AN ANALYSES OF FACTORS
ASSOCIATED WITH CHANGES IN
SCHOLASTI‘C PERFORMANCE PATTERNS

fine]: {or ”H: Degree of DH. D.
MICHIGAN STATE UNIVERSITY
Rodney T. Hartnett

1963

THESIS

This is to certify thatOthe

thesis entitled

An Analysis of'Factors Associated With
Changes in Scholastic Performance Patterns

presented by

Rodney T. ertnett

has been accepted towards fulfillment
of the requirements for

Ph.D Education

' degree in

6%??M

Date Ju {

 

0.169

 

@v

 

ABSTRACT

AN ANALYSIS OF FACTORS ASSOCIATED WITH CHANGES
IN SCHOLASTIC PERFORMANCE PATTERNS

by Rodney T. Hartnett

The purpose of this investigation was to explore the
relationship between selected variables and changes in the
academic performance patterns of students during four years
of college. The study was designed to differentiate between
students whose performance pattern improved as opposed to
those Whose pattern of achievement changed negatively. The
basic purpose was to examine the factors thought to have

association with the change in performance at the time the

 

change occurred.

 

The sample selected for the study was composed of 1,041
students who entered a large midwestern university as fresh—
men fall, 1958, and were in attendance during each fall,

winter, and spring term through spring of 1962,

Subjects were classified into change groups on the basis
of the difference between their performance predicted from
the previous year's grade point average (determined by

linear regression) and their actual performance. Any

Rodney T. Hartnett

discrepancy greater than a confidence band of .50 standard
errors of estimate was used as the criterion for change
classification. The subjects were also classified according

to their achievement into high, middle, and low groups.

The change groups were then compared on three types of
experimental variables. These variables were: (1) ability
and aptitude, (2) attitudes and values, and (3) behavioral
characteristics. Used as measures of the aptitude variables
were The Michigan State University Reading Test, The Collgge

Qualification Test, and A Test of Critical Thinking, Form G.

—fi—‘

 

For measures of attitudes and values The Inventory of Beliefs,

 

The Differential Values Inventory, and Rokeach's Dggmatism

 

Scale were employed. The behavioral variables consisted of
degree of participation in extra-curricular activities, place
of residence, the amount of time devoted to part-time
employment during the school year, and a number of personal

factors.

Analysis of variance was used as the basis for determining
the significance of the differences between the performance
change groups on the first two types of variables, whereas
chi—square tests were employed for analysis of the data per-

taining to the behavioral variables. Throughout the study a

 

 

(D

rt

'0

If?)

(n

f

(‘7

PI‘

(I)

m

Rodney T. Hartnett
.05 level of probability was used in determining statistical

significance.

The ability-aptitude and attitude-value variables failed
to discriminate between the performance changers. An occasional
F-value exceeded by .05 level of confidence, but no more than
would be expected by chance. It was concluded that these
variables were not associated.with changes in scholastic

performance patterns.

The behavioral variables, on the other hand, were more
helpful in distinguishing between performance changers. Those
making a positive change during the sophomore year, for
example, were typically those who: (1) pledged and joined a
fraternity or sorority during that year, (2) worked fewer
hours at part-time jobs than negative changers, (3) lived in
a residence hall (whereas negative sophomore changers tended
to list a fraternity or sorority as their place of residence),
and (4) participated less in extra-curricular activities.
During the junior year, those making positive changes (as
opposed to those making negative changes) were characterized
by: (1) fewer changes of major field, (2) fewer students
becoming "active" in fraternities and sororities, (3) less
employment, and (4) less participation in extra-curricular

activities. The behavioral variables associated with changes

 

12‘. Cd

pronou

still

Rodney T. Hartnett
in academic performance during the senior year were not as
pronounced as during the sophomore and junior years, but
still provided some clues regarding behavioral differences.
Positive changers during the senior year more likely lived
in residence halls, whereas negative changers more likely

lived in fraternity or sorority houses.

AN ANALYSIS OF FACTORS ASSOCIATED WITH

CHANGES IN SCHOLASTIC PERFORMANCE PATTERNS

BY
,Cv‘,

Rodney Tf Hartnett

A THESIS

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

DOCTOR OF PHILOSOPHY
College of Education

1963

ACKNOWLEDGEMENTS

The writer wishes to express his sincere appreciation to
Dr. James Costar, Chairman of the committee, for his encourage-

ment and guidance.

In addition, the author is grateful to Dr. Arvo Juola,
Dr. Hans Toch, and Dr. Willa Norris, for valuable criticisms
and suggestions. Especial appreciation is extended to
Dr. Juola, who, in his role as Thesis Chairman, was of in-
valuable assistance from the initial planning stages of the

study through its completion.

A portion of the data reported herein was gathered as
part of a larger investigation of "Critical Thinking,
Attitudes, and values in Higher Education," sponsored jointly
by Michigan State University and the United States Department
of Health, Education, and Welfare, Office of Education. The
writer is grateful to Dr. Paul Dressel for his permission to
use some of this data, and to Dr. Irvin Lehmann for his many

suggestions and thoughtful considerations.

ii

TABLE OF CONTENTS

ACKNOWLEDGEMENT . . . . . . . .

LIST OF TABLES . . . . . . . . .

LIST OF ILLUSTRATIONS . . . . .
CHAPTER
I. THE PROBLEM . . . . . . . .

II.

III.

Introduction . . . . . .
Statement of the Problem
Factors to be Examined .
Definition of Performance

Change

Determining expected grade point

Determining change classification

The Hypotheses . . . . .

Scope and Limitations of the Study

Importance of the Study .

REVIEW OF THE LITERATURE

General Overview . . . .
Attitudes, values,

0

O

averages.

The Behavioral Variables and Academic

Changes in Scholastic Performance

Summary . . . . . . . . .

THE METHOD OF INVESTIGATION

Definition of the P0pulation
Classification of the Sample as to Change

Instrumentation . . . . .

The College Qualification Test .
The Test of Critical Thinking,
The Michigan State University Reading Test
The Inventory of Beliefs .

The Differential Values Inventory
Rokeach's Dogmatism Scale

The Experience Inventory .

iii

0

Procedure for the Collection of the Data
Procedure for the Analysis of the Data

Form G

and Scholastic Performance
Tests of Aptitude and Academic Performance

Page

ii

23

23
26
34

Performance 41

52
55

57

57
58
58
58
6O
61
62
64
66
68
72
73

IV. E

 

IV. ANALYSIS OF THE DATA. .

V. SUMMARY, CONCLUSIONS, DISCUSSION, AND

BIBLIOGRAPHY . . . . . . .

APPENDIX . . . . . . .

A Description of the Sample . . . .
Grade-point ranges . . . . . . . . . .
Correlation between annual grade

point averages . . . . . . . . .
Standard errors of estimate . .
Characteristics of the group . . .

Hypothesis I . . . . . . . . . . . . . .
Male students . . . . . . .
Female students . . . . . . . . . . .
Conclusions regarding Hypothesis I

Hypothesis II . . . . . .
Male students . . . . . . . . . . . .
Female students . . . . . . . . . . .
Conclusions regarding Hypothesis II

Hypothesis III . . . . . . . . . . . . .
Sophomore change . . . . . . . . . .
Conclusions regarding sophomore change
Junior change . . . . . . . . . . . .
Conclusions regarding junior change
Senior change . . . . . . . . . . . .
Conclusions regarding senior change .

FOR FURTHER RESEARCH . . . . . . . . . . .

Summary . . . . . . . . . . . . . . . . .
Purpose and procedure . . . . . . .
Findings . . . . . . . . . . . . . .

Conclusions . . . . . . . . . . . . . . .

Discussion . . . . . . . . . . . . . .

Implications for Further Research . . . .

iv

IMPLICATIONS

Page

75

75
76

78
79
83
84
85
87
89
9O
91
95
98
99
100
108
110
118
118
124

127

127
127
130
132
135
139

142

151

QC

 

LIST OF TABLES

TABLE

10

Grade point average ranges for
achievement groupings . . . . . . . . . . . . .
Intercorrelations of annual grade point averages,
male and female, for group maintaining continuous
enrollment, fall, 1958, to Spring, 1962 . . . . .
Grade point average discrepancy equivalents of
the .50 standard error of estimate reported for
sophomore, junior, and senior change points . . .
Means and standard deviations on instruments
used for ability-aptitude and attitudes-values
variables, with comparative data for entire
sample of students entering Michigan State
University in fall, 1958 . . . . . . . . . . . .
Descriptive break-down of students by behavioral
variables during sophomore, junior, and senior
year S O O O I O O O I O O O O O O O O O O O O O 0
Ability and aptitude mean scores of male students
classified by freshman grade point average and
direction of performance change during the
sophomore year . . .

. Ability and aptitude mean scores of female students

classified by freshman grade point averages and
direction of performance change during the
sophomore year . . . . . . . . . . . . . . . . .
Attitude and value mean scores for fall, 1958,
testing: male students classified by freshman
grade point average and direction of performance
change during the sophomore year . .

. Attitude and value mean scores for spring, 1959,

testing: male students classified by freshman
grade point average and direction of performance
change during the sophomore year . . . . . . . .

V

Page

76

78

79

81

82

86

88

92

94

 

Ty“ 2’ T
”has“.

13

H
(3"

F4
(j)

I
LL J

TABLE Page

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

Attitude and value mean scores for fall, 1958,
testing: female students classified by freshman

grade point average and direction of performance
change during the sophomore year. . . . . . . . . 96

Attitude and value mean scores for spring, 1959,
testing: female students classified by freshman
grade point average and direction of performance
change during the sophomore year. . . . . . . . . 97

Comparison of students classified by direction of
sophomore performance change, level of freshman
achievement, and applicability of personal factors
during the sophomore year . . . . . . . . . . . . 101

Comparison of students classified by direction of
sophomore change, freshman achievement level, and
hours employed per week during the sophomore year 103

Comparison of students classified by direction of
sophomore change, freshman achievement level, and
place of residence during the sophomore year. . . 105

Comparison of male students classified by direction
of sophomore change, level of freshman achievement,
and activities immersion during the sophomore year 108

Comparison of female students classified by

direction of sophomore change, level of freshman
achievement, and activities immersion during the
sophomore year. . . . . . . . . . . . . . . . . . 109

Comparison of students classified by direction of
junior performance change, level of sophomore
achievement, and applicability of personal factors
during the junior year. . . . . . . . . . . . . lll

Comparison of students classified by direction
of junior change, sophomore achievement level, and
hours employed per week during the junior year. . 113

Comparison of students classified by direction of

junior performance change, sophomore achievement

level, and place of residence during the junior

year. . . . . . . . . . . . . . . . . . . . . . . 115
vi

 

TABLE

61.

6,2.

20.

21.

22,

23.

24.

25.

*3 OJ .‘b'

,‘d
rhgltt

a:

TABLE

20.

21.

22.

23

243

255

26a

Page

Comparison of male students classified by direction
of junior change, level of sophomore achievement,
and activities immersion during the junior year . 116

Comparison of female students classified by

direction of junior change, level of sophomore
achievement, and activities immersion during the
junior year . . . . . . . . . . . . . . . . . . . 117

Comparison of students classified by direction of
senior performance change, level of junior achieve-
ment, and applicability of personal factors during

the senior year . . . . . . . . . . . . . . . . . 120

Comparison of students classified by direction of
senior change, junior achievement level, and hours
of weekly employment during the senior year . . . 122

Comparisons of students classified by direction of
senior change, junior achievement level, and place
of residence during the senior year . . . . . . . 123

Comparison of male students classified by level of
junior achievement, direction of senior academic
performance change, and activities immersion during
the senior year . . . . . . . . . . . . . . . . . 125

Comparison of female students classified by level of
junior achievement, direction of senior change, and
acitivities immersion during the senior year. . . 125

Analysis of variance data for high, middle, and low
achievement males on fall, 1958, Test of Critical
Thinking . . . . . . . . . . . . . . . . . . . . 152

Analysis of variance data for high, middle, and low
achievement of males on fall, 1958, College Quali-
fication Tests . . . . . . . . . . . . . . . . . 153

Analysis of variance data for high, middle, and low

achievement males on fall, 1958, Michigan State
University Reading Test . . . . . . . . . . . . . 154

vii

TABLE Page

9.2.

10.1.

10 2

10.3

Analysis of variance data for high, middle, and low
achievement females on fall, 1958, Test of Critical
Thinking . . . . . . . . . . . . . . . . . . . . 155

Analysis of variance data for high, middle, and low
achievement females on fall, 1958, College Quali—
fication Tests . . . . . . . . . . . . . . . . . 156

Analysis of variance data for high, middle, and low
achievement females on fall, 1958, Michigan State
University Reading Test . . . . . . . . . . . . . 157

Analysis of variance data for high, middle, and low
achievement males on fall, 1958, Inventory of
Beliefs . . . . . . . . . . . . . . . . . . . . . 158

Analysis of variance data for high, middle, and low
achievement males on fall, 1958, Differential Values
Inventory . . . . . . . . . . . . . . . . . . . . 159

. Analysis of variance data for high, middle, and low

achievement males on fall, 1958, Dogmatism Scale. 160

Analysis of variance data for high, middle, and low
achievement males on spring, 1959, Inventory of
Beliefs I I I I I I I I I I I I I I I I I I I o 161

Analysis of variance data for high, middle, and low
achievement males on spring, 1959, Differential
Values Inventory . . . . . . . . . . . . . . . 162

Analysis of variance data for high, middle, and low
achievement females on fall, 1958, Inventory of
Beliefs I I I I I I I I I I I I I I I I I I I I I 163

Analysis of variance data for high, middle, and low
achievement females of fall, 1958, Differential
Values Inventory . . . . . . . . . . . . . . . . 164

Analysis of variance data for high, middle, and

low achievement females on fall, 1958, Dogmatism
Scale I I I I I I I I I I I I I I I I I I I I I I 165

viii

 

5;;
m
E;

11.1. Are
ac":
Be.

11.2, Aid;

'v'al

TABLE

11.1.

11.2.

Page

Analysis of variance data for high, middle, and low
achievement females on Spring, 1959, Inventory of
Beliefs . . . . . . . . . . . . . . . . . . . . . 166

Analysis of variance data for high, middle, and low
achievement females on spring, 1959, Differential
Values Inventory . . . . . . . . . . . . . . . . 167

 

_—s-. -w- r -

Figure 3;,

1, Re

1

1‘

LIST OF ILLUSTRATIONS

Figure No.

Page
1. Regression of freshman grades on sophomore grades,
illustrating achievement classifications and
direction of performance change . . . . . . . . . l4
2. Grid used for weighting student activities and
specific positions or functions in that activity. 70

3. Progressive mean grade point averages, male and
female, for group maintaining continuous enrollment,
fall, 1958, to spring, 1962 . . . . . . . . . . . 78

 

 

Per
least i:
has rec
topic I
ing out
by soar
C1112e:

SC3-001:

CHAPTER I

THE PROBLEM

Introduction

 

Perhaps no area of research in higher education -- at
least in the fields of psychology, sociology, and education --
has received as much attention in the last decade as the broad
topic referred to as "research on college students." Originat—
ing out of pressures placed on the colleges and universities
by soaring enrollments, tight budgets, and questioning
citizens, inquiries into the four years of undergraduate

schooling have increased rapidly.

Joining with those already probing the area, psychologists
and sociologists have recently recognized the vast potential
for social psychological research studies of collegestudents.
Consequently, we have seen a marked increase in the investi—
gations of the environment of the college campus and its
various subcultures. Along with this, there has been a natural
growth of interest in research dealing with the personality
of the individual college student, and especially the person—

ality changes that take place during the undergraduate years,

presumably as a result of some facet of the college experience.

Prediction of college achievement, of course, has been
studied with a number of approaches. Measures of scholastic
aptitude, specific subject-matter ability, interests, socio-
cultural backgrounds, high school performance -- all of these
and others have been used as variables in attempting to pre-

dict college performance.

There have been a large number of studies dealing with
over and under-achievers, as well as "drop-out" or retention
and withdrawal of students investigations, characteristics of

diverse curricular area majors, and the like.

Limiting the discussion to research on college students‘
academic performance alone, it can be seen that roughly four
categories of current "types" of research exist: (1) census—
type studies, which generally concern themselves with per-
centages, such as percentage of students withdrawing, per—
centage of students graduating, etc., (2) autopsy-type studies,
which usually are post facto studies in the sense that they
study characteristics of students who have withdrawn or been
dismissed, (3) case—study approaches, which concentrate exten—
sively on small samples of selected students for various

reasons, and (4) prediction studies, generally attempting to

3

relate selected admissions variables to success or failure

criteria (51).

Research findings in these areas will be further discussed
in Chapter 11. They are listed here, however, to illustrate
the particular dearth of information which typifies our
knowledge of what factors are associated with patterns of
academic performance during the student's years in college.

The existing research on college students' academic perform-
ance has concentrated on two points in the student's course --
early college achievement and the eventual record at graduation
or withdrawal. This approach has overlooked the intervening
performance pattern. This assumes that scholastic perform-
ance patterns are stable, an assumption which is not necessar-
ily true. Students are generally regarded as being high
achievers, low achievers or "average" achievers. The possible
deception in this standard classification is that the student
may be all of these at different times in his undergraduate
years. The "average" achiever, for example, may actually be
one whose term-to-term pattern consists of peaks and valleys --
high performance one term, low the next term, and so on. Or

he may actually be a high achiever, but one whose over—all
grade point average is depressed by low performance during

his initial terms, or by a short—term period during later

terms.

4

The point being made here is essentially this: that
while for some students the performance pattern in college is
quite consistent from their freshman year to their time of
graduation or withdrawal, this is often not the case, and
that, in contrast, many students' patterns of attainment are
characterized by gradual downward or upward trends or a
"roller-coaster" up and down model. Consequently, to judge
one's performance pattern by a cumulative grade average may
be quite deceptive, for the process of averaging may actually
conceal the most revealing index of the present level of

achievement.

Those familiar with and involved in higher education can
readily recognize the importance of understanding more about
the student‘s performance pattern as Opposed to a cumulative
index of achievement. A student with an over-all grade point
average of 2.5 might well be overlooked as one who possibly
has academic difficulty, yet if it is understood that this
2.5 represents the average attainment, and his term-to-term
pattern consists of a number of changes in his level of
attainment, then further consideration should be given to

this student's performance status.

While this kind of performance analysis has been attempted

in modified forms at several universities (56, 87) there has

5
been, up to this point, a failure to ferret out the factors
which might have some bearing on inconsistent performance
patterns. If it can be demonstrated that student X, who has
a grade point average of 3.0, is nevertheless a student in
need of some assistance because of his instability of perform-
ance, the probem is only recognized. The question must still
be answered: What factors seem to be associated with gross

changes in college students' levels of scholastic attainment?

Various studies have indicated a relationship between
such variables as attitudes, values, and beliefs, to changes
in one's curriculum (51, 54). These same variables have been
demonstrated to bear some relationship to withdrawal and
attrition rates (43). Certainly such variables as attitudes,
values, and beliefs would possibly be associated.with sizable

changes in academic performance.

There has also been a considerable amount of research
regarding the mutual association of various student activities
(i.e., fraternities, student government, etc.) and various
levels of attainment in college, eventual graduation, or
subsequent withdrawal. Once again, such variables would con-
ceivably be important to consider in any investigation of

Chan es in one's academic erformance attern.
P

6

The academic performance of a college student may simply
be regarded as a specific sample of behavior. Like other bits
of behavior, it is susceptible to change and is affected in
various ways by different situations and previously held
diSpositions to behave in certain ways. Unfortunately,
however, too few recognize that performance patterns change,
and even fewer have attempted to isolate any factors associated

with such changes.

Statement of the Problem

The problem investigated in this research is the relation—
ship between various selected factors and changes in students'
academic performance patterns during four years of under-
graduate study at Michigan State University. The factors to
be studied include certain measured "personality" variables,
measured indices of aptitude or ability, and various bio-
graphical data pertaining to employment, the amount of
participation in extra-curricular activities, place of resi-

dence during school, and the like.

It is hoped that the findings reported herein will be of
value in better understanding some of the basic reasons re-
lated to students' inconsistent grade achievement during their

four years in college. While the investigator realizes that

7
mere association of various factors with irregular patterns
of performance does not necessarily mean causality, it is at
the same time recognized that exploratory studies must come
first if we hOpe to isolate the conditions associated with

academic inconsistency.

Factors to be Examined

The factors to be examined can be classified into three
categories: (1) abilities and aptitudes, (2) attitudes and

values, and (3) behavioral variables.

The ability and aptitude factors may be further delin-
<aated.into general aptitude, reading ability and critical
‘tliinking ability, all three which were measured by instruments

discussed in Chapter III.

The factors falling under the "attitudes and values"
<ZEttegory include dogmatism, personality rigidity, beliefs,
and values. These characteristics were also measured by

instruments described in detail in the third chapter.

Finally, those factors which have been given the label
"laehavioral" include amount of participation in extra—
curricular activities, amount of weekly employment during

the school year, place of residence during the school year,

8

and events or occurrences that may have been peculiar to the

individual student. This information was gathered partially

by questionnaire and partially by examination of the students'

records, and, once again, will be discussed in detail in

Chapter III.

Definition of Performance Change

In the final analysis, regardless of the statistical

manipulations involved, any definition of what is a sufficient

,performance change must be arbitrary. The procedure used to
determine the students whose performance pattern changed

<anough to merit further consideration, had to first cope with

tflne problem of regression. Because of errors in measurement

‘tllere is a normal tendency for any imperfect measure (in this

(38186, a grade point average) to converge toward the mean

Score for the entire group. To insure against this statistical

ar‘tifact confounding the present study, a regression analysis

Pr ocedure was utilized .

On the basis of their freshman year performance, regres-
Sion equations were calculated for the group, yielding the
e)iipected performance of the students for the sophomore year.

'Because of a sex difference in the grade point averages, the

regression formula was computed separately for males and females.

9
The formula used for determining the expected grade point
average was (59):

x
)<'==1H!¥- Y’ + (M — rq;§.My)
x q'y

where (using as an example, the prediction of
sophomore grade point average on the basis of
freshman grade point average):

o'x = standard deviation of s0phomore g.p.a.'s

v‘y = standard deviation of freshman g.p.a.'s
Mx = sophomore year mean grade point average
My = freshman year mean grade point average

Y = freshman year grade point average

After the entire group had been used in the regression
:procedure to determine their expected sophomore grade point
aaverage, the students were classified into three groups —-
liigh, medium, and low, on the basis of the scholastic perform—
ance during their first year in college. This three-way

Cijnvision'was necessary since the cognitive factors to be
Gaacamined are highly related to achievement. Therefore, it
VVEls considered that a rough achievement grouping such as the
(Drbe employed would reduce the possibility of this factor con-
tEuminating significant associations if they were to be found.
TTiestratification procedure was done by simply calling those
CH1 the lower third of the first year grade point average dis-
1liribution the low group, those on the upper third the high

group, and those falling between these two points the middle

10
group. The actual g.p.a. ranges for the groups are presented

in Chapter IV.

For the successive years, the threesway stratification
and basis for prediction of the next year's achievement level
was always moved up one year. In other words, when predicting
the sophomore level of achievement, the achievement grouping
was done on the basis of the freshman achievement and the
freshman grade point average was used as the predictor
variable. When predicting the junior year level of attainment,
the achievement grouping was done on the basis of sophomore
tachievement and the soPhomore grade point average'was used as
the predictor variable. Finally, when predicting the senior

ynaar grade point average, the achievement grouping was done
<Dr1 the basis of the junior year attainment, and the junior
Year grade point average was used as the predictor variable.
133' this procedure, it was assured that the standard "over and
‘Jlixier achiever" patterns would not be the criterion of per—
formance change, but that the student would have to change
from term-to-term on the basis of his or her most recent

level of achievement.

Comparison of a student's predicted performance with his
aCtual performance then resulted in a discrepancy score. To

illustrate: one of the students involved in this study

11

received a freshman year cumulative grade point average of 2.50
which placed him in the middle achievement category for the
freshman year. By means of the regression equation calculated
for the group, his expected score for the soPhomore year was
2.52. He received a true sophomore grade point average of
3.11, the difference between his actual grade point average

and his expected or predicted grade point average is 3.11 -

2.52 = .59.

The obtained discrepancy score was considered to be large
enough for"change" classification if it was more than .50
standard errors of estimate. The confidence band of .50
standard.errors of estimate was chosen for several reasons.
Phirst, it is large enough to require that a fairly wide

éLiscrepancy between observed and expected achievement levels
E3c2tually occur before one is labelled a "changer." Second,
lit: conveniently divides the sample in approximate thirds, and
finally, use of a larger standard error of estimate would have
rEisulted in change groups too small for statistical analysis,

especially for the female sample.

12

The formula used for determining the standard error of

estimate was (2):

 

*-

— ‘\I l r2
‘<Fest. -<Fy - xy

where: r2 =

XY validity coefficient (correlation

between predictor variable and
criterion variable)

<r'y = standard deviation of the criterion
variable.

Returning to our previous example, the discrepancy of .59

between the predicted grade point average and the true grade
point average would be sufficient to place that student in a

"change” classification, since it surpasses the .50 standard

«error of estimate band (cut-off points for grade point

eaverages using the .50 standard error are also presented in

Cflnapter IV.) Therefore his upward change (.59 grade points

115) from his expected level) would classify him as medium

achievement, positive change.

It can be seen, from what has been said, that for any one

<31‘1ange point (sophomore, junior, or senior), a student may

faillinto one of nine possible classifications:

13

Direction of Performance Change

 

 

 

 

 

 

 

 

‘— O +
(Neg.) (Freeze) (Pos.)
High 1 2 3
Achievement
Medium 4 5 6
Levels
Low 7 8 9
once

This procedure was repeated three different times,

for each of the three observed achievement points at the end

of the SOphomore, junior and senior years. The intent

I

of

<:ourse, was to examine the variation in performance patterns

.frtmlyear-to-year and not just for the freshman to sophomore

transition. Figure No.

l on page 14 presents the outline of

‘tlie regression model used to select the achievement groups and

The numbers 1 - 9 in parentheses

‘tlie performance changers.

<=<>incide with the chart on this page.

It is important to keep in mind that the grade point

average used as the indices of achievement during the sopho-

mere.

aVerage .

junior and senior years was not a cumulative grade point

It was the level of academic attainment achieved by

‘Ehe student during that particular academic year only.

Utilization of a cumulative index would not have revealed

304

14

ZDHQMZ

mem

 

.mmoouw ucofim>mflnum

mmmum>¢ ucflom momma mHOEonnom

o.m ¢.m N.m o.m m.N 9N .v.N N.N o.N m.H via a...” NJ” o.H

 

 

 

 

8v

 

 

 

 

 

 

 

 

m
V
3
\
JJ
m1
0!
\ I
h
H
V

 

 

\

 

 

\

 

 

 

eSexeAv auroa apexs uemqsaxg

15

performance patterns, since

realistic changes in the students‘
the inclusion of the past totals would either depress or raise

the present level of performance, thereby disguising radical

performance changes.

The Hypotheses

As the review of literature presented in the next chapter

demonstrates, much of the existing research bearing relevance

to this investigation is either vague, inconclusive, or even

contradictory. Moreover, the concept of scholastic performance

change used in this study has not been employed previously in

Iresearch dealing with the three classifications of variables

loeing analyzed. As a result, the hypotheses, although evolv-

iJng out of the general review of literature, lack the tradi-

‘trional framework of a stable theory. Keeping in mind, however,

‘tliat this is essentially an exploratory investigation, it was

<3<>nsidered necessary to define more clearly the direction of

tirre investigation. Consequently, taking care to avoid the

1‘11111 form, the following hypotheses were used as guidelines

3501' the study:

1) It is possible to differentiate among groups of
students classified by direction of academic
performance change, on the basis of their scores

on ability and aptitude measures.

l6

2) It is possible to differentiate among groups
of students classified by direction of academic
performance change, on the basis of their scores
on attitude and value measures.

3) It is possible to differentiate among groups of
students classified by direction of academic
performance change on the basis of the amount
of their employment, their place of residence,

degree of participation in extra-curricular
activities, and various personal factors.

Sc0pe_and Limitations of the Study

Not unlike many research studies dealing with college
satudents' academic performance, the present investigation was
heunpered by several limitations. While recognized as factors
tc>‘be considered when drawing inferences from the findings,
‘hcnwever, it is felt that the limitations listed are not

Stufficient to offset the importance of the data presented.

One of the first limitations concerns the sample. Only
tkuzse students who remained in continuous attendance (except
fOI"the summers) for four years following the term of their
enr'Ollment were included. Obviously, students who withdrew
frCHn the university were not among those for whom.it was
POSEiible to compute the attainment patterns in the sophomore,
junixor and senior years, depending on the time of their with-
draWal,

But neither were their levels of attainment used in

'U“3 regression equations used in arriving at predicted levels

17
of achievement. It was felt that inclusion of the grade point
averages of those students who dropped out of the university
would lend a spurious change pattern to those who remained in

school for the entire four years.

A second limitation is one encountered by numerous inves-
tigators employing data pertaining to obtained measures of
gpersonality. These measures, in this study, were used to
(determine such personality traits as dogmatism, rigidity,
txeliefs, and values. The typical approach to measuring these
\narious characteristics is to elicit from the respondent some
iruflication of his feeling or judgment on an empirically de-
rtrved affective scale. There is no guarantee that this
response is an honest one, nor, assuming that it is honest, is
thueir any evidence that this felt attitude has any stability.
Nevertheless, this is a limitation faced by nearly all inves-
ti43ators engaged in research of this nature, and if it is to
bereconsidered as a limitation, it certainly is not one which

is peculiar to this particular study.

There are other limitations concerned with the grades of
the= students, the variable used to construct the entire study.
PrCflDably most obvious is the problem of different grading
Staruflards. The utilization of grade point averages for

vanPious statistical manipulations sudh as part of a regression

l8
equation, or simply a rough classification procedure as that
employed here, makes the assumption of uniformity of grade
assignment. However, what is work of "A" calibre for one
‘professor may only be deserving of a ”B" for another instructor
in the same course. As a result, changes in the academic
;performance patterns of the students may possibly be the
:result of different faculty standards, not any factor dealing
vwith the student at all. Different curriculum standards (i.e.,

tuuequal standards in different academic departments) is a

stnilar difficulty.

The students who carry part-time loads introduce another
type of limitation under the grading problem. It is possible
tIHat changes in one's pattern of attainment may actually be

tIue result of his credit-hour-load. Only grade point averages

were considered here, and consequently those carrying one or

tVKD courses were not separated from those carrying full

academic schedules .

Still another limitation having to do with the grade
factuor is the problem of repeated courses. Students receiving
unsEl‘tisfactory grades in courses have the opportunity of re-
Peailing that course with the good probability that their grade
Wil-l'be improved. No extra credit is given when the course is

repea¢ed, but the honor points earned are substituted in place

19
of those deserved during the student's first time in the
course. As a result, a student could conceivably go from a
low level of achievement to a very high level, as long as his
performance in the repeated courses is much better. Again, the
change in performance pattern may actually be a function of

:repetition of courses, rather than any of the variables being

exendned in this investigation.

These limitations regarding the grading factor, while
lJending a proper note of caution to the interpretation of the
rtesults, do not render meaningless the grade-based procedure
ennployed. In spite of the recognized limitations, grade point
axnerages still represent reliable evaluation and certainly

I
reunain our best estimate of student performance.

Importance of the Study

The importance of this study is best illustrated through
coI'lSSiderations of the basic substantive and methodological
in‘Plications. Analyses of patterns of academic performance
in-lieu of a cumulative grade point average offers a different
PrOCedure for evaluating student progress and the effects that
different experiences have on this scholastic performance.
SOWKB of the techniques employed for measurement of the be—

lfifllioral variables (i.e., the Activities Index) also present

20
a relatively new way of viewing the student sub-culture. In
these respects, the importance of the study basically lies in
its method of investigation. It is important because it is
suggesting a new approach to an old problem and denotes a
further contribution of research to our body of knowledge. This
study may not provide clear-cut information of practical mean—
ing to anyone, but if it serves as a "springboard" for allied

:research, then its importance is recognized.

But its importance can also be illustrated through con-
saideration of the implications this research holds for both
tflne students and the university. While it is agreed that
rwesearch need not be of a pragmatic nature to be meaningful,
it: is also felt that an understanding of some of the appli-
cartions of the findings may lend more significance to educa—

t ional research .

In terms of the student, analysis of factors associated
‘Wiifli academic performance patterns may yield discoveries having
Valfiious implications. Students experiencing difficulty in
mailitaining a stable pattern of achievement, but who are
unalble to understand the reasons for their inconsistency.
miglrt be better able to appreciate their personal situation.
Exierting research, for example, indicates that participation

in extra-curricular activities has little or no effect on the

21
scholastic achievement of college students (74). If, however,
it is revealed that periods of peak activity in the various
campus organizations are associated with a downward trend in
the student's level of performance, then appraisal of one's

outside-of-class activities may be pertinent.

In terms of the university, the research related to
students' performance patterns has multiple implications.
Eketter counseling could be provided those students who are
txeset with difficulties related to negative fluctuation of
tflieir academic performance. If it is revealed, for example,
tluat personality rigidity is a factor associated with students
\vhuase pattern of attainment is somewhat of a "roller-coaster“
type, then it is conceivable that a more direct approach to

resolving the situation could be available.

Finally, those involved in student personnel administra—
‘ticnu might find the results of this study meaningful and
fihelqpful as guidelines in the establishment of various policies
regarding student housing, fraternities and/or sororities,

Partlicipation in out-of-class activities, student employment

and the like .

A critical reivew of the literature related to this study

Wil-lbe presented in the following chapter. The findings of

.(y
.1.“

 

22
previous investigators will be reported and discussed. In
Chapter III an account of the methodology of the study is
presented, including a description of the sample, a discussion
of the instruments used, procedure for gathering the data, and
the processing and analysis of the data. In the fourth
chapter the results of the analysis are presented, and the

fifth and final chapter of the thesis contains the summary.

CHAPTER II

A REVIEW OF LITERATURE

While it is true that research on college students has
increased notably in the past decade, it is also true that
there is a great need for more which is of high quality. This
is especially true in the area of college students' academic
performance patterns. As suggested earlier, the current
literature is addressed primarily to the "pre—" and "post-"
points, and tends to overlook completely the pattern during

college.

The present chapter is the result of an attempt to
review that research which has some bearing on the thesis. It
discusses both the findings regarding the "factors" under

examination and also the various instruments used.

General Overview

 

Prediction studies and drop—out studies have proven to
be the paths most often chosen by those interested in explor-
ing the area of college students' academic performance.

Although not directly related to an analysis of changes in

23

24
performance patterns, a number of these investigations shed
considerable light on the over-all problem of student achieve-

ment and many of the variables involved.

Probably the most comprehensive report on student with—
drawal at the time of this writing, is a census study by Robert
Iffert (42). Sponsored by the United States Office of Health,
Education and Welfare, Department of Education, Iffert surveyed
nearly 150 colleges and universities in 1950. Foremost in his
findings were the revelations that approximately 50% of a
college's student body departs from school within four years
after beginning, about 40% graduate "on schedule," and
roughly 20%.go on to graduate from some college some day. In

this regard, Iffert's study is supported by investigations of

others (58,76).

(College drop-outs apparently do not differ significantly
in seu< (46,63,75), or age (18,77), but can be distinguished
on tlue basis of certain socio—economic characteristics such
as ferther's occupation (75), parents' education (64) and size
0f htnnetown (22). Almost all studies of drop-outs have re-
Vealeua significant differences in high-school performance.
earlfir college achievement and scores on various measures of

aptiinide or ability, with the drop-outs being lower in all

three instances .

25
Of the wealth of information garnered from drop-out

studies, that which has the most relevance for the present

investigation may well be found in the area of motivation.
{That various motivational factors are extremely important in
‘the evaluation of patterns of scholastic performance is a fact
lnighlighted by most of the studies which consider motivation.

Iffert‘s study (42) revealed that of 1450 withdrawals, 48%

of the men and 33% of the women report disinterest in their

studies as a primary reason for dropping out. Freedman (24)

and.Farnsworth (18) also report lack of motivation as being

an important reason for withdrawal.

Vocational motivation is even more obviously related to

Ehzholastic performance. A number of investigators have shown,

Iftor example, that students who have made definite vocational

<2}ioices are more likely to be successful college students

Eacademically (23,42,69,81).

Generally, however, drop-out studies have been far from

£3Eitisfactory in terms of answering questions regarding college

5students' ongoing performance. They have had some influence,

‘kKDwever, on our understanding of a terminal point in one's

Collegiate career .

26

Attitudes, Values, and Scholastic Performance

Not previously considered as factors having any bearing

<3n academic performance, there has been a recent growth in

‘the amount of research attempting to relate attitudes and

‘values to academic performance in college.

Heist and Williams (35), in a study of the freshman

class at California Institute of Technology, attempted to
determine some of the correlates of different levels of

achievement within a relatively homogeneous group of high
abi l ity students .

Dividing the all-male group into three achievement

gfiroups (high, medium, and low) on the basis of their freshman

Year grade point average, they then obtained the students‘
Ssczores on various inventories, including the Omnibus Personal—
-icty'Inventory, the Strong Vocational Interest Blank, The
Ikillport-Vernon—Lindzey Study of Values, and two specific

13Giographical items relating to student interests and expec-
t at ions .

They found that the primary distinguishing feature among

lihe groups was a tendency on the part of the high achievers

‘to value more highly a strong orientation toward inquiry, and

27

speculative and creative thought.

Ehrlich (l6) explored the relationship between students‘
(degree of dogmatism and their academic achievement. Using
IRokeach's instrument to measure dogmatism, the author hypoth-
‘esized that if dogmatism implied a "closed cognitive structure,"
'then this would have some effect upon one's capacity to learn,

independent of academic aptitude. The results confirmed

Ehrlich ' s hypothesis .

Lehmann (53) examined the relationships between various
<:ognitive and affective variables to student performance in
cuemmon courses. The cognitive and affective variables in
tliis case were six in number: the Inventory of Beliefs, A
Thest of Critical Thinking, Rokeach‘s Dogmatism Scale, Prince's
JEMifferential Values Inventory (affective measures), and the
C2<>llege Qualification Test and Michigan State University

I{eeading Test (cognitive). The dependent variable of student
IDeerformance was designated as achievement in the required
jflreshman courses of Communication Skills and Natural Science,

Elsa well as the first term grade point average.

For the most part, Lehmann was unable to establish a
13'<alationship between the affective variables and the criteria

(bf scholastic performance. The correlation of greatest

28
Inagnitude for the Differential Values Inventory, for example,
xdas .11 with freshman year g.p.a., while the others using this
‘variable were in the neighborhood of .05. The correlations
ihetween ROkeach's Dogmatism Scale and academic performance

swere of a somewhat higher magnitude.

The author concludes that while the affective measures
tased did not result in any significant relationships with
zacademic performance, the inclusion of affective measures can
.increase multiple correlations between cognitive variables

and academic performance.

Findings somewhat contradictory are reported by Neel (62).
IJsing the F-scale, a measure used by Adorno, Brunswick, et_al.,
111 their study of the Authoritarian Personality (1950), Neel
Eitztempted to analyze the relationship between degree of student
IPITejudice (subsequently labelled "authoritarianism") and
<2ZLassroom performance. Her findings suggest that authoritar-
jLeanism lg related to certain kinds of learning, and that
Eintthoritarian subjects would experience difficulty in absorbing

$3Ilbject matter which is of a humanitarian nature, such as

Social science .

Howard and Warrington (40) found a relationship between

Student beliefs and academic performance. They administered

29
the Inventory of Beliefs to students at the beginning of their
freshman year, at the end of their freshman year, and once
again at the end of their senior year. For the females, there
xuas a significant positive relationship between the Inventory
«of Beliefs and senior grade point average. However, there
suas not a significant correlation between test score and grade
jpoint average for the males. The authors also report that the

«grade point average did not correlate significantly with gain

scores on the Inventory of Beliefs.

Using a modified form of the Semantic Differential to
Ineasure values, Winter (85) attempted to guage the similarity
inn values between a group of college students and their
iJistructor, and to relate this similarity to their academic
achievement in his class. The Semantic Differential, as

<3£escribed by Osgood, is designed to elicit responses regarding
1Zl'mree concepts: value, strength, and activity. Winter's study,
1‘icbwever, was concerned with the value dimension, and only

illuose adjectives with highly-intercorrelated value concepts

Vveere included.

Administering the Semantic Differential to 34 freshmen
nmen enrolled in a required course in general psychology,
VVinter compared their obtained scores with their instructor

‘Who had earlier taken the same inventory. The students were

30

also asked to predict how they thought their instructor would

answer the same items.

Winter's findings suggest that the more similar in values
a.student was to his instructor, the higher was his achieve-
Jment in class. The investigator suggests that since the
:professor is generally a scholarly person, a student with
discrepant values would be a non—scholar, one who is not
Inotivated to read and study. In any event, the findings do

indicate the importance of non-intellectual factors, especially

'values, in college achievement.

Just a little more than a year after Winter's study was
rueported, he once again published the findings of research
Ikelated to student values and academic achievement (84). In
‘tlue latter investigation, Winter used a slightly larger sample
of students enrolled in two general psychology courses, and
1Z‘vo special procedures: the first involved ranking fifteen
<3¢ifferent concepts as to their value for the individual
n[leaking the ranking, and the second involved re-ranking these
<2<Dncepts as the student thought his instructor had ranked
firmmn These "concepts" were play, professor, college, cheat-
Zing; good time, easy money, work, studying, homework, books,

<grades, athletics, tests, social activities, and research.

Once again, Winter concludes that grades in psychology were

31

correlated with the degree of similarity in values between a
student and his teacher. On the basis of his findings in the
Inore recent study, however, he also concludes that academic

jperformance is related to the degree of democratic, non-

ethnocentric values the student holds.

Red, McCary and Johnson (67) conducted a study attempting

to determine the degree of relationship between some measures

of aspiration and academic achievement. Tests administered

in the investigation included an adaptation of Cassell's Level

(of Aspiration Test, Otis Quick Scoring Mental Ability --

(Samma, and Revised Beta. These three instruments were given

tx: 101 freshmen and sophomore students and their scores on
tflnese measures were compared to their grade point averages.
The investigators report that all correlation coefficients
iirad comparisons between measures of academic achievement and
lTuaasures from the aspiration tests show little or no relation-
Egllip between the two variables. Unfortunately, however, the

(ieescription of methodology employed in this study was all too

brief. Although the reader is informed that the correlations
between measures of asPiration and measures of academic per-

fkarmance are insignificant, no correlations are presented for

the reader ' s appraisal .

32

In an interesting piece of research regarding conformity
and achievement in college, Erb (17), has attempted to develop
a technique for measuring conformity and to relate this measure
to college grade earning ability. Conformity and non-conform-
ity were determined by a procedure similar to that used in a
Rorschach ranking test. The subjects, who were 25 males and
25 females scoring immediately above and below their respec-
tive medians on the School and College Ability Test upon
entering college, were asked to respond to a series of state-
ments dealing with their self-concept. Conformity or non-
conformity was then based on pupular versus infrequent
placement of the items in a Q-sort. Popular placements were
interpreted as conformity, while idiosyncratic placements were
used as indices of non-conformity.

Erb found that the higher conformity females achieved
better than low-conformity females at beyond the .05 level
of confidence. The mean grade point average for the lower
conformity females was 2.16, while the mean g.p.a. for the .
higher conformity females was 2.61. For the males, however,
no differences were found. The mean grade point average for
the lower conformity males was 1.77 as Opposed to 1.49 for
the higher conforming males. While this difference was not

demonstrated to be more than a chance occurrence, it is

'4.
n

It.

7" .

x
u-

‘1
B1

1‘

I

111

‘i

l)!

5..

33
interesting to note that the conformity-achievement pattern

nevertheless reverses itself for the males.

Although Erb's analysis is subject to several methodokxp-
iJzal shortcomings (i.e., no statement of theory, failure to
state rationale used in sample selection and size, incomplete
information regarding the Q-sort procedure, and a confusing
report), his findings nevertheless suggest that conformity,
‘which is essentially a value regarding adherence to norms,
ndght be a meaningful factor to consider when analyzing the

complex pattern of scholastic performance.

Many other studies have been reported which have examined
tine relationship between non-intellective factors and scholas-
tLic performance in college. For the most part, however, these
iJnvestigations have concerned themselves with measures of
IENarsonality characteristics which are more temperament-oriented
JTEather than attitude or belief directed. An interesting
Eattempt at develoPing an instrument which sides with the
:Liatter approach has been reported by Fricke (27). Essentially
5i device for measuring achievement motivation, Fricke's
iJnstrument, The Opinion,Attitude, and Interest Survey, has
1been demonstrated to correlate .40 with grade point averages
\lsing a sample of 209 students. For the most part, however,

existing inventories or "tests" lie in areas other than the

34
attitude or value domain. Descriptions of the instruments

used in this investigation will be reported in Chapter III.

Tests of Aptitude and Academic Performance

The relationship between intellectual ability or academic
aptitude and scholastic performance in college has been

explored by many investigators.

A study at the University of Wisconsin by Lins and Pitt
(56) involved a group of nearly 2,000 freshmen. Using number
of semesters in school as the criterion of academic perform-
ance, the authors noted that the higher the percentile rank
of a student on the American Council on Education Psychological
Examination, the more semesters he was likely to remain in
School. The relationship was not as large as we might expect,
hO‘Mever, as the examiners reported a correlation coefficient

of .29.

A study at the State University of Iowa (44) reveals a
definite relationship between percentile rank on the S.U.I.
Placement Tests and persistence in college. Using a procedure
nearly identical to that employed by Lins and Pitt, the Iowa
investigators found that a much greater percentage of high-

ability students go on to receive a college degree. For

35
example, of 319 students whose composite placement scores were
below the 30th percentile, only 71 received degrees, while of

341 students above the 70th percentile, 173 received degrees.

Both of the foregoing have a great deal of support.
Summerskill, in a review of the research in this area (76) ,
found 19 studies of scholastic aptitude and its relationship
to college attrition rates. He reports that average scholas—
tic aptitude scores were found to be lower for drop-outs than

for graduates in 16 of those studies.

A study dealing with student performance rather than the
withdraw-remain dichotomy, is reported by Juola (47). Follow-
ing an entire entering class at a large university through
their first two years of college, correlations were computed
betWeen scores on the orientation tests (made up of an
English Test, Arithmetic Test, Reading Test, and The College
Qualification Test) and term-by-term grade point averages
OVer the two year period. In discussing his findings, Juola
concludes that while aptitude tests do bear considerable
relationship to academic performance, the measures used in
this study became substantially less precise as predictor
Variables beyond the first term. The trend suggests that

predictive validities decrease from .60+ the first quarter to

-30 and less the sixth term. There was also data presented

36
to show that this same pattern appears to function whether
specific course grades are used or whether the grade point

average is considered.

In a later study designed to examine the relative “pre-
dictability" of five college-level academic aptitude tests,
Juola administered the College Qualification Test, Form B,
the School and College Ability Test, Form A” the AHC.E., the
Ohio State Psychological Examination, and the Scholastic
Aptitude Test, to differing samples of college freshmen (48).
Using the freshmen grade point average as the criterion of
academic achievement, correlations were computed between
students' scores on the five measures of academic aptitude
and their early college performance. The resulting matrix
led Juola to conclude: l) the coefficients obtained from the
five different instruments were almost the same, 2) the
correlations were higher for the females, 3) total scores on
all the tests were generally as good an index of attainment
as the most relevant sub-score, and 4) the College Qualifi—
cation Test seemed to have more applicability for the male

population and the SCAT for the female population.

In a third investigation of the applicability of aptitude
tests for prediction of college attainment (49), Juola

examined only the College Qualification Test, but analyzed

an;
5i

-nv. -

i ‘j.

 

37
its relationship to performance in diverse curricular areas.
Using both zero—order and multiple correlations, and students
divided into six samples according to their curriculum, the
criteria used as level of academic attainment were the first
quarter grade point averages. On the basis of his analysis,
Juola concludes that the total CQT battery seems to be as
efficient as Optimal weighting of subscores in multiple
regression when predicting attainment of students in the non—
technical areas,but that multiple correlations using the sub-
scores yields coefficients of greater magnitude when predicting

grades in the technical curricula.

John L. Holland (37) performed a study testing the use-
fulness of combining a personality test with an aptitude test
in predicting scholastic success. Using the California
Psychological Inventory and the Scholastic Aptitude Test,
Holland analyzed their relationship, both separately and
combined, for a sample of exceptionally talented college

freshmen attending 291 colleges and universities.

After analysis of his findings, Holland concludes that
both the CPI and the SAT are useful in predicting college
freshman grades for a sample of high aptitude high school
seniors, but that combining the two instruments and using a

multiple regression procedure results in a coefficient that

38
is two or three times as large as the zero-order correlations.
Unfortunately, Holland neglected to include the figures for
the multiple-correlation in his report, and the reader is left

with no basis on which to make a personal judgment.

Reporting findings consistent with those presented by
Juola in an earlier study reviewed, Seashore (71) presents
data leading to the conclusion that in terms of their collegi-
ate academic performance, women are more predictable than men.
He presents evidence acquired by utilization of both the
Differential Aptitude Test battery and the College Qualifi-
cation Test, and points out that of all the correlations
developed between aptitude and academic performance by use
of these two measures, by far the greater number favor the
women. In fact, of those correlations revealing significant
differences, 71%.favor the women. As Seashore suggests, such
findings probably raise more questions than answers, and the
first question which comes to mind at this point is how do

the performance patterns of men and women compare?

In order to examine the validity of the General Aptitude
Test Battery as a predictor of college success, Sharp and
Pickett (72) analyzed the relationships between the various
aptitudes measured by the G.A.T.B. and grade point averages

of 262 college juniors and seniors. Each of the nine

39
“aptitudes“ were correlated with the grade point averages of
the entire sample and also with the performance of students‘
categorized into engineering, business administration, edu-
cation, and physical education. The investigators found
many of the correlations to be significantly greater than
zero, but none exceeded .50. Those students who were success—
ful in Engineering course work ("successful" meaning their
cumulative grade point average was above 2.10) scored signifi-
cantly higher in verbal, numerical, spacial, general "intelli-
gence" and form perceptual aptitudes, while scoring somewhat
lower in the manual dexterity aptitude when compared to

Engineering students whose cumulative g.p.a. was below 2.10.

Combining high school performance with the School and
College Ability Test, Klugh and Bierley (50) attempted to
determine the validity of the SCAT against a criterion of
first semester grades in college, and also to determine
whether a significant improvement in prediction could be ob-
tained by using the student's high school grade point average
in conjunction with the ability test. Their obtained inter-
correlations indicate that the SCAT is definitely a valid
predictor of early college achievement, as the r's range from
.51 to .67 for different years and sexes. Combining the high

school grade point average with the SCAT score, however,

40
strengthened the correlation even more. The multiple regres-
sion figures using both the SCAT and high school performance
ranged from .66 to .78, in each case being significantly
higher than the correlation Obtained by using only the ability

test.

Attempting essentially the same thing with a different
predictor variable, Hansmeier (31) examined the relationship
between the Iowa Tests of Educational Development and college
first year grade point averages. Using as the sample 401
students (102 men and 299 women) who enrolled at a midwestern
teacher's college in 1954, zero-order correlations were
computed between the ITED composite score and freshman grade
point averages. This relationship was .71. The nine individual
tests of the ITED were also correlated with freshman achieve-
ment, as the regression figures for these ranged from .49 to

.63.

Of all the multiple aptitude batteries, probably the
most research has been performed with the Differential
Aptitude Tests (DAT). There is an abundance of validity data
available for the DAT, with the general reports indicating
high validity coefficients when using college achievement as

the criterion variable (2).

41

Much more published research is available which points
to the relationship between various aptitude or ability tests
and achievement in college. But review of more relevant
research studies at this point would not add to the informa—
tion already gathered. The bulk of the research in the area
of aptitude or ability tests and academic performance in
college indicates that: (a) there is a definite positive
relationship between most of the currently popular tests of
ability or aptitude and college achievement; (b) the relation-
ship decreases steadily from the early stages of college
attainment through the senior year; (c) females are more
predictable than males; and (d) multiple correlation proce-
dures, especially when adding performance in high school,
add to the magnitude Of the correlation between ability or

aptitude and college achievement.

The Behavioral Variables and Academic Performance

 

Only recently have there been indications of an awareness
of the relationship between academic performance and such
behavioral variables as employment, marital status, member—
ship or non-membership in a fraternity or sorority, and the
like. The literature seems to be gradually accounting for

the influence of the behavioral variables and, hopefully,

42

this trend will prove fruitful.

In an attempt to explore the relationship between outside

.p

employment and achievement in college, Trueblood (7;; utilized
a sample of students in the school of business at a large
midwestern university. Classifying students according to the
amount of outside work and class credits being carried, True-
blood calculated the relationship between amount of work and
the criterion variable of grade point average. His findings

suggest very little relationship, with obtained coefficients

being either slightly negative or positive.

Criticizing Trueblood's study for failure to account f
w; «'7-

for individual student potential, Dickinson and Newbegin (135
selected for analysis a sample of all entering freshmen
students at a large western university in 1957. Here again,
students were categorized according to amount of work and
class credits. The criterion variable in this case, however,
was expected achievement, an index arrived at by means of a
multiple regression procedure utilizing scores on entrance
examinations and other data. The students in this sample
were also classified according to age and marital status.
Interestingly, the investigators report an increase in rela—

tive achievement with an increase in amount of work hours per

week. For example, the mean achievement index for students

43
indicating a 0-4 work bracket was 0.19, while the mean achieve-
ment index for those students working 25 hours or more per
week (with the same number of credit hours!) was 0.32. In
spite of these rather interesting aspects of their study,
however, a significant relationship (correlation) between

amount of work and achieved grades was not found.

Although all other studies turned up during the review
of literature were quite old, their findings lend support to?1
the conclusions of the more recent investigations. Harris (32)
and Wrennazgg) for example, both demonstrate no clear relation-

ships between amount of employment during school and academic

performance.

Turning now to the area of fraternity membership and
its effect on college academic performance, we find a little
more research evidence available. Iffert (42) reports that
men who graduated from college and were members of social
fraternities, received slightly lower grades than men who
were not members of fraternities. These differences were not
significant, however, and it should be pointed out that Iffert
admittedly made no attempt to equate the two groups. His

was a simple survey report.

44

In an investigation made at the University of Utah,
Crookston (12) found that an opposite pattern was taking place,
that is, the fraternity members were achieving slightly better
than non-fraternity members. Once again, the difference was
not statistically significant. Crookston used a sample of
112 fraternity pledges and 112 non-fraternity pledges. Data
were collected on the two groups regarding their obtained
grade point average, their predicted grade point average, and
number of credit hours carried. The mean grade point average
and mean predicted grade point average for the fraternity
pledges were 2.23 and 2.46 respectively, as compared with
2.49 and 2.17 for the non-fraternity pledges. Crookston's
findings suggest that, at the University of Utah at least,
fraternity members perform pretty much the way they were

predicted to perform before they joined a fraternity.

In an investigation designed to determine if fraternity
and residence hall living environments differentially affect
or influence the attainment of college students, Buckner (6)
used as his sample 216 first semester freshmen pledges living
in their respective fraternity houses and 266 first semester
male freshmen non-pledges who lived in the University of
Missouri residence halls. Using scores on the Ohio State

University PsycholOgical Examination and their high school

45
achievement as predictor variables, Buckner grouped the
students according to their expected.leve1 of achievement.
The first semester mean grade point averages of the two groups
of subjects falling at each level of potential were then
independently compared. No significant differences were
found between the mean grade point averages of the two matched
groups by analysis of variance. In fact, the averages were
nearly identical for the two groups at each of the various
levels. Buckner concludes that neither residence halls nor
fraternity houses have a differential effect upon the attain-

ment of freshman students at the University of Missouri.

Willingham (83) examined the relationship between
fraternity membership and three indices of scholastic per-
formance: freshman attrition rates, freshman grade point
average, and senior grade point average. Willingham used
chi-square to analyze the differences between fraternity and
"independent" men in terms of first year drop-out rates, and
found that a significant difference exists in favor of the
fraternity members. That is, significantly more non—fraternity
men drop out of college (Georgia Tech, in this case) during

or at the end of the freshman year.

For those students completing one full year of college,

a comparison was made on grade point averages. Obtaining a

46
corrected freshman year g.p.a. of 2.12 for the fraternity
members and 2.04 for the "independents" (by means of analysis
of covariance), Willingham noted that this difference would
occur by chance only one or two times in a hundred, evidence
suggesting a positive influence of fraternity living on one's
academic performance. No mention is made by Willingham of the
place of residence of the two groups, however, and it is quite
possible in this case that the living situation of the two

groups was not actually different.

Finally, Willingham compared the senior grade point
averages of the two groups. Once again correcting the aver—
ages by means of analysis of covariance, grade point averages
of 2.28 and 2.30 were obtained for the fraternity and
independent groups respectively, suggesting a slight reversal
from the trend established during the freshman year. The
senior g.p.a.'s are not statistically significant, however,
and it must be concluded that little or no difference actually

exists.

In a study designed especially to investigate the
relationship between place of residence and academic achieve-
ment, Diener (14) reports a negative influence of living in a
fraternity. Using a sample of 138 over-and—under—achievers

at the University of Arkansas, Diener defined the over-achievers

47
as those students whose T-scores for cumulative g.p.a. were
15 or more points above the T—score for their ACE Psycho-
logical Examination, taken during their freshman year. Under-
achievers, on the other hand, were those whose T-scores for
cumulative g.p.a. were 15 or more points below their ACE
scores. For both the over-achievers as a whole and over-
achieving males, dormitory was listed first as place of
residence. With both the under-achievers and under-achieving
males, the fraternity house was ranked first as a place of
residence. Diener's sample of female students was too small

for analysis.

In searching for literature dealing with the association
of participation in extra—curricular activities with academic
perfonnance, studies of student leadership proved to be the
most fertile area to explore. Isolating student leaders by
the methods of student ratings, faculty ratings, and leader—
ship records, Hunter and Jordan (41) studied a sample of 82
"leaders." Drawing a sample of non-leaders who were stratified
by year in school so as to equal the leaders, the two groups
were then administered a number of instruments, including:

(1) the ACE Psychological Exam, (2) Strong Vocational Interest
Blank, (3) Bernreuter Personality Inventory, and (4) The Test
of Social Attitudes. College grades were obtained from the

Registrar's records for the two groups.

48
The findings reported by Hunter and Jordan bearing
relevance to the present study are: (1) 80% of the leaders
and 48%.of the non-leaders were fraternity members, and (2)
leaders made significantly higher grades on courses completed
the year immediately previous to the study than did non-
leaders, and their over-all grade point averages were con-

siderably higher also.

In another study dealing with participation in extra-
curricular activities, Stright(74) administered an ll-item
questionnaire, aimed at measuring participation in student
activities, to a group of 200 college sophomores. Using a

tetrachoric correlation technique (the items called for a

simple "yes“ or no answer regarding participation in a
specific activity), Stright attempted to measure the
relationship between student participation and academic
achievement. He found, for the males, that fraternity member-
ship had a negative correlation of -.l6, being an officer in
various campus organizations a correlation of .04, but that
participation in campus organizations in general yielded a

.33 correlation. For the females, sorority membership related
to academic achievement very positively (.63), being an officer

not quite as high (.41), and participation in extra—curricular

activities in general yielded a .44 correlation.

49

In suggesting some reasons for the rather high positive
correlations for the females and opposite pattern for the
males, Stright explains that the male population of students
was not typical at the time of the study due to the war period.
In any event, when combining males and females and comparing
extra-curricular participants with non-participants, regard-
less of sex, the participants were reported to have received

higher grades.

In another study comparing campus leaders and non-leaders,
Hodges (36) used four criteria of leadership: (1) past or
present presidency of a fraternity, (2) member of "Blue Key,"
(3) representative on the campus Inter-fraternity Council, and
(4) elected membership in the campus honorary service club.
This resulted in a sample of 100 "leaders" against whom a
randomly drawn sample of 100 male non-leaders were compared
on the basis<xftheir answers to biographical data question—
naires. A chi—square technique was used to analyze the
responses, and the results having relevance for this investi—
gation were that the leader is generally a more superior
student than the non-leader. According to a grade-point sys—
tem in which C=l, B=2, and A=3, the leaders mean g.p.a. was
1.55 compared with 1.33 for the non-leaders, a difference

which HOdges reports as being significant at the .05 level of

50
confidence. No ability-equating process was performed, and
nothing was mentioned about the expected performance of the

two groups which were studied.

Studies dealing with the comparative performance patterns
of married vs. single students are conflicting. In a study of
598 married students at the University of California, Santa
Barbara, Lantagne (52) asked the students to indicate the
effect of marriage on their academic achievement. They in—
dicated that 80% of the men and 62% of the women had positive
trends in academic performance following their marriages. The
grade point averages of the students were then analyzed upon
graduation, and Lantagne reports that married women's g.p.a.
was much higher than the single women's (2.81 vs. 2.54) and
that the married men’s achievement level was slightly higher
than the single men's (2.45 vs. 2.41). The mean grade point
average for all single students in the graduating class was
2.47 as compared to the mean grade point average of 2.56 for
the married students. Unfortunately, the results have little
value for the researcher. Asking students what direction
their performance patterns took after marriage is hardly a
technique that is very reliable, and the comparison of cumu-
lative grade point averages was done with the student body as

a whole, without any classification as to age, curriculum,

51

aptitude or ability, etc.

Another study reporting superior performance for married
students is offered by Aller (1). Using a sample of 100
married student couples at the University of Idaho, Aller
found that grade point averages of student husbands and wives
were higher than the over-all g.p.a.'s of men and women in
general. Student parents earned slightly higher averages
than did non-parents in the study, and student wives earned
higher averages than did the husbands. Once again, this is
simply reporting of a gross average, and suffers from the
same methodological limitations as the study reported by

Lantagne.

Jensen and Clark (45) report findings somewhat different
than those already mentioned. Matching a sample of 36 married
men with 36 single men according to age and scores on the ACE
Psychological Examination, they found that the single men
obtained the higher grade point averages. The obtained
difference was not statistically significant, however, and
there is no basis for concluding that married students achieve

at a higher level than single students.

A second study which utilized a matching procedure is

reported by Samenfink and Milliken (70). Equating groups

52

of married and unmarried students on the basis of ability
(ACE scores),they found that there was no difference in aca-
demic achievement over the four years of college work for the
two groups. It is interesting to note that the two studies
presented here which included no ability grouping, report
higher achievement for the married students, while the two
studies which equated the students on their ACE scores report

no difference between the single and married students.

Changes in Scholastic Performance

Evidence that academic achievement is regarded as an un-
wavering, ultimate index of performance rather than an ever-
changing, temporary level of attainment, is easily found in
the literature. Frequently used as a criterion variable of
one sort or another, the grade point average is seldom con-

sidered to be part of a pattern.

One of the few reports taking exception to this View is
pmesented by Fisher (20), who studied the trends in scholastic
performance of college students. Examining the annual grade
point averages of students over a four year period (and
thereby excluding those who withdraw from school), Fisher
wished to test the hypotheses that the average grade point

for a given class of students will steadily increase from

53
year to year and that individual variation in achievement is

of very low magnitude.

Using the entering students at Hampton Institute as
subjects, (the sample size was not mentioned) Fisher found
that mean g.p.a.'s do increase steadily with each successive
year, but that the greatest increase is made from the fresh-
man year to the sophomore year, the next greatest increase is
made from the SOphomore year to the junior year and so on.

If students dismissed from college for academic reasons are
excluded from the sample, then the year-by-year increase

drops off markedly.

Of those students who were able to continue in college
for three academic years, the changes in achievement from
one year to the next were slight, with grade point average
fluctuations of 0.50 occurring in approximately one out of

every five cases.

Unfortunately, Fisher's criterion of change is an arbi-
trary .50 grade points. The important criterion -- one's
standing in the class relative to the other students —- is
ignored. She demonstrated that the mean g.p.a. for the class
increased each year. Yet she would claim that one whose

individual g.p.a. remained the same is not a "changer,' even

54
though his or her standing relative to the rest of the class

would be taking a downward turn for each successive term.

Another report dealing with patterns of academic perform-
ance is presented by Young (87). Following the 1,924 entering
male freshmen through their first four consecutive semesters
at the University of Wisconsin, Young, like Fisher, reports
a steady increase in the grade point average of the students.
The percentage of the group receiving a grade point average
between 3.00 and 4.00 moves from 18.0 for the first semester,
to 18.5, 25.4 and 29.6 for the next three successive semesters
through the sophomore year. On the other hand, the percentage
of students receiving a grade point average between 0.00 and
1.99 moves from 37.2 for the first semester, to 32.1, 21.1
and 14.0 for the next three successive semesters through the

sophomore year.

Changes in individual performance are not reported by
Young, who calls his developed charts "scholastic progress
patterns," and encourages their use in academic counseling

situations.

55

am

The foregoing review of the literature consisted of
presentations and critical examinations of some of the research
which has been published dealing with variables thought to have
some association with academic achievement in college. Begin-
ning with the relationship between attitudes and values and
academic performance, and moving through tests of aptitude and
the behavioral variables, the attempt has been made to demon-
strate the information presently available regarding some of
the instruments and variables used in the present study, and
how they are associated with scholastic performance in college.
From this review of the literature, the following conclusions
are suggested:

1. Long over-looked as factors associated with academic
performance in college, non-cognitive-personality
variables such as attitudes and values have recently
been used in a growing number of studies exploring
this area. The findings at this point are typified
by inconsistency, but enough evidence has been
gathered to justify further consideration of atti—
tudes, values and beliefs as correlates of academic

performance.

56
Tests of aptitude and ability have repeatedly demon-
strated a relationship to scholastic performance in
college. Multiple aptitude batteries seem to be
most applicable, and the instrument used in this
study -— The College Qualification Test -- is one
of those indicating consistent relationship to the

criterion of college grades.

Behavior variables are also being examined in the
research with greater frequency. Unfortunately,

many of the studies in this area have lacked adequate
methodology, and many of the findings are contra-

dictory.

Research utilizing changes in academic performance
patterns as a variable is literally non-existent.
Cumulative indices of achievement -- not achievement

patterns -- have been generally used.

No definite direction has been forthcoming as a

result of the review of the literature. As shown,

the variables used in this study have been the

objects of considerable research effort. Unfortunately,
however, their previous appearance has been in designs

quite different from that employed in this investi-

gation.

CHAPTER III

THE METHOD OF INVESTIGATION

Definition of the Population

 

The population being studied consisted of all freshmen
entering college for the first time in the fall of 1958 and
who were enrolled in school during each fall, winter and
spring quarter until spring of 1962. Students who had trans-
ferred to Michigan State University from another institution
were not included in the study and students dropping out of
school for any term were likewise eliminated. Thus the final
population consisted of 1,085 students, 690 males and 395
females. From this 1,085, cognitive and non-cognitive person—
ality data were avilable for 1,041 students (96%) made up of
655 males and 386 females. Data pertaining to the behavioral
variables were not collected until the students' senior year,
at which time information was gathered from 674 students
(255 females and 419 males) or 62% of the p0pulation being

studied.

57

58

Classification of the Sample as to Change

 

Academic performance changers were designated by a
procedure outlined in detail in Chapter I of this thesis,
under the heading "definition of performance change."
Essentially, students were classified as performance changers
if their obtained grade point average deviated by specified,
arbitrary levels from the point at which they would normally

be expected to perform as determined by a regression formula.

The method used to classify the students according to
their direction of performance change, resulted in three
change groups (positive, negative, and "freeze" or no-change)
for each ability grouping for the year. These groups were

then compared on the basis of the following instruments.

Instrumentation

 

The College Qualification Test consists of three ability

 

tests which may be combined to yield a comprehensive total
score. The three areas included are verbal facility, numerical
ability, and general information, all of which were designed

to be predictive of academic achievement in college. The
verbal test consists of 75 vocabulary items, 50 which require

identification of synonyms and 25 which require identification

59
of antonyms. The time limit for the verbal test is 15 minutes.
The numerical ability test contains 50 items designed to
measure one‘s conceptual skill in the areas of arithmetic,
algebra and geometry. This part of the College Qualification
Test has a time limit of 35 minutes. The general information
test is composed of 75 items chosen from a broad range of
subject matter areas, half of the items dealing with the
physical, biological and chemical sciences, and half of the
items dealing with social studies such as history, economics,
geography and government. The time limit for this test is
30 minutes. The total score of the CQT consists of the sum
of the three sub-tests previously discussed. Although each
of the sub-tests may be used independently, available research
suggests that the predictive power of the test is as high
when the total score is used as when sub-scores are weighted
in optional ways (e.g., multiple correlation). The total

score was used for this study.

Estimates of the reliability of the College Qualification
Test are reported in the test manual (4) and include coeffi-
cients of .97 and .96 for males and females respectively.
These reliability coefficients were determined by a split-
half method whereby a comparison was made between scores on

the odd and the even items of the test. This technique

60
results in a somewhat higher coefficient of reliability than

would be obtained using a stability method (i.e., test-retest).

The validity of the College Qualification Test has been
examined frequently. Lehmann and Ikenberry (54) report
validity coefficients ranging from .34 to .66 when using
grades in required freshman couses as the criterion. Further
evidence of the predictive quality of the College Qualification
Test is provided in the studies reviewed in the second chapter
of this paper. Correlations in the neighborhood of .50 to
.70 seem to be the usual findings when relating the College
Qualification Test total score to early college performance,
but the magnitude of the coefficients drOp off with each

succeeding year.

The Test of Critical Thinking, Form G, was developed as

 

a part of the Cooperative Study of Evaluation in General
Education (11). In order to reduce critical thinking ability
to a measurable entity, emphasis was placed on problem
solving ability. Questions were designed to measure the
ability to recoqnize the existence of a problem, to define
the problem, to select information pertinent to its solution,
to reCOgnize assumptions, to make hypotheses, to draw con-
clusions, to judge the validity of the conclusions, and to

evaluate the conclusions in life situations (11).

\‘l'l‘lfl‘ulll ll. ‘11 I.

61
While it is perhaps debatable that grades serve as the
best criteria for testing the validity of a measure of critical
thinking, the validity of any test is defined in terms of the
purpose for which it is used. As an indicator of academic

ability, the Test of Critical Thinking seems to be very

 

appropriate. Correlations between this measure and grades in
Communication Skills and Natural Science courses as well as
cumulative grade point averages, range from .21 to .57 (54),
indicating a clear relationship between the instrument and

achievement in college.

The manual for the Test of Critical Thinking reports

 

estimates of reliability of the instrument ranging from .71
to .89 (11), and a reliability coefficient of .79 was found
when the Kuder-Richardson formula 20 was applied to the Test

of Critical Thinking scores of a smaller, restricted sample

 

used in a recent investigation (43).

The Michigan State University Reading Test was developed
by the Office of Evaluation Services of Michigan State Uni-
versity, and consists of 45 items designed to measure the
subject's ability to comprehend thoughts expressed in reading
passages. The items and passages were selected on the basis
of the relationship of the items to academic success and how

well they covered the textual materials found in several

academic areas.

62
The reliability of the M.S.U. Reading Test has been
estimated by the Office of Evaluation Services of Michigan
State University on various occasions and has been found to
be approximately .80. In a doctoral thesis done at Michigan
State, Ikenberry (43) found the reliability to be .79 by

means of Kuder-Richardson formula 20.

The validity of the Reading Test has likewise been tested
frequently. Using academic performance as the criterion,
obtained coefficients have ranged from .36 to .65 for the
males and from .34 to .60 for the females. Once again, the
magnitude of the relationship is greater when the criterion
is performance in courses taken during the first term in
college, and in the case of the Reading Test, is greater

also for courses requiring a mastery of verbal materials.

The Inventory of Beliefs was developed by the Inter-

 

College Committee on Attitudes, and Personal Adjustment of
the Cooperative Study of Evaluation in General Education.
The fundamental assumption underlying this instrument is
that "...the objectives of general education can serve as a
base from which may be inferred the model organization
characterizing the personalities of those most adaptable to
the purposes of general education" (10). The present form

consists of 120 pseudo—rational, cliche-like statements to

63

which students are asked to respond by means of a four-element
scale: strongly agree, agree, disagree, and strongly disagree.
The score is based upon the number of items with which the
subject indicates disagreement or strong disagreement, and

may range from 0-120. Subjects who accept a large number of
statements, achieving a low score on the test, are regarded

as being stereotypic in their beliefs, resistive and defensive,
and possessive of tendencies toward authoritarianism. Con-
versely, individuals with high scores rejecting the majority
of the statements are characterized as flexible, adaptive.

and non-stereotypic in their beliefs.

The reliability of the Inventory of Beliefs, as reported
in the manual (10), range from coefficients of .69 to .95
with a median coefficient of .86. Ikenberry (43) applied the
Kuder-Richardson formula 20 to test score data of a restricted

population and obtained a reliability coefficient of .84.

The validity of a measure such as the Inventory of
Beliefs, of course, is considerably more difficult to estab-
lish. Concurrent validity coefficients range from .63
(negative) for the relationship between the Inventory and the
deeach Dogmatism Scale, to .28 for the relationship between

the Inventory and the M.S.U. Reading Test for females. Face

64
validity and construct validity evidence are also presented

in the manual.

Thus, it can be seen that while the Inventory of Beliefs
does seem to be related to cognitive measures to some extent,
it is nevertheless measuring characteristics more closely
related to factors measured by the dogmatism scale, an instru—
ment supposedly measuring an attitude or personality trait

similar to that of stereotypy.

The Differential Values Inventory was developed by
Richard Prince at the University of Chicago (66), and.was
designed to measure the "traditional" and "emergent“ value
categories as previously established by Spindler (73). Some
value categories outlined by Spindler were combined by Prince
to formulate four "traditional" value categories -— puritan
morality, individualism, work—success ethic, and future-time
orientation —- and four "emergent" value categories, which
include sociability, conformity, relativistic moral attitudes,
and present—time orientation. These were the categories
originally set forth by Spindler, except that Spindler in-
cluded "achievement orientation" in the group of traditional
values and "consideration for others" in the emergent value
area, both of which.were eliminated by Prince in the develop-

ment of the Differential Values Inventory.

65
The scale consists of 64 pairs of forced-choice items.
Each pair has a traditional value statement ranked against an
emergent value statement. The subject chooses either tradi—
tionally oriented items or emergent value items in each of

the 64 pairs, with one point scored for each traditional choice.

A subject receiving a high traditional score is described
as placing high value on personal respectability, thrift, self—
denial, respect for elders: as valuing hard work as good in
itself and necessary for success; as placing his individual
desires and ideas equal to or above the desires and ideas of
the group; and as being oriented toward the future to the
extent that present needs should be sacrificed for future
satisfaction and reward. On the other hand, the person
oriented toward the emergent value system is most concerned
about getting along with people, group-determined morality
standards, consideration for the group and their feelings,

and emphasis on the present.

A Kuder-Richardson formula 20 was used to estimate
reliability and resulted in a coefficient of .75 (43), while
test-retest procedures have yielded stability coefficients of

.60 and .61 (54).

66

Once again, the validity difficulty presents itself.
Although Prince has found some theoretically sound patterns
emerging from scores on the inventory, no empirical validity
investigations have been conducted. It has been demonstrated,
for example, that both teachers and students in parochial
high schools receive more traditional scores than teachers
and students in public high schools (66); that students
aspiring toward more than four years of college education are
more traditional in values than students aspiring to four
years or less of college; and that college students coming
from rural areas tend to be more traditional in values than
students coming from urban centers (54). These findings are
in general agreement with what is known about differences in
values among various social and cultural groups, and hence,
it would appear that the Differential values Inventory is
measuring characteristics similar to those described by
Prince in the purpose of the inventory. Nevertheless, evidence
of validity is lacking, for those who might hOpe to use the
instrument for predictive purposes. Rational content is
‘present, and it is largely on this factor that the instrument

is presently being regarded as promising.

ROkeach's Dogmatism Scale, Form EL_ was developed by

Dmilton J. Rokeach at Michigan State University as a measure

67

of general authoritarianism, general intolerance, and openness
of belief systems. Form E of the Dogmatism scale consists of
40 dogmatic statements with which the subject is asked to
indicate agreement or disagreement on a six point scale

ranging from "agree very much" to "disagree very much."

Rokeach defines dogmatism as representing a total
ideological defense against threat and at the same time a
c0gnitive framework for satisfying one's need to know and
comprehend the world one lives in. He argues that dogmatism
allows one to ward off threatening aspects of reality and at
the same time receive the satisfaction of feeling that one

understands it (68).

Plant, Minium, and Myers reported odd-even reliability
coefficients for the Dogmatism Scale of .84 for the males
and .85 for the females (65), while Ikenberry (48) obtained
a reliability coefficient of .76 by using an internal

consistency method.

As typical of measures of its kind, the Dogmatism Scale
has failed to demonstrate consistent evidence of validity.

It does, however, bear a negative .63 relationship1 to the

 

1The negative correlation between The Inventory of Beliefs
and Rokeach's Dogmatism Scale are resultant from the scoring
procedure. The concepts of stereotypy and dogmatism are
positively related.

68
Inventory of Beliefs, and since these two scales were con-
structed independently to measure theoretically related
phenomena and are correlated to such a degree, the coefficient

tends to support claims for validity for both measures (43).

The Experience Inventory was used to ascertain various
aspects of the students' "behavioral" characteristics. Only
the last section of the inventory, consisting of 43 items,

was used in this study.

Developed by the investigator, the items can be classi-
fied into four areas, those dealing with (1) amount of weekly
employment during the school year, (2) degree of involvement

in extra-curricular activities, (3) place of residence, and

(4) various personal factors.

The employment section simply asked the students to
indicate, by checking in provided columns, the number of
‘hours they were employed during each week of the school year.
Thus, frequencies were obtained for the number of students
falling into hourssworked ranges of (1) over 20, (2) ll-20,

(3) l-lO, and (4) didn't work.

The place of residence and personal factor items were
listed and the students were asked to indicate, by using

codes'representing the academic terms, which items applied

69
to the students during which term or terms. For example, one
item reads "became engaged" and the students were to indicate
the term or terms this particular item applied to them, if at
all. In this way, frequencies were obtained for the year-to—
year place of residence and applicability of various personal

factors.

The index of participation in extra-curricular activities
was developed in a somewhat different fashion. All the possible
student activities were listed, with the assistance of a
representative of the Dean of Student's Office. The students
were then asked to indicate if they participated in this
activity, and, if so, what specific position they held or
function they performed and what term or terms they did each.
Next, particular values were assigned the positions and
activity according to the estimated amount of time required
by each activity. These values were "experts‘ ratings" made,
once again, with the assistance of the Director of the Men‘s
Division of Student Affairs. The result was a "scorecard"
such as that presented in illustration 2 on page 70. The
value was then multiplied by the number of terms the student
indicated it applied during a given year, to get an annual

weighted score for activities participation.

70

 

 

 

 

 

 

 

.- - ,- 1 - . . i <
m a i t
J x m .1 ‘5‘?me .. .
m\\V\«\-w§\ Tau-55$? w??- Q i a M l u 2”
,1-.. . . . M . .. u x . .
. . . . \x a i
.4 {‘«kv‘B “ V-\% \\\\\x WSXTV: \\v\\.. \\ V s \\ «1‘ \Wfl. ~\1fl\s\ ._. «N
.‘\\\‘ J .‘k‘x‘AJ _ s :\.\«.\ 3% w .\ §\\m\‘\v . , \\ _ \v‘\o\\\\\ rug . .\\ ~QQU \0\ \‘Vb ‘
.33.“.th \\§\ 6% i x .. -.A.wx11§1b-.1W.1V\..-\ .Nm‘fi. 1 . a». ., . -. . . . 1 a
11112:- -, 1 - . _ _ Cots-a \xQ _
. a .kai $1.39 .M.
QR.§\\ . S {Ah-.94 Q Vuxi) “Squi- Mi xx?“ «SQ NEMMQ W
.. . . wt. . . v. ._ . . :6: - w
Eu 1% h 3.5% . .. a. SSS 536. 3. is E325 .. «Q.» xx,
YHerQ WM‘XQ :3 Q Thvfi- . . . .. 1x- _
.. . 1 I. 11. 141 - . _... .t u i
w m m% \-v\ Oasis-west"
3:xe if m h.» i .1 1..-3.. 3... Est... itww .
, . . . - . at
\.\.\t5\.\\h . VHRCQR . VN‘x-Q “\WV‘\ hm%\m VWQXQ .\\\‘U V‘sQQN . -1 vW- -.--
«kw-xx .nNtV . in...» H 1111..., 1- .- 1a-.
111- -w .. . 1 4-1...-1 . . u . . . 3.2
. _ 3i. l :2 ai. b:
m \N‘RQVKQ\QV\ U .~Q s‘<b .k9‘0\ . Q . s -
. . . s . .. “.2ch it: u .
. . xix i \ . my .-
. h \ . kaxo. ”
. _ Rust-\Q \WN. -. . .w . . .21...- ..:...~
1111111 3.11 1. .11- . .- a 1. 1:- . ital-.4 1.1 H .. AM.WMQ¥R\.\.~»U
. . w . («\wwtxfi. i. w»\u....\t \u.£o.%
\ ‘2 . _ . . \\ \. h t . . .. . «\v.
xi... . M $- _ \. V... \x x- 114 Rh...
r,.\\.i..\-s3 c. i. «xmxéwok new. - . .. . .. . - . .11-11.. .1
1BR-.- i- 1. H é . ......\.-\..-..,
. i . 3. y S...» n... u 336 . x.
i w .Jwfi £3». - . _ «9&3 Vt
“ i n t . .
W . \G.“ a... «13K. \lw\t.\ nhtV W. .w 3.. m V...%qm\.nJ-..-< X
a ._ . axx. 23-h . _ -. 1..... 1..
- .. 1 1 111,-... ”EEK-2.1.9;
_ . .. ko 2m C-NY IN \nwsfik
. m . film-sé v- x \c \\ .b m
. . _ .
_ W . \ .v _
. . . R Q V fix \ N _ r. w .19; (Dalia-I’ll)... . I - 1 \13
l . . .1.11L.-1111111-. {.11-1.1.2.1111 so _ . . .
, r --_.- .hV.-:..s-w- as S 331%...” .3. a: Mrs: H 3% tar-Mos 3- s q
kmvvfow Hanan ,1 tht -m.,\_a~\\\..\\\. . M VVQMM§ H S o\\ M Rixuwuust-an \w w.
M . N a a... \ \hv‘fil WWQVQN Q ._ G‘NVXO WMQXQ
1.3.x \N-uss uqr\ .N. \- \. i Wm-«anf . W -. 111. - 1.. -. - . I 1 . L-
. a _ 1w . w .. . {1|
r 111-! 11.. -. :1..- . MUflxrflHUM “Mpg. a...” NGOfiUUCUM HO mGOflHflma m
vanaummw 0cm mwflufl>fluom ucmpsum unflufimflOB How pwlfi GHHO N 0H5 Hm

71
As an example, assume student X belonged to several
activities during his junior year. According to his experience
inventory, he was a member of the junior class council for the
entire year, serving as a committee chairman for one term,
and he was a reporter for the student newspaper for two terms.

His activities index score would be calculated as follows:

Member of Class Council = 2 points x 2 terms = 4 points
Chmn. of Class Committee = 3 points x 1 term = 3 points
Reporter for newspaper = 2 points x 2 terms é_4 points
Total Annual Activities Index . . . . . . . =11 points

It was found that the distribution of activities index
scores obtained by this weighting procedure had a highly
skewed distribution, with a large number of students receiving
0, l, or 2 points for an academic year, and gradually dropping
off to a small number of students with high activities index
scores. Consequently, no assumptions of normality or linear-
ity were made with the scores, and they were analyzed with a
non-parametric technique, as were the other non—cognitive-be-

havioral variables.

One of the obvious limitations of the Activities Index
was its necessary reliance on memory. Students were asked to
recall not only what activities they participated in and what
functions they performed, but also what terms they did each.

Nevertheless, the procedure was probably better than going to

72
the individual student records, since information of this kind

is not complete and often inaccurate.

Procedure for the Collection of the Data

 

During freshman orientation week in the fall term, 1958,
a battery of instruments was administered to 2,973 entering
freshmen at Michigan State University. The test battery

included The College Qualification Test, The Michigan State

 

University_Reading_Test, A Test of Critical Thinking, the

 

 

Inventory of Beliefs, The Differential Values Inventory, and

 

Rokeach's Dogmatism Scale. Complete and usable test data were

 

gathered for each of the 1,085 students who maintained

continuous attendance for the next four years.

In May, 1959, or one academic year later, a similar test-

ing program was completed. At this time, The Inventory of

 

Beliefs and The Differential Values Inventorijere administered,

 

and usable data were obtained from 1,041 students who were to

maintain continuous attendance.

Finally, a third testing program was conducted at the end
of the students' senior year in college in May, 1962. At this

time, the students were administered The Differential Values

 

 

Inventory, The Inventory of Beliefs, Rokeach's Dogmatism

73

Scale, and The Experience Inventory. Usable data were gathered

from 674 students at this time.

The annual grade point averages for the students were
obtained from the cumulative summary records of the Record‘s
Office of Michigan State University. These records of
student attainment are recorded on a term—to-term basis with
the cumulative information also indicated. Calculating the
students' annual grade point average was simply a matter of
adding the term totals for a given academic year, and dividing
annual points earned by the annual credits carried. This
calculation was done on the IBM 604, Operated by the Michigan

State University Data Processing Center.

Procedure for Analysis of the Data

The students‘ scores on the various instruments of the
test battery were coded and key-punched into IBM cards. The
achievement information obtained from the Record‘s Office was
already on IBM cards, and consequently all the data relevant
to the study was processed by machine, with each card con-
taining the student number, sex, and curriculum as the

identification characteristics.

74

Because of the nature of the data gathered, it was
impossible to use a single statistical technique for the
analysis. Some of the data, for example, is discrete, and
hence could not be analyzed by techniques which assume
normality of the population distributions, while other vari-
ables met the necessary assumptions for parametric analysis.
It was therefore decided to use analysis of variance in
examining the data relating to hypotheses one and two (i.e.,
the non-COgnitive-personality and the COgnitive variables),
and chi-square for the data relating to hypothesis three

(or the behavioral variables).

The calculations necessary for the analysis of variance
were performed on the IBM 604, while the facilities of the
Michigan State University Computer Center were used for both
determination of the scholastic change groups and the chi-

square analysis for the behavioral variables (MISTIC).

CHAPTER IV
ANALYSIS OF THE DATA

The data gathered by the procedure discussed in the
preceding chapter was analyzed in various ways, depending on
the nature of the data and variables in question. It will be
the purpose of this chapter to present in detail the analysis
of the data, as well as some of the inferences such analyses

make possible.

The analysis of the data will be presented in four parts:
(1) a description of the sample of students involved in terms
of the variables being studied, (2) Hypothesis I, (3) Hypoth-

esis II,and (4) Hypothesis III.

A Description of the Sample

In Chapter I it was mentioned that students were classified
into achievement groupings on the basis of their obtained
grade point averages, by simply dividing them into upper,
middle, and lower thirds of equal size. The resulting grade

point average ranges are presented in Table l.

75

76
Thus, it can be seen that if a male student's first year
grade point average was above 2.60, he would have been classi—
fied as a high achiever, or if 2.16 or below, he would have

been classified as a low achiever.

 

 

 

TABLE 1: Grade point average ranges for
achievement groupings.
Classification Points
Achievement Freshman Sophomore Junior

Groups Male Female Male Female Male Female
High 2.60- 2.70- 2.59- 2.71- 2.68- 2.82-
above above above above above above
Medium 2.17- 2.28- 2.19- 2.27- 2.26- 2.36-

2.59 2.69 2.58 2.70 2.67 2.81
Low 2.16- 2.27- 2.18- 2.26- 2.25- 2.35-
below below below below below below

 

 

 

 

 

 

 

 

The ranges differ for males and females (with the latter typi-

fied by higher grade point averages) and from year-to—year.

For example, a grade point average of 2.63 would result in a

high achievement classification for the males during the fresh-

man year, but only a medium achievement grouping for females.

Moreover, the same grade point average deserved a medium

achievement placement for both males and females during the

junior year.

 

77
A better understanding of this situation might be pro-
vided by Figure 3,which indicates the annual increase in grade
point averages from year to year for both males and females,
but with female averages always being somewhat higher.

Figure 3 is located on page 78.

A final description concerning the grade point averages
is presented in Table 2, where it is shown that the relation-
ship between annual grade point averages decreases as the
time lapse increases. For example, for the females (below
the diagonal) there is a .79 correlation between freshman
g.p.a.'s and sophomore g.p.a.'s. However, the correlation

between freshman averages and the senior averages drops to .56.

It has been previously indicated that a standard error
of estimate of .50 was used in selecting the performance
changers. The rationale for the .50 standard error was
presented on pagell of the first chapter. The actual grade
point average discrepancies which resulted from this figure

are presented in Table 3.

Figure 3:

78

Progressive mean grade point averages,
male and female, for group maintaining
continuous enrollment, fall,1958, to

     

  

  

 

 

 

 

spring, 1962.
2
2.70 ( '7E9’
2.65 (2.63)”/”
I”
2 .60 l/
I
I”
2.55 I;
l” 2
2.50 __________________ El '54)
(2.49) (2.49) (2.50) .46)
2.45
2.40 Males
.40) (2.39) Females—---
2.35
2.30 (2.32)_
Fall '58 Fresh. Soph. Junior Senior
TABLE 2: Intercorrelations of annual grade point
averages, male and female, for group
maintaining continuous enrollment,
fall, 1958, to spring, 1962.
Fresh. Soph. Junior Senior
Freshman .73 .55 .54
Sophomore .79 .67 .64
Junior .64 .71 .70
Senior .56 .61 .68

(Male correlations above the

diagonal,

females below.)

 

79

TABLE 3: Grade point average discrepancy equivalents
of the .50 standard error of estimate reported
for sophomore, junior, and senior change points.

 

 

 

Male Female
Sophomore .1978 .1715
Junior .2064 .1840
Senior - .2001 .2010

 

Consequently, if a student's sophomore grade point aver—
age was .1978 points above or below his expected sophomore
level of achievement (or .1715 if the student were female)
he or she would be classified as a performance changer. As
the table reveals, a discrepancy of greater magnitude would
be necessary for change classification during the junior and

senior years.

The size of the group being studied is varied. For
analysis of the variables pertaining to the first two hypoth—
eses, data were available for 655 males and 386 females. For
analysis of the behavioral variables, data were available for
419 males and 255 females. These groups were further divided
into achievement groups and change groups, so that in some

cases, especially for the behavioral variables, the specific

80
number of students in a particular grouping became quite
small. The specific group sizes are included in the tables
presented in the analysis of the data pertaining to the

hypotheses.

Descriptive data pertaining to the variables used in the
investigation are presented in Tables 4 and 5 which appear on
pages 81 and 82 respectively. In later tables dealing with
analysis of data pertaining to the hypotheses, mean scores
revealing characteristics of the entire group are omitted.
This information is presented here so that the reader has
some awareness of the entire group's status on these variables
before being divided by level of achievement and direction

of performance change.

Table 4 contains the means and standard deviations for
the entire group‘s scores on the measures used for the cogni-
tive and non-cognitive—personality variables, as well as the
means and standard deviations found for the entire entering
class in 1958 on these same measures. The sample being
studied here is a part of the larger group, and therefore
data for the larger group is presented for purposes of

comparison.

81

TABLE 4: Means and standard deviations on instruments
used for ability-aptitude and attitudes-values
variables, with comparative data for entire
sample of students entering Michigan State

University in fall,

1958.

 

 

 

 

 

 

 

MALES FEMALES
7 Present Entire Class Present Entire Class
Sample Entering 1958 Sample Entering 1958
Mean S.D. Mean S.D. Mean S.D. Mean S.D.
Ability and Aptitudes:
C.Q.T. 132.35 24.33 126.46 26.28 122.63 23.49 116.47 24.65
Reading 28.35 6.02 27.04 6.40 29.39 5.36 28.14 6.37
Crit.
Think. 32.84 6.98 31.51 7.18 32.54 6.51 31.29 7.18
Attitudes and values:
I.B. 64.06 14.16 62.86 14.08 65.55 12.50 64.77 12.84
D.V.I. 35.01 6.84 34.67 6.93 33.92 6.99 33.59 6.99
Dogmatism
166.72 25.22 168.19 25.36 162.89 24.38 163.56 25.47

 

 

82

TABLE 5: Descriptive break-down of students by
behavioral variables during sophomore,
junior and senior years.

 

 

 

 

 

 

(N=674) (N=664) (N—662)
Sophomore Junior Senior
ACTIVITIES : N % N % N %
Much 210 31 244 37 181 27
Some 237 35 210 32 225 34
Little or None 225 34 210 32 256 39
100 101 100
Place of Residence
Residence Halls 262 39 211 32 153 23
Fraternity or Soror. 98 15 111 17 109 16
Co-ops 87 13 94 14 86 13
Married Housing 83 13 83' 13 100 15
Off-Campus 134 20 165 25 214 33
100 101 100
Employment
Over 20 hrs. per week 79 12 80 12 103 16
Between 10 and 20 148 22 144 22 120 18
Between 1 and 10 102 15 105 16 175 26
None 345 51 335 50 264 40
100 100 100
Personal
Made 1st choice of
major 89 13 34 5 7 1
Changed major field 161 24 84 13 29 4
Pledged fraternity
or soror. 91 14 22 3 9 1
Became "active" in
F. or S. 93 14 55 8 20 3
Became engaged 39 6 65 10 108 16
Got married 12 2 31 5 58 9
Became "pinned" 57 8 95 14 80 12
Began "going steady“ 85 13 68 10 39 6

83
Considering for the moment, only those students who are
part of the present study, it can be seen from Table 5 that
the males and females are quite similar with respect to the
variables under consideration. Males appear to have higher
College Qualification Test scores and higher Dogmatism scores,

but other differences are minimal.

When compared to the original entire entering class, it
is evident that the present sample has higher mean scores on
the measures used as indices of ability and/or aptitude. This
is to be expected, since drop-outs and withdrawals are not

included in the present sample of students.

Table 5 presents the number and percentage of students
for whom the various behavioral variables were applicable.
Table 5 is interesting in several respects. The most notice-
able characteristics of the entire group are:

1) Participation in extra-curricular activities drops

off markedly during the senior year,

2) The number of students living in residence halls
drops off sharply from the sophomore through the
senior years, while the number of students living
in off—campus housing increases,

3) The amount of weekly employment increases during the

senior year, and

84
4) Engagements and marriages show steady increases

from the sophomore through the senior year.

It should be pointed out here that while the percentages
reported under the activities, place of residence, and employ-
ment sections always total 100%” such a total is not to be
expected from the items under the personal category, since

these items are independent of one another.

Hypothesis I
It is possible to differentiate among groups
of students classified by direction of academic
performance change, on the basis of their
scores on ability and aptitude measures.

After classifying the students according to freshman
achievement level and sophomore direction of performance
change (as outlined in Chapter I), the analysis of variance
technique was employed to ascertain whether the differences
between the change groups and their scores on the cognitive
variables were large enough to be more than simply chance
occurrences. The .05 level of confidence was chosen as the
confidence level for the hypothesis, and was the critical
level used throughout the entire dissertation. Data for the
males and females were analyzed separately on the cognitive

data because of the initial sex differences found to be associ—

ated with their performance on these particular tasks.

85

Male Students

 

Table 6 shows the results of the comparison of the change
groups on the ability and aptitude variables for the males
during the sophomore year. The cognitive variables were
examined at the end of the students' second year (or their
first change point), while the measures used to obtain the
data were administered during the fall term of the students'
first year in college. In this sense, then, the analysis is
post facto. The low, middle, and high categories, indicates
the student's standing relative to the rest of his group in
terms of his academic achievement during his freshman year.
This separation was particularly relevant when analyzing the
ability and aptitude data, since academic performance and the
instruments used as measures of a cognitive trait have been

demonstrated to be related (47).

Table 6 reveals the expected pattern of scores for the
groups who are classified according to their first year
scholastic performance. The highest scores on the ability
and aptitude measures were obtained by those falling in the
top third of the grade point distribution, followed in order
by the middle third and the low third. This difference
between groups was expected, and, in fact, was the major
rationale for the division of the sample into this three-

dimension type of arrangement.

.0>onw @ manna CH pmumfia
modam>lm mnu mo mdem> soaumufluo mgu paw .mmHMSUm same .Eoummum mo mmmummp .mmumsvw mo
mEdm .COflDMHHm> mo mousom msfipummmu sump How Napswmmm on» Ca m.o paw .N.@ .H.@ moanma 0mm

11 1 1

.msoum mmcmfio

ammo How A=Z=v wNHm mHmEMm mfiu op Homo“ msfidaoo mflu m>onw mommnusmumm msu CH nudges:

pcm .mmcamdonm uswEm>mH£om any ou Momma .mmusoo mo :.£mH£= paw =.macpfiE: =.3OH= mpuo3 mgB
.mummcmno m>fluamom u =+= swam mfiu paw .mmcmfiu on u :0: amen m£u .mummcmgo m>flummmn n :1: swam
mflu .umma msflpmmm huflmum>flas opmum cmmflnoflzn.m.n.m.z .umma coaumofiwflam5o mmmaaoou.a.0.u
.mcflxcflnB HMUfiUHuUn.B.U “poms on Hafl3 maonemm paw msoaumfl>munnm mafizoaaom mfiu .3OHHom swap
moans“ mgu CH can .muom .mmanmu ucmsvmmnsm may mo mama How Hmpoe m mm m>umm HHHB manmu mwnB

.mocwpflmsoo mo Hm>ma mo. on» pcommn unmoflmflsmflm mum mdem>1m w>onm mg» no 0:02

86

gv.a hm.mm OH.Nm vm.am mo.m om.mm m¢.hN mo.hN om.N mm.mm mm.m~ m®.¢m .m.D.m.2

mm.H mv.HmH 00.0va mm.hva 00.0 ma.mNH mo.mNH Hm.mNH fib.~ mo.HNH fim.¢HH mN.NNH .B.O.U

 

 

 

 

 

 

 

mm.o m¢.6m mm.mm om.sm mm.o mm.~m sv.am mm.Hm m~.o ma.m~ m¢.m~ mm.mm .a.o
m nose home Ammo m Ammo Aomv “may m ance lame name 2
+ o I: + 0 II + o I.
mem mannHz son

 

.ummh muoeonm0m ozu mcflusp
mmsmno madeHomumm mo soauowuflp paw mmmum>m assom mpmum cmEfimem
>3 poemsmmmHo mucmpdum mama mo mmuoom same mpsuflumm paw muflaflnm no mamma

87
Within the groups, however, there are still no differences.

That is, changes in performance pattern do n2£_seem to be
related to any of the ability and aptitude measures employed
in this study. Although there does seem to be a rather con-
sistently higher mean score on all three variables for those
whose performance has changed positively, this consistency is
not borne out to be more than a chance occurrence at the .05

level of confidence.

Female Students

 

Table 7 shows the results of the comparison of the change
groups on the ability and aptitude variables for the females
during the sophomore year. As was the case with the male
data, those in the top or high third of the grade point dis—
tribution during the freshman year tend to obtain higher
scores on the cognitive variables. In fact, in every case
where a mean score is presented, the higher the achievement
grouping, the higher the females' score on the cognitive

variable in question.

Once again, however, the differences occurring within the
three achievement groupings are small. Though relative con-
sistency seems to prevail, with the positive changers tending
to score somewhat higher than the negative or downward

changers, only one F-value exceeds the established critical

I:

.m>onm h mHQuB CH UmpmHH mmCHm>1m

msu mo moCHC> CoHHwUHHo may pCm .mmuwsvm Came .Eopmmnm mo mmmummp

.mmHMCUm mo mECm

.CoHuMHHm> mo wUHCOm mCHUHmmmH sump How xHUCmmmm wan CH m.h pCm .N.h .H.h mmHQMB mom

 

.moCmpHmCoo mo Hm>mH mo. wfiu pCome quUHwHCmHma

88

oo.N No.vm om.mm mv.mm w¢.N mm.mm mm.mm No.mm bo.H hm.mm mo.®N

mo.mm .m.D.m.2

00.0 mm.mma oo.©mH Nm.H¢H No.0 hm.NNH mm.HNH mm.HNH mm.H m¢.HHH mm.moa Cm.COH .B.0.U

 

 

 

 

 

 

 

 

 

so¢.¢ OH.mm mo.¢m gm.om hm.o HH.mm om.Hm on.mm No.0 mo.mm vm.m~ mm.hm .9.0
m Aasv Ammo AHmv m Ammo “was have m Ammo Ammo “sea 2
+ o 1. + o .I + o I:
mem WAQQHS 30A
.HmmS mHoEOCQOm may mCHHCp mmCmno
moCmEComnmm mo CoHuooqu pCm wmmnm>m uCHom mpmum Cmagmmum mg
pmHmHmmmao mUCmpsum mHmem mo monoum Came mpsuHumm pCm mpHHHQC an mamas

89
level. In this particular case, the females in the high
achievement group whose performance pattern for the sophomore
year changed upward, have a significantly higher score on the
test of critical thinking than those whose performance pattern
changed negatively. Close examination of the difference in
question, however, suggests that the significant F-value is
actually a result of the variation between the students
witnessing an upward change in grades versus those whose
pattern "froze" or did not change during the sophomore year,
and caution must be used in drawing conclusions from the

significant difference.

Conclusions Regarding Hypothesis I

The measures used to gauge the students' ability to
perform tasks of an academic aptitude or ability nature (i.e.,
a cognitive variable) were administered during the fall term
of 1958 only. When analyzed regarding their relationship to
changes in performance patterns occurring during the sophomore
year therefore, the criterion is nearly a two-year—old index.
Such characteristics have been demonstrated to possess a good
degree of stability, however, and it seems safe to conclude,
on the basis of the data examined, that the cognitive variables
employed bear little association with changes in one's scholas—

tic performance pattern. Nine male change groups and nine

  

. ..~fl‘.|fi....l~r..Y-‘\i}n.c. ... Juli-41‘

M». II

90
female change groups were analyzed on three different measures,
and although some consistency is apparent, statistical signifi-

cance appears only once.

Hypothesis II

 

It is possible to differentiate among groups

of students classified by direction of academic
performance change, on the basis of their
scores on attitude and value measures.

The analysis of variance technique was again employed to
analyze the data, and the procedure for grouping students the
same as that discussed under Hypothesis I. Males and females
'were again considered separately, owing to previous research
on college student populations using these same instruments

(54). The instruments used to test Hypothesis II -— The

Differential Values Inventory, The Inventory of Beliefs, and

 

Rokeach's Dggmatism Scale -- were administered several times

 

during the students' course of study. Consequently, analysis
was possible more than one time. In the case of both the
male and female groups, the performance changes taking place
during the sophomore year were compared with non—cognitive-
personality test scores recorded at two different points:

(1) during fall term, 1958, or the beginning of the students'
freshman year, and (2) during spring term, 1959, or at the

end of the students freshman year at Michigan State University.

91

Male Students

 

Table 8 shows how the sophomore year male change groups
compare on the basis of their scores on the attitude and
value measures taken during their first term in college. As
was the case with the ability and aptitude data, differences
between achievement groups are more apparent than differences
occurring within achievement groups for the performance
changers. For example, those in the high achievement group
tend to be more flexible and/or adaptive than their counter-
parts in the low achievement group (as reflected in higher

mean scores on the Inventory of Beliefs), have a slightly

 

higher traditional value orientation (suggested by their
somewhat higher scores on the Differential Values Inventory),
and be somewhat less dogmatic. Ferreting out differences for
the performance change groups within the achievement groupings

proves to be a more difficult task, however.

None of the male F-values for the sophomore changers
achieve a level of statistical significance. Not only is
significant variation absent, but consistency is also lacking.
Clearly, the measures of a non-cognitive-personality type are
not related to direction of change in performance patterns
for males, at least when comparing on the basis of freshman

test scores, and sophomore performance change.

92

meHm>Im mfiu mo

men-um

.Co UMHMC> mo

mmCHm> COHkuHHU mfiu pCm

.mmumCUm CmmE .EOpmme mo momummp
moudom mCHpummmH sump How prCmmm< may CH mwm pCm .N.m .H.m mmHQma 0mm

.m>onm m magma 2H umumHH

.mwumsqm mo

.mHmum EmHumEmOQ w.£Ummxomu.m.Q.m UCm .huoqu>CH mmflHm> HCHHCOHMHMHQN.H.>.Q .mmeme mo

hHOHCm>CHn.m.H

.moCmpHMCoo mo Hm>0H

"pom: ma HHH3 mCOHumH>mHQQm mums»

.3oHHom pas» manna“ may 2H can

.mumm

mo. m3“ UCOMCQ “CMUHMHCmHm mum mmsHm>1m m>onm mnu mo 0Coz

 

 

 

 

 

 

 

 

 

hm.H m¢.mmH Hm.moH mm.omH om.o mm.moH mo.mmH wo.va 00.0 mm.mmH mH.N>H mm.NhH .m.n.m
mm.H m~.hm om.mm NH.mm mm.o mm.¢m mm.¢m NH.mm mN.H Co.vm MH.mm H¢.mm .H.>.Q
mm.H mm.mo mm.oo vm.wm Hm.o v>.mm on.mo mm.¢o mm.o Co.mm mm.mm mm.mm .m.H
m Amhv name “awe m Ammv homo Amen m “see news Avmv z
+ o I. + o I. + o 11
mem HAQQHZ 304
.mmm» mHoEosmom msu mCHHCp mmCmfio mUCmEHomumm mo
CoHuumHHp pCm mmmum>m DCHom opmum Cmenmmnm ha pmHMHmmmHo mqupCum
mHmE umCHumou .mmmH .HHmm How mmuoom CmmE mCHm> pCm mpCqupm .m mamCB

. 93
Table 9 shows how the sophomore year male change groups
compare on the basis of their scores on the attitude and
value measures taken at the SEQ-0f their freshman year in
college. Their scores in this instance, have been altered
with exposure to the college environment for one year, where-
as the scores used in Table 8 are indicative of the variable

present when they came to the university. The make-up of the

performance change groups is the same, of course.

Here we find one difference large enough to be more than
a chance difference. In the middle achievement group, the
positive performance changers are significantly more flexible
and/Or adaptive than the "freeze" group or negative changers.
Although not verified by statistical significance, this
variable also seems to favor the positive performance
changers in the low and high achievement groups. That is,
there seems to be evidence in support of the theory that those
whose performance pattern takes a positive turn will tend to
be less rigid or compulsive than those who exhibit a downward

academic trend.

‘The Differential Values Inventory fails to differentiate

 

between any of the groups significantly, and really portrays

no hint of consistency or trend.

94

.m>onm m mHnma CH pmumHH
mmCHm>lm gnu mo mmCHm> COHHmuHHU 03¢ pr .mMHMCgm mefi .Eopowuw mo mmwummp .mmumsvm no
mean .COHDMHHM> mo mousom mCHpHmmmn dump How prCmmm< wnu CH N.m UCm H.m mmHQwB mom

.moCmpHmCoo mo Hm>mH mo. mfiu pCOth quoHMHCmHma

.Hmmm Cmfiflmmnm mo pr pm pwuwpmHCHEpm uoz .m.n.m

 

 

 

 

 

 

 

mm.o s¢.am mm.m~ om.mm -.~ mm.om om.Hm mm.m~ Hm.o mm.m~ so.om ¢¢.mm .H.>.n
sm.o ma.om Ha.mm m~.mo :Hm.m Ho.oo mo.mo no.8m om.~ oo.ao m6.mm mm.mm .m.H
m Amss Ammv Ammo m Ammv Acme Amps m Chev Asmv “46v 2

+ o .1 + o 11 + o I:

mem HAQQHE BOA

.umm> mHoEonmom 03» mCHHCp mmCm£U moCmEHomumm mo
CoHuowqu pCm mmmum>m uCHom mpmum Cmfinmmum >9 UCHMHmmmHo mqupsum
mHmE «mCHumou .mmmH .mCHumm How mmuoom Came mCHm> pCm opCuHuum um mqm<a

95

It should be pointed out here that the difference ob-

tained on The Inventory of Beliefs in Table 8 is consistent

 

with the direction of scores found in Table 9, the scores for
the same groups on the Inventory taken at the beginning of
the freshman year. Though none of the latter differences
achieved the level of significance, the consistency of trend,
both on the freshman and sophomore results, may be suggestive
of an association between flexibility and constancy of

academic performance.

Female Students

 

Table 10 relates the mean scores for the freshman year
administration of the attitude and value instruments to the
females. As was the case with the males, the scores on these
instruments at the beginning of the freshman year seem to
bear no relationship to changes in scholastic performance.
The magnitude of difference necessary for high confidence
interpretation isn't even approached, and trends in scores

toward some consistency are not to be found.

Table 11 considers the attitude and value traits as
measured at the end of the freshman year. While the end-of—
the—year testing indicates some possible associations for
the male group, the females' scores shed no more light on the

associative factors than did their scores recorded at the

m£u mo moCHm> COHCmuHHo on» pCm
.COHUMHHC> mo wuusom mCHpummwu sump Com prCmmm< wfiu CH m.0H pCm .N.OH .H.OH mmHQma mmm

.wUCmUHMCOU mo Hm>mH

.m>onm 0H mHQma CH pmumHH mmCHm>1m

.mmumsvm Came .Eooomum mo ammummp

.mmuwsvm mo mesa

mo. may UCOth quoHMHCmHm mum meHm>Im m>onm may mo mCoz

 

 

 

 

 

 

 

6
9 o~.o o~.ooa ov.HoH Hm.HmH mo.H em.oma «m.mmfl mm.omH om.o ov.~HH HH.H6H HH.moH .m.n.m
mo.o Hm.mm H¢.vm mm.mm m~.o m~.¢m oo.¢m mm.¢m mm.a vm.mm ms.mm m¢.Hm .H.>.n
6H.o Hm.mo m~.ms m~.mm mm.o Hm.so os.ms H¢.s6 40.0 «m.om mH.Ho H¢.Ho .m.H
m laws Ammo Ammv m Ammv Amvv Apes m Ammo name news 2
+ o 1 + o 1. + o .l
mem quQHz 30A

.Hmwm mHoEonmom any mCHHCU mmCmflo mUCmEHOMme mo
COHuomHHp pCm mmmum>w uCHom mpmum Cmfifimmum ma pmHMHmmmHo mqupCuw
mHmEmm «mCHummu .mmmH .HHmm How monoum CmmE 05Hm> pCm mpCquufl ”0H mqméa

97

.m>onm HH wHQMB CH
pmHmHH meHm>Im wsu mo waHm> CoHHmuHHU mflu pCm .mmumsvm meE .EOpmmHm mo mmmummp .mmumsvm
mo mECm .CoHumHum> mo moHCom mCHUHmmmH mnmp How prCmQQC mfiu CH N.HH pCm H.HH mmHQmB mom

.moCmprCoo mo Hm>oH mo. 0:» pCOhwfl HCMUHHHCmHm mum mmCHm>1m m>0Qm may mo oCoz

.Hmmm CmECmmum mo pCm um pmumumHCHfipm H02 .m.n.m

om.o mm.mm mN.om mo.om mm.N _oo.mN N®.mN mv.mm mn.o hm.om mm.mm 0H.hm .H.>.Q

 

 

 

 

 

 

 

00.0 00.0H 00.H0 00.HH 0H.H 00.HH H0.00 00.0s HH.H HH.H0 H0.0m 00.00 .m.H
a idea 100V Ammv m Amme A000 Assn m Ammo Ammv 100V 2
+ 0 1. + 0 I. + 0 I.
mem MAQQHS 30A

1 1 111 11 1

1 .111 11 111 1 111 1 b P 111 111i

.Hwoh mHOEOCCOm mgu mCHHCp mmCmfio CUCMEHomem mo
CoHuUmHHU pCm mommm>m uCHom mpmnm Cmfinmwum an pmHmHmmeo mqupsym
mHmEmm umCHummu .mmmH .mCHumm How mmuoom meE mCHm> UCC mpCuHupd .HH mamma

98
beginning of the freshman year. None of the differences are

statistically significant, and consistency is totally absent.

Conclusions Regarding Hypothesis II

The attitude and value variables bear little relationship
to changes in students' academic performance patterns in
college. There are literally no associations between scores
on such measures and change in performance for the females,
and only slight association for the male students. Of the
three instruments administered -- The Inventory of Beliefs,
The Differential Values Inventory, and Rokeach's Dogmatism
Scale -- it appears that The Inventory of Beliefs warrants

some consideration as a possible index for this purpose.

The obvious lack of relationship between the attitude
and value variables and changes in performance patterns,
prompted the investigator to omit further use of these
variablesin analysis of performance changers. The analysis
up to this point points out the random association between
the students' scores on the instruments employed and their

direction of performance change in college.

99

Hypothesis III

 

It is possible to differentiate among groups

of students classified by direction of academic
performance change on the basis of the amount
of their employment, their place of residence,
degree of participation in extra—curricular
activities, and various personal factors.

Because the variables to be examined under Hypothesis III
do not meet the assumptions necessary for use of a parametric
statistic, chi-square was selected as the tool for analysis
of the data in this section of the thesis. In only one case --
degree of participation in extra—curricular activities -- the
male and female students were analyzed separately. Analysis
of the data pertaining to the cognitive and nonécognitive-
personality variables was made for only one change point. The
present hypothesis was examined at fliree different change

points —- the sophomore, junior, and senior years.

It was the intention of the investigator to examine the
association of behavioral variables with performance pattern
change at a time of mutual occurrence. That is, behavioral
variables were analyzed for relationship during the year the
performance pattern change toOk place. For example, if
student X changed negatively during his or her SOphomore year,
then the behavioral variables relative to this student's

sophomore year were examined.

100

Chapter III presents a more detailed discussion of the
procedure followed in investigating Hypothesis III. It
should be pointed out here that some of the items had to be
omitted because of inadequate size of the sample to which the
items applied. For example, students were asked to indicate
the term of their getting married. During the early change
points, this item did not receive enough checks to be worthy
of analysis, whereas for the senior change point it became
adequate. Some of the other items were not applicable for

any of the three change periods in question.

Sgphomore Change

Table 12 shows how the sophomore year change groups, male
and female, compare on the basis of their responses to a
number of personal factors. The change groups have been
divided into three achievement groups according to their
cumulative freshman grade point average, and the table lists

both the observed and expected frequencies, along with the

chi-square value (9).

Of the 24 chi-squares, four are of sufficient magnitude
to be significant beyond the .05 level of confidence. Three
of these, it is interesting to note, occur for one variable.
All three change groups show a significant chi—square for the

item asking whether or not they pledged a fraternity or

A.wmmm£quumm mflu CH

mm.H
hm.o

@h.H

wm.m
M .
lmfi m
amm.m

ho.m

w®.m

AHV H

Ame 0

Ave m

“we 0

Ame 0H

Ame 0H

Aeavon

Aosvms

Amy 0

Amy 0H

Ame m

Ame 0

Amy 0

Amy m

Amsvom

AHHVHH

Rev

Amy

Amy

Amv

on

Amv

Ame

AHV

.moCmUHHCoo mo Hw>mH mo. mnv pCOka quoHMHCmHm¥

mum mmHoCmCUmum pouommxm “pomOHoCm DOC was mmHoCvamuw pm>uomnov

N

CH

NH.~_

mm.o

@N.H
mm.¢
Hm.o
rho.h
mm.m

go.o

nae 0

Ass 0

Amv m
Ame 0
Ame NH
Ame ma
Aeneas

Amy 0

10V m

Aonvaa

10V CH

va m

“nevus

AHHVmH

A0HV0H

Aosvofl

Amy m

Ame 0

Amy 0

Amy m

ANHVHH

Aoave

Amavsm

Amv OH

mm.o

mm.m

mm.o

mH.N

hm.m

smh.m

swh.o

wo.H

va m

Ame m

Ame 0

“we 0

Ame ma

Aoavma

Amavvm

A00 an

10H00H

Amsvma

Ame H
x00 0.
Assess
“mavmfl
Aomvmm

inflame

AHV H

Hoaema

Ame 0

“00 m

Aoavo

AHHVH

AHNVHH

Amy m

zhpmmum
mCHom:
pwmmoum

:mpmmum
mCHom=
Cmmmm

guacaHm=
050 0 mm

zpmmmmCm=

mEsomm

= ”NV.“ pom .-
memowm
.m Ho .m
ummnmam
HOnmE
UmmCan
Honda
mmOCU

 

 

 

pCm

 

 

 

 

.Cmm> muoEOCQOm ms» mCHHCp mHOHUMM HCCOmHmm mo muHHHQMOHHmmm

.UCmEm>mH£0m CMEfimmHm mo Hm>mH .mmCmnu mUCmEHomHmm

OHOEOCQOm mo COHuomHHp ha pmHMHmmmHU mqupCum mo ComHHmmEou

”NH figmfifi

102
sorority during the sophomore year. The results, however,
are inconsistent. For the low achievement group, those whose
performance pattern went down during their sophomore year
pledged a greek society less than expected. This same
situation holds for the middle achievement group, but is
altered somewhat by the high achievement group, where we see
the downward changers also pledging more than we would expect.
This occurrence becomes more understandable, however, if we
are familiar with the achievement regulations of the greek
society aspirants. Only those with grade point averages
above 2.2 are allowed to pledge. Table l on page 76 reminds
us that a 2.2 grade point average for the freshman males
falls somewhere in the middle of the medium achievement group.
The g.p.a. restriction would therefore result in fewer numbers
from the medium achievement group and eliminate students
entirely from the low achievement group. In the case of the
females, a 2.20 restriction on pledging also eliminates a

sizalfle portion of the low achievement group.

The other personal factors which apply to the students
during their sophomore year really give no evidence of the
existence of a common thread or factor which might be con-

sidered as being associated with a particular type of perform-

ance change.

103
Table 13 presents the chi-square data, male and female,
pertaining to the sophomore changers' amount of weekly employ-
ment during the sophomore year. For the low freshman achieve—
ment group, the amount of weekly employment seems to be a
meaningful idex, although interpretation of the data is

difficult. Of the positive changers, fewer than expected

TABLE 13: Comparison of students classified by
direction of sophomore change, freshman
achievement level, and hours employed
per week during the sOphomore year.

 

 

Hours Employed Per Week

 

 

 

 

 

 

Over 20 10-20 1-10 None X2
- 9 (8) 15 (12) 7 (7) 22 (25)
H O 9 (13) 18 (20) 14 (12) 45 (40) 5.41
-+ 16 gle) 18 (l8)_ 9 (11)_ .35 (37)
34 51 30 102
- 4 (8) 21 (17) 10 (ll) 43 (40)
M O 11 (9) 14 (19) 12 (12) 47 (43) 7.28
+ 9 (6) 16 (14) 10 (8) 27 4(32)
24 51 32 117
- 6 (6) 15 (13) 19 (12) 30 (38)
L O 10 (9) 22 (l9) l6 (17) 52 (54) 13.79*
+ 5 (5) 9 (12 ) 5 (l l) 44 (14)
21 46 40 126

 

(Observed frequencies are not enclosed; expected
frequencies are in the parentheses.)

*Significant beyond the .05 level of confidence

 

104
worked more than ten hours a week, while more than expected
didn't work at all. The negative changers for that group,
on the other hand, worked more than expected (had more students
working over ten hours per week) and consisted of fewer
students than expected who didn't work at all. Although
extreme caution must be used here, this does suggest an
association between amount of employment and direction of

academic performance change in college.

For the high and middle achievement groups, however, the
trend is in the Opposite direction. That is, m2£e_than the
expected number of positive changers work over ten hours per
week, and fewer than expected did not work at all. Once
again, as was the case with the fraternity and sorority
pledging, the significant difference obtained with the low
group may beaafunction of their poor freshman year achievement,
rather than their Change classification at the end of the

sophomore year.

A comparison of the sophomore change groups on the basis
of their sophomore place of residence is presented in Table 14.
As indicated, the chi-square value for the high achievement
group is high enough to be regarded as more than a chance
occurrence. The largest amount of variation here seems to be

contributed to those living in the residence halls and

105

 

 

 

 

 

 

 

 

TABLE 14: Comparison of students classified by
direction of sophomore change, freehman
achievement level, and place of residence
during the sophomore year.

Residence Sororities Married Other

Halls or Co-ops Housing Off-Campus 2

Frats. X

11 (20) 16 (10) 10 (8) 8 (8) 8 (7)

33 (32) 18 (16) 10 (12) ll (13) 14 (12) 17.60*
37 (29) 7 (15) ll (11) 15 (12) 8 (ll)

81 41 31 34 30

28 (29) 18 (14) 10 (8) ' 8 (7) 14 (17)

30 (31) 13 (16) 9 (9) 9 (8) 23 (18) 4.43
27 (23) 12 (ll) 6 (6) 5 (6) 12 (13)

75 43 25 22 49

29 (31) 6 (4) ll (9) 7 (8) 17 (16)

47 (45) (6) 10 (13) 11 (11) 27 (23) 4.99
30 (28) 3 (3) 10 (8) 9 (7) 11 (14)
106 14 31 27 55

 

*Significant beyond .05 level of confidence.

 

106
fraternities (or sororities). It can be seen that a much
smaller than expected number of negative changers lived in
the residence halls during the sophomore year, whereas the
residence halls housed a larger-than-expected number of
students characterized by positive performance changes.
Conversely, the fraternities and sororities housed a greater-
than—expected number of negative or downward changers and a
much smaller than expected number of positive performance
changers. Differences for those living in the other types

of housing appear to be negligible.

The pattern witnessed in the high achievement group for
the residence hall students and the students living in
fraternities or sororities, can also be seen, though to a
smaller extent, in the middle and low achievement groups.
This suggests that rises in performance patterns seem to be
associated with residence hall living, while decreases or
negative changes are associated with fraternity or sorority
living. These factors will be examined again at the junior

and senior change points.

The final variable examined for its association with
sophomore change patterns was the degree of student partici-
jpation in extra-curricular activities during the sophomore year.
As explained earlier. the amount of time required for these

activities has been used for the index of activities immersion.

107
Tables 15 and 16 present the activities data for the
males and females respectively for the sophomore year. The
classifications of "much," "some," and "very little or none"
were derived by simply dividing the range of weighted immersion
values into thirds which would most closely approximate an
equal distribution of students. The weighting procedure for

the activities index is described in Chapter III.

For the males (Table 15), a significant chi-square value
was obtained for the middle achievement group, Where we find
the negative changers participating more than expected in
extra-curricular activities (24 are in the "much" category,
where only 14 were expected). The positive or upward
changers, on the other hand, reflect an opposite pattern.
Fewer than expected participated in the "much" end of the
scale, whereas more than expected were found in the "none"
category. The no-change group is characterized by a minimum
of variation in the middle achievement group.

For the females (Table 16), none of the chi-square values

surpassed the chance level.

It is interesting to note that both males and females
participate in extra—curricular activities according to their
achievement level, regardless of their present direction of

performance change. Note, for example, that for both males

108

TABLE 15: Comparison of male students classified by
direction of sophomore change, level of
freshman achievement, and activities
immersion during the sophomore year.

 

 

Degree of Participation

Very little

 

 

 

 

 

 

Much Some or none X2
- 17 (15) 11 (ll) 10 (11)
H 0 20 (19) 15 (l4) l4 (14) 0.54
+ 19 (20) 16 (15) 16 g(14)
56 42 40
- 24 (17) 20 (20) 7 (14)
M 0 15 (17) 20 (19) 15 (13) 11.10*
+ 9 (14) 15 (l6) 16 (ll)
48 55 38
— ll (9) l7 (14) 14 (17)
L 0 ll (13) 19 (21) 30 (24) 3.62
+ 10 .(8)_ 13 .112) 13 (14)
32 49 57

 

*Significant beyond the .05 level of confidence

 

and females, the high achievement group is high on the “much"
end of the participation scale, and low on the "very little
or non" end of the scale, While the opposite is true for the

low achievement group.

Conclusions Regarding Sophomore Change

 

The association of behavioral variables with performance
change during the sophomore year was discovered to be slight
but encouraging. Place of residence was the most promising

variable, Where positive academic performance trends seemed

109

TABLE 16: Comparison of female students classified
by direction of sophomore change, level
of freshman achievement, and activities
immersion during the sophomore year.

 

 

Degree of Participation

Very little

 

 

 

 

 

 

Much Some or none X2
- 5 (5) 5 (6) 5 (3)
H 0 16 (l4) l6 (l4) 5 (7) 3.09
+ 9 ((19) ll (10) 7 (5)
3O 32 17 ‘_
- ll (8) 8 (9) 8 (9)
M O 10 (10) 15 (ll) 8 (10) 6.05
+ 5 ((7) 6 (7); ll (7)
26 29 27
«- 4 (5) 10 (8) 14 (13)
L O 9 (7) .10 (12) 21 (19) 1.99
+ 5 (4) 10 (8) ll (12)
18 30 46

 

None of the above X2 values are significant beyond
the .05 level of confidence.

 

to be associated with residence hall students, while those
living in a fraternity or sorority were characterized by‘
downward patterns in achievement. Pledging a fraternity or
sorority, amount of weekly employment, and degree of partici-
pation in extra-curricular activities, all seemed to bear some
relationship to the phenomenon of performance change, but the
data regarding these variables was typified by inconsistency
and great caution must be used in drawing conclusions regard-

ing their value.

110

Junior Change

 

The junior change groups were obtained by grouping
students into high, middle, and low groups according to their

sophomore grade point averages, predicting junior performance

 

on the basis of the sophomore g.p.a., and placing them into
change groups according to the discrepancy between their ob-
tained junior grade point average and their predicted junior
grade point average. A more detailed explanation of the
grouping procedure and isolation of the performance change

groups is presented in the first chapter.

Table 17 presents the data relating to the association
of various personal factors and students' academic perform-
ance change for the junior year. When applied to the sopho-
more change groups, pledging a fraternity and sorority seemed
to be the key personal factor. For the junior changers,
however, this variable does not apply. Four of the obtained
chi-square values rise above the .05 level of probability and
some of the tentative indications revealed by the data in
Table 17 are:

1) Downward performance changers are more apt to change

their major field. become "active" in a fraternity
or sorority, terminate a stable heterosexual relation-

ship, whereas,

1

.mOCOOHMCOO mo Hw>mH mo. 0:» pCOth uCMOHMHCmHma

A.wmmm£qu£uumm mflu CH mum mmHOCmCUmHH pmuommxm upmmoHOCm HOC mum mmHOCOCUmHm po>nmmflov

 

 

 

 

 

 

mm.mmim0 m 100 0 H40 0 .00.H WAHV m Ame 0H Has 0 0H.H 1H0 m 100 0H Hmv H .N0mmum
mCHom=pmmmonm
0N.m_ E HH 2: m 7: m H00 :3 m 8va 23 NH 00.1. E 0 E 0 H3 0 $003....“
M mCHom: Cmmmm
00.0” HNV N Na N H: m 0H.0 H3 0 H8 0 H3 0 .30.: H3 0 E H E 0 =92:ng
# mxonm
mm.0 HHV 0 Hmv 0 Hoe 0 HN.N .HoHcH HNHch HNHV0H 0H.0 HHHVN HchNH Hoe mH =0mcch=
wEmomm
00.0 Hoe m HHV H Ame s mm.0 HHV NH Hmv 0 Has 0 00.N Hoe 0 HHV H Hmv m 000m0am
mEmOOm

l s e
u mN N Hwy 0 Ame 0 Hmv 0 :mo H Hwy N Ame 0 HHV NH No.0 Hoe 0 HHV m Ame m =m>Huom=
mfimomm
HH.0 HNV m Hmv N HNV m Hm.m HNV H HNV m HNV H 0N.H Hes H Amy 0 Amy N .m uo.m
pmmcmHm
:HN.H H05 0 H0H00H HHV NH 0H.0 HHV 0 Hmc 0 Hmc m N0.H Hova HNHvsH Hmc 0H uona
meCmgu
mm.0 HNV H HNV N HNV N 00.N Hmv 0 Ass N Hmc H 0H.0 Ame 0 H0c 0 Ave 0 Hone
mmogu

Nx + 0 Nx + 0 1 Nx + 0 .1
mem mannHz 30H

1

.HCO% HOHCCH any mCHHCO muouomm HMCOmHmm mo MHHHHQCOHHmmm

OCm .uCoam>mHCOm OHOEOCaom mo Hm>mH
uoHCdn mo CoHuomHHp >9 OOHMHmmmHO mqupCum mo COmHHsmEOU

.omCmno OOCmEuomumm

uhH mflmde

112
2) positive performance changers were less likely to
change their major field, become "active" in a
Greek social society, or break a "pinning'I or going

“steady“ arrangement.

Examination of the junior change personal factor table
(Table 17) reveals, by way of supporting the points listed
above, that negative performance changers changed their major
during the junior year in greater than expected numbers and
the positive changers changed choice of major field less than
expected (this was significant for the high achievement group);
that they became “active" in fraternities or sororities more
than expected while the positive changers did so less than
expected (significant for the middle achievement group): that
they brdke “pinnings” and "steady" relationships more often
than expected, while the positive changers did not (signifi—

cant for low and middle achievement groups).

Consistency seems to be present for the significant vari-
ables discussed. Although in each instance a significant
chi-square value was achieved in only one of the three achieve-
ment groupings, similar results, though not as large, were

usually the case in the other achievement levels.

113

Data pertaining to amount of weekly employment as a

variable associated with junior year academic performance

change is presented in Table 18. As was the case with the

sophomore changers, they are classified according to achieve-

ment grouping (based on sophomore grade point average),

direction of performance change, and amount of hours employed

per week during their junior year.

 

 

 

 

 

 

 

 

TABLE 18: Comparison of students classified by
direction of junior change, sophomore
achievement level, and hours employed
per week during junior year.

Hours Employed Per Week
Over 20 10 - 20 1 - 10 None X2
- 12 (9) 12 (13) 10 (9) 27 (28)
H O 12 (12) 16 (18) 14 (13) 42 (39) 3.26
+ 9 (10) 20 (15) 10 (ll) 33 £33)
33 —48 34 102
- 11 (9) 15 (18) 10 (10) 42 (39)
M O 12 (10) 17 (18) 13 (10) 37 (40) 5.28
+ 5 (8) 19 (14) 6 (8) 33 (32)
28 51 29 112-
— lb (5) 9 (12) 10 (11) 32 (32)
L 0 5 (7) 14 (18) 21 (17) 52 (49) 13.80*
+ 4 (6) 22 (14) ll (13) 37 (39)
19 45 42 121

 

*Significant beyond the .05 level of confidence.

 

For the low adhievement group, a significant chi-square

'value was obtained. Most of the O—E differences contributing

114
to this large X2 value seem to occur in the negative and
positive change groups. Here we see the downward changers
having a greater-than-expected number of students working
over 20 hours per week and the upward changers having fewer
than expected working over 20 hours per week. It can also
be shown that this pattern is the same for all achievement
groupings, although the deviation from the expected frequencies

is not as great in those cases.

Once again this suggests that the more hours a student
is employed during the academic period, the greater is the
likelihood that this student will witness a downward change
in his or her academic performance pattern. At this point,
however, such a conclusion cannot be warranted, and further

analysis seems to be called for.

A comparison of the junior changers on the basis of their
place of residence during their junior year is presented in
Table 19. None of the obtained differences were large enough
to be attributable to more than chance occurrences. In each
of the achievement groups, those who lived in the residence
halls seemed to experience positive performance changes more
often than expected, while those living in other off-campus
living units placed a fewer than expected number in this

category. In any case, the over—all variation was not

TABLE 19:

115

Comparison of students classified by

direction of junior change,

sophomore

achievement level, and place of residence
during the junior year.

 

 

 

 

 

 

 

 

Residence Frat. or Married Other
Halls Sorority Co-Op Housing Off—Campus X
17 (19) 16 (12) 9 (8) 7 (7) 12 (14)
25 (26) 16 (16) 11 (ll) 8 (10) 24 (19) 6.47
26 (22) ll (14) 9 (9) 12 (8) l4 (16)
68 43 29 27 50
21 (21) 15 (15) ll (10) 9 (10) 22 (18)
20 (22) 16 (15) 10 (10) 11 (11) 22 (19) 4.68
21 (17) 13 (12) 8 (8) 11 (8) 10 (15)
62 44 29 31 54
20 (21) 7 (6) 12 (9) 8 (6) 14 (16)
34 (32) (9) 11 (14) 10 (10) 28 (24) 3.11
27 (26) 8 (7) 13 (11) 7 (8) l9 (19)
81 24 36 25 61

 

None of the above X2 values are

the .05 level of confidence.

significant beyond

 

116
sufficient to warrant any conclusions regarding the positive
or negative effects of any of the living units being con-

sidered.

Tables 20 and 21 contain the data regarding the amount
of participation in extra-curricular activities during the
junior year. Table 20 presents this information for the male
students, and Table 21 for the females. For the male high
achievement group a significant difference was found, a
difference which seems to be due, once again, to the variation

occurring in the positive and negative change groups.

TABLE 20: Comparison of male students classified by
direction of junior change, level of
sophomore achievement and activities
immersion during the junior year.

 

 

Degree of Participation

very little

 

 

 

 

 

 

Much Some or none x2
— 20 (13) 6 (12) 9 (9)
H O 21 (23) 22 (20) 16 (15) 10.63*
4— 10 4(14) l7g_(lg) 9 .(9)
5T 45 34
- 18 (l7) l4 (l4) l7 (16)
M 0 l6 (14) 12 (12) 12 (13) 1.03
4- 13 (15)_ 14 (13) 16 (14)
47 40 45
- 15 (14) 9 (10) 14 (12)
L 0 20 (21) 17 (15) 18 (17) 1.08
+ 20 (18L 15 L14) 14 (15)
55 41 46

 

H

*Significant beyond the .05 level of confidence.

 

117

TABLE 21: Comparison of female students classified
by direction of junior change, level of
sophomore achievement, and activities
immersion during the junior year.

 

 

Degree of Participation

very little

 

 

 

 

 

 

Much .Some or none X2
-- 15 (10) 5 (7) 6 (8)
H O 12 (10) 5 (6) 8 (7) 8.09
+ 10 _(l5)_ l3 (9) 13 (ll)
37 23 27
- 12 (11) 9 (9) 8 (7)
M - 0 16 (15) 12 (12) ll (10) 0.59
+ 7 (7) 8 (6). 5 ¥(5)
35 29 24 -
— 7 (5) 8 (8) 8 (9)
L 0 8 (8) 14 (13) 15 (14) 1.48
+ 4 (5) 10 (9) 11 (10)
19 32 34

 

2
*None of the above X values are significant
beyond the .05 level of confidence.

 

For every level of achievement in both the male and
female tables (Tables 20 and 21), it can be seen that the
negative performance changers participated in greater—than-
expected numbers in activities at the "much” end of the
continuum, while in every case but one, the positive perform-
ance changers were represented in this category less often
than expected. This observed minus expected discrepancy is
great enough to result in a significant X2 value for only the
male high achievement group, but it is nevertheless present

in every case._

118
The findings regarding degree of participation in extra-
curricular activities are in keeping with the findings at the
end of the sophomore year. In both of these cases, there
seems to be a tendency for the negative performance change to
be associated with a greater—than-expected frequency in the
“much” participation category. The opposite pattern is con-

sistently found for the positive performance changers.

Conclusions Regarding the Junior Change

 

On the basis of the data presented regarding the associ-
ation of various behavioral variables with academic perform-
ance change during the junior year, these trends seem to be
present: in comparison with the positive changer, the
negative performance changers are more apt to change their
major during the year, more likely to become active in a
fraternity or sorority, terminate a stable, heterosexual
relationship, participate heavily in extra-curricular
activities, and hold a job requiring over 20 hours of weekly

employment.

Senior Chang§_

 

The senior change groups were obtained by grouping
students into high, middle, and low groups according to their
junior grade point averages, predicting senior performance on

the basis of this level of attainment, and placing them into

119
change groups according to the discrepancy between their
obtained senior grade point average and their predicted senior
average. A more detailed explanation of the grouping procedure
and isolation of the performance change groups is presented

in the first chapter.

Table 22 contains the data regarding the association of
various personal factors with students' academic perfonnance
change during the senior year. The first obvious difference
between this type of analysis for the senior change group as
opposed to the change groups for both of the first two years,
is the reduction in the number of personal variables which
apply. Sudh factors as "changed major field," "made first
choice of major field," "pledged fraternity or sorority," and
"became active in a fraternity or sorority," have all lost
application to the students during their senior year in
college. On the other hand, such items as "became ‘pinned,'"
and "became engaged" are characterized by a marked increase
in applicability (as evidenced by larger number of students
checking them) and the item “got married" now enters the

analysis for the first time.

For the middle achievement group there is a significant
difference on the becoming engaged factor, where fewer than

expected became engaged for the negative change group

 

.mocwowmcoo mo Hm>ma m0. 039 020%09 HGMUHMHGmHms

 

 

 

 

 

 

 

RN 2: a E N E m mH.H x3 4 E m as o 34 E m H3 H E m 163$
maHow=
Gmmmm
4N.¢ Ame m Amy 4H Hos m H¢.o HHV 6 Amy HH Hos m HH.¢ HoHVOH HOHVmH HHV 4 .6mcaHm.
oamomm
m Ho.m Hmv m HoHVHH Hoe oH HH.o Has 4 Hmv m Hmv m mm.H Ame m Hoe 6 Ave H anuumz
How
mo.H HNHVNH HoHVmH HoHVOH me.H Hmv m AHHVNH HHHVo No.4 ANHVHH HNHVmH Has 6 nmmmmcm
mfimomm

mx + o .l mx + 0 II «x + 0 II

mon mAQQHS 304

 

.Hmw> Hoacmm mgu msHHSG muouomm anaemumm mo hnflaanmofiammm

0cm

.ucmEm>mN£Um Hoacsm mo Hwbwa
Reason «0 sofluomnao ma omHMHmmmHo mucmosum mo confiummfiou

.mmsmnu mUGMEHowumm

“NN mflmdfi

121
and more than expected did so for the "freeze" or no change
group. This can hardly be interpreted as evidence that failure
to establish stable heterosexual relationships leads to
negative performance changes, however. For both the high and
low achievement groups, the negative changers are typified by
greater-than-expected numbers getting married and beginning
"going steady" relationShips. The association of these factors,
then, seems to be random and sheds no light on the identifi-
cation of the various types of academic performance changers,

at least for those changes occurring during the senior year.

Comparison of the senior change groups on the basis of
amount of weekly employment during the school year is presented
in Table 23. This variable proved to be a promising one When
applied to the sophomore and junior changers, but it yields
no meaningful index of association when analyzed for the
senior changers. As noticed before, the negative changers,
regardless of achievement level, seem to work over 20 hours
per week more often than expected, while the opposite is true
of the positive performance changers. This consistency is
encouraging, but, nevertheless, the differences obtained
cannot be claimed to be more than chance occurrences at the
.05 level of probability. It can also be pointed out that the

relationship between level of achievement and employment still

122

TABLE 23: Comparison of students classified by
direction of senior change, junior
achievement level, and.hours of weekly
employment during the senior year.

 

 

Hours Employed Per Week

 

 

 

 

 

 

Over 20 10 - 20 l - 10 None X2
- 11 (9) 10 (10) 15 (14) 23 (24)
H o 14 (14) 15 (16) 20 (21) 39 (36) 1.11
+ 12 (12)_ 16 (14) 19 (18L 3o (31)
37 41 54 92
— 15 (12) 16 (14) 17 (20) 33 (33)
M o 13 (12) 14 (14) 21 (20) 34 (34) 2.76
+ 7 (9) 9 (10) 18 (15) 27 (25)
35 39 56 94
- 10 (8) 13 (10) 18 (17) 16 (20)
L o 12 (11) 14 (14) 22 (23) 31 (28) 3.35
+ _9 (11) 13 (14) 25 (2;) 31 (2g
31 4o 65 78

 

None of the above X2 values are significant beyond
the .05 level of confidence.

 

holds, with the higher achievement group working more than the

middle and low achievement groups Who follow in that order.

How place of residence is associated with one's senior
year performance change is presented in Table 24. The 21.07
chi-square value obtained for the high achievement group is
significant beyond the .05 level of confidence, and is in
agreement with the nature of the differences reported for both
the sophomore and junior change groups. For every negative

change group, regardless of achievement level, there is a

123

 

 

 

 

 

 

 

 

TABLE 24: Comparison of students classified by
direction of senior change, junior
achievement level, and place of
residence during the senior year.

Residence Frat. Married Other Off-
Halls Soror. Co-ops Housing Campus X
- 12 (14) 17 (9) 9 (7) ll (9) 10 (18)
H 0 18 (21) 12 (14) 10 (11) 12 (14) 36 (27) 21.07*
+ 24 (18) 7 (12) 9 (9) 1,3 (12) 24 (24)
54 36 28 36 7O
— 17 (18) 16 (14) 9 (9) 8 (10) 31 (27)
M 0 19 (18) 15 (15) 10 (19) 10 (10) 28 (28) 3.06
+ 14 (13) 10 (ll) 8 (j) ll (7)_» 18 (20)
50 41 27 29 77
- 12 (13) 13 (8) 10 (8) 7 (9) 15 (17)
L 0 17 (18) ll (11) 10 (11) 12 (12) 29 (24) 7.26
+ 20 (17) 8 (ll)_ 11 (ll) 16 (12) 23 (24)
49 32 31 35 67

 

*Significant beyond the

.05 level of confidence.

 

smaller than expected number

of students who lived in the

residence halls, and for every positive change group, regard—

less of the level of achievement, there is a larger than

expected number of students who lived in the residence halls.

This pattern is reversed for those who spent their senior

years residing in a fraternity or sorority, and, once again,

this finding is in keeping with.what was reported for the

sophomore and junior changes.

Those living in married housing

are also found in the positive change groups more often than

expected.

124

For the high achievement group, the largest O-E discrep-
ancy is found for those living in “other off-campus” housing
and occupying the "freeze" or no-change group. Review of the
tables for the sophomore and junior change groups will show
that those living in other off-campus housing consistently
placed more than the expected number of students in the
"freeze“ or no—change category. Stability of academic perform-

ance seems to be their characteristic in this regard.

Tables 25 and 26 present the data regarding the associ—
ation of participation in extra-curricular activities and
direction of academic performance change, for males and
females respectively. None of the obtained chi—square

values surpass the .05 level of probability.

Conclusions Regarding Senior Changg_

The factors found to be associated with changes in
scholastic performance patterns during the senior year were
fewer in number than those associated with sophomore and
junior year performance changes. One‘s place of residence
seemed to be the best index, Where it was noticed that resi—
dence hall living was associated with more than the expected
number of positive performance changers, fraternity and/or
sorority living during the senior year was associated with

more than the expected number of downward performance

125

 

 

 

 

 

 

 

 

TABLE 25: Comparison of male students classified
by level of junior achievement,direction
of senior academic performance change, and
activities immersion during senior year.
. 2
Much Some Very Little X
- 14 (12) 13 (13) 11 (11)
H 0 14 (16) l7 (17) 18 (15) 1.24
+ 20 (18) 20 (191. 16 g(11)
48 50 45
— T7 (12) 12 (14) 17 (18)
M 0 13 (13) 14 (15) 22 (19) 6.97
+ 6 (}0)_fi_ 18 (13);f l6 (16)
36 44 55
— “7 (7) 14 (12) 16 (16)
L 0 1J. (11) l6 (16) 25 (23) 0.78
+ 11 4(9) l4 (14), 21 (21)_
29 44 62

 

None of the above X2 values are significant beyond
the .05 level of confidence.

 

 

 

 

 

 

 

 

 

TABLE 26: Comparison of female students classified
by level of junior achievement, direction
of senior change, and activities immersion
during the senior year.

Very little 2
Much Some or none X
- 9 (6) 6 (7) 6 (7)
.H 0 11 (11) ll (l3) l7 (14) 4.73
+- 4 (6) 10 (7); 7 (7)
24 27 30
— 9 79) 13 (14) 13 (11)
M 0 8 (8) 14 (13) ll (11) 0.54
+ 6 451 9 18) 6 (D
23 36 '30 1.
- 7 (5) 6 (6) 7 (8)
I. 0 7 (7) 9 (8) ll (11) 1.63
+ '7 (8) 9 (9) l6 (13)
21 24 34

 

None of the above X2 values are significant beyond
.05 level of confidence.

 

126
changers, and those living in other off-campus units were

characterized by achievement stability.

One's involvement in extra-curricular activities, amount
of employment during the school year and other personal
variables seemed to be of little utility in ferreting out
those factors which might have some relationship to dips and
rises in the scholastic performance patterns of college

students.

 

CHAPTER V

SUMMARY, CONCLUSIONS, DISCUSSION, AND
IMPLICATIONS FOR FURTHER RESEARCH

Summary

Purpose and Procedure

 

The purpose of this investigation was to explore the
relationship between selected variables and changes in the
academic performance patterns of students during four years
of college. The study was designed to differentiate between
students whose performance pattern improved as opposed to
those whose pattern of achievement changed negatively.
Students whose patterns of performance did not change at all
were also compared. The basic purpose was to examine the
factors thought to have association with the change in
performance at the time the change occurred. If, for example,
a student's performance changed negatively during the junior
year, then "factors" having relevance for that student during

the junior year were examined.

The sample selected for the study was composed of 1,041
students who entered Michigan State University as freshmen

fall term, 1958, and were in attendance during each fall,

127

128

winter, and spring term through spring term of 1962.

Subjects were classified into change groups on the basis
of the difference between their performance predicted from
previous year‘s grade point average (determined by linear
regression) and their actual performance. Any discrepancy
greater than a confidence band of .50 standard errors of
estimate was used as the criterion for change classification.
Since the instruments used for measures of the cognitive
variables have been demonstrated to correlate highly with
academic achievement, the groups were further divided into
high, middle, and low achievement groups. This procedure
for determination of the performance change groups was con-
ducted separately at the beginning of the sophomore, junior,
and senior years. As a result, there were nine groups -—
three change groups at each of the three levels of achieve-

ment, analyzed at the end of each academic year.

The change groups were then compared on three types of
experimental variables. The variables were: (1) ability
and aptitude, (2) attitudes and values, and (3) behavior.
Used as measures of the ability and aptitude variables were
The Michigan State University Reading Test, The College

Qualification Test, and A Test of Critical Thinking, Form G.

 

For measures of attitudes and values The Inventory_9f Beliefs,

 

129

The Differential values Inventory and Rokeach's Dogmatism

 

Scale were employed. The behavioral variables were measured

by an experience inventory developed by the investigator.

The instruments used as measures of ability and aptitude
were administered only at the beginning of the freshman year.
The instruments used as measures of attitudes and values were
administered at the beginning of the freshman year (fall,
1958) and at the end of the freshman year (spring, 1959),
while the instrument used to gather information regarding the
behavioral variables was administered only during the spring
term of 1962, Which was the last term in school for students

who had progressed at a normal rate.

Analysis of variance was used as the basis for determining
the significance of the differences between the performance
change groups on the first two types of variables, whereas
chi-square tests were employed for analysis of the data per-
taining to the behavioral variables. Data regarding the
ability-aptitude and attitude-value variables were analyzed
separately for males and females, while the data pertaining
to the behavioral variables were usually not divided accord-
ing to sex. Throughout the study a .05 level of probability

was used in determining statistical significance.

130

Findings

 

The instrumentsemployed as measures of ability and
aptitude failed to discriminate between male students classi-
fied according to level of achievement and direction of
performance change, and in only one instance differentiated
between these groups for the females. Marked differences

were revealed between achievement groupings —— the high

 

achievement group recorded consistently higher cognitive
scores than the middle or low achievement groups -— but not

between performance changes within a given achievement group.

This lack of association is true also for the attitude
and value variables. There were no significant F-values
forthcoming from the analysis for the females. and the
association between such factors and change in academic
performance was suggested on only one occasion for males.

In addition to no differences between change groups, differ—
ences between achievement levels were also lacking for the

personality measures.

The behavioral variables were more helpful in distinguish—
ing between performance changers. Those making a positive
change during the sophomore year, for example, were typically
those who: (1) pledged and joined a fraternity or sorority

during that year, (2) worked fewer hours at part—time jobs

131
than negative changers, (3) lived in a residence hall (where-
as negative sophomore changers tended to list a fraternity or
sorority as their place of residence), and (4) participated
less in extra-curricular activities. During the junior year,
those making positive changes (as opposed to those making
negative changes) were characterized by: (l) fewer Changes
of major field, (2) fewer students becoming "active“ in
fraternities and sororities, (3) less employment, and
(4) less participation in extra—curricular activities. The
behavioral variables associated with changes in academic
performance during the senior year were not as pronounced as
during the sophomore and junior years, but still provided some
clues regarding behavioral differences. Positive changers
during the senior year more likely lived in residence halls,
whereas negative changers more likely lived in fraternity

or sorority houses.

0f the three “types“ of variables then, it was found that
the behavioral variables were most frequently associated with

academic performance changes of college students.

132

Conclusions

 

It seems evident that, of the variables analyzed in this
investigation, some show consistent trends of association with
changes in scholastic performance patterns, While others are

not associated in any discernible way.

Measures of ability and/or aptitude and measures of
dogmatism and values all seem to be typified by this lack of
association. Though in some cases related to academic achieve-
‘mgntJ they seem to be completely unrelated to changes in
scholastic performance. Although several F—values obtained
from the analyses of variance did reach a level exceeding
significance at the .05 level of confidence, the infrequency
and inconcistency indicate that these associations are probably
random and merely chance occurrences. The only variable which
might be an exception to this is the variable of beliefs.

Male students in the positive performance change groups con-
sistently recorded higher mean scores on the Inventory of
Beliefs. This suggests that the positive changers are those
'with less rigid personality structures and that this flexi—
bility is somehow related to positive changes for males.

The differences reported, however, are not of such a magnitude

to be regarded as statistically significant differences, and

133
it must be concluded that, on the basis of this study, vari-
ables of attitude and value or ability and aptitude nature
(as described in this investigation) are 22£_associated with
changes in scholastic performance patterns of college

students.

Behavioral variables, however, show considerable associ-
ation with scholastic performance change. The type of
behavioral variables associated with changes varies from year
to year and, in some cases, consistency is lacking. Neverthe-
less, the behavioral variables certainly warrant consideration
as factors being associated with college scholastic perform-
ance changes. Some of the conclusions regarding the behavioral
variables which seem to be suggested on the basis of the
data are:

1) Emplgyment:

 

The amount of weekly employment during school is
consistently greater for the high achievement group
and, except for the sophomore changers, seems to be
associated with negative changes in scholastic
performance. That is, the more time one devotes to
a part-time job While in school, the greater the
likelihood that that student's grade point average

will drop.

2)

3)

134
Place of Residence:
Dormitory or residence hall living appears to be
associated with positive changes in achievement,
while fraternity or sorority living seems to be
associated with negative changes in academic
performance. For each of the three years and for
each of the three achievement levels, the negative
changers live in fraternity or sorority housing in

greater-than-expected numbers, while the positive

 

“" {nu-43' ”I." Q '

changers live in residence halls more often than
expected. Associations are not as obvious or con-
sistent for the other types of housing. but it does
seem that married housing dwellers are typified by
positive changes in academic performance, while off—
campus students are found in the negative change

category more frequently than expected.

Activitievammersion:

The degree of involvement in extra-curricular
activities, while not as consistent as the place of
residence variable, apparently does have some associ—
ation with change in scholastic performance. Those
whose grade point average changed negatively were

consistently involved in extra-curricular activities

135

more often than expected, while positive changers

were characterized by a lesser degree of activities

immersion.

4) Personal Factors:

 

The association between scholastic performance
change and the personal factors examined in this
investigation vary for each of the sophomore, junior,
and senior years. During the sophomore year, there
is a significantly greater likelihood that those

who pledged a fraternity or sorority will be positive
changers. During the junior year, this factor loses
its applicability, but becoming ”active" in a
fraternity or sorority takes its place and seems to
be associated with negative changes. Junior year
negative changers are also more likely to change
their major field. These personal factors all lose
their applicability during the senior year, where
one finds little or no association between scholastic

performance change and the personal factors considered.

Discussion

 

Out-of-class activities have often been overlooked

as factors possibly influencing the scholastic attainment

136
patterns of college students. Organized student activities
have generally been encouraged on the basis of the socialization
value they hold for the student; part-time employment has been
generally accepted as vitally necessary for many, with the
apparent assumption made that it seldom affects academic
achievement; and the many personal factors affecting each
student have usually been regarded as research possibilities
only for those interested in group dynamics or other forms of
social psychology. The few attempts that have been made to
explore any relationships that might exist between scholastic
performance and the above variables have generally reported

no negative effects of any of these factors.

The findings of this investigation suggest a some-
What different conclusion. Exploring this from the standpoint
of changes in patterns of scholastic performance (rather than
a cumulative grade point index), the data leads one to the
hypothesis that such outside—of-class activities as those
mentioned above dg_affect one‘s scholastic performance and,
further, that this effect is a negative one. To accept this
as conclusive would be erroneous, for only associations have
been explored and to say that a cause and effect relationship
exists would be going beyond the justifiable limits of the

data presented. Nevertheless, association between two

137
variables always suggests the possibility of one causing the
other, and in this case, particiaption in various out—of-class
activities naturally leads to questions regarding its impact

on scholastic performance.

The facts are these: examination of student partici-
pation in extra-curricular activities from the standpoint of
amount of time required, reveals that negative scholastic
performance changers were consistently involved in extra-
curricular activities more often than expected. Living in a
fraternity or sorority house is also consistently associated
with a negative change in scholastic performance patterns,
and, though not to the same degree, part—time employment is

associated with a downward change.

The basic nature of the relationship between the
student's outside—of-class activities and in-class performance
is challenged by these data. And since the findings seem to
contradict the findings reported by other researchers, one
cannot help wondering if the techniques employed in this
investigation are not worthy of consideration in future
research of this kind. Particularly the pattern analysis

approach to achievement seems to be promising.

138

In any event, factors such as those examined in this
investigation must be examined more thoroughly in terms of
their relationship to academic performance. Only a fraction
of the variance in academic achievement is accounted for by
such standard indices as ability, aptitude, and previous
performance, and even the relationships between these variables
and scholastic achievement diminishes rapidly after one year
in college. In fact, little attention has been given to
achievement beyond the first one or two years, and it is here
that behavioral variables of the kind used in this investi—
gation might prove especially helpful. The steady decrease
in relationship between scholastic achievement and measures
of ability and aptitude suggests the possibility of other
factors contributing to the variance in performance. New
demands on the student's time, suoh as part-time employment,
or more involvement in organized activities, could well be

the factors in question.

If this is the case -- and the data resulting from this
study seem to give this indication - then a new look at the
student out—of-class seems to be in order. Extra-curricular
activities, it seems should still be regarded as "extra," and
closer examination must be given to the effect they have on
the academic performance of those who consider such activity

as a major purpose or goal of higher education.

 

139
The findings of this investigation suggest continued use
of behavioral variables as possible factors in scholastic
performance, and indicate that the concept of performance

change analysis provides a promising methodological technique

for such an endeavor.

Implications For Further Research

An attempt was made in this study to identify factors
which seem to be related to changes in the scholastic perform—
ance patterns of college students. The variables chosen for
this study have by no means eXhausted the possibilities in
this area. On the other hand, a number of the variables
analyzed in this investigation have been demonstrated to
lack association. The results of thiszstudy, then, suggest

the following considerations for further research in this area.

1) A more concentrated emphasis on behavioral variables
would be a desirable approach, with the further
suggestion of an interview technique to ascertain

the factors associated with performance changers.

2) The present investigation considered only associ-
ations between various factors and performance

change. Ideally a study would be designed to

3)

4)

140
isolate the causes of the performance change.
This is closely related to the first suggestion,
since an interview technique would probably be
one of the more promising techniques employed in
attempting to ferret out the reasons and causes

of academic performance change.

A replication of the present study or any approach
which has a similar purpose, should make a thorough
analysis of such factors as repeated courses, credit-
hour load, curriculum, and level of courses taken,
and control on these variables in the design of

the research.

Elimination of the achievement grouping would be
desirable. This technique was employed in the
present study because of the known relationship
between some of the variables analyzed and the
level of student achievement. In further research,
however, since it is likely that ability and
aptitude factors would be omitted, the adhievement
grouping would also be a technique to omit. This
would considerably reduce the number of groups to
be analyzed. and yield easier interpretation to the

data.

5)

6)

141

If possible, a larger confidence band should be used
in determining changes in scholastic performance.

The standard error of estimate chosen will determine,
to a great extent, the nature of the performance
change groups. Utilization of a larger standard
error would result in larger actual performance
changes for those being classified in one of the

change groups. 1

Continued use of changes in performance patterns
as a methodolOgical technique is encouraged, along
with such analyses in the students' later terms in

school.

10.

BIBLIOGRAPHY

Aller, Florence D. "Some Factors in Marital Adjustment
and Academic Achievement of Married Students," The
Personnel and Guidance Journal. XLI (1960), 217-220.

 

Anastasi, Anne. Psychological Testing. New York: The
Macmillan Company, 1961.

 

Bach, Mary L. "Factors Related to Student Participation
in Campus Social Organizations," Journal of Social
Psychology. LIV (1961). 337-348.

 

Bennett, George K., Bennett, Marjorie G., Wallace,
Winburn L., and Wesman, Alexander G. College Qualifi-
cation Tests, Manual, 1957. New York: The PsycholOgi-
cal Corporation, 1957.

 

 

Bloom, B. S., and Heyns, I. Dev. "Development and
Applications of Tests of Educational Achievement,"
Review of Educational Research. XXVI (1956), 72-88.

 

Buckner, Donald R. "A Comparison of the Scholarship of
Freshmen Fraternity Pledges and Residence Hall Men,"
Journal of College Student Personnel. III (1961),
20-25.

 

Clark, Burton R. "The 'Cooling-Out' Function in Higher
Education," American Journal of Sociology, LXV (1960),
569-576.

 

Computer Laboratory, Michigan State University. "KS-M,
Correlation, Means, Standard Deviations, Variances:
Card Input," MISTIC Library Index. East Lansing,

 

Michigan: 1959.

Computer Laboratory, Michigan State University. ”K6-M,
Chi-Square for K x L Tables," MISTIC Library Index.
East Lansing, Michigan: 1957.

 

Cooperative Study of Evaluation in General Education,
Paul L. Dressel, Director. Instructor's Manual for the
Inventory of Beliefs. The American Council on Education,
Committee on Measurement and Evaluation, 1953 (Mimeo-
graphed).

 

 

142

ll.

12.

13.

14.

15.

l6.

17.

18.

19.

20.

143

Cooperative Study of Evaluation in General Education,
Paul L. Dressel, Director. Instructor's Manual for the
Test of Critical Thinking, Form G. The American Council
on Education,rCommittee on Measurement and Evaluation,
1953 (MimeOgraphed).

 

 

Crookston, Burns B. “Academic Performance of Fraternity
Pledges," The Journal of College Student Personnel.
I (1960), 19-22.

 

Dickinson, Carl, and Newbegin, Betty. "Can Work and
College Mix," Personnel and Guidance Journal. XXXVIII
(1959), 314-318.

 

Diener, Charles L. "Similarities and Differences Between
Over-Achieving and Under-Achieving Students," The

Personnel and Guidance Journal. XXXVIII (1960), 396—400.

 

Dressel, Paul L., and Mayhew, Lewis B. General Education:
Explorations in Evaluation. washington, D.C.:
American Council on Education, 1954.

 

 

Ehrlich, H. J. "Dogmatism and Learning," Journal of
Abnormal and Social Psychology. LXII (1961), 148-149.

 

 

Erb, Everett D. "Conformity and Achievement in College,"
Personnel and Guidance Journal. XXXIX (1961), 361-368.

Farnsworth, Dana S. "Some Non-Academic Causes of Success
and Failure in College Students," College Entrance
Examination Board, College Admissions, No. 2. Princeton,
New Jersey: The Board, 1955, 72-78.

 

Fink, Joseph. "Collecting and Quantifying Social Partici-
pation Information," The Personnel and Guidance Journal.
XXXVI (1958), 417-420.

 

Fisher, Margaret B. "Trends in College Students' Grades,"
The Personnel and Guidance Journal. XXXIX (1961),
491-494.

 

21.

22.

23.

24.

25

26.

27.

28.

29.

30.

144

Fishman, Joshua A. "Non-Intellective Factors as Predictions,
as Criteria, and as Contingencies in Selection and Guid—
ance of College Students: A Socio—Psychological Analysis,"
in Conference on Selection and Educational Differentiation.
Selection and Educational Differentiation: Proceedings.
Berkeley, California: The Center for the Study of
Higher Education, 1960, 55-73.

 

Fredericksen, N., and Schrader, W. B. Adjustment to
College. Princeton, New Jersey: Educational Testing
Service, 1951.

 

Fredericksen, N., and Schrader, W. B. "The A. C. E.
Psychological Examination and High School Standing as
Predictors of College Success," Journal of Applied
Psychology. XXXIX (1955), 39, 49-52.

 

 

Freedman, N. B. "The Passage Through College," Journal
of Social Issues. XII (1956), 13-28.

 

Freeman, Edward M., and Johnson, Palmer 0. "Prediction
of Success in the College of Agriculture, Forestry, and
Home Economics," in University of Minnesota Studies in
Predicting Scholastic Achievement, Part I. Minneapolis:
University of Minnesota, 1942.

 

 

Frick, J. W. "Improving the Prediction of Academic
Achievement by Use of the M.M.P.I.," Journal of
Applied Psychology. XXXIX (1955), 49-52.

 

 

Fricke, B. G. The Opinion, Attitude and Interest Survey.
Minneapolis: Investors Diversified Services, 1955.

 

Garrett, Harley F. "A Review of and Interpretation of
Investigations of Factors Related to Scholastic Success
in Colleges of Arts and Science and Teachers Colleges,"
Journal of Experimental Education. XVIII (1949), 91-138.

 

Gough, H. G. "Factors Related to Differential Achievement
Among Gifted Persons." Paper read at the annual meeting
of the American Psychological Association, San Francisco,
September, 1955.

Gough, H. G. "The Construction of a Scale to Predict
Scholastic Achievement," Journal of Applied Psyghology.
XXXVII (1953). 361-366.

 

 

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

145

Hansmeier, Thomas W. "The Iowa Tests of Educational
Development as Predictors of College Achievement,"
Educational and Psycholpgical Measurement. XX (1960),
843-847.

Harris, P. "The Relation of College Grades to Some
Factors Other Than I.Q." Archives of Psychology, XX
(1931), 136-141.

 

Heist, Paul A. "Diversity in College Student Character—
istics." Journal of Educational Sociology. XXXIII
(1960), 217-220.

Heist, Paul A. "Variations in the Personality Character-
istics of Underachieving College Students." Paper
presented at the annual meeting of The American College
Personnel Association, Philadelphia, 1960.

Heist, Paul A., and Williams, Phoebe, A. "Variation in
Achievement Within a Select and Homogeneous Student
Body." Unpublished research report. Center for The
Study of Higher Education, University of California,
1958.

Hodges, Harold M. "Campus Leaders and Non-Leaders,"
Sociology and Social Research. XXXVII (1953) 251-255.

Holland, John L. "The Prediction of College Grades from
the California Psychological Inventory and Scholastic
Aptitude Test," Journal of Educational Psychology.

L (1959), 135-139.

Horst, P. "A Technique for the Development of a
Differential Prediction Battery," Psychological
Monographs. LXIX (1955).

 

 

Horst, P. "The Differential Prediction of Success in
Various College Course Areas," College and Universipy.
XXXII (1956), 539-546.

 

Howard, Victor, and Warrington, Willard. "The Inventory
of Beliefs: Changes in Beliefs and Attitudes and
Academic Success Prediction," Personnel and Guidance
Journal. XXXVII (1958), 299-302.

 

41.

42.

43.

44.

45.

46.

47

48.
49

50.

146

Hunter, E. C., and Jordan, A. M. ”An Analysis of Qualities
Associated with Leadership Among College Students,"
Journal of Educational Psychology. XXX (1939), 497-509.

Iffert, Robert E. Retention and Withdrawal of College
Students, Bulletin No. l, 1958. Office of Education,
Washington: U.S. Government Printing Office, 1957.

 

Ikenberry, Stanley 0. "A Multivariate Analysis of the
Relationship of Academic Aptitude, Social Background,
Attitudes and Values to Collegiate Persistence."
Unpublished Ph.D. dissertation, Michigan State University,
1960.

Iowa, State University. A Study of Student Persistence
at the State University of Iowa. Iowa City, Office of
the Registrar, 1959.

 

Jensen, Vern H., and Clark, Monroe H. “Married and Un-
married College Students: Achievement, Ability and
Personality," The Personnel and Guidance Journal.

 

XXXVII (1958). 123-126.

JOhnson, G. B. "A Proposed Technique for the Analysis of
Dr0p—Outs at a State College,“ Journal of Educational
Research. XLVII (1954), 381-387.

 

Juola, Arvo E. "Freshman-Level Ability Tests and Long-
Range Prediction," Paper presented at the annual
meeting of the National Council on Measurements Used
in Education, Chicago, 1963.

Juola, Arvo E. "Predictive validity of Five College-
Level Academic Aptitude Tests at One Institution,"
Personnel and Guidance Journal. XXXVIII (1960), 637-641.

Juola, Arvo E. "The Differential Validity of the College
Qualification Tests for Diverse Curricular Groups,"
Personnel and Guidance Journal. XXXIX (1961), 721-725.

Klugh, Henry E., and Bierley, Robert. "The School and
College Ability Test and High School Grades as Predictors
of College Achievement," Educational and Psychological
Measurement. XIX (1959), 625-628.

 

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

147

Knoell, Dorothy. "Institutional Research on Retention and
Withdrawal," Research on College Students. Co-sponsored
by The Western Insterstate Commission for Higher
Education, Boulder, Colorado, and The Center for Higher
Education, Berkeley, California. Hall T. Sprague
(editor), December, 1960. '

 

Lantagne, Joseph E. "College Marriages," The Journal of
College Student Personnel, III (1962), 98-105.

 

 

Lehmann, Irvin J. "The RelationShip Between Scholastic
Performance and Attitudes and Values," 17th Yearbook,
National Council on Measurements Used in Education.
Ames, Iowa: National Council on Measurements Used in
Education, 1960, 83-92.

 

Lehmann, Irvin J., and Ikenberry, Stanley 0. "Critical
Thinking, Attitudes, and Values in Higher Education."
A Preliminary Report of Research, Paul L. Dressel,
Principle Investigator. Michigan State University, 1959.

Lins, L. J. "Probability Approach to Forecasting Uni-
versity Success with Measured Grades as Criterion,"
Educational and Psychological Measurement. X (1950).
386-391.

 

Lins, L. J., and Pitt, Hy. "The 'Staying Power' and
Rate of Progress of University of Wisconsin Freshmen,"
College and Universipy. XXIX (1953), 86-99.

 

Maier, G. E. "The Contribution of Interest Test Scores
to Differential Academic Prediction," Unpublished
Ph.D. dissertation, University of Washington, 1957.

McNeeley, John H. College Student morgality. U.S.
Department of Interior Bulletin, No. 11, 1937.

 

McNemar, Quinn. Psychological Statistics. 2nd edition.
New York: John Wiley & Sons, Inc., 1955.

 

Morgan, H. H. "A Psychometric Comparison of Achieving
and Non-Achieving College Students of High Ability,"
Journal of Consulting Psychology. XVI (1952), 292-298.

 

 

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

148

Mueller, Kate H. "The Married Student On The Campus,"
College and University. XXXV (1960), 155-163.

 

Neel, Ann Filinger. "The Relationship of Authoritarian
Personality to Learning: F-scale Scores Compared to
Classroom Performance," Journal of Educational
Psycholog . L (1959), 195-199.

 

 

Pattishall, E. G., Jr., and Banghart, F. W. "A Comparative
Analysis of School of Education Graduates and‘Withdrawals,"
Educational Research Bulletin. University of Virginia,
April, 1957.

 

Pearlman, S. "An Investigation of the Problem of Academic
Underachievement Among Intellectually Superior College
Students." Unpublished Ph.D. dissertation, New York
University, 1952.

Plant, Walter T., Minium, Edward W., and Myers, Celestine.
"An Analysis of the Rokeach Dogmatism Scale Used with
a Sample of American College Students." Paper read at
the annual meeting of The Western Psychological
Association, San Diego, 1959.

Prince, Richard. "A Study of the Relationship Between
Individual Values and Administrative Effectiveness in
the School Situation." Unpublidhed Ph.D. dissertation,
University of Chicago, 1957.

Red, S. B., McCary, J. L., and Johnson, Bette. "A Study
of the Relationship Between Aspirational Levels and
Academic Achievement," Journal of Educational Research.
LV (1962), 159-163.

 

Rokeach, Milton. "Political and Religious Dogmatism:
An Alternative to Authoritarian Personality,“
Psychological Monographs. LXX (1956), 1-10.

 

Rust, R. M., and Ryan, F. J. "Personality and Academic
Achievement: A Questionnaire Approach." Paper read
at the annual meeting of the American Psychological
Association, San Francisco, 1955.

Samenfink, J. Anthony, and Milliken, Robert L. "Marital
Status and Academic Success: A Reconsideration,"
Marriage and Family Living. XXIII (1961), 226-227.

 

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

149

Seashore, Harold G. "Women Are More Predictable Than Men."
Journal of Counselipg Psychology. IX (1962), 261-264.

 

Sharp, H. C., and Pickett, L. M. "The General Aptitude
Test Battery as a Predictor of College Success,"
Educational and Psychological Measurement. XIX (1959),
617-620.

 

Spindler, George. "Education in a Transforming American
Culture," Harvard Educational Review. XXV (1953), 156-163.

 

Stright, I. L. "Some Factors Affecting College Success,"
Journal of Educational Psychology. XXXVIII (1947),
232-240.

 

Suddarth, Betty M. "Factors Influencing the Graduation
of Freshmen Who Enroll at Purdue University." Unpub-
lished report, June, 1957.

Summerskill, thn. "Dropouts from College," Chapter 19
in The American College (Nevitt Sanford, Editor).
New York: Wiley and Sons, 1962.

 

Thompson, Martha. "Admission Information as Predictors
for Graduation." Unpublished Master's Thesis, Cornell
University, 1953.

Travers, Robert M. "Significant Research on the Prediction
of Academic Success," in The Measurement of Student
Adjustment and Achievement (Wilma Donahue, Editor).

Ann Arbor: University of Michigan Press, 1949.

 

 

Trueblood, Dennis L. "Academic Achievement of Employed
and Non-Employed Students at The Indiana University of
Business," Dissertation Abstracts. XIV (1954),
643-644.

 

Walker, Helen M., and Lev, Joseph. Statistical Inference.
New York: Holt, Rinehart and Winston, 1953.

 

 

81.

82.

83.

84.

85.

86.

87.

150

Weigand, G. "Motivational Factors Associated with Success
and Failure of Probational Students." Unpublished
Ph.D. dissertation, University of Maryland, 1951.

Williamson, E. G., and Hoyt, D. "Measured Personality
Characteristics of Student Leaders," Educational and
Psychological Measurement. XII (1952), 65-78.

 

 

Willingham, Warren W. "College Performance of Fraternity
Members and Independent Students," Personnel and Guid-
ance Journal. XL (1962), 29-34.

Winter, William D. "Student Values and Grades in General
Psychology,“ Journal ofAEducational Research. LV
(1962), 331-334.

 

Winter, William D. "Values and Achievement in a FreShman
PsycholOgy Course," Journal of Educational Research.
LIV (1961), 183-185.

 

Wrenn, C. Gilbert. Student Personnel Work in College.
New York: The Ronald Press Company, 1961.

 

Young, F. Chandler. Scholastic Progress Forecasts.
University of Wisconsin, 1961 (Mimeoqraphed).

 

APPENDIX

151

 

IT

152
TABLE 6.1: Analysis of variance data for high, middle,

and low achievement males on fall, 1958,
Test of Critical Thinking.

HIGH ACHIEVEMENT

 

 

 

 

 

 

Component of variabilipy §§_ 'gf V_ F. 52
Among 22.71 2 11.35 0.25 3.89
Within 10,270.89 224 45.85
Total 10,293.60 226
MIDDLE ACHIEVEMENT
Component of Variability. §§_ HE. E; F ES
Among 69.05 2 34.52 0.89 3.04
Within 8,118.87 210 38.66
Total 8,187.92 212
LOW ACHIEVEMENT
Component of Variability §§_ 8E. V_ ‘F ES
Among 15.86 2 7.93 0.23 3.04
Within 7,323.47 212 34.54

Total 7,339.33 214

153

TABLE 6.2: Analysis of variance data for high, middle,
and low achievement males on fall, 1958,
College Qualification Tests.

HIGH ACHIEVEMENT

 

 

 

 

 

 

Component of variability SS_ SS y_ .E 'SE
Among 1,264.12 2 632.06 1.32 3.89
Within 107,104.12 224 478.15
Total 108,368.90 226
MIDDLE ACHIEVEMENT
Component of Variability .§§ _SS 2 S. So
Among 48.54 2 24.27 0.06 3.04
Within 86,691.75 210 412.82
Total 86,740.29 212
LOW ACHIEVEMENT
Component of Variability SS ‘QS ‘V S ‘SS
Among 2,365.52 2 1,182.76 2.74 3.04
Within 91,545.10 212 431.82

Total 93,910.62 214

 

154

TABLE 6.3: Analysis of variance data for high, middle,
and low achievement males on fall, 1958,
Michigan State University Reading Test.

HIGH ACHIEVEMENT

 

 

 

 

 

 

Component of Variability SS if y_ S £9
Among 62.08 2 31.04 1.44 3.89
Within 4,823.46 224 21.53
Total 4,885.54 226
MIDDLE.ACHIEVEMENT
Component of Variability SS_ {SS Y. .3 So
Among 189.85 2 94.92 3.03 3.04
Within 6,562.72 210 31.25
Total 6,752.57 212
Low ACHIEVEMENT
Component of Variability SS 1 _d_f_ _\_I_ S ES.
Among 138.10 2 69.05 2.50 3.04
Within 5,851.83 212 27.60

Total 5,989.93 214

 

155

TABLE 7.1: Analysis of variance data for high, middle,
and low achievement females on fall, 1958,
Test of Critical Thinking.

HIGH ACHIEVEMENT

 

 

 

 

 

 

Component of Variability SS_ .8: X .E .52
Among 292.18 2 146.09 4.46 3.07
Within 4,089.69 125 32.72
Total 4,381.87 127
MIDDLE ACHIEVEMENT
Component of Variability SS_ “SS V S SS
Among 37.18 2 18.59 0.57 3.07
Within 4,034.57 124 32.54
Total 4,071.75 126
LOW'ACHIEVEMENT
Component of Variability SS. SS ‘2 S SE
Among 61.71 2 30.85 0.92 3.07

Within 4,301.57 128 33.61
Total 4,363.28 130

156

TABLE 7.2: Analysis of variance data for high, middle,
and low achievement females on fall, 1958,
College Qualification Tests.

HIGH ACHIEVEMENT

 

 

Component of Variability SS_ g£_ _! §_ ‘So
Among 649.44 2 324.72 0.66 3.07
Within 61,730.28 125 493.84
Total 62,379.72 127

MIDDLE ACHIEVEMENT

 

 

 

 

Component of Variability SS_ S: V; S ‘So
Among 13.51 2 6.75 0.02 3.07
Within 44,919.70 124 362.26
Total 44,933.21 126

LOW ACHIEVEMENT

Component of Variability SS_ SE_ 2' E. .E2
Among 950.65 2 475.32 1.39 3.07
Within 43,810.88 128 342.27

Total 44,761.53 130

 

157

TABLE 7.3: Analysis of variance data for high, middle,
and low achievement females on fall, 1958,
Michigan State University Reading Test.

HIGH ACHIEVEMENT

 

 

 

 

 

 

Component of Variability SS g; ‘2' S .SE
Among 69.97 2 34.98 2.00 3.07
Within 2,183.58 125 17.47
Total 2,253.55 127
MIDDLE ACHIEVEMENT
Component of Variability SS. SS ‘1 §_ .52
Among 103.12 2 51.56 2.46 3.07
Within 2,596.96 124 20.94
Total 2,700.08 126
LOW ACHIEVEMENT
Component of variability SS_ SS .2 ,E SE
Among 46.15 2 23.07 1.07 3.07
Within ‘ 2,753.90 128 21.51

Total 2,800.05 130

158

TABLE 8.1: Analysis of variance data for high, middle,
and low achievement males on fall, 1958,
Inventory of Beliefs.

HIGH AcngVEMENT

 

 

Component of variability SS_ 1g: 2 ‘S ES
Among 647.09 2 323.54 1.82 3.89
Within 39,882.43 224 178.05
Total 40,529.52 226

MIDDLE ACHIEVEMENT

 

Component of Variability SS_ SS .2 ‘S ‘SE
Among 317.14 2 158.57 0.81 3.04
Within 41,250.59 210 196.43
Total 41,567.73 212

LOW ACHIEVEMENT

 

Component_9f variability_ SS_ ‘SS .2 'S SE
Among 337.63 2 168.81 0.83 3.04
Within 43,341.55 212 204.44

Total 43,679.18 214

 

159

TABLE 8.2: Analysis of variance data for high, middle,
and low achievement males on fall, 1958,
Differential Values Inventory.

HIGH ACHIEVEMENT

 

 

Component of variability. SS_ ‘SS y_ 'S SE
Among 201.29 2 100.64 1.99 3.89
Within 11,314.43 224 50.51
Total 11,515.72 226

MIDDLE ACHIEVEMENT_

 

 

 

 

Component of Variability SS 9}: X S So
Among 30.34 2 15.17 0.35 3.04
Within 9,226.94 210 43.94
Total 9,257.28 212

LOW ACHIEVEMENS

Component of variability» SS .85 y: ‘S Fc
Among 113.39 2 56.69 1.29 3.04
Within 9,345.87 212 44.08

Total 9,459.26 214

 

160

TABLE 8.3: Analysis of variance data for high, middle,
and low achievement males on fall, 1958,
Dogmatism Scale.

HIGH ACHIEVEMENT

 

 

 

 

 

Component of Variability SS_ SS .2 S .ES
Among 1,609.50 2 804.75 1.37 3.89
Within 131,609.50 224 585.78
Total 132,825.07 226
MIDDLE ACHIEVEMENT
Component of Variability SS. SS .2 .5 SS
Among 1,023.98 2 511.99 0.80 3.04
Within 135,119.01 210 643.42
Total 136,142.99 212
LOW ACHIEVEMENT
Component of variability SS_ ‘SS 1!» S SS
Among 796.73 2 398.36 0.60 3.04
Within 139,635.20 212 658.66

Total 140,431.93 214

161

TABLE 9.1: Analysis of variance data for high, middle,

and low achievement males on spring,

Inventory of Beliefs.

HIGH ACHIEVEMENT

 

 

 

 

 

 

Component of Variability
Among 451.
Within 150.282.
Total . 150.734.
MIDDLE ACHIEVEMENT
Component of Variability
Among 3.802.
Within 100,555.
Total 104,358.
LOW ACHIEVEMENT
Component of Variability
Among 2,181.
Within 98.078.

Total 100,259.

99.

28
80
08

9.8

71
58
29

.85

31
67
98

2
212
214

31

225.64
670.91

1’.

1901.35
478.84

1’.

L090.65
462.64

IW

0.34

IW

2.36

1959.

3.89

3.04

162

TABLE 9.2: Analysis of variance data for high, middle,
and low achievement males on spring, 1959,
Differential Values Inventory.

HIGH ACHIEVEMENT

 

 

 

 

 

 

Component of Variability SS. SS. .2 S SS
Among 128.79 2 64.39 0.39 3.89
Within 37,044.76 224 165.38
Total 37,173.55 226
MIDDLE ACHIEVEMENT
Component of Variability SS_ ‘SS 1! S SS
Among 538.11 2 269205 2.22 3.04
Within 25,401.55 210 120.96
Total 25,939.66 212
LOW ACHIEVEMENT
Component of Variability SS ‘SS S S. SS
Among 97.88 2 48.94 0.37 3.04
Within 27,714.75 212 130.73

Total 27,812.63 214

 

163

TABLE 10.1: Analysis of variance data for high, middle,
and low achievement females on fall, 1958,
Inventory of Beliefs.

HIGH ACHIEVEMENT

 

 

 

 

 

 

Component of Variability SS_ SS_ .2 S .52
Among 43.60 2 21.80 0.16 3.07
Within 17,283.64 125 138.27
Total 17,327.24 127
MIDDLE ACHIEVEMENT
Component of Variability SS_ ‘SS .2 S SS
Among 190.68 2 95.34 0.85 3.07
Within 14,982.58 124 112.53
Total 15,173.26 126
LOW ACHIEVEMENT
Component of Variability SS_ .8: ‘S S SS
Among 15.49 2 7.74 0.04 3.07
Within 23,573.41 125 184.17

Total 23,588.90 127

TABLE 10.2:

164

Differential Values Inventory.

HIGH ACHIEVEMENT

 

 

Component of Variability
Among 69.
Within 6,352.
Total 6,421.

MIDDLE ACHIEVEMENT

 

 

Component of Variability
Among 23.
Within 5,661.
Total 5,685.

LOW ACHIEVEMENT

 

 

Component of Variability
Among 133.
Within 6.484.
Total 6,618.

9.9.

47
41
88

99.

51
67
18

.89

43
62
05

l<

34.73
50.82

|<

11.75
45.66

|<

66.71
50.66

Analysis of variance data for high, middle,
and low achievement females on fall,

1958.

2:: Es
0.68 3.07
11 31c.
0.26 3.07
S_ Fc
1.32 3.07

165

TABLE 10.3: Analysis of variance data for high, middle,
and low achievement females on fall, 1958,
Dogmatism Scale.

HIGH ACHIEVEMENT

 

 

Component of Variability SS_ .SS ‘2 S_ .32
Among 243.96 2 121.98 0.26 3.07
Within 59,001.34 125 472.01
Total 59,245.30 127

MIDDLE ACHIEVEMENT

 

 

 

 

Component of Variability SS_ SS S .E. (ES
Among 1,279.03 2 639.51 1.06 3.07
Within 74,950.84 124 604.44
Total 75,229.87 126

LOW ACHIEVEMENT

Component of Variability SS ‘SS ‘2’ S SS
Among 479.73 2 239.86 0.36 3.07
Within 84,732.80 128 661.97

Total ~85,212.53 130

 

166

TABLE 11.1: Analysis of variance data for high, middle,
and low achievement females on spring. 1959.
Inventory of Beliefs.

HIGH ACHIEVEMSNT

 

 

 

 

 

 

Component of Variability SS .SS .2 S. SS
Among 838.70 2 419.35 0.99 3.07
Within 52,847.52 125 422.78
Total 53,686.22 127
MIDDLE ACHIEVEMENT
Component of Variability SS_ ‘SS ‘S S SS
Among 1,239.43 2 619.71 1.74 3.07
Within 44,089.63 124 355.56
Total 45,329.06 126
LOW ACHIEVEMENT
Component of Variability SS. ‘SS 2; S SS
Among 1,218.96 2 609.50 1.17 3.07
Within 66,484.58 128 519.41

Total 67,703.54 130

 

167

TABLE 11.2: Analysis of variance data for high, middle,
and low achievement females on spring, 1959.
Differential Values Inventory.

HIGH ACHIEVEMENT

 

 

 

 

 

 

Component of Variability SS_ 'SS ‘2 ‘S SS
Among 163.20 2 81.60 0.96 3.07
Within 10,591.42 125 84.73
Total 10,754.62 127
MIDDLE ACHIEVEMENT
Component of Variability ‘SS 'SS 2: S_ SS
' Among 435.27 2 217.63 2.35 3.07
Within 11,492.60 124 92.68
Total 11,927.87 126
LOW ACHIEVEMENT
Component of Variability SS_ SS_ 2. ‘S SS
Among 202.63 2 101.31 0.73 3.07
Within 17,766.12 128 138.80

Total 17,968.75 130

P’ 'H' O : C' r ‘
J'Sf‘ .’ 394“," I .5" 1‘.
2 3 ‘5 (1' a r ' 1' f '
' - L H. , ‘
l\ d “j“,

 

 

 

 

 

‘ 9

. ‘ J
.r.. .

II. t w.

I“. ...‘ m .

. 1‘.

 

ERSI

WTITIIIWINWTIWITUTW WI! W "111111111111“
3 1293 0301] 9743