THE RELATIONSHIP BETWEEN COLLEGE ACADEMIC
PERFORMANCE AND EXPECTANCIES (Abstract)

William Edward Clarke

Two questionnaires were developed by the investigator
for the purpose of assessing the academic expectancies held
for selected freshman, male college students by certain in-
dividuals and groups of individuals assumed to be significant
others for the students. A ten item, multiple-choice Student
Questionnaire was designed to determine a student's perception
of the academic expectancies held for him by members of his
home and home community group; and, a Peer Group Questionnaire
was designed as a rating form upon which a rating of the
academic motivations of a student's university peer group
could be made.

Complete questionnaire data were available for 836
freshman residents of men's residence halls at Michigan State
University during the 1957—1958 school year. The total sample
was composed of 369 non-probationary students whose grade-
point-averages during their freshman year were consistently
2.00 or higher, 340 probationary students whose grade-point-
averages were consistently below 2.00, and 127 raisers whose
grade-point-averages for the fall term were below 2.00 but
whose cumulative grade-point-averages for the full year were
2.00 or higher.

The Student Questionnaires were distributed by residence
hall assistants during the spring term and each student

completed the questionnaire privately. The separate items on

William Edward Clarke

the questionnaire dealt with the student's perception of the
attitudes and feelings of his parents, high school friends,
high school staff members, relatives, and neighbors regarding
his going to college and success at college. The residence
hall assistants completed the Peer Group Questionnaire by
rating a given student's university peer group as either
highly academically motivated or poorly academically motivated.

Normative comparisons were made between the three
groups of students with respect to their responses to the
Student Questionnaire and with respect to the ratings of their
peer groups made by the residence hall assistants.

Analysis of the data revealed that probationers and non-
probationers were significantly differentiated with respect to
each of the ten items on the Student Questionnaire as well as
the total expectancy score derived from the Student Question-
naire. Probationers and raisers were significantly differen-
tiated with respect to the total expectancy score, but only
three of the ten Student Questionnaire items were found to dif-
ferentiate these groups significantly. The ratings of the
students' peer grOups which were made by the residence hall
assistants significantly differentiated probationers from non-
probationers and probationers from raisers. Throughout the
study, however, the observed differences between probationers
and non-probationers were systematically greater than the

differences between probationers and raisers.

THE RELATIONSHIP BETWEEN COLLEGE ACADEMIC
PERFORMANCE AND EXPECTANCIES

By
WILLIAM EDWARD CLARKE

A THESIS
Submitted to the School for Advanced Graduate Studies of
Michigan State University of Agriculture and
Applied Science in partial fulfillment of
the requirements for the degree of

DOCTOR OF EDUCATION

Department of Administrative and Educational Services

Guidance and Counselor Training

1960

ACKNOWLEDGMENTS

The author wishes to thank Dr. Raymond N. Hatch who
served as his major professor and chairman of his guidance
committee. Dr. Hatch greatly assisted the writer in evalu-
ating his educational objectives and planning his doctoral
program. Similarly, the author wishes to express his
appreciation to the other members of his committee, Dr.
William Baten, Dr. John Krumboltz, and Dr. Buford Stefflre,
for their suggestions relating to this study.

Gratitude is expressed to Dr. John W. Truitt, Director
of the Men's Division of Student Affairs, for permitting the
study to be conducted in the men's residence halls. The
assistance in collecting the data which was given by the
head resident advisors and the 68 resident assistants was
greatly appreciated. Without their help, the study could
never have been completed.

Finally, to his wife, Louise, whose expectancies for
the author served as an ever present and increasing source
of motivation, special thanks for long hours of tabulating,

card sorting, and typing.

TABLE OF CONTENTS

Chapter

I.

II.

III.

FORMULATION AND DEFINITION OF THE PROBLEM

Definition and Clarification of Terms .

Theoretical Framework of the Problem and Basic

Hypotheses . . . . . . . .
Importance of the Problem . .
Scope of the Study . . . . .
Limitations of the Study . .
Organization of the Study . .

REVIEW OF THE RELATED LITERATURE

Social Psychological Foundations of

Review of Related Research. . . .
METHODOLOGY AND PROCEDURE . . . .
Description of the Samples . . .

The Original Population
The Non-Probationers .
The Probationers . . .
The Raisers . . . . . .

Development of the Instruments .

The Student Questionnaire . . .
The Peer Group Questionnaire .

Data Collecting Process . . . . .
Methods of Analyzing the Data . .
Reliability of the Instruments .

The Student Questionnaire . . .
The Peer Group Questionnaire .

Validity of the Instruments . . .

The Student Questionnaire . . .
The Peer Group Questionnaire .

Behavior.

Chapter

IV.

ANALYSIS OF THE DATA . . . . . . . . . . . . . .

Comparisons between Probationers and Non-
Probationers with Respect to Their Perceptions
of the Academic Expectancies held for Them by
Members of Their Home and Home Community Group .

Comparisons with Respect to Each of the Items
on the Student Questionnaire . . . . . .
Comparisons with Respect to the Total Expectancy
Scores Derived from the Student Questionnaire.

Comparisons Between Probationers and Non-
Probationers with Respect to the Ratings of
Their Peer Groups Made by the Resident
Assistants . . . . . . . . . . . . . . .

Comparisons between Probationers and Raisers
with Respect to Their Perceptions of the
Academic Expectancies Held for Them by Members
of Their Home and Home Community Group . . . . .

Comparisons with Respect to Each of the Items
on the Student Questionnaire . . . .
Comparisons with Respect to the Total Expectancy
Scores Derived from the Student Questionnaire.

Comparisons between Probationers and Raisers
with Respect to the Ratings of Their Peer
Groups Made by the Resident Assistants . . . . .
The Relationship between Total Expectancy Scores
and Grade-Point-Average for the Combined
Samples . . . . . . . . . . . . .
The Relationship between Academic Performance
and a Combination of Peer Group Rating and
Total Expectancy Score Derived from the Student
Questionnaire . . . . . . . . . . . . . . . . .

SUMMARY AND CONCLUSIONS . . . . . . . . . . . . .

Summary and Interpretation of the Findings . . . .
Implications for Further Research . . . . . . . .

BIBL IO GRAPI'IY O O O O O O O O O O O O O O O O O 0
APPENDIX C O O O O O O O O O O O O O O O O O O O

The Instruments . . . . . . . . . . . . . . . . .

Page
85

96

97

97
101

103

1014

109
115

117
131

1314
1140
141

Table

1.

2.

10.

11.

12.

LIST OF TABLES

Differences between Responses on Two Successive
Administrations of the Student Questionnaire .

Distribution of Expectancy Scores on Two
Successive Administrations of the Student
Questionnaire O O O O O O O O O O O O O O O 0

Differences between Ratings on Two Successive
Administrations of the Peer Group Questionnaire

Correlations between Item Response and Total

Expectancy Scores on the Student Questionnaire .

Relationship between Academic Performance of Peer
Group and Expectancies Held for Group Member
as Perceived by Member . . . . . . . . . . . .

Relationship between Ratings of Peer Group and
Grades Earned by Peer Group as Perceived by
Group Member . . . . . . . . . . . . . . . . .

Comparisons between Probationers and Non-
Probationers with Respect to Their Responses to
Each of the Items on the Student Questionnaire

Comparisons between Probationers and Non-
Probationers with Respect to Total Expectancy
Scores on the Student Questionnaire . . . . .

Comparisons between Probationers and Non-
Probationers with Respect to Total Expectancy
Scores-~Stratified by Scholastic Aptitude Level

Comparisons between Probationers and Non-
Probationers with Respect to Peer Group
Rat ings O O O O O O O O O O O O O O I O O O 0

Comparisons between Probationers and Raisers
with Respect to Their Responses to Each of the
Items on the Student Questionnaire . . . . . .

Comparisons between Probationers and Raisers with
Respect to Total Expectancy Scores on the
Student Questionnaire ,

Page

66

68

72

77

80

82

87

93

95

96

98

101

Table Page

13. Comparisons between Probationers and Raisers with
Respect to Total Expectancy Scores--Stratified
by Scholastic Aptitude Level . . . . . . . . . . . 102

14. Comparisons between Probationers and Raisers with
Respect to Peer Group Ratings . . . . . . . . . . 10A

15. The Relationship between Total Expectancy Scores
and Grade-Point-Average . . . . . . . . . . . . . 106

16. The Relationship between Scholastic Aptitude Scores
and Total Expectancy Scores Derived from the
Student Questionnaire . . . . . . . . . . . . . . 108

17. The Relationship between Peer Group Ratings and
Total Expectancy Scores on the Student Question-
naire O O O O O O O O O O O O I O O O O O O O O 0 ll]-

18. Comparisons between Probationers and Non-
Probationers of "Average" Scholastic Aptitude
with Respect to a Combination of Peer Group
and Home Group Expectancies . . . . . . . . . . . 113

CHAPTER I
FORMULATION AND DEFINITION OF THE PROBLEM

This study dealt with an analysis of selected Michigan
State University freshmen who were placed on academic pro-
bation at the end of fall term, 1957. Comparisons were made
between those probationary students who persisted as pro-
bationers throughout the year and students who were never
placed on academic probation during their freshman year.
Comparisons were also made between those probationary
students who improved their grades sufficiently to be re-
moved from academic probation by the end of the year and
those who persisted on probation.

The groups were compared with respect to (1) their
perception of the academic expectancies held for them by
members of their home and home community groups, and (2)
the academic expectancies held for them by members of their
university peer groups. Expectancy data were collected by
the use of two questionnaires developed by the investigator.

The home and home community expectancies were assessed
through the use of a questionnaire administered to the
students. The peer group expectancies were obtained from
ratings made by resident assistants assigned to the personnel
staffs of the men's residence halls.

It was hypothesized that there would be a positive

relationship between the academic expectancies held for a

student and his academic performance. The basic hypotheses
of the study were tested by comparing the students' responses
to the questionnaires with their academic performance and by
comparing the peer group ratings made by the resident

assistants with the students' academic performance.

Definition and Clarification of Terms

Probationers. The word probationers was used to

 

describe those students who were placed on academic probation
at the end of the fall term and persisted on probation
throughout the year. In general, their grades fell below

"C" or 2.00 grade-point-average.

Non-Probationers. Non-probationers were students
whose grades were consistently above 2.00 grade-point-
average. They were never placed on academic probation.
Raisers. For want of a better term, the term "raisers"
was applied to those students who, at the end of fall term,
were placed on academic probation but improved their cum-
ulative grade-point-average to 2.00 or higher by the end of
the year.

Expectancies. Expectancies were considered to be a
product of the attitudes and feelings which individuals or
groups of individuals have with respect to the behavior of

another person. They are communicated to the other person

through subtleties of language and association and

observation. Expectancies, as the term was used in this
study, involve some valuing of certain forms of behavior in
terms of relative goodness, badness or appropriateness.
Moreover, it is a person's knowledge of a given individual's
attitudes about his behavior and the meanings which the
person attaches to those attitudes which are important.

Academic Expectancies. The expectancies which indi—
viduals hold for others with respect to the importance,
necessity, or appropriateness of the other person's attending
and succeeding in an educational institution were called
academic expectancies.

Scholastic Aptitude. Scholastic aptitude was used
as a more appropriate term for what is typically called
intelligence. Tests of scholastic aptitude purport to
measure a student's general ability to do school work.

Resident Assistant. The resident assistants who
participated in the study were student assistants employed
as members of the residence hall advisory staff. Each
resident assistant had the direct responsibility for the
supervision and advisement of approximately 50 students who
resided on a given floor or wing of a men's residence hall.
He resided on the same floor or wing as the students he
served.

Precinct. The term precinct was used to describe a
subdivision of a residence hall presided over by a resident

assistant. A precinct housed approximately 50 students and

was confined to one floor or wing of the hall.

Theoretical Framework of the Problem and Basic Hypotheses

In order to give direction to the study, it was
essential that some consideration be given to a theoretical
framework within which the various aspects of the study
could be viewed.

This study was about students-~students who were
enrolled in a major, middlewestern, state supported, land-
grant university. The particular students who were of
concern were those who encountered academic difficulty
during their freshman year and, as a result, were placed on
academic probation. It was believed that there were deter-
minable differences that existed between probationary and
non-probationary students. Similarly, it was believed that
among the probationary students there were certain signifi-
cant differences that operated to determine the likelihood
of a student's being removed from probation at some later
time.

Of primary importance was the fact that grade-point-
average was used as a criterion for the original definition
of the groups involved in the study. Students with poor
grades were placed on academic probation. Others, with
better grades, avoided probation. Some, during the course

of their probation, improved their grades and were removed

from probation. Some of the questions raised in the mind of
the investigator were: What does the grade-point-average
mean? What are the factors that influence it? Can it be
said that some students are predisposed to earn low grades
and others higher grades? What factors are involved which
enable a student to improve his grades to an acceptable
level once he has been placed on academic probation?
Although it would be impossible to answer all these questions
in their entirety, it was felt the theoretical approach
developed for this study might provide new insights into

the problem.

While this study may be regarded as purely a study
about students, this approach falls short of the investi-
gator's intentions. The study is better described as a
study of human behavior. When a student comes to the
university he exhibits many different behaviors. He eats;
he sleeps; he studies; he goes to classes; he takes exam-
inations; and he engages in a multitude of social inter-
actions with individuals and groups. Some students get
higher grades than others; and some perform so poorly
academically that they are asked to withdraw from the
university. Other students are suspended for disciplinary
offenses. In this context no distinction is made between
"academic" behavior and other types of behavior. Behavior
is viewed simply as a phenomenon characteristic of the

human organism.

As we consider the various behaviors that students
exhibit, we immediately begin to speculate upon what causes
these behaviors. It may be hypothesized that, taken in the
aggregate, the various behaviors exhibited by a college
student may be attributed, at least, in part, to the set of
expectancies held for him by certain significant individuals
and groups of individuals in his life to whom he looks for
sanction and validation of his behavior. These "significant
others", as they have been called, are the persons in whom
the student places confidence; he respects their opinions;
he may like them or fear them; but he strives to be accepted
by them and to please them.

In analyzing the role that expectancies might play
in determining behavior it is important to consider the
sources of these expectancies. Two groups of significant
others may be considered as primary sources of expectancies
for university students. The first group is the home and
home community group. The second is the peer group at the
university. No doubt there are many individuals and other
groups of individuals whose expectancies for a particular
student could be studied; but, for purposes of this study
we have considered the two groups mentioned as being of
particular relevance.

It may be hypothesized that if a student comes from
a home and home community background where a premium is

placed on higher education, he is more likely to be successful

academically than a student who does not come from such a
background. If a student comes to the university with the
knowledge that there are people back home who are expecting
him to perform at a high level academically, it is reasonable
to suspect that this will have a positive influence in
determining the attention he gives to his course work. On
the other hand, the student who comes to the university
without such a set of expectancies being held for him may
have had a major source of motivation removed, and the
removal of this factor may be enough to materially influence
his grades in a negative direction.

A second group of significant others whose expectancies
of the student may be important in determining his academic
behavior is the peer group he is associated with at the
university. For the students considered in this study, this
group included roommates and friends who lived in close
proximity in the same residence hall. Again it may be
hypothesized that if a student is involved with a peer group
that places a premium on conscientious study this will
influence his grades in a positive direction: And similarly,
if the peer group is the kind that values a lot of things
more highly than studying this may have its effect in a
negative direction as far as his grades are concerned.

In summary, then, it may be hypothesized that infor-
mation about the expectancies and values held by the

significant others in a student's life may provide insight

into his academic behavior as well as any other type of
behavior he may exhibit. In addition, for purposes of this
study, two groups of individuals were identified who may be
regarded as significant others for the majority of students,
namely the home and home community group and the university
peer group. It was felt that information about the expec-
tancies held by those two groups for a given student would
be particularly relevant in attempting to explain his
academic behavior.

Stated in positive hypothesis form, the four basic
hypotheses tested in this study were:

(1) There is a positive relationship between a

student's academic performance and the academic

expectancies held for him by the significant

others in his home and home community group;

(2) There is a positive relationship between a

student's academic performance and the academic

expectancies held for him by the significant

others in his peer group at the university;

(3) There is a positive relationship between a

student's potential for being removed from

academic probation and his perception of the

academic expectancies held for him by the

significant others in his home and home com-

munity group; and

(A) There is a positive relationship between

a student's potential for being removed from

academic probation and the academic expect-

ancies held for him by the significant others

in his peer group at the university.

Hypotheses (3) and (4) are essentially corollaries of
hypotheses (l) and (2) but may involve somewhat different
dynamics. In any event, it was felt they should be tested

separately.

Importance of the Problem

Working with students who have academic difficulties
is a responsibility which confronts almost everyone working
in a university. Instructors, deans, counselors, and
residence hall personnel workers spend a considerable portion
of their time trying to help students overcome academic
deficiencies. In many instances these efforts are in vain.
Perhaps the problem is lack of time‘or lack of meaningful
information about students. Perhaps it is the inability of
the counselor or instructor to see the student's problem as
it really is or to take into account the student's value
system. It is quite possible that a lot of time is spent
working with students who might be better off both psych-
ologically and sociologically if they were not enrolled in
college at all, but engaged in some more satisfying

occupation elsewhere. Conceivably many probationary students

10

will later succeed regardless of any efforts to help them.
Answers to the questions implied here are needed if we are

to develop realistic programs of assistance to probationary
students. It was the purpose of this study to attempt to
discover some of those answers by investigating the relation-
ship between the academic expectancies held for a student

and his academic performance.

Most of the research on academic performance among
university students reported to date has dealt with predicting
grades from scores on tests of scholastic aptitude and other
standardized instruments. Studies are also reported which
have related "non-intellectual" factors such as personality
characteristics to academic success. Attention has been
attracted to the importance of first term's and first year's
grades in predicting future academic progress. However,
little research has been reported which relates academic
performance to the social dimensions which were prescribed
for this study. Previous studies have shown that scores on
tests of scholastic aptitude typically fall short of precise
prediction of academic performance. The explanation has been
that other factors operate which limit accurate prediction
from one or even a battery of standardized tests. The
investigation of some of these "other factors" provided the

justification for the present study.

ll
Scope of the Study

The study was restricted to freshman, male residents
of men's residence halls at Michigan State University and
an investigation spanning the first three terms of university
study for these men. This limitation in scope was planned
purposefully for the following reasons:
(1) The university policy requiring all freshmen
(except commuters) to live in university residence
halls;
(2) The availability of data concerning residents
of men's residence halls;
(3) The assistance in conducting the study which
could be given by the advisory staffs of the men's
residence halls;
(4) The high degree of interest shown in such a study
by administrative officials responsible for the
personnel program in the men's residence halls; and
(5) The necessity of imposing realistic boundaries
in order to keep the study within manageable pro-
portions from the standpoint of size of pOpulation

and time factors.

Limitations of the Study

It is evident from the outset that any research

dealing with human behavior is subject to certain inherent

12

weaknesses and uncertainties. Furthermore, the scope of this
study, as defined in another section, necessarily imposed its
own limitations on the general applicability of the findings.
As a result of these factors, the following limitations of
the study are noted:
(1) The students under consideration in this study
were freshman male residents of men's residence halls
at Michigan State University during the 1957-1958
school year. Thus the findings are directly appli-
cable only to this population. Any generalization
of the findings to other populations must be made
with caution;
(2) Grade-point-average was the determining factor
in academic probation and, as such, was subject to
the known lack of objectivity in the granting of
grades; and,
(3) Certain data were gathered by the use of
questionnaires developed by the investigator and
were subject to the usual limitations of imperfect

reliability and validity.

Organization of the Study

This thesis is divided into five chapters. In Chapter
I the problem is defined and presented in a theoretical
framework. In Chapter II the theoretical background is

13

presented from the literature of social psychology and a
report made of related research. The methods and procedures
used in the study are reported in Chapter III. Reliability
and validity studies of the instruments are also included in
Chapter III. The analysis of the data is presented in
Chapter IV. In Chapter V the study is summarized and the

significant findings and conclusions are reported.

CHAPTER II
REVIEW OF THE RELATED LITERATURE

Behavior is caused. Moreover, as Prescott (40) puts
it, "the causes of behavior are multiple, interrelated, and
complex." That these two statements are axiomatic is
attested to by the volume of literature dealing with the
subject of human behavior. Hardly ever have so many written
so much about a common problem--the explanation of why
pe0ple behave as they do. Philosophers have approached the
problem. Psychologists, sociologists, and anthropologists
have attacked the problem scientifically, each from his own
point of view. Today we have the social psychologist pon-
dering, hypothesizing, and studying the complexities of
human behavior.

The social psychologist, as an ”emergent form" of
scientist has certain advantages over his predecessors. He
has the laboratory and the experiment with which to test his
hypotheses. Thus, he exceeds the philosopher in his presen-
tation of evidence supporting (or negating) his convictions.
He has at his disposal the body of knowledge already accum-
ulated by traditional psychologists and social scientists.
Thus, he is free to select from what commonality does exist
in the various disciplines only those considerations which
help integrate theories of behavior and assist in the

development of new, unified explanations of human behavior.

15

The social psychologist has yet another distinct
advantage over his predecessors-~as well as some of his con-
temporaries. He can explain himself: That is, the emergence
of the "social psychologist" over the last half century is
readily explained by the application of social psychological
principles. The social psychologist will argue that the
human organism acquires most of its characteristics through
interaction with other human organisms. Through a process of
communication predicated upon common knowledge of certain
significant symbols, human beings acquire many strikingly
similar characteristics, attitudes, and behaviors. Further-
more, the extent of similarity depends upon the quality of
communication which takes place, which, in turn, depends on
the opportunity to communicate. Certainly the opportunity
for communication between the philos0phical, psychological,
and sociological disciplines has been present, particularly
since the advent of mass media of communication and rapid
transportation. Therefore, it is not at all surprising that
philosophers are observed acting like psychologists and
sociologists; psychologists have ventured into philosophy
and sociology; and sociologists have taken on some of the
attitudes of philosophers and psychologists. The outcome of
it all is the very emergence of the social psychologist, who
proceeds to study human behavior from an inter-disciplinary
frame of reference. In discussing the discipline of the

social psychologist, Newcomb (35:7) comments: "Social

16

psychology leans heavily upon contributions from the neigh-
boring disciplines of individual psychology, sociology and
cultural anthropology."

The foregoing "explanation of the social psychologist”
may be an over-simplification of the dynamics of the situation
and it doesn't explain why everyone in society isn't a social
psychologist, but it is indicative of the social psychological
point of view which has emerged in recent years as a new
approach to the study of human behavior. It should also be
added that the present writer is not certain that social

psychologists would subscribe to his ‘explanation of their
behavior" or that, in their pre-occupation with explaining
other people's behavior, they would even attempt to turn
their principles on themselves. Nevertheless, in the
opinion of this writer, the sheer existence of the social
psychologist in society today gives testimony and lends
credence to the principles to which they claim to adhere.

Some of the principles of social psychology are
presented in this chapter as a theoretical background for
the present study. Adaptation of the theoretical background
to the educational setting is attempted and examples of

educational research related to the study are presented.

17
Social Psychological Foundations of Human Behavior

It would be impossible to cite all the contributors to
social psychological theory. Perhaps, the best that can be
done is to present selected examples from the writings of a
few authors which demonstrate basic principles.

A convenient starting point is found in the contri-
butions of George H. Mead and John Dewey. From 1900 on,
Mead devoted his life to the development and elaboration of
a theory of intelligence and mind. During the same period,
John Dewey was actively engaged in attempts to place the
educative process in a social perspective and thus enhance
the effectiveness of formal education as practiced in schools.
Morris (3l:x-xi), contrasts the work of Mead and Dewey
thusly: "The work of Mead and Dewey is in many respects
complimentary, and so far as I know, never in significant
opposition . . . If Dewey gives range and vision, Mead gave
analytical depth and scientific precision."

George H, MEEQ: Let us now consider some of the
thinking of Mead. He stood in Opposition to the then pre-
valent theories of mind and self which did not explain how
minds and selves arose within conduct. Quoting Morris (31:
xiv) again:

This criticism breaks into two parts: (1) they
all in some sense presupposed antecedently existent

minds or selves to get the social process under way;
(2) even in respect to the phases of mind or self

18

which they did attempt to account for socially, they
failed to indicate the mechanism involved.

Mead's "social behaviorism" in contrast to Watsonian
behaviorism begins with the society and not\with the indi-
vidual. The organism is born into a society, é‘presently

‘\
existing and on-going society. It is the influencéspf the
social group on the individual that determines what it§xk
responses will be and the direction its behavior will takeP\Q\
Minds and selves are essentially social products facilitated
by physiological mechanisms. Quoting Mead (29:2) exactly:

Individual experience and behavior is, of course,
basic to social experience and behavior: the processes
and mechanisms of the latter (including those which are
essential to the origin and existence of minds and
selves) are dependent physiologically upon the pro-
cesses and mechanisms of the former, and upon the
social functioning of these.

Thus we see Mead's attempt to bridge the gap between
Watsonian "individual behaviorism" and his own "social
behaviorism".

Mead accepted Wundt's conceptualization of the term,
gesture, as the beginning of the social act which provides
stimulus to other forms involved in the same social act. He
extended the idea of gesture to include the notion that
gestures among human forms have meanings, ideas, or attitudes
attached to them which bring about the adjustment of the

responses of the different forms involved in the social

process. He extended the concept a step further by

19

suggesting "that the gestures mgag these attitudes on the part
of the form, that is, they have that meaning for us." (29:A5)
(underscore added by present writer)

In contrast to animal forms, whose gestures elicit
certain instinctive responses in each other, the gestures of
the human form provide symbols which answer to "meaning in
the experience of the first individual and call out that
meaning in the second individual. Where the gesture reaches
that situation it has become what we call 'language'. It is
now a significant symbol and it signifies a certain meaning."
(29:46)

Without elaborating more fully here on the content of
gestures and significant symbols, suffice it to add Mead's
summary statement. "Only in terms of gestures as significant
symbols is the existence of mind or intelligence possible;
for only in terms of gestures which‘are significant symbols
can thinking . . . take place." (29:47)

The concept, gesture, includes the possibility of
vocal gestures and thus we have the origin of spoken language.
"What language seems to carry is a set of symbols answering
to certain content which is measurable identical in the ex-
perience of the different individuals." (29:54) Mead notes
the point that when one makes use of a vocal gesture he tends
to respond to it. That is, it calls out in him the same
response that it calls out in others. Thus, only those vocal

gestures of the child are retained which call out a response

20

in him which is like the response he calls out in others;
"consequently, giving greater weight to those responses than
to the other responses, and gradually building up those sets
of responses into a dominant whole." (29:66) Infantile
babbling which may call out a response in the infant, but
which is ineffective in calling out the same response in
adults, soon gives way to more articulate vocalization,
depending, of course, on the extent to which the child finds
vocal gestures which call out in others the same responses
which are called out in him.
Mead summarizes his conception of the mind in the
following paragraphs:
We must regard mind, then, as arising and de-
veloping within the social process, within the
empirical matrix of social interaction . . . The pro-
cesses of experience which the human brain makes
possible are made possible only for a group of inter-

acting individuals . . . not for the individual or-
ganism in isolation from other individual organisms.

29:33)
Regarding the emergence of the mind he states:

Mind arises in the social process only when
that process as a whole enters into, or is present in,
the experience of any one of the individuals involved
in that process. When this occurs the individual
becomes self-conscious and has a mind . . . (29:13A)

He states further:

It is by means of reflexiveness--the turning
back of the experience of the individual upon him-
self--that the whole social process is brought into

21

the experience of the individuals involved in it; it
is by such means, which enable the individual to take
the attitude of the other toward himself, that the
individual is able consciously to adjust himself to
that process . . . (29:134)

And finally: ”Reflexiveness, then, is the essential
condition, within the social process, for the development of
the mind." (29:134)

Regarding the self, Mead says, ”the self is something
which has a development; it is not initially there, at birth,
but arises in the process of social experience . . ." (29:
135). He distinguishes between the self and the body in
that the self "is an object to itself". The very word,
"self", is a reflexive and indicates that it can be both
subject and object. The individual experiences himself
indirectly from the standpoints of other individuals of the
same social group. He enters his own experience as a self
only insofar as he first becomes an object to himself just
as other individuals are objects to him and "he becomes an
object to himself only by taking the attitudes of other in-
dividuals toward himself within a social environment of ex-
perience and behavior in which both he and they are involved."
(29:138) In this situation, communication becomes a form of
behavior which enables the individual to become an object to
himself. It is where one responds "to that Which he addresses
to another and where that response of his own becomes a part

of his own conduct, where he not only hears himself but

22

responds to himself . . . that we have behavior in which the
individuals become objects to themselves." (29:139)

The organized community or social group in which the
individual finds himself was termed the "generalized other"
by Mead. Speaking of the generalized other, he states:

It is in the form of the generalized other that
the social process influences the behavior of the in-
dividuals in it . . . for it is in this form that the
social process or community enters into the individual's
thinking. In abstract thought the individual takes the
attitude of the generalized other towards himself . . .
and in concrete thought he takes that attitude insofar
as it is expressed in the attitudes towards his behavior

of those other individuals with whom he is involved in
the given social situation or act. (29:155-156)

In summary, Mead concludes: "The self-conscious indi-
vidual, then, takes or assumes the organized social attitudes
of the social group or community to which he belongs . . .
(and) he governs his conduct accordingly." (29:156) And
further:

The self reaches its full development by or-
ganizing these individual attitudes of others into the
organized social or group attitudes, and by thus be-
coming an individual reflection of the general systematic
pattern of social or grou behavior in which it and
others are all involved. 29:158)

Thus we have, in a highly abbreviated form, the
thoughts of George H. Mead regarding the emergence of mind
and self, through language, and the impact of the social group

as a determinant of individual human behavior.

John Dewey. As has been previously indicated, John

23

Dewey's contribution to social psychological thought was
expressed in terms of its implications for education. He

saw education as a social function carried on in a social
environment. Like Mead, he was a social behaviorist. He was
concerned with the "way in which a social group brings its
immature members into its own social form." (13:12) His

general answer to the question was:

By means of the action of the environment in
calling out certain responses. The required beliefs
cannot be hammered in; the needed attitudes cannot be
plastered on. But the particular environment in which
an individual exists leads him to see and feel one
thing rather than another; it leads him to have
certain plans in order that he may act successfully
with others; it strengthens some beliefs and weakens
others as a condition of winning the approval of
others. Thus it gradually produces in him a certain
system of behavior, a certain disposition of action.

(13:13)

In discussing the impact of the expectations of others
for the individual, Dewey states:

A being whose activities are associated with
others has a social environment. What he does and
what he can do depend upon the expectations of others.
A being connected with other beings cannot perform
his own activities without taking the activities of
others into account. For they are the indispensable
fondifiions of the realization of his tendencies.

13:1

Dewey stresses the importance of group members'
having shared experiences and cites language as the chief

facilitator of sharing as well as being a product of shared

experience itself. His explanation of the development of

24

language is essentially the same as Mead's. He summarizes
his discussion thusly: "The bare fact that language consists
of sounds which are mutually intelligible is enough of

itself to show that its meaning depends upon connection with
a shared experience." (13:18)

Dewey denounces the theory that the social control of
individuals arises from an innate tendency of humans to
imitate and thus conform. "But imitation throws no light on
why they act so; it repeats the fact as an explanation of
itself . . . This social fact (imitation) is then taken for
a psychological force, which produced the likeness." (13:41)
In discussing the wide variation in customs among social
groups he says:

- The mere fact that customs are different means
that the actual stimuli to behavior are different.
(There is no need to appeal to imitation as an ex-
planation.) Conscious instruction plays a part; prior
approvals and disapprovals have a large influence.
Still more effective is the fact that unless an in-
dividual acts in the way current in his group, he is
literally out of it. He can associate with others on
intimate and equal terms only by behaving in the way
in which they behave. The pressure that comes from
the fact that one is let into the group action by
acting one way and shut out of it by acting in another
way is unremitting. (13:41-42)

Dewey's concern was that the processes of instruction
in school be unified in that they concentrate on the pro-
duction of good habits of thinking. The necessary conditions

for effective, democratic instruction were simply put and

derived from a social psychological base.

25

They are first that the pupil have a genuine
situation in which he is interested for its own sake;
secondly, that a genuine problem develop within this
situation as a stimulus to thought; third, that he
possess the information and make the observations
needed to deal with it; fourth, that a suggested
solution occur to him which he shall be responsible
for developing in an orderly way; fifth, that he have
the opportunity and occasion to test his ideas by
application, to make their meaning clear and to
discover for himself their validity. (13:192)

That Dewey and Mead subscribed to the same basic
orientation is clear. Philosophically they were both prag-
matists; psychologically they were both social behaviorists.
They were close friends. As Morris (3l:xi) puts it: "A
natural division of labor at a common task was the result."
Dewey's contribution was through the medium of educational
philosophy and psychology and Mead's contribution has been

described as a landmark of social psychology.

Contemporary Authors. Many writers have amplified and

 

refined the basic foundations of mind and self which Mead
laid down. An example is a statement regarding the feelings
of security of the individual made by Kimball Young.

It (security) begins with the social act involving
the child and the mother, and in time it includes re-
lations with the father, the siblings, other relatives,
playmates, neighbors, teachers and other adults. The
source of security arises from the consistency of the
reactions of others towards the child, which he, in
turn, introjects and uses as a basis for defining his
own role and hence for determining his actions and
attitudes toward himself and others. Self-assurance,
self-reliance, and self esteem--three basic components
of the self-~depend for their origin and continuance
upon the ability to meet demands defined at first by
others and later by ourselves in advance. (55:42)

26

Gardner Murphy (33:479) offers a simpler view of self:

. . . There is an organism, which among its many
functions includes the function of observing and knowing
. . . Being a more or less integrated system of responses,
the organism appropriately orders these diverse impres-
sions into an integrated whole and agrees to call it
by the name which others have given to it . . . From
the diverse knowing and thinking processes a conceptual
unity is deduced. The self is a thing perceived . . .
conceived . . . responded to. It comes gradually
into being as the process of differentiation goes on
within the perceptual field.

Murphy discusses the process of identification by
which the self is integrated or gains unity.

. . . Selfhood is interwoven with experiencin
other individuals. As fast and as deeply as (mother%,
father, brother, sister make their impression upon the
child, they become parts of him too. This dependence
of self upon the perception of others is a primary
clue to the social nature of man and to his utter in-
capacity for any complete autonomy of either perception
or action . . . Man reacts not only to the behavior of

others but to their thoughts regarding himself. (33:
491-492)

H

Newcomb (35:5-7) paraphrases Mead thusly: . .
every human infant enters a society that is already a going
concern . . . In the process of social interaction . .
peOple perceive and respond to one another and are them—
selves changed in so doing."

Newcomb makes a fine distinction between the reference
groups and membership groups of an individual. Membership
groups are simply groups to which a person is recognized by

others as belonging; while reference groups are those groups

27
whose norms affect his attitudes and behavior.

All membership groups probably serve as reference
groups to some degree and in some ways. But not all
reference groups are membership groups; most of us are
influenced by the norms of some groups in which we are
not recognized by others as belonging. (35:224)

The inter-relation of position and role are discussed
by Newcomb (35:28). "The ways of behaving which are ex-
pected of any individual who occupies a certain position
constitute the role associated with that position."

Linton (26:77) extends the meaning of the word,
position, by substituting the word "status". By status he
means "the place in a particular system which a certain in-
dividual occupies at a particular time . . ." In this con-
text, the "particular system" may be the family group or any
of a number of "association groups" including age-sex groups,
friendship groups, or work groups. The association groups
or, for that matter, the family groups, through adoption or
marriage, may change with time. A person's role, then,
becomes "the sum total of the culture patternsassociated
with a particular status. It thus includes the attitudes,
values and behavior ascribed by the society to any and all
persons occupying this status." (26:77)

It is interesting to note Newcomb's use of the generic
term, "reference group", to denote the dynamic qualities of

Linton's family group and association groups.
Tolman (49:350) provides a psychological model for

28

analyzing the matter of status and expected roles. He
conceptualizes role thusly: :f‘;c“va

A role is thus a series of appropriate and ex:
pected ways of behaving relative to certain objects, by
v r ue of a given individual's status in a given social
structure or institution.

Further, these expectations that individuals in
a given status will behave in such and such ways are
called role expectations. This term has a double
meaning. It applies not only to the expectations of
the alters . . . that ego will behave in certain ways
but it applies as well to the expectations of ego that
if he behaves in these expected ways, the alters will
meet his behavior with approval . . .

 

Talcott Parsons and Edward Shils (36:190-191) comment

along the same line of thought:

The primary ingredients of the role is the role-
expectation . . . Role-expectations organize (in accor-
dance with general value-orientations) the reciprocities,
expectations, and responses to those expectations,
in the specific interaction systems of ego and one or
more alters . . . It is in this reciprocity or com-
plimentarity that sanctions enter and acquire their
place in systems of action.

Insofar as ego's set of role-expectations is
institutionalized, the sanctions which express the
role-expectations of the other actors (alters) will
tend to reinforce his own need-dispositions to conform
with these expectations by rewarding it and by punishing
deviance.

Cottrell (10) offers a list of sixteen basic propo-
sitions concerning inter-personal behavior. Slightly

abbreviated, a few are reproduced here which indicate the

general tone of the complete listing.

29

I. When human organisms respond to each other

over a period of time, the activity of each becomes the
' stimulus pattern for a more or less stabilized response
pattern in the other (s) . . .

II. . . . each member of an inter-personal
relationship . . . is conditioned to respond to his own
response series as a stimulus series . . .

III. The impact upon one human organism, A, of
the activities of another, B, . . . conditions in A
the response pattern of B to A‘ag A has perceived that
action . . .

IV. The self—other patterns of each member of
an interact system are frequently not congruent. The
more intimate the contact through time, the greater
will be the tendency for the patterns to coincide.

X. The personality system includes self-other
patterns developed by other persons and acquired through
the process of identification.

XIV. The person responds in a social situation
according to his own definition of the situation.

Simon (45) has developed a mathematical model for
translating the behavior of group members into mathematical
language. His formulas require four variables; (1) the
intensity of interaction among members; (2) the level of
friendliness among them; (3) the amount of activity carried
on by members of the group; and (4) the amount of activity
imposed on the group by the external environment. A formi-
dable group of equations is the result, which according to
Simon, "help in the clarification of concepts . . . and the
derivation of new propositions . . ."

Snygg and Combs (46:15) present the most adamant and

unequivocable view of the determinants of human behavior

3O

enCountered by this writer. They say, "All behavior, without

exception, i§_comp1etely determined by_and pertinent to

 

the phenomenal field g£_the behaving organism." In this

 

context, the "phenomenal field" means "the entire universe,
including himself, as it is experienced by the individual at
the instant of action."

The "phenomenal self”, as Snygg and Combs refer to it,
is a product of the phenomenal field of a person and includes
that portion of the field which has a personal referent, that
is, has a special reference to self. Furthermore, it is the
phenomenal self that is of prime importance in determining the
majority of an individual's behaviors. The phenomenal self
is not thought of as an entity or object, but as "a pure ab-
straction created for convenience and understanding." (46:

57)

The basic human need is defined as "the preservation

 

and enhancement 2: the phenomenal self . . . When each act

 

is seen as an attempt to preserve or to fortify the indivi-
dual's concept of himself, behavior becomes meaningful."
(46:58)

Thus, the tendency of individuals to identify with
groups is explained in terms of need satisfaction. The
following principles laid down by Snygg and Combs are relevant
to this point:

31

(1) Individuals tend to seek self enhancement
through identifying themselves with and winning the
approval of groups or individuals they believe to be
important.

(2) People tend to withdraw from groups whose
approval they are unable to win and from groups which
no longer satisfy their needs.

(3) Identification of an individual with a group
leads him to adopt and defend the standards and be-
havior of the group. (46:187-188)

Cartwright and Zander (8) discuss in detail the
questions of group cohesiveness, group pressures, and group

standards in their volume, Group Dynamics. In general their

 

thesis is that group membership provides a source of need
satisfaction for group members and that persons will remain
in or leave a given group depending upon the satisfactions
they derive. The original motivation for becoming identified
with a group may be any of a number of reasons all related

to need satisfaction. Speaking of groups in general, they
say:

. . . attraction to the group will depend upon
two sets of conditions: (a) such properties of the
group as its goals, programs, size, t pe of organization,
and position in the community; and (b) the needs of the
person for affiliation, recognition, security and other
things which can be mediated by groups. Both the nature
of the groups and the motivational state of the persons
must be treated in any adequate formulation of group
cohesiveness. (8:76)

Once a person has become identified with a group,
however, he is immediately subject to group pressures and

group standards. If he is to remain in the group, some of

his previously held values which may be at variance with

32

major group values must give way to acceptance of group
values. Otherwise, he is likely to be rejected by the group
and its attractiveness for him diminished. Thus, the source
of his need satisfaction which originally prompted him to
identify with the group is curtailed and withdrawal from the
group in search of another is his only recourse.

After noting certain research findings relative to
conformity among group members, Cartwright and Zander cite
the following as plausible explanations:

(a) Membership in a group determines for an
individual many of the things he will learn, see, do,
think about, and so on. The nature of the stimuli in
the environment of a person are in large part affected
by his roup membership.

b) An individual may act like others in the
group because they are attractive to him.and he wants
to be like them.

(c) A person may behave in a manner similar to
the rest of the group because he fears punishment,
ridicule, or rejection by members of the group unless
he does act as they do. (8:139)

With slight modification, Cartwright and Zander have
stated one of the major hypotheses of the present study:
Among college students . . . the "eager beaver"
is reminded that "a C is a gentleman"s grade.” In this
fashion average students indicate to the honor scholar

that his behavior is too different from theirs to be
acceptable, despite their professor's hopes in the

matter. (8:137)
Yoshino (54) concluded, after a study of college
dropouts, that despite differences in high school rank and

scholastic aptitude scores, "there are socio-economic and

33

motivatiOnal factors that must be taken into consideration
in accounting for the success or failure of any given
student."

Combs (9) approaches the problem of school learning
by focusing on the concept "intelligence”. Intelligence is
thought of as the effectiveness of a person's behavior rather
than an inborn organic quality. Since effectiveness of
behavior is determined by the perceptions a person is able to
make, intelligence becomes a function of the adequacy of "the
perceptions that he can make in his own unique perceptive
field". The limitations on the development of intelligence,
then, become synonomous with the limiting factors upon the
perceptions of individuals.

Limiting factors on perception include certain physical
conditions such as pre-natal or congenital damage or mal-
formation of the organs of the nervous system; environmental
deprivement or lack of Opportunity to make perceptions and
thus increase differentiation in the perceptive field; lack
of time to build new perceptions; depressed levels of an
individual's goals or values; inadequate self-concept or in-
ability to see one's self as capable of achievement; narrowing
of the perceptual field due to perceived threat to the in-
dividual; and the draining off of energy in defense of the

self under threat.

Combs does not suggest that physiological limits do
not exist in respect to intelligence, but he is suggesting

34

"that we may have conceded too early that we had approached
those limits." The implications of this conception of in-
telligence, particularly for schools are "(a) to free
individuals from the restraints upon perception and (b) to
provide the opportunities for perception to occur."

Several provocative questions relevant to educational
practices are raised by Combs. A few are reproduced here:

What effects might we be able to produce by
providing experiences that provide adequate concepts
of self in children and adults? . . .

What would happen if we were consciously and
carefully to set about the task of providing exper-
iences that would lead peOple to conceptions of
themselves as adequate, worthy, self-respecting
people? . . .

Finally, if threat to the individual has as
important effects as seem indicated in this dis-
cussion, the removal of threat would seem a most
important factor to consider in the release of the
individual to perceive more adequately.

w- OH..-“_'
‘AM‘ u...

fl/””7Brookover’/(6), proceeding from a definition of

\\_’\Mu‘_~r..

education which makes education virtually synonymous with
socialization, is led to analyze school learning on the same
basis that social learning and behavior of all kinds is
analyzed and explained. With regard to the specific role

of the school as an agent of socialization, he comments:

In addition to the social climate, which defines
the general norms and conditions of behavior, and the
models of behavior presented to the pupils, the school
functions in socialization by defining the specific
expectations of the students. In general, these expec-
tations of the student role are primarily defined for
the student by three groups: the teachers, the parents,
and other students. (6:348) (underscore added by
present writer)

35

Brookover suggests further that with a few exceptions
the expectations of parents and teachers are fairly conver-
gent with respect to the scholastic performance of elementary
school children. That is, they both expect the child to
learn the three R's, work hard, and stay out of "trouble".
Notable exceptions occur when parents' definitions of
' "trouble" conflict with teachers' definitions. However,
secondary school and college students, in contrast to elemen-
tary pupils, are confronted more realistically with role
expectations which derive from interaction with a peer group.
These expectations may be clearly in opposition to previously
internalized values as well as the current desires of teachers
or parents. The "curve raiser" or the "grind" may be praised
by teachers and parents but condemned by his peers. The
extent to which peer approval is preferred to adult approval
at this stage may be a critical variable in determining
scholastic performance.

In a subsequent journal article, Brookover (7) con-
ceptualizes a social psychological basis for classroom
learning. He notes the dilemma of the present-day educator
who is faced with societal demands for more highly trained
scientists, but at the same time has predominately "average"
students to train. The panic produced by this apparently
impossible task has resulted in frenzied re-organization of
school curricula with primary emphasis on honors courses and

large-scale programs for identification and placement of the

36

"gifted learners".

The educator is seen as stranded with a learning
theory, derived largely from individual psychology, which
assumes that the human organism is innately equipped with a
fixed capacity to learn. Intelligence tests reflect this
theory. Students of average IQ are assumed to be incapable
of becoming research scientists and engineers.

Brookover does not deny the possibility that variation
in school learning may be influenced by organic differences,
but he suggests "that these organic differences have not
been identified. We know, however, that human beings have
a tremendous range of learning possibilities and that no one
has reached the end of learning."

After considering certain assumptions and general
hypotheses regarding human learning in a social context, the
following specific but tentative hypotheses are suggested by

9/ “\
"Brookover:;>

_. __ -.

1. Persons learn to behave in the ways that
each considers appropriate to himself. Thus, each
child and each adult learns to do those things that
are viewed as proper, required, necessary and
desirable by the individual . . .

2. Appropriateness of behavior is defined by
each person through the internalization of the
expectations of significant others . . .

3. The functional limits of one's ability to
learn are determined by his self-conception or self-
image as acquired in social interaction . . .

4. The individual learns what he believes
significant others expect him to learn in the class-
room and other situations . . .

37

Brookover summarizes the social psychological con-
ception of school learning thusly:
If the foregoing hypotheses are valid, the lim-
itations we have placed upon our human resources by

our conception of individual limitations on learning
would be greatly expanded.

This requires the creation of a climate of
learning in schools which defines maximum achievement
as the norm . . . If the educational system recognizes
this and applies a social conception of learning to
the school situation as we have long since done to
other kinds of learning endeavors, a high level of
educational achievement throughout the society may be
realized.

Review of Related Research

Studies relating various factors to educational
achievement have frequently involved correlations between
assumed independent variables and observed, but still assumed,
dependent criteria. The present study is no exception since
it depends on statistical inference from population descrip-
tions based on normative data. A danger inherent in all
studies of this sort is that concomitance of events may be
mistaken for causation. To be explicit: in the present
study the academic performance of a student's peer group
was found to be related to his own academic performance.
What assurance is there that like-achievers do not naturally
group together as a result of their common characteristic,

like-achievement? Or, on the other hand, can like-achievement

38

among group members be attributed, at least in part, to some
degree of pressure to conform to group standards (be they
high or low) among group members?

A study relevant to these questions was conducted by
Hoffman (23) in a study of small group interaction in a
classroom setting. The question considered was, ”do people
select each other as friends because they are similar, or do
they become similar because of the interactions connected
with their friendships?" Groups of four persons each were
assembled in seven laboratory sections in an undergraduate
psychology course at the University of Michigan in 1955. The
basic purpose of the grouping was to study the effects of
group composition on problem solving performance, however, a
partial answer to the question mentioned above was an added
outcome.

On the basis of similarity or dissimilarity of person-
ality as measured by the Guilford-Zimmerman Temperament Survey,
students were assigned to homogeneous or non-homogeneous
groups. The correspondence between the personality profiles
of the students on the GZTS was determined by using Kendall's
tau as a measure of profile correlation. Thus, homogeneous
groups were formed from students whose personality profiles
were positively inter-correlated and non-homogeneous groups
were constituted of persons whose profile inter-correlations
were zero or negative. Data on 70 students in homogeneous

groups and 128 students in non-homogeneous groups were

39

reported. The groups worked together in class on role playing
problems and case discussions and thus got to know each other.
At the end of the course, a measure of group attractiveness
was obtained for each member by asking him to indicate on a
sociometric questionnaire the three people from the total
laboratory section he would most like to be in a group with
if he were to take the course over again. The frequency

with which a member chose his own group members provided an
index of the member's attraction to his group. The mean
number of in-group preferences for the homogeneous groups was
1.49 and for the non-homogeneous groups it was 1.36 with
standard deviations of approximately 1.00 in each case. The
difference between the means were not significant at the .05
level. Thus, the hypothesis that similarity of personality
leads to personal attraction was not supported. It was shown,
however, that quality of group problem solving was greatest
among homogeneous groups with high in-group preferences. No
relation between in-group preferences and quality of group
performance was noted for the non-homogeneous groups.

Perhaps the most provocative experiment of all which
involved conformance of group members to group norms was the
classic experiment conducted by Sherif (44) in 1935 at
Columbia University. He utilized the autokinetic phenomenon
as a medium to study the relationship of group interaction
to individual perceptions. A stationary pinpoint of light,

exposed for two second intervals in a dark room, was perceived

40

by subjects to move each time it was flashed on and off.
Subjects, taken individually, were told to indicate in inches
how far they thought the light had moved each time it was
exposed. It was found that while there were wide variations
among individuals as to their perceptions of movement, each
subject soon fell into a characteristic pattern of response.
That is, his perception of the movement was restricted to a
range of only a few inches. When groups of two and three
were simultaneously viewing the light and each one reported
orally his estimate of the movement, there was a tendency for
the group to begin to agree on the perceived movement and
thus establish a group norm for their various estimates.
Most notable, however, was the fact that groups formed from
subjects who had experienced the phenomenon previously (and
had established their own individual norms) were also found
to converge toward a new group norm which in many instances
was quite'different from their original individual norms.
These groups took longer to establish the group norm, but in
time, the group norm was clearly established. Twenty-four
hours later, the subjects were again exposed individually to
the flashing light and their estimates of movement still
clustered around the previously established grgup_norms.
Subsequent experiments by various investigators have
confirmed Sherif's findings. Schonbar (43) used a line
length as the stimulus and found similar results. Bovard (4)

found individual subjects still influenced by the group norms

41

28 days after the group norms were established. Rohrer, gt
.§1. (42) found the group norms to be stable after a period

of one year. Bovard (5) also found that subjects who had
previous class experience in a course in psychology where
group discussion was emphasized shifted to a common group

norm more quickly than students who had been in teacher cen-
tered classrooms. In this experiment the stimulus was a green
rectangle the length of which was to be estimated by the
groups.

While these studies are basically laboratory studies,
they are indicative of the many studies of group cohesiveness
and group conformity which have been reported. Hare, §t_al,
(22) have summarized and reported 584 separate studies and
articles which deal with interaction within small groups.

The essence of these studies cannot be spelled out in a few
words, but, in general, the social determinants of general
human behavior are widely documented by research findings.

Studies more specifically related to the present study
include investigations relating academic performance to
social factors.

Haggard (21) reported in 1957 a longitudinal study of
45 children who had been studied from the time they entered
the third grade until completing the seventh grade. The
students were all enrolled in the Laboratory School of the
University of Chicago. For the most part the children were

from professional homes and many of the fathers were

42

university professors. Among other things, measures of
parental pressure to achieve academically were studied in
relation to the students' academic performance.

"By the time they had arrived in Grade III, the high.

general achievers (in contrast to low achievers) were sen-

 

sitive and responsive to socialization pressures, had largely
accepted adult values, and were striving to live up to adult
expectations." They showed a higher level of overall adjust-
ment than did the low academic achievers. By Grade VII, the
high achievers continued to respond to adult standards of
behavior, but "had developed strong antagonistic attitudes
towards adults"--attitudes which were not expressed by the
low achievers. By Grade VII, the high achievers had emerged
as social leaders of their peers.

Two summary statements are particularly noteworthy.

Our findings indicated clearly that, in the
setting of this study, it is not only the parents and,
to a lesser extent, the teachers who exert strong
pressures on the child to achieve academically. The
other Children, coming from similar backgrounds, also
exert strong pressures which are not openly observable
but nonetheless add to the pressures imposed by adults.

. . . Is it necessary for these parents to
exert strong pressures upon their children to achieve
academically. Our knowledge to date suggests that,
if the parents are themselves intellectually oriented
and are not ambivalent about the value of intellec-
tuality as a way of life, the child, as a matter of
course, will accept the same value system and acquire
the appropriate behavior patterns.

43

Drews and Teahan (15) tested the hypothesis that the
parents of high academic achievers are more demanding, more
autocratic and less accepting in the treatment of their
children than the parents of low academic achievers. Samples
of 40 gifted students and 28 average IQ students were iden-
tified with gifted defined as Stanford-Binet IQ 130 or above
and average defined as IQ from 93 to 120. Samples were drawn
from junior high school students in Lansing, Michigan in 1957.
Within the gifted and average groups, half of each group had
maintained "A" averages for two previous years and were
classified as high achievers. The other half of each group
were classified as low achievers. Gifted low achievers had
grades of "B minus" or below and average low achievers had
grades of "C minus or D".

A Parental Attitude Survey was constructed using items
from an instrument devised by Shoben. The instrument was
administered to the mothers of the children with the following
results: mothers of high achievers were more authoritarian
and restrictive in the treatment of their children than the
mothers of low achievers; and, the parents of gifted achievers
also seemed to have more punitive attitudes in rearing their
children.

In a further report on the Drews and Teahan data,
Hurley (25) pointed out that while the mothers of the high

academic achievers were more dominating and ignoring than the

 

low achievers' mothers, the mothers of the intellectually

 

44

gifted group were markedly less dominating and possessive
than the average children's mothers. As a result he suggests,
"the interpretation that some middle degree of maternal
domination has an 'optimal positive impact' upon both the
intelligence and the academic success of the child does
appear to be consistent with these findings."

In a survey of recent research on achievement among
gifted students, Gowan (20) gleaned the following common
findings which differentiate achievers from non-achievers:
'clearness and definiteness of academic and vocational goals;
strong ego controls and strength; parents who took pains to
motivate their children; some tension in tasks demanded in
childhood; enthusiastic, socialized, activity oriented view
of life. In contrast to the Drews and Teahan findings,
Gowan found permissive parental attitudes to be associated
with achievers and autocratic parental attitudes to be
associated with non-achievers. In summary, Gowan states:

Thus achievement and under achievement in the
gifted may be viewed as social and asocial responses
of the individual to proper stimulation . . . either
tendered or denied by the parental and educational
environments.

Ford (17), in a study of Kentucky junior high school
students in 1956, points out the following factors found to
differentiate over- and under-achievers: membership of
parents in PTA; attitudes towards school; Occupational am-

bition; perceptions of occupational ambitions held for them

45

by parents; and punishment and scolding received for making
poor marks. Other strong factors were: marital status of
parents; employment of mothers; time spent studying at home;
and types of persons named as models, "ideals", or objects
Of admiration. Ford noted that social class differences, as
such,-between over- and under-achievers were not significant.
He comments thusly: ". . . the chief inference which should
be drawn is that explorations of social factors affecting
academic achievement which fail to go beyond the mere recog-
nition of social—class differences have been brought to a
premature conclusion."

Several investigators (30, 32, 48) have studied the
efficiency of the Rorschach Test as a predictor of academic
achievement and found significant relationships. Sopchak (47)
included the Rorschach along with other predictors in a study
of the academic success of 356 Adelphi College freshmen in
1957. The students were administered the California Reading
and Language Tests, and American Council on Education Psych-
ological Examination, and the Harrower Multiple Choice
Rorschach. Data on high school grades were also available.
High school grades were found to correlate most highly with
freshman grades (.64). ACE and California Test scores
correlated from .23 to .49 with grade-point-average and
Rorschach-GPA correlations ranged from -.20 to .24. Total
scores, human movement responses, and small detail responses

were the only Rorschach scores found to be correlated with

46

GPA at the .05 level of confidence out of the eleven separate
Rorschach scores obtained.

Holtzman, e; §_l_. (214) have designed a Survey 93 M
Habits and Attitudes "to meet the need for an easily admin-
istered, valid measure of study methods, motivation for
studying, and related attitudes of importance in scholastic
success". Correlations from .26 to .66 between Survey scores
and GPA are reported for a sample of 1756 men and 1118 women
in ten different colleges which were studied in 1954.

Ahman, £32. 53;. (1) studied the validity of the Survey
2f Study Habits and Attitudes devised by Holtzman and col-
leagues. A sample of 301 male freshman at Cornell University
in 1956 were administered the Survey. The correlation
between Survey scores and fall grade-point-average was found
to be .08. Survey scores correlated with Ohio State Univer-
sity Psychological Examination to the extent of .15. A group
of 50 over-achievers and 46 under-achievers were identified
from freshman males who enrolled in the fall of 1955. These
two groups were compared with respect to their responses to
each of the 36 items on the Survey. Chi-square was employed
to demonstrate the significance of the observed relationships.
Only two of the items were found to differentiate signifi-
cantly (.05 level) between the under- and over-achievers as
a group. These two items dealt with the students' interest
in school work. A study of item discrimination using the

total score on the Survey as the criterion resulted in

47

discrimination indices which were judged "satisfactory" by the
investigators. Indices ranging from .40 to .60 were found

for the majority of the items. In summarizing the study, the
investigators concluded: "In this instance the SSHA did not
display predictive validity to any noticeable degree al-
though the test items did consistently exhibit satisfactory
discriminating power (internal consistency)."

Garcia and Whigham (18), in a study of 509 Emory
University freshman in 1956, reported a correlation of .25
between scores on the Brown-Holtzman Survey 9£_Study_Habits
and Attitudes and grade-point-average for the freshman year.
The Survey scores were also fOund to correlate with total
scores on the American Council on Education Psychological
Examination to the extent of .14. Administering the Survey
later in the school year, after the students had experienced
study conditions at the university, resulted in slightly
higher (but not significantly) correlations with grade-
point-average.

Myers and Schultz (34) devised a questionnaire which
sampled motivation for attending college, intellectual
interests, teacher relations, and study habits and admin-
istered it to entering freshmen at an eastern women's
liberal arts college. The classes of 1951 and 1952 were
chosen for study. A correlation of .10 was found between
questionnaire scores and grade-point-average.

In the early 1950's, Gough (19) developed a personality

48

scale to predict undergraduate grades. The title of the
scale was Honor-Point-Ratio Scale (Hr) and it was included
in the California Psychological Inventory along with other
personality scales. A mean correlation of .38 in eleven
cross validating college samples totalling 1253 cases was
attained. The mean correlation with scholastic aptitude
scores was .26. Some of the more prominent social-interactional
implications of higher and lower scorers were: "high scorers
tend to be seen as capable, intelligent, and reliable and low
scorers as dissatisfied, dull, rigid, and shy."

Bendig and Klugh (2) validated Gough's Hr Scale
against students' self-reported grade-point-averages and
found a median correlation of .32 between Hr and the cri-
terion for four samples of approximately 100 each at the
University of Pittsburgh in 1954-1955. In the same study they
related scores on Taylor's Manifest Anxiety Test to grade-
point-average and found no significant relationships.

Scores on the various scales of the Kuder Preference
Record were correlated with the academic performance of over
400 University of Georgia male students in 1946-1947 by
Phillips and Osborne (38). They found no significant re-
lationships. Correlations ranging from -.11 to 4.08 were
reported. Normative comparisons between academic probationers
and non-probationers likewise revealed no statistically
significant differences.

Wilson (53) studied the personality characteristics of

49

100 Los Angeles high school students in 1947. Fifty of the
students had failed two or more subjects the previous semester
and fifty had never failed. The students' responses to the
Thematic Apperception Test and the Bell Adjustment Inventory
were used for comparing the two groups. No significant
differences between the two groups were demonstrated, and it
was concluded "that failure in the high school does not
necessarily indicate a maladjusted personality . . ."
Bledsoe (3) related size of high school to grade-
point-average for freshmen enrolled in all the state colleges
and universities in the state of Georgia between 1924 and
1951. All high schools that had been accredited by the
Georgia Accrediting Commission were involved in the study.
He concluded, "students from Georgia high schools who are
members of large graduating classes tend to make signifi-
cantly higher average marks during the first year of college
than do students who attended small and middle sized high
schools."

DiVesta, §§.§1, (14) devised an Orientation Inventory

 

to be administered to students for purposes of predicting
their grades. Samples of 24 and 28 respectively were iden-
tified as under-achievers or over-achievers from a population

of Cornell University students in 1949. Inventory items

 

dealing with the number of hours the students worked; ability
in note taking; and satisfaction with present courses; were

found to differentiate between the two groups as defined.

50

A study to test the hypothesis that over- and under-
achievement among students is directly related to the reward
value of high marks for the individual student was conducted
by McDavid (28). In this study scholastic marks were viewed
as some kind of external, non-objective standard which con-
stitutes a form of interpersonal evaluation of a student by
his teachers. High grades were assumed to be a form of
approval on the part of the teacher for a given student's
performance and, hence, a source of social reinforcement.

The Situational Interpretation Test was used to
determine the effective value of approval and disapproval as
reinforcers of an individual's behavior. In effect, the
§Ig_assesses a person's psychological need for approval of
his behavior. A sample of 23 over-achievers and 21 under-
achievers was selected from among the juniors and seniors in
a New Jersey high school in 1957. Scores on the SIT ranged
from zero to eight, with the mean for under-achievers being
3.62 and the mean for the over-achievers being 5.78. The
differences between the means was significant at the .001
level. Thus it was concluded that scholastic over- and
under-achievement do appear to be related to a student's
need for social approval as a reinforcer for his behavior.
In this case the teacher's evaluation of the student, as
reflected in the grades he "gives" him, was the source of
assumed approval or disapproval.

Studies relating tests of scholastic aptitude or

51

achievement to success in school are among the most frequent
forms of educational investigation. In contrast to studies
of "non-intellective" factors such as those reported above,
studies involving prediction of grades from standardized
tests have usually shown more significant results. However,
even the most rigorously designed aptitude and achievement
tests have systematically fallen short of precise prediction.
A few examples of research relating "intellective" factors to
educational achievement of college students are presented
here.

Wantman (52) in describing the procedures used at
the University of Rochester for predicting academic perfor-
mance in 1954, reported multiple correlations ranging from
.50 to .69 using different combinations of predictors.
Predictors included the ACE Psychological Examination and
English, reading, and mathematics tests. In attempting to
predict those students who were likely to have difficulty
academically, it was found that the multiple cut-off tech-
nique worked better than multiple regression equations. The
most useful single predictor was the Iowa English Placement
test which correlated as high as .58 with the criterion fOr ‘
one entering class. Correlations of .70 to .80 were reported
between freshman grades and grades earned in succeeding
years. Efforts to predict grades in specific courses usually
resulted in correlations around .50, seldom higher than .60.

Manuel (27) reported a series of correlations between-

52

various scholastic aptitude tests and college grade-point-
average at the University of Texas in 1955. The ACE Psych-
ological Examination, College Ability Test, and the Word-
Number Ability Test were administered to 1501 entering
freshman and zero order correlations with freshman grade-
point-averages computed. Correlations ranged between .45 to
.61. The Word-Number total score was found to have the
highest correlation with grade-point-average. Comparisons
between high school grades and admissions test scores prompted
Manuel to comment, "particularly puzzling are the cases where
grades and test scores are widely different . . . it will be
interesting to see how these students perform at the college
level when their high school marks point one way and their
test scores point another."

Pounds (39) conducted an extensive investigation into
the predictive efficiency of certain information about stu-
dents which college admissions officers have available at the
time a candidate applies for admission. The study involved
predicting the freshman grades of students in the Teachers
College, University of Cincinnati during the years 1951-1954.
The following eight factors were chosen for study as promising
predictors: (1) high school rank; (2) college interview
rating; (3) high school principal's rating; (4) L-score, ACE
Psychological Examination; (5) Q-score, ACE Psychological
Examination; (6) total-score, ACE Psychological Examination;

(7) scores on the COOperative Reading Test; and (8) scores on

53

the Cooperative English test, Mechanics of Expression. A
total of 196 students were included in the study and all of
these were enrolled in "regular" training programs, excluding
such specialized programs as art education, music education,
physical education and health education. It was found that
including students enrolled in the "special" programs ma-
terially reduced the correlations with the criterion scores.
Correlations between selected factors and grade-point-average
ranged from .42 to .55. The correlation with English scores
was .42; with principal's rating .45; with high school rank
.47; with ACE L-score .53; and with reading test scores .55.
The other factors were used in multiple regression equations.
The highest multiple correlation was .65 and involved high
school rank, ACE L-score, and an adjusted reading score as
predictors.

The highest correlation between grade-point-average
and a combination of other factors found in the present review
of research was obtained by Robinson (41) at Bradley Univer-
sity around 1950. He found the multiple correlation between
the grades of graduate students in the Department of Natural
Science and a combination of undergraduate grades, science
test scores, and the Miller Analogies Test to be .746. The
study was based on a sample of 21 graduate students. The
other correlations relating grades of graduate students to
test scores were similar to the findings for undergraduates,

that is, they were of the order of .50 to .60.

54

The theoretical background and research studies re-
ported in this chapter have necessarily been selected from
a larger volume of material, all of which could not possibly
have been reported. It was felt however, that the selections
included were sufficiently varied to represent an adequate
sample of social psychological theory and relevant educa-

tional research.

CHAPTER III
METHODOLOGY AND PROCEDURE

The basic hypotheses tested in this study involved
comparisons between three groups of students with respect to
two measured criteria. The three groups of students were
identified by their academic performance and were classified
as probationers, non-probationers, or raisers. The two cri-
terion measures were (1) the students' perceptions of the
academic expectancies held for them by members of their home
and home community group, and (2) the academic expectancies
held for them by members of their university peer group. The
criterion measures were obtained from questionnaires submitted
to the students and ratings made by members of the residence
halls personnel staff. Data relevant to the reliability and
validity of the criterion measures are presented in this

chapter along with a description of the general methodology.

Description of the Samples

The Original Population. The original population from

 

which cases could be drawn included all first term freshman
males residing in residence halls at Michigan State University
during the fall term of 1957. Thirteen students carrying

less than twelve credits were considered part-time students

and were excluded from the population. Likewise students who

dropped out of school during the fall term were not considered.

56

At the end of the fall term, there were remaining in school a
total of 1341 full-time students from the original population.
Of these students, 700 earned a grade-point-average of 2.00
or higher during the fall term and 641 earned a grade-point-
average of less than 2.00. These 641 students were placed on
academic probation by the dean of the Basic College.

The Non-Probationers. For purposes of this study, a
non-probationer was defined as a student whose grade-point-
average never fell below a 2.00 for any term of his freshman
year. Of the original group of 700 non-probationers, 240
fell below 2.00 grade-point-average during winter or spring
terms and were excluded from the sample. By the end of the
year, 45 of the original group had drOpped out of school.
Incomplete questionnaire and test data accounted for a loss
of 46 caSes. Total attrition in the non-probationer group
amounted to 331, leaving a final sample of 369.

The Probationers. For purposes of this study, a pro-
bationer was defined as a student who was placed on academic
probation at the end of the fall term and was still on pro-
bation at the end of his freshman year. Of the original
group of 641 probationers, 138 improved their grades suffi-
ciently to be removed from probation by the end of the year.
Those students were excluded from the probationer sample and
reclassified as raisers. By the end of the year, 145 of the
original group had dropped out of school. Incomplete

questionnaire and test data accounted for a loss of 18 cases.

57

Total attrition in the probationer group amounted to 301,
leaving a final sample of 340.

The Raisers. For purposes of this study, a raiser was

 

defined as a student who was placed on academic probation at
the end of the fall term and subsequently improved his grades
sufficiently to be removed from probation by the end of the
year. The raisers were derived out of the original pro-
bationer group and represented a total of 138 students. In-
complete questionnaire data accounted for a loss of 11 cases,

leaving a final sample of 127.

Development of the Instruments

The Student Questionnaire. In order to test the
hypotheses relating the expectancies held for a student by
his home group and his academic performance, it was necessary
to develop an instrument to assess quantitatively the in-
tensity of these expectancies as perceived by the student
himself. A ten item, multiple-choice questionnaire was de-
veloped for this purpose. (see Appendix)

The innocuous title, "Student Questionnaire", was
selected for the instrument purely for the sake of giving
it a name without connoting its true purpose. Multiple-
choice type items were used in order to facilitate the
analysis of the responses to each item as well as the total

score on the questionnaire.

58

In accordance with the theoretical framework of the
study, each of the ten items was designed to elicit from the
student his perception of how certain significant others in
his home and home community would feel about his going to
college and his academic performance at college. Certainly,
items dealing with all the significant others in the home
community were not included on the questionnaire. However,
those items which were included dealt with persons and groups
which are easily identified. Furthermore, to insure maximum
return, it was essential to include only those items to which
it could be reasonably assumed each student could and would
respond.

Of the 10 items included in the questionnaire, 4 dealt
with parents; 2 dealt with friends from high school; 2 dealt
with high school staff members; 1 dealt with neighbors and
friends of the family; and 1 dealt with relatives. It was
felt that items distributed thusly would yield adequate data
for purposes of this study. A typical item on the question-
naire was:

If I did not go to college, my close friends from
high school

1 would wonder quite a bit why I did not go

2 would wonder a little why I did not go

3 really wouldn't wonder at all why I did not go

4 would probably take the attitude that I was

better off if I didn't go

Each of the items was constructed in such a way that

the response numbered (1) was indicative of a high expectancy.

59

The other responses signified lowered expectancies. The
total score on the questionnaire was determined by counting
the number of (1) responses. Therefore, the total expectancy
score represented a composite picture of a student's per-
ception of the academic expectancies held for him by certain
members of his home group. The possible range of scores was
from zero to ten. Information relevant to the reliability
and validity of the Student Questionnaire is presented in
later sections of this chapter. A

In addition to the ten items in the body of the ques-
tionnaire, two additional items were included to provide data
for studying the validity of the Peer Group Questionnaire.

The Peer Group Questionnaire. In order to test the

 

hypotheses relating the expectancies held for a student by
his university peer group and his academic performance, it
was necessary to develop an instrument to assess the quality
of these expectancies. The Peer Group Questionnaire was
developed for this purpose.

~This questionnaire was a simple rating form upon which
a rating of a given student's university peer group could be
made. Criteria for rating the peer groups were established
by the investigator and became part of the written instruc-
tions to the raters.

The raters who completed the Peer Group Questionnaire
were the 68 resident assistants assigned to the eight men's

residence halls. Their instructions were to develop a

60

composite image of the group of students a given student
associated with and rate this reference group as a highly
academically motivated group or as a group who seemed to be
poorly motivated academically. The ratings involved check—
ing the tendency of a given student to associate more with
one type of group than the other. The Peer Group Question-
naire and the instructions to the raters are reproduced in
their entirety in the appendix.

Normally it would be desirable to obtain independent
ratings of the peer groups by several raters in order to get
valid ratings. However, in this instance, the resident
assistants were the only persons in a position to have the
necessary information to make the ratings. Since one of the
prime responsibilities of the resident assistants is to know
the residents of their precincts well, it was felt that they
were in a particularly advantageous position to make the peer
group ratings and that their ratings could be relied upon.
Some evidence of the validity of the peer group ratings is
reported in the section on instrument validity included in a
later section of the present chapter. The reliability of the
ratings is also presented in a later section.

The primary assumption in the development of the Peer
Group Questionnaire was that the expectancies which a group
holds for its members are reflected in the major values held
by the majority of the group members. If it is possible to

obtain a qualitative assessment of a group's values with

61

respect to a given criterion, it is possible to make infer-
ences regarding the expectancies held by the group for its
members. If the assumptions hold and the ratings are ac-
curate, then the necessary conditions for assessing the effect
of the expectancies held for an individual by his peer group

upon his academic performance are met.

Data Collecting Process

Shortly after the close of the fall term, 1957, the
grade sheets of all full-time, first term freshman residents
of the men's residence halls were inspected. The students
whose grade-point-averages were 2.00 or higher were placed
in one group and those whose averages were below 2.00 were
placed in another group. The names of these students, their
grade-point-averages, student numbers, residence halls, and
precincts were transcribed to 5x8 cards, one for each student.

During the winter term, the scores obtained by the
students on the orientation test battery administered to all
freshmen were entered on the cards.

Early in the spring term, the grade sheets of all
students in the sample were again inspected. The winter
term's grades were entered on the cards and certain cards
were discarded. The explanation of the attrition in the pOp-
ulation is explained in the section describing the samples;
therefore, it is not duplicated here.

62

Toward the middle of the spring term, the expectancy
data were cOllected. The investigator scheduled meetings
with the resident assistants and advisors of each of the
eight men's residence halls. Approximately an hour was taken
in each meeting to explain the purpose of the research and
to enlist the co-operation of the resident assistants. Con-
siderable time was spent answering questions and making sure
the resident assistants knew exactly what was expected of

them. The resident assistants were provided with a Student
Questionnaire for each student remaining in the sample and
a Peer Group Questionnaire. The names of the students had
been entered on the questionnaires and checked for accuracy
to account for the students who had Changed precincts during
the course of the year. During the meetings, considerable
emphasis was placed on the instructions for the Peer Group
Questionnaire. The resident assistants were asked if the two
types of groups as described in the instructions did, in
fact, exist in their precincts. There was general agreement
that the two groups were identifiable and that it would not
be difficult to indicate which of the two groups any given
student was identified with. The questionnaire did not
necessarily require a strong rating in one direction or the
other. Therefore, the task of making the ratings was facil-
itated since the resident assistants were free to check
intermediate positions where necessary.

All the questionnaires were distributed within the

63

space of three days and were returned to the investigator
within one week after initial distribution. The data for
each student were transcribed to the 5x8 cards and tabulation
of the results begun.

In order to gather data for reliability studies of
the instruments, the test-retest method was employed with a
sample of the original respondents. Approximately three
weeks from the time the questionnaires were originally com»
pleted, a random sample of 52 students was asked to complete
the Student Questionnaire again. Also, a group of ten resi-
dent assistants was asked to respond once again to the Peer
Group Questionnaire. The ten resident assistants re-rated
the peer groups of 168 students. The results of these re-
liability studies are reported in a later section of the
present chapter.

Following the close of the spring term, the grade
sheets of all the students remaining in the sample were
examined for the third time. The grades were entered on the
5x8 cards and the cards were sorted into the three classifi-

cations of students required for the study.

Methods of Analyzing the Data

The analysis of the data gathered for this study was
accomplished by using traditional methods of statistical

inference. Many of the data were of such a nature that only

64

the number of cases which fell in specified categories was of
interest. Chi-square tests of significants were employed in
such instances. In other situations, more precise distinc-
tions were attempted and parametric statistics were employed.
These included tests of the significance of the difference
between two means; tests of the hypothesis that two variances
were equal; and tests of the hypothesis that obtained cor-
relations were significantly different from zero. Some of
the hypotheses required one-tailed tests and others required
two-tailed tests of significance.

Primary statistical references were Walker and Lev (51),

Statistical Inference and Peters and VanVoorhis (37), Sta-

 

tistical Procedures and Their Mathematical Bases. Tables of

 

the percentile values of the chi-square distribution, "Stu-
dent's" distribution, and the F distribution were entered to
determine significance levels. The tables reported in
Walker and Lev (51:464-469) were used for this purpose.
Flanagan's table of correlations corresponding to given pro-
portions of successes in the 27 percent scoring highest and
the 27 percent scoring lowest on a continuous variable, which
is reported in Walker and Lev (51:472-475), was used in
determining the contributions of the respective items on the

Student Questionnaire to the total expectancy score.

65
Reliability of the Instruments

The Student Questionnaire. The Student Questionnaire
was designed to yield a measure of a student's perception of
the academic expectancies held for him by members of his home
group. The reliability of this instrument was studied in
two ways. First, the reliability of each of the ten items
on the questionnaire was studied separately. Secondly, the
reliability of the total score was determined. Data for these
reliability studies were obtained by re-administering the
questionnaire to 52 students randomly selected from the total
population. Three weeks elapsed between the first and second
administrations of the questionnaire.

Table 1 shows the distribution of the differences in
responses between the first and second administrations for
each of the ten items. Positive differences indicate that
the respondents shifted to a lower numbered response the
second time. Negative differences indicate selection of
higher numbered responses the second time the questionnaire
was completed. The means and standard deviations of the
ten distributions of differences were computed and are also

reported in Table 1.

66

TABLE 1. DIFFERENCES BETWEEN RESPONSES ON TWO SUCCESSIVE
ADMINISTRATIONS OF THE STUDENT QUESTIONNAIRE

 

 

 

 

Differences Standard
Item -2 ‘1 0 *1 ‘2 Mean Deviation N
1 O 4' 38 10 O +.115" .506 52
2 1 4 41 6 O .000 .519 52
3 1 13 29 9 O -.ll5—- .698 52
4 0 ll 33 8 0 -.058 ' .602 52
5 l 5 38 8 O 4.019 .572 52
6 O 5 41 5 1 +.038” .518 52
7 1 IO 32 8 1 -.O38 / .706 52
8 O 5 40 7 O +.O38 .480 52
9 O 2 41 9 O +.135 .439 52
10 1 3 4O 8 O 4.058 2 :.534 52

 

 

The standard deviations reported in Table l are measures
of the variability of the responses to the questionnaire items
on two successive administrations of the questionnaire.
Standard deviations ranging from .439 to .706 were observed.
Thus, the chances that a given student's responses did not
differ by more than one standard deviation on any one item
are two to one and the chances are nineteen to one that dif-
ferences no larger than two standard deviations occurred.
Therefore, a measure of the reliability of each item is ob-

tained by referring to the standard deviation of the

57

differences for that item. In general, the standard devia-
tions did not greatly exceed 0.50. No cases were observed
where the actual differences were +3 or -3, which would be
the maximum possible difference since each item contained
four possible responses. Only seven of the 52 students show-
ed differences of +2 or —2. For every item, the majority of
the students showed on differences in their responses.

The interpretation of Table 1 should also take into
consideration the fact that a reported standard deviation is
a form of standard error of estimate and not standard error
of measurement. We have evidence regarding the variability
of a student's responses to an item from one administration
of the questionnaire to the other, but we do not have a
measure of the scatter of the individual student's responses
around his hypothetical "true" response. The nature of the
items makes the computation of the standard errors of meas-
urement for each of the items inappropriate. However, in
situations like this, the standard error of measurement (if
it can be computed) is smaller than the standard error of
estimate. Therefore, any inferences drawn from Table 1 re-
garding the reliability of the items, if in error, will be in
the direction of under-estimating the true reliabilities.
Several authors have discussed the vagaries of the standard
error of estimate and the standard error of measurement as
measures of instrument reliabilities and the suggestions of

Cureton (11) and Peters and VanVoorhis (37) are followed in

68

presenting the reliability of the total expectancy score.’
The total expectancy score derived from the Student
Questionnaire was determined by counting the number of (1)
responses. Expectancy scores ranged from zero to ten. The
bivariate distribution of the total expectancy scores ob-
tained by the 52 students on the two administrations of the

questionnaire is reported in Table 2.

TABLE 2. DISTRIBUTION OF EXPECTANCY SCORES ON TWO SUCCESSIVE
ADMINISTRATIONS OF THE STUDENT QUESTIONNAIRE

;

 

Second Expectancy Scores Obtained on First Administration
Admin. f2
Scores 0 1 2 3 4 5 6 7 8 9 10

 

 

10 1 l 2 4
9 l 2 3
8 2 2 2 6
7 1 l 5 7
6 1 1 1 l 2 6
5 2 1 1 1 5
4 1 4 1 1 7
3 l 3 l 5
2 2 3 5
1 1 l 1 3
0 1 1
fl 1 4 2 14 4 6 2 10 7 0 2 52

 

 

69

The Pearson Product-Moment correlation between the two
sets of obtained scores reported in Table 2 was found to be
.866. This reliability coefficient is spuriously low as a
measure of the true reliability since it was based on the
correlation of two sets of fallible measures. Cureton (11:
685) advocates that reliability of an instrument be reported
in terms of the index of reliability. This index is the cor-
relation between a set of fallible (obtained) measures and
the hypothetical set of true scores for the individuals in
the group. True scores have been defined by Peters and Van
Voorhis (37:204) as ”the scores yielded by an instrument ap-
plied an infinite number of times and the average taken so
that the scores have been completely stabilized." This hypo-
thetical distribution of obtained scores is impossible to
secure; however, Peters and VanVoorhis (37:205) have shown
that the correlation between a set of fallible scores and a
set of true scores is the square root of the reliability
coefficient of the instrument. Therefore, the index of re-
liability of the total expectancy score obtained on the Student
Questionnaire was computed and found to be .931. According
to Walker (50:25), an index of reliability of .931 is
sufficiently high for the study of group behavior.

The standard error of measurement of the total ex-
pectancy score derived from the questionnaire was computed
and found to be .94. Thus, the chances are two to one that

an individual's obtained score did not differ from his true

70

score by more than .94 points and the chances are nineteen

to one that differences no larger than 1.88 occurred. The
computation of the standard error required a measure of the
variance in the sample. For this purpose, the variance of

the two distributions was pooled and used as the estimate of
the population variance. The standard deviation derived from
the pooled variances was 2.58 and the combined means was found
to be 5.14.

An indication of the representativeness of the sample
used for the reliability studies of the Student Questionnaire
was obtained by comparing the mean and standard deviation
(pooled) of the sample with the mean and standard deviation
of the total population. (Treatment of the data for the
total population is reported in Chapter IV.)

The means of the sample and the population were 5.14
and 5.17 respectively. The critical ratio was .083, which
is well below the critical ratio of 1.96 required to reject
the hypothesis that the means are different at the five
percent level of confidence. It may be concluded that there
was no significant difference between the two means.

The standard deviation of the sample and the pOpulation
were 2.58 and 2.62 respectively. For the hypothesis that the
standard deviations are different to be rejected at the five
percent level of confidence, the value of the statistic, F,
must equal or exceed 1.44 when there are 835 degrees of

freedom in the numerator and 51 degrees of freedom in the

71

denominator. The obtained value of F was 1.03 which is below
the value required to reject the hypothesis that the standard
deviations are different at the five percent level of con-
fidence. It may be concluded that there was no significant
difference between the two standard deviations.

The Peer Group Questionnaire. The Peer Group Ques-
tionnaire was designed to assess the quality of the academic
expectancies held for an individual by members of his univer-
sity peer group. The reliability of this instrument was)
studied by having a group of ten resident assistants respond
again to the questionnaire three weeks after they made their
original ratings. The ten resident assistants re-rated the
peer groups of 168 students. Table 3 shows the distribution
of the differences between the original ratings and the re-
ratings of the peer groups. For purposes of analysis, the
rating "Associates mostly with Group A" was assigned position
1; "Tends to associate with Group A" was assigned position 2;
"Tends to associate with Group B" was assigned position 3;
and "Associates mostly with Group B" was assigned position 4.
Positive differences indicate a shift in the direction from
"B" towards "A", and negative differences a shift from "A"
towards "B". These numbers do not represent equal-interval'
measurements, and, therefore, may be thought of as ordinal

only.

72

TABLE 3. DIFFERENCES BETWEEN RATINGS ON TWO SUCCESSIVE
ADMINISTRATIONS OF THE PEER GROUP QUESTIONNAIRE

 

 

 

 

Differences Standard
Mean N
-2 -l 0 +1 +2 Deviation
l 11 122 31 3 .14 .570 168

 

 

The standard deviation of .570 is a measure of the
variability between the original ratings of the peer groups
by the resident assistants and their re-ratings of the same
groups three weeks later. The chances that two successive
(ratings did not differ by more than .570 are two to one and
the chances are nineteen to one that differences no larger
than 1.14 occurred. No cases were observed where the actual
differences were +3 or -3, which would be the maximum possible
differences since the ratings were made with respect to four
classifications. Only four of the 168 pairs of ratings
showed differences of +2 or -2. Seventy three percent of

the ratings were exactly the same both times.

Validity of the Instruments

One of the limitations of this study was the lack of
criteria for validating the instruments. If adequate criteria
had been available if would not have been necessary to develop

the instruments at all. Instead, the criterion measures

73

themselves could have been used. Therefore, the validity of
the instruments, in the strict sense of the word, was un-
certain. The best that could be done was to draw on the
theoretical framework of the study and attempt to develop
instruments in accordance with the requirements of the
theoretical framework. The theoretical framework is treated
in Chapter I and the rationale for the development of the
instruments is included in Chapter III. While it was im-
possible to study systematically the validity of the in-
struments against true outside criteria, some supportive
evidence of their validity was obtained.

A necessary (but not sufficient) condition for valid
measurement is reliable measurement. On the basis of the
reliability studies, it was felt that the requirements of
reliability of measurement were adequately met. Data re-
garding reliability of the instruments are reported in
Tables 1, 2, and 3.

The Student Questionnaire. Cureton (11:643-649) has

 

discussed the problem of the validity of an instrument such
as the Student Questionnaire. His approach to the problem
involves establishing that the summary scores or total scores
on an instrument do, in fact, represent "measurements" and
not simple "appraisals". "Appraisals" are described as the
aggregate value of the responses to a series of items which
may or may not be positively related to one another or the

summary score. On the other hand, the term "measurement" is

74

more restrictive. Quoting Cureton, "unless the item . . .
scores are themselves related, in the sense that they vary
consistently rather than independently, the summary (total)
score is an appraisal but not a measurement." Measurement
requires (1) an approximately linear score continuum, and (2)
a set of approximately equal score units. For clarity,
Cureton's remarks on these two requirements for measurement

are quoted.

The linearity of the continuum need not be per-
fect, and the score units need not be exactly equal.
But the continuum must at least exist, and every unit
(item or observation score) must have some relation
to it. This means that when we score a set of . . .
items or criterion observations for each of the indi-
viduals of an appropriate group, every item score must
exhibit some significant positive correlation . . .
with the total score.

The particular sense in which they (item scores)
must be equivalent is that each one of them must be
an equivalent indicator of the total summary score.
Ideally each item must have the same validity as every
other, as an indicator of the summary score, and these
validities must all be above zero. In practice we
are likely to accept evidence that each item score
possesses some significant positive validity as a work-
ing approximation to the ideal.

Not only must every item score be related to the sum-
mary score, but the correlations between the item scores and
the summary scores should be approximately equal. Quoting
Cureton again, "no one of the essential conditioning factors

of the test (instrument) or observational situation may be

of overriding importance, and neither may any one reaction of

75

the subject to the test (instrument) or observational con-
ditions." If the responses to an instrument fulfill the re—
quirements specified above, we have what may be appropriately
called a quasi—measurement. According to Cureton, "quasi-
measurements may for practical purposes be treated as
measurements. They may be added, subtracted, averaged,
correlated, etc."

Cureton suggests further that whenever the sum of
the item scores on an instrument can be defended as a quasi-
measurement, "it (the sum) is a quasi-measurement of 'what-
ever', in the reaction-systems of the individuals, is evoked
in common by the . . . items as presented . . ." Referring
once again to the theoretical framework of this study and
the rationale for the development of the Student Question-
naire, it can be seen that each of the ten items was, in
fact, designed for the express purpose of evoking a student's
perception of the academic expectancies held for him by
members of his home and home community. Thus, it may be
presumed that " 'whatever' . . . is evoked in common by the
. . . items" is quite likely to be related to the expectan—
cies as defined. It remains to be demonstrated, however,
that the scores on the ten items possess some degree of
consistency and that they measure to some substantial degree
the same thing. Once this has been demonstrated, it may be
assumed that the expectancy data derived from the Student

Questionnaire did, in fact represent "measurements" and that

76

confidence may be placed in the total expectancy score as an
internally consistent measure of what it purports to measure.
It was reported earlier in the present chapter that
the total expectancy scores derived from the Student Question-
naire distributed themselves along a continuum from zero to
ten with a mean of 5.17 and a standard deviation of 2.62.
In order to determine the contribution of each of the ten
questionnaire items to the total expectancy score, it was
necessary to obtain the correlations between the responses
to the items and performance on the whole questionnaire. A
method devised by Flanagan (16) was used for this purpose.
The 836 total expectancy scores derived from the questionnaire
were sorted into two groups comprised of the upper 27 percent
of the scores and the lower 27 percent of the scores. The
middle 46 percent were put aside. For each of the ten items,
the proportions of (1) responses in each of the high and low
27 percent groups was obtained. Flanagan's table of product-
moment correlations, which is reported by Walker and Lev
(51:472-475) was entered and the required coefficients

noted. The results are shown in Table 4.

77

TABLE 4. CORRELATIONS BETWEEN ITEM RESPONSES AND TOTAL EXPEC-
TANCY SCORES ON THE STUDENT QUESTIONNAIRE

 

 

 

Item 1 2 3 4 5 6 7 8 9 10 Median
Correlation
with .67 .72 .73 .80 .67 .70 .62 .82 .61 .68 .69

Total Score

 

 

The correlations reported in Table 4 are all positive
and may be assumed to be statistically significant. (No
method of computing the standard error of a correlation de-
rived from Flanagan's table was known to the investigator;
however, the standard error of the traditional Pearson
Product-Moment correlation derived from a sample of 836
cases is .03. Since all the correlations reported in Table
4 were at least twenty times .03, the assumption of signi-
ficance seemed justified.) The median correlation was found
to be .69. The range was from .61 to .82. While the re-
spective correlations differed somewhat, it was assumed that
the differences were not so great that they did not represent
a working approximation to the ideal which is that no dif-
ferences would be observed between the correlations.

As a result of the item score-total score correlation
studies, it was concluded that the Student Questionnaire was
measuring something. It was postulated that it was providing

a valid measurement of a student's perceptions of the academic

78

expectancies held for him by members of his home group.

The foregoing rather laborious treatment of the valid—
ity of the Student Questionnaire was not intended to be a
substitute for a more desirable validation of the instrument
against an acceptable outside criterion, but under the cir-
cumstances that was impossible. According to Cureton (12)
the validity data are simply "mathematical rationalizations",
and are, at best, only suggestive. But it was felt that if
it were possible to demonstrate that the instrument met the
criterion of internal consistency and possessed the character-
istics for measurement, it should be reported as such and
included in the discussion of instrument validity.

The Peer Group Questionnaire. It was possible to
obtain some measure of the validity of the Peer Group Ques-
tionnaire against an outside criterion. The responses to
items 11 and 12 included at the bottom of the Student Ques-
tionnaire provided a convenient source of information against
which to validate the peer group rating made by the resident
assistants. Item 11 dealt with the student's knowledge of
the grades earned by his close friends in the precinct.

Item 12 dealt with the student's perception of the grades he
felt his close friends expected him to get.

A random sample was drawn from the population by
pulling all the cards of students whose student numbers
ended in 7 or 1. Since student numbers are assigned con-

secutively as students apply for admission to the university,

79

it was felt that this method would provide a random sample of
the population. The procedure resulted in a sample of 158
cases. Three cases were unusable since the students did

not respond to one or both of the items, leaving 155 cases
for the validity study. Inspection of the 155 cases revealed
that cases had been drawn from 58 of the 68 precincts in the
eight men's residence halls. The sample was deemed adequate
for purposes of the study.

The rationale for the development of the Peer Group
Questionnaire assumed that there would be a correspondence
between the academic performance of a student's peer group
and his perception of the academic expectancies which his
peers held for him. Data relevant to this assumption are
presented in Table 5. The numbers, 1, 2, 3, and 4, refer
to the numbered responses in items 11 and 12; and the numbers
in the cells indicate the observed frequencies of the pairs

of responses.

80

TABLE 5. RELATIONSHIP BETWEEN ACADEMIC PERFORMANCE OF PEER
GROUP AND EXPECTANCIES HELD FOR GROUP MEMBER AS PERCEIVED
BY MEMBER

 

 

 

 

 

Perceived Peer Group Performance f
E
Expectancies l 2 3 4

4 O 2 3 6 11

3 l 8 45 8 62

2 2 23 20 5 50

l 6 14 11 l 32

fP 9 47 79 20 155

 

Inspection of Table 5 shows that the cases tended to
accumulate in the lower-left and upper-right hand quadrants.
This accumulation suggests a positive relationship between
the variables. This relationship was tested for its sig-
nificance by using the chi-square test of independence in
contingency tables. Since the theoretical frequencies in
some of the cells were found to be less than five, both
variables were dichotomized between 2 and 3, and chi-square
was computed for the 2x2 contingency table. The resulting
value of chi-square was 26.6 (chi-square .999 : 10.8, df : 1).
The hypothesis of independence was rejected beyond the .001
level; and, it was concluded that a positive relationship

existed between the academic performance of a student's peer

81

group and his perception of the academic expectancies which
his peers held for him. This conclusion is tenable so long
as the responses to item 11 are assumed to be valid. Since
knowledge of one's college friends' grades is quite common
among undergraduates, it was assumed that the students were,
in the main, able to report fairly accurately the general
level of their friends'grades.

A measure of the amount of the relationship demonstrat-
ed in Table 5 was obtained by computing the tetrachoric
correlation from the 2x2 contingency table resulting from
dichotomizing the data. The correlation was found to be .647.

The resident assistants were instructed to rate the
peer groups of the subjects in the study in terms of the
academic motivations exhibited by the majority of the members
of the respective groups. This involved a subjective judgment
on the part of the resident assistants. It was felt by the
investigator that there would probably be a relationship
between the resident assistants' rating and the quality of
the grades earned by the various peer groups. The responses
to item 11 provided a criterion against which to test the
validity of the peer group ratings made by the resident
assistants. Item 11 was particularly valuable for this pur-
pose, not only because it dealt with a given student's
knowledge of the grades of his peers, but, perhaps, what is
more important, it permitted the student himself to decide

who his peers really were. Thus, in a way, we have a check

82

on the correspondence between a student's perception of who
his peers were and the resident assistant's perception of
who the student's peers were. Table 6 presents the data
comparing peer group ratings and the students' responses to
item 11. The numbers, 1, 2, 3, and 4, refer to the numbered
responses in item 11 and the positions of the ratings on the
Peer Group Questionnaire. The same positioning scheme for
the ratings was used in the reliability study of the ratings.
A rating of 1 indicates that the rater thought the student's
peer group was highly motivated academically and ratings 2,
3, and 4 indicate progression toward the lower end of the

continuum.

TABLE 6. RELATIONSHIP BETWEEN RATINGS OF PEER GROUP AND GRADES

EARNED BY PEER GROUP AS PERCEIVED BY GROUP MEMBER

 

 

 

 

 

Peer Group Peer Group Ratings f

P
Performance 1 2 3 4

4 3 8 3 6 2O

3 15 23 19 22 79
2 14 13 14 6 47

1 6 l 2 O '9

fR 38 45 38 34 155

 

 

83

Analysis of Table 6 reveals that the cases tended to
accumulate in the lower-left hand quadrant and the upper-right
hand quadrant over and above what might be normally expected
if no relationship existed. It was noted that peer group
ratings 2 and 3 contributed little to the relationship. This
is in keeping with the commonly found situation where inter-
mediate values of a variable are frequently less valid than
extreme values. Due to the nature of the Peer Group Question-
naire and the complexity of the dynamics being rated it may
be assumed that the ratings reported on it were particularly
prone to be less valid in the intermediate range than at the
extremes. Thus, it was decided to employ the method of
widely spaced groups to overcome this difficulty.

The responses on item 11 were dichotomized between 2
and 3, and peer group ratings 2 and 3 were excluded. The re-
maining data were arranged in a 2x2 contingency table of 72
total cases. Chi-square was computed and found to be 9.58
(chi-square .995 . 7.8, df . 1). The hypothesis of indepen-
dence was rejected beyond the .005 level; and, it was con-
cluded that a positive relationship existed between the
academic motivations of a student's peer group as perceived
by the resident assistant and the academic performance of the
student's peer group as reported by him.

A measure of the amount of the relationship demon-
strated in Table 6 was obtained by computing the tetrachoric

correlation from the 2x2 contingency table. The method

84

described by Peters and VanVoorhis (37:375-382) for determining
tetrachoric correlation from widespread classes was employed.
The correlation was found to be .371.

On the basis of these studies, it was concluded that
the ratings reported by the resident assistants on the Peer
Group Questionnaire, while inappropriate for strict individ—
ual interpretation, could be relied upon for study of group

tendencies.

CHAPTER IV

ANALYSIS OF THE DATA

The data derived Irom the Student Questionnaire and
the Peer Group Questionnaire were of such a nature that
cOmparisons between probationers and non-probationers and
comparisons between probationers and raisers could readily
be made. For the most part, the data were arranged in con-
tingency tables and appropriate statistical tests applied
to determine the significance of the observed relationships.
The five percent level of confidence was established as the
requirement for statistical significance.

The comparisons that were made between the various
groups are reported in the first sections of the chapter and
supplementary analyses involving correlations with academic
performance, inter-instrument correlations, and the relation-
ship between academic performance and a combination of peer
expectancies and home expectancies are presented in later

sections.

Comparisons between Probationers and Non-Probationers
with Respect to Their Perceptions of the Academic
Expectancies Held for Them by Members of Their
Home and Home Community Groups

Comparisons with Respect £2 Each pf the Items pg the

 

 

Student Questionnaire. The Student Questionnaire was con-

structed primarily to yield a total expectancy score which

86

could be used in comparing the academic performance of the
three groups of students identified for this study. However,
it was felt that it would be of interest to analyze the
students' responses to each of the ten items separately.
Table 7 shows the distributions of the responses of the pro-
bationers and the non-probationers for each item. Chi-
square was computed for each of the ten contingency tables
to determine the significance of the relationships, if any,
between the students' responses and their academic status.
The numbers, 1, 2, 3, and 4, refer to the numbered responses
to the questionnaire items. As they proceed from left to
right, they are indicative of lowered expectancies. Some of
the theoretical frequencies in the ”4” cells were found to
be below five. In this case the observed frequencies were
added to cell "3" in order to compute chi-square. Two de-
grees of freedom (df) in the df column indicates this
grouping. The significance level indicates the probability
with which the distributions could have occurred by chance

from an uncorrelated population.

TABLE 7.

87

COMPARISONS BETWEEN PROBATIONERS AND NON-PROBATIONERS
WITH RESPECT TO THEIR RESPONSES TO EACH OF THE ITEMS ON THE
STUDENT QUESTIONNAIRE

 

 

 

 

Chi- Si ifi-
Item Groups Responses N df Square ngce
1 2 3 4 Level
1 Probationer 216 71 48 5 340 2 6.7 .05
Non-Probationer 266 65 36 2 369
2 Probationer 214 116 8 2 340 2 5.3 .10
Non-Probationer 262 98 7 2 369 (1 5.2 .025)
3 Probationer 141 128 67 4 340 2 1h,o .001
Non-Probationer 200 122 45 2 369
Probationers 147 130 61 2 340
4 Non-Probationers 227 110 30 2 369 2 27-7 '001
Probationers 200 99 41 0 340
5 Non-Probationers 263 68 37 l 369 2 13°2 '005
Probationers 199 99 37 5 340
6 Non-Probationers 255 65 47 2 369 2 13°2 '005
7 Probationers 162 98 6O 20 340 3 14.3 .005
Non-Probationers 225 89 40 15 369 (2 14.2 .001)
Probationers 181 120 38 1 340
8 Non-Probationers 228 113 27 1 369 2 6'2 '05
Probationers 21 92 159 68 340
9 Non-Probationers 135 139 60 35 369 3 147.0 '001
Probationers 24 97 159 60 340
10 Non-Probationers 136 148 63 22 369 3 147'0 '001

 

Analysis of Table 7 shows that for every questionnaire

item the cases tended to accumulate in the lower-left and

upper-right hand quadrants over and above chance occurrence.

This suggests that there was a positive relationship between

88

the students' responses and their academic status. It can be
said with considerable certainty that the questionnaire items
did differentiate significantly between probationers and non-
probationers as a group. In all cases but one (item 2), the
probability with which the observed distributions could have
occurred by chance was less than .05. (The right hand tails
of the distributions of responses to item 2 were erratic and
the frequencies small. By lumping responses 2, 3, and 4 to-
gether and computing chi-square with one degree of freedom,
significance at the .025 level was obtained. This finding

is reported parenthetically in the table.)

Items 1 and 2 dealt with the student's perceptions of
the attitudes and feelings of his parents regarding his going
to college and his attainment of a college degree. The items
are reproduced here for ease of reference.

1. My parents

(1) have planned since I was quite young for me to

go to college

(2) have planned since I was in high school for me

to go to college

(3) have never really planned for me to go to college

4 would rather have me get a job or work at home
than go to college
2. My parents

(1) would be extremely disappointed if I did not

get a college degree

(2) would be mildly disappointed if I did not get

a college degree
(3) really don't care whether I get a college

degree or not
(4) would prefer that I not get a college degree

89

There was a tendency for the non-probationers to in-
dicate that their parents had planned over a longer period
of time for them to go to college. Similarly, the non-
probationers tended to feel that their parents would exhibit
greater disappointment over their failure to attain a college
degree.

Items 3, 4, and 8 were worded the same, except that the
stems were different. Each of the three items dealt with the
student's perception of what others would think if he did not
go to college. Item 3 dealt with neighbors and friends of
the family; item 4 dealt with high school friends; and item

8 dealt with relatives. The items were:

3. If I did not go to college, the neighbors and
friends of the family
1 would wonder quite a bit why I did not go
2 would wonder a little why I did not go
3 really wouldn't wonder at all why I did not go
4 would probably take the attitude that I was
better off if I didn't go

4. If I did not go to college, my close friends from
hi h school
1 would wonder quite a bit why I did not go
2 would wonder a little why I did not go
3 really wouldn't wonder at all why I did not go
4 would probably take the attitude that I was
better off if I didn't go

8. If I did not go to college, my relatives
1 would wonder quite a bit why I did not go
2 would wonder a little why I did not go
3 really wouldn't wonder at all why I did not go
4 would probably take the attitude that I was
better off if I didn't go

90

Both probationers and non-probationers tended to agree
that their relatives would wonder more than their high school
friends or neighbors if they did not go to college. A total
of 409 students (probationers and non-probationers combined)
indicated that their relatives would wonder "quite a bit" if
they did not go to college, while 374 and 341 respectively
indicated that their high school friends and their neighbors
would wonder "quite a bit" at their not going to college.

Despite the fact that the students ranked relatives
highest in their prOpensity to wonderment about their not
going to college, it was the propensity to wonder of the
students' high school friends that proved to differentiate
probationers from non-probationers most significantly. Since
there were two degrees of freedom in the computation of the
three chi-squares, the rank order of the chi-squares provides
a Clue to the relative importance of the expectancies held
by these three groups of "significant others". High school
friends ranked first (chi-square - 27.7); neighbors and friends
of the family ranked second (chi-square - 14.0); and relatives

ranked third (chi-square : 6.2).

Items 5 and 6 dealt with the amount of encouragement
the student felt he received from high school staff members
regarding his going to college. Item 5 involved teachers and

item 6 involved principals and counselors. The items were:

91

5. My high school teachers
1 strongly encouraged me to go to college
2 mildly encouraged me to go to college
3 didn't encourage me at all to go to college
4 discouraged me from going to college

6. My high school principal or counselor
1 strongly encouraged me to go to college
2 mildly encouraged me to go to college
3 didn't encourage me at all to go to college
4 discouraged me from going to college

The distributions of responses to items 5 and 6 were
quite similar. In both instances, the non-probationers seem
to have been encouraged to go to college more strongly. Of
particular interest is the fact that more than half of the
probationers felt they had been strongly encouraged to go
to college by high school staff members.

Item 7 dealt with how many of a student's high school
friends were going or contemplating going to college. Chi-
square, calculated with three degrees of freedom, was sig-
nificant at the .005 level. When responses 3 and 4 were
lumped together, significance at the .001 level was found.
The item read as follows:

7. Of my close friends from high school
(1) most of them are now going or planning to
go to college a
(2) quite a few of them are now going or planning
, to go to college
(3) a few of them are now going or planning to
go to college

(4) practically none of them are now going or
planning to go to college

92

More than half of the non-probationers reported that
most of their high school friends were going to college and
less than half of the probationers did so. Twenty-five more
probationers than non-probationers indicated that few or none
of their close high school friends were going to college.

Items 9 and 10 involved the expectancies which father
and mother held for their son in terms of actual grade-point-
averages. The relationship of the expected grade-point-
averages to the numbered responses may be determined from

the items themselves.

9. My father

(1) expects me to get at least a 3.00 grade-point
average

(2) expects me to get at least a 2.50 grade-point-
average

(3) expects me to get at least a little better
than a 2.00 grade-point-average

(4) is satisfied as long as I make my 2.00 grade-
point-average

10. My mother

(1) expects me to get at least a 3.00 grade-point-
average

(2) expects me to get at least a 2.50 grade-point-
average

(3) expects me to get at least a little better
than a 2.00 grade-point-average

(4) is satisfied as long as I make my 2.00 grade-
point-average -

The non-probationers clearly indicated that their
parents held higher academic expectancies for them than did
the probationers. Nevertheless, 121 probationers indicated

that one or both parents expected them to earn a 2.50 grade-

point-average or higher.

93

Comparisons with Resppct §g_the Total Expectancy Scores
Derived from the Student Questionnaire. The total expectancy
score derived from the questionnaire was intended to yield
a composite measure of a student's perception of the academic
expectancies held for him by members of his home group. Table
8 shows the distributions of the total expectancy scores for
the probationers and the non-probationers. A score of 10
indicates high expectancies and a score of 0 indicates very

low expectancies as perceived by the student.

TABLE 8. COMPARISONS BETWEEN PROBATIONERS AND NON-PROBATIONERS
WITH RESPECT TO TOTAL EXPECTANCY SCORES ON THE STUDENT
QUESTIONNAIRE

 

 

 

 

Academic Total Expectancy Scores
Sig.
O 1 2 3 4 5 6’ 7 8 9 10 N Mean C-R- Level
Status
Non-
Prob'ners 13 14 17 20 43 5O 44 45 61 26 36 369 5.69

6.81 .0005
Prob'ners 19 28 31 44 49 46 53 29 32 6 3 340 4.44

 

 

The mean expectancy score for the non-probationers was
5.69 and the mean for the probationers was 4.44. The critical
ratio test was performed to test the hypothesis that the mean
of the non-probationers was higher than the mean of the pro-
bationers. The value of the critical ratio, 6.81, was found

to be significant beyond the .0005 level of confidence

94

(t.9995 - 3.29, one-tailed test). This indicates that, as a
group, probationers and non-probationers were differentiated
from each other with respect to their total expectancy scores
derived from the Student Questionnaire.

In order to test further the relationship between
academic status and total expectancy scores, it was decided
to stratify the samples at three separate levels of scholastic
aptitude. The students‘ scores on the American Council on
Education Psychological Examination (ACE) were available as
measures of scholastic aptitude. The scores were reported
in MSU derived score units. ACE scores of 5 and 6 were
labeled as "average"; those above 6 as "above average"; and
those below 5 as "below average". These dividing lines are
in keeping with common practice at the university counseling
center in interpreting ACE scores to students. In general,
44 percent of the students would be expected to have "average
scores" and 28 percent respectively would be expected to have
"above average" and "below average" scores. Table 9 shows
the findings that resulted from this stratification by ACE

SCOI’ES .

95

TABLE 9. COMPARISONS BETWEEN-PROBATIONERS AND NON-PROBATIONERS
WITH RESPECT TO TOTAL EXPECTANCY SCORES--STRATIFIED BY
SCHOLASTIC APTITUDE LEVEL

 

 

Academic Total Expectancy Scores

 

 

Sig.
N Mean C.R.
Status 0 1 2 3 4 5 6 7 8 9 10 Level
ABOVE AVERAGE SCHOLASTIC APTITUDE
Non-
Prob'ners 6 5 4 9 16 16 16 24 31 13 25 165 6.62
4.29 .0005
Prob'ners O 4 2 5 4 3 8 6 2 l O 35 4.83
AVERAGE SCHOLASTIC APTITUDE
Non-
Prob'ners 2 6 12 8 23 3O 22 19 24 10 8 164 5.60
3.65 .0005
Prob'ners 7 8 11 21 28 23 22 l6 l7 3 1 157 4.69
BELOW AVERAGE SCHOLASTIC APTITUDE
Non-
Prob'ners 5 3 1 3 4 5 6 2 6 3 3 41 5.12
2.00 .025

Prob'ners 12 16 18 l8 17 20 23 7 13 2 2 148 4.07

 

 

Examination of Table 9 reveals that at each level of
scholastic aptitude the mean for the non-probationers was
higher than the mean for the probationers. In each case the
critical ratio was significant beyond the .025 level of con-
fidence (t .975 a 1.96, one-tailed test). Thus at each of
the three levels of scholastic aptitude, the total expectancy
score derived from the Student Questionnaire was found to
differentiate between probationers and non—probationers as a

group.

Comparisons Between Probationers and Non-Probationers
with Respect to the Ratings of Their Peer Groups
Made by the Resident Assistants
Each of the students in the study had his university

peer group rated by the resident assistant in charge of his
residence hall precinct. The ratings indicated the resident
assistant's opinion of the academic motivations of the group
of students with whom a given student most closely associated.
A rating of 1 indicated that the student associated mostly
with highly motivated students, and ratings of 2, 3, and 4
indicated progression downward. Data on the peer group

ratings are reported in Table 10.

TABLE 10. COMPARISONS BETWEEN PROBATIONERS AND NON-
PROBATIONERS WITH RESPECT TO PEER GROUP RATINGS

 

 

 

 

Academic Peer Group Ratings
N
Status 1 2 3 4
Probationers 26 . 86 97 131 340
Non-Probationers 153 119 57 40 369

 

 

The frequencies presented in Table 10 tended to accumu-
late in the lower left and upper-right hand quadrants. Chi-
square for the table was found to be 153. This value is
significant well beyond the .001 level of confidence with

three degrees of freedom. Thus, as a group, the probationers

97

were differentiated from the non—probationers with respect

to the peer group ratings.

.Comparisons between Probationers and Raisers with
Respect to Their Perceptions of the Academic
Expectancies Held for Them by Members of
Their Home and Home Community Group

Comparisons with Respect £2_Each 93 the Items gg_the

 

 

Student Questionnaire. Data regarding the responses of the

 

probationers and the raisers are presented in Table 11. The
numbers, 1, 2, 3, and 4 refer to the numbered responses to
the items on the Student Questionnaire. As they proceed
from left to right, they are indicative of lowered expect-
ancies. Some of the theoretical frequencies in the right
hand tails of the distributions were quite small, therefore,
adjacent cells were grouped for purposes of computing chi-
square. The number in the degrees of freedom (df) column
indicates the extent of this grouping. The significance
level indicates the probability with which the distribution

could have occurred by chance from an uncorrelated population.

TABLE 11.

98

COMPARISONS BETWEEN PROBATIONERS AND RAISERS WITH

RESPECT TO THEIR RESPONSES TO EACH OF THE ITEMS ON THE

STUDENT QUESTIONNAIRE

 

 

Chi-

 

 

 

Item Groups Responses N df 31%-
l 2 3 4 Square Level
1 tasters 2%: a at 2 its; 2
2 3:22:11?“ 2:211?) 2 3 $23 1 .57 ~50
a sham 1%; 2; i as; 2
4 12:22am 12713: s 3 s? 2
s mam 288 a it 2 s: 2
6 teams 122 a i: f it; 2
7 Probationers 162 98 60 \20 340 9.1 .05
Raisers 79 29 12 7 127 (2 8.3 .02)
8 $33233...“ 12313? I? 3 $33 2 “'7 '95
9 Eii‘éiiimers it 75123 2? $33 3 '2-8 ~005
10 tassel”: 2:“: 2:12; a :23 3
In general, the frequencies reported in Table 11 tended

to accumulate in the lower-left and upper-right hand quadrants

of the various contingency tables.

(Since the N's are quite

different, the relationships are not readily detected by

inspection; however, all except item 8 were positive albeit
low in some cases.) This finding parallels the findings
obtained when probationers and non-probationers were compared
with respect to their responses to each of the items on the
Student Questionnaire (Table 7). However, none of the re-
lationships demonstrated in the present table was found to

be as significant statistically as those reported in Table 7.

Items 6 and 8, dealing with high school principals and
relatives respectively, were clearly ineffective in differ-
entiating between probationers and raisers. The probability
that the observed distributions of responses to these two
items could have occurred by chance was greater than .90,
which is so near unity that the hypothesis of independence
could not possibly be rejected.

The responses to items 3 and 5, dealing respectively
with neighbors and high school teachers, could have been
explained by the operation of chance factors or accidents of
sampling with probability of .75. That is, there are three
chances in four that the observed relationships could have
occurred in samples from uncorrelated populations despite the
fact that in the samples the variables were slightly positively
correlated.

The relationship observed with respect to the responses
to item 2, which dealt with parents' disappointment at the
student's not getting a college degree, was significant at

the .50 level. Thus, it could have been explained by chance

100

one time out of two.

Item 1 dealt with how long a student's parents had
planned for his going to college. Item 4 involved the stu-
dent's perception of how much his high school friends would
wonder at his not going to college. The chi-squares for these
distributions were both significant at the 25 percent level.
While this level of significance is inadequate to reject the
statistical hypothesis of independence with any great con-
fidence, it may be thought of as being somewhat suggestive
that a true relationship may have existed between the
variables.

Of the first eight items on the questionnaire the
responses to item 7 were the only ones found to differentiate
between probationers and raisers with any degree of statis-
tical significance. On item 7 the student reported how many
of his high school friends were going or planning to go to
college. Chi-square for this contingency table was 9.1 with
three degrees of freedom and was significant beyond the .05
level.

Items 9 and 10 involved the expectancies which father
and mother respectively held for their son in terms of actual
grade-point-averages. Probationers and raisers were signif-
icantly differentiated as a group with respect to their
responses to these items. Of note was the fact that the
academic expectancies held by the fathers seemed to be more

important than the expectancies held by the mothers in

101

differentiating between probationers and raisers. Chi-square
for item 9 (fathers) was 12.8 with three degrees of freedom
and chi-square for item 10 (mothers) was only 7.8 with three
degrees of freedom.

Comparisons with Respect 39 the Total Expectancy Scores

 

Derived from the Student Questionnaire. Table 12 shows the
distributions of the total expectancy scores derived from the
Student Questionnaire for the probationers and raisers.
Scores of 10 indicate high expectancies and scores of 0 in-

dicate very low expectancies.

TABLE 12. COMPARISON BETWEEN PROBATIONERS AND RAISERS WITH RE-
SPECT TO TOTAL EXPECTANCY SCORES ON THE STUDENT QUESTIONNAIRE

 

 

Academic Total Expectancy Scores N Mean C.R. gig 1
Status 0 l 2 3 4 5 6 7 8 9 10 eve

 

Raisers 3 11 ll 15 17 11 22 18 9 5 5 127 4.93
1
Prob'ners 19 28 31 44 49 46 53 29 32 6 3 340 4.44

 

 

The mean expectancy score for the raisers was 4.93 and
the mean for the probationers was 4.44. The critical ratio
test was performed to test the hypothesis that the mean of
the raisers was larger than the mean of the probationers.

The value of the critical ratio, 1.90, was found to be sig-
nificant beyond the .03 level of confidence (t .97 = 1.88,

102

one tailed test). It was concluded that as a group, pro-
bationers and raisers were significantly differentiated from
each other with respect to their total expectancy scores.

In order to test further the relationship between
total expectancy scores and academic performance, it was
decided to stratify the samples by scholastic aptitude level.
Scores on the American Council on Education Psychological
Examination were used for the stratification of the groups
into three levels. ACE scores of 5 and 6 were called "av-
erage"; above 6 was called "above average"; and below 5 was

called "below average”. Table 13 shows the results of this

stratification.

TABLE 13. COMPARISONS BETWEEN PROBATIONERS AND RAISERS WITH RE-
SPECT TO TOTAL EXPECTANCY SCORES--STRATIFIED BY SCHOLASTIC
APTITUDE LEVEL

 

 

Academic Total Expectancy Scores
Sig.
N Mean C.R. L 1
Status 0 1 2 3 4 5 6 7 8 9 10 eve

 

ABOVE AVERAGE SCHOLASTIC APTITUDE
Raisers O O 3 2 2 O 4 3 O 2 2 18 5.72 1 23 20
Prob'ners O 4 2 5 4 3 8 6 2 1 ° '

AVERAGE SCHOLASTIC APTITUDE -
Raisers 3 8 5 8 5 8 14 12 3 2 3 71 4°82 35 36
Prob'ners 7 8 ll 21 28 23 22 16 17 3 1 157 4.69 ' ’

BELOW AVERAGE SCHOLASTIC APTITUDE
Raisers 0 3 3 5 10 3 4 3 6

Prob'ners 12 16 18 18 17 20 23 7 13 148

[Di—J
h.)

 

103

Examination of Table 13 reveals that at each level of
scholastic aptitude the mean for the raisers was higher than
the mean for the probationers. At the above average and average
scholastic aptitude levels the means of the raisers were not
found to be significantly higher than the means for the pro-
bationers. _But, at the below average level of scholastic
aptitude the difference in means was significant beyond the
.05 level of confidence (t .95 - 1.64, one-tailed test).

It was concluded from the data reported in Table 13
that probationers and raisers were not significantly dif-
ferentiated as a group with respect to their total expectancy
scores when they were above average or average in scholastic
aptitude although the differences between the means were in
the predicted direction. However, those probationers and
raisers who were below average in scholastic aptitude were
significantly differentiated as a group with respect to their

total expectancy scores.

Comparisons between Probationers and Raisers with Respect
to the Ratings of Their Peer Groups
Made by the Resident Assistants
Data relating the ratings of the peer groups of the
probationers and raisers is presented in Table 14. A rating
of 1 indicates that the resident assistant in charge of the
student's precinct felt that the student associated mostly

with students who were highly motivated academically.

104

Ratings of 2 and 3 are intermediate positions and ratings of
4 indicate association mostly with students who were poorly

motivated academically.

TABLE 14. COMPARISONS BETWEEN PROBATIONERS AND RAISERS
WITH RESPECT TO PEER GROUP RATINGS

 

 

 

 

Academic Peer Group Ratings

N
Status 1 2 3 4
Probationers 26 86 97 131 340
Raisers 21 30 34 42 127

 

 

The frequencies presented in Table 14 tended to accumu-
late in the lower-left and upper-right hand quadrants. Chi—
square for the table was found to be 8.2. This value is
significant beyond the .05 level with three degrees of freedom.
Thus, as a group, the probationers and raisers were signif-
icantly differentiated with respect to the peer group ratings

made by the resident assistants.

The Relationship between Total Expectancy Scores
Derived from the Student Questionnaire and
Grade-Point-Average for the
Combined Samples

The data relating total expectancy scores derived from

the Student Questionnaire to academic performance dealt

105

primarily with comparisons between the three groups of
students identified for purposes of the study. It was felt
that one more step should be taken in testing the relation-
ship. It was decided to combine the three groups into one
population of 836 cases and compute the Pearson Product-
Moment correlation between the students' total expectancy
scores and their cumulative grade-point-averages for their
freshman year. The grade-point-averages were computed and
found to range from .67 to 3.95 (D a 1.00; A a 4.00). For
purposes of computation, the distribution of grades were
grouped into eleven classes with a class interval of 0.3
units. Eight cases falling below 1.00 were grouped together.
The bivariate distribution of the total expectancy scores and
the grade—point-averages is reported in Table 15.

The correlation between total expectancy scores and
grade-point-averages was found to be .284. The standard error
was .03. Since the obtained correlation was more than nine
times the standard error, it was concluded the obtained cor-
relation was statistically significant. This correlation is
not high by certain criteria, but it is indicative that there
was a positive correlation between the total expectancy scores
derived from the Student Questionnaire and actual grade-

point-average.

TABLE 15.
AND GRADE-POINT-AVERAGE

106

THE RELATIONSHIP BETWEEN TOTAL EXPECTANCY SCORES

 

 

Grade-Point

Total Expectancy Scores

 

 

 

Average 0 1 2 3 4 5 6 7 8 9 10 GPA
3.71 - 4.00 1 ' O 0 1 O 1 0 1 1 1 3 9
3.41 - 3.70 0 1 1 O 3 4 "4 4 5 0 7 29
3.11 - 3.40 2 1 2 3 5 6 4 5 9 5 9 51
2.81 - 3.10 2 1 3 5 11 13 6 15 15 10 8 89
2.51 - 2.80 1 5 9 7 11 13 15 8 13 7 6 95
2.21 - 2.50 7 9 3 8 18 13 20 11 16 4 3 112
1.91 - 2.20 4 10 12 17 21 22 27 24 15 5 5 162
1.61 - 1.90 8 15 I9 23 22 16 24 15 13 3 2 160
1.31 - 1.60 7 9 9 ll 12 15 12 5 12 2 1 95
1.01 - 1.30 3 1 0 3 4 4 6 2 3 O O 26
0.00 - 1.00 0 l 1 1 2 O 1 2 O O O 8

fE 35 53 59 79 109 107 119 92 102 37 44 836

 

 

The correlation between total expectancy scores and
grade-point-average was known to have been influenced sta-
tistically by the inter-correlation of expectancy scores with
scholastic aptitude scores and scholastic aptitude scores with
grade-point-average.

This trend was noted throughout the

study. That is, students with high expectancy scores seemed

to have higher scholastic aptitude scores than students with

107

low expectancy scores and conversely. It is well known that
there is a positive correlation between scholastic aptitude
scores and grade-point-average and it is not unthinkable
that there might be a relationship between the academic ex-
pectancies held for a student and his scholastic aptitude as
measured by a standardized test.

In order to deal with the problem of the mutual depen-
dence of grade-point-average upon expectancies on the one
hand and scholastic aptitude scores on the other, it was
decided to apply the method of partial correlation to the
data. This method enabled the investigator to compute the
correlation between expectancy scores and grade-point-
average with the effect of their common statistical dependence
upon scholastic aptitude scores "partialled out". In effect,
the influence of the scholastic aptitude level of the student
was held constant. Aspects of this method were employed
earlier in the study when a report was made of group ten-
dencies when the samples were stratified by scholastic ap-
titude level (Tables 9 and 13).

The computation of the partial correlation required
knowing the correlation between scholastic aptitude scores
and grade-point-average, and the correlation between scho-
lastic aptitude scores and total expectancy scores as well as
the original zero order correlation between expectancy scores
and grade-point-average. The correlation between scholastic

aptitude scores and grade-point-average was computed and

108

found to be .505 which is in keeping with the commonly found
value for such correlations. The correlation between scho-
lastic aptitude scores and total expectancy scores was com-
puted from the data presented in Table 16. Scholastic aptitude
is reported in MSU derived scores on the American Council on

Educational Psychological Examination.

TABLE 16. THE RELATIONSHIP BETWEEN SCHOLASTIC APTITUDE SCORES
AND TOTAL EXPECTANCY SCORES DERIVED FROM THE STUDENT
QUESTIONNAIRE

 

 

 

 

 

Expectancy Scholastic Aptitude Score fE
Score 1 2 3 4 5 6 7 8 9 10

10 0 0 2 3 4 8 12 8 '4 3 44

9 0 O 0 6 4 11 6 3 4 3 37

8 1 0 11 13 22 22 13 15 3 2 102

7 0 1 6 5 23 24 24 8 1 O 92

6 2 5 4 22 29 29 22 4 2 0 119

5 1 l 9 17 3O 31 10 6 2 0 107

4 1 4 7 19 35 21 13 6 2 1 109

3 1 3 7 15 21 16 8 6 2 O 79

2 1 1 10 10 11 17 8 O 1 0 59

1 O 3 6 13 10 12 6 2 O 1 53

O O 2 6 9 5 7 4 1 1 0 35

fA 7 2O 68 132 194 198 126 59 22 10 836

 

 

109

The correlation between expectancy scores and scholastic
aptitude scores was found to be .258 with a standard error of
.03. Since the obtained correlation was more than eight times
the standard error, it was concluded the correlation was sta-
tistically significant.

The partial correlation between total expectancy scores
and grade-point-average with scholastic aptitude held constant
was found to be .185. The partial correlation was lower than
the zero order correlation as would be expected; however, the
partial correlation was still six times the standard error
of .03 which indicates significance well beyond the .0005
level. That is, the observed partial correlation could not
have occurred by chance from an uncorrelated population more
than 5 in 10,000 times. Thus it was concluded that the cor-
relation between total expectancy scores derived from the
Student Questionnaire and grade-point-average--with scholastic

aptitude held constant--was significantly different from zero.

The Relationship between Academic Performance and
a Combination of Peer Group Rating and
Total Expectancy Score Derived from
the Student Questionnaire
A systematic analysis of the peer group ratings and
their relation to actual grade-point-average was not felt to
be apprOpriate since it would require treating the peer group

data as if they were cardinal data. Also, there seemed to

be little to be gained by stratifying the samples by

110

scholastic aptitude level and making additional comparisons
between the groups with respect to the peer group ratings on
that basis. The result of it all was that the data from the
Student Questionnaire were studied fairly extensively and the
peer group ratings not so extensively. Of course, the nature
of the questionnaires partly determined this difference.
However, it was felt some further consideration should be
given the peer group data. With this in mind, it was decided
to compare selected students' academic performance when their
peer group ratings and their total expectancy scores were
both taken into consideration.

In order to minimize any effect that scholastic apti-
tude might play and, at the same time, maintain a sizeable
sample, it was decided to restrict the study to comparisons
between probationers and non-probationers all of whom had
"average" scholastic aptitude scores. This resulted in sam-
ples of 157 probationers and 164 non-probationers for a total
of 321. Thus, the samples were nearly equal and of sufficient
size to assure comfortable statistical treatment.

Of particular interest at this point was the question
of whether or not there was any relationship between the
peer group ratings and the total expectancy score derived
from the Student Questionnaire. The bivariate frequency
distribution presented in Table 17 shows this relationship
for the combined probationers and non-probationers all of whom

had scholastic aptitude scores of five and six on the American

111

Council on Education Psychological Examination. Peer group
ratings of (1) indicate that a student associated mostly with
a highly academically motivated university peer group and
ratings of (4) indicate association with a poorly motivated
group. Total expectancy scores range from zero to ten with
high scores indicating that high academic expectancies were
perceived by the student as being held for him by members of

his home and home community group.

TABLE 17. THE RELATIONSHIP BETWEEN PEER GROUP RATINGS AND
TOTAL EXPECTANCY SCORES ON THE STUDENT QUESTIONNAIRE

 

 

 

 

Peer
Group Total Expectancy Scores fPG
Ratings 0 1 2 3 4 5 6 7 8 9 10

1 3 2 5 5 14 10 6 15 4 1 5 70
2 2 6 6 7 15 17 15 5 l3 5 0 91
3 3 1 6 7 9 12 11 6 8 2 3 68
4 1 5 6 10 13 14 12 9 16 5 1 92

fE 9 14 23 29 51 53 44 35 41 13 9 321

 

 

Inspection of Table 17 reveals that there is very little
relationship between the variables. The tetrachoric corre-
lation when the data were dichotomized at the medians was
computed and found to be -.O3. The standard error or r(tet)

in an uncorrelated population is .09 for sample size 321; thus,

112

it was concluded that the observed correlation was not signif-
icantly different from zero. Furthermore, for this sample the
peer group ratings and the total expectancy scores may be
thought of as "uni-dimensional traits". That is, since they
are uncorrelated, they must be tapping different factors if
they are tapping anything at all. It was shown in previous
sections of this chapter that both variables were positively
related to academic performance.

The combined effect of peer group ratings and total
expectancy scores in differentiating probationers and non-
probationers (at the "average" scholastic aptitude level)
was determined thusly: First, the combined distribution of
peer group ratings was dichotomized at the combined median
and cases falling above and below the median were labeled
"high" and "low". The median peer group rating was 2.5.
Second, the combined distribution of total expectancy scores
was dichotomized at their combined median and cases above and
below labeled "high" and "low" The median total expectancy
score was found to be 5.15. And thirdly, from the groups
thus dichotomized, it was possible to identify 16 non-
probationers who were "low" on both variables and 62 non-
probationers who were "high" on both variables. Similarly,
there were 62 probationers who were "low" on both variables
and 16 probationers who were "high" on both variables. Table
18 shows the relationship. For ease of interpretation, the

column headings are in terms of the expectancies inferred

113

from the peer group ratings and the Student Questionnaire

SCOI'GS.

TABLE 18. COMPARISONS BETWEEN PROBATIONERS AND NON-PROBATIONERS

0F "AVERAGE" SCHOLASTIC APTITUDE WITH RESPECT TO A COMBINA—
TION OF PEER GROUP AND HOME GROUP EXPECTANCIES

 

 

 

 

Academic Expectancies
f
Status Low Home Group High Home Group
Low Peer Group High Peer Group
Non-Probationers 16 62 78
_Probationers 62 16 78
fE 78 78 156

 

 

Direct computation of the tetrachoric correlation for

Table 18 resulted in a correlation of .80. Thus, a high degree

of relationship was shown when perceived "high" home expectan-
cies were accompanied by association with a university peer
group whose academic expectancies for a given student may be
inferred to be "high", and obversely.

Since the direct computation of the tetrachoric cor-
relation from Table 18 denies the existence of intermediate
positions between "high-high" and "low-low” namely "high-low"
and "low-high", computation by the principle of widely-
spaced groups provides a more appropriate estimate of the true

population correlation. The tetrachoric correlation thus

114

computed was found to be .574.

Incidental to this finding, the tetrachoric correlation
between peer group ratings alone and academic performance for
the sample was found to be .644. Also, the tetrachoric cor-
relation between total expectancy score alone and academic
performance for this sample was found to be .223. These data
suggest that at the "average" level of scholastic aptitude
the academic motivations of a student's peer group as rated by
a resident assistant are more significantly related to his
academic performance than his perception of the academic ex-
pectancies held for him by members of his home group.

This section concludes the analysis of the data. A
summary and interpretation of the findings are presented in

Chapter V.

CHAPTER V
SUMMARY AND CONCLUSIONS

Two questionnaires were developed by the investigator
for the purpose of assessing the academic expectancies held
for selected freshman, male college students by certain in-
dividuals and groups of individuals assumed to be significant
others for the students. A ten item, multiple-choice Student
Questionnaire was designed to determine a student's per-
ception of the academic expectancies held for him by members
of his home and home community group; and, a Peer Group
Questionnaire was designed as a rating form upon which a
rating of the academic motivations of a student's university
peer group could be made.

Complete questionnaire data were available for 836
freshman residents of men's residence halls at Michigan State
University during the 1957-1958 school year. The total
sample was composed of 369 non-probationary students whose
grade-point-averages during their freshman year were consis-
tently 2.00 or higher, 340 probationary students whose grade-
point-averages were consistently below 2.00, and 127 raisers
whose grade-point-averages for the fall term were below 2.00
but whose cumulative grade-point-averages for the full year
were 2.00 or higher.

The Student Questionnaires were distributed by resi-

dence hall assistants during the spring term and each student

116

completed the questionnaire privately. The resident assis-
tants completed the Peer Group Questionnaire by rating a
given student's university peer group as either highly
academically motivated or poorly academically motivated.

Normative comparisons were made between the three
groups of students with respect to their responses to each
item on the Student Questionnaire as well as the total ex-
pectancy score derived from the questionnaire. The separate
items on the questionnaire dealt with the student's per-
ception of the attitudes and feelings of his parents, high
school friends, high school staff members, relatives, and
neighbors regarding his going to college and success at
college. Responses to the items were numbered 1, 2, 3, and
4. Responses numbered (1) were indicative of high expectan-
cies, (2) and (3) were intermediate responses, and responses
numbered (4) indicated very low expectancies. The total
expectancy score derived from the Student Questionnaire was
determined by counting the number of (1) responses a student
made while responding to the ten items.

Comparisons were also made between the three groups
of students with respect to the ratings of their peer groups
made by the resident assistants. Ratings of (1) indicated
that the resident assistant thought a given student asso-
ciated mostly with students who were highly motivated aca—
demically, ratings of (2) and (3) were intermediate positions,

and ratings of (4) indicated association with a peer group

117

judged to be poorly motivated academically.

Additional analyses of the data involved the corre-
lation between the total expectancy score derived from the
Student Questionnaire and actual grade-point-average for the
combined samples; and, a comparison between probationers and
non-probationers of "average" scholastic aptitude with respect
to a combination of peer group expectancies and home group
expectancies.

A summary and interpretation of the findings and im-

plications for further research are included in the following

sections.

Summary and Interpretation of the Findings

Stated in positive hypothesis form, the four basic
hypotheses tested in this study were:

(1) There is a positive relationship between a stu-

dent's academic performance and his perception of the

academic expectancies held for him by significant

others in his home and home community group;

(2) There is a positive relationship between a stu-

dent's academic performance and the academic ex-

pectancies held for him by the significant others in

his peer group at the university;

(3) There is a positive relationship between a stu-

dent's potential for being removed from academic

118

probation and his perception of the academic expectan-

cies held for him by the significant others in his

home and home community group; and

(4) There is a positive relationship between a stu-

dent's potential for being removed from academic

probation and the academic expectancies held for him
by the significant others in his peer group at the
university.

Translated into the operational language of the study,
testing the first two hypotheses involved comparisons between
probationers and non-probationers with respect to their re-
sponses to the Student Questionnaire and comparisons between
the ratings of their peer groups made by the resident as-
sistants. Hypotheses (3) and (4), which are really corallaries
of the first two, were tested by making comparisons between
probationers and raisers with respect to the two variables.

Comparisons between probationers and non-probationers
with respect to their responses to each of the items on the
Student Questionnaire yielded the following major findings:

(1) Each of the ten questionnaire items was found to

differentiate significantly between probationers and

non-probationers beyond the five percent level of
confidence (chi-square tests).

(2) Items with the greatest differentiating power

were items dealing with the attitudes of neighbors

and friends of the family, the attitudes of high

119

school friends, the expectancies of mothers and
fathers in terms of actual grade-point-average, and
the number of a student's high school friends who
were going or planning to go to college. Each of
these items differentiated the groups at the .001
level of confidence.

(3) Items with intermediate differentiating power
were items dealing with the encouragement to go to
college the student felt he had received from high
school staff members and the student's estimate of
his parents' disappointment at his not getting a
college degree.

(4) Items differentiating at the .05 level of con-
fidence were those dealing with the length of time a
student's parents had planned for him to go to college
and the attitudes of his relatives if he did not go
to college.

The total expectancy score on the Student Questionnaire

was found to differentiate probationers and non-probationers

at the .0005 level of confidence (one-tailed, critical ratio

When the groups were stratified by scholastic ap-

titude scores into "above average", "average", and "below

average" levels, the total expectancy score was still found

to differentiate beyond the .025 level of confidence for each

level of scholastic aptitude.

The data summarized above support the hypothesis that

120

there is a positive relationship between a student's academic
performance and his perception of the academic expectancies
held for him by the significant others in his home and home
community group.

The data also suggest that the attitudes, feelings and
expectancies of certain individuals in the home community
may be more important than others. For example, questionnaire
items dealing with the attitudes of parents, neighbors and
friends of the family, high school friends, and the college
plans of high school friends tended to differentiate pro-
bationers from non-probationers more significantly than
items dealing with high school principals, teachers, and
counselors or family relatives.

For most of the questionnaire items there was a ten-
dency for probationers and non-probationers alike to choose
the response numbered (l)--which was the response indicating
the highest expectancy represented by the item. This finding
suggests that the very fact that a student has been enrolled
in the university for almost a year when he completed the
questionnaire was indicative that academic expectancies of
some magnitude were probably held for him. In every case,
however, a greater percentage of non-probationers than pro-
bationers chose item responses numbered (1). Similarly, the
students seemed to avoid the response numbered (4)-~which
was the response indicating the lowest expectancy represented

by the item. Typically however, the probationers exceeded

121

the non-probationers in their choice of responses numbered (4).

In every statistical test employed, the differences
between probationers'and non-probationers with respect to the
academic expectancies inferred from their responses to the
Student Questionnaire were in the predicted direction and
Statistically significant. Thus, support of the first major
hypothesis seems warranted.

The ratings made by the resident assistants of the
academic motivation of the students' university peer groups
were found to differentiate probationers from non-probationers
at the .001 level of confidence (chi-square test). That is,
students who seemed to associate most closely with a peer
group which was rated as highly academically motivated were
more likely to be non—probationers and students who were
identified with peer groups which were rated as poorly aca-
demically motivated were more likely to be probationers.

Interpretation of this finding requires the assumption
that the academic expectancies which peer group members hold
for each other are reflected in the apparent academic moti-
vation of the group as rated by the resident assistants in
this case. Data presented in Chapter III which dealt with the
validity of the Peer Group Questionnaire would seem to support
this assumption. It was demonstrated that a student's per-
ception of the academic expectancies held for him by his close
friends were positively correlated (.647) with the general

level of grades he thought his friends were receiving. It was

122

also demonstrated that the peer group ratings made by the
resident assistants were positively correlated (.371) with
the student's perception of the grades his friends were re-
ceiving. Therefore, it was concluded that the correspondence
between the resident assistants' ratings and the students'
own reports of the academic performance of their peers was
sufficiently high to warrant fairly valid group comparisons.

On this basis, then, the data suggest that there is a
positive relationship between a student's academic performance
and the academic expectancies held for him by the significant
others in his peer group at the university. Thus, support of
the second major hypothesis tested in the study was obtained.

Comparisons between probationers and raisers were
employed to test the hypotheses relating a student's potential
for being removed from academic probation to the academic
expectancies held for him. The same basic procedures were
employed as were employed when probationers and non-probationers
were compared.

Unlike the findings when probationers and non-probationers
were compared, the differences between probationers and raisers
with respect to the various variables were not so great. The
relative smallness of the sample of raisers (127) compared to
the sample of probationers (340) was partly responsible for
this since even small differences between large samples are
oftentimes more significant statistically than larger dif-

ferences between smaller samples. Another interpretation of

123

this finding, however, is that true differences between pro-
bationers and raisers may not actually be as great as dif-
ferences between probationers and non-probationers. The
non-probationers had never been on academic probation while
all the raisers had been on probation for at least one term.
One might assume that since all raisers had experienced
academic probation, the expectancies held fbr them were
logically not as high as for students who had never experi-
enced academic probation. This assumption is supported by
the findings when probationers and non-probationers were
compared.

While the differences between probationers and raisers
were not as great as the differences between probationers and
non-probationers, the observed differences were typically in
the predicted direction. The single exception was the dis-
tribution of responses to item 8 on the Student Question-
naire. The responses to this item were negatively related
to academic performance but statistically insignificantly.

Student Questionnaire items found to differentiate
most significantly between probationers and raisers were items
dealing with the number of a student's high school friends who
were going or planning to go to college and the expectancies
of mothers and fathers in terms of actual grade-point-average.
Each of these items differentiated the groups beyond the .05

level of confidence. A clue as to the relative importance of

124

the expectancies inferred from these items is Obtained by
comparing the values of chi-square for each item. (Since each
chi-square was computed with three degrees of freedom, direct
comparisons are justified.) The expectancies of fathers
ranked highest (chi-square . 12.8). The propensity to college
going of high school friends ranked second (chi-square - 9.1).
And the expectancies of mothers ranked lowest (chi-square a
7.8). This difference between the students' perceptions of
the expectancies held for him by his two parents was not ob-
served when probationers and non-probationers were compared.
If any confidence can be placed in the students' responses at
all, one might conclude that the probability of a student's
being removed from academic probation is more closely related
to the expectancies held by fathers than those held by mothers.
Two questionnaire items were found to differentiate
probationers and raisers at the .25 level of confidence. That
is, the observed differences could have occurred by chance
only once in four times. These items dealt with the length
of time a student's parents had planned for him to go to
college and the attitudes of his high school friends if he
did not go to college. Of note is the fact that the attitudes
of a student's high school friends as he perceived them were
not as highly related to academic performance as his report
of how many of them were planning to go or going to college.
Chi-square computed for the item dealing with attitudes of
high school friends was significant at the .05 level and the

125

significance level of chi-square for the item dealing with
the number of friends going to college was only .25 as just
noted. It might be supposed that a student was more able to
report objectively how many of his friends were going to
college than what their feelings would be if he did not go.
In which case, his report of how many friends were going to
college would logically be more valid than his perception of
his friends' attitudes towards his not going. In any event,
for this sample, the students' reports of how many of their
high school friends were going to college were found to dif-
ferentiate probationers from raisers more significantly than
their perceptions of the attitudes of their high school
friends regarding their not going to college.

In contrast, the comparisons between probationers and
non-probationers revealed that it was the attitudes of high
school friends which differentiated those groups more sig-
nificantly. Chi-square computed for the item dealing with
attitudes was 27.7 (df . 2) and chi-square for the item
dealing with the prOpensity to college going of high school
friends was only 14.2 (df a 2). Both chi-squares were sig-
nificant beyond the .001 level, but the former was nearly
twice the latter. It may be presumed that a student's per-
ception of his high school friends' attitudes towards his
going to college had more affective overtones than his more
or less objective or cognitive report of how many of his

friends were going to college. In which case, one might

126

assume that a student's perception of his friends' attitudes
would quite logically differentiate probationers from non-
probationers more significantly. The data support this
assumption. However, why doesn't the same assumption obtain
with respect to comparisons between probationers and raisers?
Perhaps it does and the data do not reveal it. Accidents of
sampling, lack of objective reporting, or factors completely
undetected may all have contributed to the apparent inconsis-
tency in the data.

The distributions of the responses to questionnaire
items dealing with the encouragement to go to college the
student felt he had received from high school staff members,
the attitudes of neighbors and relatives towards his not
going to college, and his estimate of his parents' dis-
appointment at his not getting a college degree, while typ-
ically in the predicted direction, were not significantly so.

The total expectancy score on the Student Questionnaire
was found 'to differentiate probationers from raisers at the
three percent level of confidence (one-tailed, critical ratio
test). This finding parallels the finding when probationers
and non-probationers were compared and lends further support
for acceptance of the first major hypothesis of the study.

When the samples were stratified by scholastic aptitude
scores, only those students who were at the "below average"
level were significantly differentiated. The mean expectancy

‘Ascores of the raisers were higher than the mean scores of the

127

probationers at all three levels, however, at the "average"
and "above average" levels not significantly so. These
findings suggest that students on academic probation who are
among the lower echelons in scholastic aptitude as measured
by a standardized test are more likely to be removed from
probation when their studying is accompanied by a perception
that relatively high academic expectancies are held for them
by members of their home group. Thus they exceed other pro-
bationers in the attention they give their course work and
succeed in raising their grades to an acceptable level.
Despite the smallness of the samples involved in the present
study, the data support this conclusion.

In general, probationers and raisers taken as a group
were significantly differentiated with respect to their total
expectancy scores derived from the Student Questionnaire.
Using this general finding as a criterion, the hypothesis
that there is a positive relationship between a student's
potential for being removed from academic probation and his
perception of the academic expectancies held for him by the
significant others in his home and home community group is
supported by the present study.

The ratings made of the academic motivations of the
students' university peer group were found to differentiate
probationers from raisers at the five percent level of con-
fidence. Once again the results were not as significant as

when probationers and non—probationers were compared, but the

128

findings are consistent with each other and lend mutual sup-
port for the second major hypothesis of the study. Further-
more, the data suggest that the fourth major hypothesis of
the study is also tenable. That is, it would seem that there
is a positive relationship between a student's potential for
being removed from academic probation and the academic ex-
pectancies held for him by the significant others in his peer
group.

The overall predictive efficiency of the Student
Questionnaire was determined by computing the correlation
between total expectancy scores and actual freshman grade-
point-average for the combined sample of 836 cases. The
Pearson Product-Moment Correlation was found to be .284. The
correlation between scholastic aptitude scores and grade-
point-average was found to be .505, and the inter-correlation
of scholastic aptitude scores with total expectancy scores
was found to be .258. Computation of the partial correlation
between total expectancy scores and grade-point-average--with
scholastic aptitude held constant--resulted in a correlation
of .185. A partial correlation of .185 in a sample of 836
cases is significantly different from zero beyond the .0005
level of confidence. And, of course, the zero order cor-
relation of .284 between total expectancy scores and grade-
point-average is also statistically significant. These
findings are consistent with the findings previously reported

regarding the relationship between academic performance and

129

total expectancy scores. However, it must be conceded that
the total score on the Student Questionnaire did not match

the predictive efficiency of the American Council on Education
Psychological Examination as a predictor of freshman grade-
point-average.

The most significant relationships between academic
performance and academic expectancies were found when a com-
bination of peer group ratings and total expectancy scores
were used to compare probationers and non-probationers all
of whom had ”average" scholastic aptitude scores. A total
sample of 156 probationers and non-probationers was identified
whose peer group ratings and total expectancy scores were
both "high" or both "low".’ All the students had scores of
five or six on the American Council on Education Psychological
Examination. Sixty-two probationers and 16 non-probationers
were identified who fell below the combined sample medians
with reSpect to peer group ratings apthotal expectancy
scores. Similarly, 16 probationers and 62 non-probationers
who had peer group ratings and total expectancy scores above
the combined sample medians were identified. The tetrachoric
correlation computed by the principle of widely-spaced groups
was used to determine the extent of the relationship between
academic performance and the combination of expectancy fac-
tors as defined. The tetrachoric correlation was found to
be .574. This observed correlation was influenced largely by

the correlation between academic performance and peer group

130

ratings which was found to be .644 for the selected sample.
The correlation between academic performance and total ex-
pectancy scores derived from the Student Questionnaire for
the selected sample was found to be .223. This value approxi-
mates the previously obtained correlation between actual
grade-point-average and total expectancy score.

These data would seem to indicate that the academic mo-
tivations of a student's university peer group as rated by
a resident assistant are more closely related to his academic
performance than his perception of the academic expectancies
held for him by members of his home group. At least, this
would seem to be the case among students of "average" scho-
lastic aptitude as measured by a standardized test. This
conclusion is also supported by theoretical considerations.
The conformance of group members to group norms with respect
to many forms of social behavior has been widely observed--
even when conformance to group norms was in conflict with
previously established norms and values. From this, it could
be assumed that college students would be particularly prone
to yield to new found group pressures when entering college
and adjust their academic performance accordingly-—the ex-
pectations of parents and others at home notwithstanding. In
any event, the evidence from the present study would seem to
support this assumption.

In review, it may be concluded that the four major hy-

potheses defined for this study were supported by the evidence.

131

No evidence was found which would seriously discredit either
the major hypotheses or any of the required assumptions. The
study has raised many questions and implications for further
research. Some implications for further research are reported

in the final section of this thesis.

Implications for Further Research

Probably the most important single implication for
further research is cross-validation of the Student Question—
naire. On the basis of the present study, it would seem the
Student Questionnaire is a potentially useful device for the
study of students' academic performance. The study revealed
a correlation of .284 between the total expectancy score de-
rived from the questionnaire and cumulative freshman grade-
point-average in a sample of 836 resident, male, Michigan
State University freshmen. Cross-validation with other pop-
ulations, women included, is needed to test the presently
reported correlation.

The questionnaire should be administered to samples of
entering freshmen. In the present study the students had been
enrolled for nearly a year when they completed the question-
naire. There may be differences in the manner in which stu-
Cients respond depending upon the length of time they have
‘been enrolled. Improvement in the items themselves could

LDPobably be achieved and the questionnaire might be lengthened

132

to include items regarding how the student himself feels about
the expectancies held for him. It is possible that the effec-
tiveness of the present questionnaire was limited since it did
not take into account the student's own definition of the
situation but considered only his perception of the expec-
tancies held for him.

A study should be undertaken to test the predictive
efficiency of the Student Questionnaire over time. That is,
correlations between expectancy scores of entering freshmen
and their sophomore, Junior, and senior grades might reveal
clues as to the continuing influence of the expectancies held
by the home group in determining a student's academic per-
formance. It might be hypothesized that the longer a student
is in college the less would be the influence of his home group
on his academic performance. This hypothesis needs to be
tested.

More research is needed regarding the dynamics of the
interactions within undergraduate peer groups. Residence hall
personnel workers might profitably undertake systematically
to study the peer group structure of their halls and develop
action research projects designed to define maximum academic
performance as the norm for all students. This would require
including in the residence hall personnel program activities
for individuals and groups of individuals which would clearly
legitimize high academic achievement as a major value and

observing the educational outcomes obtained.

133

Another implication for further research would be in
the area of counseling tools and techniques. An instrument
similar to the Student Questionnaire could be devised and
experimented with by college counselors working with students
who experience academic difficulty. The instrument could
include items regarding the expectancies held for the student
by parents and other persons at home, the student's own
interpretation of these perceived expectancies, and some in-
formation regarding the academic motivation and achievements
of the student's close college friends.

In conclusion, it should be stated that the findings of
the present study should not be interpreted as evidence that a
causal relationship between academic performance and academic
expectancies was clearly demonstrated. The observed differ-
ences in questionnaire responses might have been influenced
by the fact that the questionnaires were administered after
the students had been in school for some time and had already
achieved certain grade-point-averages. In which case, the
achievement of the students might have been causing certain
expectancies to be held for them as Opposed to the interpre-
tation that it was the expectancies of others which were
causing the students to achieve. A definitive analysis of the
relationship between college academic performance and academic
expectancies would require gathering expectancy data before
students had earned any college grades and would probably
require student interviews of some depth as well as data from

questionnaires and other psychometric devices.

134

BIBLIOGRAPHY

10.

ll.

135
BIBLIOGRAPHY

Ahman, J. 3.; Smith, W. L.; and Glock, M. D., "Predicting
Academic Success in College by Means of a Study Habits
and Attitude Inventory," Educational and Psychological
Measurement 18:853-857, Winter, 1958.

 

Bendig, A. W., and Klugh, Henry E., "A Validation of
Gough's Hr Scale in Predicting Academic Achievement,"
Educational and Psychological Measurement 16:516-523,
Winter, 1956.

Bledsoe, Joseph C., "An Analysis of the Relationship of
Size of High School to Marks Received by Graduates in
First Year of College," Journal of Educational
Sociology, 27:41u-418, Magi—igsut‘

 

 

Bovard, Everett W., Jr., "Social Norms and the Individual,”
Journal gf_Abnormal and Social Psychology 43:62-69,
January, 1938.

Bovard, Everett W., Jr., "Group Structure and Perception,"
Journal of Abnormal and Social Psychology, 46:398-405,
July, 1951.

 

American Book Company, 1955, 36 pp.

 

Brookover, Wilbur E., A Sociology of Education, New York:

Brookover, Wilbur B., "A Social Psychological Conception
of School Learning," School and Society 87:84-87,
February, 1959.

 

Cartwright, Dorwin and Zander, Alvin, Group Dynamics,
Ezanston, Illinois: Row, Peterson and Company, 1953,
2 pp.

 

Combs, Arthur W., "Intelligence from the Perceptual Point
of View," Journal of Abnormal and Social Psychology
47:662—673, July, 1952.

 

Cottrell, Leonard 8., Jr., "The Analysis of Situational
Fields in Social Psychology," American Sociological
Review 7:370-382, June, 1942.

 

Cureton, Edward E., in Educational Measurement, edited
by E. F. Lindquist, American Council on Education,
Washington D. C., 1951, 819 pp.

 

l2.

13.

14.

15.

l6.

17.

18.

19.

20.

21.

22.

23.

136

Cureton, Edward E., ”The Principal Compulsions of Factor-
Analysts," Test Service Notebook Number 3, World Book
Company, Yonkers on Hudson, (reprinted, with revisions
from the Harvard Educational Review, Vol. 9. No. 3.)

 

Dewey, John, Democracy and Education, New York: The
MacMillan Company, 19167’fi34 pp.

 

DiVesta, Francis J., Woodruff, Asahel D., and Hertel,
John P., "Motivation as a Predictor of College
Success," Educational and Psychological Measurement

9:338-348, Autumn, 19H9.

Drews, Elizabeth M. and Teahan, J. E., "Parental
Attitudes and Academic Achievement," Journal of_
Clinical Psychology» 13:328-332, October, 1957.

 

Flanagan, John C., "General Considerations in the
Selection of Test Items and a Short Method of Esti-
mating the Product-Moment Coefficient from the Tails
of the Distribution," Journal of Educational
Psychology 30:674-680, December, 1939.

 

 

Ford, Thomas R., ”Social Factors Affecting Academic
Performance," The School Review 65:415-h22, Winter,

1957.

Garcia, Dolores and Whigham, Neil, "Validity of SSHA
Administered Before and After College Experience,"
Educational and Psychological Measurement 18:845-857,
Winter, 1958.

 

 

Gough, H. G., "The Construction of a Personality Scale to
Predict Scholastic Achievement," Journal 9£_Applied
Psychology 37:361-366, October, 1953.

 

Gowan, John 0., "Dynamics of the Underachievement of
Gifted Students," Exceptional Children 24:98-101,
November, 1957.

 

Haggard, Ernest A., "Socialization, Personality and
Academic Achievement in Gifted Children," The School
Review 65:388-414, Winter, 1957.

 

Hare, A. P.; Borgatta, E. F.; Bales, R. F.; Small Groups,
New York: Alfred A. Knopf, 1955, 666 pp.

 

Hoffman, L. Richard, "Similarity of Personality: A Basis
for Interpersonal Attraction?," Sociometry 21:300-307,
December, 1958.

 

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

137

Holtzman, W. H.; Brown, W. F.; and Farquhar, W. G.,
"The Survey of Study Habits and Attitudes," Educational
and Psychological Measurement 14:726-732, Winter, 1954.

Hurley, John R., "Maternal Attitudes and Children's
Intelligence," Journal of Clinical Psychology
15:291-292, July, 1959.

Linton, Ralph, The Cultural Background 9£_Personality,
New York: Appleton-Century—Crofts, Inc., 1945, 157 pp.

Manuel, Herschel T., "Aptitude Tests for College
Admission," Fourteenth Yearbook of the National
Council on Measurements Used in_Education, The Council,

1957, p "Tao-27..

McDavid, John Jr., "Some Relationships between Social
Reinforcement and Scholastic Achievement," Journal of_
Consulting Psychology 23:151-154, April, 1959.

 

 

 

 

Mead, George H., Mind, Self and Societ , Chicago:
University of Chicago Press, 1934, 400 pp.

 

 

Montalto, Fannie D., "An Application of the Group
Rorschach Technique to the Problem of Achievement in
College,” Journal of Clinical Psychology, 2:254-260,
July, 1946.

 

Morris, Charles W., in Introduction to Mind, Self and
Society, by George H. Mead, Chicago: University of
Chicago Press, 1934, 400 pp.

 

Munroe, Ruth, "Adjustment and Academic Performance of
College Students," Applied Psychology Monographs,
Stanford University Press, 1945.

 

Murphy, Gardner, Personality, New York: Harper and
Brothers, 1947, 999 pp.

 

Myers, R. C. and Schultz, D. C., "Predicting Academic
Achievement with a New Attitude-Interest Questionnaire,"
Educational and Psychological Measurement 10:654-663,
Winter, 1950.

 

Newcomb, Theodore M., Social Psychology, New York: The
Dryden Press, 1950, 690 pp.

 

Parsons, Talcott and Shils, Edward A., Toward a_General
Theory g£_Action, Cambridge, Mass.: Harvard University
Press, 1951, 506 pp.

37.

38.

39.

40.

41.

42.

43.

an.

45.

46.

47.

138

Peters, Charles C. and VanVoorhis, Walter R., Statistical
Procedures and Their Mathematical Bases, New York:
McGraw-Hill Book Co., 1940, 516 pp.

Phillips, W. S. and Osborne, R. T., "A Note on the
Relationship of the Kuder Preference Record to College
Marks, Scholastic Aptitude, and Other Variables,"
Educational and Psychological Measurement 9:331-337,
Autumn, 1949.

Pounds, Ralph L., "Prediction of Academic Success at the
University of Cincinnati Teachers College," Twelfth
Yearbook gf_the National Council on Measurement Used
§2_Education, Part II, The Council, 1955, pp. 12-31.

 

Prescott, Daniel A., The Child in the Educative Process,
New York: McGraw-Hill Book Company, 1957, 502 pp.

 

Robinson, Donald W., "A Comparison of Two Batteries of
Tests as Predictors of Success in the Graduate School
of Bradley University," Fifteenth Yearbook 9£_§h§_
National Council gn_Measurements Used in_Education,
The Council, 1958, pp. 118-127.

 

 

Rohrer, J. H.; Baron, S. H.; Hoffman, E. L.; and Swander,
D. V., "The Stability of Autokinetic Judgements,"
Journal of Abnormal and Social Psychology 49:595-597,
October,-1954.

Schonbar, Rosales A., "The Interaction of Observer-Pairs
in Judging Visual Extent and Movement," Archives of_

Psychology 37:No. 299, 1945.

Sherif, M., "A Study of Some Social Factors in Perception,"

Archives of Ps cholo 27:No. 187, 1935. (reported in
Newcomb 35:264-2655

Simon, Herbert A., "A Formal Theory of Interaction in
Social Groups," American Sociological Review 17:202-
211, April, 1952.

Snygg, Donald and Combs, Arthur, Individual Behavior,
New York: Harpers and Brothers, 1949, 386 pp.

 

Sopchak, Andrew L., "Prediction of College Success by
Commonly Used Tests," Journal of Clinical Psychology
14:19u-197’ April, 1958.

 

48.

49.

50.

51.

52.

53.

54.

55.

139

Thompson, Grace M., "College Grades and the Group
Rorschach," Journal 23 Applied Psychology 32:398-407,
August, 1948.

Tolman, Edward C., in Toward a_General Theorygf Action,
edited by Talcott Parsons and Edward A. Shils,
Cambridge, Mass.: Harvard University Press, 1951,

506 pp.

Walker, Helen M., Statistical Methods, New York: Henry
Holt and Company, 1943, 368 pp.

 

Walker, Helen M. and Lev, Joseph, Statistical Inference,
New York: Henry Holt and Company, 1953, 510 pp.

 

Wantman, Morey J., "Development of Measures for Prediction
of College Success." Eleventh Yearbook of the National

Council on Measurements Used ig_Education, The Council,

1954, ppT—82-87.

Wilson, Claude E., "Differences between Personal Char-
acteristics of Students Who Have Failed in High School
and Those Who Have Not Failed," The Ninth Yearbook of,
the National Council gn_Measurements Used in_Education,
The Council, 1951, pp. 42-50. (mimeographed)

3

 

 

 

 

 

 

Yoshino, 1. Roger, "College Drop-outs at the End of the
Freshman Year," Journal 9£_Educational Sociology
32:42-48, September, 1958.

 

Young, Kimball, Social Psycholo , New York: Appleton-
Century-Crofts, Inc., 1944, 57 pp.

 

140

APPENDIX

141

APPENDIX

The Instruments

STUDENT QUESTIONNAIRE

Instructions to Resident Assistants

This questionnaire is to be administered to the same
men whose names appear on the Peer Group Questionnaire.
Note that you have a questionnaire for each man. The in-
structions to the student are simple and precise, and it
should take him Just a few minutes to respond to the twelve
items.

(1) Approach each man individually when he is alone.
(2) Enlist his cooperation and wait for him to complete the
uestionnaire. (3) Make sure he has completed all the items.
?4) Return the completed questionnaire along with the PEER
GROUP QUESTIONNAIRE to your Head Resident Advisor.

If necessary, assure the student that his responses
will be treated in strictest confidence and will in no way
affect him as a resident of the hall or as a student of the
university.

142

STUDENT QUESTIONNAIRE

Instructions: Please respond as quickly as possible to the

1.

following twelve items. Your first reaction is all that
is required. Circle the number of the response that
indicates your first reaction. Please answer all the items.

My parents
(1) have planned since I was quite young for me to go
to college

(2) have planned since I was in high school for me to
go to college

E3; have never really planned for me to go to college

4 would rather have me get a Job or work at home
than go to college

My parents

(1) would be extremely disappointed if I did not get a
college degree

(2) would be mildly disappointed if I did not get a
college degree

(3) really don't care whether I get a college degree
or not

(4) would prefer that I not get a college degree

If I did not go to college, the neighbors and friends
of the family
1 would wonder quite a bit why I did not go
2 would wonder a little why I did not go
3 really wouldn't wonder at all why I did not go
4 would probably take the attitude that I was better
off if I didn't go

If I did not go to college, my close friends from high
school
1 would wonder quite a bit why I did not go
2 would wonder a little why I did not go
3 really wouldn't wonder at all why I did not go
4 would probably take the attitude that I was better
off if I didn't go

My high school teachers
1 strongly encouraged me to go to college
2 mildly encouraged me to go to college
3 didn't encourage me at all to go to college
4 discouraged me from going to college

10.

143

My high school principal or counselor
1 strongly encouraged me to go to college
2 mildly encouraged me to go to college
3 didn't encourage me at all to go to college
4 discouraged me from going to college

Of my close friends from high school

(1) most of them are now going or planning to go to
college

(2) quite a few of them are now going or planning
to go to college

(3) a few of them are now going or planning to go
to college

(4) practically none of them are now going or planning
to go to college

If I did not go to college, my relatives

1 would wonder quite a bit why I did not go

2 would wonder a little why I did not go

3 really wouldn't wonder at all why I did not go

(4 would probably take the attitude that I was
better off if I didn't go

My father
1 expects me to get at least a 3.00 grade-point-average
2 expects me to get at least a 2. 50 grade-point- average
3 expects me to get at least a little better than a

2. 00 grade- point- average
(4) is satisfied as long as I make my 2. 00 grade-

point average

 

My mother
1 expects me to get at least a 3.00 grade-point-average
2 expects me to get at least a 2.50 grade-point-average
3 expects me to get at least a little better than a
2.00 grade-point-average
(4) is satisfied as long as I make my 2.00 grade-
point-average

(Note: Items 11 and 12 were included with the STUDENT
QUESTIONNAIRE to gather data for validity studies of the
PEER GROUP QUESTIONNAIRE.)

11.

Most of my close friends in the precinct

 

1 usually get at least a 3.00 grade-point-average
2 usually get at least a 2.50 grade-point-average
3 usually get at least a little better than a 2.00

grade-point-average
(4) usually get a 2.00 grade-point-average or below

144

12. Most of my close friends in the precinct

ti

(4)

expect me to get at least a 3.00 grade-point-average
expect me to get at least a 2.50 grade—point-average
expect me to get at least a little better than a
2.00 grade-point-average

expect me to get a 2.00 grade-point-average or below

145

The Instruments
PEER GROUP QUESTIONNAIRE
Introduction and Instructions

This questionnaire is being used to gather some confi-
dential information about some of the men in your precinct.
The information will be used in a research study that is
currently being conducted and will in no way affect a resident's
standing in school or his official university record.

When you think of the men of your precinct, you often-
times think of them as members of a group 23 clique. Most
precincts have several fairly well defined groups that
operate in the precinct. These groups are usually made up
of men who have something in common. Sometimes the only thing
they seem to have in common is the desire to make noise or
complain. Other groups are bound together by common academic
interests. Still others Just seem to hang around together,
play cards together, eat together, or go out together.

The research that you are participating in has to do
with the various groups you have in your precinct and the way
in which some of your men are identified with these groups.
We are concerned with two basic kinds of groups:

Group A includes

groups made up of men who seem to be pretty highly
motivated academically; most of the men in the group
seem to really understand what they are here for; they
take their studies rather seriously; they know what it
means to put their studies first; it could be said that
they expect each other to measure up pretty well
academically.

Group B includes

groups made up of men who seem to be rather poorly
motivated academically; most of the men Just seem to be
putting in their time here; they don't take their studies
very seriously; they tend to put their studies second
and other things first; it could be said that they ex-
pect each other to be good guys rather than scholars

The information requested is easily indicated. The men
you are to consider are listed for your convenience. (1) Think
about the man: (2) Think about the group of men with whom he
most closely associates: (3) Decide whether this group is
most like Group A or Group B: (4) Check your decision in the
appropriate column: (5) Check one column only for each man.

PEER GROUP QUESTIONNAIRE

 

 

146

 

 

 

 

Hall
Precinct Resident Assistant
Group A Group B
Associates Tends to Associates Tends to
RESIDENTS m9§§1y_ associate gg§t%y_ Associate
with with w t with

Group A Group A Group B Group B
l.
2.
3.
4.
5.
6.
7.
8.
9.
10.
ll.
12.
13.
l4.
15.

l--'
O\

 

 

 

 

 

ROOM USE 0““

 

3 ROOM U$£ ‘hmv’

%

u.—.