explanation of academic achievement.

To provide some clarity to the direction of his
enquiry, Farquhar defined academic motivation as "a combin—
ation of forces which initiate, direct and sustain behavior
toward a scholarly goal."2 Based on empirical evidence from
previous studies and on theory from research on the achieve-
ment motive2 Farquhar constructed a battery of objectively
scorable instruments (M-Scales) to reflect motivational

factors underlying school achievement. One sub-scale, the

 

1William W. Farquhar, Motivation Factors Related to

Academic Achievement, Cooperative Research Project 846
(East Lansing, Michigan: Michigan State University, Office
of Research and Publication, 1963).

2Ibid., p. 3.

 

3D. McClelland and J. Atkinson, et a1., The Achieve-
ment Motive (New York: Appleton—Century-Crofts, 1953).

 

Generalized Situational Choice Inventory (GSCI), in par-
ticular was directed towards the academic achievement motive
construct.

Initial research with these instruments indicated
scores on the M-Scales contribute to ability-based predic-
tions of academic achievement. Factor analytic study of
those sub-scales constructed on the basis of theory, pro-
duced factors interpretable as dimensions of the achieve-
ment motive construct. Subsequent studies have dealt with
concurrent validity,3 "factor validity"u and predictive

validity5 of the M—Scales.

Two Problems of Motive Assessment

 

There are two areas of concern about the M-Scale
validities as evidence that the instrument reflects the
operation of an underlying motivation construct.

The first area of concern is discussed by McClelland6

and directly relates to the procedure used by Farquhar in

 

3David A. Payne, "The Concurrent and Predictive
Validity of an Objective Measure of Academic Self-Concept,"
Educational and Psychological Measurement, 22 (1962),
773-780.

“Marion D. Thorpe, "The Factored Dimensions of an
Objective Inventory of Academic Motivation Based on Eleventh

 

Grade Male Over- and Under-achievers." (Unpublished doc-
toral dissertation, Michigan State University, 1961).
5

Robert L. Green, "The Predictive Efficiency and
Factored Dimensions of the Michigan State M-Scales for
Eleventh Grade Negro Students, An Exploratory Study."
(Unpublished doctoral dissertation, Michigan State Uni-
versity, 1962).

6John W. Atkinson (ed), Motives in Fantasy Action and
Society (D. Van Nostrand Co., Inc., 1958).

 

the development of the M-Scales. McClelland describes the
method of working with differences in motivation through
the location of two groups of subjects that differ in some
conspicuous way and treating them as representing different
stages of aroused motivation. McClelland states that, as
in the case of over- and under-achievers assumed to differ
in levels of achievement motivation, "over-achievers may
differ from under-achievers in other ways than in achieve-
ment motivation, and these differences may be picked up
in the measure."7
That the M-Scales do reflect differences other than
those attributable to motivation is evident through sub-
stantial correlation with scholastic ability and with group
differences in demographic characteristics.8 Attempts to
minimize the influence of scholastic ability on M-Scale
scores through revision of the scales, has resulted in
unacceptable levels of scale reliability.9
The second area of concern about the M-Scale vali-

dities as evidence that the instrument reflects the oper-

ation of an underlying motivation construct, relates to the

 

71bid., p. 14.
8Farquhar, op. cit.

9Virgil B. Sterling, "A Pilot Factor Analytic Study
of Academic Motivation and Achievement Levels in Eleventh
Grade Male Students." (Unpublished doctoral dissertation,
MIchigan State University, East Lansing, Michigan, 1962).

use of achievement criteria in both the construction stages
and validation stages of research on that instrument.

The development of the M—Scales proceeded on the
assumption that over-achievers and under-achievers could be
used to define the extremes of a motivation continuum.lO

Level of achievement was, therefore, a part of the
original item selection technique. Subsequent findings
of relationships between M-Scale scores and teacher-
assigned grades contribute to the evidence for the predic-
tive validity of the instrument but add little to the
clarification of its construct validity base.

It appears that confusion may result if instruments
purporting to measure scholastic ability and those
attempting to measure motivation both use the same achieve—
ment criterion (teacher-assigned grades) for validation of
these conceptually distinct constructs. Additional con-
fusion appears likely when achievement criteria are used
in both the construction stage of an instrument thought to
reflect motivation, e.g., the M-Scales, and in the valida-
tion stage.

Therefore, an attempt is reported in the following
section to select criteria of motivation, other than teacher-

assigned grades, by attention to the statements of Atkinson

 

loFarquhar, op. cit.

concerning the empirical and theoretical problem for the
study of motivation and to Farquhar's definition of the role

of motivation in the academic situation.

The Problem of Motivation

 

The empirical problem of motivation as stated by
Atkinsonll is that of "accounting for the direction, vigor
and persistence of behavior."12 The problem for theory is
the development of a "coherent conception of the contemp-
oraneous determinants of direction, vigor and persis-
tence of action."13

Atkinson draws on the work of LewinlLl to distinguish
the problem of learning theory with its emphasis upon the
frequency of past associations from the problem of moti-
vation with its emphasis upon the "contemporaneous" deter-

minants of behavior. Atkinson credits Tolman15 with the

clarification of the characteristics of behavior to be

 

11John W, Atkinson, An Introduction to Motivation
(D. Van Nostrand and Co., Inc., 1964).

12

 

Ibid., p. 274.

l31b1d.

1MK. Lewin, The Conceptual Representation and the

Measurement of Psychological Forces (Durham, North
Carolina: Duke University Press).

15E. C. Tolman, Purposive Behavior in Animals and
Men (New York: Century Co. By permission of the Univer-
sity of California Press, 1932).

 

 

explained by motivation theory--direction, vigor and
persistence.

Farquhar has drawn on the above empirical and theoret—
ical legacy for the construction of the GSCI. Farquhar's
application of theory to the study of achievement in school
was guided by the following definition of academic moti-
vation: "a combination of forces which initiate, direct, and

sustain behavior toward a scholarly goal."l6

17

While research by Atkinson and his co-workers has
been greatly involved with exploration of the relation-
ships between theory, instruments and the direction, vigor
and persistence of behavior (construct validity studies),
little has been done by Farquhar and his co-workers to
relate theory and instruments directly to the initiation,
direction and sustaining of behavior in the academic
situation.

This void is particularly surprising since one of the
most persistent problems for Farquhar and his co-workers
has been the difficulty in distinguishing intellective and

non-intellective influences in their findings. Other

18 . . .
researchers have shown such characteristics of behav1or

 

l6Farquhar, op. cit., p. 3.
17

02. cit.

18Ibid.; Atkinson, An Introduction to Motivation,

02. cit.

Atkinson, A Theory of Achievement Motivation,

 

as goal setting and persistence to be less influenced by
intellective factors than such criteria as grade point
average or achievement test performance. These findings
suggest the potential usefulness of similar behavioral
criteria in attempts to gain greater clarity concerning the
constructs underlying the GSCI.

19 and his co-

The laboratory studies by Atkinson
workers have been attempts to gain such clarity for both
their theoretical notions and the instruments thought to
index the operation of various terms in their theory.
However, the laboratory setting of many of these studies
involving the use of simple response measures, "fictitious"
instructions and "contrived" tasks has resulted in a lack
of clarity in the application of their findings to "real-
life" situations.

On the other hand, the efforts of Farquhar and his
co-workers have been centered in "real—life" academic
situations and concerned with "worth-while" accomplish-
ments. While Farquhar's findings hold considerable utility
when applied to the prediction of accomplishment in school,

they suffer most in a lack of clarity for the basic

constructs involved in his research.

 

l9
op. cit.

Atkinson, A Theory of Achievement Motivation,

 

It may be possible to gain greater clarity for the
constructs involved in Farquhar's research through observa-
tion of certain classroom behavior relevant to the study of
motivation. If academic motivation is thought to involve
the Operation of "forces which initiate, direct, and sustain
behavior toward a scholarly goal"2O then it may be possible,
under apprOpriate conditions, to observe the influence of
these "forces" in the behavior of students performing an
academic task flithe classroom situation.

The academic classroom is viewed as a competitive
situation in which performance is evaluated and feelings
of satisfaction contingent upon the performance of certain
instrumental acts. Such a situation is described by
Atkinson21 as appropriate for the expression of "a rela—
tively stable disposition to strive for achievement."

Atkinson22 suggests that performance in such a
situation is a function of certain personality disposi—
tions (motives) apd certain cues in the situation (expec-
tancy and incentive values) i.e., Behavior = f (motive x
expectancy x incentive).

One implication of these formulations is that situ-

ational characteristics (expectancies and incentives) are

 

 

2OFarquhar, op. cit., p. 3.
21Atkinson, An Introduction to Motivation, op. cit.
22

Ibid.

thought to interact with individual differences in stable
personality characteristics (motives) to determine the
expression of motivation in behavior.

However, there has not been adequate exploration of
the influence of motives, interacting with situational
characteristics, on the initiating, directing and sus-
taining of achievement-related behavior in the classroom

situation.

Need for the Present Study

 

No study, using the GSCI, has been designed in which
Atkinson's theory of achievement motivation is used to
predict theory-relevant behavior within the classroom
situation. Furthermore, no study, using the GSCI, has
investigated the contentions of Atkinson's theory that
individual differences in stable personality traits (motives)
interact with situational characteristics to influence
behavior.

Confirmation of the above relationships and inter-
active effects could aid in clarifying the constructual
basis of the GSCI, provide support for Atkinsons's theory
and contribute information concerning the relative pre-
dictive efficiency in the use of the GSCI.

The accepted approach for the study of the implica-
tions personality test scores may have for theory-relevant,
extra-test, behavior is that of construct validation which

is discussed in the next section.

10

Construct Validation

 

Statements in The Technical Recommendations for

Psychological Tests and Diagnostic Techniques23

imply
that "construct validation" is appropriate when an inves-
tigator believes that his instrument reflects the Opera-

tion of a particular construct. That is, he makes infer—

 

ences concerning the hypothetical syndrome, trait, or
personality dimension he believes reflected in the scores
of his instrument. The testing of the inferences is the
purpose of construct validation.

Theory is given the guiding role in formulating
inferences about a construct and in specifying the proper-
ties of observables which can test the accuracy of these
inferences. Cronbach and Meehl2u use the term "nomological
network" to describe an interlocking system of laws which
can form the basis for inferences concerning the inter—

pretation of a test. It is a necessary condition, in their

view, that at least some of these laws relate to observables

 

thereby allowing the collection of confirming or discon-

firming evidence.

 

23American Psychological Association, Committee on
Psychological Tests. Technical Recommendations for Psycho-
%pgica1 Tests and Diagnostic Techniques. Washington,
.C.: APA, 1954.

2“Lee J. Cronbach and Paul E. Meehl, "Construct
Validity in Psychological Tests," Psychological Bulletin,
Vol. 52, No. 4, July, 1955, 281-302.

 

 

 

11

However, construct validation studies are undertaken
precisely when no ppg observable property is accepted as an
entirely adequate criterion to define the quality to be
measured. Cronbach and Meehl stress that while some observ-
ables may be regarded as "criteria" the construct validity
of the criteria themselves is regarded as under inves-
tigation.’

Since no ppe criterion is thought to be adequate to
establish validity for a construct, Cronbach and Meehl
suggest that:

numerous successful predictions dealing with pheno-
typically diverse "criteria" give greater weight to
the claim of construct validity than do fewer
predictions involving very similar behaviors. 5

While Cronbach and Meehl stress the importance of the
number and diversity of successful predictions to deal with
the problem of inadequate criteria, Sarason26 has sug-

gested an "analytic" strategy in which inferences regarding

a trait and inferences regarding criteria are separately

 

analyzed.

 

251bid., p. 295.

261. G. Sarason, Contemporary Research in Person-

alit (Princeton, New Jersey: D. Van Nostrand Co., Inc.,
1962;.

 

l2

Sarason suggests that, in construct validation

studies:

A two-stage inference is involved; first, there is an

inference from the criterion requirement to the

traits that are relevant to that performance (the

"criterion analysis"); and secondly, an inference

from the subject's observed behavior and test per-

formance to his status on the trait dimensions

(the assessment). Research on the validity of these

inferences requires two separate studies; one of the

validity of the analysis of the criterion require-

ments and the criterion indices, and one of the

validity of the tests as predictors of the

criterion.2

For the present study, the statements of Cronbach and

Meehl together with those of Sarason suggest that motiva—
tion in the classroom situation would be more adequately
investigated by criteria of initiating, directing and sus-
taining behavior than by any one of these criteria and
that a separate analysis of the criteria themselves could
aid in clarifying the locus of potential problems asso-
ciated with either the trait assessment or criterion

analysis.

Construct (Trait) Validity of the M-Scales
Hofmann28 has explored the construct validity of the
M-Scales through multitrait-multimethod matrix analysis.

He intercorrelated scores from several instruments thought

 

fir

27Ibid., p. 90.

28Louis J. Hofmann, "An Application of the Multitrait-
Multimethod Matrix to the Study of the N-Achievement
Construct," (unpublishedmdoctoraldissertation, Michigan
State University, East Lansing, 1965).

 

13

to measure traits that were similar (achievement motive) and
dissimilar (affiliation motive) by assessment techniques
that were similar (objective) and dissimilar (projective).

Hofmann's results were interpreted in support of the
convergent validity of the achievement motive scales with
the projective methods of assessment less related to aca-
demic achievement and intelligence than were the objective
methods.

The type of construct validity study reported by
Hofmann is described by Campbell29 as appropriate when
"theory, if any, goes no farther than indicating a hypo-
thetical syndrome, trait, or personality dimension."3O
Campbell labels this type of construct validity study as a
"trait validity" study to distinquish it from a "nomo-
logical validity" study which he describes as:

. . . the very important and novel emphasis of
Cronbach and Meehl on the possibility of vali-
dating tests by using the scores from a test

as interpretations of a certain term in a formal
theoretical network and, through this, to gen-.
erate predictions which would be validating if

confirmed when interpreted as still other
Operations and scores.3

 

29D. T. Campbell, "Recommendations for APA Test
Standards Regarding Construct, Trait or Discriminant
Validity," American Psychologist, Vol. 15, No. 8,August
1960.

 

301bid., p. 547.

3lIbid.

14

Hofmann's purpose was to find support for the achieve-
ment motive construct and for the Farquhar instrument
(M—Scales) as an index to the operation of that construct.
Hofmann followed Campbell's recommendation in using "a joint
criteria of independence of method and relevance to the
trait"32 in the selection of validating measures. That is,
theory was considered sufficiently developed to distinguish
instruments purportedly measuring the same traits from
instruments measuring different traits.

Hofmann concluded that theory, underlying each of the
several instruments used in his study, was not sufficiently
developed to allow predictions concerning performance in
the academic situation, i.e., a nomological validity study.

However, theory is at present sufficiently developed
to allow a nomological validity study of at least one of
the sub-scales used in Hofmann's study; the Generalized
Situational Choice Inventory (GSCI).

Construct (Nomological) Validity of the

Generalized Situational Choice
Inventory (GSCI)

 

 

 

The GSCI, one of the sub-scales within the M-Scales
battery, was constructed to conform specifically to the
dimensions of an achievement—related motive, the Motive for

Success (Ms), explicated by McClelland33 and Atkinson.3u

 

r

32Ibid., p. 548.
33

34

McClelland and Atkinson, 0p. cit.

Atkinson, op. cit.

l5

Farquhar35

extended, polarized and applied these dimensions
to the academic situation.

Scores derived from the GSCI were thought to reflect
individual differences in strength of the Motive for Success
which is one of the terms occurring in Atkinson's36 Motive
x Expectancy x Incentive formulation of the determinants of
achievement motivation.

The academic classroom was considered to be a compet—
itive situation in which personality (motives) and charac-
teristics of the situation (expectancies and incentives)
interact to influence performance.

Therefore, it appeared possible to make inferences
from Atkinson's theory, using scores from the GSCI as inter-
pretations of the Motive for Success (Ms) term in that
theory, and test these inferences on criteria of behavior,
relevant to the study of motivation, within the classroom
situation.

Atkinson's statements concerning the empirical problem
of motivation and Farquhar's definition of motivation within
the academic situation were considered sufficiently explicit
for the selection of dimensions of behavior relevant to the

eXpression of motivation in the c1assroom--initiating,

directing and sustaining of achievement-related behavior.

 

35Farquhar, op. cit.

36Atkinson and Feather, op. cit.

l6

Purpose of the Present Study

The major purpose of the present study was to seek
evidence concerning the construct (nomological) validity
of the Motive for Success, as (a) represented within
Atkinson's theory of achievement motivation, (b) indexed
by the Generalized Situational Choice Inventory (GSCI)
test scores and (c) expressed in theory-selected behavior
within the classroom situation.

A second purpose was to investigate both the theo-
retical and practical consequences of three theory-
selected criteria of motivated behavior within the class—
room for level of achievement in the academic situation.

A third purpose was to investigate both the
theoretical and practical consequences of predictive
efficiency for the Generalized Situational Choice Inven-
tory (GSCI) with level of achievement in the academic

situation.

Theory

Theory is drawn from the formulations of Atkinson

37

and McClelland38 as extended and applied to the academic

situation by Farquhar.39 Atkinson and McClelland view

37Atkinson, Motives in Fantasy Action and Society and
53 Introduction to Motivation, op. cit.; Atkinson and

Feather, op. cit.
38

 

 

McClelland and Atkinson, 0p. cit.

39Farquhar, op. cit.

l7

motivation, as expressed in the direction, magnitude and
persistence of behavior, to be a positive function of certain
personality dispositions (motives) and certain character-
istics of the situation (expectancy and incentive values).
These authors have given particular attention to one
personality disposition, the Motive for Success, and its
expression in competitive situations.

Atkinson“Q defines the Motive for Success as a rela-
tively stable (though latent) disposition to strive for
feelings of satisfaction through achievement-related
activities. The influence of individual differences in
strength of the Motive for Success depends upon, (a) the
presence of situational cues which determine the strength
of an expectancy of satisfying that motive through some
action instrumental to achievement and (b) the value of the
specific goal or incentive offered in the situation.

Atkinson defines an achievement-related "expectancy"
as a cognitive anticipation, aroused by cues in the compet-
itive situation, that performance of some act will be
followed by success. In the competitive situation, the
strength of the expectancy for success is assumed to vary
inversely with the difficulty of the task. That is, the
more difficult the task, the lower the strength of the

expectancy for its successful accomplishment.

k

qutkinson, An Introduction to Motivation, pp. cit.

 

18

“1 defines the achievement-related "incentive"

Atkinson
value as the relative attractiveness of a specific goal that
is offered in the competitive situation. The attractiveness
of success, in the competitive situation, is assumed to be a
positive function of the difficulty of the task. That is,
the more difficult the task, the higher the incentive
value for its successful accomplishment.

Atkinson has assumed that the strength of the expec-
tancy for success can be indexed by the "subjective prob-
ability" of success. Thus a task for which an individual has
a strong expectancy of success is one for which his subjective
probability of success is high (an easy task). Conversely,

a task for which the individual has a weak expectancy of
success is one for which he holds a low subjective prob-
ability of success (a difficult task).

Atkinson has also assumed that, in the competitive
situation, the attractiveness of success (incentive value)
is dependent upon the probability of success (expectancy)
such that the accomplishment of a difficult task (low
probability of success) is more attractive (has higher
incentive value) than the accomplishment of some trivial
or easy task (high probability of success).

Thus, the more difficult the task, the lower the

expectancy (probability) for success but the higher the

 

‘7

”lipid.

19

attractiveness (incentive value) of its accomplishment.
Conversely, the easier the task the higher the expectancy
(probability) for success ppp the lower the attractiveness
(incentive value) of its accomplishment.

The relationships between the terms of the theory are
given in Table 1.1.- In Table 1.1 it can be seen that, as
the difficulty of a task decreases from "Very Difficult"
to "Very Easy," the probability of success (Ps, in values

ranging from 0 to 1.00) inpreases while the incentive value

 

of success (Is, in values ranging from 0 to 1.00) decreases.

 

The product of Ps and Is, as specified in theory, can be
seen to have their highest values (.25) for the task of
an "Intermediate" difficulty level.

It is, therefore, in the range of "Intermediate" task
difficulty that individual differences in strength of the
Motive for Success (Ms) are most strongly influenced by the
product of expectancy and incentive values (Ps x Is) and the
resultant expression of the Motive for Success (Ms) in
behavior is also strongest.

42 43

Atkinson and Cartwright and Weiner have recently

given attention to Atkinson's 1957 theory of achievement

 

u2J. W. Atkinson and D. Cartwright, "Some Neglected
Variables in Contemporary Conceptions of Decision and Per-
formance," Psychological Reports, 14 (1964), 575-590.

 

u3Bernard Weiner, "The Effects of Unsatisfied Achieve—
ment Motivation on Persistence and Subsequent Performance,"
Joupnal of Personality, 33 (1965), 428—442.

 

20

TABLE l.1--Motivation (Expressed in Behavior) as a Joint

Function of the Motive for Success (Ms), Expectancy of

Success (Ps) and Incentive Value of Success (Is) for
Individuals in Whom Ms = 1 and Ms = 2.

 

Strength of

Motivation when:

 

 

TASK
Ps x Is x Ms = l, or Ms = 2
A. Very Difficult .10 .90 .09 .18
B. Intermediate .50 .50 .25 .50

C. Very Easy .90 .10 .09 .18

 

21

motivation[”4 and have criticized the "stimulus bound"
condition of that theory. Criticism by the above authors
is based on the fact that Atkinson's 1957 model requires.
specification of certain situationally determined factors
(expectancy and incentive values) which allow prediction of
behavior 921x when these factors are objectively present.
Therefore, that model is "stimulus bound" or restricted to
prediction of behavior in situations in which certain
stimulus conditions, specified in the theory, are available
to influence behavior.

Atkinson and Cartwright argue that a theory of moti-
vation must attempt to account for the change from one
activity to another rather than for only the instigation of
action (initiation) or for only the duration of activity
(persistence) and view a "stimulus bound" model of achieve-
ment as inadequate for prediction or description of a
change in activity.

Atkinson and Cartwright have expanded the 1957 model
of achievement motivation to allow theory to account for a
change in activity from one situation to another. They
posit the Operation of an "inertial tendency," a goal-
directed tendency which, once aroused, persists until satis-

fied or dissipated.

 

”NJ. W. Atkinson, "Motivational Determinants of
Risk-Taking Behavior," PsychologicalpReview, 64 (1957).

.359-372.

22

The formulations of Atkinson and Cartwright imply that
if the achievement motive is aroused by instructions or the
performance of an achievement—oriented activity and then
prevented from obtaining satisfaction, either by interrup-
tion or failure, some portion of the tendency persists to
influence subsequent behavior.

Atkinson,4S has attempted to maintain a sharp distinc-
tion betweentflneterm "motive," conceived as a relatively
stable personality trait, and "motivation," which is con-

ceived as the activated state of a person to engage or not

 

engage in a particular activity. Atkinson has recently
suggested the use Of the term "tendency" as a substitute for
the more commonly used term "motivation" to aid in clari-
fying the distinction between stable traits and fluctu-
ations in behavioral expressions which depend upon charac-
teristics of the particular situation.

Atkinson has assumed that the consequences of the
multiplicative relationship of the Motive for Success (Ms)
with particular expectancy and incentive values will be
reflected in the strength of a "tendency" (motivation)
such that:

1. When the difficulty of a task is held constant,

the tendency to achieve success is stronger
when Ms is strong than when it is weak, but

 

r v Y‘ W w

uSAtkinson and Feather, op. cit.

23

2. the difference in strength of tendency to achieve
success that is attributable to a difference in
strength of the achievement motive (MS) will be
substantial only whfip the task is one of inter-
mediate difficulty.

Atkinson's statements, 1. and 2. above, suggest that
there should be a relationship between strength of the
Motive for Success and strength of the tendency (motiva-
tion) to achieve success when differences in the difficulty
of success are held constant, and the above relationship
should be stronger when success is perceived to be within

an intermediate range of difficulty than when perceived to

be easier or more difficult.

Prediction of Behavior invthe Classroom Situation

 

Atkinson's statements, 1. and 2. above, were thought
to hold implications for prediction of relationships among
three different types of variables--personality, motivation
and achievement--within an academic situation in which
students are presented with both a non—achievement task and
an achievement task and motivation is reflected by criteria
of the initiation, direction and sustaining of achievement-
related behavior. Descriptions of the relationships among
these three types of variables are presented below.

The statements by Atkinson, 1. and 2. above, were
thought to hold implications for the prediction of relation-

ships between strength of the Motive for Success and behavior

 

u6Atkinson and Feather,_op. cit., p. 329.

24

within the academic situation when strength of the tendency
(motivation) to achieve success was reflected by criteria
of the initiation, direction and sustaining of achievement-
related behavior.

Atkinson's statements, 1. and 2. above, were also
thought to hold implications for the prediction of relation-
ships between strength of the tendency (motivation) to
achieve success, reflected by criteria of the initiation,
direction and sustaining of achievement-related behavior,
and level of achievement within the academic situation, if
level of academic achievement is itself assumed to be a
criterion of strength of the tendency (motivation) to
achieve success within the academic situation.

The statements by Atkinson, l. and 2. above, were
furthermore thought to hold implications for the prediction
of relationships between strength of the Motive for Success
and level of achievement within the academic situation, if
level of academic achievement is itself assumed to be a
criterion of strength of the tendency (motivation) to
achieve success within the academic situation.

Hypotheses, formulated on the basis of Atkinson's
statements 1. and 2. above, were applied to prediction within
three Stages of enquiry corresponding to the relationships
among the three types of variables described above. The
three Stages of enquiry, and the hypotheses derived from

Atkinson's statements, are presented below.

25

Hypotheses for Stage I--Nomologica1 Validity

Enquiry with this stage was thought to be most closely

relevant to the construct (nomological) validity of the

Motive for Success within the academic situation.

Hypotheses were formulated concerning the influence of

strength of the Motive for Success on three criteria of

strength of the tendency (motivation)tx>achieve success within

the academic situation. Hypotheses were also formulated con-

cerning the strength of the above influence as differentially

affected by the perceived difficulty of success in school.

10

With differences in the perceived difficulty of
success in school held constant, differences in
strength of the Motive for Success should

influence the initiation, directing, and sus-
taining of achievement-related behavior.

The influence of differences in strength of the
Motive for Success on the initiation, direction and
sustaining of achievement-related behavior should
be stronger when success in school is perceived to
be within an intermediate range of difficulty than

when perceived to be easier or more difficult.

Hypotheses for Stage II-—Criterion Validity

Enquiry within this stage was thought to at least

allow for a separate analysis of the validity of three

behavioral criteria of strength of the tendency (motivation)

to achieve success within the academic situation.

26

Hypotheses were formulated concerning the influence of
three criteria of strength of the tendency (motivation) to
achieve success on level of achievement in the academic
situation. Hypotheses were also formulated concerning the
strength of the above influence as differentially affected
by the perceived difficulty of success in school.

3. With differences in the perceived difficulty of
success in school held constant, differences in
the initiation, direction and sustaining of
achievement-related behavior should influence
the level of academic achievement.

4. The influence of differences in the initiation,
direction and sustaining of achievement-related
behavior on the level of academic achievement
should be stronger when success in school is
perceived to be within an intermediate range of
difficulty than when perceived to be easier or
more difficult.

Hypotheses for Stage III--Predictive
and Nomological Validity

 

 

Enquiry within this stage was thought to allow for
evidence concerning the relative predictive efficiency of an
index of strength of the Motive for Success with indices of
academic achievement, as well as allowing for evidence con-
cerning the operation of this construct, as predicted by

theory, within the academic situation.

27

Hypotheses were formulated concerning the influence of
strength of the Motive for Success on level of achievement
within the academic situation. Hypotheses were also formu-
lated concerning the strength of the above influence as
differentially affected by the perceived difficulty of
success in school.

5. With differences in the perceived difficulty of
success in school held constant, differences in
strength of the Motive for Success should influence
the level Of academic achievement.

6. The influence of differences in strength of the
Motive for Success on the level of academic achieve-
ment should be stronger when success in school is
perceived to be within an intermediate range of
difficulty than when perceived to be easier or

more difficult.

Overview of the Study

In Chapter I, the need for behavioral criteria,
relevant for a construct (nomological) validity study of the
Generalized Situational Choice Inventory (GSCI) was met by
the selection of classroom behaviors that conform to
Atkinson's statements concerning the empirical problem of
motivation and to Farquhar's definition of the role of moti-
vation within the academic situation. Theory was obtained

from the Motive x Expectancy x Incentive formulations of

28

Atkinson and hypotheses stated concerning relationships
among the Motive for Success, initiating, directing and
sustaining behavior and level of academic achievement.

In Chapter II, a selected review of the literature
will be presented with particular attention to construct
validity studies of the Motive for Success and to behavioral
criteria of motivation similar to those used in the present
study.

In Chapter III, the design of the study will be
described with reference to the sample, method of data
collection, statement of hypotheses and the statistical
techniques used.

In Chapter IV, the results of the statistical analysis
of the data will be presented along with statements of
hypotheses accepted and rejected. I

In Chapter V, the findings of the study will be dis-
cussed as related to theory, previous research and as
limited by methodology.

In Chapter VI, the summary, conclusions and implica-

tions for future research will be presented.

CHAPTER II
REVIEW OF THE LITERATURE

Each of the studies reviewed in this chapter offers
some empirical evidence, theoretical View or experimental
design feature bearing on the relationship between
achievement-related motives and behavior in competitive
situations.

To provide some measure of clarity in presentation,
the studies reviewed in this chapter are presented in sep-
arate sections according to the major criterion variables
of the present study. That is, studies concerned with the
expression of motivation through the initiation, direction
and sustaining of activity are reviewed in the first
sections. In later sections attention is given to those
studies in which a third variable was allowed to differen-
tially influence relationships between personality and
behavior.

Achievement Related Motives and the
Initiation of Behavior

Measures of speed or latency in approaching a consum-
matory situation, have been extensively researched by psychol-
ogists interested in the effects of variations in basic need

states on the behavior of animals in a laboratory situation.

29

30

Reviews1 of animal studies suggest that the effects
of increasing need states are to increase speed of leaving
a start box, speed of traversing a runway and speed of
initiating consummatory activity.

Kimble2 concludes, from his review of the effects of
variation in basic need states on the behavior of animals,
that "motivational variables seem to have more influence
on speed than on measures of accuracy." He cites a study
by Cotton3 in which the runway behavior of rats was studied
under varying lengths of food deprivation. Cotton found
the effects of high drive were to (a) increase running
speed and (b) decrease competing responses (e.g., stopping
to sniff, wash or explore).

Kimble interprets the findings of Cotton in support
of the contention that one reason speed measures may more
clearly reflect variation in motivational variables is
because speed measures are particularly susceptible to

disruption by the effects of competing responses.

 

Gregory A. Kimble, Conditioning and Learning (New
York: Appleton-Century-Crofts, Inc., 1961; Francis W.
Irwin, "Motivation and Performance," Annual Review of .
Psychology, Vol. 12 (Palo Alto: Annual Reviews, Inc., 1961).

2Kimble, op. cit.

3J. W. Cotton, "Running Time as a Function of Amount
of Food Deprivation," Journal of Experimental Ppychology,
46 (1953), 188-198. V I

 

31

While psychologists interested in the effects of
motivational variables on the performance of human subjects
have rarely studied speed of initiating performance, con-
siderable attention has been given to the role of competing
responses during the task performance of subjects thought
to differ in the level of need or drive state.

Studies by Taylorl4 and Mandler and Sarason5

working
with reSponse-defined measures or adversive drive states,
suggest that in simple response situations high drive states
result in faster performance but in complex situations
competing responses may interfere most with the performance
speed and accuracy of high drive subjects.

While the above studies indicate the possible effects
of varying drive states on speed of activity during the
performance of a particular task, little attention has been
given by the above researchers to the possible effects of
either drive state or strength of competing responses on
the speed with which their subjects initiated task behavior.

The importance of the behavior of human subjects at

the point of initiating "worth while" activities has, none-

theless, been discussed by several authors.

 

”Janet A. Taylor, "The Relationship of Anxiety to the
Conditioned Eyelid Response," Journal of Experimental
Psychology, 41 (1951), 81-92.

 

5G. Mandler and S. B. Sarason, "A Study of Anxiety
and Learning," Journal of Abnormal and Social Psychology,
“7 (1952), 166-173. '

32

Tiebout6 concluded, from his study of under-achieving

college girls, that they appear to experience particular

difficulty at the point of beginning to study. Bruner,7

 

in discussing "learning blocks" in the academic situation
suggests that a distinction needs ulbezmade between those
behaviors that "cope" with the problems of achievment and

those that "defend" against entry into the learning

 

situation.

Brown, Abeles and Iscoe8 conclude from their study of
the motivational differences between high and low achieving
college students, that the poor students seems to be char-
acterized by "what we shall call here 'activity delay' that
is, a lack of decisiveness of action, a tendency to
procrastinate . . ."

The studies above, concerned with the behavior of
students in the academic situation, suggest that motiva-
tional variables may influence behavior at the point of
initiating achievement-related activity, that behavior at
this point may have coping or defensive characteristics,
and that tendencies to procrastinate may have an influence

on later academic achievement.

 

6H. M. Tiebout, "The Misnamed Lazy Student," Educa—
tipnal Record, 24 (1943), 113-129.

7Jerome Brunner, Toward a Theory of Instruction,
(Cambridge, Massachusetss: Harvard University Press, 1966).

8William F. Brown, Norman Abeles and Ira Iscoe,
"Motivational Differences Between High and Low Scholarship
College Students," The Journal of Educational Psychology,
45 (1954), 215-223. I

 

 

33

Atkinson,9

in reviewing his own studies, and those of
Feather,10 concerned with the effects of the interaction of
personality and environmental factors on the persistence of
activity during the performance of an achievement-related
task, has noted the rather arbitrary distinction made
between time measures labeled "persistence" and those called
"latency."

With both "persistence" and "latency" measures a
clock is used to measure a period of activity. If the
interest of the psychologist happens to be in a particular
task performance, he calls his measure "latency" if he stops
his clock when the performance begins, but he calls his
measure "persistence" if he stops his clock when the
performance ceases.

Atkinsonll has argued that psychologists must attend
to both latency and persistence measures if motivational
theory is to account for what he believes to be of crucial

importance, that is, "change" in behavior.

fiv— 7i

9John W. Atkinson, An Introduction to Motivation,
D. Van Nostrand and Co., Inc., 1964.

10Norman T Feather, "The Relationship of Persistence
at a Task to Expectation of Success and Achievement Related
Motives, " Journal of Abnormal and Social Psychology, 63
(1961), 552-561; "Persistence at a Difficult Task with
Alternative Task of Intermediate Difficulty," Journal of
Abnormal and Social Psychology, 66 (1963), 604-609.

llAtkinson, pp. cit.

 

 

34

If behavior "change" is viewed as crucial for moti-
vational theory, then description or prediction restricted
to performance of a single task will be inadequate and may
require attention to the effects of the interaction of
personality and environment on behavior preceding and
following that of immediate concern to the researcher.

Atkinson describes a situation in which a college
professor is working at his desk when a colleague invited
him to the coffee room. Atkinson concludes that "the
latency of R2 (i.e., the length of time before the professor
gets up to go to the coffee room if he does, in fact, get up
-to go at all) should be proportionate to the strength of the
achievement motive (Ms) and inversely proportionate to the
strength of affiliative motive (MAff)."12

That is, the relative strength of the above two motives
as well as situationally—determined expectancies and incen-
tives for satisfying each motive must be taken into consid-
eration to account for the point at which a "change" will
take place in the professor's behavior.

Therefore, Atkinson stresses the importance of attention
to alternative activity, from which and to which, a subject
turns as well as knowledge concerning the strength of possible

competing motives, for an accurate prediction of behavior.

If the alternative activity is considered to be one appealing

 

‘71 Vfi

12Ibld.

35

to different motives than the activity in question then
measures of latency or speed of initiating the activity of
interest may reflect motivational influences in addition to
’those reflected by measures of task persistence.
Achievement Related Motives and the
' Direction chBehavior"

Atkinson suggests that one of the important problems
for a theory of motivation is to account for the selection
or direction of one path of action among a set of possible
alternatives. He suggests that the "problem of selection
arises in experiments which allow the individual to choose
a task among alternatives which differ in difficulty (level
of aspiration)."13

However, when subjects are presented with only one
task and asked to express a "hoped for" or "expected" level
of achievement, their aspirant behavior can be character-
ized as involving a choice between higher or lower direc-
tions of future achievement. That is, a subject who states
expectations toward higher levels of future achievement
could be considered to have chosen a more difficult direc—
tion (lower probability of attainment) than a subject who
states expectations toward a lower level of achievement on-

objectively the same task.

 

T'r'v Y "‘ V V ' ‘

l3Ibid.

36

One paradigm for investigation of level of aspiration
involves a subsequent level of performance which an indi—
vidual states he anticipates achieving following performance
on a task (future goal minus past achievement). While such
a procedure attempts to control for the effects of reality
determinants and equate discrepant levels of past attain-

ment, Hillslu

and Siegel15 have demonstrated the complexity
of such goal-discrepancy measures and cast doubt on the
clarity of such techniques particularly as measured within
the academic situation.

Because of the complexity and limitations of the
goal-discrepancy measures of level of aspiration and because
of their susceptibility to "wishful" or "unrealistic"
influences, the following review will center on studies in
p which expectations were defined by direct statements con—
cerned with future achievement and reality determinants
dealth with, if at all, in ways other than as discrepancies
from previous achievement.

Atkinsonl6 administered the TAT n—achievement test to
approximately one-half (N = 38) of the students in his

college class who volunteered to appear one-half hour before

 

5-7 a? v—w— fifi

1“John R. Hills, "The Measurement of Levels of Aspir—
ation," {ournal of Social Psychology, 41 (1955), 221-229.

15Sidney Siegel, "Level of Aspiration and Decision
Making," Psycholpgical Review, 64 (1957), 253-262.

16D. McClelland and J. Atkinson, et a1., The Achieve-
ment Motive (New York: Appleton—Century-Crofts, 1953). Y

 

 

37

the final exam. He asked them to indicate the score
(possible total of 100 points) that "you expect to make on
the exam." He also asked them to report the quarter of the
class in which they scored at mid-term time and their
overall grade average.

Atkinson assumed that the level of aspiration measure,
"what grade do you expect," would be influenced by both
reality factors, e.g., past level of achievement in class
and in college, and by motivational factors, e.g., achieve-
ment needs. An attempt was made to eliminate the influence
of reality determinants by partial correlation technique.

Atkinson found level of aspiration scores (expected
grades) related to mid-term standing (specific past per-
formance), r = .41 (p > .05), but not significantly related
to overall grade average (general past performance),

r = .30 (p < .05). The relationship between n-achievement
test scores and eXpected grades was not significant, nor
was this relationship significant when the influence of
either past specific, or past general, performance was held
constant by partial correlation technique.

Atkinson further sought to reduce the influence of
reality determinants, on the relationship of n-achievement
test scores with expected grades, by a technique other than
partial correlation. He selected those students who appeared
in discrepant thirds of the separate distributions of mid-

term standings and general grade standing. He assumed that

38

this subgroup of students should be in some greater doubt
or conflict concerning what to eXpect on the basis of past
experience. Such an interpretation has typically been
applied to the expectations of subjects faced with a task
of "intermediate" difficulty defined as 50% probability of
success.

For this subgroup (23 of 38 students) with discrepant

past achievement, the correlation Of n-achievement test

scores with expected grades was r = .45 (p < .05), while this
relationship was r = -.23 (p > .05) for the remaining 15
students.

Atkinson interprets the above findings in support of
his contention that both reality determinants and motiva-
tional determinants influence level of aspiration measures.
However, when reality determinants are minimized by the
selection of subjects assumed to be most uncertain about
future success, the influence of motivational determinants
is more pronounced.

Pottharstl7

reports findings similar to those above
using high school boys and a novel task designed to reduce
the influence of past experience on future expected levels

of achievement. Pottharst found those high school boys

with high TAT n-achievement test scores reported higher

 

178. C. Pottharst, "The Achievement Motive and Level
of Aspiration After Experimentally Induced Success and
Failure," (unpublished Doctoral thesis, University of
Michigan, Ann Arbor, 1955).

39

expected levels of achievement on a novel task than those
boys with low TAT n-achievement test scores.

Atkinson, Bastian, Earl and Litwinl8 also report an
upward shift in level of expectation attributable to higher
level of achievement needs. Atkinson, pt gl., asked 59
college males to state the number of persons they expected
to excel before engaging in a shuffleboard contest. While
there were no differences in the number of successful shots,
those males with high n—achievement test scores (Test of
Insight [TOI]) reported they expected to excel more persons
than those with low n-achievement test scores (X2 = 6.17,
df = l, p < .02).

While the studies reported above were designed to
investigate the influence of achievement-related motives
on level of expectation when the influence of previous

19

experience was controlled or minimized, Izard investi-
gated the relationships of Edwards Personal Preferences
Schedule (EPPS) test scores with both expected future
achievement and actual later achievement within an academic

situation in which past experience was assumed to play a

central role.

 

vvi

18John W. Atkinson, Jarvis R. Bastian, Robert W. Earl,
and George H. Litwin,-"The Achievement Motive, Goal Setting,
and Probability Preferences," Journal of Abnormal and Social
Psychology, 60 (1960), 27—36. '

 

19Carroll E. Izard, "Personality Characteristics
(EPPS), Level of Expectation, and Performance," Journal of
Counseling Psychology, 26 (1962), 394.

 

 

40

Izard assumed that EPPS test scores would not relate
to expected grades on a course exam, described as comparable
to one recently taken and reviewed, but would relate to
actual course achievement (number of items correct on ail
exams in the course).

Izard found no relationship between EPPS n-achievement
scale scores and expected exam grade but found a signifi-
cant correlation for 33 males between EPPS n-achievement
scale scores and actual level of course achievement, r - .40
(p < .05). Izard also reports a subsequent study in which
a significant correlation was obtained between EPPS n-
achievement scores obtained on 180 males as freshman and
their GPA's as seniors, r = .28 (p < .01).

While the findings in Izard's studies indicate some-
thing of the potential predictive efficiency for EPPS scores
with achievement criteria in the academic situation, his
procedure did not allow for evidence that EPPS n-achievement
scale scores reflect the operation of a Motive for Success
as this term appears in Atkinson's theory of achievement
motivation. That is, there is no evidence in Izard's
study to indicate the EPPS n-achievement scale scores relate
to behaviors thought to be more clearly reflective of moti-
vational influences than college grades (e.g., aspirations
or expectations) or that such relationships are influenced

by situational variables.

441

The above studies indicate somewhat less support for
the predictive efficiency of the projective (TAT and TOI)
measures of achievement-related motives than for the objec-
tive (EPPS) measures, but somewhat greater support for the
construct validity of the projective measures.

However, even those studies demonstrating construct
validity for the projective measures of achievement motives
through relationships with the direction of expectations
concerned with future achievement, and with stronger rela-
tionships in situations in which reality determinants were
minimized, failed to investigate either the reliability
of the indices of eXpectation or the influence of these
variables on future achievement.

The findings of the studies reviewed above were
thought to at least allow sufficient clarity for the selec-
tion of "Expected Grades" as a criterion for the direction
of activity in the classroom situation. The direction of
activity toward higher or lower academic achievement
appears relevant to the construct validity of the Motive
for Success and previous findings support the contentions
of theory that relationships between the Motive for Success
and expected future achievement should be stronger when

reality-determinants are controlled or minimized.

42

Persistence as a Personality Trait

Webster's Collegiate Dictionary20 states that to
persist is: "To go on resolutely in spite of opposition,
importunity, or warning; to persevere." This quality of
"sticking to a task once undertaken" has received exten-
sive attention in the literature21 and has been the focus
of a number of experimental studies.22

Woodworth23 centers attention on the importance of
persistence in the academic situation and states: "In
school work there is a good reason to believe that persis-
tence, or sticking to a task, is one of the main factors

that helps to supplement or compensate for ability . . ."

Research efforts concerned with the contribution of

 

20Webster's Collegiate Dictionary (Springfield,
Massachusetts: G. and C. Merriam Co., 1944).

21H. J. Eysenck, The Structure of Human Personality
(London: Methuen, 1953); R. S. Woodworth, Psychology
(New York: Henry Holt, 1940); M. Hartshorne, M. A. May
and J. B. Maller, Studies in Service and Self-Control
(New York: Macmillan Co., 1929).

22G. R. Thornton, "The Use of Tests of Persistence
in the Prediction of Scholastic Achievement," Journal of
Educational Psychology, 32 (1941), 266-273; R. S. Mac-
Arthur, "An Experimental Investigation of Persistence in
Secondary School Boys," Canadian Journal of Psychology
9 (1955), 42-54; T. H. Howells, "An Experimental Study
of Persistence," Journal of Abnormal and Social
Ppychology, XXVIII (1933), 14_29.

 

 

 

 

 

 

 

 

23Woodworth, op. cit.

43

persistence to academic achievement have been summarized by
Thornton,2u MacArthur25 and Eysenck.26

The general paradigm of research concerned with
persistence is that in which a person is presented with an
extremely difficult or impossible task and is unrestricted
in either the time or number of attempts he can work on the
task. Persistence is usually measured by the total time
spent or the number of attempts made before the subject
turns to an alternative activity.

A study by Howells27 is indicative of the type of
research concerned with the role of persistence in the
prediction of academic achievement. Howells found the
correlation between scores on a battery of persistence
tests (such tests as "the length of time holding one foot
in the air" and "time spent studying a page of printed
materials to answer later test questions") and average
past grades for one hundred college students to be r = .44.
The relationship between intelligence test scores and grades
for the same students was r = .51, while the multiple corre—

lation of intelligence and persistence with grades was

r = .64.

fl

 

2“Thornton, op. cit.
25

MacArthur, op. cit.
26Eysenck, op. cit.

27Howells, op. cit.

44

While the procedure and purpose of the study by

Howells was typical of many studies concerned with persis-

tence as a personality trait, the coefficient obtained

between persistence test scores and grade averages indicates

28

a stronger relationship than is usually found.

Some of the conclusions reached by MacArthur

29 based

on a comprehensive review of early studies of persistence

are worth noting:

1.

Correlations between persistence tests have
usually been low but positive.

Persistence measures usually have very low posi-
tive relationships with intelligence and age,
up to the young adult level.

Success at a specific task encourages persistent
behavior at the task, but a mixture of success
and failure at initial stages may further
increase persistence.

Persistence in group situations may be affected
by the performance of other members of the
group.

Initial motivation influences performance on
persistence tests.

Persistence is approximately normally distributed
in unselected populations.

Persistence measures are usually related to school
success, the relation being roughly in proportion
as the measures resemble school situations.

 

28

MacArthur, Op. cit.

29Ibid.

45

30 has also emphasized the importance of the

Thornton
similarity between persistence test situation and school
situation for the prediction of school performance (Mac-
Arthur's final conclusion listed above). Thornton specifies
two aspects of the similarity between persistence test
situation and school situation: (1) Similarity of tasks and
(2) similarity of social relationships.

Thornton has argued that the means of increasing the
relationship between persistence test scores and indices
of academic achievement "is not by making the tests better
measures of persistence but rather by planning the test
situations to resemble more closely the tasks and social

"31 Thornton concludes

relationships found in the classroom.
that '“the best test situation for determining a student's
reactions to classroom tasks and classroom social rela-
tionships would be the classroom itself."32

Feather33 in a recent review of studies of persistence
places the type of study reported by Howells, Thornton and
MacArthur in a class of persistence studies concerned with

persistence as a trait or uniformity in behavior. Studies

in this class attempt to locate consistencies in the

 

30Thornton, op. cit.

3lIbid., p. 270.
321bid., p. 271.

33Norman T. Feather, "The Study of Persistence,"
Psychological Bulletin, 59 (1962), 94-115.

46

behavior (persistence) of subjects across a variety of
laboratory tasks and then seek to establish relationships
between laboratory behavior and some measure of "worth
while" accomplishment, e.g. school achievement. Consis-
tency in behavior across tasks and situations is thought
to indicate a relatively stable personality character-
istic.

Feather criticized these "persistence as a person-
ality trait" studies for their failure to account for the
influence of task and situation variables on behavior.
Thornton has discussed a similar point, above, in his
recognition of the importance of the similarity between
predictor and criterion Situations. However, Feather's
criticism goes beyond that concerned with a similarity that
is constant for all subjects between these two situations
by suggesting that individual differences in personality

34

may interact with situational differences. Atkinson has
suggested that simultaneous consideration of both differ-

ences in personality and differences in situational influ-
ences would provide for a more adequate test of explanatory

concepts and for increased precision in the prediction of

behavior.

 

3“Atkinson, op. cit.

47

Feather has described Atkinson's35

theory of achieve-
ment motivation as an approach which considers the inter-
action of stable personality dispositions (motives) and
situationally-defined variables (expectations and incentive
values) in the prediction of behavior. Persistence, for
Feather and Atkinson becomes a behavioral phenomenon to be
explained by a theory of motivation rather than as a trait
holding explanatory import.

In the following sections, the literature is reviewed
in which persistence is viewed as a behavioral phenomenon
in the laboratory, classroom and, as defined by attrition,
in the college situation.

Persistence as a Motivational Phenomenon

in the Laboratory
36

Feather has investigated the relationship of persis-

 

 

tence on a laboratory task to both its apparent difficulty
(situational variable) and the relative strength within the
individual of the motives to achieve success and to avoid
failure (personality variables).

Feather presented 89 male undergraduates with a series
of "Perceptual Reasoning Tests" (unicursal puzzles) intro-

duced as measures of "important skills and abilities." All

 

35J. W. Atkinson, "Motivational Determinants of Risk-
Taking Behavior," Psychological Review, 64 (1957), 359-372.

 

36Feather, Op. cit.

48

subjects experienced failure on the first puzzle (objec-
tively insoluble) and were allowed to "try again" (persist)
by selecting another card from a pile containing COpies of
the same geometric figure pp "move on" (not persist) by
selecting a card from a second pile of similar geometric
figures which might be either "easier" or "more difficult."

Apparent difficulty of the initial task was manipu-
lated by the use of "fictitious norms" designed to lead
half of the subjects to expect the first task to be "easy"
(70% chance of success) and half of the subjects to expect
the first task to be "difficult" (5% chance of success).
Subjects were classified as relatively strong in the Motive
for Success or relatively strong in the Motive to Avoid
Failure on the basis of discrepant appearance in the sepa-
rate distributions of TAT n-achievement scores and TAQ (Test
Anxiety Questionnaire--developed by Mandler and Sarason)
scores (above the median on one test and below the median of
the other).

Feather assumed total motivation to perform the
initial task to be a consequence of the resultant summation
of extrinsic motivation (non-achievement related motives,
eXpectancies and incentives) and achievement motivation
(achievement related motives, expectancies and incentives).
Change in the resultant total motivation to perform the

initial task was attributed to changes in achievement

49

related expectations (apparent task difficulty) as the
subject worked unsuccessfully on the initial task.

Feather reasoned that when the individual strong in
the Motive for Success is presented with an "easy" task,
his expectations are high (above .50 probability) that
he will attain success. Repeated failure at the task then
tends to lower his expectations concerning success with a
consequent increase in motivation (the result of the product
of decreasing expectations and increasing incentive values)
to continue performance (persist) on the initial task.

When the individual's expectations concerning success
fall to within the "intermediate" range of apparent diffi-
culty (50% chance of success) his motivation to continue
(persist) is maximum (the product of Expectancy and Incen—
tive values is highest) and individual differences in
motive strength have their maximum influence on behavior
(the product of Expectancy and Incentive values maximally
enhances differences in motive strength).

Not until the individual's expectations concerning
success decrease (due to repeated failure) below the "inter-
mediate range of apparent difficulty," does motivation to
continue work on the initial task begin to decrease and at
some point become weaker than his motivation to work on the

alternative task.

50

In contrast, if the individual strong in the Motive
for Success is presented with a "difficult" task then his
expectations concerning success are already at a low level
(below the "intermediate" range of apparent difficulty)
and failure experiences immediately begin to weaken the
tendency to persist.

A similar line of reasoning was used concerning the
behavior of individuals strong in the Motive to Avoid
Failure, when the task was "easy" or "difficult" and
assuming that the "intermediate" range of apparent diffi-
culty of the task was the range of maximum motivation to
93219 performance of the task.

Feather predicted that individuals strong in the
Motive for Success would persist longer when the task was
presented as "easy" rather than as "difficult" and would
persist longer on the "easy" task than subjects strong in
the Motive to Avoid Failure. Conversely, those subjects
strong in the Motive to Avoid Failure were expected to
persist longer when the task was presented as "difficult"
rather than as "easy" and to persist longer on the
"difficult" task than those subjects strong in the Motive
for Success.

Feather failed to find significant differences in

persistence for subjects differing 1J1 either motivation

51

(relative strength of the Motive for Success and Motive to
Avoid Failure) or in expectations ("easy" or "difficult")
of success, alone.

However, Feather did find a significant triple
interaction effect for motivation x expectation x persis-
tence (X2 = 7.89, with 3 df, p < .05).

Feather interprets his findings in support of his
contentions that persistence increases for subjects strong
in the Motive for Success, but decreases for subjects
strong in the Motive to Avoid Failure, as tasks vary from
"difficult" to "easy."

However, Feather's failure to obtain significant
relationships between achievement motivation, when the
Motive for Success and the Motive to Avoid Failure con-
structs were interpreted as TAT n-achievement scores and
TAQ scores respectively, and persistence as time spent on
an achievement task, provides no support for Atkinson's
theory nor for the techniques of motive assessment.

Feather37 has given further attention to the influ—
ence of the apparent difficulty of a task on persistence
and to the assumption that a task of an "intermediate"
level of difficulty maximizes the expression of achieve-

ment related motives in behavior.

 

37Feather, "Persistence at a Difficult Task. . .,"
op. cit.

52

Feather noted that in the experiment above the diffi-
culty level of the alternative task was not specified. He
assumed that differences in persistence on the initial task
were accounted for by changes in apparent difficulty as
subjects worked unsuccessfully on the initial task. However,
changes in apparent difficulty were assumed, on the basis
of theory, to influence motivation for the initial task
relative to motivation for the alternative task.

Feather hypothesized that an individual relatively
strong in the Motive for Success should terminate work on
a difficult task (5% chance of success) sooner, when the
alternative task is one of an "intermediate" level of
difficulty (50% chance of success), than subjects relatively
strong in the Motive to Avoid Failure.

Since the principle of change in motivation was based
on changes in the apparent difficulty of the task, as in the
study above, Feather also hypothesized that individual
differences in apparent difficulty (subjective probability
of success) should also influence persistence on the initial
task.

Feather presented 60 male undergraduates with the
perceptual reasoning task (unicursal puzzle) used in the
above study. Fictitious norms were used to describe the
initial task as difficult (5% chance of success) and the
alternative task as one of an "intermediate" level of

difficulty (50% chance of success).

53

After reporting the fictitious norms, estimates of
the individual's own subjective estimate of his probability
of success were obtained by having the subject check on a
20-point rating scale numbered from 0 to 100 in steps of
five.

Subjects were classified as relatively strong in the
Motive for Success if they scored above the median of TAT
n-achievement scores and below the median of TAQ scores.
Conversely, subjects were classified as relatively strong
in the Motive to Avoid Failure if they scored above the
median of TAQ scores and below the median of TAT n-
achievement scores.

Feather failed to find support for the expected
greater persistence of subjects relatively strong in the
Motive to Avoid failure, than subjects relatively strong in
the Motive for Success, when both groups were presented
with a "difficult" task and an alternative task of an "inter-
mediate" level of difficulty.

However, Feather found a significant relationship
between individual differences in apparent difficulty (sub-
jective estimates) and persistence on the initial task.
Using all 60 subjects tested, Feather found a statistically
significant tendency (Chi square = 8.08, df = 1, p <
.01) for subjects with high initial probability estimates
(above the median) to persist longer at the initial task

than subjects with low initial probability estimates.

54

Further analysis considering the relative strength of
the achievement-related motives indicated a positive rela-
tionship (p < .005, Fisher exact test, one-tailed)
between initial estimates of probability of success and
persistence on trial 1 for those subjects relatively strong
in the Motive for Success. No such relationship was found
among those subjects relatively strong in the Motive to
Avoid Failure.

Feather's findings suggest that his attempts to
manipulate the apparent difficulty level of the tasks by
the use of "fictitious" norms were not uniformly effective
for all subjects. Individual differences in subjective
estimates of the difficulty of the task remained (as for
the total group) related to persistence among those subjects
relatively strong in the Motive for Success.

38

The two studies reported by Feather involve measures
of persistence (motivation) in a laboratory situation in
which the subject experiences repeated failure on an
achievement-oriented task with an alternative task available
that is plpp achievement-oriented.

39

Weiner presents a study which involves persistence

(motivation) in a laboratory situation in which the

 

38Feather, "The Relationship of Persistence. . .",
op. cit. and "Persistence at a Difficult Task. . ." op. cit.

39Bernard Weiner, "The Effects of Unsatisfied
Achievement Motivation on Persistence and Subsequent
Performance," Journal of Personality, 33 (1965), 428-442.

 

55

alternative to an initial achievement-oriented task is a
task appealing to motives ptppp than achievement—related
motives. Such a situation is thought to hold special
relevance for testing the implications of an "inertial
tendency," i.e. a goal-directed tendency that persists until
satisfied.

Atkinson and Cartwrightl4O have proposed that in a
situation in which a tendency to attain some particular
goal (e.g. success) is aroused and thwarted, an "inertial
tendency" persists, i.e. a general tendency toward that
class of goal. The "inertial tendency" is thought to
equally enhance performance of an initial and alternative
task when both are achievement tasks (the situation

explored by Feather). However, the "inertial tendency"

should selectively influence persistence in achievement-

 

oriented activity following failure (thwarting), rather than
success, and this influence should be particularly apparent
when the alternative is to engage in non-achievement
oriented activity.

“1 had 60 male college students engage in a

Weiner
non-achievement oriented activity (asked to state prefer-

ences for various advertisements) then interrupted them and

 

“OJ. W. Atkinson and D. Cartwright, "Some Neglected
Variables in Contemporary Conceptions of Decision and'
Performance," Psychological Reports, 14 (1964), 575—590.

41

 

Weiner, op. cit.

56

presented an achievement-oriented activity (digit-symbol
substitution). Half of the subjects were told that this
"test of important skills and abilities" was difficult (30%
chance of successful completion) and given cards containing
too many substitutions to allow success on any trial
(failure condition). Half of the subjects were told the
task was easy (70% chance of successful completion) and
were subsequently given cards containing enough substitu—
tions to allow success on every trial (success condition).
Subjects were allowed to return to the non-achievement

task (advertisement preferences) at any time. Measures of
performance level (actual time to complete a uniform number
of substitutions on each trial) and persistence (number of
trials before returning to the non-achievement task) were
obtained. Relative strength of achievement—related motives-
was assessed by subtraction of "Z" scores obtained from the
TAT and TAQ instruments.

Weiner's results failed to attain statistical signif-
icance except in those analyses involving so few, highly
selected subjects, as to raise considerable question con—
cerning the influences producing his findings. Weiner sug-
gested.that failure to find significant differences in the
behavior (persistence) of groups differing in achievement-
related motives, may have been due to expectations by
members of both groups that the experimenter "was more

interested in performance at the substitution task, and a

57

'good' S was one who continued at the activity in progress."
In support of this argument, Weiner notes that 23 of the 60
subjects did not voluntarily quit the substitution task.

Strengths and Weaknesses of Laboratopy
Studies of Persistence

 

 

The value of the studies by Feather and Weiner lies
in the opportunity afforded by the laboratory situation to
specify, control and manipulate theoretically important
variables within that situation. Such explicit operations
are generally thought to provide greater clarity for the
relationships obtained.

The problem for the studies by Feather and Weiner
concerns the generality of their findings for Situations
outside the laboratory. The manipulation of variables,

e.g. the apparent difficulty of the task, assumed the exper—
ience of the subjects outside the laboratory is irrelevant

to performance of the tasks within the laboratory, an assump-
tion which restricts the findings of the above authors to
relatively few competitive situations (those for which the
subjects'previous experience is irrelevant). The results of
Feather's study indicate his attempts to manipulate the
apparent difficulty of the task were only partly successful
and suggests that his subjects may have drawn on their pre-
vious eXperience outside the laboratory to make their own

estimate of the difficulty of the task.

58

Persistence as a Motivational Phenomenon
in the Classroom

Atkinson and Litwin“2 have investigated relationships

 

between individual differences in achievement-related
motives and both persistence and achievement in the class—
room situation.

Atkinson and Litwin assumed that individuals rela-
tively strong in the Motive for Success should remain longer
in a situation that is potentially rewarding for them than
individuals relatively weak in the Motive for Success.
These authors further assumed a college course final exam-
ination to be an achievement situation offering potential
satisfaction for the Motive for Success while activities
immediately outside this situation would appeal to non-
achievement related motives.

Atkinson and Litwin obtained TOI n—achievement scores
for 44 male college students. Edward's PPS n—achievement
scores and TAQ scores were also obtained. The achievement
task consisted of a multiple—choice and short-answer final
examination for a course in Psychology. Persistence scores
(time spent in the examination room) were obtained as the
students left the test situation presumably to engage in
non-achievement related activities. Grades attained on the

final examination were also available.

 

u2John W. Atkinson and G. H. Litwin, "Achievement
Motive and Test Anxiety Conceived as Motive to Approach
Success and Avoid Failure," Journal of Abnormal and Social
Psychology, 60 (1960), 52-63.

 

 

59

Atkinson and Litwin found that those subjects above the
median of TOI n-achievement scores spent a greater amount of
time in the examination room than those subjects below the
median (U = 156.5 p < .03). Furthermore, those subjects
above the median of TOI n-achievement scores obtained higher
grades than those subjects below the median (U = 152.5,

p < .02). The relationship between the number of minutes
Spent working on the final examination and the grade
obtained (two behavioral measures of motivation) was not
significant (r = .27, p < .10).

Atkinson and Litwin found that when subjects were
simultaneously classified on TOI n-achievement and TAQ scores
the High TOI n-achievement-Low TAQ scoring group (assumed
relatively strong in the Motive for Success) spent a longer
time on the examination and obtained higher grades than the
Low TAT-High TAQ scoring group (assumed relatively strong in
the Motive to Avoid Failure). No significant relationships
were found between EPPS n-achievement scores and time spent
on the examination or examination grades.

Atkinson and Litwin conclude that the relationships
between TOI n-achievement scores and persistence and grades
on the examination, and between TAQ scores and these vari—
ables, contribute evidence for the construct validity of
their theory of motivation and for the use of the T01 and
the TAQ as interpretations of the Motive for Success and

Motive to Avoid Failure, respectively.

60

The failure of Atkinson and Litwin to find significant
relationships between the Motive for Success and either per-
sistence or examination grades when EPPS n-achievement
scores were used as an index of the Motive for Success, was
taken by these authors as evidence that such "self-report"
measures do not reflect the operation of their theoretical
construct.

Smith“3 attempted to replicate the findings of
Atkinson and Litwinuu using intelligence test scores as an
index to the apparent difficulty of an examination and pre-
dict relationships between achievement-related motives and
persistence on the basis of the findings in Feather's

study.”5

Smith”6 failed to find significant relationships
between TOI n—achievement scores and Otis gamma intelligence
test, examination grades or time spent on either a mid-term
examination or a final examination for 146 undergraduate
students.

However, Smith did find significant relationships

between a measure of "resultant motive strength" (TOI "Z"

 

u3Charles P. Smith, "Relationships Between Achievement-
Related Motives and Intelligence, Performance Level, and
Persistence," Journal of Abnormal and Social Psychology,

68 (1964), 523-533.
uuAtkinson and Litwin, op. cit.
uSFeather, "The Relationship of Persistence . . ."

op. cit.

u6Smith, op. cit.

61

scores minus TAQ "Z" scores) and intelligence (r = .18,

p < .05) and time Spent on a mid-term exam (r = -.l8,

p < .05). These results imply that subjects with rela-
tively stronger Motive for Success obtain higher intel-
ligence test scores and leave the examination room earlier
than subjects relatively strong in the Motive to Avoid
Failure.

The findings by Smith concerning persistence were
directly contrary to those of Atkinson and Litwin.u7 Smith
suggested, however, that it might be possible to reconcile
these discrepant findings through consideration of estimates
of apparent difficulty, the situational variable shown by
Feather to influence the relationship between personality
(motives) and motivation (persistence).

Smith reasoned that:

. .-.subjects with relatively high Motive for Success
will persist longer at a task of intermediate diffi-
culty than subjects with relatively high Motive to
Avoid Failure (the result of the Atkinson-Litwin
study) but will persist less long 8% a very easy task
(the result of the present study).

While no measure of perceived difficulty was used in
the study by Atkinson and Litwin or in the study by Smith,
the former authors had suggested that differences in ability

may have influenced their results.

“7Atkinson and Litwin, op. cit.

uaSmith, op. cit., p. 527.

62

Smith, therefore, assumed that differences in the
apparent difficulty of an examination might be indexed by
individual differences in intelligence test scores. That
is, subjects high in intelligence might be expected to View
the examination as "easy" while subjects low in intelli-
gence might be eXpected to view the examination as
"difficult."

Smith found a significantly greater (negative) rela-
tionship between resultant strength of the Motive for
Success and persistence on the final exam among subjects
with high Otis scores than among subjects with low Otis
scores (Z = 2.01, p < .05, two-tailed test).

Smith interprets his findings in support of the notion
that among subjects presented with an "easy" task (the high
Otis score group) those subjects relatively strong in the
Motive for Success tend to be less persistent than subjects
relatively strong in the Motive to Avoid Failure. However,
among subjects presented with a "difficult" task (the low
Otis score group) those subjects strong in the Motive for
Success tend to be more persistent than subjects relatively
strong in the Motive to Avoid Failure.

While the results of the study by Smith appear to
require explanation through a fortuitous manipulation of the
terms of theory, there is evidence to suggest that perceived

difficulty of a task influences the relationships between

63

achievement-related motives and persistence, or, at least,
that a measure of intelligence acts as a "moderator
variable" (i.e., a third variable that influences relation-
ships between two other variables) in the prediction of
persistence from indices of achievement-related motives.

The results of the study by Smith appear to warrant
further study using both measures of ability 229 of per-
ceived difficulty such that the assumed relationship between
these measures can be tested as well as their implications
for the prediction of behavior.

Strengths and Weaknesses of Studies of
Persistence in the Classroom

The value of the studies by Atkinson and Litwin and
by Smith lies in the selection by these authors of a com-
petitive task and situation common in the experience of
their subjects; a task for which the past experience of the
subjects is relevant and a situation which allows the
application of their findings to theory gpd the generali-
zation of their findings to frequently occurring, important
situations.

The limitations of the studies by Atkinson and Litwin
and by Smith have been discussed by Atkinson“9 in which

Atkinson argues that performance in most "real-life"

 

ugJohn W. Atkinson and W. R. Reitman, "Performance
as a Function of Motive Strength and Expectancy of Goal
Attainment," Journal of Abnormal and Social Psychology,
53 (1955), 351-366.

64

situations is "over-determined," i.e. performance is the
consequence of many factors, such that relationships
between measured variables are often slight or "washed-
out" entirely.

It would appear that Atkinson and Litwin and Smith
might have increased the precision in their studies by
closer attention to the control of variables assumed to be

50 and Weiner.51

so important in the studies by Feather
Neither Atkinson and Litwin nor Smith used a measure of the
apparent difficulty of the task (exam) even though both
authors assumed that such a variable would be expected to
influence performance. Only Smith attempted some degree of
control of the difficulty factor by the use of a measure of
intellective ability. Furthermore, neither author attempted
to Specify, control or manipulate the alternative activity
to which their subjects could turn, but only assumed that
the alternative activity involved different motivations

than performance in the examination room.

Studies of Attrition in the
Academic Situation

 

Heilbrun52 has investigated the influence of person-

ality on continuation (persistence) in college following the

 

_,.

50Feather, "The Relationship of Persistence . . .," pp.
cit. and "Persistence at a Difficult Task . . .,' 'op. cit.

51Weiner, op. cit.

52Alfred B. Heilbrun, Jr., "Personality Factors in
College Dropout," Journal of Applied Psychology, 49 (1965),
l- 7

 

65

freshman year. Heilbrun assumed that both ability and per-
sonality would influence attrition and that situational
conditions at a large university might pose particular
adjustment problems for students of high ability.

Heilbrun matched 304 male students and 306 female
students who returned for the SOphomore year with a student
of the same sex and academic ability test score who did not
return. He compared mean scale scores on the Adjective
Check List between students who did not drop out with those
who did drop out within each of three levels of ability
(bottom 43%, middle 32% and top 25%).

Heilbrun found significant mean scale score differ-
ences primarily among the high ability students of each sex.
The personality variables found to influence attrition of
high ability males included n-achievement which occurred
along with n-order and n—endurance in a group of scales
which Heilbrun had classified as "task relevant" behaviors
indicating problems of "specific conformity to academic
values." Heilbrun concluded that "personality makes an
important systematic contribution to college attrition for
high ability students only . ..."53

Heilbrun emphasized that his results are based on
findings in a large university setting and might not be

replicated in a smaller college setting which might present'

 

53Ibid., p. 4.

66

different environmental demands. These results appear to
support the contentions of Smith5u and the theoretical

55 that personality and situational

notions of Atkinson
variables interact and that the relationships between per-
sonality (motives) and motivation (persistence) may be
influenced by measures of ability.

56 has also investigated the relationship

Spielberger
between personality and college attrition (persistence)
covering a three year period. Spielberger assumed that:
"college life is characterized by conditions and expecta-
tions which may heighten anxieties already present in
students or may induce new anxieties."57

Spielberger obtained ACE scores, MAS scores and an
index of drOpout rate due to academic failure for 267 male
college students. Relationships between MAS scores and
attrition over a period of three years were investigated
within three levels of academic ability (lower 20%, middle
60% and upper 20%).

Spielberger found that, excluding those students in
the lowest 20% of ability, 18% of the relatively able high

anxiety students had dropped out of the university due to

 

5”Smith, op. cit.-

55Atkinson and Litwin, op. cit.

56C. D. Spielberger, "The Effects of Manifest Anxiety
on the Academichchievement of College Students," Mental
Hygiene, 46 (1962), 420-426.

57Ibid., p. 421.

67

academic failure while only 4.5% of the low anxiety students

had dropped out for this reason.

58 59

The above studies by Heilbrun and Spielberger
reflect something of the current interest in exploring
relationships between personality variables (motives or
needs) and criterion measures (e.g. persistence) which hold
greater relevance for the personality constructs involved
than criteria of academic achievement (GPA) alone. Atten-
tion is also directed in these studies to importance of
control of ability factors which may tend to alter or
obscure the relationships of personality variables.
Similarities and Differences in the
Study of Persistence

The studies of persistence as a motivational phenom-
enon reviewed above, represent attempts to explore relation-
ships between individual differences in personality and'
behavior in competitive situations when attention is given
to the characteristics of the situation in which personality
is expressed and behavior observed.

The findings of the above studies in which persistence

is treated as a dependent variable, are thought to hold

:hnplications for the validity of certain constructs

58Heilbrun, op. cit.

59Spielberger, op. cit.

68

occurring in Atkinson's6O theory of achievement motivation.
However, the results of these studies are varied and often
appear contradictory.

Variation and apparent contradiction in findings may
be taken as non-confirming evidence for theory, however it.
must be recognized that each of the above studies represents
a particular instance of investigation involving a selected
approach to the assessment of certain personality traits,
restrictions in the selection of subjects and the selec-
tion of a particular situation in which behavior takes
place. Each of these variables--tests, subjects and
situations--may influence the findings of a particular study
in ways not mentioned in theory and differences in these
variables make comparison of results hazardous.

A brief comparison of the above studies follows in
which similarities and differences in personality measure-
ment, subjects and behavioral situations amediscussed. The
purpose of the following comparison is to draw attention to
the variety of motivationally relevant influences operating
in even these few studies. Recognition of the potential 5
influence of these variables is thought to be a necessary
precursor to attempts to distinguish apparent contradictions
in findings, which tend to disconfirm theory, from differ-
ences in the conduct of studies which tend to restrict the
generality of findings to disparate spheres of theory

discourse.

k

6OAtkinson, "Motivational Determinants. . .,"op. cit.

69
The Measurement of Individual Differences
in Achievement Related Motives

The studies of persistence as a motivational phenom-
enon reviewed above, used scores derived from six different-
instruments to index the operation of personality variables
thought to influence performance in competitive situations.
The instruments used in the above studies were: the Thematic
Apperception Test (TAT), the Test of Insight (TOI), the Test
Anxiety Questionnaire (TAQ), the Manifest Anxiety Scale
(MAS), Edward's Personal Preference Schedule (EPPS) and the
Adjective Check List (ACL).

Two of these instruments (TAT and T01) are projective
methods of measurement while the remaining four instruments
(TAQ, MAS, EPPS and ACL) represent objectively scorable
self—report inventories. Four of these instruments (TAT,
TOI, EPPS and ACL) provide scores assumed to reflect the
strength of an approach motive (Motive for Success) which
is thought to facilitate performance in competitive situations.
Two of the instruments (MAS and TAQ) provide scores assumed
to reflect the strength of an avoidant motive (Motive to Avoid
Failure) which is thought to interfere with efficient perfor-
mance in complex competitive situations. In addition, two
indices, thought to reflect the resultant of conflict between
the Motive for Success and the Motive to Avoid Failure, were
derived by combining scores (from discrepant halves of sep-
arate scale score distributions or by subtraction of "Z"

scale scores).

70

Statistically significant differences in performance
attributable to differences in personality (motives) were
found by Atkinson and Litwin61 using scores derived from
one projective instrument (TOI); using scores derived from
one self-report instrument (TAQ); and using scores derived
from a combination of both instruments (scores above the
median on the TOI and below the median of the TAQ, and the
reverse); but failed to find significant differences using
scores derived from a second self-report instrument (EPPS).

Smith62 failed to replicate the findings of Atkinson
and Litwin using the same methods of measurement and instru-
ments (TOI and TAQ) used by Atkinson and Litwin but using a
different method of combining scores ("2" score subtrations).

63 failed to replicate his own earlier findings6u

Feather
using identical methods of measurement (projective and self-
report), instruments (TAT and TAQ) and score combination
technique (scores above the median of the TAT and below the
median of the TAQ, and the reverse).

Weiner65 failed to find significant performance dif-

ferences attributable to motive classification using scores

 

61Atkinson and Litwin, op. cit.

62Smith, op. cit.

63Feather, "Persistence at a Difficult Task. . .,"
op. cit.

6“Feather, "The Relationship of Persistence. . .,"'
op. cit.

65Weiner, op. cit.

71

derived from the methods and instruments (TAT and TAQ) used
by Feather but using a different score combination technique
("2" score subtractions). Both Weiner and Feather report
their findings only for the-combined score technique.
Although Atkinson and Litwin failed to find signifi-

cant differences in persistence related to scores on one
self-report instrument (EPPS) both Heilbrun66 and Spiel-

67

berger obtained significant relationships between measures
of college attrition and self-report inventory scores (ACL
and MAS, respectively)..

The few studies of persistence reviewed above, indi-
cate something of the variety of methods, instruments and
score combination techniques used for the assessment of
achievement-related motives. The differences between
studies in-the measurement of personality and the varia—
bility of the findings preclude drawing conclusions con-
cerning the superiority of one method, instrument or score
combination technique.

However, the failure of various authors to replicate
the findings of previous research combined with the rela-

tively low relationships that have been obtained suggests

caution in the interpretation of results as indicating the

 

66Heilbrun, op. cit.

67Spielberger, op. cit.

72

Operation of theoretical constructs independent of the

particular instance of measurement.

Subjects

Research by Atkinson68 and his co-workers69

has shown
differences in the responses of male and female subjects on
personality tests and differences in the relationships
between the test responses of these two groups and behavior
on a wide variety of performance criteria. These findings
by Atkinson and his co-workers have resulted in the concen—
tration of their efforts on research typically involving
only male subjects.

70 and his co-workers anticipated sex differ-

Farquhar
ences in the assessment of personality by the construction
of separate instruments for males and females and by separate
analysis, by sex, of behavior on performance criteria.

All of the studies of persistence as a motivational
phenomenon reviewed above, have used college students as
subjects. All of the above studies but one, Smith,71 have

used only college males or have analyzed the results for

males and females separately.

 

68John W. Atkinson (ed.), Motives in Fantasy Action and
Society, D. Van Nostrand Co., Inc., 1958.

69John W. Atkinson and Norman T. Feather (eds.), A
Theory of Achievement Motivation (New York: Wiley and Sons,
Inc., 1966).

70
71

 

Farquhar, op. cit.
Smith, op. cit.

73

Smith failed to replicate the findings of Atkinson and

Litwin72

concerning the relationships between personality
(motives) and behavior (persistence) in the classroom situa-
tion. However, Smith included both college males and
females in his analyses while Atkinson and Litwin used only
data based on college males (though it appears likely
females were present in the examination situation investi-
gated by the latter study). The possible influence of
differences in subjects included in the two studies compli-
cates direct comparison and evaluation of their-contradic-
tory findings. The locus of differences could be in per-
sonality assessment, criterion behavior or differences in
relationships between test and task measurements.

Differences attributable to sex were found by Heil-
brun.73 Heilbrun separately analyzed results for male and
female college students and found twice as many of the
Adjective Check List scale scores to significantly discrim-
inate between male college students, than between female
college students, who did and who did not drop out of
college after the first year.

The findings of the above studies and the conclusions

of various reviewers suggest caution in the comparison of

studies using only male subjects with studies using combined

 

72Atkinson and Litwin, op. cit.

73Heilbrun, op. cit.

74

data from male and female subjects. There is further
concern that results based on college subjects may not
generalize directly to high school subjects or older

employedsadults.

The Situation

Atkinson's?“ theory of achievement motivation attaches
importance to one situational characteristic (apparent
difficulty of the task) which is thought to differentially
influence the expression of stable personality traits
(motives) in competitive activity (persistence).

Each of the studies reviewed above has given consid-
eration to the "apparent difficulty" of the task. However,
there are wide differences in the manner and extent each
study has dealt with this aspect of the situation. In
addition, there are differences in other aspects of the
situations used in the above studies—-group or individual
performance, novelty of the task, availability of knowledge
of results, and type of alternative activity--which may also
influence performance in competitive situations and limit

the generality of particular findings.

 

7“Atkinson, "Motivational Determinants. . .," Op. cit.

75

Apparent Difficulty

75 76

Feather and Weiner manipulated the apparent
difficulty of a task by the use of "fictitious norms"
designed to present the task as "easy" or "difficult."-
While their manipulation of this variable allows clarity in
the findings obtained by these specific operations, it
prevents consideration of the influence of previous exper-
ience, or of consistency in relationships between past

and present experience, on task performance.

In contrast, Atkinson and Litwin77 and Smith78
assumed that measures of ability might index initial expec-
tations of the apparent difficulty of an examination. Such
an index was assumed to be related to the previous exper—
1ence of college students with "easy" or "difficult" exam—
inations and allow consistency in the relationship between
previous experience and present expectations. However, only

Smith tested this assumption and attempted to relate his

findings to theory.

 

75Feather, "The Relationship of Persistence. . .," pp,
cit. and "Persistence at a Difficult Task . . .," op. cit.

76
77
78

Weiner, Op. cit.
Atkinson and Litwin, op. cit.

Smith, op. cit.

76

Heilbrun79

and Spielberger80 found measures of ability
to influence the relationships between personality and per-
sistence but neither author attempted to provide a system-
atic rationale for the differential prediction of these
relationships.

While each of the above studies appears to confirm the
importance of consideration of objective or subjective
measures of task difficulty, there appears to be little
basis for the comparison of findings obtained through such

disparate approaches to the control and manipulation of task

difficulty.

Group and Individual Performance

Feather81 and Weiner82 explored persistence in a

situation in which subjects worked individually in com-
petition with "fictitious norms." In contrast, the subjects
used by Atkinson and Litwin83 and Smith814 worked in groups

85

while the subjects in the studies by Heilbrun and

79Heilbrun, op. cit.
80

Spielberger, op. cit.
81Feather, "The Relationship of Persistence. . .," pp.
cit. and "Persistence at a Difficult Task. . .," op. cit.

82Weiner, op. cit.
83

Atkinson and Litwin, op. cit.
8“Smith, op. cit.
85Heilbrun, op. cit.

77

Spielberger86 can be assumed to have experienced a mixture
of individual and group competitive situations.
87

MacArthur has noted above that the performance of

individuals in groups has been found to be influenced by

other members of the group and suggests caution in comparing

the results of performance from individualemm.group situations.
While comparison of results obtained from group and

individual performance situations may be unwarranted it

would appear that each situation occurs in the school envir-

onment and findings should be applicable to these respec-

tive situatiOns. Experimental isolation of differences

within group situations and differences within individual

performance situations would appear to hold greater poten-

tial clarity than studies, e.g. those by Heilbrun and

Speilberger, in which the effects of both situations are

allowed to influence measures of outcome in unknown ways.

Novelty of the Task
The tasks used by Feather88 (unicursal puzzles) and
by Weiner89 (digit-substitutions) were unusual ones in the

previous experience of their subjects. While the use of

 

86
87

88Feather, "The Relationship of Persistence. . .," pp.
cit. and "Persistence at a Difficult Task. . .," op. cit.

89

Spielberger, op. cit.
MacArthur, op. cit.

Weiner, Op. cit.

78

such simple response tasks allows a considerable degree of
control over extraneous influences on performance it also
raises questions of the application or generality of findings.
The very fact that these tasks are "new and different" may
provide motivational influences which differentially affect
the performance of individuals in ways not considered by
theory and in ways which restrict the application of findings
to relatively few "real-life" competitive situations.

In contrast, the tasks used by Atkinson and Litwin90

and by Smith9l

(college examinations) were common tasks in
the previous experience of their subjects and tasks which
are likely to occur repeatedly in the future. The use of
commonly occurring tasks for the observation of differences
in motivation allowed previous eXperience to influence
performance in ways considered by theory and allows the
application of findings to at least these frequently
occurring "real-life" competitive situations. A

The problem in the use of "real-life" competitive
situations has been discussed by Atkinson.92 Atkinson

suggests that performance in such situations is "over—

determined," i.e. the consequence of the resultant influence

 

90Atkinson and Litwin, op. cit.

ngmith, op. cit.

92Atkinson and Reitman, op. cit.

79

of many factors, such that relationships between a rela-
tively few measured variables are often slight or "washed

out" entirely.

Knowledge of Results

93 94

Feather and Weiner provided subjects with imme-
diate knowledge of either "success" or "failure" throughout
initial task performance. Knowledge of competitive standing
was therefore available to influence persistence. Further-
more, both Feather and Weiner contrived a situation in which

subjects experienced either continuous success or continuous

 

 

failure.

In contrast, subjects included in the studies by
Atkinson and Litwin95 and by Smith96 did not have objective
knowledge of results at the time they left the examination
room and may be assumed to have varied in their subjective
estimates of relative degrees of success and failure as they
worked on the examination and at the point they ceased work.

The subjects used by Heilbrun97 and Spielberger98 can

be assumed to have had knowledge of relative degrees of

 

93Feather, "The Relationship of Persistence. . .," pp.
cit. and "Persistence at a Difficult Task. . .," op. cit.

9“Weiner, op. cit.

95Atkinson and Litwin, op. cit.

96Smith, op. cit.

97Heilbrun, op. cit.

98Spielberger, Op. cit.

80

success or failure at the point they either continued or
"dropped out" of college. However, the final event can also
be assumed to have been influence by an accumulation of
instances in which they persisted or failed to persist in
competitive situations without objective knowledge of
results. The influence of knowledge of results on contin-
uation or persistence is thus confounded in these latter

two studies.

The Alternate Activity
The characteristics of an alternative activity and

the relationship of that activity to an initial activity

99

holds considerable importance for Atkinson's theory of

achievement motivation which attempts-to account for
"change" in behavior on the basis of conflict between

various motives and their relevant expectancy and incentive

values.

_ Feather100 and WeinerlOl specified and controlled the

characteristics of an alternative activity for their sub-
jects. Feather Specified an alternative activity appealing
to the same motives (achievement-related motives) as an

initial task. In a second study, Feather also specified

 

99Atkinson, "Motivational Determinants. . .," op. cit.
lOOFeather, "The Relationship of Persistence. . .," pp.
cit. and "Persistence at a Difficult Task. . .," op. cit.~

101Weiner, op. cit.

81

the difficulty of the alternative activity. Weiner provided
an alternative activity (non-achievement related activity)
which was different than the initial activity (achievement
related).

Atkinson and LitwinlO2 and Smith103

assumed that
activity outside the examination room was nonrachievement
related while activity inside the examination room was
assumed to be only achievement related. Neither Atkinson
and Litwin nor Smith attempted to specify or control alter-
native activity for their subjects nor did they attempt to
provide any check on their assumption that the activity
inside the examination room appealed to different motiva-
tions than that immediately outside the room. Failure to
adequately control for the influence of alternative activity
necessitated a change in an experiment by Hartshorne and
Maylou when these authors discovered that persistence on
their task was being influenced by the motivation of those
subjects they expected to be high persisters, to leave the

laboratory situation to return to the performance of impor-

tant school tasks (behavior they had hoped to predict).

 

102Atkinson and Litwin, op. cit.
103Smith, op. cit.

lOuHartshorne,May and Maller, op. cit.

82

Moderator Variables and the Prediction
of Achievement

Two events in the field of industrial psychology,

106 107

discussed by Porter,105 Dunnette, Saunders, and

Ghiselli,108 hold implications for the present study.

The first event has to do with the findings of Dunnette
and Ghiselli that the performance of some individuals and
groups are more predictable than others and that these dif-
ferences in predictability can themselves be predicted. That
is, knowledge of a third variable can enhance the validity
of certain predictors for some individuals or groups.

109

Saunders has given the name "moderator" to such an

independent variable that influences the relationship between

 

105Lyman W. Porter, "Personnel Management," Annual
Review of Psychology, 17 (1966), 295-422.

‘ 106Marvin D. Dunnette, "A Modified Model for Test
Validation and Selection Research," Journal of Applied
'Psych61pgy, 47 (1963), 317-323.

107David R. Saunders, "Moderator Variables in Pre-
diction," Educational and Psychological Measurement, 16
(1956), 209—222.

108Edwin E. Ghiselli, "Moderating Effects and Dif-
ferential Reliability and Validity," Journal of Applied
Psychology, 47 (1963), 81-86.

logsaunders, op. cit.

83

another independent variable and a dependent variable. Both

110 have suggested that moderators may

Saunders and Berdie
operate by sorting a heterogeneous aggregation of individ—
uals into homogeneous groups.‘

Thus the findings of Farquharlll and Atkinson112 that
separate consideration must be given to the prediction of
the achievement-related behavior of males and females is an
example of the conception of a third variable, sex, oper-
ating to influence the relationships between personality
test scores and achievement.

The search for such moderator variables has largely
proceeded on a trial and error basis. Empirical or logical
methods have been used to specify rather obvious moderators,
e.g. age, sex, or level of education. Theory has been
little involved in attempts to specify in advance which
moderator variables might be useful.

Atkinson's theory Of achievement motivation appears
to offer one theoretical approach to the selection of a

potentially useful moderator. That is, Atkinson suggests

that characteristics of the particular achievement situation

 

110Ralph F. Berdie, "Intra-Individual Variability and
Predictability," Educational and Psychological Measurement,
21 (1961), 663-676.

lllFarquhar, op. cit.

112Atkinson, Motives in Fantasy. . .," Op. cit.

 

84

may moderate the relationships between measures of rela-
tively stable personality characteristics (motives) and
behavior (motivation). Specifically, Atkinson has theo-
rized that measures of the difficulty of a task should
enhance the predictability of achievement-related behavior
for individuals faced with a task of an intermediate level
of difficulty (based on the theoretical relationship
between expectancy and incentive values in competitive
situations).

The second event in the field of industrial psychol-
ogy closely associated with the selection of moderator
variables is the problem of criterion selection. Wallace113
has suggested that industrial psychologists have been too
concerned with proving the predictive utility of their
techniques and-have thus centered.on-a global criterion of
organizational worth, e.g. success or satisfaction. These
attempts have overlooked events intervening between gross
predictor and global criterion.

Wallace's criticism of the industrial psychologists‘
concern with "the criterion" can be appropriately leveled
at those psychologists using the "ubiquitous grade point
average" to validate personality tests. It is unreasonably

rare that personality test developers ask "what sorts of

 

113s. R. Wallace, "Criteria for What," (Presented at

Amgfitcan Psychological Association Meeting, Los Angeles,.
19 .

85

behavior should be most intimately related to the dimension
of personality I am trying to measure" and then, having
selected that behavior and evaluated its relationship to
personality, go a step further to evaluate the influence

of that behavior on some worthwhile outcome..

The work of Farquhar and Atkinson appears to offer
guide lines for a theoretical exploration of variables
which may intervene between gross predictor (personality)
and global criterion (GPA) in the academic situation.
Thus, the criteria most intimately related to a study of a
Motive for Success has been suggested by the above authors
to involve "the direction, vigor and persistence of

"11“

behavior or "the initiating, directing and sustaining of

behavior."115
It would appear that the above authors imply that

construct validity investigation for their instruments would
center on eXploring the relationships between test scores
and evidence of the initiation, direction, vigor and persis-
tence of behavior. Concern with indices of such behavior,
would relate to the practical problems of how adequately one
has sampled and measured such behavior and the theoretical

problems of the relationships between motivation and

quality of achievement.

r

ll”Atkinson, An Introduction to Motivation, op. cit.

115Farquhar, Op. cit.

86

The studies by Smith, Heilbrun and Spielberger,
reviewed above, represent attempts to employ a third variable
(ability) to moderate relationships between personality
(motives) and criteria of motivation (persistence on an exam-
ination or attrition in college) that at least may hold
greater relevance for a theory of motivation than the exclu-
sive use of achievement criteria (GPA).

However, some authors have also investigated relation-
ships between personality variables and level of achievement
criteria and allowed measures of ability to moderate these
relationships

Goodstein and Heilbrun116 have argued for the impor-
tance of using levels of ability as a control variable in
studies of non-intellective factors in achievement. They
suggest that the relationship between personality and some
measure of academic achievement may not be the same for all
levels of ability.

Goodstein and Heilbrun correlated scores on the EPPS
with semester grade averages for 206 male and 151 female
college sophomores. Initial analyses indicated approxi-
mately 20% of the EPPS scales correlated significantly with

a brief measure of verbal ability. When the variance

 

116Leonard D. Goodstein and Alfred B. Heilbrun, Jr.,
"Prediction of College Achievement from the Edwards
Personal Preference Schedule at Three Levels of Intellectual
Ability," Journal of Applied Psychology, 46 (1962), 317-
320.

 

87

attributed to ability was controlled by partial correlation
technique, the only significant correlation between EPPS
scales and GPA was for the n-Achievement scale.

However, when these authors further analyzed their.
data by level of ability (three equal—sized subgroups) with
the influence of ability within levels controlled by par-
tial correlation, they found over one-half of the scales on
the EPPS significantly related to GPA and the majority of-
these were within the middle ability group of male students.
Similar analyses for females did not appear to produce the
differential influence of personality by ability_1evel.

These authors interpret their findings in support of
the notion that personality factors are related to academic'
achievement when the influence of academic ability is.
statistically removed, but that the nature of the relation-
ships depends upon the general ability level Of the group
being studied.

Goodstein and Heilbrun further interpret their
findings as suggesting that personality factors are more
important in determining the success or failure of average
ability males while intellectual factors are more important
for both the relatively bright and relatively dull male
college students. Among their findings was a significant
relationship between the neAchievement scale of the EPPS
and GPA pply within the middle ability group of males,

P = .29 (p < 005)0

88

While these authors did not attempt to relate their
findings to a theory of-motivatiOn; it appears that the
relationship between n-Achievement scores and GPA would be
predictable by Atkinson's117 theory of motivation if it is
assumed that school tasks are typically perceived by the
average ability males as tasks within an "intermediate"
range of apparent difficulty. Atkinson's theory would
predict the strongest relationship between individual
differences in achievement-related motives and behavior for
tasks of an "intermediate" range of apparent difficulty.
This prediction would assume that GPA with_the influence of
ability statistically removed, represents a measure of
motivation.

Speilberger and Katzenmeyer118 have also investigated
the relationship between personality and GPA within three
levels of academic ability (lower 20%, middle 60% and
upper~20%).

Spielberger and Katzenmeyer obtained ACE scores, MAS
scores and GPA's over one semester for 640 male college
students. They report that tests for linear and curvilinear

regression indicated that GPA's were not related to MAS

scores for either the high or low academic ability groups.

 

117Atkinson, "Motivational Determinants. . .," Op. cit.v

118C. D. Spielberger and W. C. Katzenmeyer, "Manifest
Anxiety, Intelligence, and College Grades," Journal of
Consulting Psychology, 23 (1959), 278.

89

However, the test for linear regression for the groups of
an intermediate level of academic ability yielded an F of
13.06 (df = 1,39% p < .001).

Spielberger and Katzenmeyer concluded that academic
achievement (GPA) varies inversely with anxiety for the
students of an-intermediate level of academic ability.

They attribute failure by low ability students to their
limited ability and the success of students of superior
ability to the relative ease with which these students can
obtain grades irrespective of anxiety level.

While Spielberger and Katzenmeyer did not attempt to-
relate their findings to a theory of motivation, it appears
that these relations (as were those of Goodstein and
Heilbrun) are predictable by Atkinson's theory of achieve-

ment motivation.

Summary

The review of the literature was centered on studies
in which relationships between achievement-related motives
and the initiating, directing and sustaining of achievement--
related activity were interpretable by a theory of achieve-
ment motivation.

The importance of latency measures.in animal labora—
tory studies of basic need states was linked to the potene
tial importance of motivation as expressed by initiating

achievement-related activity in humans. The arbitrary

90

nature of the distinction between persisting and initiating
activity was cited and both related to a theory of motiva-
tion that attempts to describe and predict "change" in-
behavior.

Studies were also reviewed in which the direction of
expectations toward higher or lower levels of achievement
were viewed as influenced by both reality and motivational~
determinants with stronger influences ascribed to motiva-
tional determinants when reality determinants were controlled
or minimized.

Findings from laboratory studies using persistence as
a behavioral criterioncfl‘motivation emphasized the impor-
tance of attention to the "apparent difficulty" of success
and specification of an alternative-activity. While the
laboratory studies were able to manipulate both the diffi—
culty of the task and the nature of the alternative activity
their findings were thought to hold little relevance for
"real life" competitive situations.

Findings from studies of persistence as a behavioral
criterion of motivation within the classroom situation-
emphasized the importance of motivational criteria other
than level of academic achievement and demonstrated the.
influence of persistence on the level of achievement.

While studies of persistence in the classroom situation

gave attention to task difficulty and the nature of the

91

alternative activity in theory they failed to manipulate or
control these variables in their research.

Consideration of studies of persistence in both the
laboratory and in the classroom situation indicated varia—
tion in_findings which could be attributable to differences
in motive assessment techniques, subjects and the type_of‘
experimental situation including such variables as diffi-
culty of the task, group vs. individual performance,
novelty of the task, knowledge of results.and the nature of~
the alternative activity.

Attention was also focused on a current interest in-
the field of Industrial Psychology concerned with the
selection of variables which influence (moderate) relation-
ships between two other variables and which may "mediate"
relationships between gross predictors and global criteria.-

Empirical studies of attrition and achievement in the
academic situation were also reviewed with implications that
findings from these studies could be predicted by Atkinson's
theory of achievement motivation using academic aiblity
test scores as one index of the difficulty of success to
moderate relationships between personality and academic

achievement.

CHAPTER IIIA
-DESIGN OF THE STUDY

This study was designed to investigate evidence for-
the construct validity of the Motive for Success, (a) as
represented within Atkinson's theory of achievement moti-
vation,(b) as interpreted by scores on the Generalized.
Situational Choice Inventory (GSCI) and, (c) as expressed
through activity within the classroom situation.

The setting for the present study was a classroom
situation in which students were presented with both a
non-achievement task and an achievement task. Instructions
were designed to emphasize the importance of the achieve-
ment task but all students were prevented from immediately
starting the achievement task by the requirement that they
first spend some time on the non-achievement task.

The design of this study was that of:a correlational
approach to construct validity through investigation of the
influence of individual differences in strength of the
MotiVe-for Success (GSCI scores) on initiating, directing
and sustaining achievement-related activities within the

classroom situation.

92

93

However, a quasi-experimental approach was also incor-
porated through an investigation of the relative strengths
of the above relationships among subgroups of students
assumed to be performing achievement tasks of.varying levels
of difficulty (Stage I--Nomological Validity).

The correlational approach and the quasi-experimental.
approach also allowed investigation of the validity of
criteria for initiating, directing and sustaining activity
through relationships with indices of the level of
academic achievement (Stage II--Criterion Validity)..

Combining the correlational and quasi-experimental;
approaches also allowed investigation of the predictive
validity and construct validity of the GSCI through rela-
tionships with indices of the level of academic achievement

(Stage III--Predictive and Nomological Validity).

Sample.
The theoretical population of individuals from which~

the sample was selected is 9th grade Junior high school
students. Only those schools of one large metropolitan
school system which had previously administered the M-
Scales were considered for inclusion in the sample. The
first three schools to accept the invitation to participate

in this study were selected.

94

The total 9th grade pOpulation of the three schools
(855 subjects in 32 English classes) was used for this
study. Subjects were discarded before statistical analyses
on two bases: pilot study to develop experimental pro-
cedures (5 complete classes consisting of 118 subjects);
and missing data from GSCI, Grade Reports, DAT-V scores,
task measures and questionnaire items (227 subjects).

The final sample available for statistical analyses
consisted of 510 subjects (252 boys and 259 girls) from 28

English classes in the three schools.
Test Instruments

Differential Aptitude Test--
VerbaITScale

The Differential Aptitude Test--Verbal Scale (DAT-V)
is administered routinely, by schools participating in the
present study, to all students in the second semester of
the 8th grade. Raw scores for the DAT-V scale were obtained
from the cumulative records maintained by the school.

Generalized Situational
Choice.Inventory

 

The Generalized Situational Choice Inventory (GSCI)
is one of the four scales of the M-Scales developed by

Farquhar.l The GSCI is a forced choice instrument

 

le. W. Farquhar, Motivation Factors Related to
Academic Achievement, COOperative-Research Project 846
(East Lansing, Michigan: Michigan State University,
Office of Research and Publication, 1963).

 

95

consisting of 53 items on the Male Form and #6 items on
the Female Form. The GSCI was designed to elicit prefer-
ences between the polar dimensions of the achievement
motive when cast in the form of school activities and
situations.

The GSCI was administered to students in the 9th
grade during February of 1967 (approximately two months
before the present investigation) through Project MEMO.*
Scores for the GSCI were computed from data supplied by

Project MEMO.

Research Instruments
Three instruments-were developed for use in this

study.

Non-Achievement Task

A booklet (see Appendix_A) containing NU jokes
printed one to a page, was constructed. These jokes were
reproduced from "School Daze," a publication of Scholastic
Magazines, recommended for elementary school reading

ability level. A separate "Joke Rating Sheet" was

 

*Project.MEMO is a COOperative venture of community
colleges in the state of Michigan and Michigan State Univer-
sity to identify and assist high school students who may
have difficulty continuing their education. The project
is funded by the U. S. Office onyducation, under section-
“08 of the Higher Education Act of l965.

Appreciation is expressed to Gordon Sabine, Vice
President, Michigan State University, and to David Shultz,‘
Director Project MEMO.

96

constructed to allow the subjects to rate these jokes

"Like" or "Dislike."

Achievement Task

 

A packet of three duplicate pages (see Appendix B)
containing printed instructions and 7 blank lines was
provided for the use of each student in writing an assigned
essay. The instructions directed each student to write
an essay about a "make believe" student named Jim (a
separate form for girls differed only in the use of the
name Jane). The instructions stated that "Jim thinks it is
important to try to do his best in English Class," and each
student was directed to describe a believable past, present
and future for Jim by answering the following questions:

1. What could have happened in the past to make

Jim think it is important to try to do his
best in English class?

2. While in English class, what does he do?

3. What could happen to Jim in the future because

he thinks it is important to try to do his best
in English class?

The essay was to be written in ink, answering all three
questions within one paragraph of seven lines. Opportunity
to revise or correct the essay was available through the‘
use of more than one of the three duplicate pages. Instruc-
tions encouraged attention to clarity of eXpression,
grammar and neatness of work and informed the students that

their_essays would be graded by their teacher and the grade~

contribute to their final grade in English.

97

Achievement-Questionnaire

A six-item questionnaire (see Appendix C) was con-
structed for use.in this study to elicit responses concerned
with the level of past achievement in school, the perceived
difficulty of success in the present English class and
expectancies concerned with future achievement.

Only two of the six items were designed for analysis
in the present study. The remaining four items were
designed to assure that the student was immediately aware
of his previous level of achievement and that he was
attending to time as an objective factor in the performance
of the non-achievement task and achievement task.-

The first-of the two items analyzed asked each
student to estimate the "difficulty of doing well in this
English class." An 18 point scale was provided with descrip-‘
tive labels ranging from "Very Easy" at the left, to "Very
Difficult" at the right. Responses to this item provided
one basis for the formation of the three subgroups of
students thought to vary in the difficulty of success in
school. ‘

The second of the two items analyzed asked each
student to state "the grade you really expect to get on'
today's paragraph." Responses to this item were used for
'the "Expected Grade".criterion of the direction of achieve-

Inent activity (described below).

98

Instructions

The instructions necessary to administer the achieve-
ment task and the non-achievement task and to collect data
concerning the difficulty of success as well as latency,
expectancy and persistence indices, are presented in
AppendixI).

The instructions were designed to accomplish two
purposes. The first was, of course, to inform the students
of what was expected of them and how they might proceed in
the complex situation confronting them.

The second purpose for the instructions was to
emphasize the importance of the achievement task (graded
by teacher and grade contributing to final course achieve-
ment) and the unimportance of the non-achievement task‘

(representing only a predilection of the investigator).

Time Measurement Procedure

The following procedure was used to obtain behavioral
measurements of the time spent on the non-achievement task
(latency) and the time spent on the achievement task (per-
sistence).

An artists' spiral sketch-pad (9 x 12 inches)
supported by a metal music stand, was displayed in the
.front of the classroom., A large number (3% inches high)
<3ut from a calendar, was pasted on each page-of the sketch-

IDad.t Numbers ranged from 1 to 61 but their order of

99

appearance on consecutive pages was rearranged such that
no two_numbers appeared in their normal progression (this
was done to avoid the possibility that a student might
anticipate numbers or devise his own system for assigning
numbers).

The experimenter began turning pages of this sketch-
pad when the signal was given for the students to begin
the non—achievement task. The E continuously turned
these pages at the rate of one page every 20 seconds
throughout the 20 minutes available for the_achievement
task. The duration of each 20 second period was deter-
mined by the use of an ordinary wrist watch with a sweep
second hand.

The students were instructed to copy the number
displayed (on the sketch-pad) into the space provided on
the first line when they started to work on the first
theme—page. When they stopped writing on the first page,
they again copied the number appearing at that time into
the space provided on the last line of the first theme-
page. They were to continue in this manner if they used
additional pages to revise or correct their themes. The
students were cautioned that the numbers were in "code" and
while they were not supposed to "make sense" to them they
would not be given credit for the assignment unless the*
numbers were properly filled in (this threat was not carried

out).

lOO

Pilot Study Modifications,

An initial procedure had been devised using the same
jphysical arrangement described above but requiring each
student to read only one joke during each 20-second period.
This procedure had the advantage of controlling for the
effects of_individual differences in reading speed on the
number of jokes read (amount of reinforcement). However,
the confusion which resulted from the complexity of the
instructions used to accomplish this necessitated a_change
to the procedure described above with the consequent loss,
to analysis of the first 5 classes tested under the'

initial procedure.
Motivation Criteria

Latency (Initiating Behavior)

The difference between the number (decoded by order
of appearance rather than numerical value) appearing at the
start of the paragraph on the first page and zero, was
taken to represent the amount of time (number of 20-second
periods) spent on the jokes before starting the theme.-
Observation of the studentSYbehavior indicated they wrote
down a number as soon as they turned to the first theme—
page., They also began to write immediately, suggesting
the effects of the prior detailed explanation of the theme
task provided the student with sufficient clarity to be able

to proceed on the theme task as soon as he chose to do so.

101

iPersistence (Sustaining Behavior)

The difference between the numbers (decoded) appearing
at the start of the first paragraph and at the end of the
last paragraph was taken to represent the time spent working
on the Paragraph-Essay. In those cases involving the use
of more than one theme-page, there was a frequent discrep-
ancy between the number at the end of one paragraph and the
number at the beginning of the next. It was assumed that
the time represented by these discrepant numbers was spent:
in reformulating the next writing and was included in the

persistence score.

Expected Grades (Directing Behavior)
Grade level expectations were obtained from responses
to item #5 on the Achievement Questionnaire (see Appendix C).

Item #5 stated . . . show the grade you really expect to

 

get on today's paragraph. Responses to Item #5, in the form
of letter grades, were converted to a l2-point scale (A =

12, A- = 11.. . .E = 1) for statistical analysis.
Indices of Academic Achievement Level

Level of Achievement on the
Academic Task

 

The original copy of each student's "best paragraph"
(selected by the student) was extracted from the test
materials and returned to his teacher for grading. The

teacher was allowed maximum latitude in the assignment of

102

letter grades other than to caution that grading was to
:reflect the use of skills commonly develOped in English
courses with reference to grammer, neatness and clarity of
expression. It was assumed that such a letter grade would
reflect some combination of achievement and motivation-
variables as viewed by the teacher. Theme Grade--grades
assigned by each student's teacher were.converted to.a 12-
point scale (A = 12, A- = 11.. . . E = l) for analysis.

Level of Achievement in the
AcademiC‘Situatign

 

 

Two indices of the level of achievement in the aca-
demic situation were obtained consisting of teacher-,
assignedvcourse~grades.

A. Semester English Grade-econsisted of the grade_
assigned for English Class approximately one
month after the present investigation was
conducted. Letter grades were converted to,a
12-point scale (A,= 12, A- = 11 . . . E =~l) for
statistical analysis.

B. Cumulative GPAe-consisted of grades assigned for
courses requiring homework (English, Foreign,
Language,‘ Mathematics, Natural Science and
Social Studies) covering a period of two years
(in those cases in which grades-were not

available for the full.two-year period, grades

103

covering a lesser time period were used). Letter
grades were converted to a 12-point scale (A = 12,
A- = 11 . . . E = l) summed, and divided by the

number of courses involved.

Two Indices of "Apparent Task Difficulty"

Academic ability test scores (DAT-V) and Difficulty
Ratings (difficulty of "doing well in English Class"), rated
on an eighteen-point scale containing six categories ranging
from "Very Easy" to "Very Difficult" were both assumed to
index the "apparent difficulty" of school achievement.

The coefficient of correlation obtained for DAT-V scores
with Difficulty Ratings was found to be only r = -.139
(p < .05) for 252 males, and r = -.288 (p < .01) for.

259 females. While both of the above coefficients are
statistically significant, indicating that students of
higher ability tend to rate "doing well" as less difficult
than students of lower ability, the coefficients are both
low, indicating relatively little common variance.
The Formation of Groups Assumed to Vary in
Level of "Apparent Task Difficulty"

The distribution of DAT-V scores for males was arbi-
trarily tri-chotomized to form three equal sized groups
varying in level of academic ability (High Ability, N = 8M;
Middle Ability, N = 8“; Low Ability, N = 8N) and the distri-
bution of DAT-V scores for females was tri-chotomized to form

three nearly equal sized groups varying in level of perceived

104

difficulty (East, N = 87; Intermediate, N = 86, Difficult,
N = 86).

The separate distribution of Rated Difficulty for males
was also tri—chotomized to form three equal sized groups
varying in level of perceived difficulty (Low Difficulty, N==84;
Intermediate Difficulty, N = 8M; High Difficulty, N = 8N) and
the distribution of Rated Difficulty for females was tri-
chotomized 1x3 form three nearly equal sized groups varying in
level of perceived difficulty (Lovaifficulty, N = 87; Inter-
mediate Difficulty, N = 86; High Difficulty, N = 86).

The middle one-third of subjects in each of the four
distributions of scores was assumed to repreSent students most
likely to view success in school as of an "intermediate"

level of difficulty.

Reliability of the Generalized Situational
Choice Inventory (GSCI)

 

 

Estimates of test reliability were computed by Kuder-
Richardson formula #20 for the GSCI, Separately for the
total samples of males and females and separately within
each of the six subgroups. The results of these computations
are presented in Table 3.1 for both males and females.

It-can be seen in Table 3.1 that reliability estimates
based on the total separate samples of males and females are
equal, r = .82. Within subgroups of males and females varying
in level of academic ablity (DAT-V scores) reliability esti-
mates range from a high of r = .87 to a low of r = .76 for
males and from a high of r = .87 to a low of r = .75 for

females. Within subgroups of both males and females there

105

TABLE 3.1—~Kuder-Richardson estimates (Formula #20) of
reliability for male and female forms of the Generalized
Situational Choice Inventory (GSCI) for the total samples
of subjects and for each of the subgroups formed by
trichotomizing the distributions of DAT-V scores and
Difficulty Ratings of school success.

 

Reliability Estimates

 

 

 

 

Males Females
N r N r
Total Sample 233 .82 247 .82-
High Ability (DAT-V) 77 .87 85 .87
Middle Ability (DAT—V) 81 .80 81 .82-
Low Ability (DATeV) 75 .76 82 .75
Low Difficulty (Rated) 82 .80 83 °82-
Intermediate Diff. (Rated) 77 .8u 8A .82

High Difficulty (Rated) 74 .78 80 .81-

 

106

appears to be a consistent trend toward slightly lower
reliability with decreasing level of academic ability.

Within subgroups of males and females varying in the
level of perceived difficulty of success in English (Rated
Difficulty), reliability estimates range from a high of
r = .84 to a low of r = .78 for males and a high of r = .82
to a low of .81-for females. It appears that while the
subgroup reliability estimates remain fairly constant for
females they appear to be highest for the "Intermediate
Difficulty" subgroup of males, r = .8H, and decrease for
both the "Low Difficulty," r = .80, and "High Difficulty,"
r = .78, subgroups.

Reliability Estimates for Three Indices
for Academic Achievement

Information was gathered concerning the reliabilities
of the three indices of academic achievement used in this
study--Theme Grade, Semester English Grade and Cumulative
GPA. It was assumed that the reliabilities of these three
indices would vary within the total samples of males and
females and that each index might also vary between the
subgroups of students formed by trichotomizing the separate
distributions of DAT-V scores and Difficulty Ratings of
School Success. Reliability estimates are displayed in

Table 3.2.

107

Theme Grade

 

Evidence concerning the reliabilities of the Theme
Grade index of academic achievement was obtained by corre-
lating Theme Grade with the grade obtained at the completion
of the semester of English during which the theme (achieve-
ment task) was written (Current Semester English Grade).

It can be seen in Table 3.2 that reliability estimates
for the Theme Grade index of academic achievement were low
for males, ranging from a high of r = .U91 to a low of!

r = .362, and consistently lower than for females, ranging

from a high of r = .648 to a low of r = .501.

Semester English Grade

 

Reliability estimates for the Semester English Grade
index of academic achievement were obtained by correlating
the grade obtained for the current semester of English with_
the grade obtained for the previous semester of.English.

It can be seen in Table 3.2 that reliability estimates
for the Semester English Grade index of academic achievement
.378.
.68“ to a low of r = .298.

for males range from a high of r .659 to a low or r

and for females from a high of r

Cumulative GPA
Evidence concerning the reliability of the Cumulative-
GPA index of the level of academic achievement ( which

included grades for courses over a period of two years)

108

 

was. mma. 0mm. mme. Hmm. mom. mm :m Aeeeemv zefiseaeeae swam

 

 

 

 

amp. mm». mm:. mam. Hem. ass. mm em Aeceemv .cedm cemHUeELcecH
mm». ama. 2mm. mam. mom. mos. am am . neceemv sefiscaecea seq
mam. mma. mam. mac. saw. mam. ow em A>ue<ov meflaaea sea
was. .mme. ewe. ass. mmm. mes. em em A>ue<ov sedaflea cfieefiz
meet mad. Hme. amt. mam. see. am am. A>ne<ov seeadea swam
amp. was. . 3mm. 3mm. wee. Has. mmm mmm mefidemm fleece
moamsom mode: mmHMEmm mode: mmHmEmm mmamz mmamsmm wmamz
. mwmmm>< .mpmmu cmfiamcm opmam cmflamcm
woman new» topmoEmm msoa>omm popmmEom pampazo mpomwpsm mQSOtwnzm ocm
msoa>oam cpHB npflz ucoppso cpflz opwam mEmLB do amnasz mmadsmm Heuoe

meow unmeaso

 

mceeefiemm aefiafieeaacm

 

.mmooodm Hoonom
mo muHSOHMMfiQ woumm one mmhoom >| B¢Q mo wcofipsofippmflp cowhmaom esp wcHNHEOpOQOfiap an
weapon masoanSm on» no some Canvas Ucm mpchSpm mawsom ocm mama mo mmHQEmm deco» Canvas
mowmpm>m mpmhw mama w30fi>mpa cpfiz mommao>m macaw atom pcopaso one modem cmfiamcm topmoEmm
m50H>oaQ spas compo cmfiawcm ampmmEmm unmanso .mpwam Qmfiawcm topmoEmm pcopmso Sufi: oompw
msona wcfipmaopaoo an omcfiwuno ucoEo>mHnom oHEopmom mo Hm>oa mo moofipzw =<mo o>HpmHSESQ=
one secede emaawem teemcEcmz =.ceeto esteee one see mceesaeme aefiHHemHHcmuum.m mamae

109

was obtained by correlating grades obtained during 1965—
1966 with grades obtained during 1966-1967.

It can be seen ianable 3.2 that reliability esti-
mates for the Cumulative GPA index of the level of academic
achievement for males range from a high of r = .748 to a
low of r'= .639, and for females, from a high of r = .795

to a low of r = .699.

Analysis
The descriptive statistics which formed the basic

data for testing hypotheses derived from theory consisted
of coefficients of simple correlation and partial corre—
lation (DAT-V scores controlled) for GSCI scores with each
of the three theory-selected criteria of motivation (Stage
I-—Nomological Validity), for each of the three theory-
selected criteria of motivation with two indices of aca—
demic achievement (Stage II--Criterion Validity) and for
GSCI scores with two indices of academic achievement
(Stage III--Predictive and Nomological Validity).
Hypotheses concerning relationships for the total
group of males and females were tested in each of the three
Stages by means of one-tailed tests of the significance of-
the F statistic. Hypotheses concerning differential rela-
tionships among subgroups of subjects, formed on the basis
of varying levels of DAT-V scores or Rated Difficulty, were
tested by both one-tailed tests of the significance of the F

statistic and by one-tailed tests of the significance of the

110

difference between Fisher's "z" (r to z transformation)
statistic.

The use of Pearson coefficients of correlation, F
tests of significance from zero and Fisher's r to z trans-.
formation for comparison of significance between coeffi-
cients, all require assumptions of a normal bivariate dis-
tribution.

Attempts were made to determine the extent to which
the normal bivariate assumption might‘have been violated
among all of the variables involved in the present study
through: (a) inspection of plots of all bivariate distri-
butions; (b) analysis of variance technique for evaluating
the contribution of linear, quadratic and "other" variance
to the obtained regressions; and (c) the computation of a
non-parametric statistical test of association (phi-
coefficient, with scores above the median assigned "1" and
scores below the median assigned "0").

Inspection of the plots for all bivariate distribu-
tions indicated an occasional "atypical" grouping of scores
outside a smooth eliptical shape, but there did not appear
to be sufficient evidence to warrant the conclusion that
other than rectilinear relationships were usually involved.

The analysis of variance technique indicated a
preponderance of evidence to suggest that linear variance

accounted for the significant amount of relationship with

111

only an occasional contribution by either quadratic or
"other" variance.

The phi-coefficients consistently indicated only the
expected effects of their decreased power through uniformly
lower coefficients.

With the variety and number of variables involved in
the present study, an arbitrary decision was made to keep
the data in its raw score form rather than converting toua
normalized form and to proceed with the parametric tests of
association and significance even though in particular
instance, with some variables, the assumptions underlying
these tests may have been violated. The Opportunity to
hold academic ability constant by partial correlation was
an added advantage of the Pearson coefficient.

It was thought that the relatively large sample size
combined with the unknown effects of departure from the
assumption of normality tended to favor the use of the more
powerful parametric statistics.

Because of the eXploratory nature of this study,_the
.05 level of significance was selected for rejection of all
null hypotheses. The conclusions based on the analyses pre—
sented in Chapter IV assume graduated variables, restricted.
by the assumption of linearity with but slight, if any,

skewness.

112

Research Hypotheses

 

Hypotheses were formulated on the basis of theory for
each of three Stages of enquiry. Descriptions of the three

Stages and the research hypotheses are presented below.

Stage I--Nomological Validity

 

Hypotheses were formulated concerning the influence
of strength of the Motive for Success (GSCI scores) on the
initiating, directing and sustaining criteria of strength
of the tendency (motivation) to achieve in the academic
situation. Hypotheses were also formulated concerning the
above influence as differentially affected by the perceived
difficulty of success in school.

With differences in academic ability (DAT-V scores)
held constant:-

1. There is an inverse relationship between strength
of the Motive for Success (GSCI scores) and time
spent on a non-achievement task before starting
an achievement task,

2. There is a direct relationship between strength
of the Motive for Success (GSCI scores) and
expected level of academic achievement,

3. There is a direct relationship between strength
of the Motive for Success (GSCI scores) and

time spent on an achievement task,

113

4. The inverse relationship between strength of the
Motive for Success (GSCI scores) and time spent
on a non—achievement task before starting an
achievement task will be stronger for subjects
performing tasks of an "intermediate" level of.
difficulty than for subjects performing easier
or more difficult tasks,

5. The direct relationship between strength of the
Motive for Success (GSCI scores) and eXpected
level of academic achievement will be stronger
for subjects performing tasks of an "intermediate"
level of difficulty than for subjects performing
easier or more difficult tasks,

6. The direct relationship between strength of the
Motive for Success (GSCI scores) and time spent
on an achievement task will be stronger for
subjects performing tasks of an "intermediate"
level of difficulty than for subjects performing

easier or more difficult tasks.

Stage II--Criterion Validity

Hypotheses were formulated concerning the influence
of the initiating, directing and sustaining criteria of
strength of the tendency (motivation) to achieve in the
academic situation on the level of achievement in the
academic situation. Hypotheses were also formulated con-

cerning the above influences as differentially affected by

114

the perceived difficulty of success in school. With dif-

ferences in academic ability (DAT-V scores) held constant:

7.

10,

11.

There is an inverse relationship between time
spent on a non-achievement task before starting

an achievement task and level of academic achieve—
ment,

There is a direct relationship between expected
level of academic achievement and obtained level
of academic achievement,

There is-a direct relationship between time spent
on an achievement task and level of academic-
achievement,

The inverse relationship between time spent on.a
non-achievement task before starting an achieve-7
ment task and level of academic achievement will
be stronger for subjects performing_tasks of an
"intermediate" level of difficulty than for
subjects performing easier or more difficult
tasks,

The direct relationship between expected level

of academic achievement and obtained level of
academic achievement will be stronger for subjects:
performing tasks of an "intermediate" level of
difficulty than for subjects performing easier

or more difficult tasks,

115

12. The direct relationship between time spent on an
achievement task and level of academic achieve-
ment will be stronger for subjects performing
tasks of an "intermediate" level of difficulty
than for subjects performing easier or more
difficult tasks.

Stage III--Predictive and
Nomological Validity

 

 

Hypotheses were formulated concerning the influence of
strength of the Motive for Success (GSCI scores) on level,
of achievement in the academic situation. Hypotheses were
also formulated concerning the above influence as differ-
entially affected by the perceived difficulty of success in
school.

With differences in academic ability (DAT-V scores)
held constant:

13. There is a direct relationship between strength

of the Motive for Success (GSCI scores) and
level of achievement in the academic situation,

14. The direct relationship between strength of the

Motive for Success (GSCI scores) and level of
achievement in the academic situation will be
stronger for subjects performing tasks of an
"intermediate" level of difficulty than for
subjects performing easier or more difficult

tasks.

116

Summary

The sample for the present study consisted of five~
hundred ten students in 28 ninth-grade English classes
within three junior high schools of one large metropolitan
school system.

The Differential Aptitude Test-~Verbal Scale (DAT-V),
thought to reflect individual differences in academic
ability, and the Generalized Situational Choice Inventory
(GSCI), thought to reflect individual differences in
strength of the Motive for Success within the academic»
situation, had been previously administered to all_students
in the sample.

Three instruments were designed specifically for this
study. One instrument consisted of a Theme Writing Task
which allowed students to write and revise a seven-line
essay and which was thought to provide an achievement
activity typical for students in their English classes.

The second instrument consisted of a Joke Rating Task on
which students were required to spend some time, rating
jokes "Like" or "Dislike," before starting the achievement
task. The Joke Rating Task was thought to provide a non-
achievement activity holding little satisfaction for
achievement needs in the classroom situation.

The third specially constructed instrument consisted
of an Achievement Questionnaire which asked students to

estimate "the difficulty of doing well in this English

117

class," which was used to provide one basis (DAT-V scores
provided the second basis) for separating students into
three levels of "apparent difficulty" of school success.
The second Achievement Questionnaire item asked students,
to state "the grade you really expect to get on today's
paragraph" which was used to reflect the direction of
achievement activities, i.e., toward higher or lower
levels of future achievement.

Three indices of level of academic ahcievement were
also obtained--Theme Grade, which consisted of the teacher—
assigned grade for the achievement task, Semester English
Grade, which consisted of the final grade for the current
semester of English, and Cumulative GPA, which.consisted
of grades in all courses requiring homework during a period
of two years.

Criteria of the tendency (motivation) to achieve in
the classroom situation consisted of:‘ time spent on the
non-achievement task before starting the achievement task
(initiating), grades expected on the achievement task
(directing), and time spent on the achievement task (sus-
taining.

Subgroups of males and females, thought to vary in
level of "apparent difficulty" of success in school, were
formed by separately trichotomizing the distributions of

DAT-V scores and Rated Difficulty scores.

ll8

Reliability estimates (K-R Formula #20) for the male
and female forms of the GSCI~were found to be equal, r =
.82, but to range within subgroups from a high of r,= .87,
to a low of r = 76 for males and from a high of r = 87
to a low of r = .75 for females.

Estimates of reliability for the indices of academic
achievement were found to be generally higher for females,
r = .795 to r = .298, than for males, r = .748 to .362,
and generally higher for the Cumulative GPA, males r = .748
and females r = .784, then for the Theme Grade index, males
r = .491 and females r = .648.

Statistics consisted of coefficients of correlation
and partial correlation evaluated for significance from
zero by the F-test and evaluated for significant differences
between coefficients by use of Fisher's r to z transforma-
tion referred to a table of normal probability. The .05
level of significance was used for rejection of all null
hypotheses.

Research hypotheses were formulated on the basis.of'
theory within each of three Stages of investigation.

Hypotheses within Stage I—-Nomological Validity, con-
cerned relationships between strength of the Motive for
Success (GSCI scores) and criteria of the tendency (moti-
vation) to achieve in the classroom-~length of time spent
on the non-achievement task before starting the achievement

task (initiating), expected level of academic achievement

119

(directing) and length of time spent on the achievement
task (sustaining). The above relationships were also
hypothesized to be stronger for subgroups of students
performing school tasks of an "Intermediate" level of
difficulty than for students performing easier or more
difficult tasks.

Hypotheses within Stage II--Criterion Validity,
concerned relationships between the criteria of strength
of the tendency (motivation) to achieve in the classroom-—
initiating, directing and sustaining activities--and
indices of academic achievement. The above relationships
were hypothesized to be stronger for subgroups of students
performing school tasks of an "Intermediate" level of
difficulty than for students performing easier or more
difficult tasks.

Hypotheses within Stage III--Predictive and Nomo-
logical Validity, concerned relationships between GSCI
scores and indices of academic achievement. The above
relationships were hypothesized to be stronger for sub-
groups of students performing school tasks of an."Inter-
mediate" level of difficulty than for students performing

easier or more difficult tasks.

CHAPTER IV
ANALYSIS OF RESULTS

The analysis of results proceeded in three stages.
In Stage I, hypotheses were tested concerning relation-
ships of personality_test performance with three theory-
selected criteria of motivation in the classroom situation.
In Stage II, hypotheses were tested concerning relationshipS»
of three theory-selected criteria of motivation in the
classroom with level of academic achievement. In Stage
III, hypotheses were tested concerning the contribution of
personality test scores to ability-based predictions of

academic achievement.

Stage I--Nomological Validity

Analyses in Stage I were concerned with relation-
ships of the Motive for Success (GSCI scores) with three
theory-selected criteria of motivation in the academic
situation: Initiating Behavior (Joke Rating Time);
Directing Behavior (Expected Grades); and Sustaining
Behavior (Theme Writing Time). Further analyses evaluated
differential predictions for the above relationships when
measures of a situational characteristic (apparent task

difficulty) were allowed to influence the expression of

120

121

the Motive for Success in task performance. All analyses
were performed separately for male and female subjects.
The Strength and Direction of

Relationships Stated in
Hypotheses l, 2 and 3

 

 

Null Hypothesis 1: With Dat-V scores (V ) held constant,
there is no correlation for GSCI scores (V1) with
Joke Rating Time (V2).

Alternative Hypotheses 1: With Dat-V scores (V ) held
constant, there is a negative correlation for GSCI
scores (V1) with Joke Rating Time (V2).

H : -r > 0

1 12°6
Null Hypothesis 2: With DAT-V scores (V6) held constant,

there is no correlation for GSCI scores (V1) with
Expected Grade (V3).

Alternate Hypothesis 2: With DAT-V scores (V ) held
constant, there is a positive correlation for GSCI
scores (V1) with Expected Grade (V3).

H > 0

2‘ rl3-6

Null Hypothesis 3: With DAT-V scores (V ) held constant,
there is no correlation for GSCI sgores (V1) with
Theme Writing Time (V4)°

Ho‘ rlu°6 = O

122

Alternate Hypothesis 3: With DAT-V scores (V ) held
constant, there is a positive correlation for GSCI
scores (V1) with Theme Writing Time (V4)'

H3: rl4°6 > 0

It was stated in the null form of statistical
Hypotheses l, 2 and 3 that relationships of GSCI scores
with each of the three theory-selected criteria of motivae
tion would not be significant. The direction of the above
relationships was stated, in the alternate form, to be
negative in the case of GSCI scores with Joke Rating Time
(Hypothesis 1) and positive in the case of GSCI scores
with both eXpected Grade (Hypothesis 2) and Theme Writing
Time (Hypothesis 3).

The relationships stated in Hypotheses l, 2 and 3
were evaluated by tests of significance (F test) for
coefficients of partial (DAT-V scores controlled) corre-
lation computed separately for males and females.

Relationships of GSCI scores with Three Theory-
Selected Criteria of Motivation for Males.—-The results of
the computation of coefficients of partial (DAT—V scores
controlled) correlation and tests of significance (F test,
with 249 df for partial coefficients), using 252 junior
high school males, are displayed in Table 4.1.

It can be seen by inspection of Table 4.1 that the
partial coefficient of correlation for GSCI scores with Joke
Rating Time (Hypothesis 1) was not significantly greater

than zero.

123

Ho. v Q**

 

 

 

 

 

 

mo. v Q*

moo. awe. **smfi.n >ue<o Ammmuzv

moan:

*ONH. **>wa. *xzfim. **mHm. :mo.l moo.l Homw HmuOE

amassed madsam amassed cassam amassed cfidsam

memce aeaaaea

mEHB wsfipfipz osmce modem Umpomcxm mEHB wcflpmm mxow 0cm mpomwnsm
eofiee>aeoz cc easceaso sememsem c>aeoz

 

.mchE Hoonom nwfiz MOHQSn mmm pom coaom>fipos mo mfipmpfiao pmpomfiomuzmomnu
conga mo comm cpH3.mmaoom pmoe mpOpcm>cH oofioso HmcofipmSpHm UmNflHmamsmo
on» no Aomaaoapcoo mohoom >IB<QV soapmamaaoo assumed one mHQEHmIIH.: mqm<9

124

However, it-can also be seen in Table 4.1 that the
coefficient of partial correlation for GSCI scores with
Expected Grade (Hypothesis 2), partial r = .214 (p < .01),
and for GSCI scores with Theme Writing Time (Hypothesis
3), partial r = .170 (p < .05), were both low, but sig-
nificant in the expected direction.

Support for the nomological validity of the Motive
for Success, as indexed by GSCI scores, was therefore not
found (Null Hypothesis 1 not rejected) with Joke Rating
Time (initiating behavior), but support was found
(Alternate Hypotheses 2 and 3 accepted) with Expected
Grade (directing behavior) and with Theme Writing Time
(sustaining behavior), using males.

Further-inspection of Table 4.1 reveals that the
relationships of academic ability (DAT-V scoreS) with
Expected Grade (directing behavior) and with Theme Writing
Time (sustaining behavior) are both nonsignificant while
the relationship of academic ability (DAT-V scores) with
Joke Rating Time (initiating behavior), r = -.l94
(p < .01) is negative and significant.

Relationships of GSCI Scores with Three Theory:
Selected Criteria of Motivation for-Females.--The results

 

of the computation of coefficients of partial (DAT-V scores
controlled) correlation and tests of significance (F test,
with 256 df for partial coefficients) using 259 junior

high school females, are displayed in Table 4.2.

125

 

 

mo. v Q *
moo. **mma. mwo. >IB<Q Ammmuzv
mmflmsmm
H30. H30. *emem. **mmm. mmo. ago. Home fleece
Hsfieted maesam amassed cassam amassed caesam

 

cane messes: mates

 

 

memes seafiaea

mpmpw empomdxm use whomnpsm

mefie mcflpmm oxow

 

semecsem esaeoz
eoaee>fleoz co sentence

 

 

.mmHmEmm Hoozom swan LOHcsh mmm com coapm>flpoe do maamufiao

Umpomammlmaomnp coach mo comm Spas mLOpco>cH moaono HmCOHumSpHm UmNHHmamcmu on» mo
Apmfiaoapsoo mopoom >IB<QV coapmHmaaoo mo mpcmflofimdmoo Hmflosmd pcm oHQEflmnlm.: mqm<e

126

It can be seen in Table 4.2 that relationships of
GSCI scores with the three criteria of motivation are
significant gnly with the EXpected Grade (directing
behavior) criterion, partial r = .265 (p < .01), as stated
in Hypothesis 2.

Support for the nomological validity of the Motive
for Success, as indexed by GSCI scores, was therefore not
found (Null Hypotheses l and 3 not rejected) with Joke
Rating Time (initiating behavior) and Theme Writing Time
(sustaining behavior) but support was found (Alternate
Hypothesis 2 accepted) with Expected Grade (directing
behavior), using females.

Further inspection of Table 4.2 reveals that the
relationship of academic ability (DAT-V scores) with the
three criteria of motivation was significant only-with the
Expected Grade (directing behavior) criteria, r = .199
(p < .01), for females.

The Relative Strength and Direction

Of‘RelatiOnShips Stated in
Hypotheses 4A, 5A andySA

 

 

 

Null Hypothesis 4A: With DAT-V scores (V ) held constant,
the correlation of GSCI scores (V ) with Joke Rating
Time (V ) within a "Middle Abilit " ) subgroup
equals aero and is equal to corresponding corre—
lations within both "High Ability" (G1) and "Low
Ability" (G3 ) subgroups.

HO: P1206 G2 = O and = r12.6 G1 and r12.6 G3

127

Alternate Hypothesis 4A: With DAT-V scores (V ) held
constant, the correlation for GSCI scores (V ) with
Joke Rating Time (V ) within a-"Middle Ability" (G )
subgroup is negativg and greater than correspondin

correlations within both "High Ability" (G1) and
"Low Ability" (G3 ) subgroups.

H4A: -rl2o6 G2 > 0 and > -rl2°6 G1 and -rl2°6 G3

Null Hypothesis 5A: With DAT-V scores (V ) held constant,
the correlation of GSCI scores (V ) with Expected
Grade (v3) within a "Middle Ability" (G ) subgroup is

zero and equal to corresponding correla ions within.

both "High Ability" (G1) and "Low Ability" (G3 )
subgroups.

O: r13 6 G2 = 0 and = r1306 G1 and r1396 G3

Alternate Hypothesis 5A: With DAT-V scores (V ) held
constant, the correlation of GSCI scores (V ) with

Expected Grade (VT) within a "Middle Abilitl" (G )

y
subgroup is posit ve and greater than correspond ng

correlations within both "High Ability" (G1) and
"Low Ability" (G 3)subgroups.

HSA: r1306 G2 > 0 and > r13,6 G1 and r1306 G3

Null Hypothesis 6A: With DAT-V scores (V6) held constant,
the correlation of GSCI scores (V ) with Theme
Writing Time (Vu) within a "Middle Ability"(
subgroup is zero and equal to the corresponding)
correlations within both "High Ability" (G1) and
"Low Ability" (G3 ) subgroups.

O: rl4°6 G2 = O and = rl4°6 G1 and 1311406 G3

Alternate Hypothesis 6A: With DAT-V scores (V ) held;
constant, the correlation of GSCI scores (V ) with
Theme Writing Time (v ) within a "Middle Ability"
(G2) subgroup is posi ive and greater than corre—
sponding correlations within both "High Ability"
(G1) and "Low Ability" (G3) subgroups.

128

H6A:

r1406 G2 > 0 and > r1406 G1 and rl4°6 G3

It was stated in the null form of statistical
Hypotheses 4A, 5A and 6A that relationships, significantly
greater than zero, do not exist for GSCI scores with the
three criteria of motivation within the "Middle Ability"
subgroup of students gpd that the above relationships
will be equal to the corresponding relationships within
"Low Difficulty" and "High Difficulty" subgroups of students.

It was stated in the alternate form of statistical
Hypotheses 4A, 5A and 6A, that relationships significantly
greater than zero do exist for GSCI scores with the three
criteria of motivation within the "Middle Ability" subgroup
of students (negative in the case of Joke Rating Time and
positive in the case of Expected Grade and Theme Writing
Time) gpd that the above relationships will be signifi-
cantly greater than the corresponding relationships within

the "High Ability" and "Low Ability" subgroups of students.

 

Relationships of GSCI Scores with Three Theory-
Selegged Criteria of Motivation for Subgroups of Males

Varying in Level of DAT-V Scores.--Relationships stated in

 

Hypotheses 4A, 5A and 6A were evaluated by coefficients of
partial (DAT-V scores controlled) correlation computed
within each of three subgroups of males varying in level
of DAT—V scores, tested for significance (F test, with 81

df for partial coefficients), and tested for significant

129

differences between coefficients (Fisher's r to z trans-
formation) obtained within the "Middle Ability" subgroup
of males and (from) coefficients obtained within "High

Ability" and "Low Ability" subgroups (zl - 2 statistic

2
evaluated with df = 80 for partial coefficients).

The results of the above computations and signifi-
cance tests are displayed ianable 4.3.

It can be seen in Table 4.3 that, within the "Middle
Ability" subgroup of males the only significant coefficient
of partial correlation (DAT-V scores controlled) for GSCI
scores with the three theory-selected criteria of motiva-
tion occurs with the EXpected-Grade (Hypothesis 5A) criter-
ion, r = 0350 (p < .01).

While the significant partial coefficient for GSCI
scores with Expected Grade within the "Middle Ability"
subgroup is significantly greater than the corresponding
coefficient obtained within the "High Ability" subgroup
(z = 2.00, p < 005) it is apt significantly greater than
the corresponding coefficient obtained within the "Low
Ability" subgroup (z = .051, p > .05).

Therefore, Null Hypotheses 4A, 5A and 6A were not
rejected using males.

Further inspection of Table 4.3 shows that signif-
icant relationships for GSCI scores with Theme Writing

Time (sustaining behavior) were obtained only within

130

 

 

 

 

enhance 304 n w d mo. v as
He. V a ** enhance swam a w a mo. v an
omo.u mas. *iomm.n >ueaa Aamuzv
seaflaee
mflo.u mfio.u *efimm. **Hmm. Hao.u mae.- Home son
flee. mmo. **mmm.u >neaa Asmuzv
enhance
com. mom. n**omm. n**Hmm. and. woe. Home oases:
mac. mmo. mmo.u >ne<a Asmuzv
seflfiaba
**mom. **Ham. Hmo. mso. HOH.- woe.u Home roam
amended ofidsam Heathen cfidsam Heathen mfidsam
hence enhance
oEHB mcfipfiaz oEoEB opoaw pouooqu oEHB mcflpom oxoh Ugo mpoomnzm

 

eoaem>aeoz no mesmeano

 

nemeonem o>aeoz

 

 

.monE Hoonom zwfln LOflQSn mmm pom Amoaoom
>IB¢QV mpfififipm mo mHo>oH oopnu mo Sumo pm Compm>flpoe do ofiaoufipo popooaomlzaoonp

oops» do zomo spas moaoom pmoe hp0p2o>cH ooflono Hocofipmzpflm voNHHmaocoo och mo

AUOHHOLHPCOO mmhoom >IB<QV COHUQHmcfiHOO MO me®HOH%.HmOO HQHPVHMQ ECG ®HQEHWIIM.: Mdm¢8

131

the "High Ability" subgroup of males, simple r = .311
(p < .01) and partial r = .309 (p < .01).

An additional finding, displayed in Table 4.3, was
that DAT-V scores were significantly (negatively) related
to Joke Rating Time (initiating behavior) within the
"Middle Ability" subgroup, r = -.282 (p < .01), and within

the "Low Ability" subgroup, r = .208 (p < 001) of males.

Relationships of GSCI Scores with Three Theory-.
Selected Criteria of Motivation for Subgroups of Females

Varying in Level of DAT-V Scores.—-Relationships stated in

 

Hypotheses 4A, 5A and 6A were evaluated separately for
females by coefficients of partial (DAT-V scores controlled)
correlation computed within each of three subgroups of~
females varying in level of DATeV scores, tested for signif-
icance (F test, with 83 df for partial coefficients), and
tested for significant differences between coefficients
(Fisher's r to z transformation) obtained within the "Middle
Ability" subgroup of females and (from) coefficients
obtained within the "High Ability" and "Low Ability" sub-.
groups‘(zl - Z2 statistic evaluated with df~= 82 for partial
coefficients).

The results of the above computations and signifi-

cance tests-are displayed in Table 4.4.

132

 

 

 

 

 

Ho. v a** seafiaea son a m a .mo. v on
mo. v as oeafieee swam n m n .mo. V cm
oHH. ssfi.n moo. >ne<o Aoonzv
message
moo.u oao.n moo. moo. ooo.u moo.u Homo seq
:ao.n moH. amo. >ue<o Aoouzo
enhance
oHo. ooo.- cameo. cemam. Nod. oHH. Homo eaten:
amo. *tzmm. mmo.l >IB<Q Anwnzv
message
mad. Hod. esoas. erode. Foo. moo. Homo swam
Hearted ofidsam Heathen madehm Hearted ofiaefim
menoe enhance
oEHB manpoaz oEocB oumau popoooxm oEHB weapom oxow one mooonnsm

 

soaeo>fieoz to annotate

someonem o>aeoz

 

.moHoEom Hoonom swan pooc3m mmm pom Amoaoom
>IB<QV zpfiHHno mo mHo>oH oops» mo sumo um cooum>flpoe no mfipopfipo oopooaomlzaoonp
oonnu mo nomo spa: monoom umoa Apopco>CH oofioco HmCOfiomSpHm couoampococ on» no
Avoaaopucoo mopoom >IB<QV cofiumaonpoo mo mucofiofimwooo Hofipmmo cam oHQEamulz.z mqm<e

133

It can be seen in Table 4.4 that, within the "Middle
Ability" subgroup of females, the only significant coeffi-
cient of partial correlation for GSCI scores with the
three theory—selected criteria of motivation occurs with
the Expected Grade criterion, partial r = .263 (p < .01).

However, the above partial coefficient obtained within
the "Middle Ability" subgroup is numerically less than-the
corresponding coefficient obtained within the "High Ability"
subgroup, partial r = .419 (p < .01).

Therefore, Null Hypotheses 4A, 5A and 6A were not
rejected using females.

It can also be seen in Table 4.4 that significant
correlations were obtained, within the "High Ability" sub-
group of females, for GSCI scores with the Expected Grade
criterion of motivation, simple r = .418 (p < .01), and
partial r = .419 (p < .01), and for DAT-V scores with the
Expected Grade criterion, simple r = .334 (p < .01).

The Relative Strength and Direction

of Relationships Stated in
Hypotheses 4B, 5B and 6B

 

 

 

Null Hypothesis 4B: With DAT-V scores (V6) held constant,
the correlation of GSCI scores (V1) with Joke Rating
Time (V2) within an "Intermediate Difficulty" (G )
subgroup equals zero and is equal to correspondi g
correlations within both "Low Difficulty (G4) and
"High Difficulty" (G6) subgroups.-

HO: 131206015 = Dand=r12°6 GL} and r1206 G6

134

Alternate Hypothesis 4B: With DAT-V scores (V ) held
constant, the correlation of GSCI scores (V ) with
Joke Rating Time (V ) within an "Intermedia e
Difficulty" (G ) suggroup is negative and is greater
than corresponding correlations within both "Low
Difficulty" (G4) and "High Difficulty" (G6) subgroups.

H4B: -r12 6 G5 > 0 and > ”r1206 G4 and "r1206 G6

Null Hypothesis 5B: With DAT-V scores (V ) held constant,
the correlation of GSCI scores (V ) with Expected
Grade (v3) within an "Intermediate Difficulty" (G )
subgroup equals zero and is equal to corresponding
correlations within both "Low Difficulty" (G4) and
"High Difficulty" (G6) subgroups.

HO: r1306 G5 = 0 and = r13 6 G4 and r13 6 G6

Alternate Hypothesis 5B: With DAT—V scores (V ) held
constant, the correlation of GSCI scores (V1) with
Expected Grade (V ) within an "Intermediate Difficulty"
(G ) subgroup is ositive and is greater than corre-
sp nding correlations within both "Low Difficulty"

(G4) and "High Difficulty" (G6) subgroups.

HSB: r1306 G5 > 0 and > r13 6 G“ and r1306 G6

Null Hypothesis 6B: With DAT-V scores (V6) held constant,
the correlation of GSCI scores (V ) with Theme
Writing Time (V ) within an "Intermediate Difficulty"
(G ) subgroup is equal to zero and is equal to
co responding correlations within both "Low
Difficulty" (G4) and "High Difficulty" (G6) subgroups.

HO: r1406 G5 = O and = r1u06 G“ and rlu°6 G6

135

Alternate Hypothesis 6B: With DAT-V scores (V ) held
constant, the correlation of GSCI scores (V ) with-
Theme Writing Time (V4) within an "Intermed ate
Difficulty" (G ) subgroup is positive and is greater
than corresponding correlations within both "Low
Difficulty" (G4) and "High Difficulty" (G6) subgroups.

H6B: r14o6 G5 > 0 and > r1406 G4 and r14o6 G6

It was stated in the null form of statistical
Hypotheses 4B, 5B and 6B, that relationships significantly
greater than zero do not exist for GSCI scores with the
three criteria of motivation within the "Intermediate
Difficulty" subgroup of students apd that the above
relationships will be equal to the corresponding relation-
ships within "Low Difficulty" and "High Difficulty" sub-
groups of students.

It was stated in the alternate form of statistical
Hypotheses 4B, 5B and 6B that relationships significantly
greater than zero do exist for GSCI scores with the three
criteria of motivation within the "Intermediate Difficulty"
subgroup of students (negative in the case of Joke Rating
Time and positive in the case of Expected Grade and Theme
Writing Time) apd that the above relationships will be
significantly greater than the corresponding relationships
within the "Low Difficulty" and "High Difficulty" subgroups

of students.

136

Relationships of GSCI Scores with Three Theory-
Selected Criteria of Motivation for Subgroups of Males
Varying in Level of the Rated Difficulty of Success in
School.--Relationships stated in statistical Hypotheses 4B,
SB and 6B were evaluated, separately for males, by coeffi-
cients of partial correlation computed within each of three
subgroups of males varying in the level of Rated Difficulty
(difficulty of "Doing Well in English Class," rated on an
eighteen-point scale containing six categories ranging-from
"Very Easy" to "Very Difficult"). Coefficients were tested
for significance from zero (F test, with 81 df for partial
coefficients), and tested for significant differences
between coefficients obtained within the "Intermediate
Difficulty" subgroup and coefficients obtained within "Low
Difficulty" and "High Difficulty"_subgroups (zl - 22.
statistic evaluated with df = 80 for partial coefficients).

It can be seen in Table 4.5 that none of the coeffi-.
cients-of partial correlation for GSCI scores with any of
the three theory-selected criteria of motivation were
significantly different from zero within the "Intermediate
Difficulty" subgroup of males.

Therefore, Null Hypotheses 4B, 5B and 6B were not‘
rejected using males.

However, it can be seen in Table 4.5, that within the
group of males who rate school success as easiest ("Low

Difficulty" subgroup), significant coefficients were

137

somsoaomao swam n

a .mo. v n

 

 

Ho. v o** w

mo. v as seasoaccao son h m n .mo. v on
oHo. ooo.- eeoom.- >ue<o Aemuzo
seasoam
ems. ems. ooa. omm. smo. meo.u Homo loam swam
Aemuzo
moH. moo. mmH.n >ue<o mefisoaooao
opmHUoE
men. HHH. mma. mmm. ooo.- moH.- Homo ssoeeH
oso. Hao.u omo. >ue<o Aemuzo
meanest
scmmm. *Hmm. mos. mom. zoo. meo. Homo (one son

Hearted omoemm Hearted odoeHm amended omasem

 

 

mafia message esoee

opmhw popcooxm

 

mane oedema oxoo

 

soaemeaooz mo manoeaso

memos moaaaoe
paw

someonem oeaeoz meooooom

 

.moHoE Hoonom two: aOHQSm mmm pom pcoso>omnom Hoonom
mo meadommmmo cocoa on» mo mHo>oH oops» mo sumo pm coopm>fipoe mo mmpopmmo popcomom
lagoon» oops» mo homo sums monoom pmoB mAOpco>cH oomono Hosompmsumm noummmpocou on»
mo Apoaaohpcoo moaoom >IB<QV mucofiommgooo compomoanoo Hampton use omdsmmllm.= mqm<9

138

obtained for GSCI scores with Theme Writing Time, simple
r = .231 (p < .05) and partial r = .288 (p < .01). The
relationship of DAT-V scores with Theme Writing Time for
this "Low Difficulty" subgroup was not significant.

It can also be seen in Table 4.5 that a significant
(negative) relationship was found for DAT-V scores with
Joke Rating Time, r = -.309 (P < .01), only within the
subgroup of males who rate success in school as most

difficult ("High Difficulty" subgroup).

Relationships of GSCI Scores with Three Theory-
Selected Criteria of Motivation for Subgroups of Females
Varying in Level of the Rated Difficulty of Success in
School.--Research Hypotheses 4B, 5B and 6B were separately
evaluted using females, by partial (DAT-V scores controlled)
correlations computed within each of three subgroups of
females varying in the level of Rated Difficulty. Coeffi-
cients were tested for significance from zero (F test, with
83 df for partial coefficients) and tested for significant
differences between coefficients obtained within the "Inter-
mediate Difficulty" subgroup and (from) coefficients
obtained within "Low Difficulty" and "High Difficulty" sub-
groups (Fisher's r to z transformation with 21 - Z2 statis-
tic evaluated with df = 82 for partial coefficients). The

results of the computations, described above, are presented

in Table 4.6.

139

 

Ho. v o** mopsoppppo roam p w p .mo. v no
mo. v dc oopsopppfio son p m p .mo. v on
pao.- omo.u emo.n >upeo Aomuzv
someone
room. *pmm. *Hmm. *ppm. Hop.- mop.- Homo loam swam
Ammuzv
omo. poo. :mo. >np<o spade
seeped mom
ooo.u mop.- oom. oop. mop. oop. Homo upooEpooeH
omo. *omm. omp.u >up<o Apmuzo
sepsopp
Hmo. poo. *omm. *empm. mom. oop. Homo nope soc

 

poached opoepm

deposed opdeam poached oHoEHm

 

 

mane meaoppz esoep

whoop seappoa

ocopo cocoooxm oEHB wompmm oxoo Ugo

 

meoooosm
compo>mpoz mo oHpopHpo compoppm o>mpoz

 

.monEop Hoonom smog pOHCSn mmm pop pcoEo>omcom Hoonom

po meadomppmo poump on» po mHo>oH oopnp mo coco pm comoo>mpoe po oppopfipo oopooaom
Impoocp oops» po nooo cums mopoom umoe zpouco>sH oomoco HopomuoSpHm pouomopocou on»
mo Acoaaoppcoo mopoom >IB<QV mocomofippooo compoaoppoo Homupoo use oHQEHmIIm.: mqmge

140

It can be seen in Table 4.6 that none of the coeffi-
cients of partial correlation for GSCI scores with the
three theory-selected criteria of motivation were signifi-
cant within the "Intermediate Difficulty" subgroup of
females.

Therefore, Null Hypotheses 4B, 5B and 6B were not
rejected using females.

However, it can be seen in Table 4.6 that significant
relationships were found for GSCI scores with the Expected
Grade criterion within_the subgroups of females who rate
success as easiest ("Low Difficulty"), simple r = .278
(p < .01) and partial r = .226 (p < .01), 229 within the
subgroup who rate success as most difficult ("High‘
Difficulty"), simple r = .217 (p < .05) and partial
r = .221 (p < .05).

It-can further be seen in Table 4.6 that a signifi-
cant relationship for GSCI scores~with Theme Writing
Time was found only within the subgroup of females who
rate success in school as most difficult ("High Diffi-
culty"), simple r = .257 (p < .05) and partial r = .266

(p < .05).

141

Stage II-—Criterion Validity

 

Analyses in Stage II were concerned with evaluating
relationships of three theory-selected criteria of motiva-
tion: Initiating Behavior (Joke Rating Time); Directing
Behavior (Grade Expectations); and Sustaining Behavior
(Theme Writing Time) with level of academic achievement.
Further analyses evaluated differential predictions for
the above relationships when measures of a situational
characteristic (apparent task difficulty) were allowed to
influence the relationships of strength of motivation with
level of academic achievement. All analyses were performed
separately for male and female subjects.

The Strength and Direction of

Relationships Stated in
Hypotheses 7, 8 and 9

 

 

 

Null Hypothesis 7: With DAT-V scores (V ) held constant,
there is no correlation for Joke Rgting Time (V2) with
teacher-assigned grades (V5).

0: r25.6 = 0

Alternate Hypothesis 7: With DAT-V scores (V ) held
constant, there is negative correlation for Joke
Rating Time (V2) with teacher-assigned'grades,(VS).

H7: -r25 6 > O

Null Hypothesis 8: With DAT- V scores (V ) held constant,
there is no correlation for Expected Grade (V3 ) with
teacher-assigned grades (V5 ).

H : r

0 35.6 =

142

Alternate Hypothesis 8: With DAT-V scores (V ) held
constant, there is a positive correlatign for
Expected Grade (V3) with teacher-assigned grades,
(V ).

5

H8: r > 0

3506

Null Hypothesis 9: With DAT-V scores (V ) held constant,
there is no correlation for Theme Qriting Time (V4)
with teacher-assigned grades (V5).

HO: ru5°6 = 0

Alternate Hypothesis 9: With DAT-V scores (V ) held
constant, there is a positive correlati n for Theme
Writing Time (V4) with teacher-assigned grades (V5).

H9: ru5°6 > 0

It was stated in the null form of statistical
Hypotheses 7, 8 and 9 that relationships of the three theory-
selected criteria of motivation with indices of academic
achievement would not be significantly greater than zero.
The direction of the above relationships was stated, in the
alternate hypotheses, to be negative in the case of Joke
Rating Time (Hypothesis 7) and positive in the case of both
Expected Grade (Hypothesis 8) and Theme Writing Time
(Hypothesis 9).

Two indices of the level of academic achievement were
used to evaluate Hypotheses 7, 8 and 9. One index, "Theme
Grade," consisted of the grade assigned by each student's

teacher to the theme written for the present project. The

143

second index, "Semester English Grade," consisted of the
grade assigned by each student's teacher at the completion

of the current semester of English.

Relationships of Three Theory-Selected Criteria of
Motivation with Two Indices of Academic Achievement for
Maie§.--The relationships stated ianypotheses-7,_8 and 9
were evaluated, using males, by tests of significance (F
test, with 249 df for partial coefficients) for coeffi-
cients of partial (DAT-V scores controlled) correlation-
computed for the three theory-selected criteria of motiva-
tion with each of two indices of academic achievement.

The results of the computations and significance
tests are displayed in Table 4.7.

Coefficients of partial correlation computed for
each of the three theory—selected criteria of motivation
with the "Theme Grade" index of academic achievement pro-.
duced a significant partial correlation only for the
Expected Grade criterion, partial r.= .271 (p < .01).

Therefore, support was not found for the predicted,
relationship of Joke Rating Time with academic achievement
(nullvapothesis-7 not rejected) nor for the predicted
relationships of Theme Writing Time with academic achieve-l
ment (null Hypothesis 9 not rejected) but support was found
for the predicted relationship of Expected Grade with:

academic achievement (Alternate Hypothesis 8 accepted),

144

Ho.v o**

 

 

 

 

mo.v 9*
ssemm. *emem. >upeo
epop. *mmp. mpo. Heo. ospp
Ammmuzo
mesons: oEoee homo:
.m 0
*epom. **mom. eeppm. *eoam. oompo ooeoodxm p e e
HHH.I eeppp.u omo.u Hpo. ease meaemm oxoo
poached opoepm popepmi opoeHm coop meappoa
ppm
oompo nmmmwsm popmoEom oompo oEocB soapo>mpoz mooonnsm

 

mo oppoopno
upoEo>oH£o< omEocoo< po moomccH

 

.onoE Hoozom swan pomczn mmm pop
psoEo>oosoo omEocooo po moomosm 03» some coapo>mpoe mo mmpopmpo cocooaomnmpoonu oopsp
po Acoaaopucoo mopoom >IB<QV mocoHOprooo compmaoppoo Hmmppoo com oHQEHmIIo.: mqmee

145

using the "Theme Grade" index of academic achievement
for males.

With the influence of academic ability (DAT—V scores)
removed by partial correlation technique, the relation-
ships involving "Semester English Grade" with Expected
Grade, r = .301 (p < .01) and Theme Writing Time, r = .161.
(p < .05) both were significant while the relationship of-
"Semester English Grade" with Joke Rating Time, r = —.111
(p > .05), failed to show an acceptable level of signif-
icance.

Therefore, using "Semester English Grade" index of
academic achievement for males, null Hypothesis 7 was not
rejected with partial coefficients of correlation but
alternate Hypotheses 8 and 9 were accepted.

However, coefficients of simple correlation computed
for each of the three theory-selected criteria of motiva-
tion with the "Semester English Grade" index of academic
achievement produced significant findings in all three
cases: Joke Rating Time,-r = -.l7l (p < .01); Expected
Grade, r = .292 (p < .01) and Theme Writing Time,

r = .152 (p < .05).

146

Relationships of Three Theory—Selected Criteria of

 

Motivation with Two Indices of Academic Achievement for

 

Females.--The relationships stated in Hypotheses 7, 8 and

9 were evaluated, using females, by tests of signifi-

cance (F test, with 256 df for partial coefficients), for
coefficients of partial (DAT-V scores controlled) corre-
lation computed for the three theory-selected criteria

of motivation with each of two indices of academic achieve—
ment ("Theme Grade" and "Semester English Grade").

The results of the above computations and signifi—
cance tests are displayed in Table 4.8.

It can be seen in Table 4.8 that coefficients of
partial correlation computed for each of the three theory-
selected criteria of motivation with the "Theme Grade"
index of academic achievement produced significant results
only for Expected Grades, partial r = .230 (p < .01).

Thus, null Hypotheses 7 and 9 are not rejected while
only alternate Hypothesis 8 was accepted using the "Theme
Grade" index of academic achievement for females.

It can also be seen in Table-4.8 that coefficients of
partial correlation computed for each of the three theory-
selected criteria of motivation with the "Semester English
Grade" index of academic achievement produced significant
results only for Expected Grades, partial r = .354

(p < .01).

147

.1

 

 

 

 

Ho. v 0**
mo. v ax
*szmz. *smmz. >IB¢Q
mmo.u mmo.u mpo.- moo.- mane
weapon; metre
Aommuzo
*eemm. *eoom. eeomm. eeaom. ooono ooeooaxm emanate
Hmeoe
omo.- aop.- ooo.u oeo.u tape meaemm oxoo
poached opasam poached opossm coop essence
oco
oompo zmfiawpm popmosom ooopc oEonB compo>mpoz ooomosm

 

pcoEo>oH£o< omEonoo< mo mooHUcH

no mppoeapo

.mommsop Hoonom swan poocsn mmm pop pcoso>omzom
oHEoomoo no modaccfi 03p mo sumo cums coaum>mpos po ompopmpo cocoomonnzpoonp oops»
mo Avoamoppcoo mopoom >IB<QV mucoHOprooo compomoppoo momupoo cam oHosHmlum.z mqm<e

148

Thus, null Hypotheses 7 and 9 are not rejected while
only alternate Hypothesis 8»is accepted using the "Semester

English Grade" index of academic achievement for females.

The Relative Strength and Direction~
of Relationships Stated in
Hypotheses 10A, 11A and 12A '

Null Hypothesis 10A: With DAT—V scores (V ) held constant,
the correlation of Joke Rating Time (V 3) with
teacher-assigned grades (V ) within a "Middle
Ability" (G ) subgroup equals zero and is equal to
correspondigg correlations within both "High
Ability" (G1) and "Low Ability" (G3) subgroups.

O: r2506 G2 = O and = r2506 Gland r25 6 G3

Alternate Hypothesis 10A: With DAT-V scores (V ) held
constant, the correlation of Joke Rating Time (V )
with teacher- -assigned grades (V ) within a "Midd e
Ability" (G ) subgroup is negative and greater than
correspondigg correlations within both "High
Ability" (G1) and "Low Ability" (G3 ) subgroups.

HIOA: -r25°6 G2 > 0 and > -r25°6 G1 and -r25.6 G3

Null Hypothesis 11A: With DAT-V scores-(V ) held constant,
the correlation of Expected Grade (V ) with teacher-
assigned grades (V ) within a "Middlé.Ability" (G2)
subgroup equals zego and is equal to corresponding
correlations within both "High Ability" (Gl ) and
"Low Ability" (G3 ) subgroups.

O: r3506 G2 = O and = r35 6 G1 and r35 6 G3

149

Alternate Hypothesis 11A: With DAT-V scores (V ) held
constant, the correlation of EXpected Grade (V )
with teacher-assigned grades (V ) within a "Middle
Ability" (G ) subgroup is.posit§ve and greater than
correspondigg correlations within both "High.
Ability" (G1) and "Low Ability" (G3) subgroups.

H r35 6 G2 > 0 and > r35.6 G1 and r35.6 G3

11A:

Null Hypothesis 12A: With DAT-V scores (V ) held constant,
the correlation of Theme Writing Time (V ) with
teacher-assigned grades (V?) within a "Mi dle
Ability" (G2) subgroup equgls zero and is equal to
corresponding correlations within both "High Ability"
(G1) and "Low Ability" (G3) subgroups.

HO: r45o6 G2 = 0 and = r45-6 G1 and r45-6 G3

Alternate Hypothesis 12A: With DAT-V scores (V ) held
constant, the correlation of Theme Writing Time (Vu)v
with teacher—assigned grades (V') within a "Middle
Ability" (G ) subgroup is positive and greater than
corresponding correlations within both "High
Ability" (G1) and "Low Ability" (G3) subgroups..

H12A: ru5.6 G2 > O and > r45-6 G1 and ru5,6 G3

It was stated in the null form of statistical
HypothesesleA, 11A and 12A, that relationships signifi-
cantly greater than zero do not exist for the three
criteria of motivation with indices of academic achieve-
ment within the "Middle Ability" subgroup of students apd
that the above relationships are equal to the corresponding
relationships within "High Ability" and "Low Ability" sub-

groups of students.

150

It was stated in the alternate form of statistical
Hypotheses 10A, 11A and 12A, that relationships signifi-
cantly greater than zero do exist for the three criteria of
motivation with indices of academic achievement within the
"Middle Ability" subgroup of students (negative in the case
of Joke Rating Time and positive in the case of Expected
Grade and Theme Writing Time) 223 that the above relation-
ships will be significantly greater than the corresponding
relationships within the "High Ability" and "Low Ability"

subgroups of students.

Relationships of Three Theory-Selected Criteria of

 

Motivation with Two Indices of Academic Achievement for

 

Subgrogps of Males Varying in Level of DAT-V Scores.-e

 

Research Hypotheses 10A, 11A and 12A were evaluated by
coefficients of partial correlation computed within each

of three subgroups of males (formed by trichotomizing the
distribution of DAT-V scores), tested for significance from
zero (one-tailed F test, with 81 df for partial coeffi-
cients), and tested for significant.differences between
coefficients (Fisher's r to z transformation) obtained
within the "High Ability" and "Low Ability" subgroups

(z - .statistic evaluated with df - 80 for partial

1 z2
coefficients).
The results of the above computations and signifi-

cance tests are displayed in Table 4.9.

 

151

 

 

 

..l A 0
Ho. v d** suppose son a w p mo v no
mo. v as seepage swam p w p .mo. v on
mmp. mos. >upeo Aemuzv
mmo. ppo. ooo.- oop.- oepe oedema: oEoee appease
esoom. scopm. amp. mop. oompo ooooooxm son
moo.- omp.u oHp. mmo. ospp mepomm oxoo
Hop. sop. >ue<o Aemuzv
mmp. amp. opo.u ooo.- oefip mesons: esoep appease
esmmm. esmmm. mom. pom. oompo ooeoooxm opoofiz
oeo.u moo.- HHH.- HmH.n mane mepomm oxoo
omp. meo.- >npem Aemuzo
emom. eeoom. app. esp. oepe menses: esoee suppose
. . mo . . o op o coax
*emom *emmm me e *Hpe o o o e m swam
omH.u mop.n zoo. moo. oepp mepomm oxoo
poached opasam poached opdeHm coop penance
ppm
oompu cmfifiwcm popmoEom ooopo oEone somom>mpoz mpoowozm

 

wo mmpopfipo
psoEo>oHco< omEoooo< mo moomocH

 

.monE Hoonom swan

pOmCSn mmm pop Amopoom >IB¢QV anemone mo mHo>oH oopnp mo como Cpnpfis psoEo>omnom
anoomOm mo moomncfl 03» go sumo cums compo>mpoe mo ofipopfipo oopooaomlzpoonp oops»
ho Aooaaoppcoo mopoom >IB<QV mucoHOprooo compofioppoo Hmfippmd ppm omqsmmllm.: mqm<e

152

It can be seen in Table 4.9 that, within the "Middle
Ability" subgroup of males, the only significant correla-
tions occur for the Expected Grade criterion of motivation
with the Semester English Grade index of academic achieve-
ment, partial r = .299 (p < .01).

However, the above coefficient was not significantly
greater than-the corresponding coefficients obtained within
the "High Ability" subgroup or within the "Low Ability"
subgroup.

Therefore, null Hypotheses 10A, 11A and 12A were not
rejected for males.

It can also be seen in Table 4.9 that the Expected
Grade criterion correlated significantly with both the
Theme Grade and Semester English Grade indices of academic
achievement within the "High Ability" subgroup of males
but only with the Semester English Grade index within the
"Low Ability" subgroup of males.

Further inspection of Table 4.9 reveals that the only
remaining significant correlation occurred for Theme Writing
Time with Semester English Grade within the "High Ability"
subgroup of males, simple r - .266 (p < .01) and partial

r = .263 (p < .05).

153

Relationships of Three Theory-Selected Criteria of
Motivation with Two Indices of Academic Achievement for
Subgroups of Females Varying in Level of DAT-V Scores.--i
Research Hypotheses 10A, 11A and 12A were evaluated,
separately for females, by coefficients of correlation,
computed within each of three subgroups of females (formed
by trichotomizing the distribution of DAT—V scores),
tested for significance from zero (one-tailed F test, with-
83 df for partial coefficients) 289 tested for significant
differences between coefficients (Fisher's.r to z trans-
formation) obtained within-the "Middle.Ability" subgroup.
of females and (from) coefficients obtained within the
"High Ability" and "Low Ability" subgroups(zl - Z2
statistic evaluated with 82 df for partial coefficients).

The results of the above computations and signifi--
cance tests are displayed in Table 4.10.

It-can be seen in Table 4.10 that, within the.
"Middle Ability" subgroup of females, significant.relation-
ships between the three criteria of motivation and the two
indices of academic achievement occurred only-for the.
Expected Grade criterion with the Semester English Grade
index, partial r = .251 (p < .05).

However, the above coefficient was not significantly
greater than the corresponding coefficient obtained within

the-"High Ability" subgroup or "Low Ability" subgroup.

154

 

 

 

 

monopoe sod p u p .mo. v o
Ho. v nee v o
mo. v dz oopzaoe swam p w p .mo. V cm
ooz.u ozo. >upeo Aomuzo
pmo.- mzo. omo. mmo. ospp months: esoep anemone
*ozm. *mmm. *Nmm. *mmm. tempo ooeoodxm son
moo.- ooo.- Hmo.u mmo.- ospp oedema oxoo
**Nmm. *xmmz. >IE<Q Aowuzv
mmo.- smo.n mzo.- ozo.u ooze mezczpz oEoep spoofing
*zmm. *zomm. mop. pop. tempo ooeooaxm epoch:
mmo.- zpz.- omo.- Hzo.u maze wepeom oxoo
exemz. *xmma. >IB¢Q Ammuzv
moo. ozo. mmo.- zzo.- maze weapon: esoee enhance
*zozm. zempo. *zopm. *zmpm. tempo eoeoooxm swam
mmo.- omo.- zoo. mmo.- oepp meaoom ozoo
o o o o
zmpepmm oz can z and i oz saw some appease
new
oompm cmmawcm popmosom ooopw oEoQB compm>mpoz mpoomosm

 

pcoEo>oH£o¢ oHEocmo¢ mo moomocH

do mppoozpo

 

.moHMEop Hoonom swan
poHGSn mmm pop Amopoom >|B¢QV humafino po mao>oa oopcp po nooo Campos psoEo>omnom
anopMom mo moomosm 03» mo nooo com: compo>muoe po ompopopo cocooaomumpoonp oopcu
mo Aboaaoppcoo mopoom >IB<QV mpcofiomppooo compoaoppoo Hofippod com oaosmmlnoa.: mqm<e

155

Therefore, null Hypotheses 10A, 11A and 12A were not
rejected using females.

It can also be seen in Table 4.10, within both the
"High Ability" and "Low Ability" subgroups of females, the
Expected Grade criterion significantly correlated with
both the Theme Grade and Semester English Grade indices of
academic achievement.
The Relative Strength and Direction

of Relationships Stated ir1
Hypotheses 10B, 11B and 12B

 

 

 

Null Hypothesis 108: With DAT-V scores (V )-held constant,
the correlation of Joke Rating Time V ) with teacher-
assigned grades (v ) within an "Intermédiate Diffi-
culty" (G ) subgrodp equals zero and is equal to . .
corresponding correlations within both "Low Difficulty"
(G4) and "High Difficulty" (G6) subgroups.

HO: r25 6 G5 = 0 and = r25.6 G“ and r25 6 G6

Alternate Hypothesis 10B: With DAT-V scores (V ) held
constant, the correlation of Joke Rating Time (V )
with teacher-assigned grades (V ) within an "Inter-
mediate Difficulty" (G ) subgroép is negative and
greater than corresponding correlations within both
"Low Difficulty" (G4) and "High Difficulty" (G6)
subgroups. 7

HlOB: “r2566 G5 > O and > -r25.6 G“ and -r25.6 G6

Null Hypothesis 11B: With DAT-V scores (V6) held constant,‘
the correlation of Expected Grade (V ) with teacher—
assigned grades (V ) within an "Integmediate Diffi-
culty" (3.) subgrodp equals zero and is equal to
corresponding correlations within both "Low Difficulty"-
(G4) and "High Difficulty" (G6) subgroups..

HO: r35.6 65 = 0 and = r35 6 G4 and'r35.6 G6

156

Alternate Hypothesis 11B: With DAT-V scores (V ) held
constant, the correlation of EXpected Grade (V ) with
teacher-assigned grades (V ) within an "Intermédiate_
Difficulty" (G ) subgroup Ts positive and greater
than corresponding correlations within both "Low
Difficulty" (G4) and "High Difficulty" (G6) subgroups.

H > 0 and > r3506 G4 and r3506 G6

113‘ r35o6 G5

Null Hypothesis 12B: With DAT-V scores (V ) held constant,
the correlation of Theme Writing Timé (V ) with
teacher-assigned grades (V ) within anw" ntermediate
Difficulty" (G ) subgroup Equals zero and is equal to
corresponding dorrelations within.both "Low Difficulty"
(G4) and "High Difficulty" (G6) subgroups..

HO: r4506 G5 = O and = r45°6 G4 and'r45.6 G6

Alternate Hypothesis_12B: Wtih DAT—V scores (V ) held
constant, the correlation of Theme Writin Time
(V‘) with teacher-assigned grades (V ) within.an
"Intermediate Difficulty" (G ) subgrgup is positive
and greater than correspondigg correlations within
both "Low Difficulty" (Gu) and "High Difficulty"
(G6) subgroups..

H128: ru5°6 G5 > 0 an > ru506 Gu and r45.6 G6

It was stated in_the null form of statistical
Hypotheses 10B, 11B and 12B, that relationships significantly
greater than zero do not exist for the three-criteria of
motivation with indices of academic-achievement within the
"Intermediate Difficulty" subgroup of students~§pdthat the.
above relationships are equal to the corresponding relation—
ships within "Low Difficulty" and "High Difficulty" sub-

groups of students.

157

It was stated in the alternate form of statistical
Hypotheses 10B, 11B and 12B, that relationships signifi-
cantly greater than zero do exist for the three criteria of
motivation with indices of academic achievement within the
"Intermediate Difficulty" subgroup of students (negative
in the case of Joke Rating Time and positive in the case
of Expected Grade and Theme Writing Time) apd that the
above relationships will be significantly greater than the
corresponding relationships within the "Low Difficulty"

and "High Difficulty" subgroups of students.

Relationships of Three Theory-Selected Criteria of

 

Motivation with Two Indices of Academic Achievement for-
Subgroups of Males Varying in Level of Rated Difficulty

of Success in School.--Statistical Hypotheses 10B, 11B and

 

12B were evaluated by coefficients of correlation computed
within each of three subgroups of males (formed by trichot-
omizing the distribution of ratings of the difficulty of
success in English Class), tested for significance from

zero (one-tailed F test, with 81 df for partial-coefficients),
and tested for significant differences between coefficients
(Fisher's r to z transformation) obtained within the "Inter—
mediate Difficulty " subgroup of males and (from) coeffi-
cients-obtained within the "Low Difficulty" and "High
Difficulty" subgroups-(zl - Z2 statistic evaluated with 80

df for partial coefficients).

158

The results of the above computations and significance
tests are displayed in Table 4.11.

It can be seen in Table 4.11 that, within the "Inter-
mediate Difficulty" subgroup of males, coefficients of
partial correlation significantly greater than zero were
obtained only for the EXpected Grade criterion of-motiva-
tion with Theme Grade, partial r = .238 (p < .05), and with
Semester English Grade, partial r = .235 (p < .05).

However, neither of the significant partial corre-
lation coefficients obtained within the "Intermediate
Difficulty" subgroup were significantly greater than the
corresponding coefficients obtained within either the "Low
Difficulty" or "High Difficulty" subgroups of males.

Therefore, null hypotheses 10B, 11B and 12B were not
rejected using males.

It can also be seen in Table 4.11-that significant
positive relationships were obtained for the Expected Grade
criterion with Theme Grade within the "Low Difficulty" sub-
group of males, simple r = .334 (p < .01) and partial
r = .344 (p < .01). A significant negative relationship
was obtained for the Joke Rating Time criterion, with
Semester Enlgish Grade, only within the "High Difficulty"
subgroup of males, simple r = -.290 (p < .01) but this
latter relationship was not significant with the influence
of academic ability removed by partial correlation tech--

nique, partial r = -.157 (p > .05).

159

 

 

 

 

 

mezsoppppo empm p n p .zo. v a
no. v one v o
mo. V dc mezsozppao son a w p .mo. v on
*xamm. **wwm. >IB¢Q Azwuzv
pzo. mzo. zmo. zmo. maze weapons metre mezsoapppo
*zmm. mzz. mmo. mmo. oompo ooooooxm swam
mmo.- *zoom.- moo.- mmo.- oezp meaoom omoo
zmmm. omz. >npeo Azmuzo
zomm. mop. zmo. mzo. maze mopeapz esoep mozsozpppo
emmm. *mzm. *mmm. *mzm. oompo ooeooaxm one
mmo.- mmzm.- omo. poz.- maze meaeom ozoo nzooEpoeep
*zmm. mom. >upao Azmuzo
mzz. mmz. omo. mzo.- oepe moaned: mates mozsozppzo
ozm. mop. *zzzm. zezmm. tempo cocoooxm zoo
mmo. mmo. omo. mmo. ease meaomm omoo
poached ozoszm Hmpopmm ozoezm chop mepzpea
pom
oompu cmfiawcm popmoEom ooopo oEocB compm>mpoz mpoonnsm
pcoEo>oH£o< anoomo¢ mo moooocH mo ofipopmpo

 

.monE Hoozom swan p0H23n mmm
pop mmooosm Hoonom mo meadomppmo oopmp po mHo>oH oops» mo nooo Cocoa: pcoEo>oHcoo
oHEoomoo no moofiosfi 03» mo homo sums compo>aoos no ofipopmpo couoomomlmpoonp oopnp

mo Aooaaopusoo mopoom >IB<QV mucoHOprooo cofiuoaoppoo Homppoo cam oHQEHmIIHH.: mqm<a

160

Relationships of Three Theory-Selected Criteria of
Motivation with Two Indices of Academic Achievement for‘
Subgroups of Females Varying in Level of Rated Difficulty

of Success in School.--Research Hypotheses 108, 118 and

 

12B were evaluated, separately for females, by coefficients
of correlation computed within each of three subgroups of
females (formed by trichotomizing the distribution of
ratings of the difficulty of success in English Class),
tested for significance from zero (one-tailed F test, with
83 df for partial coefficients), and tested for signifi-
cant differences between coefficients (Fisher's r to.z
transformation) obtained with the "Intermediate Difficulty"
subgroup of females and (from) coefficients obtained within
"Low Difficulty" and "High Difficulty" subgroups (zl - Z2
statistic evaluated with 82 df for partial coefficients).

The results of the above computations and signifi-
cance tests are displayed in Table 4.12.

It can be seen in Table 4.12 that, within the "Inter-
mediate Difficulty" subgroup of females, significant coeffi-
cients were obtained only for the expected Grade criterion
with both Theme Grade, partial r.= .326 (p < .01), and with
Semester English Grade, partial r = .]80 (p < .01).

However, of the above significant coefficients,
obtained within the "Intermediate Difficulty" subgroup of
females, only the coefficient for the Expected Grade cri-

terion with the Theme Grade index of academic achievement

161

 

 

 

 

po. v ace mopsoppppo empm .p w p .mo. v no
mo. v o* haddoppppm 304 p m p .mo. v at
*mmm. *xmmm. >IB<Q Ammuzv
mmo. mmo. mmo. mmo. oepp mopeppz esoep mopsoppppo
*zmpm. zeoom. ppo. poo.- oompo ooeooaxm empm
mmo.- mop.u app. zzp.u tape mepoom oxoo
comm. *zmpm. >upao Ammuzo
pmo.u pzo.- mmo.- moo. tape mepcppz esoep mepooppppo
*zomm. *zzpm. ozzmmm. zemom. oompo ooeoooxm one
mpp.- mop.u mmo. mzo. ospe mepomm opoo upooEpooep
*xmmz. *somm. >IB<Q Ammuzv
mmo.- mmo. poo.u moo. ospp mepoppz mamas mepooppppo
mzp. *mzm. mmo. app. tempo ooeooaxm sop
poo. mmo.: mmo. zpo. oepp mopemm oxoo
popcnmm opQEpm popepmm opeEpm coop mopppoe
Ugo
ooMpm nmppwcm popmoEom ooopw oEoze compm>mpoz muoonnsm

 

psoso>opno< opEoooo< no moopocH

po mpnoeppo

 

.moposop poonom amps pOHQSn mmm
pom mmooosm poonom mo muHSOprpc popmp mo mpo>op oopnu po homo spnppz ucoso>opnom
opEooMom po moopoCH 03p mo sumo cup: compo>ppos po oppopppo vopoopomsmpoonp oopzp

mo Aooppoppcoo mopoom >Ie<ov monopopppooo COHompoppoo popppmo use opdspmlump.z mqm<e

162

was significantly greater than the corresponding coeffi-
cient obtained within the "High Difficulty" subgroup, but-
it was not significantly greater than the corresponding
coefficient obtained within the "Low Difficulty" subgroup
of females.

Therefore, null Hypotheses lOB, 11B and 12B were not
rejected using females.

Stage III-—Predictive and
Nomological Validity

 

 

Analyses in Stage III were concerned with relation-
ships of the Motive for Success (GSCI scores) with academic
achievement. Further analyses evaluated differential
predictions concerning the relationship of the Motive for
Success (GSCI scores) with academic achievement when
measures of academic ability (DAT-V scores) and of Rated»
Difficulty of success in English were used to form groups
of students thought to vary in the "apparent task diffi-
culty" of success in the school situation.

The Strength and Direction of the

Relationship Stated in
Hypothesis 13

 

 

 

Null Hypothesis 13: With DAT— V scores (V ) held constant,
there is no correlation for GSCI scores (V1) with
teacher- -assigned grades (V5 ).

H : r = 0

1506

163

Alternate Hypothesis 13: With DAT-V scores (V ) held
constant, there is a positive correlation for GSCI
scores (V1) with teacher-assigned grades (V5).

H13: r15°6 > 0

It was stated in the null form of statistical Hypoth-
esis 13 that the:relationship of GSCI scores with indices.
of academic achievement would not be significantly greater
than zero.

It was stated in the alternate form of statistical
Hypothesis 13 that the relationship of GSCI scores with
indices of academic achievement would be significantly
greater than zero.

Two indices of the level of academic achievement were
used to evaluate Hypothesis 13. One index,_"Semester‘
English Grade," consisted of the grade assigned by each
student's teacher at the completion of the present semester
"of English. The second index, "Cumulative GPA," consisted
of the sum of grades obtained in courses requiring home-8
work over a period of two years, divided by the number of.

such courses.

The Relationship of GSCI Scores with Two Indices of

Academic Achievement for Males.--The relationship stated in

 

Hypothesis 13 was evaluated, using males, by tests of sig-

nificance (one-tailed F test, with 249 df for partial-

164

coefficients) for correlations computed for GSCI scores
with "Semester English Grade" and "Cumulative GPA."

The results of the computations and significance
tests are displayed in Table 4.13.

It can be seen in Table 4.13 that the coefficient
of partial correlation for GSCI scores with the Semester
English Grade index of academic achievement is signifi-.
cant in the expected direction with the influence of DAT-V
scores controlled by partial correlation technique,

r = .299 (p < .01).

It can also be seen in Table 4.13-that-the coeffi-
cient of partial correlation for GSCI scores with the
Cumulative GPA index of academic achievement is-signifi-v
cant in the expected direction with the influence of DAT-V
scores controlled by partial correlation technique, r =
.301 (p < .01).

Therefore, alternate Hypothesis 13 is accepted using

both indices of academic achievement for males.

The Relationship of GSCI Scores with Two Indices of

Academic Achievement for Females.--The relationship stated~

 

in Hypothesis 13 was evaluated, separately for females, by
tests of significance (one-tailed F test,_with 256 df for
partial coefficients) for correlations computed for GSCI~

scores with "Semester English Grade" and "Cumulative GPA."

165

po. v see

 

 

 

 

mo. v Q*
**o:m. *Nmm. >IE<Q Ammmuzv
mopoz
*xmom. **Hmm. **mmm. *xfimm. Homw .HdeB
popcpmm opoepm popepmm opospm
, momma mopppoe
<mo o>ppopssso ooopo coppwcm popmoEom ppm mpoohnsm

 

pcoEo>opco< omEocoo< mo moOpocH

someopem oepeoz

.mopoE poonow amps pOHCSn mmm
pop osoao>opnom omaoomom p0 moop©Cp 03p mo sumo pups opoom ponpo>llpmoe oesoppa¢

poppcopopppo opp Ugo AHomwv mpOpco>cH oopono pocomuoSppm oomppopocou one mo
ApoppoquOO mopoom >IB<QV mucopompmooo compmpoppoo popppmd use opospmlump.z mpm<e

166

The results of the computations and significance tests
are displayed in Table 4.14.

It can be seen in Table 4.14 that the coefficient of
partial correlation for GSCI scores with the Semester English
Grade index of academic achievement is significant in the
expected direction with the influence of DAT-V scores
controlled by partial correlation technique, r = .285
(p < 001).

It can also be seen in Table 4.14 that the coefficient
of correlation for GSCI scores with the Cumulative GPA
index of academic achievement is significant in the expected
direction with the influence of DAT—V scores controlled by
partial correlation technique, r = .289 (p < .01).

Therefore, alternate Hypothesis 13 is accepted using
both indices of academic achievement for females. V
The Relative Strength and Direction

of Relationships Stated in
Hypothesis 14A

 

 

 

Null Hypothesis 14A: With DAT-V scores (V ) held constant,
the correlation of GSCI scores (V ) with teacher-
assigned grades (V ) within a "Middle Ability" (G2)
subgroup equals ze?o and is equal to corresponding
correlations within both "High Ability" (Gl ) and
"Low Ability" (G3 ) subgroups.

= 0 and =

O: r1506 G2 r1506 G1 and r1506 G3

167

po. v dez

 

 

 

 

mo. v Q...
*zmmz. *zzmz. >up<o Ammmuzv
momeom
**mmm. **Hpm. **mmm. **opm. Homo Houoe
pmpepmm opoEpm popepsm opaEpm
memos mopppom
<mo o>ppmpseso opopu cmppwsm popmoEom use

 

neocosopeoe opEoomoa mo woopoep

someopom o>peoz neoomosm

.mopoaop poocom mmHn

pOpQSn mmm pop pcoso>opnom opEoomoo po moopocp 03p po sumo pups opmom pmnpo>nupmo9
oosepoaa popeeopomppo one one ppomov mnoeeoeep oopoeo pmeopomsepm oonppmpoeoo one
MO AUQHHOLpCOO mmkoom >IB<QV mpCTHOHMeroo COHpmflmcfiHOO HMHPLGQ USN ®HQEHWIIJH.: mnHmsz

168

Alternate Hypothesis 14A: With DAT-V scores (V ) held
constant, the correlation of GSCI scores 7V ) with
teacher-assigned grades (V ) within a "Midd e
Ability" (G2) subgroup is Bositive and greater than
corresponding correlations within both "High
Ability" (G1) and "Low Ability" (G3) subgroups.

H

14A: r1506 G2 > 0 and > r1506 G1 and r1506 G3

It was stated in the null form of statistical
Hypothesis 14A that a relationship, significantly greater
than zero, does not exist for GSCI scores with academic
achievement within the "Middle Ability" subgroup of
students apd that the above relationship is equal to the
corresponding relationships within "High Ability" and
"Low Ability" subgroups of students.

It was stated in the alternate form of statistical
Hypothesis 14A that a relationship, significantly greater
than zero, does exist for GSCI scores with academic achieve-
ment within the "Middle Ability" subgroup of students apd
that the above relationship is equal to the corresponding
relationships within "High Ability" and "Low Ability"

subgroups of students.

169

Relationships of GSCI Scores with Two Indices of

Academic Achievement for Subgroups of Males Varying in Level

 

of DAT-V Scores.--The relationships-stated in Hypothesis 14A

 

were evaluated, for males, by coefficients of correlation
computed within each of three subgroups of males (formed by
trichotomizing the distribution of DAT-V scores), tested
for significance from zero (one-tailed F test, with 81 df
for partial coefficients), apd tested for-significant
differences between coefficients (Fisher's r to z trans—
formation) obtained within the "Middle Ability" subgroup of
males and (from) coefficients obtained within the "High

Ability" and "Low Ability" subgroups (zl - z statistic

2
evaluated with 80 df for partial coefficients).

The results of the above computations and signifi-
cance tests are displayed in Table 4.15.

It~can be seen in Table 4.15 that, within the "Middle
Ability" subgroup of males, the partial correlation of GSCI
scores with the two indices of academic achievement are not
significantly greater than zero.

Therefore, null Hypothesis 14A was not rejected using~
either index of academic achievement for males.

It can also be seen in Table 4.15 that coefficients
of correlation, significantly different from.aero, were
found for GSCI scores with both indices of academic achieve-.

ment within both the "High Ability" and "Low Ability"

subgroups of males.

170

 

 

 

 

m p 30 p u p . . a
po. v o** opp.o< p v mo v o
mo. v o* mopppoe empm p m p .mo. V cm
zop. mmp. >up<o pzmuzo
mopppoa
*mzm. momm. *mzm. *omm. Homo 30A
zoom. pmp. >np<o pzmuzo
mopppoa
omp. mmp. mmp. mmp. pomo opoopz
mmp. mmp. >up<o pzmuzv
mopppoa
*zpmz. *zmzz. *zmmz. *zmmz. pomo empm
popepmm opaepm poponmm opQEpm
whoop mopppoa
«m0 o>moopseso ooopw coppwcm popmoEom one mpoomnsm

 

ucoEo>op£o< opEocoo< mo moopch

someopem oepeoz

 

.mopoE poocom swan pompzn 2mm pop Amopoom >IB<QV mppppno po mpo>op oopzu

po sumo spcppz ucoso>opcoo omEopoom po mooppsp 03» mo homo cups opoom poopo>ulpmoe
oesopooe popeeopopppo one one ppomov mpoeeoeep oopoeo pmeopemsepm oomppopoeoo one
po Aooppoppcoo mopoom >IB¢QV mpcopopppooo compopoppoo pomppmo use opQEmmnlmp.z mpm¢e

171

Relationships of GSCI Scores with Two Indices of
Academic Achievement for Subgroups of Females Varying in
Level of DAT-V Scores.--The relationships stated in
Hypothesis_l4A were evaluated, separately for females, by
coefficients of correlation computed within each of three
subgroups of females (formed by trichotomizing the distri-
bution of DAT-V scores), tested-for significance from
zero (one—tailed F test, with 83 df for partial coeffi-
cients), gpd tested for significant differences between
coefficients (Fisher's r to z transformation) obtained
within the "Middle Ability" subgroup of females and (from)
coefficients obtained within the "High Ability" and "Low

Ability" subgroups (zl - 2 statistic evaluated with 82 df

2
for partial coefficients).
The results of the above computations and signifi-
cance tests are displayed in Table 4.16.
It_can be seen in Table 4.16 that, within the
"Middle Ability" subgroup of females, partial correla-
tions for GSCI scores with Semester English Grade, partial
r = .383 (p < .01), and partial correlations for GSCI
scores with Cumulative-GPA,-partial r = .445 (p < .01),
are significantly greater than zero.
Furthermore, the coefficients of partial correlation,

obtained within the "Middle Ability" subgroup of females,

are significantly greater than the corresponding.

172

 

 

 

 

% H a . Q
po. epppoe zop p v a mo v o
mo. mopppoa empm p w p .mo. v on
mmp.- mmp.- >ne<o Ammuzo
mopppoa
zmo. mmo. opp. mmp. pomo sop
**mmm. *xwmm. >IB¢Q Amwuzv
mopppoe
D**m::. **mm:. Q**mmm. ##3H2. Homm mHUUHE
*zmom. *zpmz. >Ip<o ppmuzo
mopppoe
*pmm. *zzmm. *emmm. *zmmm. Homo empm
pmpcpmm opospm popepmm opaepm
momma mopppoa
<mo o>mptpsezo oompm nmppmcm popmoEom ppm mpoompsm

 

pcoEo>op£o< opEoUoo< po moOpch

someopem o>peo2

 

.wopoaop poonom 3mm: pOpCSn mmm pop Amopoom >IB<QV mumppno mo mpo>op oops»
p0 sumo spzups pcoao>op£om opaocooo po moopocH 03» mo cooo naps opoom pmnpo>nlpmoe
opSppr< poppcopopppo on» new AHomuv mpOpco>cH oopozo poCOHpoSupm oouppmpocou on»

mo Aooppoppcoo mopoom >IB<QV mucomopppooo COHpmpoppoo pomppoo ppm opQEpmllop.z mpm<e

173

coefficients obtained within the "Low Ability" subgroup
(GSCI with Semester English Grade: 2 = 1.86, p < .05 apd
GSCI with Cumulative GPS; 2 = 2.70, p < .01)

However, the partial coefficients for GSCI scores
with both indices of academic achievement, obtained within
the "Middle Ability" subgroup, are ppp significantly
greater than the corresponding coefficients obtained within
the "High Ability" subgroup of females.

Therefore, null Hypothesis 14A was not rejected
using either index of academic achievement for females.
The Strength and Direction of

Relationships Stated in
Hypothesis l4B

 

 

 

Null Hypothesis 14B: With DAT—V (V ) held constant, the
correlation of GSCI scores (V ) with teacher-
assigned grades) (v ) within ah "Intermediate
Difficulty" sdbgroup equals zero and is equal
to correspondigg correlations within both "Low
Difficulty" (G4) and "High Difficulty" (G6)
subgroups.

H : r G = O and = r1506 G“ and r1506 G6

1506 5

Alternate Hypothesis 14B: With DAT-V scores (V ) held
constant, the correlation of GSCI scores V) with'
teacher— —assigned grades (V ) within an "Intérmediate
Difficulty" ) subgroup 7s positive and greater
than corresponding correlations within both "Low
Difficulty" (Gu)and "High Difficulty" (G6) subgroups.

HluB: r1506 G5 > O and > r1506 G4 and r1506 G6

174

It was stated in the null form of statistical
Hypothesis 14B that a relationship, significantly greater
than zero, does not exist for GSCI socres with academic
achievement within the "Intermediate Difficulty" subgroup
of students-Egg that the above relationship is equal to
the corresponding relationships within "Low Difficulty"
and "High Difficulty" subgroups of students.

It was stated in the alternate form of statistical
Hypothesis 14B, that a relationship, significantly greater
than zero, does exist for GSCI scores with academic achieve-
ment within the "Intermediate Difficulty" subgroup of
students apd that the above relationship is significantly
greater than the corresponding relationships within "Low

Difficulty" and "High Difficulty"-subgroups of students.

Relationships of GSCI Scores with Two Indices of
Academic Achievement for Subgroups of Males Varying in
Level of Rated Difficulty of Success in School.--The
relationships stated in Hypothesis 14B were evaluated, for
males, by coefficients of correlation computed within each
of three subgroups of males (formed by trichotomizing
the distribution of ratings of difficulty of success in
English Class), tested for significance from zero (one-
tailed F test, with 81 df for partial coefficients), Egg
for significant differences between coefficients (Fisher's

r to z transformation) obtained within the "Intermediate

175

Difficulty" subgroup of males and (from) coefficients
obtained within the "Low Difficulty" and "High Difficulty"
subgroups (zl - Z2 evaluated with 80 df for partial
coefficients).

The results of the above computations and signifi-.
cance tests are displayed in Table 4.17.

It can be seen in Table 4.17 that, within the
"Intermediate Difficulty" subgroup of males, coefficients-
for GSCI scores with Semester English Grade, partial
r = .417 (p < .01), and for GSCI scores with Cumulative
GPA, partial r = .362 (p < .01), are both significant in
the expected direction.

It can also be seen in Table 4.17 that the coeffi—
cient of partial correlation for GSCI scores with Semester
English Grade, within the "Intermediate Difficulty" sub-
group of males, is significantly greater than the corre-
sponding coefficients obtained within the "Low Difficulty"
subgroup (z = 2.19, p < .05) and within the "High
Difficulty" subgroup (z = 1.36, p < .05).

However, the partial correlation for GSCI scores
with Cumulative GPA, obtained within the "Intermediate
Difficulty" subgroup of males, was not significantly
greater than the corresponding coefficients obtained

within the "Low Difficulty" or "High Difficulty" subgroups.

176

 

 

 

 

I n .

po. v oer mopsoppopo empm p w p mo v no

mo. v dz mopsoppmpo sop p m p .mo. V cm
*zmmz. *zpmm. >np<o pzmuzo
mopsoppppo
zmp. *zzmm. pmp. *zmm. pomo empm
*pmm. zmmm. >up<o pzmnzo
mopsopmppo
**mmm. *zmoz. omzzmpz. *eomz. pomo oompoospooep
zom. *pmm. >upao pzmuzv
mepooppppo
mmp. mmp. oop. mop. pomo mop

popenom opoepm pmponmm opdepm
whoop mopppoa

ado o>ppmpsaso oompo cmppwsm popmoEom one mpoomnzm

 

pcoEo>op30< opEoooo< po moopocp

someopem o>pooz

 

.mome poonom

amps pOpCSn 3mm pop mmooodm poonom mo mpp30ppppc oopmp one no mpo>op oopnp po homo
apnops pcoao>opnom opEoomoo po moopocm 03p po zomo cups A>IB<QV opmom ponpo>nnpmoe
oodpppd< poppcopopppo on» cam Apommv mpOpco>cH oopono poCOHpmSppm Uomppopocou on»

p0 Aooppoppsoo mopoom >|B<QV mucopopppooo COHpopoppoo popppmd ppm opospmllmp.: mam<e

177

Therefore, alternate Hypothesis 14B was accepted using"
the Semester English Grade index of academic achievement for
males, but Hypothesis 14B Was not rejected using the Cumu-

lative GPA index of academic achievement for males.

Relationships of GSCI Scores.with Two Indices of_
AcademicAchievement fpr Supgroups of Females Varying in
Level of Rated Difficulty of Success in School.--The
relationships stated in Hypothesis 148 were evaluated,
separately for females, by coefficients of correlation;
computed within each of three subgroups of females (formed
by trichotomizing the distribution of ratings of.difficu1ty
of success in English Class), tested for significance from
zero (one-tailed F test, with 83 df for partial coeffi-
cients), apd tested for significant differences between
coefficients (Fisher's r to z transformation) obtained
within the "Intermediate Difficulty" subgroup of females
and (from) coefficients obtained within the "Low Difficulty"
and "Higthifficulty" subgroups (zl - 22 evaluated with
82 df for partial coefficients).

The results of the above computations and signifi-
cance tests are displayed in Table 4.18.

It can be seen in Table 4.18 that,-within the "Inter-
mediate Difficulty" subgroup of females, neither of the
coefficients of partial correlation for GSCI scores with‘

either index of academic achievement were significantly

greater than zero.

mopsoppppo empm n p .mo. v a

 

178

 

 

 

po. v aze m o
mo. v on mopsoppppo sop p w p .mo. v no
*emmm. *mmm. >np<o Ammuzv
mopsoppppo
*zpmm. *zmmm. *zmmm. *zmpm. pomo empm
zepmm. *omm. >up<o Ammuzo
mopsoppppo
mpm. zmp. pmo. omo. pomo oompooEpooep
**pmm. **mmz. >IB<Q Amwuzv
mopsoppppo
:mH. **mmm. **:Nm. **:mm. Homw 30d
pmpenmm opoepm pmpopmm opospm
menop mopppoa
¢mw o>HpmHSESO oompo zmmpwcm popmoEom ppm mpoomnsm

 

someopem oepoos
psoEo>opno¢ opsoooo¢ mo moopocH

 

.mopmaop poonom
nwps pOHCSh mmm pom mmooosm poonom mo appsoppppp oopmp on» po mpo>op oopnp po sumo
spnppz psoEo>opnoo opaocooo po moopccp 03p mo nooo cops A>IB<QV opoom ponpo>lupmoe

occuppoa poppcopopppa opp one AHomwv mpOpco>cH oopono poCOHpozppm oomppopocou one
po Aooppoppsoo mopoom >IB<QV meoHOprooo COHpopoppoo popppmd ppm opdspmlump.: mam¢e

179

Therefore, null Hypothesis 148 was not rejected with
either index of academic achievement for females.

It can also be seen in Table 4.18 that coefficients
of correlation, significantly different from zero, were
found for GSCI scores with both indices of academic achieve-
ment within both the "Low Difficulty" and "High Difficulty"

subgroups of females.
Summary of the Analysis of Results

Results within Stage I Analyses

Support was not found for the predicted negative
relationships of the Motive for Success (GSCI scores) with
the Joke Rating Time criterion of motivation (initiating
behavior), as stated in Hypothesis 1, using males or
females. However, a significant negative relationship was
found for academic ability (DAT—V scores) with the Joke
Rating Time criterion of motivation (initiating behavior)
using males.

Support was found for the predicted positive relation-
ship of the Motive for Success (GSCI scores) with the
Expected Grade criterion of motivation (directing behavior),
as stated in Hypothesis 2, using both males and females
separately.

Support was also found for the predicted positive
relationship of the Motive for Success (GSCI scores) with
the Theme Writing Time criterion of motivation (sustaining

behavior), as stated in Hypothesis 3,-but only using males.

180

Support was not found for the differential relation-
ships of the Motive for Success (GSCI scores) with any of
the three criteria of motivation, as stated in Hypotheses
4A, 5A and 6A or 4B, 5B and 6B, using indices of "apparent
task difficulty" (DAT-V scores and Rated Difficulty scores)
to influence (moderate) these relationships among subgroups
of males and females.

Within analyses using subgroups of males varying in
level of academic-ability (DAT-V scores), a significant
positive relationship was obtained for the Motive for Success
(GSCI scores) with the Theme Writing Time criterion of
motivation (sustaining behavior) ppiy among males within-
the highest one-third of the distribution of DAT-V scores.

Among males within the lower two-thirds of the distri-
bution of DAT-V scores, significant negative relationships
were obtained for academic ability (DAT-V scores) with the
Joke Rating Time criterion of motivation (initiating
behavior).

Within analyses using subgroups of males varying in
either academic ability (DAT-V scores) or perceived diffi-
culty (Rated Difficulty), significant positive relation-
ships for the Motive for Success (GSCI scores) with the
Expected Grade criterion of motivation (directing behavior)
were obtained only among males within the lower two-thirds

of the distribution of DAT-V scores.

181

Within analyses using subgroups of females varying in
level of academic ability (DAT-V scores), significant
positive relationships were found for the Motive for Success
(GSCI scores) with the Expected Grade criterion of motiva-
tion (directing behavior) only among females within the
highest two-thirds of the distribution of DAT-V scores.

Among females within the highest one-third of the
distribution of DAT-V scores, a significant positive rela-
tionship was found for academic ability (DAT-V scores)
with the Expected Grade criterion of motivation (directing
behavior).

Within analyses using subgroups of females varying
in level of perceived difficulty (Rated Difficulty) of
success in English, significant positive relationships
were obtained for both the Motive for Success (GSCI scores)
and acdemic ability (DAT-V scores) with only the Expected
Grade criterion of motivation-(directing behavior).v

Among the one-third of females who rate success in
English as "most" difficult ("High Difficulty" subgroup),
significant positive relationships were obtained for the
Motive for Success (GSCI scores) with both the Expected_
Grade criterion (directing behavior) and the Theme Writing

Time criterion (sustaining behavior) of motivation.

182

Results within Stage II Analysis

Support was not found for the predicted negative
relationship of the Joke Rating Time criterion of motiva-
tion (initiating behavior) with the level of academic
achievement, as stated in Hypothesis 7. A significant
relationship was found using a simple correlation but only
using one of the two indices of academic achievement
(Semester English Grade) and only with males. The above
relationship was not significant with academic ability
(DAT-V scores) held constant by partial correlation
technique.

Support was found for the predicted positive rela-
tionship of the Expected Grade criterion of motivation
(directing behavior) with the level of academic achieve-
ment, as stated in Hypothesis 8, using both indices of
the level of academic achievement (Theme Grade and Semester
English Grades), for both males and females.

Support was also found for the predicted positive
relationship of the Theme Writing Time criterion of motiva-
tion (sustaining behavior) with the level of academic
achievement, as stated in Hypothesis 9, using one of the
two indices of academic achievement (Semester English
Grade), but only for males.

No support was found for the predicted relationships
of Joke Rating Time (initiating behavior) and Theme

Writing Time (sustaining behavior) criteria of motivation

183

with either of the two indices of the level of academic
achievement, as stated in Hypotheses 7 and 9 respectively,
for females.

Support was not found for the differential relation-
ships of the three criteria of motivation with either of the
indices of the level of academic achievement, as stated in
Hypotheses 10A, 11A and 12A or 10B, 11B and 12B, using
indices of "apparent task difficulty" (DAT-V scores and
Rated Difficulty scores) to influence (moderate) these
relationships among subgroups of males and females.

Within analyses using subgroups of males varying in
level of academic ability (DATzV scores), significant
positive relationships were obtained for the Expected
Grade criterion of motivation (directing behavior) with at.
least one index of the level of academic achievement
(Semester English Grade) within each of the three levels
of academic ability (DAT-V scores) even with the influence
of DAT-V scores held constant (within subgroups) by
partial correlation technique.

Among males within the highest one-third of the
distribution of DAT-V scores (High Ability), a significant
positive relationship was obtained for the Theme Writing
Time criterion of motivation (sustaining behavior) with
one index of the level of academic achievement (Semester

English Grade).

184

Within analyses using subgroups of males varying in
level of perceived difficulty (Rated Difficulty) of success
in English, a significant positive relationship was
obtained for the Theme Writing Time criterion of motivation
(sustaining behavior) with one index of the level of
academic achievement (Semester English Grade) only among
the one-third of males who rate success in English within
a moderate (Intermediate) range of difficulty.

Among the one-third of males who rate success in
English as least easy (Low Difficulty), a significant
negative relationship was obtained for the Joke Rating
Time criterion of motivation (initiating behavior) with
one index of the level of academic achievement (Semester
English Grade). However, the above relationship was not
significant with the influence of academic aiblity (DAT-V
scores) held constant by partial correlation technique.

Within analyses using subgroups of females varying
in level of academic ability (DAT-V scores) or in level of
the perceived difficulty of success in English (Rated
Difficulty), significant relationships were obtained for
the EXpected Grade criterion of motivation (directing
behavior) with at least one of the two indices of the level
of academic achievement (Theme Grade or Semester English
Grade) within each of the subgroups and, in most cases,
even with the influence of academic ability (DAT-V scores)

held constant by partial correlation technique.

185

Within analyses using subgroups of females varying in
level of academic ability (DAT—V scores) or in level of
the perceived difficulty of success in English (Rated
Difficulty), significant relationships were p23 obtained
for either the Joke Rating Time (initiating behavior) or
Theme Writing Time (sustaining behavior) criteria of moti-
vation with either of the two indices of the level of

academic achievement within any of the subgroups.

Results within Stage III Analyees

 

Support was found for the predicted positive rela-
tionship of the Motive for Success (GSCI scores) with the
level of academic achievement, as stated in Hypothesis 13,
using both indices of the level of academic achievement
(Semester English Grades and Cumulative CPA), with the
influence of academic ability (DAT-V scores) held constant
by partial correlation technique, for males and females
separately.

Support was not found for the predicted differential
relationships of the Motive for Success (GSCI scores) with
the level of academic achievement, as stated in Hypothesis
14A, using either index of the level of academic-achieve-
ment (Semester English Grades or Cumulative GPA), with the
influence of academic ability (DAT-V scores) held constant
(within subgroups) by partial correlation technique, using
subgroups of either males or females varying in level of

academic ability.

186

Support was found for the predicted differential
relationships of the Motive for Success (GSCI scores) with
the level of academic achievement, as stated in Hypothesis
14B, using one of the two indices of the level of academic
achievement (Semester English Grade) with the influence of
academic ability (DAT—V scores) held constant (within sub-
groups) by partial correlation technique, using subgroups
of males varying in level of Rated Difficulty of Success
in school.

Within analyses using subgroups of males varying in
level of academic ability (DAT-V scores), significant
positive relationships (with DAT-V scores held constant
within subgroups) were obtained for the Motive for Success
(GSCI scores) with both indices of the level of academic
achievement (Semester English Grades_and Cumulative GPA)
ppiy within the extreme subgroups (one-third highest, and
one—third lowest in distribution of DAT-V scores) but ppp
within the moderate (middle one-third in distribution of
DAT-V scores) ability subgroup.

Within analyses using subgroups of males varying in
level of perceived (Rated) difficulty of success in English,
significant positive relationships (with DAT-V scores held
constant within subgroups) were obtained for the Motive for
Success (GSCI scores) with both indices of the level of
academic achievement (Semester Englisthrades and Cumulative

GPA) only within the moderate (middle one—third of

187

distribution of Rated Difficulty scores) difficulty group
but 222 within the extreme subgroups (one-third highest,

and one-third lowest in distribution of Rated Difficulty

scores).

Within analyses using subgroupscfi‘females varying in
level of academic ability (DAT—V scores), significant posi-
tive relationships (with DAT-V scores held constant within
subgroups) were obtained for the Motive for Success (GSCI
scores) with both indices of academic achievement (Semester
English Grade and Cumulative GPA) within the two-thirds of
females highest in ability (High Ability and Middle
Ability subgroups) but not witin the one-third of females
lowest in ability (Low Ability subgroup).

Within analyses using subgroups of females varying in
level of perceived (Rated) difficulty of success in English,
significant positive relationships (with DAT-V scores held
constant within subgroups) were obtained for the Motive for
Success (GSCI scores) with both indices of academic achieve-
ment (Semester English Grade and Cumulative GPA) within the
one-third of females who rate success as most difficult
(High Difficulty subgroup) and within the one—third of
females (Low Difficulty subgroup) who rate success as least
difficult, using the Semester English Grade index, but
significant relationships were not found using either index
among the one-third of females who rate success as of

moderate difficulty (Intermediate Difficulty subgroup).

CHAPTER V
DISCUSSION OF RESULTS

The discussion of results is presented below as it
relates to the three main purposes of the study which were
investigated within three Stages of analysis: Stage I-—
Nomological Validity, Stage II--Criterion Validity and

Stage III--Predictive and Nomological Validity.

Stage I--Nomological Validity

The first purpose of this study was to obtain evi-
dence relevant to the nomological (construct) validity of the.
Motive for Success, as the term (a) appears within-Atkinson's.
theory of achievement motivation, (b) is interpreted by
scores on the GSCI and, (c) is expressed in theory-selected
behaviors within the classroom situation.

It was assumed that Atkinson's statements, concerning
the influence of the Motive for Success on strength of the
tendency (motivation) to achieve, and Farquhar's definition
of the role of motivation in the academic situation, were
sufficiently explicit to allow selection of, and prediction.
for, three criteria of motivated behavior in the classroom--.
the initiating, directing and sustaining of achievement-

related activity. It was further-assumed that evidence

188

189

based on all three criteria of motivation could be more
confirming than evidence based on any one criterion.

Confirmation of directional hypotheses, concerning
relationships between strength of the Motive for Success
(GSCI scores) and the three behavioral criteria of strength
of the tendency (motivation) to achieve in the classroom
situation, was thought to provide one source of support for
the nomological validity of the Motive for Success.

A second source of validating evidence was sought
through evaluating hypothesized stronger relationships,
between strength of the Motive for Success (GSCI scores) and
the three behavioral criteria of strength of the tendency
(motivation) to achieve, among students thought to be
performing tasks of an "Intermediate" level of difficulty
than among students performing easier or more difficult

tasks.

Discussion of Relationships Between
GSCI Scores and Three Criteria of
Motivation for Males

 

 

 

Support for the nomological validity of the GSCI, as
an interpretation of the strength of the Motive for Success,

was not obtained using the Initiating criterion of strength

 

of the tendency (motivation) to achieve in the classroom
situation (Hypothesis 1).
However, support was found for.the nomological val—

idity of the GSCI, as an interpretation of strength of the

190

Motive for Success, through confirmation of predictions
(Hypotheses 2 and 3) concerning relationships of GSCI scores

with both the Directing, partial r = .214 (p < .01) and

 

Sustaining, partial r = .170 (p < .05), criteria of strength
of the tendency (motivation) to achieve in the classroom
situation.

While the coefficients of correlation, for the success-

ful prediction involving GSCI scores with both the Directing

 

and Sustaining criteria of strength of the tendency (moti-

 

vation) to achieve in the classroom, were low they were
significant both with and without the influence of academic
ability (DAT-V scores) held constant. Furthermore, signif-
icant relationships were not found for academic ability

(DAT-V scores) with either the Directing or Sustaining

 

 

criteria of motivation.

The above findings suggest that it was possible to
use statements concerning the role of motivation in the
classroom to select behavioral criteria relevant to a theory
of motivation and confirm predictions, based on a theory of
achievement motivation, concerning relationships between
GSCI scores, as an index of strength of the Motive for
Success, with two of the three criteria of strength of the
tendency (motivation) to achieve in the classroom situation.

The finding that neither the Directing nor Sustaining

 

 

criteria of motivation were significantly influenced by

191

academic ability (DAT—V scores) adds further clarity to the
nomological validity interpretation of relationships between

GSCI scores and these two behavioral criteria.

Discussioncfl‘lnitiating Activity.--Failure to find

 

significant relationships between GSCI scores and Initiating

 

activity did not confirm the prediction that males with
higher GSCI scores (assumed stronger in the Motive for
Success) would more quickly cease performance of a non-
achievement task and thus, more quickly undertake performance
of an achievement task than males with lower GSCI scores
(assumed weaker in the Motive for Success).

The above prediction was based on the theoretical
speculations of Atkinson and Cartwrightl concerning the Oper-
ation of an "Inertial Tendency" which once aroused is thought
to persist and influence subsequent performance. Postulation
of the "Inertial Tendency" was necessary for prediction of
behavior in a situation (non-achievement task) in which
"expectancy" and "incentive" values, relevant to the Motive
for Success, are not objectively present.

Prediction of behavior during performance of the non-
achievement task was thus based on the assumption that
"expectancy" and "incentive" values relevant to expression
of the Motive for Success were established during the explan-

ation of the achievement task and multiplicatively combined

 

1J. W. Atkinson and D. Cartwirght, "Some Neglected
Variables in Contemporary Conceptions of Decision and Per-
formance," Psychological Reports, 14 (1964), 575-590.

 

192

with stength of the Motive for Success to determine the
strength of a tendency (motivation) to achieve which would'
be reflected in the length of subsequent time spent on the
non-achievement task. Thus it was cues associated with the
achievement task that were used for the prediction of
behavior on the non-achievement task.

The present findings, however, suggest that when cues,
of the task being performed do not allow for satisfaction-
of the Motive for Success then performance on that task is
unrelated to differences in strength of the Motive for
Success. Such a finding is in accord with findings of
other researchers which suggest that when "chance" rather
than "skill" is involved in task performance2 or when task
performance appeals to motives other than the Motive for

3 then relationships between strength of the Motive

Success
for Success and behavior are not found.

However, academic ability (DAT-V scores) was found,
in this study, to be significantly correlated (negatively)
with time spent in performance of the non-achievement task'
before starting to perform the achievement task. That is,
males of lower_ability tended to spend more time-in per-

formance of the non-achievement task before starting the

achievement task than students of higher ability.

 

2NormanT. Feather, "Subjective Probability and
Decision Under Uncertainty,"Psychological Review, 66
(1959), 150-164.

 

3Elizabeth G. French, "Some Characteristics of Achieve-
ment Motivation," Journal of Experimental Psychology, 50‘
(1955), 232-236.

193

The above finding could be interpreted as suggesting
that time spent in performance of the non—achievement task
was reflecting the strength of an "Inertial Tendency" based
on characteristics of the achievement task, but that the
tendency (motivation) had an-"avoidant" quality rather than
the proposed "approach" quality.

Atkinson“ has prOposed the operation of a tendency to
avoid performance of an-achievement task that is a conse-
quence of the multiplicative relationship of a Motive to
Avoid Failure, expectations concerned with the likelihood
of failure and the negative incentive value of failure.

The operation of a tendency (motivation) to avoid
failure had not been considered in this study since the
concern was with the nomological validity of GSCI scores as
an interpretation of the strength of the Motive for Success
thought to influence the strength of a tendency (motivation)
to approach success.

However, the finding of a negative relationship
between DAT-V scores and time spent on the non-achievement
task before starting performance of the achievement task is
reminiscent of the negative relationship found by Smith5
between intelligence (Otis test scores) and time spent in

both a mid-term and final examination room.

 

“J. W. Atkinson, "Motivational Determinants of Risk-
Taking Behavior," Psychological Review, 64 (1957), 359-372.

 

5Charles P. Smith, "Relationships Between Achievement-
Related Motives and Intelligence, Performance Level, and
Persistence," Journal of Abnormal and Social-Psyphology,

68 (1964). 523—533.

194

Though Smith proposed other interpretations for his
findings, it appears that one interpretation of his findings
could be that, with the multiple-choice and short-answer
examination task given his subjects, differences in total
time spent in the examination room might be attributable to
differences in time spent performing "extra-" achievement
task activities.

While in the present study, it may be assumed, that
the attractiveness of the alternative activity (Joke Rating)
minimized the influence of unaccounted for, "extra-task,"
behavior on the total time spent in performance of the
achievement task.

This interpretation of the present findings and those
of Smith's suggest the importance of providing an alternative
activity (non—achievement task) in an investigation that
attempts to locate relationships between the Motive for
Success and behavior and raises the possibility that behavior
on a non—achievement task preceeding performance of an

achievement task may be a consequence of avoidant tendencies.

Discussion of Directing Activity.--Significant rela-

 

tionships were found for strength of the Motive for Success

(GSCI scores) with the Directing of achievement-related

 

activities as reflected by expectations concerned with

future achievement (Expected Grades).

195

That is, males with high GSCI scores (assumed strong
in the Motive for Success) stated they expected to achieve
higher grades (chose more difficult objectives) than males
with low GSCI scores (assumed weak in the Motive for
Success).

The choice of more difficult taskscn°the choice of a
higher level of ahcievement on the same task was thought to
reflect one influence of strength of the Motive for Success
on behavior in the classroom situation.

The interpretation of differences in expectations
concerned with grades as reflecting a Directing influence
of the Motive for Success assumes that there is a path or
sequence of activities associated with one level of academic
achievement that differs from the path or sequence of
activities associated with another level of achievement.
The statement of the level of eXpected achievement can then
be interpreted as indicating that students are following
one path rather than another.

Whether one accepts this interpretation of expecta-
tions concerned with future levels of achievement, the
findings of this study were in accord with previous research
involving relationships between strength of the Motive for

Success and "Expected Grades."

196

Atkinson6 had proposed that significant relationships
would be found between TAT n-achievement scores and grades
expected on a college course final examination when differ-
ences in ability were held constant by partial correlation
technique. One interpretation of the effect of the partial
correlation technique, in this situation, is that it
reflects the extent of the relationship, occurring as a
discrepancy between expected achievement and predicted.
achievement, with motive scores.

While Atkinson failed to find significant relation-
ships between TAT n—achievement scores and expected grades,
the present study found significant relationships both with
and without the effects of academic ability (DAT-V scores)

controlled by partial correlation technique.

Discussipn of Sustaining Activity.--Significant rela-
tionships were found for the strength of the Motive for

Success (GSCI scores) with the Sustaining criteria of

 

strength of the tendency (motivation) to achieve in the
academic situation, That is, males with high GSCI scores
(assumed strong in the Motive for Success) tended to spend
more time in the performance of an achievement task than
males low in GSCI scores (assumed weak in the Motive for

Success).

 

6D. McClelland and J. Atkinson, et al., The Achievement

Motive (New York: Appleton-Century-Crofts, 1953).

197

The above findings are in accord with the_relatione
ships hypothesized by Feather,7 Atkinson and Litwin8 and
Smith.9 However, only Atkinson found significant rela-
tionships between scores on a projective measure of the
Motive for Success and time spent on an achievement task
(college course examination). Atkinson failed to find
significant relationships with time spent on an examina-
tion using an objective (EPPS) technique for assessment
of strength of the Motive for Success.

Discussion of Relationships Between

GSCI Scores and Three Criteria of
Motivation for Females

 

 

 

Support for the nomological validity of the Motive
for Success and for Atkinson's theory of achievement motiva-
tion, using female subjects, was found only in the predicted
relationship (Hypothesis 2) of GSCI scores with the

Directing, partial r = .265 (p < .01), criteria of strength

 

of the tendency (motivation) to achieve in the classroom

situation.

 

7Norman T. Feather, "The Relationship of Persistence
at a Task to Expectation of Success and Achievement Related
Motives," Journal of Abnormal and Social Psychology, 63
(1961), 552- 561.

8John W. Atkinson and G. H. Litwin, "Achievement
Motive and Test Anxiety Conceived as Motive to Approach
Success and Avoid Failure," Journal of Abnormal and Social
Psychology, 60 (1960), 52-63.

9

 

 

 

Smith, op. cit.

198

That is, females with higher GSCI scores (assumed
strong in the Motive for Success) state they expect to
achieve higher grades than females with lower GSCI
scores (assumed weak in the Motive for Success).

However, the "Expected Grade" index of the influence
of the Motive for Success on the Directing of achievement-
related activities provides, (a) the least explicit deri-
vation from theory (differences in expected level of achieve-
ment holds questionable status as a directional concept),

(b) the least clear distinction from achievement criteria
(both based on teacher-assigned grades) and, (c) the least
clarity for nomological validity interpretations of GSCI
scores with "Expected Grades" (DAT-V scores correlate equally
well with this criterion).

The failure to find significant relationships between
GSCI scores for females and their behavior during the per-
formance of either the achievement task or the non-achievement
task suggests that GSCI scores for females may not be
reflecting the operation of the same construct as GSCI scores
for males.

GSCI scores for females may be reflecting the operation
of some personality construct, as indicated by the relation-
ship between GSCI scores and "Expected Grade," but there is
little evidence to suggest that the construct reflected by

GSCI scores for females has the characteristics of the

199

Motive for Success which appears to be, at least, more
adequately reflected by GSCI scores for males.

In any event, it appears reasonable to assume that
combining females in the analyses for males would have
served to diminish the strength_of the relationships found
with males alone in the present study and may have accounted
for Smith'slO failure to find significant relationships
between motive-test scores and persistence on his achieve-
ment task.

Discussion of Relationships of GSCI Scores
with Three Criteria of Motivation for Sub-
groups of Males or Females Varying in

Objective or Subjective Estimates of
the Difficulty of School Success

 

 

 

 

Two indices of the difficulty of success in school, an
objective measure of academic ability(DAT-V scores) which
assumed an inverse relationship between ability test scores
and "perceived" difficulty of success, and a subjective
measure of eXpressed difficulty (Rated Difficulty) which
assumed a direct relationship between stated difficulty and
"perceived" difficulty, were used to evaluate the predicted
differentially stronger influence of the Motive for Success
on the behavior of students performing tasks of an "Inter-
mediate" level of difficulty than on the behavior of students

performing easier or more difficult tasks.

 

lOIbid.

200

No support was found, using either males or females,
with either objective or subjective estimates of the
"perceived" difficulty of school success, for the predicted
differentially stronger influence of the Motive for Success
(GSCI) scores on the initiating, directing and sustaining
criteria of motivation for students performing tasks of an
"Intermediate" level of difficulty than for students
performing easier or more difficult tasks.

The reasons for the failure to find differentially
stronger relationships as predicted by theory could lie
with, (a) the method of selecting subgroups, (b) the
psychometric characteristics of the GSCI within different
scale ranges, (c) the low validity of both the DAT—V scores
and Rated Difficulty scores as indices of "perceived"
difficulty or, (d) with theory. These possibilities are
discussed below.

(a) The selection technique used for the formation
of subgroups thought to vary in the "perceived" difficulty
of school success consisted of trichotomizing the separate
distributions of DAT—V scores and Rated Difficulty. This
technique assured subgroups composed of equal thirds of
the samples of males and females and was used to correspond
with the technique used by Goodstein and Heilbrunll and to

provide for the least "manipulation" of data.

 

*lLeonard D. Goodstein and Alfred B. Heilbrun, Jr.,
"Prediction of College Achievement from the Edwards Personal
Preference Schedule at Three Levels of Intellectual Ability."
Journal of Applied Psychology, 46 (1962), 317-320.

 

201

While alternative selection techniques might have
produced the predicted stronger relationships within
"middle" groups, it was thought that such findings would
hold less relevance for theory.

(b) Because GSCI scores are correlated with both
DAT-V scores and with Rated Difficulty scores, selection
of subgroups on the basis of the latter two instruments
automatically selects subjects within varying ranges of
GSCI scale scores. The consequence of comparing relation-

ships based on scores within varying ranges of GSCI scores

 

is, of course, the confounding of potentially differing
scale invalidities with invalidity attributable to theory.

However, it was anticipated on the basis of theory
that students within differing levels of objective and sub-
jective estimates of school success would differ in both
motive strength and motivation as a consequence of past
experience and relationships found would therefore, at
least provide the closest correspondence with "real-life"
competitive situations.

(c) The failure of DAT-V scores anad Rated Diffi-
culty scores to correlate highly with achievement indices
or with each other suggests that neither may hold validity
as indices of "perceived" difficulty of school success.
The brevity of the DAT—V scale, dealing with only the
Verbal factor of academic ability, may have contributed to

its failure to function as predicted. While the relatively

202

crude index of Rated Difficulty may have been more effective
if attention had been given to a definition of "success" or
by comparing student perceptions of what constitutes success
for them.

(d) While failure to find relationships as specified
by theory might be attributable to the inadequacies of
theory, it appears that sufficient possibilities exist in
the inadequacies of the assessment techniques for the
variables involved in this study to make such speculation
hazardous.

However, relationships between GSCI scores and the
three behavioral criteria of motivation were found to vary
between subgroups of students, though not as predicted by
theory. The findings hold implications for theory or at
least for methodology of future investigations.

Significant relationships, between GSCI scores
(assumed to reflect differences in strength of the Motive
for Success) and time spent in the performance of the
achievement task (assumed to reflect differences in strength
of the tendency to achieve success), were found for males
only among subgroups objectively or subjectively pppp likely
to obtain success (High Ability and Low Rated Difficulty
subgroups).

On the other hand, significant relationships, between
DAT-V scores (assumed to reflect differences relative to a

Motive to Avoid Failure) and time spent in performance on

203

a non-achievement task (assumed to reflect strength of a
tendency to avoid failure), were found for males only among
subgroups objectively or subjectively ieapp likely to
obtain success (Middle and Low Ability, and High Rated
Difficulty, subgroups).

The above two findings imply that differences in the
difficulty of school success influence the expression of
achievement-related motives (approach and avoidant) in
behavior and the effects of these different motives may be
seen when alternative activities (non—achievement and
achievement tasks) are provided and behavior separately
analyzed.

Thus, the present findings incorporate both the posi-
tive relationship between the Motive for Success and time
spent on an examination, found by Atkinson and Litwin,l2
and the negative relationship between ability and time
spent on examinations, found by Smith.13 The discrepancy
between the findings of Atkinson and Smith may be attrib-
utable to their failure to provide and measure alternative
activities for males.

The analysis of relationships between GSCI scores and
the three behavioral criteria of motivation, among subgroups

of females, tended to produce significant findings only with

 

a“, _W~= ., . _ .__.- .. “t

12Atkinson and Litwin, op._cit.

l3Smith, pp___e_ip

204

the Directing_ (Expected Grade) criteria, thus, confirming
the findings based on the total sample of females and
supporting the importance of separate analyses for males and
females.

However, an interesting relationship was found

between GSCI scores and the Sustaining of activity during

 

the performance of an achievement task among females who
rate success in English class as mppp difficult (High Rated
Difficulty subgroup).

That is, among the one—third of females who rate
success as 3223 difficult, those females with higher GSCI
scores tended to spend longer on the achievement task than
females with lower GSCI scores.

The above finding with females stands in marked
contrast to the finding with males who rate success as p233
difficult, for among that subgroup, males with lower aca-
demic ability (DAT—V scores) tended to spend more time in
performance of the non-achievement task than males with
higher academic ability.

The above findings suggest a fundamental difference
in the way males and females respond to a task of high
difficulty. That is, females tend to devote increasing
amounts of time to performance of a difficult achievement
task possibly in attempts to ggip satisfaction for needs to

achieve, while males, faced with a difficult achievement

205

task, devote increasing amounts of time to performance.
of a non-achievement task possibly in attempts to avoid

failure.

Stage II--Criterion Validity

 

The second main purpose of this study was to obtain
evidence concerning the validity of the three.behaviora1-
criteria strength of the tendency (motivation) to achieve
in the classroom situation.

It was possible to investigate both the theoretical
and practical implications of indices of initiating,
directing and sustaining achievement—related behavior,
separately from the problems of assessment of strength of
the Motive for Success, by assuming that level of academic
achievement could be used as one criterion of strength of
the tendency (motivation) to achieve in the academic
situation.

Theory was used to derive directional hypotheses
that the three behavioral criteria of strength of the
tendency (motivation) to achieve should correlate (nega-
tively in the case of delay in Initiatingactivity, and
positively in the case of both Directing and Sustaining
activity) with indices of level of academic achievement
as would be eXpected of indices of the same construct

(strength of the tendency to achieve).

206

Theory was further used to derive differential
hypotheses that the three behavioral criteria of strength
of the tendency to achieve should correlate stronger with
indices of the level of academic achievement (also con-
ceived as a criterion of strength of the tendency to
achieve) among subgroups of students whose behavior was
thought to be maximally influenced by the Motive for
Success, i.e., for students performing tasks of.an~"lnter-
mediate" level of difficulty, than for students performing
easier or more difficult tasks.

Two distributions of teacher-assigned grades were
used to evaluate relationships between the three behavioral
criteria of motivation and academic achievement in English
class.

One index consisted of the grade (Theme Grade)
assigned by each student's teacher for the achievement task
on which the initiating, directing and sustaining criteria
of motivation were also obtained. The second index of-
academic-achievement consisted of the grade assigned by
each student's teacher for the semester-of English (Semester
English Grade) during which the achievement task.was adminz
istered. Both grades were assigned after collection of the
behavioral criteria of motivation.

~The "Semester English Grade" index was used because

it was thought to provide a more stable estimate of

207

achievement than was possible for the "Theme Grade" index~

(based on only one grade for one brief assignment).

Discussion of Relationphips for the Three
Behaviorainriteria of Motivation-with
indices of Academic Achievement

for Males

 

Support for the construct validity of the Initiating
criteria of strength of the tendency (motivation) to
achieve (Hypothesis 7) for males was obtained only with
the coefficient ofsimple correlation for this criterion
with one index (Semester English Grade) of academic
achievement, simple r = .171 (p < .01), but this relation-
ship was not significant with the influence of academic~
ability (DAT-V scores) held constant by partial correla-
tion, partial r = -.111 (p > .05).

Support for the construct validity of the Directing
criteria of strength of the tendency (motivation) to
achieve (Hypothesis 8) for males was obtained with coeffi-
cients of partial correlation for this criterion with both
the "Theme Grade," partial r-= .271 (p < .01), and with
the-"Semester Englisthrade," partial r = .301.(p < .01),
indices of academic achievement.

Support for the construct validity of the Sustaining
criteria of strength of the tendency (motivation) to achieve
(Hypothesis 9) for males was obtained with the coefficient
of partial correlation for this criterion with only the
"Semester English Grade,".partial r = .161 (p < .05), index

of academic achievement.

208

The significant relationships, found with the three
behavioral criteria of strength of the tendency (motivation)
to achieve, were thought to provide support for the power
of Atkinson's theory of achievement motivation to select
characteristics of behavior relevant to theory and to
confirm predictions for these criteria with indices of-
academic'achievement.

While the coefficients of correlation, for the
behavioral criteria of motivation with academic achieve-
ment, were all low, they were of approximately the same

magnitude as coefficients found for academic ability

(DAT-V scores) with "Theme Grade," r .242 (p < .01),

.357 (p < .01),

and with "Semester English Grade," r
indices of academic achievement in English class.

The failure of the measure of academic ability
(DAT-V scores) to correlate highly with indices of academic
achievement in English class for males is, in part,
attributable to the relatively low reliability (stability)
of the teacher-assigned grades as_indices of a consistent
level of academic achievement. However, the low relia-
bility of these teacher-assigned grades may also reflect
a variability in student classroom behavior which influ-
ences achievement in particular situations.

The significant relationships, found for the three
behavioral criteria of motivation with indices of academic

achievement, suggests that such criteria may have practical

209

implications as descriptions of events (motivational.
phenomena), within "real-life" competitive situations,
intervening between gross predictor variables and global

achievement criteria.

Discussion of Initiating Activity.--The relation-

ships found between the Initiating criteria of strength

 

of the tendency (motivation) to achieve and level of
academic achievement indicates that those males who spend
more time in performance of a non-achievement task, before
starting to perform an achievement task, tend to obtain
lower grades than males who spend less time in performance
of a non-achievement task. However, this relationship was
not significant with the influence of academic ability
(DAT-V scores) held constant by partial correlation tech-
nique.

When an alternative (non-achievement) task is pro-
vided in a competitive situation, the behavior of males on
the non-achievement task reflects tendencies to avoid per-
formance of the achievement task and these avoidant
tendencies have an influence (negative) on later achieve-
ment.

The failure to find a significant influence of
avoidant tendencies on later achievement with academic-
ability (DAT-V scores) controlled can be explained as a
consequence of the significant (negative) relationship

found between DAT-V scores and behavior on the non—achievement

210

task (reported in Stage I analyses). Partialing DATgV
score out of the above relationship, therefore, amounts to
controlling the unique characteristic of behavior expressed
in behavior on the non-achievement task.

Recent research by Atkinsonlu was concerned with the
construct of "Avoidant Tendencies" through attempts to
separately assess strength of a Motive to Avoid failure.
(indexed by TAQ scores) and predict the combined influence
of approach and avoidant motives on achievement task
behavior.

7 Findings in the present study indicate that both
approach and avoidant tendencies may influence behavior
in competitive situations but in addition to the need for
the separate assessment of personality characteristics~
(motives) there is a need for separate evaluation of the
effects in behavior. The provision of an alternative
(non-achievement) activity appears to provide clarity in.
the expression of avoidant tendencies and allows opportunity
for the separate assessment of the influence of personality

on separate_task behaviors.

Discussion of Directing Activity.—-The relationships'

found between the Directing criteria of strength of the

 

tendency (motivation) to achieve and indices of level of

academic achievement indicate that males who expect to

 

1“Atkinson and Litwin,'0p. cit.

211

obtain higher grades receive higher grades from their
teacher than males who expect to obtain lower grades.
The above relationships confirm the findings of

Atkinson15

using college students. However, Atkinson used
grades on a previous examination and grades accumulated
over previous semesters. The present study used grades
assigned 23222 the statements concerned with grade expec-
tations were collected. The findings of the present study
provide greater clarity for the potential influence of
grade expectations on future achievement in the academic-
situation.

Combining the findings of both Atkinson and the
present study suggests that statements concerned with
expected grades are tied to both past achievement and
future achievement and require rigorous analysis to iden-
tify those characteristics holding uniquely motivational
implications. The finding that relationships between
expected grades and future grades were significant with

academic ability (DAT-V) controlled by partial correlation,

was only a step in that direction.

Discussion of Sustaining Activity.--The relationships

between the Sustaining criteria of strength of the tendency

 

(motivation) to achieve and level of academic achievement

indicate that males who spend more time in the performance

 

15John W. Atkinson (ed.), Motives in Fantasy Action
and Society (D. Van Nostrand Co., Inc., 1958).

 

212

of an achievement task obtain higher grades from their
teacher than males who spend less time in the performance
of an achievement task..

The above findings were similar to those sought by
Atkinson and Litwinl6 and by Smith.17' However, both
previous researchers failed to find significant-relation-
ships between time spent on an achievement task (college
course examinations) and 1eve1.of academic achievement.~

The difference between the previous non—significant
relationships and the present significant relations could
be attributed to the fact that both Atkinson and Litwin
and Smith used achievement tasks consisting of short-
answer and multiple—choice examinations. Furthermore,
they failed to provide an alternative (non-achievement)
activity. The present study provided an achievement task
consisting of a theme, with opportunity for correction and
revision, and provided an alternative (non-achievement)
task.

It is difficult to see just how the-strength.of‘
tendencies to achieve might have-been effectively expressed
in the short-answer and multiple-choice tasks particularly
if the college students were influenced by the oft heard

dictum that "your first guess is probably your best guess."

 

16Atkinson and Litwin, op. cit.

l7Smith, op. cit.

213

Discussion of Relationships for the Three
BehavioraI Criteria of Motivation with
Indices of Academic Achievement

for Females

 

Support for the construct validity of the three
criteria of strength of the tendency (motivation) to
achieve with indices of academic achievement for females

was found only in the relationships for the Directing

 

criteria with "Theme Grade," partial r = .230 (p < .01),
and with "Semester English Grade," partial r = .354
(p < .01), indices of academic achievement.

Females who eXpected to obtain higher grades were
assigned higher grades by their teacher than females who
expected to receive lower grades.

The coefficients of correlation for grades expected
with grades obtained by females were significant with the
influence of academic ability controlled. The finding
that expected grades were themselves significantly corre-
lated with academic ability (finding from Stage I analyses
with females but not with males) suggests that females
may have stated grade expectations, and teachers assigned
grades for females, on a similar basis.

However, the failure to find significant relation-
ships between the behavior of females, on either the achieve-
ment task or the non-achievement task, with teacher-assigned
grades suggests that achievement-related behavior for

females differs from that for males.

214

If achievement-related activities for females differ
from those for males then it would appear'that,in.addition
to problems of assessment of the Motive for Success in.
females (suggested above in Stage I Discussion) there may
also be a problem of identifying the activities or dimen-
sions of behavior that influence academic achievement for
females.

The above discussion implies additional problems in
dealing with motivational influences for females. The
findings also point to a direction for solving the problem
by suggesting that identification of achievement-related
behaviors for females could serve as the basis for the
development or refinement of motive assessment techniques.

Observation of the behavior of females in the class-
room, preceding and following administration of_the experi-
mental tasks for this study, indicated that females were
often engaged in washing blackboards-and assisting the
teacher with various tasks. These casual observations
suggest that the achievement-related activities of females
may have to be considered in the context of the total
classroom situation rather than in the performance of

specific assignments.

215

Discussion of Relationships for the Three
Behavioral Criteria of Motivation with
Indices of Academic Achievement for
Subgroups of Males or Females Vary-

ing in Objective or Subjective

Estimates of the Difficulty of
School‘Success

 

No support was found, using either Objective (DAT-V
scores) or Subjective (Rated Difficulty) estimates of the
"perceived" difficulty of school success, for the predicted
(Hypotheses 10A, 11A, 12A or 10B, 118 and 12B) stronger
relationships of the behavioral criteria of strength of the
tendency (motivation) to achieve with indices of academic
achievement (also assumed to reflect strength of the.
tendency to achieve) among males or females assumed to be
performing tasks of an "Intermediate" level of difficulty
than for students performing easier or more difficult
tasks.

A discussion of the possible reasons for the failure
to find differentially stronger relations as predicted by
theory was presented above (Stage I Discussion of Results).

Significant relationships between the behavioral
criteria of motivation and level of academic achievement
were located within some of the subgroups of students
varying in either Objective or Subjective estimates of
school success.

Only among males objectively (DAT-V scores) pppt'
likely to obtain success in school (High.Ability subgroup)

were significant relationships-found for time spent in

216

performance of the achievement task (Sustainingactivity)
with level of academic achievement (Semester English Grade).

On the other hand, only among males subjectively
(Rated Difficulty) igapp likely to obtain-success in school
(High Difficulty subgroup) was a significant relationship
(negative) found for time spent on the non—achievement task
with level of academic achievement (Semester English-
Grade).

Combining the above.two findings suggests that aca—
demic achievement may have different behavioral correlates,
or may be more clearly expressed, for males pppp and ipapp
likely to obtain success in school.

That is, males most likely to obtain success display
differences in the time they spend performing an achieve-
ment task, a difference which may influence (positively)
their later level of achievement, while males least likely
to obtain success display differences in the time they
spend performing a non-achievement task, a difference which
may influence (negatively) their later level of achievement.

One interpretation of the above findings is that, by
analyzing behavior on a non-achievement task separately
from behavior on an achievement task, it is possible to see
more clearly two characteristics of behavior which are often
considered together. Males may spend similar amounts of
time between receiving a school assignment and handing it in"

for grading, however some males may spend increasingly

217

greater lengths of time in actual performance of the assign-
ment while other males may spend increasingly greater
lengths of time in extra-assignment (avoidant) activities,
depending upon the likelihood of success for the respective
groups.

Providing both achievement and non-achievement tasks
for groups of males who differ in the likelihood of school
success allows (a) greater clarity in the expression of
approach and avoidant-tendencies, (b) opportunity to.
precisely index personality correlates of approach and
avoidant tendencies and (c) increased accuracy in evaluating
the separate and combined influences of approach and avoidant
tendencies on academic-achievement.

For subgroups of females, however, relationships were
only found for Expected Grades with later academic achieve-
ment. These relationships were fairly uniform among all
levels of objective and subjective estimates of school
success.

Apparently females are fairly uniformly able to
estimate their level of future achievement, but neither
time spent on achievement nor on non-achievement tasks
appears to influence later achievement even when attention

is given to variation in the likelihood of school success.

218

Stage III--Predictive and Nomological Validity

The third main purpose of the study was to evaluate
evidence for both the predictive validity and nomological
(construct) validity of the Motive for Success as inter-
preted by scores on the GSCI and as expressed by achieve-
ment in the academic situation.

It was possible to investigate both the predictive
validity and nomological validity of GSCI scores by
assuming that indices of level of academic achievement
could be used as pp; criterion of strength of the tendency
(motivation) to achieve in the academic situation.

Theory was used to derive directional hypotheses
that, with differences in academic ability (DAT-V scores)
controlled, there should be a positive relationship
between GSCI scores, as an index of strength of the Motive
for Success, and indices of level of academic achievement,
as a criterion of strength of the tendency (motivation)
to achieve in the academic situation.

Theory was further used to derive differential
hypotheses that, with differences in academic ability
(DAT-V scores) controlled, there should be a stronger
relationship for GSCI scores with academic achievement
among subgroups of students performing school tasks of
an "Intermediate" level of difficulty, than for students

performing easier or more difficult tasks.

219

Level of academic achievement was, therefore, thought
to hold implications both for theory as a criterion of
strength of the tendency (motivation) to achieve and for
practical usefulness of the GSCI since level of academic
achievement constitutes a measure of "worth—while" accom-
plishment in the academic situation.

Finding significant relationships between GSCI scores
and indices of academic achievement could thus confirm
theory predictions and indicate the level of usefulness of

GSCI scores for predicting worth-while accomplishment.

 

Finding significantly stronger relationships as predicted
by theory could further confirm that part of theory and
identify subgroups of students for whom GSCI has maximal
and minimal predictive validity.

Two distributions of teacher-assigned grades were
used to evaluate relationships between GSCI-scores and
academic achievement.

One index consisted of the grade assigned by each
student's teacher for the semester of English (Semester
English Grade) during which the achievement task was
administered. This index of academic achievement was
used, as it was also used in Stage II analyses, because it
was thought to provide a reasonably stable estimate of
achievement in academic activities related to the experi—
mental task (Theme Writing) and because "difficulty of doing

well in this English class" constituted one of the bases for

220

the formation of subgroups of students thought to differ
in the "perceived" difficulty of school success.

The second index of academic achievement (Cumulative
GPA) consisted of grades accumulated over two years, for
courses requiring homework, divided by the number of such
courses for each student. The Cumulative GPA index of'
academic achievement was thought to be less relevant for
the nomological validity of the Motive for Success because
of the variability of situational factors involved which
theory demands be controlled, but was considered to repre-
sent a more reliable estimate of academic achievement.
Discussion of Relationships for GSCI
Scores for Males, and GSCI Scores

for Females, with Indices of
Academic Achievement

 

 

For males, support for the nomological validity of,
the Motive for Success and evidence for the predictive
validity of GSCI scores-for males was obtained in the
correlations (DAT-V scores held constant by partial corre-
lation) of GSCI scores with "Semester English Grade,"
partial r = .299 (p < .01), and with "Cumulative GPA,"
partial r = .302 (p < .01), indices of academic achievement
(assumed to reflect strength of the tendency to achieve
in the academic situation).

For females, support was found for the nomological
validity of the Motive for Success and evidence for the

predictive validity of GSCI scores for Females in the

221

correlations (DAT-V scores held constant by partial corre-

lation) of GSCI scores with "Semester English Grade,"

partial r .285 (p < .01), and with "Cumulative GPA,"

partial r .289 (p < .01), indices of academic achievement
(assumed to reflect strength of the tendency to achieve in
the academic situation).

Relationships of measures of academic ability (DAT-V
scores) with indices of academic achievement were of
approximately the same magnitude, r = .357 (p < .01) and
r = .340 (p < .01) for males and for females, r = .424
(p < .01) and r = .453 (p < .01), and were of approximately
the same magnitude as those reported above for GSCI scores
with indices of academic achievement.

While the coefficients of correlation for both GSCI
scores and DAT-V scores with the indices of academic achieve-
ment were low, the interpretation of the effects of par-
tialing DAT-V scores out of the relationship between GSCI
scores and academic achievement as that of correlating GSCI
scores with the discrppancy between predicted and obtained
levels of achievement,18 at least suggests the possibility
that the predictive efficiency estimates for the two tests
may be accumulative. That is, a more accurate estimate of

academic achievement may be obtained from a combination of

GSCI scores and DAT-V scores than either alone.

 

18Robert L. Thorndike, The Concepts of Over- and
Underachievement (New York: Teachers College, Columbia
University, Bureau of Publications, 1963).

 

222

While multiple correlation techniques necessary to

test the above hypothesis were not applied to the present

19

data, Farquhar has reported finding significant increases

in predictive efficiency by adding GSCI scores for either
males or females to DAT-V score predictions of grade

averages.

Discussion of Relationships for GSCI
Scores with Indices of Academic
Achievement for Subgroups of

Males Varying in Objective

or Subjective Estimates of

the Difficulty of School

Success

 

 

 

 

 

 

Support for the nomological validity of the GSCI as
an interpretation of strength of the Motive for Success for
males was found through confirmation of the predicted
stronger relationship (with DAT-V scores controlled) of GSCI
scores with one index of academic achievement (Semester
English Grade) using subjective estimates (Rated Difficulty
of success in English class) to locate males thought to be
performing tasks of an "Intermediate" level of difficulty,
than for males performing easier or more difficult tasks.

Support was not found for the differentially stronger
relationships (with DAT-V scores controlled) between GSCI
scores and academic achievement using objective estimates

(DAT-V scores) to locate males thought to be performing

 

19William W. Farquhar, Motivation Factors Related to
Academic Achievement, Cooperative Research Project 846,
(East Lansing, Michigan: Michigan State University, Office
of Research and Publication, 1963).

 

 

223

tasks of an "Intermediate" level of difficulty, than for
males performing easier or more difficult tasks.

While the above confirmation of the predicted differ-
entially stronger influence of the Motive for Success on
strength of the tendency to achieve (level of academic
achievement) represents the only instance of confirmation
of the differential predictions in this study, it appears
that confirmation occurred in the one most easily ration-
alized location. That is, using GSCI scores as a reliable
estimate of strength of the Motive for Success and "Semester
English Grade" as a more stable estimate of strength of the
tendency to achieve than either the behavioral criteria
(though these may be more valid) or the "Theme Grade" but
less influenced by variability in situational influences
than "Cumulative GPA," contributes to the ability to
confirm the differential predictions.

The use of subjective (Rated Difficulty of success in
the present English class) estimates of task difficulty
corresponds closely with Atkinson's statements that it is
"perceived" difficulty of the task that is expected to
influence the expression of motive strength in behavior
(motivation). Because difficulty was rated specifically in
regard to achievement in the current English class it
appears the "Semester English Grade" would represent the
most relevant criterion for strength of the tendency to

achieve in that class.

224

The failure to confirm the differential predictions
using academic ability (DAT—V scores) to form subgroups
thought to vary in the "perceived" difficulty of school
success represents a failure to confirm the findings of
Goodstein and Heilbrun.20 However, Goodstein and Heilbrun
used EPPS n-achievement scores and noted only whether the
relationship with semester grade averages was significant
or not within subgroups of male college students varying
in scores on a brief measure of verbal ability.v

The more rigorous test, that relationships be sig-
nificant app significantly greater than relationships
within other subgroups, used in the present study, was
chosen in light of Hakel's21 recent failure to confirm
Goodstein and Heilbrun's findings even though the same
instruments, procedures and samples were used in both
studies.

The impetus for attempts to use academic ability test
scores as indices of the "perceived" difficulty of school
success has been the obvious efficiency that would accrue
through the use of scores readily available in most
academic situations to both confirm theory implications and
locate subgroups of students for whom personality tests

hold maximal or minimal levels of predictive accuracy.

 

2OGoodstein and Heilbrun, 0p. cit.

21Milton‘D. Hakel, "Prediction of College Achievement
from the Edwards Personal Preference Schedule Using Intel-
1ectua1 Ability as a Moderator," Journal of Applied
Psychology, 50 (1966), 336—340.

 

 

225

While confirmation of differentially stronger rela-
tionships between the Motive for Success and indices of
strength of the tendency (motivation) to achieve was ob-
tained with only one index of academic achievement and only
using subjective estimates of the "perceived" difficulty of
success in English class, significant relationships between
GSCI score and academic achievement indices were located
within some subgroups of males, formed on the basis of
objective or subjective estimates, and not within other
subgroups.

Within subgroups of males varying in level of academic
ability (DAT-V scores), significant relationships (with
DAT-V scores held constant within subgroups) were obtained
for GSCI scores with both indices of academic achievement
ppiy within the extreme subgroups (High Ability and Low
Ability subgroups) but not within the moderate (Middle
Ability subgroup) ability subgroup.

On the other hand, within subgroups of males varying
in level of "perceived" (Rated) difficulty of school success,-
significant relationships (with DAT-V scores held constant
within subgroups) for GSCI scores with both indices of
academic achievement were obtained only within the moderate
(Intermediate Difficulty) difficulty subgroup but ppp within
the extreme subgroups (Low Difficulty or High Difficulty

subgroups).

226

The above findings suggest that GSCI scores can be
effectively used (although the effectiveness is extremely
low in all cases) for the prediction of academic achieve-
ment with some groups of males and not with others. Further-
more, it appears that relationships for groups formed on the
basis of subjective estimates of difficulty correspond with
predictions from theory and assuming that the subgroups of
students represent students faced with tasks of differing
levels of "perceived" difficulty. However, when-subgroups
of males were formed on the basis of varying levels of
objective (academic ability) estimates of success in school,
it is only those males most and least likely to obtain
success for whom GSCI scores significantly correlate with
level of academic achievement.

The above findings for subgroups of males formed on
the basis of academic ability might be explained on the
basis of previous research findings concerning the influence
of success and failure on level of aspiration. Atkinson22
states that one effect of success is to raise the level of
expected future achievement while failure tends to lower the
level of expected future achievement.

If students of "High Ability" can be assumed to have
been most likely to experience success in the past while

students of "Low Ability" can be assumed to have been most

 

22John W. Atkinson, An Introduction to Motivationv

(D. Van Nostrand and Co., Inc., 1964).

 

227

likely to experience failure in the past, then it is
reasonable that the "perceived" level of difficulty may have
been shifted for these respective groups such that each
group may contain students who "perceive" success within an
"Intermediate" level of difficulty. However, an alternative
explanation for significant relationships between GSCI~
scores and teacher-assigned grades within the "High Ability"
and "Low Ability" subgroups, but not within the "Middle.
Ability" subgroup, could lie outside theory and rest on the
fact that development of the GSCI-was based on the responses
of "over—achievers" and "under-achievers" and to some

extent on the basis of responses of "high-achievers" and

"low-achievers."23

Discussion of Relationships for GSCI
Scores with Indices of Academic
Achievement for Subgroups of

Females Varying in Objective

or Subjective Estimates of

the Difficulty of School

SUCCESS'

 

 

 

 

 

 

Support was not found for the nomological validity of,
the GSCI as an interpretation of strength of the Motive for
Success for females, through failure to confirm the predicted
stronger relationship (with DAT—V scores controlled) for GSCI
scores with indices of academic achievement, using either.

objective (DAT-V scores) or subjective (Rated Difficulty)

 

23Farquhar, op. cit.

228

indices of success to locate females thought to be per-
forming tasks of an "Intermediate" level of difficulty, than
for females performing easier or more difficult tasks.

Failure to confirm the differentially stronger rela-
tionships, predicted by theory, with females was attributed
to the problems of assessment of the Motive for Success for
females (discussed in Stage I above) and the problem of
locating achievement-related behaviors for females (dis-
cussed in Stage II above).

However, GSCI_scores were found to correlate signifi-
cantly with indices of academic achievement for females
within some subgroups, formed on the basis of objective
or subjective estimates of school success, and not within
other subgroups.

Significant relationships between GSCI scores and
indices of academic achievement (with DAT-V scores controlled
within subgroups) were found within those subgroups of
females objectively most likely to obtain success (High and
Middle Ability subgroups) but not within the subgroup of
females objectively ipapp likely to obtain success (Low
Ability subgroup).

Apparently, whatever is reflected by GSCI scores for
females, requires that females possess at least a certain
minimum level of academic ability before it contributes to

prediction of academic achievement.

229

On the other hand, significant relationships between'
GSCI scores and indices of academic achievement (with DAT-V
scores controlled within subgroups) were found within the
extreme subgroups of females (High Difficulty and Low
Difficulty subgroups) formed on the basis of subjective
estimates of the difficulty of success in school, but not
within the subgroup of females assumed to be faced with

moderate level of difficulty (Intermediate Difficulty

 

subgroup).

The above findings based on subgroups of females
differing in level of the Rated Difficulty of school success
might be explained as the result of the biasing effects of
past success and failure on future expectations such that
the extreme subgroups of females may have each contained
females for whom the."perceived" difficulty of success was
within an "Intermediate" range of difficulty. However, the
preceding explanation was previously used to explain signif-
icant_re1ationships between GSCI scores and academic-
achievement among the extreme subgroups of males varying

in objective estimates (academic ability) of school

 

success. For females, recourse was taken in this eXplana-
tion for significant relationships among extreme subgroups

differing in subjective estimates of school success.

 

It would appear that objective and subjective esti-

mates of the difficulty of school success differently affect

230

relationships between GSCI scores and indices of academic-
achievement for and among males and females.

The failure to find female GSCI scores to predict as
an index of strength of the Motive for Success or achievement-
related behavior to influence level of academic achievement
for females as expected of an index of strength of the
tendency to achieve, offers little support for attempts to
interpret relationships between GSCI scores and academic
achievement on the basis of theory.

For females, the failure to find relationships as
predicted may be attributable to, (a) problems of the assess-
ment of strength of the Motive for Success, (b) problems of
the selection and assessment of criteria of strength of the
tendency (motivation) to achieve, (c) problems with indices
of "perceived" difficulty of school success or, (d)
homogeneity in the responses of females with respect to the

above three variables.

CHAPTER VI

SUMMARY, CONCLUSIONS AND IMPLICATIONS

In this chapter, the problem, theory and instru-
mentation, sample and precedure, design and analysis, and
findings for males and females are summarized. The con-
clusions of the study are presented. Finally, implications

for future research are offered.

Summary

The summary is presented in six sections: the problem,
the theory and instrumentation, the sample and procedure, the
design and analysis, the findings for males and the findings

for females.

The Problem

 

The problem of this study was to seek evidence con-
cerning the construct (nomological) validity of the Motive
for Success, (a) represented within Atkinson's theory of
achievement motivation, (b) as interpreted by scores on the
Generalized Situational Choice Inventory (GSCI), (c) as
expressed in theory-selected behavioral criteria of motiva-
tion within the classroom and, (d) as differentially
influenced (moderated) by the "apparent difficulty" of

success in school.

231

232

Theory and Instrumentation

 

Atkinson's theory of achievement motivation conceives
of behavior in competitive situations as the resultant inter-
action of personality characteristics (motives) and charac-
teristics of the situation (expectancy and incentive values).
A multiplicative relationship is postulated among the terms
of theory (Motive x Expectancy x Incentive) such that the
greater the strength of a personality disposition to
achieve (Motive for Success) the stronger the influence on
certain behavior in competitive situations. However, the
situational characteristics (expectancy and Incentive
values) are further thought to influence the "perceived"
difficulty of a task such that differences in strength of
the Motive for Success have a stronger influence in
behavior on tasks of "Intermediate" difficulty than on
easier or more difficult tasks.

Atkinson's statements concerning the influence of the
Motive for Success on strength of the tendency (motivation)
to achieve and Farquhar's definition of the role of motiva—
tion within the academic situation were considered suffi-
ciently explicit to allow selection of three criteria of
motivated behavior in the c1assroom--the initiating,
directing and sustaining of achievement-related activities.

Relationships between strength of the Motive for
Success, interpreted as scores on the Generalized Situa-

tional Choice Inventory (GSCI), and the three theory-selected

233

behavioral criteria of strength of the tendency (motivation)
to achieve in the classroom, were hypothesized to be nega-
tive in the case of time spent on a non-achievement task
(initiating) and positive in the case of both expected

level of achievement (directing) and time spent on an
achievement task (sustaining).

Hypotheses further specified stronger relationshipp

 

between GSCI scores and each of the three theroy-selected

. Imu.a-l'mnfl

behavioral criteria of motivation for students assumed, on

the basis of academic ability (DAT-V scores) or stated

'fli‘m

difficulty (Rated Difficulty), to be performing school
tasks of an "Intermediate" level of difficulty, than for
students assumed to be performing easier or more difficult

tasks.

Sample and Procedure

 

The classroom situation was entered and a total of
511 students in 28 ninth grade English classes within 3
schools were presented with both a non-achievement task
(Joke Rating) and an achievement task (Theme Writing).
Instructions emphasized the importance of the theme (grade
would count toward final course grade) and the "spare-
time" value of rating the jokes.

A time limit of 20 minutes was set. All students
were required to read and rate "some" of the jokes before

starting to write and, if they chose, to re-write (revise)

234

a theme concerning "the importance of doing well in English
class." The final copy of the theme was limited to a
length of 7 lines.

The three theory-selected behavioral criteria of
strength of the tendency (motivation) to achieve in the
classroom consisted of: time spent on the non-achievement
task (Joke Rating) before starting the achievement task
(Theme Writing)-—.Initiating activity, expected level of
achievement on the achievement task--Directing activity
and, time spent on the achievement task (Theme Writing)
before returning to the non-achievement task (Joke Rating)--

Sustaining activity.

Design and Analysis

 

The design of this study was that of a correlational
approach to construct validity through an investigation of
relationships between GSCI scores and the three theory-
selected criteria of motivation in the classroom. However,
a quasi-experimental approach was also incorporated through
the assumption that students varying in objective (DAT-V
scores) or subjective (Rated Difficulty) estimates of school
success perform school tasks of varying difficulty.

The correlational approach and the quasi-experimental
approach were also used to investigate the validity of the
three theory—selected behavioral criteria of motivation, as

reflecting strength of the tendency to achieve, with indices

 

235

of academic achievement which were also assumed to reflect
strength of the tendency to achieve.

Both the correlational approach and the quasi-
experimental approach were further used to investigate the
relative predictive validity and nomological (construct)
validity of the GSCI with indices of academic achievement
which were assumed to reflect strength of the tendency
to achieve.

Hypotheses were evaluated, level of rejection (alpha)
of the null hypotheses was set at the .05 level of confi-
dence, within each of three Stages of analysis: Stage I--
Nomological Validity, Stage II--Criterion Validity, and
Stage III--Predictive and Nomological Validity.

Within all three stages of analysis, directional
hypotheses were evaluated by the F-statistic (F = t2) for
coefficients of partial correlation (DAT-V scores held
constant). Differential hypotheses were evaluated by
coefficients of partial correlation (DAT-V scores held
constant), converted by Fisher's r to z transformation, and
the standard error of the difference between two "2" scores

referred to a table of normal probability.

236

Findings for Males

 

l.

The relationships between GSCI scores and the

initiating criterion of strength of the tendency

 

(motivation) to achieve in the classroom situation
were not significant either with or without
academic ability (DAT-V scores) controlled by
partial correlation technique.

A significant (negative) coefficient of simple.
correlation was obtained for academic ability
(DAT-V scores) with the initiating criterion,

r = -.19u (p < .01), of strength of the tendency
(motivation) to achieve in the classroom.
Significant coefficients of partial correlation
(DAT-V scores controlled) were found for GSCI

scores with the directing criterion, r = .214

 

(p < .01), and sustaining criterion, r .170

 

(p < .05), of strength of the tendency (motiva-

tion) to achieve in the classroom.

Relationships between academic ability (DAT-V

scores) and the directing and sustaining criteria

 

 

of strength of the tendency (motivation) to
achieve in the classroom were not significant.
A significant coefficient of simple correlation

was obtained for the initiating_criteria of

 

strength of the tendency (motivation) to achieve

in the classroom with one index of level of

237

academic achievement, r = -.l7l (p < .01), but
this relationship was not significant with the
influence of academic ability (DAT-V scores)
controlled by partial correlation technique.
Significant coefficients of partial correlation
(DAT-V scores controlled) were obtained for the

directing criterion of strength of the tendency

 

(motivation) to achieve in the classroom with
two indices of academic achievement, r = .271

(p < .01) and r = .301 (p < .01).

A significant coefficient of partial correlation
(DAT-V scores controlled) was obtained for the

sustaining criterion of strength of the tendency

 

(motivation)to achieve in the classroom with one
index of level of academic achievement, r = .161
(p < .05).

Significant coefficients of partial correlation
(DAT-V scores controlled) were obtained for GSCI
scores with two indices of academic achievement,
r = .299 (p < .01) and r = .302 (p < .01).

A significant coefficient of partial correlation
(DAT-V scores controlled) was obtained between
GSCI scores and one index of academic achievement
for males who rate success in school within an
"Intermediate" range of difficulty, r = .Ul?

(p < .01), which was also significantly greater

10.

ll.

12.

238

than coefficients of partial correlation (DAT-V
scores controlled) between GSCI scores and one
index of academic achievement for males who rate
success in school as easier, r = .100 (p > .05)
with z - z = 2.39 (p < .01), or more difficult,

l 2
r = .151 (p > .05) with 21 — 22 = 2.19 (p < .05).
Significant coefficients of partial correlation
(DAT-V scores controlled) between DAT-V scores

and the initiating criterion, among subgroups of

 

males, were obtained only for those males objec-
tively or subjectively l§a§£_likely to obtain
success, Middle Ability subgroup r = —.282

(p < .01), Low Ability subgroup, r = -.280

(p < .01), and High Difficulty subgroups, r =
-.309 (p < .01).

Significant coefficients of partial correlation
(DAT-V scores controlled) for GSCI scores with

the directing criterion, among subgroups of males,

 

were found only for those males objectively
(DAT-V scores) least likely to obtain success,
Middle Ability subgroup, r = .350 (p < .01) and
Low Ability subgroup, r = .291 (p < .01).
Significant coefficients of partial correlation
(DAT-V scores controlled) between GSCI scores

and the sustaining criterion, among subgroups of

 

males, were obtained only for those males

13.

l“.

239

objectively or subjectively most likely to obtain
success, High Ability subgroup, r = .309 (p < .01),
and Low Rated Difficulty subgroup, r = .288

(p < .01).

A significant coefficient of simple correlation

for the initiating criterion with one index of

 

academic achievement, among subgroups of males,
was obtained only for those males subjectively
least likely to obtain success, High Rated
Difficulty subgroup, r = -.290 (p < .01), but
this relationship was not significant with DAT-V
scores controlled by partial correlation.
Significant coefficients of partial correlation

(DAT-V scores controlled) for the directing

 

criterion with at least one of the two indices

of academic achievement, were obtained among all
subgroups of males varying in objective or sub-
jective estimates of school success; High Ability,
r = .298 (p < .01), Middle Ability, r = .283

(p < .01), Low Ability, r = .299 (p < .01), and
Low Rated Difficulty, r = .3AU (p < .01), Inter-
mediate Rated Difficulty, r = .235 (p < .05),

High Rated Difficulty, r = .231 (p < .05).

15.

l6.

17.

240

A significant coefficient of partial correlation

(DAT-V scores controlled) for the sustaining

 

criterion with one index of academic achievement,
among subgroups of males, was obtained only for
those males objectively most likely to obtain
success, High Ability subgroup, r = .263 (p < .05).
Significant coefficients of partial correlation
(DAT-V scores controlled) for GSCI scores with
both indices of academic achievement, among sub-
groups of males varying in objective estimates
(DAT-V scores) of school success, were obtained
only within the extreme subgroups, High Ability
subgroup, r = .426 (p < .01) and r = .437

(p < .01), and Low Ability subgroup, r = .249

(p < .05) and r = .249 (p < .05).

Significant coefficients of partial correlation
(DAT-V scores controlled) for GSCI scores with
both indices of academic achievement, among sub—
groups of males varying in subjective estimates
(Rated Difficulty) of school success, were obtained
only within the moderate subgroup, Intermediate
Rated Difficulty subgroup, r =-.4l7 (p < .01) and

r = .362 (p < .01).

24l

Findings for Females

18.

19.

20.

21.

The relationships between GSCI scores and the

initiating criterion of strength of the tendency

 

(motivation) to achieve within the classroom were
not significant either with or without academic-
ability (DAT-V scores) controlled by partial
correlation technique.

A significant coefficient of partial correlation
(DAT-V scores controlled) was found for GSCI

scores with the directing criterion of strength

 

of the tendency (motivation) to achieve within
the classroom, r = .265 (p < .01).
The relationships between GSCI scores and the

sustaining criterion of strength of the tendency

 

(motivation) to achieve within the classroom
were not significant either with or without
academic ability (DAT-V scores) controlled by
partial correlation technique.

Of the three criteria of strength of the
tendency (motivation) to achieve in the class-
room, significant coefficients of partial
correlation (DAT-V scores controlled) were

obtained only for the directing criterion with

 

indices of academic achievement, r = .230

(p < .01) and r = .354 (p < .01).

22.

23.

24.

25.

242

Significant coefficients of partial correlation
(DAT-V scores controlled) were obtained for GSCI
scores with indices of academic achievement,

r = .285 (p < .01) and r = .289 (p < .01).
Coefficients of partial correlation (DAT-V scores
controlled) that were significant and signifi-
cantly greater where predicted by theory were

not found between any of the variables among
subgroups of females varying in objective or

subjective estimates of school success.‘

 

Relationships between GSCI scores and the

initiating criterion were not found to be

 

significant within any of the subgroups of
females varying in objective or subjective
estimates of school success.

Significant coefficients of partial correlation
(DAT-V scores controlled) for GSCI scores with

the directing criterion were found only for-

 

females objectively most likely to obtain success,
High Ability, r‘= .419 (p < .01) and Middle,
Ability, r = .263 (p < .Ol).and for females sub-
jectively most and least likely to obtain success,
Low Rated Difficulty, r = .226 (p < .05), and

High Rated Difficulty, r = .221 (p < .05).

26.

27.

28.

243

A significant coefficient of partial correlation
(DAT-V scores controlled) for GSCI scores with

the sustaining criterion, among subgroups of

 

females varying in objective or subjective esti-
mates of school success, was obtained only for
those females subjectively least likely to
obtain success, High Rated Difficulty, r = .266
(p < .05).

Significant coefficients of partial correlation
(DAT-V scores controlled) for the three criteria
of strength of the tendency (motivation) to
achieve in the classroom with indices of level of
academic achievement, among subgroups of females
varying in objective or subjective estimates of

school success, were found only with the directing

 

criterion and these relationships were significant
within all subgroups except for those females who
subjectively rate success as least difficult.
Significant coefficients of partial correlation
(DAT-V scores controlled) for GSCI scores with
either index of academic achievement, among sub-
groups of females varying in level of objective
estimates of school success, were found only for
those females objectively most likely to obtain
success, High Ability, r = .277 (p < .01) and

Middle Ability, r = .445 (p < .01).

244

29. Significant coefficients of partial correlation
(DAT—V scores controlled) for GSCI scores with
either index of academic achievement, among sub-
groups of females varying in level of subjective
estimates of school success, were found only for
those females subjectively asst and least likely
to obtain success, Low Rated Difficulty, r =,
.274 (p < .01) and High Rated Difficulty, r =

 

.368 (p < .01).

Conclusions

E

 

The following conclusions are based on the findings of
the study.
1. Nomological validity was demonstrated for the GSCI
as an interpretation of strength of the Motive for
Success for males through relationships with the

directing and sustaining criteria of strength-of

 

 

the tendency (motivation) to achieve within the
classroom.

2. Clarity for the nomological validity interpretation
of relationships between GSCI test scores and the

directing and sustaining criteria was provided by

 

 

the failure of academic ability to influence either
of these criteria.

3. Nomological validity was demonstrated for the
GSCI as an interpretation of strength of the Motive

for Success for males through relationships with

245

indices of academic achievement as one criterion
for strength of the tendency (motivation) to
achieve in the classroom.

Nomological validity was demonstrated for the GSCI
as an interpretation of strength of the Motive

for Success for males, using subjective estimates
of "perceived" difficulty of school success to
differentially influence (moderate) relationships
between GSCI test scores and indices of academic

achievement as predicted by theory.

IIH‘¢. ' ’05". 'U .-‘ Ll. "_ ‘-

 

Nomological validity was weakly demonstrated for
the GSCI as an interpretation-of strength of the
Motive for Success for females through relation-

ships with only the directing criteria of

 

strength of the tendency (motivation) to achieve
in the classroom.
The behavior of males during performance of a non-

achievement task (initiating activity) offered

 

support for the operation of an "Inertial Tendency"
which was aroused by expectations of failure on the
achievement task and indexed (negatively) by
academic ability test scores.

With both an achievement task and a non-achievement
task available within the classroom, males tend to
eXpress differences in the strength of a tendency

to achieve success during performance of an

10.

11.

246

achievement task and express differences in the
strength of a tendency to avoid failure during
performance of a non-achievement task.

For males differing in the level of objective or
subjective estimates of success in school, differ-
ences in strength of the Motive for Success

influence behavior (sustainingactivity) on an,

 

achievement task only for those males asst likely
to obtain success in school.

For males differing in the level of objective or
subjective estimates of success in school, dif-
ferences in expectations of failure influence

behavior (initiating activity) on a non-achievement

 

task only for those males least likely to obtain
success in school.

For males differing in the level of objective
estimates of success in school, differences in
strength of a tendency to achieve success, expressed

in behavior (sustaining activity) on an achievement

 

task, influence the academic achievement on only
those males mast likely to obtain success.

For males differing in the level of subjective
estimates of success in school, differences in
strength of a tendency to avoid failure, expressed

in behavior (initiating activity) on a non—,

 

achievement task, influence (negatively) the

l2.

l3.

14.

15.

247

academic achievement of only those males least
likely to obtain success.

Expectations concerned with the level of future
achievement in school (directing activity) are a
fairly uniform and consistent correlate of later
academic achievement within nearly all levels of
objective and subjective estimates of school
success for both males and females.

For males differing in the level of objective
estimates of school success, differences in
strength of the Motive for Success influence the
academic achievement of those males mast and least
likely to obtain success.

For males differing in the level of subjective
estimates of school success, differences in
strength of the Motive for Success influence the
academic achievement of only those males

moderately likely to obtain success.

 

For females differing in the level of objective~
estimates of school success, differences inr
strength of the Motive for Success influence the
academic achievement of those females mast and

moderately likely to obtain success.

 

l6.

17.

18.

19.

248

For females differing in the level of subjective
estimates of school success, differences in
strength of the Motive for Success influence the
academic achievement of those females mast and
least likely to obtain success.

Differences in strength of the Motive for Success
may influence different behaviors in the class-

room for females than the initiating and sustaining

 

of task behaviors found to be influenced by them
Motive for Success for males.

There may be a basic difference in the way males
and females respond to a task perceived to be
"difficult," such that males spend increasing
amounts of time on non-achievement activities
possibly in an attempt to avoid failure, while
females devote increasing amounts of time to per-
formance of a "difficult" task possibly in an
attempt to satisfy needs to achieve.

GSCI test scores appear to predict academic achieve-
ment equally well for males and females. Similarly
the GSCI test scores are about as efficient as
academic ability (DAT-V) test scores for both males

and females in predicting academic achievement.

 

20.

249

Significant relationships between GSCI scores and
indices of academic achievement, with academic,
ability controlled by partial correlation tech-
nique, for both males and females, tends to
support the findings of previous research that
predictions based on both measures may provide
greater accuracy than predictions based on either

measure alone.

Implications

 

The following implications are based on the conclusions

of the study.

1.

Future research should be directed toward refine-

ment of the techniques for quantifying initiating

 

and sustaining dimensions of the behavior of males

 

in the classroom.

Researchers, concerned with the construction or
refinement of techniques for assessment of strength
of the Motive for Success among males, may obtain
increased_clarity for the constructs indexed by
their instruments through investigation of rela-

tionships with initiating and sustaining criteria

 

 

of motivation.
Researchers, concerned with manipulating situational
variables thought to hold motivational_implications

for males in the classroom, may be able to observe

 

250

the effects of these variables in initiating and

sustaining achievement-related behavior.

 

The use of indices of one situational variable,
i.e. of the "perceived" difficulty of school
success, to influence relationships between per-
sonality and achievement in school, may allow
increased efficiency in predicting academic
achievement for some groups of males, decreased
efficiency in predicting achievement for other
groups, and provide a basis in theory to forecast
the membership of each group.

Further attention should be given to the use of
measures of academic ability, to moderate rela-
tionships between personality test scores and
indices of academic ability, in an attempt to
empirically select groups of males and females
for whom personality test scores hold varying
levels of predictive efficiency.

It may be necessary for future research, which
attempts to evaluate the influence of achievement-
related motives on behavior in the classroom, to
specify or control alternative activities
available to males.

Future research, which is based on a theory of
achievement motivation that posits both approach

and avoidant motives (personality characteristics)

 

251

and approach and avoidant tendencies (behavioral
characteristicsL may need to give attention to
both the separate assessment of personality
characteristics and the analysis of behavior on
separate tasks.

Because there may be special problems in assessing
strength of the Motive for Success among females
age in observing the effects of that motive in the
behavior of females, research concerned with the
competitive behavior of students in the class-
room should analyze results separately for males
and females.

A crucial problem for investigations concerned
with the motivation of females in the classroom

is that of identifying achievement-related
behaviors for females. Achievement-related
behaviors for females may differ from those for
males and may need to be considered within the
total context of the classroom rather than be,
confined to performance measures on iSolated

tasks.

“I!

 

-".' (v -2 ."'L

[mm

BIBLIOGRAPHY

American Psychological Association, Committee on Psycho-
logical Tests. Technical Recommendations for
Psychological Tests and Diagnostic Techniques.
Washington, D.C.: APA, 1954.

 

 

Atkinson, J. W. "Motivational Determinants of Risk-
Taking Behavior." Psychological Review, 64
(1957), 359-372.

 

, ed. Motives in Fantasy Action and Society.
D. Van Nostrand Co., Inc., 1958.

 

. An Introductiongto Motivation. D. Van
Nostrand and Co., Inc., 1964.

Atkinson, John W.; Bastian, Jarvin R.; Earl, Robert W.;
and Litwin, George H. "The Achievement Motive,
Goal Setting, and Probability Preferences."~
Journal of Abnormal and Social Psychology, 60
(1960), 27-36.

Atkinson, J. W., and Cartwright, D. "SOme Neglected
Variables in Contemporary Conceptions of Decision
and Performance." Psychological Reports, 14 (1964),
575-590.

 

Atkinson, John W., and Feather, Norman T., eds. A Theory
of Achievement Motivation. New York: Wiley and
Sons., Inc., 1966.

Atkinson, John W., and Litwin, G. H. "Achievement Motive
and Test Anxiety Conceived as Motive to Approach
Success and Avoid Failure." Journal of Abnormal
and Social Psychology, 60 (1960), 52-63.

 

Atkinson, John W., and Reitman, W. R. "Performance as a
Function of Motive Strength and Expectancy of Goal
Attainment." Journal of Abnormal and Social
Psychology, 534(1956), 361-366.

252

253

Berdie, Ralph F. "Intra-Individual Variability and
Predictablity." Educational and Psychological
Measurement, 21 (1961), 663-676.

 

 

Brown, William F.; Abeles, Norman; and Iscoe, Ira.
"Motivational Differences Between High and Low
Scholarship College Students." The Journal of
Educational Psychology, 45 (1954), 215-223.

 

 

Brunner, Jerome. Toward a Theory of Instruction.
Cambridge, Massachusetts: Harvard University
Press, 1966.

 

Campbell, D. T. "Recommendations for APA Test Standards
Regarding Construct, Trait or Discriminant Validity."
American Psychologist, Vol. 15, No. 8, August 1960,
546-553.

 

Cotton, J. W.- "Running Time as a Function of Amount of
Food Deprivation." Journal of Experimental
Psychology, 46 (1953), 188:198.

 

 

Cronbach, Lee J., and Meehl, Paul E. "Construct Validity
in Psychological Tests." Ps cholo ical Bulletin,
Vol. 52, No° 4, July, 1955, 281-302.

 

Dunnette, Marvin D. "A Modified Model for Test Vali-
dation and Selection Research." Journal of
Applied Psychology, 47 (1963), 317-323.

Eysenck, H. J. The Structure of Human Personality.
London: Methuen, 1953.

 

Farquhar, William W. Motivation Factors Related to
Academic Achievement. Cooperative Research
Project 846° East Lansing, Michigan: Michigan
State University, Office of Research and
Publication, 1963.

Feather, Norman-T. "Subjective Probability and Decision
under Uncertainty." Psychological Review, 66
(1959), 150-164.

 

"The Relationship of Persistence at a Task to
Expectation of Success and Achievement Related:
Motives."' Journal of Abnormal and Social Psychology,
63 (1961), 552-561.

 

 

254

Feather, Norman T. "The Study of Persistence."
Psychological Bulletin, 59 (1962), 94-115.

"Persistence at a Difficult Task with Alternative
Task of Intermediate Difficulty." Journal of
Abnormal and Social Psychology, 66 (1963), 604-609.

 

 

French, Elizabeth G. "Some Characteristics of Achieve-
ment Motivation." Journal of Experimental
Psychology, 50 (1955), 232-236.

Ghiselli, Edwin E. "Moderating Effects and Differential
Reliability and Validity." Journal of Applied
Psychology, 47 (1963), 81-86.

 

 

 

Goodstein, Leonard D., and Heilbrun, Alfred B., Jr.
Prediction of College Achievement from the Edwards
Personal Preference Schedule at Three Levels of
Intellectual Ability." Journal of Applied
Psychology, 46 (1962), 317-320.

 

 

Green, Robert L. "The Predictive Efficiency and
Factored Dimensions of the Michigan State M-Scales
for Eleventh Grade Negro Students, An Exploratory
Study." Unpublished doctoral dissertation,
Michigan State University, 1962.

Hakel, Milton D. "Prediction of College Achievement
from the Edwards Personal Preference Schedule Using
Intellectual Ability as a Moderator." Journal of
Applied Psychology, 50 (1966), 336-340.

 

Hartshorne, M.; May, M. A.; and Maller, J. B., Studies in
Service and Self-Control. New York: Macmillan
Co., 1929.

 

Heilbrun, Alfred B. Jr. "Personality Factors in College
Dropout." Journal of Applied Psychology, 49 (1965),
1-7 0

Hills, John R. "The Measurement of Levels of Aspiration.".
Journal of Social Psychology, 41 (1955), 221-229.

Hoffmann, Louis J. "An Application of the Multitrait-
Multimethod Matrix to the Study-of the N-Achievement
Construct." Unpublished doctoral dissertation,
Michigan State University, East Lansing, 1965.

 

255

Howells, T. H. "An Experimental Study of Persistence."
Journal of Abnormal and Social Psychology, XXVIII
(1933), 14-29.

 

Irwin, Francis W. "Motivation and Performance." Annual
Review of Psychology, Vol. 12. Palo Alto: Annual
Reviews, Inc., 1961.

 

Izard, Carroll E. "Personality Characteristics (EPPS),
Level of Expectation, and Performance." Journal
of Counseling Psychology, 26 (1926), 394.

 

Kimble, Gregory A. Conditioning and Learning.
New York: Appleton-Century—Crofts, Inc., 1961.

 

Lewin, K. The Conceptual Representation and the
Measurement of Psychological Forces. Durham,
North Carolina: Duke University Press.

 

 

MacArthur, R. S. "An Experimental Investigation of
Persistence in Secondary School Boys." Canadian
Journal of Psychology, 9 (1955), 42-54.

 

Mandler, G., and Sarason, S. B. "A Study of Anxiety
and Learning." Journal of Abnormal and Social
Psychology, 47 (1952), 166=173.

 

 

McClelland, D., and Atkinson, J., et al. The Achievement

 

Motive. New York: Appleton-Century Crofts, 1953.

Payne, David A. "The Concurrent and Predictive Validity
of an Objective Measure of Academic Self-Concept."
Educational and Psychological Measurement, 22

(1962), 773-780.

 

Porter, Lyman W. "Personnel Management." Annual Review
of Psychology, 17 (1966), 295-422.

 

 

Pottharst, B. C. "The Achievement Motive and Level of
Aspiration after Experimentally Induced Success and
Failure." Unpublished Doctoral thesis, University
of Michigan, Ann Arbor, 1955.

Sarason, I. G. Contemporary Research in Personality.
Princeton, New Jersey: D. Van Nostrand Co., Inc.,
1962.

 

Saunders, David R. "Moderator Variables in Prediction."
Educational and Psychological Measurement, 16
(1956), 209-2220

 

256

Siegel, Sidney. "Level of Aspiration and Decision
Making." Psychological Review, 64 (1957), 253-262.

Smith, Charles P. "Relationships Between Achievement-
Related Motives and Intelligence, Performance
Level, and Persistence." Journal of Abnormal and
Social Psychology, 68 (1964), 523-533.

 

Spielberger, C. D. "The Effects of Manifext Anxiety on the
Academic Achievement of College Students." Mental
Hygiene, 46 (1962), 420-426.

Spielberger, C. D., and Katzenmeyer, W. C. "Manifest
Anxiety, Intelligence, and College Grades."
Journal of Consulting Psychology, 23 (1959), 278.

 

Sterling, Virgil B. "A Pilot Factor Analytic Study of
Academic Motivation and Achievement Levels in
Eleventh Grade Male Students." Unpublished doctoral
dissertation, Michigan State University, East
Lansing, Michigan 1962.

Taylor, Janet A. "The Relationship of Anxiety to the
Conditioned Eyelid Response." Journal of
Experimental Psychology, 41 (1951), 81-92.

 

 

Thornton, G. R. "The Use of Tests of Persistence in
the Prediction of Scholastic Achievement." Journal
of Educational Psychology, 32 (1941), 266-273.

 

Thorndike, Robert L. The Concepts of Over- and Under-
achievement. New York: Teachers College,-
Columbia University, Bureau of Publications, 1963.

 

 

Thorpe, Marion D. "The Factored Dimensions of an Objective
Inventory of Academic Motivation Based on Eleventh
Grade Male Over- and Under-achievers." Unpublished
doctoral dissertation, Michigan State University, 1961.

Tiebout, H. M. "The Misnamed Lazy Student." Educational
Record, 24 (1943), 113-129.

 

Tolman, E. C. "Purposive Behavior in Animals and Men.
New York: Century Co. By permission of the University
of California Press. 1932.

Wallace, S. R. "Criteria for What." Presented at
American Psychological Association Meeting, Los
Angeles, 19640

257

Webster's Collegiate Dictionary. Springfield, Massachu-
setts: G. and C. Merriam Co., 1944.

 

Weiner, Bernard. "The Effects of Unsatisfied Achievement
Motivation on Persistence and Subsequent Performance."
Journal of Personality, 33 (1965), 428-442.

 

Woodworth, R. S. Psychology. New York: Henry Holt, 1940.

 

 

 

S
E
C
I
D
N
E
Di
P
A

258

APPENDIX A

JOKES (SAMPLE)

TEACHER: "Arthur, why does a moth eat holes in rugs?"
ARTHUR: "Maybe it wants to see the floor show."

LIKE DISLIKE A

 

TEACHER: "If you had ten potatoes and had to divide them
equally among twelve peOple, how would you do it?" a
MARGIE: "I'd mash them."

‘1‘" .V‘...‘ ' "~ b

 

Mimi“. .

LIKE DISLIKE

 

GYM TEACHER:
HOMER:
GYM TEACHER:

HOMER:

"You, there, mark time."

"With my feet, sir?"

"Have you ever known anything to mark time
with its hands?"

"Yes, sir, a clock."

LIKE DISLIKE

 

SWIMMING INSTRUCTOR: "And another reason for practicing

your swimming is that swimming is
good for the figure."

VOICE FROM BACK OF
THE ROOM: "Did you ever see a duck?"

LIKE DISLIKE

 

259

APPENDIX B

PARAGRAPH — ESSAY (Male)

You are to write a paragraph about a "make believe" student
named Jim. Maybe he is like someone you know. Jim thinks
it is important to try to do his best in English class. You
are to describe a believable past, present, and future for
Jim by answering the following questions in your paragraph: 5

 

 

1. What could have happened in the past to make Jim think
it is important to try to do his best in English class?

While Jim is in English class, what does he do?

What could happen to Jim in the future because he thinks
it is important to try to do his best in English class?

 

 

You are to write your paragraph on the 7 lines below. You
are to start where it says "Start" and you are to finish
anywhere on the seventh line. There are extra pages for
your use if you make a mistake or if you want to revise your
paragraph. Do all of your work on these pages. You are to
answer all three questions in ens paragraph of 7 lines.

 

Your teacher will grade your final paragraph, so use correct
grammar, spelling and punctuation. State your ideas clearly.
Be sure your paper is neat. Do all your writing in ink.
Remember, your paragraph is not to be longer or shorter than
7 lines and you are to answer all three questions in that one
paragraph.

When you start on this page, copy the number displayed in
the front of the room.

Your Name (Write)

 

Last First

START - NUMBER ( )

 

 

 

 

 

 

.STOP - NUMBER ( )

 

When you stop on this page, c0py the number displayed in
the front of the room.

260

APPENDIX C

ACHIEVEMENT QUESTIONNAIRE

Name (Print)

 

(Last) (First) (Middle)
F”

Your answers to the following questions will-not be seen by your teacher.

Circle the grade below which shows the final grade you received for your

 

 

 

 

 

 

1.
last semester of English.
A A- B+ B B— c+ c C- 0+ D D- E a
2. Circle the grade below which shows the final grade you really expect to 1
get for this semester of English. I
A A- B+ B B- C+ C C- D+ D D- E .
3. Circle a number below to show how difficult it is for you to do well in
your present English class.
,1 2 3,4 5 6,78 9,101112,131415,161718,
I T T I 7 T F
Very Easy Slightly Slightly Difficult Very
Easy ' Easy Difficult - Difficult
4. Circle the grade below which shows the last grade you received for a
writing assignment in this English class.
A A- B+ B B- C+ C C- D+ D D- E
5. Circle the grade below which shows the grade you really expect to get on
today's paragraph. -
A A- B+ B B- C+ C C- D+ D D— E
6. Circle the number below which shows how many minutes you expect it

will take you to complete today's paragraph.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

2651

APPENDIX D

INSTRUCTIONS

Each of you should have an envelOpe appropriately
marked Male or Female. You are not to open these until I
tell you to do so. Inside each envelope are several

packets of large mimeographed pages, a small packet labeled

— mvm is H“ A. '5'
'-

Joke Book, and a ball point pen. When I tell you to do so,

 

 

you are to take out see of the large packets of mimeographed
pages, the Joke Book and the ball point pen. Do not look
inside the Joke Book until I tell you to do so. Now, you
may open the envelope and take out those three things and
set the envelOpe aside. Remember, do not look in the Joke
Book until I tell you to do so.

First, look through the large mimeographed packet.
It should contain a first page asking for your name at the
tOp. Next there should be 3 pages (all the same) headed
Paragraph Essay Male or Female. The fifth sheet should be
headed Joke Rating Sheet. Is there anyone who does not
have at least these 5 pages? OK, turn back to the first
page and I will eXplain something about what we are going
to do today.

Today, you are to write a one paragraph essay for your
teacher. Your teacher will grade your paragraph as part of

your assignment for this class. I have some jokes for you

262

263

to read during your spare time before and after writing

 

your paragraph. You are to indicate whether you like or

dislike each joke you read so that I can learn something

about your preferences for reading styles (the jokes)

and writing styles (your paragraph). I am not interested

in your sense of humor but these materials are constructed

such that they can serve my purpose and may be enjoyable
for you.
First, print your name on the tOp line of the first
page of this large mimeographed packet, last name first.
Your answers to the following questions will not be
seen by your teacher. Your paragraph will be cut out of
these pages and only the paragraph given to your teacher.

1. Circle the grade below which shows the final grade you
received for your last semester of English. (Final
card marking 9 B or 8 A English).

2. Circle the grade below which shows the final grade you
really expect to get for this semester of English. (Final
card marking for this class).

3. Circle a number below to show how difficult it is for
you to do well in your present English class.

Now, before we finish the rest of the questions on this page,

turn to the second page so that I can eXplain your writing

assignment.
I will read from the boys' form, the girls' form is

different only in the use of the name Jane instead of Jim.

 

264

Maybe he is like someone you know. Jim thinks it is impor-
tant to try to do his best in English class. You are to
describe a believable past, present and future for Jim by
answering the following questions;
1. What could have happened in the past to make Jim think
it is important to try to do his best in English class?
2. While Jim is in English class, what does he do? 1
3. What could happen to Jim in the future because he
thinks it is important to try to do his best in English

class?

 

You are to write your paragraph on the 7 lines below. You
are to start where it says Start and you are to finish any-
where on the seventh line. There are extra pages for your
use if you make a mistake or if you want to revise your
paragraph. Do all your work on these pages. You are to
answer all three questions in one paragraph of 7 lines.

Your teacher will grade your final paragraph, so use
correct grammar, spelling and punctuation. State your ideas
clearly. Be sure your paper is neat. Do all your writing
in ink. Remember, your paragraph is not to be longer or
shorter than 7 lines, and you are to answer all three ques-
tions in that one paragraph.

Below these printed instructions is a place for your
name and seven lines for your paragraph. At the beginning
and end of these 7 lines there are spaces for numbers. Let

me explain about these numbers. Throughout the time allowed

265

for the paragraph writing I will be turning these pages up
here with numbers on them. These are code numbers and are
not in any order. When you start to write your paragraph,
you are to look up here and copy down whatever number I
have showing, putting it in the brackets marked Number -
Start. Then, again when you finish writing on this page,
look up here and copy down the number you see and put that
number in the brackets at the end of the seventh line.
If you use more than one page (because you made a mistake
or because you are trying to improve your paragraph) copy
the number I have showing when you start and stOp writing
on each page. Put a circle around the spaces for these
numbers now, to help remind you to look up for the number
later. You will not receive credit for this assignment
unless these numbers are prOperly filled in. There will
be a full 20 minutes available for writing the paragraph.
This will allow you to spend some of that time reading
jokes both before and after writing your paragraph. Are
there any questions about the writing assignment or the
placing of numbers when you start and stop writing on a page.
Alright, now that you know your writing assignment, turn
back to the first page and we will complete the questions
there.
4. Circle the grade below which shows the last grade you
received for a writing assignment in this English class.
(Your most recent grade for any writing assignment in

this class.)

 

266

5. Circle the grade below which shows the grade you really
expect to get on today's paragraph.

6. Circle the number below which shows how many minutes
you expect to get on today's paragraph. (Try to be as
accurate as possible.)

Now, turn to the last page of these large sheets, to the

one headed Joke Rating Sheet. Do not look inside the Joke

Book but c0py the number which appears on the front of your

L- anai: !- .:.Do‘i\".2 .1 :

Joke Book, on the proper line at the tOp of the rating

sheet. Next, put your name on the Rating Sheet. (Print or

 

write, your choice.)

The jokes I have brought are to be read in your EREES
tlae both before and after writing your paragraph. You are
to start by reading some of the jokes, how many is up to
you. You are to check Like or Dislike on the Rating Sheet
for each joke you read. Do not make any marks in the Joke
Book. When you decide you must start on your paragraph,
draw a line under the last Like or Dislike rating you have
checked on your Rating Sheet. (Do it like this.) Then turn
your Joke Book over and do not read any more of the jokes
until you are completely finished with your paragraph
writing assignment. Then you may return to reading and
rating the jokes.

You are to start, when I give you the signal, by
reading and rating some of the jokes before you write.

Remember to COpy down the numbers when you start and stop

267

writing on each page. Everyone look at the clock - I will
collect the jokes and your paragraphs in 20 minutes. OK,
start with joke number 1 in your Joke Book and check like
or dislike on your rating sheet.

Time is up. StOp whatever you are doing. Check to
see that your name is written above your paragraph. If
you used more than one paragraph page, place an X above
your name to indicate the see I should give to your
teacher.

OK, carefully put the Joke Book and the pen back in
the envelOpe. Do not put your large mimeographed packet
in the envelope. Pass these mimeographed packets forward.

Thank you for your cooperation.

 

268

 

<mw m>fipmH3530

compo nmfiawcm pmpmwsom pcmmopm
compo cause

AmCHcfimumzmv cede mcflpfinz mamas

chfipomnfinv compo ompoodxm.

chaeseeHch mane museum mxoe
seaseficcea steam

r-‘INm-filnmNOOON

 

 

 

>ne<a

Home "meanmfits>.
Asa u zv

H.m m.: =.m m.m m.m H.m e.m m.aa H.m m.m m.e o.mH m.H m.fia m.o e.mH -.m m.zm seasoeccfia swam
_ “em I zv

M.m m.e m.m m.s o.m o.m e.m a.sa m.H A.» m.- o.mH o.H e.m H.e p.3H m.e =.mm .ccfia deductesmecH
. ism u zv

m.m H.w e.m e.o m.m m.» m.oa m.mH o.m :.m m.» -.mH o.H m.e m.m m.mfi. =.o m.mm suasoacmao seq
A-m u 2V

o.m m.: =.m pm.: m.m m.m m.m o.ma H.m e.e m.e. m.mH o.m o.m H.m =.m m.o m.=m sefiadn< 30a
flew a 2v

m4 5.: m.m mt: o.m 5m ed TS o.m. 3.» mg m6; m.m Tm o; T? To m.mm 33d: 33::
. flaw u.zv

m.m N.s m.m m.e m.m m.e o.m o.mH m.m m.e e.w H.mH m.m m.» m.m e.mm w.s H.mm sesafie< swam
. Ammm.u zv

~.m m.m m.m m.m H.m H.e m.m =.mH H.m a.» a.» m.sH m.m e.m m.e m.eH m.s o.mm mafia: Hausa
mm m mm m mm m cm N am .m mm m mm N am m mm m eeouwhsm

o m m m H
awoanmfihm>

 

 

mmq<2 doomom mmHm mOHZDh mmm mom mmqdem<> mme mo

m XHDzmmm<

mZOHB<H>mQ Qm<QZ¢Bm Dz¢ mz¢m2

2639

<mo c>apmHSEso
compo cmfiflwcm ucpmcEcm pccmchm
. . compo cacca

.AwCacfimpmsmv.cEaB madcap: cEan

Amchocsfiav compo ocuocqu
AwsfipmfieHch cane msfiesm cxoe
seHsOHccHo ocemm

HNMfimKONQO

 

 

 

>ue<o

Home "mcflnmfitd>.

. gem u zv .

m.m o.m m.m m.m m.m m.o m.m H.mH o.m m.e o.m o.ma o.m m.o~ s.m m.mH m.m m.mm seHsOAccHo zoo
.. “mm u 21

fl.m m.m m.m A.e m.m m.c m.m a.sfl m.H m.m m.m H.NH N.H m.m e.m H.:H m.e m.Hm .ccfid cesfiecEAcesH
. Ase u 21

m.m H.m m.m m.m m.m c.m m.m e.eH m.H m.m e.m o.mH N.H m.m m.m m.mH m.m m.mm seflseHccao sod
, flow a 2o

~.m c.m m.m m.m m.m m.m m.m m.sfi m.m o.m H.m e.mfi s.m 0.5 m.H. s.m m.m o.Hm spfifluem sou
Aom n 2o

m.m m.c m.m m.o c.m H.- H.@ m.-H o.m H.m o.m A.HH m.m m.e o.m m.ma e.m s.om seaafiee cased:
. flew a z-

m.m N.m mew m.m m.~ ~.m a.m m.sH o.m s.m m.m o.NH m.m m.m m.c ~.em e.e e.mm senaaes swam
Ammm u 2o

e.m m.m o.m 3.5 m.m A.» H.m m.eH H.m m.m s.m s.mH m.m m.m m.e m.mH e.m m.Hm Atacama Hesse

mm m an m mm m mm m mm N mm N mm m mm N mm m mescnssw
, c .m e m m H
soc-omfium>

 

WMQ<zmm doomom mon mOHZDw mmm mom mmqm¢Hm¢> mmB mo monB¢H>mQ Qm¢Qz<Bm Qz¢

m XHszmm<

mz¢mz

 

"I1111111111111ES