A MULTWARIAYE DKSCRIMQNANT ANALYSES
OF THE RELATEGN$HEP fiETWEEN SELECTEB
CHARAfl'FERiSTECS 0F {ENEMNG
QUEEN??? AME 1.1-!!le HEAL
UNDERGRADUATE MAJORS

Thanks for fire chreo oi“ EJ. D.
MlCHIGAN S‘EATE [INNERSWY
Betty 3... Giuliani
3958

Tau»:

WI!WIN!JUIIHHUHHHMWW I .2? .1

This is to certify that the

thesis entitled

9" “ ..

A MULTIVARIATE DISCRIMINANT ANALYSIS OF THE
RELATIONSHIP BETWEEN SELECTED CHARACTERISTICS
' OF ENTERING STUDENTS AND THEIR FINAL
UNDERGRADUATE MAJORS

presented by

L.

Betty L. Giuliani

has been accepted towards fulfillment
of the requirements for

. fl." .‘. . ‘ ‘ . ,.
Ed . D . degigg in Ed uga I; i Qn

) ,
@7571: i2?
/ ~‘

3...
/l~/ajor professor i

Date September 6, 1968

0-169

. . “euro-o.“ -Q a,,

 

‘PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINE return on or before date due.

 

I DATE DUE

DATE DUE

DATE DUE

 

7 ,1“; If .
altfidl’ Lg

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1U" '

 

 

 

1198 MM“

 

ABSTRACT
A MULTIVARIATE DISCRIMINANT ANALYSIS OF THE
RELATIONSHIP BETWEEN SELECTED CHARACTERISTICS

OF ENTERING STUDENTS AND THEIR FINAL
UNDERGRADUATE MAJORS

BY

."L

Betty quGiuliani

Purpose

The purpose of this study was to investigate rela-
tionships which exist between students' characteristics, as
measured by orientation data, and their final undergraduate
majors. One major hypothesis was proposed: Undergraduate
major fields can be differentiated on the basis of selected
characteristics of entering freshman students who attain

senior standing or graduate in each major.

Methodology

Forty-seven measures of ability/achievement, family
and background eXperiences, self-ratings and interests, and
educational/vocational preferences were selected as relevant
student variables. The student sample was drawn from the
first-time freshmen who entered Michigan State University
during either Fall 1963 or Fall 1964 and had graduated or
attained senior standing by Spring 1967. The major sample

consisted of the undergraduate major fields represented in

Betty L. Giuliani

the student sample which contained at least 40 students of
one sex. A total of 909 men and 856 women in 27 groups rep-
resenting 23 different major fields made up the final study
sample from which a 20 percent random sample was drawn for
cross-validation purposes.

The data were subjected to four treatments: (1) a
discriminant analysis of the 27 major groups, (2) a separate
discriminant analysis of the 14 male groups and the 13
female groups, (3) classification of a cross—validation
sample based on the 27-group analysis, and (4) classifica-
tion of male and female cross-validation samples based on

the 14- and lB-group analyses, respectively.

Results

The 27-group discriminant analysis yielded 26 func-
tions, of which 10 were significant at the .001 level, a
rejection of the null hypothesis. Function one accounted
for 51 percent of the dispersion among groups and split the
majors into a verbal/humanistic group and a numerical/phys-
ical science/technical group. The verbal/humanistic group
contained only female majors; the numerical/physical science/
technical group contained all the male majors plus the mathe-
matics women. Function two, interpreted as generalist to
specialist in nature, accounted for 11 percent of the trace.
The third function separated the major groups along an
abstract to applied continuum and accounted for 10 percent

of the diSpersion. The fourth and fifth functions, each

Betty L. Giuliani

accounting for 4 percent of the trace, were interpreted but
not labeled because the heterogeneous combination of high
weight variables in each function did not lend themselves to
a common interpretation. The remaining five significant
functions were not interpreted.

The separate discriminant analyses of 14 male and
13 female majors each produced five significant functions
which accounted for 80 and 76 percent of the dispersion,
respectively, and which were interpreted in detail.

Three cluster analyses were performed using inter—
centroid distances and a computational scheme suggested by
Rao (1952). The 27-group case produced 6 multi-group clus—
ters and two isolated majors. The results of the cluster
formation supported the secondary hypothesis that men and
women in the same major have different patterns of character-
istics. For each analysis, clusters were compared on the
first three significant functions by means of a three-dimen-
sional plot of group centroids.

.A maximum likelihood procedure based on function
weights and discriminant scores was used to classify three
20 percent cross-validation samples into majors and clusters.
Assignments were scored as hits (classification into correct
major), near hits (classification into correct cluster), and
misses. Results were significantly better than chance.

Thus, it would seem that, as Holland (1966a) sug-

gests, there is a tendency for students to seek environments

Betty L. Giuliani

which permit them to exercise their skills and abilities,
to take on agreeable problems and roles, and to avoid dis-
agreeable ones. Results of this investigation demonstrated
that it is possible to differentiate among major field
environments on the basis of the dominant characteristics
of members of each environment. Further, it appears that

these characteristics are fairly stable over time.

A MULTIVARIATE DISCRIMINANT ANALYSIS OF THE
RELATIONSHIP BETWEEN SELECTED CHARACTERISTICS
OF ENTERING STUDENTS AND THEIR FINAL

UNDERGRADUATE MAJORS

1.3Y
( fl:

Betty L? Giuliani

A THESIS
Submitted to
Michigan State University

in partial fulfillment of the requirements
for the degree of

DOCTOR OF EDUCATION

Department of Counseling, Personnel Services
and Educational Psychology

1968

ACKNOWLEDGMENTS

The author is grateful to many people who, during

the course of the present study, offered and gave their help,

but eSpecially to the following:

to Dr. Buford Stefflre who, through four long years
as doctoral committee chairman and dissertation
director, has provided exactly the right amount of
encouragement, challenge, and understanding and is,
indeed, the compleat counselor;

to Dr. Joseph L. Saupe who, as co-director of the
dissertation, was more than generous with his time
and talent in the resolution of statistical and
analytical problems encountered during the course
of the study;

to Dr. Paul L. Dressel, whose confidence, patience,
and support made it possible to undertake this in-
vestigation;

to Dr. William W. Farquhar and Dr. Norman Abeles,
for their continuing interest and assistance as
members of the doctoral committee;

,to Dr. Margaret F. Lorimer, whose insightful obser-

vations and incisive comments were of great help
during the writing of the thesis;

to Mr. Lynn H. Peltier, for his invaluable and un-
stinting assistance with the processing of the data
for this study; and

to Mr. Stuart Thomas, Jr., without whom there would
have been no computer programs for the project.

ii

TABLE OF CONTENTS

Chapter
I. THE PROBLEM . . . . . .

Need . . . . . . . .
Purpose . . . . . . .
Rationale . . . . . .
Research Hypotheses .
Overview . . . . . .

II. REVIEW OF LITERATURE .

The Descriptive Studies . . . . . . . . .

The Predictive Studies

Summary . . . . . . .
III. DESIGN AND METHODOLOGY

The Sample . . . . .

Selection of Students and Final

Major Fields .
Instrumentation . . .

Aptitude/Achievement Measures . . .
Measures of Interest, Family
Background, and Educational
and Vocational Preference . . . . .
Summary of Instrumentation and

Variables Used
The Statistical Model
Summary . . . . . . .

IV. ANALYSIS OF DATA . . .

Discriminant Analysis
Criterion Groups .
Interpretation of
Functions . . .
Discriminant Analysis
Criterion Groups .
Discriminant Analysis
Criterion Groups .

iii

of Twenty-Seven
the Discriminant
of Fourteen Male

of Thirteen Female

Page

OGMWH H

on

43
43
43
44
46
47
49
51
54

55

55
61
74

85

Chapter

V.

VI.

Classification of the Cross—Validation

Samples 0 O O O O O O O O O O O O O O 0
Results of th Twenty-Seven Group
Classifications . . . . . . . . . .

Results of the Fourteen-Group
Classification . . . . . . . . . .

Results of the Thirteen-Group
Analysis . . . . . . . . . . . . .
Summary . . . . . . . . . . . . . . . . .

SUMMARY DESCRIPTIONS OF THE TWENTY-SEVEN
MAJOR GROUPS . . . . . . . . . . . . . .

The Majors . . . . . . . . . . . . . . .

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

Summary . . . . . . . . . . . . . . . . .
Purpose . . . . . . . . . . . . . . .
Methodology . . . . . . . . . . . . .
Results . . . . . . . . . . . . . . .

Conclusions . . . . . . . . . . . . . . .
Implications . . . . . . . . . . . . . .
Counseling and Academic Advisement .
Administrative Decision-Making . . .
Needed Research . . . . . . . . . . .

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . .

APPENDICES . . . . . . . . . . . . . . . . . . . .

A.

B.

Variables From the Personal Information
Inventory . . . . . . . . . . . . . . . .

Original and Final Frequency Distributions
of Rescaled Variables . . . . . . . . . .

Computation of Intra- and Inter-Cluster
Distances for Twenty-Seven Group Case . .

Computation of Intra- and Inter-Cluster
Distances for Fourteen Group Case . . . .

Computation of Intra- and Inter-Cluster
Distances for Thirteen Group Case . . . .

Group Means and Standard Deviations for

Twenty-Seven Groups and for All Groups
on Forty-Seven Variables . . . . . . . .

iv

Page

98

99
101
102
104
109
110
138
138
138
138
139
143
147
147
148
148
150

155

155

159

164

165

167

169

Table

1.

10.

11.

LIST OF TABLES

College and Major Groups Included in Study:
By Sex and Size of Study Sample, Analysis
Sample and Check Sample . . . . . . . . . .

Variables Included in the Discriminant
Analysis . . . . . . . . . . . . . . . . . .

Percent of Trace, Degrees of Freedom, and
Statistical Significance Levels for First
Ten Discriminant Functions of the Twenty-
Seven Group Analysis . . . . . . . . . . . .

Formation of Clusters Using Intercentroid
Distances From Twenty—Seven Group Analysis .

Inter-Cluster Relationships for Twenty-Seven
Group Analysis . . . . . . . . . . . . . . .

Variable Included in Tables 7 Through 25 . .

Standardized Weights, Group Centroids, and
Plot of Group Centroids for Twenty-Seven
Major Groups on Function One . . . . . . . .

Standardized Weights, Group Centroids, and
Plot of Group Centroids for Twenty-Seven
Major Groups on Function Two . . . . . . . .

Standardized Weights, Group Centroids, and
Plot of Group Centroids for Twenty-Seven
Major Groups on Function Three . . . . . . .

Standardized Weights, Group Centroids, and
Plot of Group Centroids for Twenty-Seven
Major Groups on Function Four . . . . . . .

Standardized Weights, Group Centroids, and
Plot of Group Centroids for Twenty-Seven
Major Groups on Function Five . . . . . . .

Page

45

50

57

59

60

62

63

65

67

69

71

Table Page

12. Percent of Trace, Degreesof Freedom, and
Statistical Significance Levels for First
Five Discriminant Functions of the
Fourteen-Group Analysis . . . . . . . . . . . 74

13. Standardized Weights, Group Centroids, and
Plot of Group Centroids for Fourteen Groups
on Function One . . . . . . . . . . . . . . . 76

14. Standardized Weights, Group Centroids, and
Plot of Group Centroids for Fourteen Groups
on Function Two . . . . . . . . . . . . . . . 77

15. Standardized Weights, Group Centroids, and
Plot of Group Centroids for Fourteen Groups
on Function Three . . . . . . . . . . . . . . 79

16. Standardized Weights, Group Centroids, and
Plot of Group Centroids for Fourteen Groups
on Function Four . . . . . . . . . . . . . . . 81

17. Standardized Weights, Group Centroids, and
Plot of Group Centroids for Fourteen Groups
on Function . . . . . . . . . . . . . . . . . 82

18. Formation of Clusters Using Intercentroid
Distances From Fourteen-Group Analysis . . . . 84

19. Inter-Cluster Relationships for Fourteen-
Group Analysis . . . . . . . . . . . . . . . . 84

20. Percent of Trace, Degrees of Freedom, and
Statistical Significance Levels for First
Five Discriminant Functions of the Thirteen-
Group Analysis . . . . . . . . . . . . . . . . 85

21. Standardized Weights, Group Centroids, and
Plot of Group Centroids for Thirteen Groups
on Function One . . . . . . . . . . . . . . . 87

22. Standardized Weights, Group Centroids, and
Plot of Group Centroids for Thirteen Groups
on Function Two . . . . . . . . . . . . . . . 88

23. Standardized Weights, Group Centroids, and

Plot of Group Centroids for Thirteen Groups
on Function Three . . . . . . . . . . . . . . 90

vi

Table

24.

25.

26.

27.

28.

29.

30.

Standardized Weights, Group Centroids,

on Function Four . .

Standardized Weights, Group Centroids,

on Function Five . .

Formation of Clusters Using Intercentroid

and
Plot of Group Centroids for Thirteen Groups

and
Plot of Group Centroids for Thirteen Groups

Distances From Thirteen-Group Analysis .

Inter-Cluster Relationships for Thirteen-

Group Analysis . . .

Near Hits,

Hits,

Classification of Cross-Validation Sample,
for Thirteen-Group Case Showing Hits,

Near Hits,

and Misses

and Misses

Vii

LClassification of Cross-Validation Sample
for Twenty-Seven Group Case Showing Hits,

.Classification of Cross—Validation Sample
for Fourteen—Group Case Showing Hits, Near
and Misses . .

Page

91

92

94

94

101

103

105

Figure

LIST OF FIGURES

Three Dimensional Plot of Group Centroids
on First Three Significant Functions of
Twenty-Seven Group Analysis . . . . . . .

Three Dimensional Plot of Group Centroids
on First Three Significant Functions of

,Fourteen Group Analysis . . . . . . . . .

Three Dimensional Plot of Group Centroids
on First Three Significant Functions of
Thirteen Group Analysis . . . . . . . . .

viii

Page
. . 73
. . 96
. . 97

CHAPTER I

THE PROBLEM

Need

The Michigan State University Catalog lists over 150
programs of study (majors), available to the entering under-
graduate student, administered by more than 80 departments
in 13 colleges. Individual undergraduate courses offered
during a term number in the hundreds and, in one academic
year, well over 2,000 different courses are presented to
students. Michigan State University is not atypical. A
study (Saupe, 1966) of courses listed and taught at nine of
the eleven universities which make up the Institutional
Research Council of Elevenl gave the median number of under-
graduate courses listed for a year as 2,108.

Michigan State University, recognizing the choice
problems presented to entering freshmen by so vast a selec—
tion of courses and majors as indicated above, provides a
liberal policy governing selection and change of major and

makes available academic advisement and counseling services

 

lThe IRCE is composed of universities in the Big Ten

Conference and The University of Chicago.

to assist students in making intelligent, appropriate, and
satisfying decisions. Beginning freshman students attend a
3-day summer orientation and counseling session (called a
"clinic"), during which they take academic aptitude and
placement tests. Personal information is collected on a
Counseling Center form, the Personal Information Inventory
(PII), and through individual interviews. The PII is made
available to clinic counselors, but not to academic advisers.
Test results are stated as raw scores and percentiles and
are used by clinic counselors, and later by academic advis—
ers, to assist students in selecting a major field and plan-
ning an educational program.
Recent research results (Abe & Holland, 1965; Davis,
1965; Werts, 1966; among others) indicate that the factors
influencing the career choices and selection of major field
by college undergraduates are related to a need for what
Davis calls ”social homogeneity"--a tendency for “birds of a
feather to flock together." According to Abe & Holland:
When a student's characteristics resemble those

of the typical student in his field, he is likely

to feel at home and remain in his field. Conversely,

,incongruencies between a student and his field

result in feelings of alienation and dissatisfaction

and usually lead to a change of plans (p. 1).
There is further the implication that, in addition to the
pull a student feels toward a field which contains others
who are most like himself, particular fields of study and

work attract certain kinds of people. This suggests that

orientation data, in its most useful form, would permit

comparisons of a student's pattern of abilities, interests,
and past experiences with the patterns of successful stu-
dents in a variety of majors within the University. But the
information collected during summer orientation clinics, in
its present format, sheds little light on themeaning of
patterns of student characteristics and provides no insight
into the relationship between student characteristics and

academic majors.

Purpose

The purpose of this study is toyinvestigate relation-
ships which exist between students' characteristics, as mea-
sured by orientation data, and their fipal undergraduate
majors! by identifying student variables which differentiate
among major fields! and quantifying the differentiating
relationships. From the results of the study, it should be
possible to describe a variety of majors in terms of the
kinds of students attracted to them, and to determine the
probability of a student's completing his undergraduate

study in a particular major.
Rationale

Guidelines for the selection or rejection of spe-
cific measures were sought from the body of theory and
research dealing with vocational choice and development.

Werts (1966) identified the problem of choosing the "right"

variables in his introductory remarks to a National Merit
Scholarship Corporation Research Report on the relationship
of ability and social class to career patterns, where he
said:

It is easy to find statistically significant
differences between persons sorted by occupation
(or occupational aspiration), almost regardless
of the type of sociological or psychological vari-
ables beind studied. . . . Briefly, the problem
confronting career research today is one of rele-
vance, of how to separate the theoretically mean-
ingful from the many sources of artifact (p. 1).

John Holland's theoretical formulations appear to be
most applicable to the purposes of this study. Holland's
(1966b) theory of vocational choice rests on four assump-
tions.

1. Most persons in our culture can be categorized
as one of six types, a type being defined as a
complex cluster of personal attributes. Parents,
social class, schools, and community play an
important role in the development of the person-
ality type.

2. Where people congregate, they create an environ-
ment that reflects the types they are. Thus, there
are six kinds of environments, each dominated by a
given type of personality.

3. People search for environments and vocations

that will permit them to exercise their skills and
abilities, to express their attitudes and values,

to take on agreeable problems and roles, and to

avoid disagreeable ones. Vocational decisions depend
on a great range of student characteristics including
interests, values, self-conceptions, competencies,
achievements, range of eXperience, and family resources.
To a lesser degree, environments also search for people
through recruiting practices.

4. .A person's behavior can be explained by the
interaction of his personality pattern and his
environment.

With regard to the fourth assumption, Holland says:

In the present theory, a person's first and
subsequent decisions are eXplained in terms of
personality pattern and environmental model only.

A more complete theory would incorporate economic
and sociological influence (p. 12).

Although Holland stresses the importance of person—
ality and interests as determinants of vocational choice, he
does not ignore aptitude and intelligence. These latter
characteristics are assumed to be differentially distributed
among the six personality types and environments.

Four general categories of relevant variables seem
to emerge from Holland's theoretical position. They are:

(l) aptitude/achievement, (2) self-ratings and interests,

(3) family and background experiences, and (4) educational
and vocational preferences. Research by Tiedeman and his
associates (Dunn, 1959; King, 1958; Tatsuoka, 1957; Tiedeman
& Bryan, 1954) on choice of undergraduate major utilized
variables which fell into the same four categories. The
selection of specific measures for these studies appeared to
be determined by what measures were available for the popula-
tion to be studied, in contrast to Holland's research where,
apparently, theory development and refinement of instrumenta-
tion have proceeded together.

At Michigan State University, information in each of
the four categories is collected from students during orien-

tation, but its usefulness is limited first, because of the

treatment of the data and second, because even in its current

form, the information is not made easily accessible to
counselors and advisers. A consequence of the current study
will be to re-organize the available data into a more mean—
ingful form for use by students and their advisers in the

selection of courses and the choice of a major field.

Research Hypotheses

The major hypothesis of this study is that under-
graduate major fields can be differentiated on the basis of
the multivariate discriminant analysis of aptitude, interest,
educational-vocational preference, and family background mea-
sures of students who attain senior standing or have gradu-
ated in each major. .A secondary hypothesis of the study is
that male and female students in the same major will have
substantially different patterns of characteristics.

To test the validity of the discriminant analysis
results, group centroids for each major and for each sex
will be used to predict membership in a major field for a

cross-validation sample of seniors and graduates.

Overview

 

The value of investigating factors which influence
undergraduate students in their selection of final major
fields was discussed above, and some of the theoretical and
research bases for such a study were delineated. In Chapter

II, previous research relevant to the problem is critiqued.

Chapter III sets forth the design and methodology employed
in the investigation. Results of the analysis are included
in Chapter IV. The fifth chapter contains a descriptive
summary of each of the majors included in the study. In
Chapter VI, the materials presented in Chapters I through
IV, the conclusions warranted by the findings, and some

suggestions for related research are summarized.

CHAPTER II

REVIEW OF LITERATURE

The review of literature is limited to studies of
the relationship between student characteristics and under-
graduate college majors. Results of descriptive and predic-
tive investigations of this topic are critically examined,

and pertinent recommendations reported.
The Descriptive Studies

The first group of studies reported in this section
deals in a general way with the identification of student
characteristics unique to Specific curricular groups or sub-
groups. The second set of studies concerns the relationship
of student characteristics to change or persistence in major
field.

The Research and Development Division of the American
College Testing Program has published several descriptive
studies of college-bound and undergraduate students, of
which three are concerned with the relationship between stu4
dent characteristics and undergraduate major fields.

Abe & Holland (1965) collected information on 117

student variables from 12,432 college freshmen at 31

institutions of higher education. The data-collection
instrument contained 1,004 items covering student interests,
attitudes, potential for various kinds of achievement, and
other orientations. Students were grouped by sex and antic-
ipated major field of study, which resulted in 79 male
groups and 60 female groups. Each major field was described
on the basis of its extreme characteristics, fields were
grouped into 13 conventional academic areas, and the charac—
teristics most descriptive of the major fields in each area
were summarized. Four categories of characteristics emerged
as differentiators: (1) occupational group preferences, (2)
self-ratings of competencies, (3) self-ratings of traits and
abilities, and (4) selection of important life goals and
achievements. .Abe & Holland concluded that measures of
student interests and life goals are useful discriminators
of major fields, but warned against a too-specific interpre-
tation of results because of certain limitations imposed by
the design of the study. They pointed out that: (1) their
subjects were grouped on the basis of anticipated major
field rather than actual major field, (2) the number of
subjects in a major ranged from 10 to 1,353 making some
characterizations more reliable than others, and (3) the use
of extreme characteristics may have overly-accentuated exist—
ing differences. In addition, major fields were grouped a
priori into 13 "conventional" academic areas and each area

was then treated as if it were composed of homogeneous majors.

10

Baird (1967, No. 17; 1967, No. 19) reported on two
ACT studies generated from data collected on the Student
Profiie Section of the ACT test battery administered to high
school students during 1964-65. He selected a 3 percent
representative sample which totaled 10,073 boys and 8,305
girls and, in the first study, grouped them by student-
reported family income into nine categories. Income groups
were compared in several areas including probable choice of
major. Baird found that students from higher income families
tended to move toward fields which he classified as adminis-
trative, political, or persuasive and away from social,
religious, or educational areas, while lower income students
were more likely to select the latter areas. In the second
study, Baird (1967, No. 19) used differing educational goals
to form student groups which he again compared in a variety
of areas including probable major choice. He concluded that
stated educational goals (reasons for coming to college) are
closely related to educational plans and that, in general,
different major fields tend to attract students with differ-
ent educational goals. Baird's design suffered some of the
same limitations described in the Abe & Holland study; that
is, students were grouped on the basis of anticipated major
field and major fields were classified a priori into sup-
posedly homogeneous categories. He used student-reported

family income as a classification variable in the first

11

study without any reported evidence that the student-reported
family income resembled the actual family income.

Smith (1967) compared 76 junior physical education
majors (female) with 70 junior letters and science majors
(female) on measures of ability, family background, and self-
identification by sex, family role, and career choice. She
found significant differences between the two majors on the
following variables: (1) size of community and high school,
(2) recreational pursuits of the father, (3) quality of rela-
tionship with father, (4) factors influencing choice of
major, and (5) leadership and interest in extra—curricular
activities.

Veldman, Peck, & Richek (1968) demonstrated a rela-
tionship among certain aSpects of high school experience,
personality measures, and the student teaching performance
of a group of 192 female students who graduated from the
University of Texas. The subjects were assigned to eight
criterion groups defined by a high school characteristic,
i.e., grade point average, class size, favorite high school
subject, and the like. A varied set of measures were avail—
able on each subject: (1) nine variables of attitude toward
self and others, (2) eight measures of personality, (3) six
factor scores from a pupil observation instrument, (4) four
factor scores from the Qaiifornia Psychological Inventory,

(5) college grade point average during the senior year, and

(6) an evaluation of student teaching effectiveness.

12

Results of the eight analyses of variance led Veldman _E ii.
to conclude that measures of high school eXperience warrant
further use and study as relevant variables in understanding
college students' choice of field and career.

Korn (1962) compared a group of 1959 freshmen men
majoring in the physical sciences (chemistry, mathematics,
and physics), with engineering majors and general studies
majors using their scores on the California Psychological
Inventory, the Strong Vocatignal interest Blank, and a bio—
graphical data sheet. Chi-square analyses of SVIB patterns
comparing engineers with physical science majors, general
studies with physical science majors, and general studies
with engineering majors were carried out. Engineering and
physical science majors shared an intense interest in the
physical sciences, which differentiated them from the gen—
eral studies group. In order to discriminate between
engineering majors and physical science majors, Korn iden—
tified strong interest areas which were not shared and found
a practical—theoretical polarization. Two of the 9g; scales,
Femininity and ReSponsibility, also distinguished among the
three groups of men. A replication of the study using
freshmen men who entered the same majors in 1960 produced
substantially the same results.

Korn concluded that differences between similar
majors can be better understood by considering more than one
basis for classification of student characteristics. He

emphasized the value of describing major groups in terms of

13

their members' rejection of some interests as well as their
attraction to others.

Each of the above studies utilized a large number
and variety of measures to describe the students in the
various majors and, thus, differentiate among the fields,
but in all cases, the most useful variables for this purpose
were nonintellective in nature. The following two studies
demonstrate the efficacy of ability-achievement measures in
accomplishing the purpose.

Voorhies (1966) selected a sample of undergraduate
men majoring in business administration, industrial arts,
and engineering to demonstrate a multivariate treatment of
entrance data which would be of value in the academic advise-
ment of male students at Middle Tennessee State College.
Entrance test scores and high school grade point average
were run through multiple regression analysis and discrim-
inant function analysis. Voorhies computed a separate multi-
ple R and discriminant function for each curricular group
but, apparently, did not run a cross-validation study on
the results. He recommended the inclusion of nonintellec-
tive measures in future studies of this type. Trends in the
data also suggested to him that variables with negative
discriminant weights and positive regression coefficients
might be used to identify avoidance patterns for major
fields. That is, students who scored high on such variables

appeared more likely to change major field than those who

14

had low scores on them, and changes seemed to be toward
fields more compatible with student abilities or interests.

Cullum (1966) investigated the relationships between
scores on the General Aptitude Test Battery, grade point
average, and six undergraduate major fields. He was able to
differentiate management, English, and physical education
majors on aptitude scores, and found that students who
expressed dissatisfaction with their major fields had apti-
tude scores which differed from the norms in their fields.
The latter finding tends to support the Voorhies notion that
avoidance patterns are discernible in such data, if research-
ers are alert to their possibility.

Interest measures alone have also been used success-
fully in discriminating among broadly defined curricular
groups. Baggaley (1947) divided 185 Harvard students into
two groups on the basis of their major fields at the end of
the freshman year; Group A was a cluster of natural science
majors and Group B was a cluster of social science and
humanities majors. Using Kuder Preference scores, he com—
puted a mean discriminant score for each group and got sig-
nificant results using a Ertest of mean differences. He
also computed individual discriminant scores for visual
inspection of group homogeneity. Baggaley suggested that
students with extreme scores be singled out for additional

educational-vocational guidance.

15

Matteson (1961) believed that student interests, as
measured on the Activity Check Li§5_(ACL), would differenti-
ate among curriculum groups. Prior to their enrollment as
freshmen at Michigan State University, 185 freshman men and
115 freshman women were given the A99; 262 students indicated
a specific choice of major and 38 had not yet selected a
major. Students were grouped by college membership. An
interest area was designated as characteristic of a college
group if more than half the group had high or high-average
scores in the area. Interest areas were distributed among
the various colleges about as Matteson exPected (i.e.,
Engineering = mechanical, computational, scientific; Busi-
ness and Public Service = clerical and computational), and
were combined differently for each college. He was also
able to differentiate among colleges on the basis of the
strength and diversity of student interests.

Although any student-linked variable could be
described as a "personality" measure, the next four studies
utilize scores from instruments designed to isolate various
personality traits. Interestingly enough, results from
these studies are not so clear cut as those obtained in
studies which relied on simpler or more straightforward
instrumentation.

Martoccia (1964) used a measure of authoritarianism
to differentiate among 10 major groups of students. Signif-

icant results from an analysis of variance indicated that

16

students in different majors do differ on degree of author—
itarianism. However, the author suggested that, because
there was a negative correlation between intelligence test
scores and authoritarian scores, the group differences
observed might also be related to intelligence.

Lundin & Lathrop (1963) investigated the relation-
ship between personality adjustment as measured by the MMPI
and choice of undergraduate major. Twenty Hamilton College
junior and senior men in each of three majors (biology-chem-
istry, history, and English literature) were given the MMP;
and results were analyzed by a one-way analysis of variance.
No significant differences were found among the criterion
groups on any of the scales. The authors suggested that
students at Hamilton College are too homogeneous a group to
exhibit adjustment differences based on major fields.

Elton & Rose (1967) studied all male freshman
students who entered the University of Kentucky College
of Engineering during 1963-65 and three semesters after
entrance either were still in engineering or had transferred
to commerce or arts and sciences. Six measures of student
personality, scores on five factors derived from the Omnibus
Personality Inventory (OPI) scales, and the AgT_composite
score were subjected to a stepwise discriminant analysis for
the three criterion groups. One significant function
emerged (Scholarly Orientation) which accounted for 73 per-

cent of the variation among groups. Elton & Rose found that

17

students who transferred into liberal arts had the highest
mean score on this dimension. Those who transferred into
commerce had the lowest mean discriminant score, and the
group that remained in engineering had a mean score midway
between the two transfer groups. Group separation was even
more pronounced when group centroids were plotted in 2-dis-
criminant space. The second function was not significant,
however, and the authors gave it only limited attention in
their discussion. They suggested that, with a larger sample,
the second function might have reached significance, adding
a valuable dimension to the interpretation of group differ-
ences.

A 20 percent sample of the entering freshmen at the
UniverSity of Oregon was subdivided into five broad major
fields (business, education, natural science, humanities,
and social science) which were then compared and contrasted
on the basis of Poe Inventory of Values (PIV) mean scores
(Warnath & Fordyce, 1961). The ELY yields scores in eight
areas of values: aesthetic, intellectual, material, power,
social contact, religious, prestige, and humanitarian. Six
of the eight scales discriminated among the five groups;
social contact and prestige did not. Warnath & Fordyce used
‘E-tests to establish differences between each pair of major
groups on each of the six value scales, a total of 60 37

tests, and achieved significance in 19 instances. They also

18

computed a D statistic, which revealed group profile differ-
ences between each pair of major groups.

Although the Warnath & Fordyce study displayed major
group differences based on the values of group members,
their interpretations must be viewed with some caution,
because of the confounding effects produced by running 60
iftests. Hays (1963) says:

One is never really justified in carrying out all
the (3) different Eftests for differences among J
groups, and then regarding this as some kind of
substitute for the analysis of variance. Such 5—
tests carried out on all pairs of means must neces-
sarily extract redundant, overlapping, information
from the data, and as a result a complicated pattern
of dependency must exist among the tests. Further-
more, the apparent levels of significance found from
a set of such tests have neither a simple interpreta-
tion nor a simple connection with the hypothesis
tested by the §_test in the analysis of variance.

. . . A11 in all, there is very little to recommend
such a multiple iftest procedure (pp. 375-376).

On the assumption that students and faculty in a
given field tend to have personal characteristics in common
which differ from the common characteristics of students and
faculty in other fields, Astin (1965) investigated the class-
room environments of different college courses. Freshman
students at 246 colleges completed a questionnaire in which
32 items provided information on the behavior and techniques
of instructors, student behavior, and student—instructor
interaction for the one course most closely related to the
major fields. Nineteen different fields were represented in

the sample. Differences among the 19 fields were evaluated

by chi-square tests of the proportion of "yes" responses

19

given to each of the 32 items. Astin found systematic dif-
ferences among the various fields of study based on the
behavior of the students and faculty in each field.

The approach-avoidance nature of choosing an under-
graduate field is clarified somewhat in the following series
of studies. In an attempt to determine why Michigan State
University students change their majors, Pierson (1962) sur-
veyed the 1958 seniors who were graduating in a major differ-
ent from the one they had originally selected. He found
that the primary reason given for major changes was a lack
of information (about course content, the variety of majors
actually available, and the vocational opportunities related
to the original major).

The responses selected by these seniors suggested
that changers were seeking major fields which were more com-
patible with their personal interests than their original
majors had been, i.e., "the courses really didn't interest
me," "future jobs related to my original major didn't appeal
to me," "I learned about another major that suits me better."

Warren (1961) tested the hypothesis that changes in
college major are likely to occur when a discrepancy exists
between self concept and expected occupational role. His
subjects were 525 males who were either National Merit
Scholars or Certificate of Merit winners and who entered
college as freshmen in 1956. Measures used in the study

included 18 scale scores from the Omnibus Personality

20

Inventory which was administered prior to college entrance,
plus a measure of expected occupational role, taken near the
end of the freshman year. A discrepancy score was computed
for each subject. Also available were three declarations of
major field, one taken before entrance and one each at the
end of the freshman and sophomore year. Change in field was
categorized as "No Change," "Minor Change," or "Major Change,"
and mean discrepancy scores were computed for each category.
Contrary to expectations, the mean differences were not sig-
nificant. A second hypothesis, that self-role discrepancy
scores of those who changed field twice would be higher than
for those who changed once, yielded significant differences.
Additional manipulation of the original data to which were
added College Board SAT total scores and freshman year grade
point average led Warren to conclude that self-role discrep-
ancies contribute to major field changes, but do not, by
themselves, explain the phenomenon. Other factors may
either act to inhibit change (financial considerations)
where it would otherwise occur or encourage change (low
grades).

Holland's theory of vocational choice hypothesizes
that students select and remain in major fields where the
.environment (other students) is most compatible. Holland
& Nichols (1964) tested this hypothesis in a study which
drew on National Merit Finalists for the student sample and

their answers to a National Merit Scholarship Corporation

21

questionnaire for the variables. Students were assigned to
one of six major field orientations, first on the basis of
their preference as high school seniors and then, on the
basis of their preference at the end of their freshman year
in college. Changes in major were coded and weighted for
degree of change. Results tended to confirm the hypothesis
that students who leave a field are different from the typi-
cal student in that field. Holland & Nichols recommend that
information used in the counseling of students who are in
the process of choosing or changing majors be based on "a
more comprehensive review of self-conceptions, achievements,
and personality in addition to information from aptitude and
interest inventories"(p. 242).

Super's vocational development theory suggested to
Cole, Wilson, & Tiedeman (1964) that those who are alike
choose alike and the more alike group members are, the
longer they remain together. From this, they hypothesized
that the students graduating in a particular major field
would have more homogeneous test scores than the students
who entered the same field four years ealier. Two studies
were conducted, one at the University of Rochester and one
at Harvard College, which, although not identical, were
replicative.

The Rochester study was based on a sample of 759
men who entered between 1948 and 1951 and, subsequently,

completed a degree. Fifteen variables, measured at

22

matriculation, were included in the discriminant analysis of
eight basic curricular groups. Centour scores were calcu-
lated for each group member and used to determine the degree
of divergence from the centroid of each curricular group at
graduation. Results showed that students tended to move out
of fields where they were atypical and into fields more com-
patible with their abilities and interests. The incidence
of change in field was higher for students whose original
field was not the field of highest centour, thus supporting
the hypotheses.

The Harvard study utilized freshman data on 774 stu-
dents who entered the college between 1946 and 1949 and
graduated on schedule. The sample represented 16 fields of
concentration. Ten principal-component scores, based on 36
variables, were analyzed and their multivariate distributions
used to compute centour scores for each student. Centour
scores were used to determine the degree of divergence from
the original curricular group. Results did not support the
hypotheses.

Both studies showed that students are more likely to
leave the natural sciences for the social sciences and human-
ities than to enter the natural sciences from other fields
and that, in general, the holding power of the natural
sciences is weaker than the holding power of the social
sciences and humanities. Thistlethwaite (Mimeo., ca., 1960),

on the other hand, found that although the natural sciences

23

had greater holding power for males than any other field,
net enrollment decreased because the field lost more men
than it gained. It is possible that the movement of stu-
dents out of natural science fields is a reflection of stu—
dent interests and abilities, but the lack of movement into
the natural sciences could be explained, at least in part,
by the nature of the curriculum itself. The stringent
sequencing of courses may well preclude transferring into
the natural sciences after the freshman year.

Gamble (1962) investigated the pre-college, out-of-
school, personal exPeriences associated with curriculum
changes during the first three semesters of college. Expe-
rience variables included in the study were classed as home
and family relationship, vocational, religious, peer group,
community, co-curricular, or social and totaled 61. Data
were collected on beginning students at The Pennsylvania
State University and a random sample of 365 was selected
from the 2,265 students still enrolled at the beginning of
the third semester. They were grouped by sex and number of
curriculum changes affected. Three variables were signif-
icantly related to the number of curriculum changes made by
men. A favorable but not insistent attitude of parents
toward attending college, greater age of the student, and
certainty of vocational choice were linked to fewer curric-
ulum changes. Several nonsignificant relationships were

also found for men. Men whose mothers worked outside the

24

home made more changes, as did men with no siblings. It was
not possible, however, to relate the life history or experi-
ence variables included in the study to the number of curric-
ulum changes made by women. Gamble used chi-square to estab-
lish significant relationships between change categories
and experience variables which made it impossible to consider
the interrelationships among variables.

Hasan (1966) demonstrated a relationship between pre-
college extra-curricular participation and persistence in a
major field; low aptitude test scores and change of major;
and change of major and lower grade point averages. The
study sample consisted of 727 students who entered Southern
Illinois University during 1958-59 and subsequently graduated
or left the institution. Hasan found that father's and/or
mother's educational level did not differentiate among
changers and persisters.

.Augustine (1966) identified factors related to
change and persistence in major field for a sample of engi—
neering students at three large midwestern universities.
Men who transferred out of engineering for nonacademic rea—
sons were paired with men who remained in the field. Data
were collected by questionnaires and personal interviews and
included information on family background, high school eXpe-
riences, factors influencing choice of major and vocation,
and important life goals. .Augustine found significant dif-

ferences between persisters and changers on these variables.

25

Both criterion groups reported dissatisfaction with the
highly structured curriculum, dislike of certain of the
required mathematics courses, and the importance of peers in
the decision to continue in or leave engineering.
Thistlethwaite (Mimeo., ca., 1960) investigated the
relationship between changes in major field and level of
training and the experiences of college students. He
obtained information on the relative holding power of dif-
ferent fields of study and the characteristics of students
and faculty in different fields. His sample was composed of
1,500 National Merit Scholars or Merit finalists who were
completing their third year of college. Students reported
their current major field of study and first college major
and completed a modified version of the College Characteris-

tics Index which yielded a faculty press index and a student

 

press index. Fields of study were combined into five broad
academic areas: (1) natural sciences, (2) biological sci-
ences, (2) social sciences, (4) arts and humanities, and (5)
other fields. Each student also described the faculty mem-
ber who had the greatest influence upon his "desire to learn."
Descriptions were based on the student's perception of the
instructor's behavior in a course he taught. Analyses of

the press indices and faculty descriptions led Thistlethwaite
to conclude that students were attracted to and/or held in
fields where they perceived compatible role models among the

faculty. Deterrents to continued membership in an academic

26

'area were related to inappropriate student expectations for

the area.
The Predictive Studies

In most of the following studies, the researchers
were concerned with differentiating among undergraduate
majors in order to apply the resulting information to the
guidance and advisement of college students. Prediction, in
these cases, was employed to demonstrate the validity of the
analysis procedure and the accuracy of results, rather than
as an end in itself.

Included in this section are designs utilizing a
wide variety of measures and several different statistical
treatments. The first four studies are examples of predic-
tive research which rely on a single category of variable to
differentiate among curricular groups. The remaining inves-
tigations employed two or more classes of variables.

In order to test the validity of the College Qualifi—
cation Test (CQT) subscores for differential prediction,
Juola (1961) selected six broad curriculum groups of first-
term freshmen at Michigan State University and ran simple
correlations between their subscores and first-term grade
point averages (GPA). The three Egg subscores predicted GPA
for the technical curricular group better than did the total
99: score, and as well as the total for two random groups

and the nontechnical curricular group. Juola concluded that

27

the QQT subtests show potential for differential prediction
of achievement, and therefore, differential guidance of
undergraduate students.

.An early investigation of the value of student inter-
est measures as predictors of major field was conducted at
Harvard College by Tiedeman and Bryan (1954). They employed
multivariate discriminant analysis of the nine dimensions of
the Kuder-Vocationai to predict the membership of sophomore
students in one of five fields of concentration. Tiedeman
and Bryan were primarily interested in demonstrating the
effectiveness of multivariate discriminant analysis in the
treatment of student variablestxaproduce more meaningful
information than did the original data.

Although satisfied with results, the authors empha-
sized that interests alone are not enough to fully account
for differences among fields, and suggested that other
variates be exPlored as potential predictors. They also
believed it would be useful to introduce a success factor
into the analysis, i.e., using field of concentration at
graduation.

Harder (1959) used the MMP; clinical scales with
three curricular groups of male undergraduates (business,
education, and engineering) in an attempt to differentiate
curricular groups in terms of personality characteristics.

Results were not significant. He then identified MMPI items

 

which distinguished among the groups and developed scoring

28

keys for these items. Normalized T_scores were computed for
each subject and used to classify him into a criterion group.
7C1assification results-for the normative sample were signif-
icantly better than chance expectations. No cross-valida-
tion-was performed.

The Kuder Preference Scientific scale, the Guilford-
Zimmerman verbal comprehension and general reasoning sub-
scores, and the ACE Psychological total score were selected
by Stinson (1958) to discriminate among three groups of stu-
dents who entered the Engineering School of Oklahoma State
University as freshmen and subsequently graduated as engi-
neers, graduated in a non-engineering major, or dropped out
of the University. Tests were administered during the fresh-
man year. Five years later, 30 members of each group were
selected at random and from the four scores, a discriminant
function was computed. Stinson used the mean discriminant
score for each criterion group to establish three mutually
exclusive critical regions that separated the groups from
each other. Stinson recommended that counselors employ the‘
technique as an aid in the academic advisement and counsel-
ing of succeeding freshman engineering classes.

,StroquVocational Interest Blank scores, socio-
economic status of father's occupation, and Minnesota Scho-
lastic Aptitude Test scores were compared for undergraduate
men in pre-business, pre-law, pre-medicine, and engineering

(Petrik, 1966). Petrik found that socio-economic status

29

played an important part in the interpretation of Syig
scores. Lower class persisters in pre-medicine and pre-law
had different SVIB patterns than their middle class counter-
parts. In the engineering and pre-business majors, socio-
economic status had little effect on the predictive effi-
ciency of the syig; the Strong predicted persistence equally
well for both classes and was not predictive of persistence
in engineering.

Mazak (1967) used stepwise discriminant analysis of
67 variables, (abilities, interests, and values), to dis-
tinguish among five criterion groups of students who, seven
years earlier, had entered El Camino Community College as
pre-engineering students. After completing the analysis
with five groups, Mazak combined students into four groups
and then, three groups for re—analysis and prediction of
membership. .As the number of groups decreased, the accuracy
of classification increased.

The major findings reported by Mazak were: (1) high
school achievement is the best predictor of success in engi-
neering and related majors in the junior college, and (2)
junior college achievement is the best predictor of transfer
to and success in senior institutions. Mazak recommended the
use of high school achievement records as a major factor in
counseling and academic advising with pre-engineering stu-

dents and suggested that future research of this type include

30

measures of student commitment, motivation, personality,
study habits, and attitudes.

.A unique approach to clarifying the relationship
between student personality and curricular choice was devised
by Goldschmid (1966). Fifty-five academic disciplines were
rated by 142 judges on two scales, a science continuum and a
humanities continuum. Five personality measures provided
the scores used in the regresssion analysis. Students were
grouped by final undergraduate major; two-thirds of the
sample was used in the analysis and one-third was reserved
for cross-validation. Eleven of 16 regression equations
produced significant cross-validation results. In addition,
interpretation of the predictive equations indicated that
the contributing variables described the "science personal-
ity" on different dimensions than were used to describe the
"humanities personality."

Simono (1967) used performance, environmental, and
activity indices to predict membership in one of two crite—
rion groups (science and non-science) and one of six-sub-
groups: (1) engineering; (2) pre-medicine, zoology and
chemistry; (3) mathematics and physics; (4) social studies;
(5) business administration; and (6) literature and language.
{The sample consisted of 215 academically-talented sophomore,
junior, and senior college men for whom data were available
(on high school performance, family and school background,

Eand extra-curricular activities. The activity index produced

31

significant differences between science and non—science
majors. Two variables in the index were especially strong
differentiators: (1) reading preferences and (2) interest
in tinkering with machines or building models. Simono con-
cluded that, for superior students at least, measures of
extra-curricular interests are meaningful variables to use
in studying differences in academic majors.

To demonstrate the use and advantages of multiple
discriminant analysis and related classification procedures,
Cooley & Lohnes (1962) report on one phase of the Scientific

Careers Study undertaken earlier by Cooley. In this study,

 

senior engineering and science majors were given the Study
of Values test, which measures six personality characteris-
tics that Cooley believed would differentiate among three
criterion groups. The criterion groups consisted of: (1)
science and engineering majors who planned to engage in
basic research after graduation, (2) those who planned to
continue in applied science and engineering, and (3) those
who planned to leave the field for work having more direct
involvement with people. The analysis produced significant
group differences, allowing Cooley to proceed with classifi-
cation of a check sample of sophomore engineering and sci—
ence majors drawn and tested at the same time as the senior
sample. After graduation, members of the "sophomore" sample
reported their career plans. If the classification proce—

dure, based on the senior group centroids, assigned a check

32

sample member to the criterion group he planned to enter, it
was credited with a "hit." Approximately 50 percent hits
were achieved, demonstrating that value measures are effec-
tive in differentiating among members of the same academic
area whose career plans differ.

.A concern for the curricular choice problems of
undergraduates at Harvard College led King (1958) to a pre-
dictive study of the relationship between student character-
istics and final undergraduate major. The purpose of the
study was to provide analytic comparisons of student apti-
tudes and interests with various fields of study to replace
the subjective comparisons then in use. Scores from the
Aptitude Survey Test (a battery of 21 subtests), the Kggeg
Preference Record-Vocationai, Coliege Board SAT Verbal and
Mathematical scores, predicted freshman grade average, Fall
semester freshman rank, public-private secondary school
attendance, and scholarship application constituted the 36
variables available for the study sample of men who gradu-
ated in the Classes of 1950 through 1953, as well as for the
check sample of men who graduated in the Classes of 1954 and
1955. .A total of 22 fields of concentration identified the
criterion groups in the study. _Multivariate discriminant
analysis provided the information necessary for the computa-
tion of group centroids, individual centour scores, and dis-
tance measures, from which King was able to predict group

membership for the check sample. Prediction results were

33

significantly better than chance expectations. Noting the
relatively large beta weights assigned by discriminant
analysis to interest measures, King recommended a reduction
in the number of aptitude-ability measures and an increase
in the number of interest and attitude measures in subse-
quent investigations of this type.

The purpose of Dunn's (1959) study was to compare
the relative effectiveness of multiple regression analysis
and multiple discriminant analysis in the prediction of
college major. The same data were subjected to both treat-
ments and Dunn found that discriminant analysis was far
superior for predicting choice of major. The sample con-
sisted of 1,380 men and women representing 14 different
major fields. The 13 variables included ability-achievement
measures and basic demographic data. Two functions accounted
for 86 percent of the variance and were used to predict mem-
bership for a cross-validation sample. Because the 14 major
groups overlapped each other in the discriminant Space, Dunn
grouped them into six clusters and judged prediction results
as "hits," "near hits," and "misses," "near hits" being pre-
dictions that fell in the right cluster but wrong major.

She found it much easier to predict into clusters than into
exact fields of study, and suggested that greater accuracy
might result first, by adding interest and personality mea-
sures and second, by separating major groups by sex rather

than including sex as an analysis variable.

34

Tatsuoka (1957) developed a joint probability model
which combined discriminant and regression analysis to demon-
strate a relationship among measures of aptitude; high school
grades and rank in class; type of high school and geographic
location; personality ratings; and final fields of concentra-
tion. The normative sample of M.I.T. freshmen was assigned
to one of six curricular groups, depending on field of con-
centration at graduation. A seventh group consisted of the
students who had withdrawn or been suspended for academic
reasons. Multiple regression and multiple discriminant
analyses were run, using 11 variables and seven criterion
groups, resulting in seven multiple regression equations and
one significant discriminant function. Tatsuoka used these
to predict a joint probability of membership and success in
one of the criterion groups for a check sample of the next
year's graduates.

Stahmann & Wallen (1966) incorporated several of the
suggestions made by other investigators who employed multi-
variate discriminant analysis to predict undergraduate field.
In their design, students were grouped by final undergrad—
uate field, a separate discriminant analysis was run for men
and women, and both achievement and interest measures were
included. Scores on the Cooperative Engiishy_Mathematics,
and Natural Science Tests and the Occupational Interest
Inventory were obtained from the University of Utah freshman

entrance examination battery for a selected sample of 1962,

35

1963, and 1964 graduates. Male subjects represented majors
in engineering, business, pharmacy, and letters and sciences.
Female subjects represented majors in nursing, elementary
education, and letters and sciences. Subjects in majors
with 50 or more members were randomly assigned to either an
analysis sample or a cross-validation sample. Major groups
with less than 50 members were included in the analysis
sample only. Thus, the "cross—validation" sample of males
contained no pharmacy majors and the "cross-validation"
sample of females contained only elementary education majors.
Mosier (1951) defined "cross-validation" as "weights deter-
mined on one sample and their effectiveness tested on a
second, similarly drawn sample" (p. 8). Cooley & Lohnes
(1962, p. 144) also use this definition. Because Stahmann
and Wallen actually had a validity-generalization sample
rather than a cross—validation sample, their results are
difficult to interpret. They concluded that the freshman
entrance battery was an effective predictor of major field
at graduation and their predictions were better than chance
for all majors except business. .Actually, of 50 engineers
in the check sample, 29 were classified correctly, of 39
business majors, 3 were classified correctly, of 50 letters
and science majors 23 were classified correctly, and 22 men
were assigned to pharmacy although there were no pharmacists
in the check sample. The women's check sample contained 50

elementary majors and 32 were properly classified. Clearly,

36

the business major was underassigned. Stahman & Wallen sug~
gested that, because business students are a heterogeneous
group, it might be necessary to use "variables other than
achievement and interest measures to achieve adequate dis-
crimination" (p. 444).

Cutting (1966) used 15 items from the Personal

 

Informatignyinventory to predict field of concentration for
a sample of Michigan State University students who entered
as first-time freshmen in the Fall of 1963. Students were
grouped by their Fall 1965 academic majors and majors with
less than 30 members were combined into curriculum groups.
The resulting 27 groups ranged in size from 40 to 257 stu-
dents. A 25 percent random sample was set aside for cross-
validation and the remaining 75 percent of the study pOpula-
tion was used for multivariate discriminant analysis. The
15 variables were categorized by Cutting as self-concept
measures and consisted of self-ratings of six abilities and
preference rankings of nine occupational interest groups.
The analysis yielded 15 functions, 10 of which were
significant. The first three functions, accounting for 75
percent of the trace, were used to identify 11 distinct
groups and five clusters. Function one was interpreted as
a masculine-feminine dimension, with high positive weights
for verbal facility, interest in artistic occupations and
social service occupations identifying the feminine charac-

teristics and high negative weights for numerical ability

37

and interest in physical science occupations identifying the
masculine. Function two, a verbal-scientific dimension, had
high positive loadings for interest in business detail and
verbal-linguistic occupations and high negative loadings for
biological science and mechanical—technical occupational
interests. The third function showed high positive weights
for self-ratings of general ability, verbal ability, reading
ability, anticipated grade point average, and interest in
verbal-linguistic occupations; high negative weights were
attached to interests in business detail occupations and
executive-managerial occupations.

From the positive and negative loadings, the most
masculine majors should have fallen at the extreme negative
end of the first function and the most feminine majors at
the extreme positive end, but because the 15 variables were
scaled with "0" or "1" as the high scores and "4" or "8"
as the low scores, just the opposite occurred. Thus, on
function one, a high self-rating of numerical ability (i.e.,
"0") was combined with the negative weighting associated
with that variable to suppress the negative contribution to
the discriminant score and indirectly increase the positive
aspect, while a low self-rating of verbal facility (i.e.,
"4") was combined with the high positive weighting for that
variable to increase the positive portion of the discriminant

score .

38

Two cross-validation procedures were employed. In
the first, a discriminant score was computed for each member
of the check sample (N=567), for each of the 10 significant
functions and each individual was assigned to the curricular
group whose mean discriminant score was closest to his dis-
criminant score. In other words, every individual in the
check sample was assigned 10 times. Out of a possible,
5,760 assignments, 331 "hits" resulted. In the second, a
maximum likelihood classification procedure was used, which
took simultaneous account of all 10 significant functions
and assigned each individual only once. With this technique,
Cutting reported 137 "hits." He concluded that the vari-
ables and statistical treatment selected for the study were
effective in predicting academic fields for incoming fresh-
man students and suggested ways in which counselors could
use such information with students.

The results of the study must be interpreted with
some caution, however. Although not reported in the disser-
tation, a printing error on the 1963 form of the Personal
Information Inventory made it impossible to score with cer-
tainty six of the nine occupational group rankings for each
student. This, in turn, brings into question the meaning of
the contribution made to each function by the nine occupa-
tional variables. In spite of this fact, the self-ratings
'and rankings of occupational interest groups did differen-

tiate among the curricular groups analyzed, suggesting that

39

these measures are somehow related to differences among

major fields.

Summary

The foregoing review of literature emphasized

descriptive and predictive studies of student characteris-

tics associated with differing undergraduate major fields.

The following conclusions represent the principal findings

relevant to the current study.

1.

Measures of student and/or faculty characteristics
are effective in differentiating among a variety of
undergraduate major fields.

A wide variety of characteristics has been examined
and shown to be useful. Although interest and value
measures appear to be particularly potent differen-
tiators, researchers are fairly unanimous in recom-
mending the use of a variety of measures, including
ability/achievement, family background, high school
background and experiences, and self-ratings, in
addition to interests and values.

Student characteristics are useful not only in estab-
lishing major field differences, but in identifying
similarities which cross department and college
organizational lines.

Many investigators believe that more accurate dis-
tinctions result when final undergraduate majors
are used to establish criterion groups than when
first or interim majors are used. Accuracy and
understanding are improved when criterion groups
are controlled for sex.

Multivariate discriminant analysis is particularly
well suited to research problems dealing with the
establishment and eXplanation of group differences
which are based on a variety of variables.

40

These conclusions, coupled with the rationale pre-
sented in Chapter I,.suggested the following guidelines for
the design of the present study.

1. Student characteristics should be measured as
early as possible, preferabiy_p£ior to first collegiate
attendance. For information of the type which should result
from this study to be applicable to the choice problems
faced by freshman students and their advisers, the data on
which it is based must be collected and processed for use
during the first term of enrollment. The following studies,
of those cited above, employed precollege measures to suc-
cessfully differentiate among undergraduate majors; Baird,
1967, No. 17 a No. 19; Cutting, 1966; Gamble, 1962; Hassan,
1966; Juola, 1961; Simono, 1967; Stahmann & Wallen, 1966;
Veldman _£_ai,, 1968; and Voorhies, 1966.

2. From the pre-college measures availabley those
selected for incigsion in the stgdy should represent a
variety of ciasses. Interest and preference measures
receive strong emphasis in Holland's theoretical formula-
tions (1966) as well as his research (Abe & Holland, 1965;
Holland & Nichols, 1964). In addition, studies by Baggaley
(1947), Matteson (1961), and Tiedeman & Bryan (1954) relied
exclusively on interest variables and showed useful results.
Similar evidence is available in support of measures of
ability/achievement (Cullum, 1966; Dunn, 1959; Juola, 1961;

and Voorhies, 1966), values (Baird, 1967, No. 19; Cooley &

41

Lohnes, 1962; and Warnath & Fordyce, 1961), family and high
school background (Augustine, 1966; Baird, 1967, No. 17;
Gamble, 1962; Smith, 1967; and Veldman §£_ai,, 1968), and
self-ratings (Abe & Holland, 1965 and Cutting, 1966).
Results of research studies using measures of adjustment
and/or personality traits were less consistent (Elton & Rose,
1967; Harder, 1959; Lundin & Lathrop, 1963; and Martoccia,
1964). .A number of studies were based on a combination of
two or more classes of variables. The following researchers
specifically recommended the inclusion of several categories
in future investigations: Abe & Holland, 1965; Dunn, 1959;
Holland & Nichols, 1964; King, 1958; Korn, 1962; Martoccia,
1964; Stahman & Wallen, 1966; Tiedeman & Bryan, 1954; and
Voorhies, 1966.

3. In so far as possible, criterion groups should
be kept fipure." Three considerations are relevant to this
point. When are groups identified, what academic areas
should be combined into criterion groups, and how should the
sex variable be treated? Tiedeman & Bryan (1954) were the
first to suggest that criterion groups include only those
students who graduated in each field, but others have since
made the same recommendation (Abe & Holland, 1965; Baird,
1967, No. 17 & No. 19; and Cutting, 1966). Interpretation
of results has been clearer in those studies where cluster-
ing of major fields was based on initial findings, i.e.,

Dunn, 1959; and King, 1958. Considerations of sample size

42

have sometimes made it necessary to treat men and women in

the same major as one criterion group. Other studies have

been designed to include only men. Dunn (1959) used sex as
an analysis variable, but recommended that it be used as a

criterion variable so that separate analyses could be run.

Again, clearer interpretations would undoubtedly result if

criterion groups were controlled for sex.

Applications of these guidelines led to the design

and methodology detailed in Chapter III.

CHAPTER III

DESIGN AND METHODOLOGY

The Sample

Selection of Students and Final
Major Fields

 

The restricted population from which the study
sample was drawn entered Michigan State University as first-
time freshmen during either Fall 1963 or Fall 1964 and had
attained senior standing (130 quarter credits or more), or
had graduated from Michigan State University by Spring 1967.
All students in this group who had earned more than 30 quar-
ter credits from an institution other than Michigan State
were eliminated. A total of 3,681 students, 1,934 males and
1,747 females, filtered through the first set of screens. A
second set of restrictions was placed on the sample because
of criteria established for the selection of final major
fields.

A major field is defined as a program of study which
is specifically identified by a college of the University,
and which culminates in a bachelor's degree. The 3,681
students represented 127 majors in 10 colleges. Major

groups containing less than 40 members of one sex were

43

44

eliminated, leaving a total of 2,096 students in 23 majors
from 9 colleges. Nine of these majors contained only
females, ten contained only males, and each of the four
remaining majors had both a male and a female group. The
sample, thus, consisted of 27 groups representing 23 differ—
ent final majors.

The last restriction on the sample, that all mea—
sures be available for all subjects, resulted in an initial
loss of 130 students whose orientation test scores and/or
Personal Information Inventory (PII) data were missing or
incomplete, and a subsequent loss of 201 students whose gi_
forms were no longer on file at the Counseling Center.

The study sample consisted of 1,765 students, 909
males and 856 females, in 23 majors from 9 colleges. A
20 percent sample of each major group was drawn from the
study sample and set aside for use in the cross-validation
study, while the remaining 80 percent of each group became
the analysis sample. The composition of the study sample,
the analysis sample, and the cross-validation sample is

shown in Table 1.

Instrpmentation

The maximum number of variables which could have
been included in the discriminant analysis, at the time the
study was designed, was limited by computer capacity to 50.

Only those variables which were measured before students

45

 

 

 

 

E E... ME. 2 .38
Ha me em m xuoz Hmfioom cocoaom amfloom .mm
0 mm ev m hmoHonommm oucoflum HMHUom .mm
0 mm we m mocoflom Hmfioom mocoflum amaoom .mm
oa mm me E Aneumam muwuumq a muu< .em
ma Nb om m nmflamcm mumuumq m wand .mm
oH mm me e moauumum use mumuqu a muu< .mm
m mm 0v m nuoomm muud coflumoacsfieoo .HN
m AM me n Hmaommm coHDmusem .oN
Hm mom mmm m sumscmsmam coHDmustm .m.
m Hm me E meamusz mocoflom amusumz .mH
OH av Hm m moaumEmnumz mocmflom amusumz .ha
m em Ne m .nuoau c .xoe .Hum moHEocoom oEom .oa
OH mm we m mcflcumos .m .m moHEocoom meow .ma
Hma mmh mom ea Hmuoa
«HI «NI wml. z mmoHonowmm oucmwom Hmaoom .ea
ea mm Nb 2 oocmflom Hecauflaom oocmflom Hmfloom .ma
m em we 2 qulonm mocwaum HMHUOm .NH
ea mm mm 2 mucoflom amaoom mocoflom anacom .aa
NH om No S zuoumflm mumuuoq o muud .oa
ma mo mm 2 mUHumEonumz oocmwom amnoumz .m
m hm we 2 mmoHOON mocmflom amusumz .m
m mm ee 2 AMUHCMSUDE mcwummafimcm .h
ea mm on 2 Hmofluuomam meauoocflmcm .0
ma am we 2 mcflumxumz mmocflmsm .m
m em we 2 moaeocoom mmocflmsm .e
ma mm em 2 .cAEpé .msm .cmw mmmcflmsm .m
ma mm me 2 meaucsooue mmmaflmsm .m
0H m@ am 2 mcflmmxomm musuasoflumd .H
onEmm oHQEmm mamfimm xom HOnmE omoaaoo
sesameuam> mumsamca sesum
Immouu mo ouwm mo ouflm mo oNHm
mqmz<m sommo ozm .mquEm mHmwa<z<
.mqmzdm NQDBm mo mNHm de Xmm Nm "MQDBm ZH QMQDAUZH mmbomo mObdz 02¢ MOMAQOU .H mqmde

46

began formal course work at Michigan State were admissable.
Within this framework, the theoretical considerations set
forth in Chapter I and the results of previous research
cited in Chapter II provided the rationale for selecting,
from the orientation data available, the specific variables
which were included in the study. As previously noted, the
most meaningful measures for predicting membership in an
academic field appeared to fall into four categories: (1)
aptitude/achievement measures, (2) interest and self-rating
measures, (3) family and background measures, and (4) educa-

tional and vocational preference measures.

Aptitude/Achievement Measures
Raw scores on the following tests were used as mea-
sures of academic aptitude and achievement.
1. The Collegeygpalification Test (CQT)
a. Verbal
b. Information
c. Numerical

2. MSU Orientation Tests

a. English
b. Reading

The College Qualification Test (Bennett, G. K.,
Bennett, M. G., Wallace, W. L., & Wesman, A. G., 1957) is a
ZOO-item battery of three tests: (1) a test of verbal facil-
ity, (2) a test of numerical ability, and (3) a test of
general information. A total of 75 vocabulary items provides

the verbal score. Included in the numerical test are 50

47

items measuring conceptual skills necessary in the study of
arithmetic, algebra, and geometry. The 75-item information
test deals with general knowledge in the physical, biolog-

ical, and social sciences.

The Michigan State University English Test and Read—
ing Test were developed by the Office of Evaluation Services
of Michigan State University. The former test consists of
38 items representing many aspects of English usage and was
designed to be especially discriminating at lower ability
levels. The reading test contains 50 questions based on a
series of reading passages from the four general education
areas in which University College courses are offered. This
test provides supplementary information for the selection of
low-ability students who need remedial work as well as for
selection of high-ability students who are eligible for
honors sections (Office of Evaluation Services, 1964).

High school grade point average had to be eliminated
as a measure of ability/achievement because the form in
which it is recorded on the Student Master Record does not
distinguish among high school grading scales based on differ-
ing point systems.

Measures of Interest, Family
Background, and Educational
apgyVocational Preference

The size of high school graduating class and loca-
tion of the high school (in- or out-of—state), were taken

from the Student Master Record which is compiled by the

48

Office of the Registrar. All other measures described below
came from the Personal Information Inventory which is filed
in each student's record at the Counseling Center.

The 2;; was developed by the Michigan State Univer-
sity Counseling Center and was first administered to stu-
dents during the 1963 summer orientation clinics. The 1964
,2;; is a revision of the 1963 instrument. All of the 3;;
information used in the analysis had to be re-scaled, either
to accomodate differences in the two forms of the Inventory
or to convert qualitative information into quantitative mea-
sures suitable for the analysis. The gii_items included in
the study, the original coding, and the redscaling are pre—
sented in Appendix A.

Interest and self-rating measures. Four classes of
variables were included in the interest and self-rating
measures.

1. Reading preferences for novels, technical books,
mystery books, biography/history, and literary
classics;

2. Rating of enjoyment of high school courses in fine
and applied arts and letters, science and mathe-
matics, social studies, and non-academic subjects;

3. Self-rating of general ability, numerical reasoning,
verbal reasoning, reading skill, and anticipated
college grade point average (GPA) in comparison to

other students entering MSU; and

4. Ranking of interest in three of nine occupational
groups.

49

Famiiy and backgrognd measuges. Ten variables were

included in this categroy.

1. Educational level of father;

2. Educational level of mother;

3. Occupational status of father;

4. Employment of mother;

5. Marital status of parents;

6. Number of older siblings;

7. Number of younger siblings;

8. Size of high school graduating class;

9. Location of high school in- or out-of—state; and

10. Sources of financial support for college.

Educational and vocational preference measures.
Four classes of variables were included in this category.
1. Strength of first major choice;
2. .Certainty of vocational choice;
3. Factors influencing vocational choice; and
4. Reasons for coming to college.
§ummary of Instrumentation and
_yariables Used
All variables came from one of three sources: (1)
the Student Master Record, (2) the Office of Evaluation
Services test record, or (3) the Personainnformation Inven-
my. The measurements selected fell into four categories:

(1) ability/achievement, (2) interest and self-ratings,

50

(3) family and background experiences, and (4) educational

and vocational preferences.

The variables included in the

analysis are enumerated in Table 2.

TABLE 2.

Ability/Achievement Measures
. CQT Verbal

. CQT Information

. CQT Numerical

. MSU English

. MSU Reading

m-PLAJNH

Reading Preferences

6. Novels

7. Technical books

8. Mystery books

9. Biography and History
10. Literary classics

High School Cogises

11. Fine and applied Arts
and Letters

12. Science and Mathematics

13. Social studies

14. Non-academics

Self-Ratings

15. General capacity

16. Numerical reasoning

l7. Verbal reasoning

18. Reading skill

19. Anticipated college GPA

Occupapional Interest Groups
20. Artistic

21. Physical science

22. Biological science

23. Mechanical and technical
24. Social service

25. Business detail

26. Business contact

27. verbal and linguistic
28. Executive and managerial

VARIABLES INCLUDED IN THE DISCRIMINANT ANALYSIS

Family and Background Experi-
ences
29. Educational level of

father

30. Educational level of
mother

31. Occupational level of
father

32. Employment of mother

33. Marital status of parents

34. Number of older siblings

35. Number of younger siblings

36. Size of high school
graduating class

37. High school in- or out-of-
state

38. Sources of financial
support

Educational and Vocational

Preferences

39. Strength of first major
choice

40. Certainty of vocational
choice

Factors influencing vocational

choice:

41. Family and significant
others

42. High school experience

43. Personal interests and
needs

Reasons for coming to college:

44. Educational/vocational
needs

45. Social needs

46. Family influence

47. Status needs

51

The Statistical Model

Multivariate discriminant analysis was selected as
the appropriate statistical treatment to accomplish the
purposes of this study. Saupe (1965) has presented a con-
cise description of the technique.

The data consist of several variables or measures
for individuals in a number of groups. Let M_be
the number of measures and let §.be the number of
groups.

The analysis of this type of data with the multiple—
discriminant model requires the calculation of two
matrices. .A is the MJcM matrix of among-groups sums
of products. Each entry in.A_is calculated in the
same manner as an among-groups sum of squares in an
analysis-of—variance problem or a sum of cross—
products in an analysis-of-covariance problem. It
is an exPression of the disPersion among group
means. 91 is the matrix, also MxM, of within-
groups sums of products. The entries in.W_are
calculated in the same way as are within-groups

sums of squares and cross-products in analysis-of—
variance or -covariance problems. It is an eXpres-
sion of within-groups dispersion.

Matrices A and.W are then used in the basic matrix
equation,

(A-AW)V = O

which is solved for A, a scalar, and y, a column
matrix of M elements. The resulting Ai§_and yi§_
are the principal results of the basic analysis.
Some characteristics of these results follow:

1. In general, there is more than one solution;
that is, unique vector y_and associated scalar A,
As a matter of fact, because the rank of matrix A
is the smaller of gel and M, this is the exact
number of nonzero solutions to be expected.

2. The solution for which A.is largest is said
to define the first discriminant function; the

solution with the next largest A, to define the
second discriminant function; and so forth. The
first discriminant function defines that linear

52

combination of the M_measures that maximizes the
ratio of among-groups dispersion to within-groups
diSpersion. The second discriminant function
maximizes the ratio of among- to within-groups
diSpersion for that component of the among-groups
dispersion that is not accounted for by the first
discriminant function, and so forth.

3. Each solution or function is orthogonal to
(that is, uncorrelated with) all the other func-
tions of the analysis.

4. The ifs can be interpreted in relation to their
sum as proportions of the diSpersion among groups
accounted for by the respective functions.

5. If the data are the product of random sampling
from multivariate normal populations with homoge-
neous dispersion matrices, hypotheses about group
differences can be tested. The hypothesis of over-
all group differences can be tested by means of
Wilks',A,criterion, and hypotheses concerning the
differéHCes associated with individual discriminant
functions and combinations thereof can be examined
by means of a y? approximation based upon the ifs
(Rao, 1952).

6. The vector y_of each solution consists of coef-
ficients that can be applied to the set of measures
for an individual to produce a discriminant score
for the individual. An individual has as many dis-
criminant scores as there are discriminant functions.
Discriminant scores are usually not computed for
nonsignificant functions.

7. A significant discriminant function is inter-
preted by examining the magnitudes of the standard-
ized coefficients for the various measures and
examining the mean discriminant scores for the
several groups.

8. With the resulting discriminant functions, a set
of discriminant scores can be computed for an un-
classified individual for whom there are scores on
the involved measures and he can be classified into
the group that has mean discriminant scores closest
to his discriminant scores (PP. 175—177).

As stated in Saupe's fifth characteristic, signif-

icance tests based on multivariate discriminant analysis

53

assumes that the within group distributions of raw data are
multivariate normal with equal variance and covariance
matrices. The problems associated with testing these assump-
tions were discussed by Ikenberry (1960). A letter from Dr.
David Tiedeman, which is quoted by Ikenberry, states:

I don't know of a simple, direct test of the

multivariate normal assumption. There are no

tables of which I am aware which Specify the

expected frequencies under the assumed multi—

variate normality. . . . We usually test the

normality of the discriminant scores rather

than that of the original scores (in Ikenberry,

1960; p. 72).

A perusal of the within group frequency distribution
Showed that several of the variables, as originally scaled,
were bimodal or J-shaped, while others had a more nearly
normal appearance. It was decided to rescale the bimodal
and J-shaped variables in such a way as to increase the
normalcy of the distributions without destroying the trends
originally present. In general, the scales affected were
constructed a priori to retain as much information from the
primary sources as possible so that a posteriori adjustments
could be made with a minimum loss of data. Appendix B con-
tains the original and subsequent frequency distributions of
the variables which were rescaled.
The use of discriminant scores to classify an indi-

vidual whose group identity is unknown, Saupe's eighth point,

assumes multivariate normal distribution of the discriminant

scores .

54

The data were analyzed by computer utilizing the
multivariate discriminant analysis program available from
the Michigan State University Computer Institute for Social

Science Research (CISSR Technical Report 33).

Summary

The student sample for this study was drawn from the
first-time freshmen who entered Michigan State University
during either Fall 1963 or Fall 1964 and had graduated or
attained senior standing by Spring 1967, having earned no
more than 30 quarter credits from another institution. The
majg£_sample consisted of the undergraduate major fields
represented in the student sample which contained at least
40 students of one sex. A total of 1,765 students, 909 men
and 856 women, in 23 different majors made up the final
study sample from.which was drawn a 20 percent random sample
for cross-validation purposes.

Data were collected from three sources: (1) the
Student Master Record, (2) the Office of Evaluation Services
test records, and (3) the Personal Information Inventopy. A
total of 47 variables was analyzed using multivariate dis—
criminant analysis. The results of the analysis are pre-

sented in Chapter IV.

CHAPTER IV

ANALYSIS OF DATA

The data were subjected to four treatments: (1) a
discriminant analysis of the 27 major groups, (2) a separate
discriminant analysis of the 14 male groups and the 13
female groups, (3) classification of a cross-validation
sample based on the 27-group analysis, and (4) classifica—
tion of male and female cross-validation samples based on
the 14- and l3—group analyses, respectively.

As Saupe (1965) pointed out, multivariate discrim-
inant analysis provides a number of nonzero solutions (func-
tions) equal to one less than the number of groups; each
function is orthogonal to all the other functions in the
analysis; and a function is interpreted by examining the
size and Sign of the standardized coefficients (weights) for
the variables in concert with the mean discriminant scores
(group centroids) for the various groups.

Discriminant Analysis of Twenty-Seven
Criterion Groups

 

The major hypothesis of the study was that under-
graduate major fields can be differentiated on the basis of

selected characteristics of the students who attain senior

55

56

standing or have graduated in each major. Stated in null
form:

No differences exist among undergraduate major

fields on the basis of the abilities/achievements,
interests, educational-vocational preferences, or
family and high school backgrounds of their student
members.

To test this hypothesis, 47 measures of aptitude,
interest, family and high school background, and educational-
vocational preference were collected on a student sample rep—
resenting 23 different undergraduate majors. (Nine fields
contained only females, ten contained only males, and each
of four fields had both a male and a female group, making a
total of 27 criterion groups.) A multivariate discriminant
analysis of the data produced 26 discriminant functions, 10
of which were significant (Rao's X?) at the .001 level, a
rejection of the null hypothesis. As shown in Table 3, the
first three functions accounted for 71 percent of the disper-
sion among groups (trace), the addition of functions four and
five increased the percent of trace to 79, and the remaining
five significant functions accounted for an additional 11
percent of the trace. The 16 nonsignificant functions
eXplained less than 10 percent of the variation among groups.

A secondary hypothesis of the study was that male
and female students in the same major have substantially
different patterns of characteristics. Although there is no

direct statistical test for this hypothesis, Rao (1952)

offers an application of the Mahalanobis 2? statistic which

57

 

 

 

 

TABLE 3. PERCENT OF TRACE, DEGREES OF FREEDOM, AND STATISTI-
.CAL SIGNIFICANCE LEVELS FOR FIRST TEN DISCRIMINANT
FUNCTIONS OF THE TWENTY-SEVEN GROUP ANALYSIS

Percent of Trace Tabled
Function X2 Values Camputed
Number Individual Cumulative df a = .001 X Values
1 50.78 ... 72 114.8 1,608.5
2 10.89 61.67 70 112.3 535.7
3 9.58 71.25 68 109.8 481.4
4 4.08 75.33 66 107.3 226.0
5 3.99 79.32 64 104.7 221.2
6 2.88 82.20 62 102.2 163.4
7 2.35 84.55 60 99.6 134.7
8 2.14 86.69 58 97.0 122.9
9 1.94 88.63 56 94.5 111.8
10 1.66 90.29 54 91.9 96.5

 

provides relevant descriptive information. The treatment is

based on the concept of "clusters." A cluster is defined as
two or more groups which fall close together in the discrim-
inant space. Restating the second hypothesis in terms of the
cluster concept: Males and females in the same major form
two groups which fall into mutually exclusive clusters. In
null form the hypothesis becomes:

Males and females in the same major form two
groups which fall into the same cluster.

Rao (1952) explains the computational procedures

thus:

58

No formal rules can be laid down for finding
the clusters because a cluster is not a well-
defined term. The only criterion appears to be
that any two groups belonging to the same cluster
should at least on the average Show a smaller 2?
than those belonging to two different clusters.

A Simple device suggested by K. D. Tocher is to
start with two closely associated groups and find
a third group which has the smallest average 2
from the first two. Similarly, the fourth is
chosen to have the smallest average 2? from the
first three, and so on. If at any stage the
average 2? of a group from those already listed
appears to be high, then this group does not fit
in with the former groups and is therefore taken
to be outside the former cluster. The groups of
the first cluster are then omitted, and the rest
are treated similarly (pp. 362-363).

The clusters shown in Tables 4 and 5 were formed
using intercentroid distances and Rao's computational scheme.
The computations are included in Appendix C. The four
fields which contained both a male and a female group were
mathematics (Groups 9 and 17), history (Groups 10 and 24),
social science (Groups 11 and 25), and psychology (Groups 14
and 26). They are underlined in the tables. It can be seen
that none of the pairs fell in the same cluster, and none
fell in the next nearest cluster. The male groups were
closest to other male groups, and with the exception of
women mathematics majors, the female groups were closest to
other female groups. The mathematics women were closer to
Cluster D (male zoology and psychology majors) than any
other cluster and closer to psychology women than any other
major. These results suggested that, for the study sample,

male and female members of the same major had substantially

59

 

 

 

... ma m
mm.u .... ... ME. 6
SH.H mm.a .... em. smnmm
Iamuomumaumauma m
mm.a Hm.a GN.H .... mm. .mmummulmImm m
am.s m~.a mm.a me.H .... mm. .mwum a
om.H mo.a om.a mo.H mH.H .... Hm. marmaummrlwre o
mm.a 6m.H mo.m AH.N me.a em.a .... om. mumumua m
Ha.~ mm.a nm.~ o~.m mm.a em.a Ne.a .... mm. .mysuo a
m o m m a o m m moamumun “mumsao some a. HmumsHo
scumDHOImuucH pmpdaocH mHOnmS
mocmumflm HmumsHUIHmucH ommuo>¢ wmmuo>¢

 

 

MHmMQ¢Z¢

mbomw Zm>mmIMBZm39 20mm mmUZ¢BmHQ QHOMBZmummBZH OZHmD mMflBmDQU m0 ZOHBflSmOh

.v flamma

6O

 

NNINN

 

musuo 4 Iamuomnmauoauma n ma m
mumumua m wanm a _Mfl o
nmumm
munuo a ma m uamuomumauoauma m
5.2 I .l I.

6.5.0 4 Iamuomumauoauma m omummu mumm m

hmumm II
Hmuomumauoauma m maumauaauoaue 6 «Hum a

ma a smumm .II.II
Iamuomnmauoauma m e m aaum a maumauaauoaue o
omummuamumm m maumauaauoaue o mIMINIH m

Hmumm .I
Iamuomnmaumanmu m aaum a 6-5-6 e
nonmSHU ummnunmm Hmumsao umumsao ummumoz Houmsau Hmumoao,cfi umumsao
GA muonmz umonunmm GA muoflmz ummummz popsaucH mHOnmz

 

 

mHmNA¢Z¢ mbomw Zfl>mmIMBZMZB mom mmHmmZOHedflmm MMBmDAUImmBZH

.m figm¢8

61

different patterns of characteristics and resembled other
criterion groups more than they resembled each other.
Interpretation of the

Discriminant Functions

Although the analysis produced ten Significant func-
tions, the first five functions accounted for 79.32 percent
of the trace, while the last five significant functions
accounted for only 10.97 percent of the variation among
groups. ,Consequently, only the first five functions were
interpreted in detail. Table 6 identifies variables by name
and number and should be used in conjunction with Tables 7
through 25 in which variables are designated by number only.

The first function accounted for 50.78 percent of
the variation among groups. The standardized weights as-
signed to the 47 variables for this function are shown in
Table 7. It was decided to consider, for interpretive pur-
poses, those variables assigned an absolute weight of 10.00
or more.

Four positive and six negative weights reached this
magnitude in function one. High positive loadings were
assigned to the MSU Engiish test score, reading preference
for literary classics, preference for fine and applied arts
and letters in high school, and interest in social service
occupations, while high negative loadings were assigned to
QQT Information and Numerical test scores; interest in phys—

ical science occupations, mechanical-technical occupations,

TABLE 6.

AAbility/Achievement Measures

1.
2.
3.
4.

5.

Reading

6.
7.
8.
9.
10.

CQT Verbal .
CQT Information
CQT Numerical

,MSU English

MSU Reading
Preferences
Novels

Technical books
Mystery books
Biography and History
Literary classics

High School Courses

11.

12.
13.
14.

Fine and applied Arts
and Letters

Science and Mathematics
Social studies
Non-academics

Self-Ratings

15.
16.
17.
18.
19.

General capacity
Numerical reasoning
Verbal reasoning
Reading skill
Anticipated college GPA

Occupational Interest Grogps

20.
21.
22.
23.
24.
25.
26.
27.
28.

Artistic

Physical science
Biological science
Mechanical and technical
Social service

Business detail

Business contact

Verbal and linguistic
Executive and managerial

VARIABLES INCLUDED IN TABLES 7 THROUGH 25

Familyyand Background Experi-
ences

29. Educational level of
father

Educational level of
mother

Occupational level of
father

Employment of mother
Marital status of parents
Number of older siblings
Number of younger siblings
Size of high school
graduating class

High school in- or out-of-
state

Sources of financial
support

30.
31.

32.
33.
34.
35.
36.

37.

38.

Educational and Vocatignal

Preferences

39. Strength of first major
choice

40. Certainty of vocational
choice

Factors influencing vocational

choice:

41. Family and Significant

others

High school eXperience

Personal interests and

needs

Reasons for coming to college:

44. Educational/vocational

needs

Social needs

Family influence

Status needs

42.
43.

45.
46.
47.

63

TABLE 7. STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR TWENTY-SEVEN
MAJOR GROUPS ON FUDCTION ONE

 

 

 

 

High-Weight Plot of Group
Standardized Variables Grow Centroids on
Variable Weight (x 100) + - Major Group Centroid this Function
1 8.09 — °
2 -23.25 x
3 -26.83 x
4 16.67 x
5 5.77 21. Speech -O.30
6 3.67 22. Art Practice 4.32::
7 - 2.84 23. English -0.41
8 - 2.48 19. Elena. Educ. -0.48::'
9 - 2.95 20. Special Educ. -0.50——§_,:— -0.5
10 11.71 x 27. Social Work -0.54——-——"'-
11 12.33 x 16. Ret.. Tex. 8 Clo. -0.60-——-——--
12 - 0.39 25. Soc. Sci. (r) -0.64 ;
14 - 0.63 24. History (2) -0.69;_=
15 - 1.42 18. Nursing -0.7g/"_
16 - 9.13 15. Home ac. Tchg. -o.a /"
17 - 1.92 26. Psychology (F) -o.84
l8 - 0.75 — -l.0
20 7.87
21 -22.46 x
22 - 1.37
23 -16.09 x
24 17.03 x
25 - 4.83
26 - 8.33
_ 17. Mathematics (1') -l.56 -—- -l.5
g; 4564' x 10. History (Ti) -l.66 \_
29 _ 1'42 11. Social Sci. (a) -1.76\_
. F-
30 7.44 \—
31 — 2.49
13. Political Sci. -l.84 —
32 ' 1'35 14. Psychology (M) -l.87\—..
33 0‘34 12 Pre-Law -1 99.x“?
34 1.12 ' '
35 4.30 \___-2 o
36 3.23“ '
8. Zoology -2.10
3; 3:3; 5. Marketing -2.13 \
39 1.29 3. Gen. Bun. Adm. -2.11
40 - 3.97
41 3.48 2. Accounting -2.33\_
42 - 4.09 4. Economice -2.37
43 - 2.25 9. Mathematics (M) -§.44 2 5
44 0.28 1. Packaging - .50 _—- .
45 4.18 ——'1
47 -12.60 x 6. Electrical Eng. -2.67~_
7. Mechanical Eng. -2.75\P

 

 

 

64

and executive-managerial occupations;.and coming to college
to fulfill status needs. Function one was interpreted as a
verbal/humanistic versus numerical/technical dimension.

A plot of the major group centroids on this function
(Table 7) gave further support to the verbal/humanistic
versus numerical/technical dichotomy. Majors at the extreme
end of the numerical/technical area included mechanical
engineering, electrical engineering, packaging, and mathe-
matics. Majors at the extreme end of the verbal/humanistic
area included speech, art practice, and English. The fact
that all the male majors spread out along the high negative
end of the ordinate and the women's majors (with the excep-
tion of the mathematic's women) huddled at the low negative
end suggested that the first function reflected a sex-
related difference among the criterion groups. This led to
the decision to run separate discriminant analyses of the
male majors and the female majors, the results of which are
described later in the chapter.

Function two (Table 8) accounted for approximately
11 percent of the difference among groups. High positive
weights were assigned to six variables (QQT Information test
scores; biography and history reading preferences; interest
in social studies in high school; and preferences for busi-
ness contact, verbal and linguistic, and executive and man-
agerial occupations) and high negative weights were assigned

to four variables (MSU English and Reading test scores;

65

TABLE 8. STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR TWENTY-SEVEN
MAJOR GROUPS ON FUNCTION TWO

 

 

 

 

High-Weight Plot of Group
Standardized Variables Group Centroids on
Variable Weight (x 100) + - Major Group Centroid this Function
1 9.81 -- 0
2 14.62 x
3 4.12
4 -13.74 x
5 -17.57 x
6 - 5.02
7 - 4e73
8 - 3.91
9 10.85 x 13. Political Sci. -0.34 ____._.
10 - 1.69
11 - 8.10
12 - 5.23
B 13.3? " 10. History (M) -0.55 r- —-0.5
15 1.04 12. Pro-Law -0.58—~—
1’: ' 3%; 4. Economics . -0.69\_
18 3.15 3. Gen. Bus. Adm1n. -0.71 _
11,3 : 3°32 11. Soc. Sci. (M) -0.80\___
21 21°09 x 5. Marketing -0.84-—-—-—-———:
- ° ' _ /
22 _ 7.03 25. Soc. Sc1. (F) 0.85
23 - 6.84
24 - 4.49 24. History (1’) -1.00\-—
25 1.16 15. Home Ec. Tchg. -l.06 \t- —— -1.0
26 11.55 x 2. Accounting -1.07 “2
27 20.02 x 16. Ret.. Tex. a Clo.. -l.037"‘_
28 15.15 x 23. English -1.0
29 - 1.31 27. Social Work -1.18 *:
30 5.84 14. Psychology (n) -1.19//j,_._"‘_
31 6.64 19. Elem. Educ. -1.23
32 1.44 20. Special Educ. -1.33
33 1.66 21. Speech -l.33 >:
34 2.43 26. Psychology (F) -1.37
35 0.22 1. Packaging -1.44 ——_._:
36 - 3.81 8. Zoology -1.4s ——-—-—"‘
37 " 5.18 22. Mt Practice “1.54 —-——n— —-1.5
38 - 6.49
39 - 8.04 _
‘0 _ 9.37 7. Mechanical Eng. 1.63 ——__ __
:2 :13'22 x 18. Nursing —1.70 —-——-—_
‘3 _ 6.68 9. Mathematics (24) -1.74 -———-""'"—
. . _ __________..__
44 _ 7.39 6. Electrical Eng. 1.78
45 - 5.30 .
46 _ 4.79 17. Mathematics (1’) -l.88 WP
47 - 6.05
—-200

 

 

 

66

preference for physical science occupations, and the impor-
tance of high school experiences in the selection of a voca—
tion). Although the nature of the second function was less
clear—cut than function one, the positively-weighted vari-
ables suggested a generalist orientation and the negatively-
weighted variables a specialist orientation.

The distribution of group centroids tended to con-
firm the generalist-specialist dichotomy. The five majors
at the specialist end of the scale were mathematics (women),
electrical engineering, mathematics (men), nursing, and
mechanical engineering, while the five majors at the other
extreme were political science, history (men), pre-law,
economics, and general business administration.

The third function (Table 9) was interpreted as an
abstract versus applied orientation. High positive loadings
on CQT Verbal and Information test scores, MSU Reading score,
anticipated college GPA, and interests in physical science,
.social service, and verbal/linguistic occupations were asso-
ciated with high positive group centroids for social science
(women), psychology (men and women), English, and mathe-
matics (women) majors. Variables with high negative weights
included preference for non-academic subjects in high school,
interest in business detail occupations, high school in
Michigan, certainty of vocational choice, family influence
in decision to come to college as well as in choice of voca-

tion, and the status value of a college education. Majors

TABLE 9.

67

STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR TWENTY-
SBVEN MAJOR GROUPS ON FUNCTION THREE

 

High-We ight
Variables

Standardized

Variable Weight (x 100) +

Major Group

Plot of Group
Centroids on
this Function

Group
Centroid

 

QQQOU§WNH

20.89 x
10.04 x
- 3.45
0.70
14.02 x
- 7.09
7.01
1.52
4.86
4.54
4.38
0.42
6.42
-10.07
7.84
- 8.09
- 4.85
4.77
14.96 x

 

14.
25.
26.
23.
17.
13.
24.

12.
10.
4.
8.

20.
15.

11.
27.
18.

21.
19.
16.

22.

UINUI-I

Psychology (M)
Soc. Sci. (P)
Psychology (F)
English
Mathematics (F)
Political Sci.
History (F)
Mathematics (M)
Pre-Law
History (M)
Electrical Eng.
Economics

Zoology

Special Educ.
Home Ec. Tchg.
Mechanical Eng.
Soc. Sci. (M)
Social Work
Nursing

Speech
Elem. Educ.
Ret., Tex. & Clo.

Art Practice

Packaging
Gen. Bus. Admin.
Accounting
Marketing

-- 3.0

 

 

 

TT

 

 

68

with low positive group centroids on function three were
marketing, accounting, general business administration,
packaging, and art practice.

Seven variables were assigned high negative weights
in function four. They were: CQT Verbal score; preference
for reading literary classics; interest in artistic, phys-
ical science, mechanical/technical, and executive managerial
occupations; and, vocational choice influenced by high
school experiences. This combination of variables was
labeled artistic/technical. The high positive variables
(MSU Readianscore, interest in social service occupations,
and educational level of mother) suggested a conventional
feminine orientation. From such a polarization, it was not
surprising to find that the women's majors were separated
much more than the men's majors (Table 10). Group centroids
at the artistic/technical pole were art practice, speech,
retailing of textiles and clothing, and English. Majors at
the conventional feminine end of the scale included nursing,
home economics teaching, social work, and elementary educa-
tion.

The high-weight variables in function five shared no
obvious common qualities. Positive weights were assigned to
self-rating of numerical reasoning; interest in social ser-
vice, business detail, and artistic occupations; and family
influence of vocational choice. Major groups with high

positive centroids included accounting, mathematics (men and

69

 

 

TABLE 10. STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR TWENTY-SEVEN
MAJOR GROUPS 0N FUNCTION FOUR
High—Weight Plot of Group
Standardized Variables Group Centroids on
Variable Weight (x 100) + - Major Group Centroid this Function
1 -15.04 x 18. Nursing 0.48
2 - 0.74 15. Home Ec. Tchg. 0.47 §__+°°5
3 9.44 27. Social Work 0.42 "
4 - 9.06 19. Elem. Educ. 0.37\
5 14.27 X 8. Zoology 0.35 __
6 - 4.86 20. Special Educ. 0.33\‘ —— 0 4
7 - 3.58 2. Accounting 0.29 '
8 — 2.54 26. Psychology (F) 0.29 h
9 3.69 \:
i2 -18':§ x 12. Pre-Law 0.27 _ 0.3
12 6:05 17. Mathematics (P) 0.272% _
13 3.13 24. History (I?) 0.27 +—
1; _ 1:9,} 9. Mathematics (11) 0.24 /
16 7.32 5. Marketing 030/: — 0.2
17 - 3.05 13. Political Science 0.19 /=
18 1.16 3. Gen. Bus. Admin. 0.187,.
19 2.58 7. Mechanical Eng. 0.18 _
20 -27.23 x 4. Economics 0.17/_-__ 0.1
21 -10.66 X 10. History (M) 0.17
22 9.28 25. Social Sci. (P) 0.13 '-
23 -11.91 x 11. Social Sci. (M) 0.12 f"
24 12.96 x 14. Psychology (M) 0.12
25 1.52 6. Electrical Eng. 0.07
26 - 9.09 1. Packaging 0.06 — 0
27 - 8.98
28 -13.37 x 23. English -0.05-——-——-‘-
29 - 0.10
30 10.87 x —-0.1
31 2.09
32 2.38
33 - 1.03
34 - 0.67
35 5,44 16. Ret., Tex. & Clo. -0.20-——-—--—-—-o.2
36 - (17.30
37 - .87 ______
38 1.29 21. Speech -0.25
39 - 6.00 __
40 6.44 0'3
41 1.00
42 -15.32 x
43 - 5.78
44 - 5.34 ..
45 - 2.45 0.4
46 - 1.68 L—
47 - 3.11 22. Art Practice -0.44——---—"'"
—-—--O.5

 

 

 

70

women), economics, retailing of textiles and clothing, and
social science women. High negative weights were assigned
to self-rating of reading skill, interest in biological
science occupations, and reasons for coming to college.
Majors with low positive group centroids included nursing,
zoology, and social science men (Table 11).

A comparison of group clusters on functions one
(numerical/technical-verbal/humanistic), two (generalist-
specialist), and three (abstract-applied) highlighted the
separation achieved by the discriminant analysis. Cluster A
(electrical engineers, male mathematicians, and mechanical
engineers) fell into the abstract-specialist-highly numer-
ical/technical category, although mechanical engineers were
lower on the abstract scale than were the other two majors
in the cluster. Cluster B (packaging, accounting, general
business administration, and marketing) was characterized as
applied-generalist-highly numerical/technical. Packaging
and accounting were higher on the numerical/technical scale
and lower on the generalist scale than were general business
administration and marketing. Cluster C (economics, history,
social science men, pre-law, and political science) was char-
acterized as abstract-generalist-numerical/technical. With-
in the cluster, economics majors were higher on numerical/
technical, and social science men were less abstract than
the rest. Cluster D (zoology and psychology men) was

clearly in the numerical/technical category, but zoology

71

TABLE 11. STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR TWENTY-SEVEN

MAJOR GROUPS ON FUNCTION FIVE

 

 

 

High-Weight Plot of Group
Standardized Variables Group Centroids on
Variable Weight (x 100) + - Major Group Centroid this Function
1 6.78
2 3.16
3 - 3.16 "1.5
4 4.74
5 - 5.26
6 - 6.29
7 - 1.49 2. Accounting 1.34-———___._.
8 - 4.31
9 0.41
10 - 1.10
11 - 5.96 9. Mathematics (M) 1.15
12 0.46 17. Mathematics (F) 1.09\
13 6.67 4. Economics 1.06 \-
14 - 1.55 16. Ret., Tex. a. Clo. 1.06\
15 - 4.43 25. Social Sci. (1*) 1.05 ' t:
16 21.35 x 19. Elem. Educ. 1.01 *—
17 6.28 15. Home Ec. Tchg. 1.00 " 1 0
18 -12.42 x 23. English 0.99 '
19 7.69 10. History (M) 0.96
20 11.51 x 6. Electrical Eng. 0.95
21 4.75 24. History (P) 0.94
22 -17.42 x 3. Gen. Bus. Admin. 0.92
23 2.26 7. Mechanical Eng. 0.90
24 21.85 x 22. Art Practice 0.87
25 24.70 x 13. Political Sci. 0.85 _
26 5.44 20. Special Educ. 0.85
27 6.13 5. Marketing 0.
28 1.25 26. Psychology (8) 0.84 "
29 4.26 21. Speech 0.80
30 - 5.29 12. Pre-Law 0.77
31 - 3.98 14. Psychology (M) 0.75 __
32 - 1.88 27. Social Work 0.7 ....o 5
33 3.79 1. Packaging 0.7 '
34 - 0.16 11. Social Sci. (M) 0.65 -
35 7.60 8. Zoology 0.52/
36 1.78 18. Nursing 0.44
37 0.64
38 - 5.17
39 3.31
40 0.96
41 12.33 x
42 9.24
43 7.60
44 —20.33 x
45 - 8.26
46 -11.32 x
47 -17.24 x ....o

 

 

 

 

72

majors were less abstract than the psychology men and
slightly more specialist. The first female cluster, E, was
abstract--more generalist than Specialist-verbal/humanistic.
The psychology women were the least verbal/humanistic and
generalist of the four majors in Cluster E, which also
included English, history, and social science. Cluster F
(home economics teaching, retailing of textiles and clothing,
elementary education, speech, art practice, and social work)
was more applied than abstractwemore specialist than general-
ist-verbal/humanistic. Home economics teaching majors,
special education majors, and social work majors were more
abstract than the other four majors and the art practice
group was higher on the specialist function than the rest.
Mathematics women and nursing majors were not clustered with
any other major groups. Mathematics women fell into an
abstract-specialist-numerica1/technica1 category, the only
female group classified as abstract-specialist and the only
female group not classified within the verbal/humanistic
segment of function one. Nursing was flanked by Cluster F,
but more specialist than any of the seven Cluster F-majors
and more abstract than four of them (Figure l).

The results of the 27-group analysis not only demon-
strated that undergraduate major fields could be differen-
tiated on the basis of their members' characteristics, but
that the most pronounced difference among criterion groups

was sex-related. For this reason, separate discriminant

73

.mHmNA¢24 mDOMO Zm>mmIWBZmZE ho mZOHBUZDm
BZfiUHhHZOHm mmmmfi EmmHm ZO mQHomBZMU mDomv m0 904m Q¢ZOHmZMEHQ MMMEB

L,
l
a
l~§§=////_
fl -_ N}.
2

3
2

.H mmDUHm

    
 

 

... ...:E\\|\.\.)|||)

_--i-I_

 

:
2 8

h
.—
-—--u.—_--

 

 

74

analyses were carried out for the 14 male majors and the 13
female majors.
Discriminant Analysis of Fourteen
Male Criterion Groups
The l4—group analysis produced 13 discriminant func-
tions, of which five were significant at the .001 level. The

five significant functions accounted for 79.85 percent of the

variation among groups (Table 12).

TABLE 12. PERCENT OF TRACE, DEGREES OF FREEDOM, AND
STATISTICAL SIGNIFICANCE LEVELS FOR FIRST FIVE
DISCRIMINANT FUNCTIONS OF THE FOURTEEN-GROUP
ANALYSIS

 

 

Tabled
Function X2 Values Computed
Number Individual Cumulative df a = .001 x2 Values

Percent of Trace

 

 

1 33.26 ... 59 98.3 415.6
2 21.90 55.16 57 95.8 299.6
3 11.38 66.54 L 55 93.2 171.6
4 7.53 74.07 53 90.6 118.1
5 5.78 79.85 51 88.0 92.3

 

The first function, numerical/physical science versus
verbal/social science, assigned high positive weights to the
99? Numerical score, self-rating of numerical reasoning,
interests in physical science and mechanical/technical occu-

pations, and certainty of vocational choice; high negative

75

weights were attached to the CQT Verbal score, preference
for social studies courses in high school, and interest in
verbal/linguistic occupations. High positive group cen-
troids were associated with electrical engineering, mathe—
matics, mechanical engineering, and packaging while low
positive group centroids were associated with political
science, history, pre-law, and social science. Function one
explained 33 percent of the group variation (Table 13).
Function two, applied versus abstract, closely
resembled the third function of the 27-group analysis; high
positive weights were attached to the QQT Numerical score,
preference for novels, preference for nonacademic high
school courses, self-rating of verbal reasoning, interest
in business detail occupations, and the importance of family
influence and status needs in the decision to come to col-
lege and the choice of vocation. Majors at the positive end
of the continuum were accounting, marketing, general busi-
ness administration, and packaging. Negative loadings were
associated with CQT Verbal and gpformation scores, the M§g_
Reading score, anticipated college GPA, and interests in
physical science, biological science, and social service
occupations. Majors grouped at the abstract pole included
psychology, electrical engineering, mathematics, and polit—
ical science, followed closely by pre-law, zoology, and

history (Table 14).

 

 

“an...

TABLE 13.

76

STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR FOURTEEN
GROUPS ON FUNCTION ONE

 

 

High-Weight

Plot of Group

 

 

 

Standardized Variables Group Centroids on
Variable Weight (X 100) + - Major Group Centroid this Function
1 -l4.38 x
2 2.13
3 13.68 x
4 4.03
5 9.45
6 0.88
7 4.32
8 5.88
9 - 9.34
10 - 7.31
11 - 2.24 6. Electrical Eng. 2.52~——_________
12 5.27 7. Mechanical Eng. 2.46-——_________ 2.5
13 -12.08 x
i; : 1:32 9. Mathematics 2.37-—-—-——“‘-'
16 15.58 x
17 - 3.98
{g ' 3:33 1. packaging 2.14-—————--"-
20 7.21
21 22.36 x 2-0
22 5.61
23 17'91 x 2. Accounting 1.88 f’
24 0'69 8 2001 1 88
25 6.36 ' °9Y '
26 - 2.37
27 ~19.60 x
28 3.00 5. Marketing 1.49
29 0.53 4. Economics 1.48~\\“-‘_‘__
30 - 8.79 14. Psychology 1.48:::::::>—fi__ 1.5
31 — 6.99 3. Gen. Bus. Admin. 1.40-————______.
32 - 1.15
33 - 0.36
34 - 1.18
35 2.01
36 1.07 11. Social Science 1.14 -———-———
37 3.70 12. Pre-Law 1.08---——'—‘—-
38 5.72
39 7.53 1-0
40 11.33 x 10. History 0.91
41 1.02
42 8.21 13. Political Science 0.79
43 3.39
44 - 8.33
45 - 0.67
46 - 0.36
47 1.78

 

 

 

TABLE 14 .

7'7

STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR FOURTEEN

GROUPS ON FUNCTION TWO

 

 

HighAWeight Plot of Group
Standardized Variables Group Centroids on
Variable Weight (X 100) + - Major Group Centroid this Function
l -11.98 x
2 -14.64 x
3 23.54 x
4 - 9.54
5 -18.78 x
6 10.29 x ___ +0.5
7 - 8.29
8 0.63 2. Accounting 0.31
9 1.31 5. Marketing 0.30‘P“-c\“_
10 - 4.73 3. Gen. Bus. Admin. 0.27.__________::
11 - 2.35 '-
12 0.28
13 — 1.43
i; _lé:gg x 1. Packaging 0.04 ————____—q_.
16 8.81 0
17 13.78 x
18 - 7.17
19 -26.23 x
20 - 1.39
21 -l9.15 x 11. Social Science -0.30———-—---‘-
22 ~12.46 x
23 1.13 7. Mechanical Eng. -0.41---———-“"
24 -l4.62 x .___
25 18.90 x ’0-5
26 - 3.50 4. Economics -0.59 ————fi—
27 0.14 10. History -0.71
28 4.94 12. Pre-Law '0'7ZEEEEEEEEEE+E
29 - 3.08 8. Zoology -0.73
30 5.57
31 0.87 13. Political Science —0.87 75
32 - 3.06 9. Mathematics -0.90"’,,z””_
33 1.06 6. Electrical Eng. ~0.9l '-- -1.0
34 - 3.26
35 0.30
36 - 7.14
37 8.44 14. psychology -1.21-——————-—4-
38 - 0.83
39 4.39
40 4.97
41 27.17 x
42 19.26 x -—- -1.5
43 6.96
44 9.21
45 - 1.37
46 13.97 x
47 16.25 x

 

 

 

78

The third function assigned high weights to only two
classes of variables: reasons for coming to college and
occupational preference groups. High positive weights were
associated with the former and high negative weights with
the latter, suggesting an educational-to-vocational spread.
The group separations, however, were not great. All group
centroids were negative and covered less than a one-point
range (from -.70 through -l.36). The function accounted for
11.38 percent of the trace. Economics and accounting had
the highest negative centroids and zoology had the lowest
negative centroid (Table 15). The remaining eleven crite-
rion groups fell within three-tenths of a point of each
other.

In function four, high negative loadings were
attached to preference for high school science and mathe-
matics courses, educational level of mother, coming to
college for status reasons, occupational level of father,
and number of older siblings. Mechanical engineering, pre-
law, and economics majors had the highest negative centroids.
Positive weights were attached to preference for business
detail, social service, and biological science occupations;
self-rating of numerical reasoning ability; vocational
choice influenced by family and significant others; the M§g_
Reading score; and preference for literary classics. Account-

ing had the lowest negative centroid on this function.

TABLE 15.

STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT
GROUPS ON FUNCTION THREE

759

OF GROUP CENTROIDS FOR FOURTEEN

 

 

 

 

High-Weight Plot of Group
Standardized Variables Group Centroids on
Variable Weight (x 100) + - Major Group Centroid this Function
1 - 5.10
2 - 0.40
3 1.53
4 2.84
5 0.58
6 5.97 _ -__-_____—-——— -— -0.7
7 1.33 8. Zoology 0.70
8 3.69
9 - 6.75
10 8.48
11 6.87
12 - 3.44 -0.8
14 - 1.38
15 2.91
16 - 7.79
17 - 4.65 _ __
18 5.90 1. Packaging 0.89-——————-—'+ -O.9
19 - 8.42 11. Social Science -0.92-——_______,__
20 - 5.04 14. Psychology -0.94——~_,_
21 -12.26 x
22 4.09
23 -12.77 x -1 0
24 -16.47 x °
25 -20.28 x 5. Marketing —1.02
26 -12.77 x
27 -15.70 x
28 -13.23 x
29 0.53 . _ ___________
30 _ 2.15 3. Gen. Bus. Admin. 1.09 -1.1
31 0.11 12. Pre-Law -1.12.-——————---
32 2.41 7. Mechanical Eng. -1.15
33 - 0.58 10. History -1.15-———‘_’__._::
34 - 4.25 6. Electrical Eng. -l.16 __
35 - 3.97 9. Mathematics -1.18:::;:=="“ _1 2
36 - 0.15 13. Political Science -1.18 ‘
37 0.73
38 5.38
39 - 2.99
40 - 3.57
41 4.54 _ __ _
42 4.37 2. Accounting 1.30————- 1.3
43 3.43
44 13.70 X
45 4.82 4. Economics -l.36 -
46 6.01
47 10.89 x _1.4

 

 

 

80

Psychology, mathematics, history, and zoology majors were
also at the low negative end of the scale (Table 16). The
high negative variables suggested a socio-economic status
factor, while the high positive variables were too mixed to
allow a single interpretation. Business detail occupations
and self-rating of numerical ability carried the highest
positive weights, which, combined with group means on these
variables, eXplained the position of accounting and mathe-
matics. Psychology, history, and zoology were more strongly
affected by the positive weights assigned to social service
and biological science occupational interests.

The high negative variables in function five
(Table 17) also suggested a socio-economic status factor;
self-rating of numerical reasoning, number of older and
younger siblings, occupational level of father, and sources
of financial support for college were important in separat-
ing zoology and pre-law majors from the rest of the groups.
Three of the four College of Business majors (marketing,
economics, and general business administration) also had
negative group centroids. The high positive variables were
heterogeneous in character. Interest in mechanical/techni-
cal, physical science, social service, verbal/linguistic,
and executive/managerial occupations as well as the QQT_
verbal score and the employment of the mother outside the

home were assigned positive weights over 10.00. Packaging

TABLE 16.

81

STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT
GROUPS ON FUNCTION FOUR

OF GROUP CENTROIDS FOR FOURTEEN

 

High4Weight

Plot of Group

 

 

 

 

Standardized Variables Group Centroids on
Variable Weight (X 100) + - Major Group Centroid this Function
1 1.44
2 0.55
3 - 8.18
4 - 2.78 2. Accounting -0.78—-—-——‘“‘*-
5 10.61 x “‘ -0.8
6 1.52
7 5.94
8 0.30
9 - 4.85
10 10.28 x
11 0.42 -—--0.9
12 -15.04 x 14. Psychology -0.92-—————-—--
13 2.01
14 - 1.39 9. Mathematics -0.96~—_________
15 - 6.01
16 20.35 x 10. History -0.98-—-—""'
17 - 3.20 — ‘1-0
18 - 7.07 8. Zoology -1.02-——________
19 4.65
20 7.00
21 - 5.49
22 13.80 x
23 - 2.00 _ "1'1
24 16.21 x
:2 23:33 " 13. Political Science -1.15\__
27 - 1.18 3. Gen. Bus. Admin. -l.l7
28 - 0.59 5. Marketing 4.17;: _ _1 2
29 6.58 1. Packaging -1.19"/,,r”" '
30 -15.08 x 11. Social Science -1.20
31 —12.65 x
32 6.16
33 - 1.30 6. Electrical Eng. -1.27-——-——"“'
34 —12.72 x
35 - 4.63 —- -1.3
36 3.33
37 2.20
38 - 2.63
39 - 2.72
40 - 2.97 4. Economics -1.40 -——--1.4
41 10.03 x
42 6.28 12. Pre-Law -1.43
43 7.26 __
44 - 3.08 7. Mechanical Eng. -1.45——--'"""
45 - 9.57
46 - 9.68
47 -13.68 x —'-1-5

 

 

 

TABLE 17.

STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT
GROUPS ON FUNCTION FIVE

82

OF GROUP CENTROIDS FOR FOURTEEN

 

 

High-We ight

Plot of Group

 

 

 

Standardized Variables Group Centroids on
Variable Weight (X 100) + - Major Group Centroid this Function

1 12.01 x

2 2.06

3 3.03

4 5.91

5 - 5.85

6 0.82 +0'4
7 2.87

8 1.09

9 - 1.74

1° ' 2°31 1. Packaging 0.31

11 2.48 . +0 3
12 - 6.29 °
:2 :23; 10. History 0.26 ———-——

15 - 9.44

16 -14.08 x

17 0.50 _

18 - 4.84 +0.2
19 1.70 6. Electrical Eng. 0.14

20 9.62 14. psychology 0.13\__

21 19.35 x 7. Mechanical Eng. 0.11 )—

22 2.86 13. political Science 0.10\__

23 25.16 x \_ —— +0.1
24 11.82 x

‘22: " 3:33 11. Social Science 0.06 —————-—~

g; 10:36 i 2. Accounting . 0.02 ____L

29 0.84 3. Gen Bus. Admin. -0.01\_‘L __ 0
30 - 3.79 4. Economics -0.02 ——-—--—

31 -1l.27 x 5. Marketing -0.08

32 10.49 x

33 2.42 \_

34 -13.08 x

35 -12.80 x — -0-1
36 9.63 9. Mathematics -0.11-———-———

37 1.54

38 -11.01 x

39 - 2.45

40 7.76

41 - 7.68 — '0-2
42 - 1.44 12. Pre-Law -0.22—————-—

43 - 7.75

44 7.01 8. Zoology -0.26 ~—
45 4.12
46 3.90

47 9.19 — -03

 

 

 

83

and history had the highest positive centroids for function
five.

Using the same computational scheme employed in the
27-group case, the fourteen majors formed four clusters (A'
through D') identical to the original clusters (A through D).
Clusters A' through D', intra- and inter-cluster distances,
and nearest cluster are shown in Tables 18 and 19. Computa—
tions are shown in Appendix D.

Comparing the plot of Clusters A' through D' on the
first three functions with the previous plot of Clusters A
through D revealed certain differences in their relative
positions. Economics, history, and pre-law were less
abstract on the l4-group analysis, making Cluster B' less
abstract than Cluster B. Cluster B' was clearly verbal/
social science in the second analysis, in contrast to its
position on function one of the 27-group analysis, where it
was at the low end of the numerical/technical segment.
Zoology shifted toward the numerical/physical science end
of function one on the 14-group analysis, appearing more
like accounting and packaging than in the 27-group analysis.
Thus, the absence of the female majors resulted in greater

clarification of the differences among the male groups.

84

 

mIMINIH .m MHINHIHHIOHI¢ .U fiHlm .Q

 

@1510 .4 fialm .Q manmalaaloalu .0
calm .Q malmalaaloalg .U mlmlmla .m
manmalaaloalv .0 galw .Q minim .4
Hmumsao ummnuumm Hmumsau Houmsao ummumoz Hmumsau Hmumsao as Hmumsao
CH muommz umwnuumm CM muons: pneumoz popsaocH muons:

 

 

mHmNQ4Z¢ ADOMO ZNHBMDOm mom mmHmmZOHB4qmm mmBmDAUImmBZH .mH mflm48

 

 

.... mm. calm .n
00.4 ... mm. maumalaaloalo .0
mm.H m~.H .... on. mIMINIH _m
mo.a Hm.H om.H .... mo. minim .4
.o .0 .m .4 mUGMDMHQ Houmsao, nounsau_sfl noumsau
mocmumwa HmumsaoluoucH ommuo>4 ImmucH mmmum>4 popsaocH muommz

 

 

mHmMA4Z¢ ADOMO
ZNMBMDOW.ZOM& mWOZ<BmHQ QHOMBZMUmMBZH OZHmD mmmBmDAU m0 ZOHB<Zm0m .mH mnmdfi

85

Discriminant Analysis of Thirteen
Femgle Criterion Groups
The l3—group analysis produced 12 discriminant func-
tions, of which five were significant at the .001 level and
accounted for 76.34 percent of the variation among groups

(Table 20).

TABLE 20. PERCENT OF'TRACE, DEGREES OF FREEDOM, AND STATIS-
TICAL SIGNIFICANCE LEVELS FOR FIRST FIVE DISCRIM-
INANT FUNCTIONS OF THE THIRTEEN-GROUP ANALYSIS

 

 

Percent of Trace Tabled
Function X2 Values Computed
Number Individual Cumulative df o = .001 x2 Values

 

1 25.76 .... 58 97.0 299.8
2 18.21 43.97 56 94.5 225.2
3 15.04 59.01 54 91.9 191.2
4 9.31 . 68.32 52 89.3 124.7
5 8.02 76.34 50 86.7 108.8

 

Almost 26 percent of the trace was accounted for by
the first function. All weights of 10.00 or more were posi-
tive and were assigned to the following variables: interest
in physical science occupations, the QQT Numerical score,
the MSU Reading score, factors influencing vocational choices,
and coming to college for educational and vocational reasons.
This combination of variables suggested a scientific orienta-

tion for those majors with high positive group centroids as

86

contrasted with a nonscience orientation of the majors with
low positive group centroids. Majors in the former category
included mathematics, psychology, nursing, history, social
science, and English. As shown in Table 21, the mathematics
women were much higher than any of the other groups, the psy-
chology women occupied a second, separate and distinct posi-
tion on the continuum, and the next four majors grouped
together near the midpoint of the distribution. The four
majors at the low positive end of the plot were retailing of
textiles and clothing, elementary education, speech, and art
practice. The prime differentiator, then, appeared to be an
interest in science coupled with numerical ability.

The second function separated the majors on the
strength of their verbal orientations. High positive load-
ings were associated with the QQT Verbal score, interests in
verbal and linguistic as well as social service occupations,
and self—rating of verbal reasoning ability. Negative
weights were attached to certainty of first vocational
choice, MSU Reading score, and self-rating of numerical
reasoning ability. The combination of positive and negative
loadings assigned the highest positive centroids to social
science and English majors (Table 22), followed by history
and psychology. The lowest centroids were assigned to nurs-

ing and mathematics.

TABLE 21.

87

STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR THIRTEEN

GROUPS ON FUNCTION ONE

 

 

 

 

 

 

 

 

High-Weight Plot of Group
Standardized Variables Group Centroids on
Variable Weight (X 100) + - Major Group Centroid this Function
1 2.58
2 4.07
3 18.62 x
4 - 4.78
2 13:33 x 17. Mathematics 4.25 __.__.,_..
7 0.25
8 6.63
9 2.42
10 - 1.86
11 - 0.79
12 4.06 — 4.0
13 0.06
14 - 1.83
15 5.20
16 3.29
17 - 1.50
18 - 0.80
19 0.08
20 - 4.02
21 32.95 x
22 5.13
23 0.69 _ 3 5
24 - 3.33 26. Psychology 3.48 r— '
25 0.65
26 - 0.42
27 4.02
28 - 2.84
29 3.55
30 - 5.21
31 5.38
32 - 1.32 18. Nursing 3.13 --——-~—
33 - 7.09 24. History 3.08 ...
34 1.70 25. Social Science 3.07 ______._._.
35 - 1.64 23. English 3.01—.———-—-4—- — 3,0
36 - 4.05
37 - 5.85
38 5.53 20. Special Educ. 2.89
39 4.84 15. Home Ec. Tchg. 2.85-—-——
40 - 1.81
41 10.35 x 27. Social Work 2.76 ——
42 20.90 x 16. Ret., Tex. & C10. 2.70 ‘_
43 15.63 x 19. Elem. Educ. 2.67 —
44 11.16 x 21. Speech 2.63 ———-—
45 5.05 22. Art Practice 2.60 —-p
- 46 1.30
47 7.98 — 2_5

 

88

 

 

TABLE 22. STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR THIRTEEN
GROUPS ON FUNCTION TWO
High-Weight Plot of Group
Standardized Variables Group Centroids on
Variable Weight (X 100) + - Major Group Centroid this Function
1 23.65 x
2 9.79
3 3.38 25. Social Science 1.38-————____.
4 1.38
5 -13.51 X
6 - 2.12 23. English 1.25 ———_1__"_
7 6.07
8 3.46
9 - 1.92
10 9.88
11 0.19
12 - 6.65 ———+1.0
13 6.13 24. History 0.95———~____.
14 - 4.80 26. Psychology 0.94—”r“"‘_‘
15 0.99
16 -11.14 x
17 11.88 X
18 0.39
19 1.68 16. Ret., Tex. 5 Clo. 0.73/V
20 5.03 )—
21 — 7.65 21. Speech 0.66-e—"”_
22 — 9.62 20. Special Education 0.57——________
23 - 4.91 22. Art Practice 0.55’—"*”‘EE
24 12.79 x 19. Elementary Educ. 0.54::j77ffl -———+o.5
25 0.69 27. Social Work 0.54
26 7.10 15. Home Ec. Tchg. 0.53
27 17.52 x
28 6.36 17. Mathematics 0.34-—-——-‘_-
29 - 3.68
30 - 1.99
31 6.05
32 - 0.75
33 - 1.59
34 6.50
35 6.98
36 1.56 18. Nursing -0.009 _P _ 0
37 - 6.44
38 - 2.84
39 - 0.62
40 -21.30 x
41 4.05
42 4.16
43 0.92
44 - 3.22
45 3.66
46 - 7.17
47 - 7.20
— -0.5

 

 

 

89

Function three produced a small fine arts-creative
cluster of majors at the high negative end of the scale and
a large cluster of service-oriented majors at the low nega—
tive end of the scale (Table 23). Highest negative weights
were assigned to preference for artistic occupations, voca-
tional choice influenced by personal needs and experiences,
the MSU Enqligh score, strength of first major choice, pref-
erence for literary classics, and the importance of a col-
lege education for its own sake. Art practice, speech, and
retailing of textiles and clothing had high negative cen-
troids. Nursing, English, and mathematics majors fell in
the middle of the plot and the remaining seven criterion
groups were close together at the low negative end. The
positively-weighted variables were interest in social ser-
vice occupations, the QQT Numerical score, education level
of mother, and interest in biological sciences.

Function four (Table 24) produced very little over—
all spread among the majors. Mathematics had the highest
positive centroid (2.31), and psychology had the lowest
(1.79), a difference of only .52 points. Examination of the
high-weight variables suggested no clear explanation of the
nature of the difference.

Function five assigned high positive weights to
eleven variables and high negative weights to five variables
(Table 25). Nursing received the highest positive centroid

on the function and home economics teaching and retailing of

TABLE 23.

END

STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF
GROUPS ON FUNCTION THREE

GROUP CENTROIDS FOR THIRTEEN

 

 

 

 

High-Weight Plot of Group
Standardized Variables Group Centroids on
Variable Weight (X 100) + - Major Group Centroid this Function
1 - 3.21
2 8.02
3 11.33 x --' 0
4 -13.23 x
5 1.09
6 - 9.26
7 1.27
8 - 2'43 15 Home E
9 4.56 e c. TChg. -OeZIN;
10 —11.48 x 24. History -0.25-——-———d-—
11 - 1.45 27. Social Work -0.31
12 2.38 19. Elementary Educ. -0.35 "
13 8.43 20. Special Educ. -0.37 2:
14 - 1.14 26. Psychology -0.39-——————--
15 - 5.23 25. Social Science -0.40.————“"'
16 - 1.49
i; _ 8:3; 18. Nursing -0.49 ————-‘-'- -—--0.5
19 5.04
20 -20.04 x _.
21 _ 3.66 23. English -0.59 —-—"‘-
22 10.69 x
23 - 3.57 17. Mathematics -0.69 ——.——
24 23.35 x
25 0.29
26 - 3.22
27 1.92
28 - 6.75
29 - 0.44
30 11.06 x
31 3.04 16. Ret., Tex. & Clo. -0.94 2'—
32 2.70 -——--1.0
33 — 2.76
34 — 1.41 21. Speech -1.06 -————
35 5.82
36 1.73
37 - 5.83 22. Art Practice -l.16
38 - 0.76
39 -12.79 x
40 5.76
41 - 4.61
42 —l9.04 x
43 -10.25 x
44 -12.55 x
45 - 1.54
46 - 5.70
47 - 1.22 -—--1-5

 

 

 

TABLE 24.

91

STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR THIRTEEN
GROUPS ON FUNCTION FOUR

 

 

High-Weight Plot of Group
Standardized Variables Group . Centroids on
Variable Weight (X 100) + - Major Group Centroid this Function
1 -l4.l9 x
2 15.78 x
3 - 0.94
4 8.01
2 : 2:33 17. Mathematics 2.31————-——-___ 2.3
7 - 5.82
8 -12.53 x
13 : 8:3: 19. Elementary Educ. 2.24-—————__
11 4.38
12 - 1.71 -—- 2.2
13 - 3.36
i: : 3:13 25. Social Science 2.16‘——-“—'
16 10.17 x
i; _ 1‘13 22. Art Practice 2.11-—-——‘*'___. 2.1
19 _ 5:33 24. History 2.09-———-::
20 3.01 23. English 2.08
21 7.21
22 -12.28 x
23 - 6.16 21. Speech 2.01—_______
24 3.23 27. Social Work 2.00———~—_ 2-0
:2 0:36 16. Ret., Tex. & Clo. 1.97-———————
27 0.64 20. Special Education 1.95--—‘-
28 - 0.51
29 1.20
30 0.89 -—- 1.9
31 6.05
32 0.23
33 6.61 18. Nursing 1.84-—————u—
34 - 1.75 15. Home. Ec. Tchg. 1.83-———""’
35 2.20 1 8
33 12g; x 26. psychology 1.79 -————-—-
38 0.91
39 1.06
40 - 2.00
41 17.26 x _
42 9.07 1'7
43 10.13 x
44 1.02
45 4.70
46 - 3.79
47 9.91 1.6

 

 

 

TABLE 25.

92

STANDARDIZED WEIGHTS, GROUP CENTROIDS, AND PLOT OF GROUP CENTROIDS FOR THIRTEEN

GROUPS ON FUNCTION FIVE

 

 

 

 

 

 

 

 

 

 

 

High-Weight Plot of Group
Standardized Variables Group Centroids on
Variable Weight (X 100) + - Major Group Centroid this Function
1 -13.66 x
2 - 9.50
3 0.27
4 17.90 x 18. Nursing 2.04
5 13.16 x
6 - 2.15 ‘——’2.0
7 6.71
8 0.28
9 0.52
10 1.37 _1.9
g 32.972 x 23. English 1.87—
13 -10.54 x 26. Psychology 1.84
14 -11.76 x 27. Social Work 1.84/
15 - 6.36 ....l.8
16 -13.92 x 20. Special Educ. 1.79
17 8.22 22. Art Practice 1.77
18 7.78 21. Speech 1.76 ___l
19 - 7.36 19. Elem. Educ. 1.70'——__________—__
20 - 5.34 24. history 1.67 — ‘ 1'7
21 2.07 25. Social Science 1.62 \
22 19.47 x _-
23 4.60
24 2.96 "‘
25 - 5.72 —1-6
26 -15.10 x
27 3.12 17. Mathematics 1.55
28 - 0.32
29 - 9.39
30 13.91 x "—1-5
31 6.50
32 - 1.11
33 - 3.92
34 6.21 _
35 0.73 1-4
36 13.46 x
37 5.39
38 10.10 x
39 0.99 _1-3
40 - 0.01
41 - 7.60
42 _ 7.04 16. Ret., Tex. & C10. 1.25 ———e
43 - 9.39 __1 2
44 26.35 x °
45 17.66 x
46 10.09 x 15. Home Ec. Tchg. 1.16 ———4
47 23.27 x

 

 

-‘—'1.1

 

93

textiles and clothing the lowest, with the remaining ten
majors closely grouped in the center. The highest positive
variables were the educational and status advantages of a
college education, preference for biological science occupa-
tions, and the MSU English score. High negative variables
included preference for business contact occupations, self-
rating of numerical ability, the CQT verbal score, and a
preference for social studies and non-academic subjects in
high school.

The clusters which were formed using intercentroid
distances from the lB—group analysis and the same computa-
tional procedure employed in the 27- and l4-group cases were
slightly different in composition and character from the
original 27-group clusters E through H (Tables 26 and 27).
Computations are shown in Appendix E.

The seven majors of the original Cluster F formed
two clusters, E' and F'. Cluster E' contained home econom-
ics teaching, elementary education, special education, and
social work majors. Cluster F' contained retailing of tex-
tiles and clothing, speech, and art practice majors. Clus-
ter E' was more social-service oriented and less artistic/
creative than F' and more closely knit than any of the other
clusters. Cluster F' was the least science-oriented of all
the clusters. Again, English, history, social science, and
Jpsychology formed a cluster and mathematics and nursing

Inajors were isolates.

94

 

 

 

 

 

 

mmlamloa .m ma .m 5H .H
ha .H nNIoNImHImH .m ma .m
5H .H nmlomlmalma _m omlmmlwmlmm .0
PH .H thONlmHlmH .m NNIHNI®H .m
“H .H PH. “5.9 HH¢ .m~ .mv~ ..m FNIONIGHImH .m
Hmumsao ummnuumm Hmumsau umumsao ummummz umumsao HmumSHU Ga Hmumdao
Ga mHOnmz ummsuumm CH mHOnmz ummummz oopsaocH mHOnmz
mHmNA<Z¢ mDOMO meBmHmB mom mmHmmZOHaéflmm mMBmDAUImmBZH .hm mqméfi
O O O O O O O “H .H
H¢.H .... ... ma .m
om.a Hm.a .... om. omlmmlwmlmm .0
m>.H NN.H 0H.H .... an. mmlamlma .m
H0.a 0m.o om.o om.o .... 00. hmlomlmalma .m
.H .m .0 .m .m mocmumfla umumdao HmumSHo CH “mumsao

mocmumfln HmumsHUnuoucH ommnm>¢

ImuucH mmmum>¢

oopsaocH muonmz

 

 

meNA¢Z¢

mbomw meBmHmB 20mm mHUZ€BmHD QHOmBZMUMMBZH UZHmD mmmBmDAU m0 ZOHBdSMOh

.mN.MAm<B

95

As a result of analyzing the female criterion groups
separately from the male criterion groups, greater discrimi-
nation among the majors was possible (Figures 2 and 3). In
the 27-group treatment, the greatest difference among groups
(function one) was a numerical/technical versus verbal/
humanistic Split, with the male criterion groups at one end
and all the female groups except mathematics at the other.
The l4-group analysis also identified a numerical-verbal
dichotomy as the primary difference, but those male majors
which at first had to be described as "less numerical/tech-
nical" were subsequently shown to be verbal/social science.
Similarly, the 13-group analysis emphasized the scientific,
numerical, and verbal measures in accounting for the major
portion of the variation among the female majors, but,
because the first function dealt with scientific/numerical
characteristics and the second function dealt with verbal
characteristics, it was possible to describe the female cri-
terion groups more exactly. Thus, the psychology women, who
appeared to be neither highly verbal/humanistic nor numerical/
technical in the 27-group analysis, were shown to be both

scientific/numerical and verbal in the lB-group analysis.

96

.mHmNA¢Z¢ mDOfiO ZmMBm30m ho mZOHBUZDh
BZflUHmHZOHm mmmmfi BmMHm ZO mQHomBZMU mDOMO m0 Foam AdZOHmZMEHD mmmme

 

 

 

 

 

 

 

 

 

 

 

.N MMDUHW

 

97

.mHmMQ<Z< m50m0 meBmHmB m0 mZOHBUZDm
BZflUHmHZUHm mmmmfi BmMHm ZO mQHomBZmU mbomw m0 Roam AdZOHmZMSHQ mmmmfi

 

 

.m MMDUHm

 

98

Classification of the.Cross—
Validation Samples

A maximum likelihood procedure, based on the results
of the discriminant analyses, was used to classify three 20
percent cross-validation samples. The significant discrim-
inant functions which resulted from a discriminant analysis
and the corresponding discriminant scores for each of the
students in the analysis sample were used to compute maximum
likelihood weights and a constant for each major group.
(The MSU Computer Institute for Social Science Research pro—
vided the computer programs used in the classification of
the cross-validation samples. Because these programs had to
be written for the present study, technical reports were not
available on either NORM CURV, the program which computed
maximum likelihood weights, or CLASS, the program which
calculated classification scores.) Discriminant scores were
then calculated for each of the students in the check sample.
The maximum likelihood weights and discriminant scores were
used to compute one classification score per major group for
each member of the sample. A student was assigned to the
major group for which he had the highest score. .An assign-
ment was termed a hit if the student was classified into his
final major, a near hit if he was classified into a major
within the same cluster as his final major, and a miss if he
was classified into a major outside the cluster which con—

tained his final major.

99

Results of the TwentyrSeven
Group Classification

The cross-validation sample for the 27-group classi-
fication contained 353 students, 181 men and 172 women. Ten
significant functions and the corresponding 10 discriminant
scores were used in developing the maximum likelihood
weights and classification scores. The results of the
classification are shown in Table 28.

0f the 353 students in the check sample, 62 (18 per—
cent) were assigned to the correct major (hits), a rate sig-
nificantly better than chance (p<<.001). This finding gave
additional support to the major research hypothesis of the
study. An additional 107 students were assigned to the
correct cluster (near hits). Combining hits and near hits
gave a 48 percent accuracy rate.

The classification results also supported the secon-
dary hypothesis of the study (that men and women in the same
major have substantially different patterns of characteris-
tics). The four major fields with both a male and a female
group are mathematics (9 and 17), history (10 and 24),
social science (11 and 25), and psychology (14 and 26).
These groups are circled in Table 28. Of the five mathe-
matics men assigned to women's majors, one was classified in
mathematics and, of the six mathematics women assigned to
men's fields, two were classified in mathematics. 0f the

seven psychology men assigned to women's majors, two were

100

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mo T: on SH 3 No mmm 3 NH 3 3 NH c 3 HN mN NH m 6H 3 HH 3 N HH 3 o N w NH N3 NH o NH 233.
mm m m OO 42 mm m m m H O O O H O O N N O O O O O O O O O O O O O O O O O ma 3
om U m OO (Z om m CH O m O O O O H O O O O O O O H O O O O O O O H H @H O Kid 0

8 3 N .N o H HH H . H . . . . . m H N H . H . . . . . . . . . . . . H N

om m OH H OH . . . H m . . c H . . . . . . . . . . . . . H . . . . NN

2 m NH H m . . H . H H H . . . . H m . . . . . . . . . . . . . . HN
v¢ ¢ OO o m N O O H O O O O m H O H I O O O O H O O O O O O O O O ON h

3 ON NN HH Hm m N N H . H HH o N. m H o N . H . . . . . . . N . . . . mH

mN N om v w H O O H H O O V O O O O O O O O O O O O O H O O O O O 0H

.1 oo o S H oH H N H N . . N H H . . . . . . . . . . . . . . . . . . .3

MV ON mm m OO o o O H H O H O O O O O O N m G O O O O O O O O O O H O O oN
HH H HH H w O O O H N O H O N H H O O O O O O G O O O O H O O O O m” m

ON N om m OH O O O O N O O AH H H O m H O O O O H O O O O O O O O O v~

m S m N m 3 . . H N H H . H . H H H m . H . H . H . . . . . . . . NN
cm N 3 N S H 3 H H . . . . . H H @ H . . H N . . H . H . . . H . . H «H a

O OO o vw v m O O O O O O O O H O O O O O ¢ O O H O O O H O O H H O m

no N 2. o cH N 3 . H . . . . . . H . H . . . . N N N H . . H . . H . H NH

NH H NH H m O O O O O O H H H O ® O O O H O H O O H O O O O H O O NH
2. o .. o S . H . . H . . . . N mu . . . N N . H H H . . H H . . HH 0

No m m H NH . . . . . . H . . . . H N . N H H H H H . . . H . . 0H

o N N .. o m . . . . . . . . H . . . . H H . H . H . . H H . . . H v

vm Vm Gm N. OO 0 MH O O H O O O O O H O O O I O N H H O O O O m m H O O O m
N... o N m 3 . . . . . . . H H . . . . H N . . N . . H m c H . H . m m

ON m mm m mH O I O O O O O N O O O O O O H O O O O O N O m H O A" O N

m o H N o 3 N . H . . . . H . H . . . . N . H . . H H . . o . H H H

mm 2 Hm m .. o 3 . Q . . . . . . H N H . . N . . H . H H . . . H . N m o
mm m OO o o O O O O O O O O O O O O H O O O O O H O O H N H O N NI 4

4 HN m ”H N *H O m O O O O O O O O O O O N O O O O O O O O H m n O N w
x 3330 z x z x z 3953 3 S N NN HN S 3 6H ...H 3 N «N N S m NH NH HH 3 c m m N H c N m .8330
...on nooz q 3H: 3H: uooz 33 NO x w m w o o 4 a Soho:
Hmuoa umumSHo uomuHQ oNHm ucgcmwmmd HoumsHU paw uOfimZ 40399:.

$322 oz... .251 32 .95: 05395 mm<o cacao zgmmumezmze mom 52m ons<BH<>um8mo mo 20353328 .mN 592.

lOl

assigned to psychology. For the remaining five groups, no
cases were assigned to the opposite-sex major group.

In general, men were more often misclassified into
women's majors than were women misclassified into men's
majors. Forty-three men were assigned to women's majors,
while only 20 women were assigned to men's majors. The most
"popular" women's majors for men were home economics teach-
ing (eight men assigned), mathematics (eight men assigned),
psychology (seven men assigned), and retailing of textiles
and clothing (six men assigned). Five women were assigned
to accounting, the only men's group to receive a substantial
number of women. At least one member of each male group was
assigned to a female group; the extreme example was psychol-
ogy, in which half the men were assigned to women's majors.
Three of the female groups (home economics teaching, speech,
and nursing) had no members assigned to male groups. Over
half of the mathematics women, however, were assigned to
male majors. Apparently, psychology men have important
characteristics in common with a number of female majors,
and mathematics women share important characteristics with
a number of male majors.

Results of the Fourteen-Group
Classification

Five significant functions and the corresponding

five discriminant scores from the fourteen-group discrim-

inant analysis formed the basis for a maximum likelihood

102

classification of the 181 men in the cross-validation sample.
Twenty percent of the men were correctly assigned to their
final majors and, when hits were combined with near hits, 52
percent of the check sample fell into the correct clusters
(Table 29). This was an increase of 5 percent and 9 percent
respectively, when compared to the accuracy achieved in
classifying the men in the 27-group case. As to where the
increase came from, the 43 men who, in the 27-group classi-
fication, had been assigned to female groups, were now forced
into male groups. Twenty-five were still misses, but ten of
them were assigned to the correct majors and eight were put
into the correct clusters. In general, it was these addi—
tional hits and near hits which accounted for the increase

in accuracy.

Sixty percent of the twenty-five misses were now
found in Cluster D', nine in the psychology group and six in
zoology, a second indication that the characteristics asso-
ciated with the male psychology majors in this study are
highly compatible with the characteristics of a variety of

female majors.

Results of the Thirteen-Group
Analysis

The cross-validation sample contained 172 women who
were classified on the basis of the five significant func-
tions produced by the l3-group discriminant analysis. Forty-

four assignments were direct hits (26 percent), an increase

103

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Nm mm Nm mm 0N hm HmH mH 0N 0H HH m HH NH m m MN NH mH HH mHmuoa
we HH HN m HN m «H m m H .. N .. .. .. .. .. H m H eH a
.Q NN N mm m w N m 00 00 In 00 00 H CI 00 N H O m -
Ho em H5 0H N H eH H .. H m m m H .. .. .. .. H . NH
hm m mN N m .. H H N .. H H H .. .. .. H . NH
me o b H «H v .. N H H H N .. N .. H . . HH .0
Ne m NH N NH N H m H .. N H .. H .. .. H . 0H
.0 hm m NH H m .. H .. H .. N H .. H H .. . . ¢
hm mm «m h mH N mH .. N H H .. .. .. N m m H . . m
mm 0H .. o mH N .. .. H N H H v .. m H . N m m
OH H 50 OH mH to .H O. .0 H no no 00 H OH on o N N .
.m NH N mN v 0H N e .. .. .. .. .. .. .. N v . N H
mm eH mH m mN w oH N .. H H .. H N .. .. .. N e N m
.0 0 HH H m 0. H O. O. I. 0. m C. O. H m 0 H h -¢
.¢ HN m HN m «H H m .. .. .. .. H .. .. H N N H o
x umumDHo z x z x z onEmm @H m mH NH HH 0H v m m N H m h umumsHU
3.5 582 can. 33 35 ~82 33 mo ..n b .m i .o a 3.3:
HauOB HoumsHu . uomuHQ mNHm unmecmHmm¢ umumsHo can u0nmz Hmsuo<
mummHz 02¢ .mBHm mdmz .meHm cszomm mmdo abomwuzmmemaom mom mqmzcm onadoHH¢>nmmomo mo ona<onHmm¢Ho .mN mqmda

104

of 6 percent over the 27-group result. Accuracy of the
cluster assignments was not directly comparable because, as
described earlier in the chapter, the clusters formed from
the 27-group analysis were not identical to the clusters
formed from the l3-group analysis. The latter analysis pro-
duced five clusters instead of four. Even with the more
rigorous test of the discriminatory power of the analysis,
however, accuracy increased to 55 percent correct cluster
assignments (Table 30).

Twenty women had been assigned to men‘s majors in
the 27-group classification. Nine of them were direct hits
in the 13-group classification and three were assigned to
the correct clusters. Seventy-five percent of the misses
were found in the mathematics group, and this, taken with
the fact that 90 percent of the mathematics women were di-
rect hits after the male groups were eliminated, provides
further support for the conclusion that mathematics women
share more important characteristics with the men in the

study than the women.

Summary

The major hypothesis of the study was that undergrad-
uate major fields can be differentiated on the basis of
selected characteristics of their students. Results of the
27-group discriminant analysis permitted rejection of the

null hypothesis:

105

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

mm mm Om Hm ON ev NNH eN HH NH HH m mH m N HN OH N mH NN mHmuoe
00 .H m I I ‘2 om 0 OH 0 I I I I H I I I I I I I I I I I I I I I I I I NIH .H
mm .m N I I <2 mm N m N N H I I I I I I I I I I H I I I I H N m‘H am
am mN «a O NN N m N .. N .. H m .. H .. .. .. .. .. ON
HH H NN N m H .. H N .. .. H H H .. .. .. N mN o
Om m ON N OH .. .. N H N .. H .. m H .. .. .. ON .
.0 ON m mm O OH N .. N N H O N .. N .. .. H .. NN
Om NH ON O OH H OH H N .. .. .. N H .. v .. .. .. .. NN
I I o NH H m I I I I I I N I I m I I H I I N I I I I I I HN .m
.m mN N NO m m .. .. .. .. .. .. H H m .. H .. .. OH
Nm we NN m m H HH H H N H N .. .. .. .. H H .. N NN
mm m II o 0 II m H I I II II I I II I I H II II q ON m
ON ON NN HH Hm m N N N N H .. H m O .. HH HH OH _
.m OO O ON N OH H H .. .. .. .. .. N .. N .. N N mH
x noumsHo z x z x z mHmsmm NH OH ON mN ON ON NN HN OH NN ON OH OH “mumsHu
muHm ummz wan muHm muHm ummz muHm m0 .H .m .o . .m .m w uoflmz
Hmuoe umumDHU . uoman mNHm ucmEcmHmm< HmumSHU 0cm Mona: Hmzuom
mmmmHz O24 .meHm mamz .meHm oszomm mmmo msomouzmmemHma mom mHNZNm onecoHH4>nmmomo mo onecuHmHmm<qo .Om mHmNO

106

No differences exist among undergraduate

fields on the basis of the abilities, interests,

educational-vocational preferences, or family

and high school backgrounds of their student

members.
It was, therefore, concluded that undergraduate major fields
were different with respect to the kinds of students who
persist in them.

A multivariate discriminant analysis of 47 variables
for 27 major-sex groups produced 26 functions, 10 of which
were significant beyond the .001 level. The first five
significant functions accounted for 79 percent of the vari-
ation among groups and were analyzed in detail.

There was a marked separation of majors into two
distinct groupings on the first function, a verbal/humanis-
tic versus numerical/technical Split. The verbal/humanistic
grouping contained only female majors; the numerical/techni-
cal grouping contained all the male majors and the mathemat-
ics women. The differences identified by function one
accounted for 51 percent of the total variation among groups
and appeared to be heavily influenced by the sex differences
measured by the variables.

The high-weight variables in function two were
interpreted as generalist to specialist in nature and spread
the majors along the continuum in gradual steps. Mathemat-
ics, engineering, and nursing were closest to the specialist

end of the scale and political science, history men, and pre-

law were at the extreme generalist end.

107

Function three identified a set of variables which
appeared to differentiate between abstract fields and
applied fields.

The fourth and fifth functions were difficult to
interpret. They each accounted for about 4 percent of the
trace, and although they were significant, they were greatly
overshadowed by the first three functions.

The combined effects of the first three functions
were demonstrated in a three—dimensional plot of group cen-
troids. Common characteristics cut across college lines.
Majors fell into more or less distinct clusters which shared
similar interest and ability patterns. The greatest differ-
ences, of course, were the sex-related ones defined by the
first function. This finding led to the decision to run
separate discriminant analyses on the 14 male groups and the
13 female groups.

The general effect of separating the sexes was to
clarify the nature of the differences among major fields.
I?or example, all the male majors were originally described
:33 numerical/technical and all the female majors (except
mathematics) were described as verbal/humanistic. The
separate analyses positioned all majors on both qualities.

To test the validity of the discriminant analyses,
discriminant weights and scores were used to compute maximum
likelihood weights from which cross-validation samples were

Classified. Separate classifications were performed for the

108

27-, 14-, and lB-group cases. Results in each case were
significantly better than chance.

A secondary hypothesis of the study was that male
and female students in the same major had substantially
different patterns of characteristics. No direct statis-
tical test of this hypothesis was possible. Descriptive
information was available, however, and it was proposed that,
if men and women in the same major were substantially the
same, their groups would fall close together in the discrim-
inant space defined by the 27-group analysis. In fact, none
of the four pairs (mathematics, history, social science, and
psychology) were together in discriminant space.

In Chapter V, a brief description of each of the 27

major groups is provided.

CHAPTER V

SUMMARY DESCRIPTIONS OF THE

TWENTY-SEVEN MAJOR GROUPS

In this chapter, the raw data and the results of the
appropriate discriminant analyses and maximum likelihood
classifications are drawn together in a descriptive summary
for each major group. Specifically mentioned are mean scores
on variables (Appendix F), discriminant analysis and classi-
fication outcomes cited in Chapter IV, intercentroid dis-
tances and cluster formations (Appendices C, D, and E), and
frequency distributions of raw data not previously included.

A few of the variables had distributions which demon-
strated sex-related trends. Women's majors generally scored
higher on the QQTfVerbal and MSU Reading tests than men's
groups. Women also tended to give higher ratings than did
men to the following: literary classics, biography, and
history books; fine and applied arts and letters courses in
high school; social service, artistic, and biological science
occupations; and the importance of family influence and the
social aspects of a college education in their decision to
attend college. Parents of women in the sample generally had

higher mean educational levels than did parents of men in the

109

110

sample. Men's majors generally scored higher on the QQT
Information and Numerical tests than women's groups. They
also tended to give higher ratings than did women to the
following: technical and mystery books; high school science
and mathematics courses; physical science and mechanical-
technical occupations; and the importance of the prestige of

a college education in their decision to attend college.

The Majors

1. Packaging. Men in this major scored somewhat

 

below the mean for men on the ability/achievement measures
and tended to rate themselves even lower. They preferred
science, mathematics, and nonacademic subjects in high
school and were more likely than most groups to read tech-
nical books for pleasure. Highest occupational preference
was given to physical science and mechanical-technical pro-
fessions; they tended to turn away from social service and
verbal-linguistic occupations. Their parents had less edu-
cation than most of the other major groups and father's
occupational level was lower. Most of them were in-state
students who planned to be partially self-supporting during
college. They were more certain of their vocational choice
prior to entering college than many groups, but a majority
of them changed majors at least once before graduating.

They were least likely of all groups to come to college for

lll

educational reasons and most likely to come to college for
the status and prestige of a college education.

This combination of characteristics placed packaging
in the cluster of majors which included accounting, general
business, and marketing, although they were also close to
the engineering-mathematics cluster in discriminant space.
Their interests and occupational preferences were similar to
the latter cluster, but their scores on the ability/achieve-
ment measures and their self-ratings of ability were consid-
erably lower. They appeared to be a group whose occupational

aspirations outdistanced their abilities.

2. Accounting. Accounting majors scored below
average on the CQT tests and MSU English and Reading tests
and their self-ratings reflected their scores except for
numerical reasoning, where they overrated their ability.
They liked social studies and nonacademics in high school.
Business detail occupations were very popular with these
men; executive and managerial occupations ranked second. Of
least interest were artistic occupations. Their parents had
the lowest educational level of all groups and father's
occupational level was also the lowest in the study. They
came primarily from.Michigan high schools and eXpected to
work for a part of their college expenses.

The accounting majors most resembled packaging,
general business, and marketing majors, and were somewhat

close to zoology and mathematics men in discriminant space.

112

Nevertheless, they were different enough from all other

groups to produce classification rates of 53 and 67 percent.

3. General businessgagministration. Although their
test scores were very low in comparison with other male
groups, the general business administration majors rated
their abilities as average or slightly above. Their occu-
pational preferences were varied, although executive—manage-
rial and business detail occupations were selected by more
of the group than any others. They had no strong likes or
dislikes among high school subjects, but showed a slight
preference for social studies. Mean educational level of
father was the highest of all male majors; mean level of
mother was slightly lower. Father's occupational level was
second highest of all groups, male and female. A high pro-
portion of these men were in-state students. They were un-
certain about choice of major and vocation prior to college
entrance, and were more likely to be influenced in these
decisions by family and significant others than most of the
male groups.

The general business administration major was charac-
terized in the discriminant analysis as generalist-applied.
The members of this major appeared to be diverse and heterge-
neous in their characteristics, suggesting that they might
fit comfortably into any of several environments. The

results of the classification supported this view. General

113

business males were assigned to all clusters and to almost

all majors in the l4-group case.

4. Economics. Economics majors scored above the
mean on all five ability measures and rated themselves above
average in all categories. They preferred physical science
and executive-managerial occupations. Their parents had
higher than average educational levels. The father's occu-
pational level was high and in general, the mother did not
work outside the home. A substantial minority of the group
came from out-of-state high schools, a higher proportion
than for any other male group. These men came to college
primarily for educational reasons, but they were also moti-
vated by social and status needs.

Economics was the only major from the College of
Business which fell outside the "business" cluster in dis-
criminant space. It was closer to the male majors from the
College of Social Science than to the majors from the Col-
lege of Business. This was apparently due to their consis-
tently higher abilities and greater interest in abstract

concepts.

5. Marketing. Marketing majors, like general busi—
ness administration majors, had very low mean scores on the
orientation tests, but rated their abilities as average and
above. They tended to come from larger high schools in

Michigan; they liked nonacademic subjects best and arts and

114

letters least. As a group, they were uncertain about a
vocational choice. This uncertainty was reflected in the
range of occupational interest expressed by them. Most
popular were occupations of the executive-managerial type,
but seven of the remaining eight occupational groups were
selected by at least 20 percent of these men. Only artistic
occupations were unpopular. They viewed a college education
as important for its own sake, but, like packaging majors,
they were more likely to value a college education for
prestige and status reasons than most groups. A majority of
their parents had continued their education beyond high
school. Fathers were likely to be in supervisory or manage-
rial positions and mothers were generally not employed out-
side the home. Their major source of financial support for
college was their parents.

Marketing majors were most like general business
administration majors. In both the 27-group discriminant
analysis and the 14-group analysis, these two groups remained
closer together in discriminant space than any other two
major groups in the study. In both groups, there were few
strong distinguishing characteristics, and consequently, it
was difficult to classify them into the correct major.

Assignment to the right cluster was much more likely.

6. .Electrical engineering. Electrical engineers,
of all male groups, had the highest mean scores on the CQT

Verbal and Information Tests and the MSU Reading Test and

115

were second on QQT Numerical and MSU English. Their self-
ratings of ability were lower than one—third of the male
groups. Science and mathematics were highly favored sub-
jects in high school. .Almost 40 percent of the group came
from out-of-state. They were generally sure of their major
and vocational choices and few changed majors during their
four years. All of the men in this group selected physical
science occupations; it was the first choice of all but
three of them. Mechanical-technical and biological science
occupations were their next preferences. Educational level
of parents was about average for the male groups. Father's
occupational level was somewhat below the male groups' aver-
age. The mother tended not to work outside the home.
Electrical engineers were characterized in the dis-
criminant analyses as strongly attracted to the physical
sciences and mathematics with specialized, abstract inter-
ests, placing them close to mathematics men and mechanical

engineers.

7. Mechanical engineers. This group scored well
above average on the QQT Numerical, above average on the QQT_
Information, and about average on the other three tests.
Their self-ratings were realistic. Just as with electrical
engineers, the mechanical engineers selected the physical
science occupations first and mechanical-technical occupa-
tions second, but more of the latter group selected execu-

tive-managerial occupations as their third choice. They

116

were very sure that engineering was the correct major and
vocation, decisions reached with the help of family and high
school eXperiences. These men had very positive feelings
about their high school experiences and, although science
and mathematics were their favorite high school subjects,
they tended to enjoy the other academic areas too. Three-
quarters of them came from Michigan high schools. A college
education was valued by this group as a way of getting
desired prestige and position.

The discriminant analyses showed these men to be
physical science-numerical-technical in interests and abil-
ities. Their interests were specialized and they were more
concerned with application of their specialty than with the
theoretical aspects of it. Thus, they appeared to fall
between electrical engineers and packaging majors, albeit,

closer to electrical engineers.

8. 2001992. On the COT and MSU tests, this group
generally scored above the mean for men's groups. Their

strongest showing was on the CQT Numerical, where they

 

ranked fifth out of fourteen and, in rating their abilities,
they viewed numerical reasoning as their strongest area. In
high school, science and mathematics were their favorite
courses; nonacademic subjects were not popular with this
group. Two-thirds of them were Michigan residents. Of all
the male groups, zoology majors most valued a college educa—

tion for its own sake. They were primarily interested in

117

the biological science occupations (many of them had started
in the pre-medical major and subsequently transferred to
zoology) and secondarily in the physical science occupations,
but were somewhat uncertain of a specific vocational choice.
Over half the group had college-educated fathers who held
executive-managerial positions or were in the professions.
.Zoology majors did not closely resemble any other
major. In the 27-group analysis, they were closest to psy-
chology men in discriminant Space, but even this was a dis—

tant relationship.

9. Mathematics. Mathematics men scored higher than
the other male groups on the QQT Ngmerical and the Egg
English tests; on the remaining three orientation tests,
they were well above average. Self-ratings of general;
capacity, numerical reasoning, and anticipated grade-point
average were also the highest of the male groups. They were
fairly certain of a major choice before entering college,
and over half of the group persisted in that choice. Phys-
ical science occupations were strongly preferred, followed
by mechanical-technical and business detail occupations.
Business contact work was most strongly rejected. In high
school, they enjoyed mathematics and science and read tech—
nical books, but did not like the nonacademic subjects.

They came to college primarily for educational—vocational
reasons. Parents' educational levels were about average;

father's occupational level was below average.

118

Over all, mathematics men and engineers were very
much alike, falling into a close-knit cluster in discrim-
inant space. Forty-four percent of the classification

sample was assigned to this cluster.

10. History. The history men scored high on the
CQT Verbal and Information tests, low on the QQT Numerical,
and slightly above average on.MSU Reading and English. Self-
ratings of ability tended to match these positions. As
eXpected, they favored social studies in high school and
were most likely to read biography and history books for
pleasure. Unlike most male groups in the study, they en-
joyed arts and letters in high school. No one occupational
preference group was strongly favored by history men. Their
choices were distributed over a number of groups. Least
popular with these men were artistic, mechanical—technical,
business detail, and physical science occupations. Most
popular were social service, verbal-linguistic, and business
contact occupations. Parents' educational level was about
average and father's occupational level also tended toward
the mean. Almost half of the mothers were working ouside
the home. Twenty percent of these men expected to be fully
self-supporting in college.

History men were the only male major from the Col-
lege of Arts and Letters included in the study. They
closely resembled political science majors. They were char-

acterized in the discriminant analyses as highly verbal with

119

abstract, general interests. From the classification results,
it appeared that these men could fit comfortably into any of
the four social science majors, but particularly into polit-

ical science.

11. Social science. Social science men had the
lowest mean score on the QQT Numerical of all male majors
and were below average on the remaining four orientation
tests. Their scores were consistently lower than the other
three male groups from the College of Social Science. Self-
ratings of ability were quite realistic. They appeared to
have no strong educational-vocational preferences.' In high
school, they liked all subject areas. They were uncertain
about selecting a major in college; all but two of them
changed majors at least once. There was no consensus of
opinion on occupational group preferences and the choices
they made were viewed as tentative. .Educational level of
parents was above average, although father's occupational
level was close to the mean for male majors.

In the 27-group analysis, these characteristics
placed social science men in the center of a large swarm of
majors in such a way that as a group, they were not close to
any one major.

.Aside from their position on function one of the 14-
group analysis, where they were clearly characterized as
verbal-social science, this group never came to the surface.

The assignment of the cross-validation sample of social

120

science men suggested that the group had a general identity,

but not a Specific one.

12. Pre-law. Social science pre-law majors consis-
tently overrated their abilities. Group scores on the Egg
Verbal and MSU Reading tests were higher than two-thirds of
the men's majors, but the pre-law group rated itself higher
on these two areas than did any other male group. They
recognized that numerical reasoning was their weakest area,
but again, they tended to be optimistic. They enjoyed all
high school subjects taken, especially social studies.
Almost 40 percent of the group were out-of-state students.
One-third of these men selected verbal-linguistic occupa-
tions as most suitable and another one-third selected phys-
ical science occupations. Among male groups, they had the
most highly educated parents; 62 percent of the fathers and
52 percent of the mothers had at least a bachelor's degree.
Almost two-thirds of the fathers were either in executive—
managerial positions or in the professions. Few pre-law
majors planned to work for college expenses during the school
year. They were more likely than other male groups to view
a college education as a social necessity and as eXpected by
the family.

Pre—law men were characterized in the discriminant
analyses as verbal-social science in their orientation and
generalist and abstract in their interests. They were most

Similar to political science majors. In the 27-group

121

classification, half of the pre-law check sample was
assigned to women's majors. In the l4-group classification,
one of these was correctly assigned to pre-law and two were
placed in the correct cluster, giving them 62 percent hits

and near hits.

13. Political Science. Political science majors
had average or above average mean scores on all orientation
tests except the QQT Numerical, on which they had the second
lowest mean score among the male groups. In general, self-
ratings of ability appeared consistent with scores. High
school social studies courses were most popular with these
men and science and mathematics courses were least popular.
Forty-one percent of the group was from out-of-state, the
highest proportion of out—of-state students in any of the
male majors. Political science also had the highest propor-
tion of students who expected to be either entirely self—
supporting or primarily self-supporting through college.
Father's educational level was somewhat below average,
although occupational level was average. Fifty-five percent
of the mothers were employed outside the home, the highest
proportion for any major in the study. Political science
majors Showed preferences for verbal-linguistic, business
contact, executive-managerial, and social service occupa-
tions. They came to college for educational-vocational

reasons with little family influence in the decision.

122

This group was characterized as highly generalist
and abstract in interests and strongly verbal-social science
in orientation. Their position in discriminant space was
very close to history, suggesting a high degree of overlap
between these two groups. .Classification results also

pointed to overlapping groups.

14. Psychology. Psychology men had high mean
scores on the orientation tests and realistic self-ratings
of ability. ~AS a group, they out-scored the other three
social science majors. Although they liked all their high
school courses more than they disliked them, they had no one
favorite subject matter area. They came from large high
schools. Thirty-six percent were out-of-state students.
Physical and biological science occupations were selected
most often by these men, while at the same time, they
rejected mechanical-technical occupations. Social service
was the next most popular choice. They said they were un-
certain of their first major choice and all but five of them
changed majors at least once. Father's educational and occu-
pational levels were above average. Mother's educational
level was slightly below average. Forty-two percent of the
mothers were employed. Although psychology men valued the
college experience for the educational benefits, the social
aspects of campus life were also important to them.

Psychology men had both numerical-physical science

abilities and interests and verbal-social science abilities

H

‘, .‘

123

and interests. The discriminant analyses located them mid-
way between the extreme majors at each pole. In both the
27— and l4-group analyses, they were placed at the extreme
abstract end of the applied-abstract function. They-were
more Specialist than generalist. This unique combination
of characteristics separated them from close association
with any other major. Even zoology, with which they were
clustered, was a distant partner. In the 27-group classi—
fication, half the men in the cross-validation sample were

assigned to women's majors.

15. figmefeconomics teaching. Their mean scores on
four of the orientation tests were about average for female
majors, but on the MSU English test they were lower than the
other women's groups. They rated themselves as average or
above in all categories and expected to maintain at least
a 3.00 grade point average. Nonacademic subjects (which in-
cluded home economics) were most popular with these women in
high school, followed by social studies. They came primar-
ily from very small Michigan high schools. Twenty-seven per-
cent of them, the highest percentage in the study, lived on
farms. Occupational preferences centered strongly on the
social service professions; only two women failed to select
this group. The biological science occupations ran a poor
second. Parents of these women were least likely to have a
college education; fathers"occupational level was also low-

est of all female groups. Forty-eight percent of the students

124

expected to provide most or all of their own financial sup—
port for college. The family was seen, however, as an impor-
tant influence in their decision to come to college and

their choice of vocation. The status value of a college
education was also viewed as important.

In general, this major contained students who were
neither highly verbal nor highly numerical in interests and
abilities. They were strongly service oriented and conven-
tional, resembling special and elementary education more
than other majors. Sixty percent of the cross-validation
sample of home economics teaching majors were assigned to

the cluster composed of these three majors.

l6. Retailing of textiles and clothing. Except for
an average mean score on the CQT Verbal, they had low orien-
tation test scores. Self-ratings of ability were all above
average, but they did not expect to maintain much more than
a 2.50 grade point average. Retailing majors were predom-
inantly concerned with business and artistic interests in
college, but in high school they preferred social studies
courses and disliked science and mathematics. They were
less interested in social service occupations than most
women, but more interested in artistic and business contact
occupations. Father's educational level was about average,
while mother's was below average. One-fourth of the fathers
were Skilled laborers. These women came to college for edu-

cational and vocational advantages.

125

Although more verbally oriented than numerical,
retailing majors were not strongly verbal. Business contact
activities coupled with a relatively weak interest in social
service suggested occupations that would deal with people at
some distance. They were not really close to any major in
discriminant Space, although they formed a loose cluster
with Speech and art practice majors in the l3-group analysis.
The cross-validation sample of retailing women was assigned
more accurately in the 27-group case than any other female
group and only mathematics women had a better hit rate in

the l3-group classification.

17. Mathematics. Mathematics women were unique
among the women's majors. On the orientation tests, they
outscored all female groups on the QQT Information test, all
male and female groups on the MSU Reading and English tests,
and all but electrical engineers and mathematics men on the
QQT Numerical test. On the.CQT Verbal, they were well above
the over-all mean. Self-ratings of general capacity and
numerical reasoning reflected their actual abilities, but
they down-graded their real position on reading skill. They
anticipated at least a 3.00 grade point average through col-
lege. Twenty-nine percent came from out-of-state high
schools. As a group, they enjoyed high school courses;
mathematics and science courses were strongly favored, of
course. They came to college to pursue an education, less

influenced by family expectations and more by their high

it

126

school experiences than other women. They were certain of
a major choice prior to college entrance. Sixty-nine per-
cent of those who graduated in mathematics had originally
entered in mathematics. The overwhelming choice of occupa-
tional group was physical science. Unlike the male majors
whose first choice was physical science, the mathematics
women rejected mechanical-technical occupations. Social
service and biological science professions ranked second
and third.

Father's educational and occupational levels were
above average; mother's educational level was below average.

Mathematics women occupied a Singular position in
discriminant Space. They resembled no other major of either
sex. Being highly numerical-scientific in orientation drew
them away from the women's groups and the verbal-social sci-
ence men's groups. Rejection of mechanical-technical inter-
ests kept them out of the numerical-SCientific-technical
men's groups. In the 27—group classification, they were
more likely to be misassigned to male majors than to female
majors. In the 13-group case, 90 percent of the sample was

correctly classified.

18. Nursing. Nursing majors scored above the mean
on the numerical and reading tests, approximately at the
mean on the information test, and below the mean on the
verbal and English tests. In high school, they liked arts

and letters most and social studies least. Biological

bl

127

sciences were enjoyable, but physical sciences and mathe-
matics were not. Ninety-one percent of these women were
in-state students. Father's educational level was about
average; mother's educational level was above the mean for
women. Thirty—nine percent of the mothers worked outside
the home. Nursing majors expressed few doubts about major
field or vocational choice. Seventy-four percent of these
graduates started in the nursing major. As expected, their
occupational preference group was biological science.
Nurses were characterized as moderately scientific-
numerical and nonverbal in nature with Specialized and
applied interests. In both discriminant analyses, nursing
was too far from the other women's groups to be included in
a cluster. Of the majors in the l3-group analysis, Special
education and social work were the Shortest distance away.
The cross—validation sample had 33 percent direct hits in
the 27-group case and 22 percent direct hits in the l3-group

case .

19. Elementary educatigg. The elementary education
group had 253 members, 51 percent of whom had originally
chosen elementary education. The performance of the group
was consistently low in comparison with other female majors.
Their verbal and reading scores ranked thirteenth out of
thirteen. Nevertheless, they rated themselves as average in
each area. They came from large high schools. Their favor-

ite subjects were in the arts and letters area. Ninety

128

percent were in-state students. Both parents had higher
educational levels than average; father's occupational level
was at the mean. Although they came to college for educa-
tional-vocational reasons, campus life and the attendant
social eXperiences were more important than for many majors.
Social service occupations were preferred by most of these
women. Seventy—five percent of the group was certain that
teaching was the proper vocation for them. They were more
strongly influenced by family in this choice than any other
group.

Elementary education majors were characterized as
conventional-feminine. They were nonscience, nonnumerical,
and more verbal than nonverbal in their interests and abil—
ities. In discriminant space, they fell near social work
majors and special education majors. They appeared to be
a loosely-knit hetergeneous group which overlapped a number
of women's groups. Results of the classification supported
this view. Members of the cross-validation sample were
assigned to every female major except speech. Elementary
education was a major which accommodated a variety of women

with ease.

20. Special educatigg, Their performance on orien-
tation tests was about average in all areas. Ratings of
ability Showed that Special education majors viewed them-
selves as average in general ability, verbal reasoning, and

reading skill and below average in numerical reasoning.

129

They eXpected to maintain at least a B- average in college.
Two-thirds of them changed majors at least once during their
four years. Although they enjoyed their high school courses,
as a group they were not strongly committed to an academic
area. Their common commitment was to service, apparently.
In making occupational preference selections, they consis-
tently chose the social service occupational group, but
Spread their remaining choices over the other areas, reject-
ing only mechanical-technical occupations. Father's mean
educational and occupational levels were Slightly below
average, while mother's educational level was above average.
Forty—one percent of the mothers were employed outside the
home. Special education majors were more likely to feel
that they had been influenced in their choice of profession
by personal eXperienceS outside of home and school than were
other women, and yet they saw family influence as important
in the decision to come to college.

It was difficult to characterize special education
women on the basis of their position on the discriminant
functions, for they remained buried among other groups near
the middle of the scale. They were closest to social work
and elementary education majors in the discriminant Space.
When the cross-validation sample was classified, no direct
hits were scored in either the 27-group or the l3—group case.
It appeared, in fact, that the "typical" Special education

major was a diSplaced home economics teaching major.

 

130

21. Speech. Speech majors had low orientation test
scores in all areas except English, where they were average.
They thought of themselves as above average in verbal and
reading ability, however, and at least average in general
capacity. They disliked mathematics and science in high
school and particularly enjoyed arts and letters. They had
a strong reading preference for novels and literary classics.
Ninety-five percent of the group attended Michigan high
schools. They were more likely to emphasize the educational
value of a college degree than most major groups. Father's
educational level was about at the mean, while mother's edu-
cational level was below the mean. Twenty percent of the
fathers were engaged in office, clerical, and sales work.
Seventy-five percent of the mothers did not work. Speech
majors were less interested in social service occupations
than most women, although over half of them selected this
group. The next most popular areas were the artistic occu-
pations and the verbal-linguistic occupations.

In the 27-group analysis, speech was characterized
as highly verbal-humanistic and artistic—technical in orien—
tation. Theirs was also an applied rather than abstract
group. In the l3-group analysis, it appeared that the
"highly-verbal" label had resulted from their strongly non-
scientific/nonnumerical character. The group was near the
middle of the verbal function. The artistic quality held up

in the l3-group case. The members of the cross-validation

 

131

sample were more likely to be assigned to the English group

than any other.

22. Art practice. Art practice majors did not do
well on the orientation tests and they knew it. AS a group
they expected to maintain about a 2.50 grade point average
in college. Throughout high school, they enjoyed the fine
arts, disliked mathematics and science, and tolerated the
other subjects. They attended larger high schools than most
of the groups in the study. Sixty—five percent of them
entered college in the art practice major. Their vocational
choices were based on personal needs and interests and only
one woman in the group failed to select artistic occupations.
Their only other consistent choice was, of course, social
service occupations. Parents' educational levels were well
below average. Twenty-four percent of the fathers and 18
percent of the mothers had not completed high school. One-
fourth of the men were Skilled laborers; 10 percent were
unskilled laborers. Mean occupational level of fathers was
well below the average for women's groups.

Compared to other female groups, art practice majors
were least scientific—numerical in character and about at
the mid-point on the verbal distribution. They had Special-
ized interests, preferred application to abstraction, and
were highly artistic. They were closest to speech majors
in discriminant space and when classified, the cross-valida-

tion sample appeared more like retailing of textiles and

132

clothing majors. The individuals in the analysis sample,

however, looked very much like art practice majors.

23. English. English majors outscored all other
major groups on the CQT Verbal test and were well above the
women's means on the remaining four tests. Their self-
ratings were high except for numerical reasoning on which
they considered themselves average. They anticipated a GPA
of at least 3.00. English women, more than any other major
group, favored arts and letters in high school, although
they also liked social studies. Science and mathematics
courses were rated low and nonacademic subjects they neither
liked nor disliked. They came primarily from.Michigan high
schools. Their parents were above average on educational
level and fathers were also above average on occupational
level. Fifty-seven percent of these men were either in
executive-managerial positions or in the professions.
English majors chose social service occupations first and
verbal-linguistic occupations second.

The discriminant analyses Showed them to be highly
verbal-abstract with low scientific—numerical interests.
They were closest to history and social science women in
discriminant space. The cross-validation sample had about

30 percent direct hits.

 

133

24. History. History women were above the female
mean on all five orientation tests. Their best performance
was on QQT Information; only mathematics women did better.
As expected, their favorite high school subjects were in the
social studies and more of them read history and biography
books for pleasure than any other group. Two occupational
preference groups received most of their attention, social
service and verbal-linguistic. None of these women selected
mechanical-technical occupations. Parents' educational
level was interesting. Fifty percent of the fathers had at
least a bachelor's degree, which was higher than the major-
ity of women's groups, but 27 percent had not completed high
school, which was more than any other women's major. Only
10 percent of the mothers had not completed high school.
Although father's mean educational level was below average,
mean occupational level was above average. Forty-two per-
cent of the mothers worked (the highest percentage among the
female majors) which may be partially explained by the fact
that 17 percent of the history majors came from families
with only one parent in the home. Fewer history women
expected strong financial support from parents than did
other groups. They came to college first for an education
but the social contacts and campus life were given more
weight by these women than by any other major.

The plot of group centroids for history women Showed

them to be more verbal than numerical-scientific in general

134

orientation, abstract rather than applied and.more general—
ists than specialists. They were closest to English majors
in discriminant space. The cross-validation sample was

classified 25 percent direct hits and 25 percent near hits.

25. Social science. Social science women ranked
third among the women's majors on the orientation tests and
rated themselves accordingly. They expected to maintain at
least a B average in college. Thirty percent of the group
came from out-of-state high schools. Social science women
preferred arts and letters courses in high school and were
much more likely to read literary classics than biography
and history for pleasure. Parents' mean educational level
was high; 62 percent of their fathers and 48 percent of
their mothers had at least a bachelor's degree and only 6
percent had not completed high school. Father's occupational
level was also above average. Nevertheless, 45 percent of
these women did not expect to receive strong financial sup—
port from their families. They came to college uncertain
about a major and a vocational choice. Ninety-one percent
changed majors at least once before graduating in social
science. Three occupational groups drew their interest:
social service, verbal-linguistic, and business contact.
None selected mechanical-technical occupations. They came
to college for educational and social reasons.

The group was characterized by strong verbal—

abstract interests. They were more service oriented than

135

creative-artistic in occupational outlook. They were most

closely associated with English majors in discriminant space.

26. Psychology. Psychology women ranked second
among female majors on the CQT Verbal and Numerical tests
and the MSU Reading test and fourth on the two remaining
tests. Self-ratings were consistent with performance. This
was the only group of the 27 to contain a majority of out-of-
state students. They favored arts and letters in high
school but liked science and mathematics courses better than
most women's groups and particularly the other two social
science majors. Their reading preferences were atypical;
they were the only group that did much reading of technical
books or mystery books. Social service occupations and
biological science occupations interested them most. Like
mathematics women, they were not interested in verbal-lin-
guistic professions. Parents' mean educational levels were
above average and father's occupational mean was high.
Sixty-four percent of the men were in the professions or in
executive-managerial positions. Their reasons for coming to
college were strongly centered on educational—vocational
needs. Family influence played more of a role than in most
groups and the status advantages of a college education were
less important than for any of the other 26 groups.

Psychology women were not really close to any group
in discriminant Space, but English majors seemed to share

more common characteristics than the others. Psychology

{’7

136

majors were high on scientific interests, numerical ability,
and verbal ability, but low on verbal interests. They were

also characterized as highly abstract and fairly Specialized
in their interests. In the lB-group classification, two-

thirds of the check sample were assigned to the correct

 

 

cluster, but only 22 percent were direct hits. '3
27. Social work. Social work majors scored lower

on the orientation tests than two-thirds of the women's i

majors. They rated themselves as about average in relation 1

to other college students, except in numerical ability where
they said they were somewhat below average. Most of the
group was neutral about science and mathematics courses in
high school and enjoyed social studies. Vocational choices
fell primarily into two groups, social services and biolog-
ical sciences. Fifty-five percent of their fathers had
college degrees; 30 percent were in the professions and 30
percent were in executive-managerial occupations. Mother's
educational level was above average, primarily because two—
thirds of them had additional education beyond high school.
Half of the students expected to receive all necessary col-
lege support from their parents, while none expected to be
entirely self-supporting. Family influence and eXpectations
figured in their choice of vocation and decision to come to

college.

137

Social work women appeared to be other-directed
rather than inner-directed and, in this respect, resembled
home economics teaching, elementary education, and special
education majors. Their abilities and interests were

neither strongly verbal nor numerical.

 

CHAPTER VI

SUMMARY AND CONCLUSIONS

Summary

Purpose
The purpose of this study was to investigate rela-
tionships which exist between students' characteristics, as
measured by orientation data, and their final undergraduate
majors. Measures of ability/achievement, family and back-
ground eXperienceS, self-ratings and interests, and educa-
tional/vocational preferences were selected as relevant
student variables. One major hypothesis was proposed.
Undergraduate major fields can be differen-
tiated on the basis of selected characteristics
of freshman students who attain senior standing
or graduate in each major.
Methodology
The student sample was drawn from the first—time
freshmen who entered Michigan State University during either
Fall 1963 or Fall 1964 and had graduated or attained senior
standing by Spring 1967, having earned no more than 30 quar-
ter credits from another institution. The major sample con-

sisted of the undergraduate major fields represented in the

student sample which contained at least 40 students of one

138

 

139

sex. .A total of 909 men and 856 women in 27 groups repre-
senting 23 different major fields made up the final study
sample from.which a 20 percent random sample was drawn for
cross-validation purposes.

Data for the 47 student variables selected were
available from three sources: (1) the Student Master Record,
(2) the Office of Evaluation Services test records, and (3)

the Personal Igformation Inventory. Multivariate discrim-

 

inant analysis and a maximum likelihood classification tech—
nique based on discriminant scores were chosen as the appro-

priate statistical treatments.

Results

The 27—group discriminant analysis yielded 26 func-
tions, of which 10 were Significant at the .001 level, a
rejection of the null hypothesis. Each of the first five
significant functions was interpreted using the standardized
discriminant weights of greatest magnitude. These five func-
tions eXplained 79 percent of the diSpersion among the
majors.

The first function, a verbal/humanistic-numerical/
technical Split, accounted for 51 percent of the trace.
There was a marked separation of majors into two distinct
groupings; the verbal/humanistic majors clustered together
at the low negative end of the scale with an appreciable gap
between them and the numerical/technical majors at the high

negative end of the scale.

 

140

The verbal/humanistic grouping contained only female
majors, while the numerical/technical grouping contained all
the male majors plus the mathematics women. Although the
first function Spread the verbal/humanistic majors out so
that no two group centroids occupied the same point in dis—
criminant Space, the numerical/technical majors were dis-
persed over a greater range of discriminant space. Thus,
the first function provided finer discriminations among the
majors characterized by numerical, scientific, and technical
interests and abilities than it did among the verbal/humanis-
tic majors.

Function two, interpreted as generalist to Special-
ist in nature, accounted for 11 percent of the trace. Group
centroids were distributed in gradual steps on this function,
from political science (-.34) at the extreme generalist end
to mathematics women (-l.88) at the extreme specialist end
of the scale.

The plot of group centroids on the third function
produced a series of small clusters along an abstract to
applied continuum. Psychology men and women, social science
women, English, and mathematics women majors formed the
extreme abstract cluster and art practice, packaging, gen-
eral business administration, accounting, and marketing
formed the extreme applied cluster. There was a distinct,
although not large, break at the midpoint of the centroid

distribution.

 

141

The high negative weights in function four were
labeled artistic/technical; only four major groups had nega-
tive centroids. They were art practice, Speech, retailing
of textiles and clothing, and English. The remaining twenty-
three majors were densely concentrated within the positive
half of the scale, covering four-tenths of a point, from
packaging with a group centroid of .06 to nursing with a
group centroid of .48.

Function five detached accounting from one end of
the centroid distribution and zoology and nursing from the
other, leaving 24 majors packed into the middle of the dis-
tribution. The nature of the discrimination was unclear and
accordingly, function five was not labeled.

The remaining five Significant functions together
accounted for less than 12 percent of the dispersion and
were not interpreted.

The separate discriminant analyses of 14 male and
13 female majors each produced five significant functions
which accounted for 80 and 76 percent of the dispersion,
respectively. In the male analysis, the first function
spread the majors from high numerical-physical science inter-
ests and abilities to high verbal-social science interests
and abilities. Function two was similar to the third func-
tion in the 27-group analysis (labeled applied to abstract)
and separated the men's groups in much the same way. The

third function suggested a distinction between major groups

 

142

with strong educational goals and groups with strong voca—
tional goals. Function four identified a socio-economic
status factor associated with college attendance. Zoology
and pre-law students were separated out at the high status
end of the scale and packaging and history were at the low
status end.

In the female analysis, function one distributed the
groups along a continuum of numerical ability coupled with
an interest in natural sciences. The second function was a
verbal continuum. Function three suggested a fine arts-
creative versus social service division of majors. Func—
tions four and five were not labeled because the heteroge—
neous combination of high weight variables in each function
did not lend themselves to a common interpretation.

Three cluster analyses were performed using inter-
centroid distances and a computational scheme suggested by
Rao (1952). The 27-group case produced Six multi-group
clusters and two "clusters" which contained only one major
each. The results of the cluster formation supported the
secondary hypothesis that men and women in the same major
have different patterns of characteristics. The l4-group
case produced four multi-group clusters of male majors
identical in composition to the first four clusters in the
27—group case. The lB-group case produced three multi-group
clusters and the same one-major "clusters" as the 27-group

analysis.

 

143

For each analysis, clusters were compared on the
first three Significant functions by means of a three-dimen—
sional plot of group centroids.

A maximum likelihood procedure based on function
weights and discriminant scores was used to classify three
20 percent cross-validation samples into majors and clusters.
Assignments were scored as hits (classification into correct
major), near hits (classification into correct cluster), and

misses. Results were Significantly better than chance.

 

Thus, it would seem that as Holland (1966a) suggests,
there is a tendency for students to seek environments which
permit them to exercise their Skills and abilities, to take
on agreeable problems and roles, and to avoid disagreeable
ones. Results of this investigation demonstrate that it is
possible to differentiate among environments on the basis of
‘the dominant characteristics of members of each environment.
Further, it appears that these characteristics are fairly

stable over time.
Conclusions

The following general conclusions were drawn from
the results of the study.
1. Of the student characteristics included in the study
the greatest differences among the major fields were
related to two dimensions of student character: (1)

the strength of their verbal and ngmerical interests

144

and abllities and (2) preference for either the
ngtural sciences or the socigl,sciences. In the
27-group analysis, the first function separated the
majors on a numerical-scientific-technical to verbal-
humanistic continuum and these dimensions accounted
for 51 percent of the dispersion. In both the 14-
and l3—group analyses, these same dimensions ac-
counted for 33 and 44 percent of the trace respec-
tively.

Ability/achievement measures and educational-voca-
tional preference measures were the mostlpotent
discriminators of major field differences. Across
all significant functions of the three discriminant
analyses test scores, occupational group preferences,
reasons for coming to college, and influence on voca-
tional choice were assigned high function weights
more often than the other classes of variables.
Major fields varied in their tolerance of incongru-
ent student members. Borrowing a concept from
hematology, there appeared to be "universal donor"
majors and "universal recipient" majors. The former
can be described as having narrowly defined and
specific environments which reject or are rejected
by incongruent student types, although congruent

student types are accepted in a variety of environ-

~ments. The latter can be described as having diffuse

 

145

and broadly defined environments which accept and
are accepted by a variety of student types, but
whose congruent types are Spread over a narrow range
of similar environments. In this study, general
business administration, among the men's groups, and
elementary education, among the women's groups,
would be classified as "universal recipients." Both
these majors contained students who most closely
resembled a variety of other fields, but had very
few students misassigned to them. Zoology and mathe-
matics women would be classified as "universal
donors." These majors contained relatively few
incongruent students but had some members from each
of the other majors misassigned to them.

The interest and ability differences between men and
women in the same major were differences of degree
rather than kind. The discriminant analysis and
classification procedure demonstrated that men and
women in the same major were more like members of
their sex group than like members of their major
group. When men and women in the same major were
each described in relation to other majors of their
sex group, however, the descriptions shared many
common characteristics and the major appeared to

hold the same relative position among its peers.

 

5W"

146

More preciae meaagrea of student characteristics are
needed to clarify major field differences suggested
by this study. Verbal and numerical interests and
abilities and preferences for the natural sciences,
social sciences, and creative arts were measured
more directly and Specifically than were abstract, Ems
applied, general, or Specialized interests and orien- I
tations. Nevertheless, these latter dimensions were

implied from the data. Understanding of the impact

 

of these characteristics on major field differences t;
requires more specific instrumentation than is cur-
rently used with incoming freshmen.

The general characteristics and composition of the
clusters which were formed in thla study were in
general agreement with results of previous research.
The Similarities between the male—group clusters of
the present study and Holland's (1966a) classifica-
tion scheme for men's vocations were particularly
striking. The fourteen male groups in the present
study were clustered as follows: mechanical engi—
neering, electrical engineering, and mathematics;
packaging, accounting, general business, and market-
ing; economics, history, social science, pre-law,
and political science; and, zoology and psychology.
Holland's taxonomy would have grouped mechanical

engineering, electrical engineering, mathematics,

147

and zoology; accounting, economics, general business,
and marketing; and, history and political science.
Holland's study did not include the packaging major
and the remaining three (psychology, pre—law, and
social science) were not grouped with any of the

above majors.
Implications

The results of the study and conclusions derived
from them have implications for three related areas: (1)
counseling and academic advisement, (2) administrative deci-

sion-making, and (3) needed research.

Counseling and Academic Advlsement

The methods employed in this study re-organized the
available data into patterns of inter-related characteristics
descriptive of major fields. The data, in this form, provide
comparative models which could be used by students and their
advisers in the planning of academic programs. The more gen-
eral information provided by the cluster analysis should be
of particular value to no-preference students in the develop-
ment of exploratory programs; it would also be applicable for
students contemplating a change of major in that it could be
used to identify major fields with Similar potentials for
interest satisfaction but differing intellectual demands or
to identify fields with Similar intellectual demands but

differing potentials for interest satisfaction.

 

148

AdministrativelQecisionsMaking

The decision to reorganize existing units within a
university structure (e.g., separating or combining depart—
ments within a college) or to create new units (e.g., start-

ing a small liberal arts college within the university)

 

encompasses a variety of concerns. One of these ought to be Fm
a consideration of the effect of the "mix" of congruent and
incongruent human characteristics that will result from the
re-structuring. The results of this study suggest that g!

selection and retention of students within a unit are par—
tially related to the dominant human characteristics found
there. There is some evidence in previous research to
suggest that selection and retention of faculty are also
related to this factor. It is important to mention that
information of this sort can be used either to perpetuate
the status quo or to introduce new qualities into an envi—
ronment. A discussion of the relative merits of these

alternatives is beyond the scope of this investigation.

Needed Research

AS indicated earlier, there is a need to improve the
precision of the most potent variables in the orientation
data. In the interests of economy, those measures currently
collected but seldom used could be eliminated. Research
time could be profitably Spent on the identification of
those human characteristics which play an important role in

defining major field similarities and differences. Results

149

of this investigation suggest a need for more directly inter-
pretable interest and personality measures.

An important aspect of major field environment, which
was not touched on in the current study, is the impact of
faculty characteristics. While it is unlikely that parallel
measures of faculty members' characteristics would be avail-
able for analysis, it seems logical to assume that some of
the information routinely collected on faculty would prove

discriminating. Indirect measures of faculty characteristics,

 

such as Thistlethwaite's (Mimeo., ca., 1960) faculty press
index based on student perceptions of faculty behavior, have

been successful in distinguishing among major fields.

B IBLIOGRAPHY

BIBLIOGRAPHY

Abe, C., & Holland, J.L. A description of college freshmen.
I. Students with different choices of major field. Iowa
City: American College Testing Program, 1965. Research
Report No. 3.

Astin, A.W. Classroom environment in different fields of
study. Journal of Educational Psychology, 1965, 56,
275-282.

Augustine, R.D. Persistence and change in major field of
academically proficient engineering students at three
midwestern universities. Unpublished doctoral disserta-
tion, Michigan State University, 1966.

Baggaley, A.R. The relation between scores obtained by
Harvard freshmen on the Kuder Preference Record and
their fields of concentration. The Journal of Educa-
tionaleaychology, 1947, 38, 421-427.

Baird, L.L. Family lncome and the characteristics of col-
lege-bound students. Iowa City: American College
Testing Program, 1967. Research Report No. 17.

Baird, L.L. The educational goals of college-bound youth.
Iowa City: American College Testing Program, 1967.
Research Report No. 19.

Bennett, G.K., Bennett, M.G., wallace, W.L., &'Wesman, A.G.
College anlification Tests: 1957 Manual. New Ybrk:
The Psychological Corporation, 1957.

Cole, J.W., Wilson, K.M.-, &Tiedeman,D.V. Dispersion analy-
sis and the search for educational goals in college: A
study in duplicate. Journal of Educational Psychology,
1964, 55, 121-131.

Cooley, W.W., & Lohnes, P.R. Multivariate procedures for the
behavioral sciences. New York: Wiley, 1962.

Computer Institute for Social Science Research, Michigan
State University. Technical Report No. 33. Discrim:
Multiple discriminant analysis. Programmed by Stuart
Thomas.

150

151

Cullum, F.W. The isolation of GATB aptitude patterns for
six major fields of study. Dissertation Abstracts, 1966,
26(9), 5223.

Cutting, D.J. Predicting the selection of a field of
concentration at Michigan State University from the
Personal lnformationylnventory. Unpublished doctoral
dissertation, Michigan State University, 1966.

Davis, J.A. Undergraduate career decisions: Correlates of
occupational choice. Chicago: Aldine, 1965.

Dunn, Frances E. Two methods for predicting the selection
of a college major. Journal of Counseling Psyghology,
1959, 6, 15-26.

Elton, D.V., & Rose, H.A. Personality characteristics of
students who transfer out of engineering. Personnel and
Guidance Journal, 1967, 45, 911-915.

Gamble, G.W. Pre-college eXperienceS and curriculum changes
by college students. Personnel and Guidance Journal,
1962, 40, 561-564.

Goldschmid, M.L. The prediction of college major in the
sciences and the humanities by means of personality
tests. Dissertation Abstracts, 1966, 26(7), 4073-4074.

Harder, D.F. Differentiation of curricular groups based
upon responses to unique items of the MMPI. Journal of
Counseling Psychology, 1959, 6, 28-34.

Hasan, S. A comparison between students who changed and
those who did not change their majors during their
undergraduate training at Southern Illinois University.
Dissertation Abstracts, 1966, 26(11), 6513.

Hays, W.L. Statistics for psychologists. New YOrk: Holt,
1963.

Holland, J.L. A psychological classification Scheme for
vocations and major fields. Journal of Counseling
Psychology, 1966a, 13, 278-288.

Holland, J.L. The psychology of vocational choice. Waltham,
Massachusetts: BlaiSdell, 1966b.

Holland, J.L., & Nichols, R.C. Explorations of a theory of
vocational choice: III. A longitudinal study of change
in major field of study. Personnel and Guidance Journal,
1964, 43, 235-242.

152

Ikenberry, 8.0. A multivariate analysis of the relationship
of academic aptitude, social background, attitudes and
values to colleglate persistence. Unpublished doctoral
dissertation, Michigan State University, 1960.

Juola, A.E. The differential validity of the College Qual-
ification Tests for diverse curricular groups. Person-
nel and Guidance Joprnal, 1961, 39, 721-724.

King, R.G. The prediction of choice of undergraduate field

of concentration in Harvard College. Unpublished doc-
toral dissertation. Harvard University, 1958.

Korn, H.A. Differences between majors in engineering and
physical sciences on CPI and SVIB scores. Journal of
Counseling Psychology, 1962, 9, 306-312.

Lundin, R.W., & Lathrop, W. The relationship between field
of major concentration and personality adjustment in
college males. Journal of General Paychology, 1963, 69,
193-196.

Martoccia, C.T. Authoritarianism and college major.
Psychological Reports, 1964, 15, 510.

Matteson, R.W. Educational experiences, academic interests,
and curriculum choices. Personnel and Guidance Journal,
1961, 39, 717-720.

Mazak, R.M.J. The relationship of selected characteristics
of junior college pre-engineering students to their
success and persistence in upper-division professional
education for engineering. Dissertation Abstracts, 1967,
27(2), 470-A.

Mosier, C.I. Problems and designs of cross-validation.
Educational and Psychological Measurement, 1951, 11,
5-11.

Office of Evaluation Services, Michigan State University.
Comparative standings of the various college ang_ourrlgr
glam groups on the orlentatlon week examinations for the
Fall of 1963. Ca., 1964. Mimeo.

Petrik, N.D. Socio-economic status, vocational interests
and persistence in selected curricula. glssertation
Abstracts, 1966, 26(9), 5242.

Pierson, R.R. Changes of majors by university students.
Personnel and Guidance Journal, 1962, 40, 458-461.

153

Saupe, J.L. Factorial-design multiple-discriminant analysis:
A description and an illustration. American Educational
Research Joarnal, 1965, 2, 175-184.

Saupe, J.L. Npmbers of coarses llsted and taught at nine
CIC gniversitles: 1962-63, 1963-64. East Lansing,
Michigan State University, Office of Institutional
Research, 1966.

Simono, R.B. The prediction of science and nonscience
undergraduate majors among a group of male high School
students. Dissertation Abstracts, 1967, 27(1), 133-A.

Smith, N.A. Social and cultural factors in the backgrounds
of women physical education majors. Dissertation
Abstracts, 1967, 27(2), 472-A.

Stahmann, R.F., &'Wallen, N.E. Multiple discriminant pre-
diction of major field of study. Educational and
Psychological Measurement, 1966, 26, 439-444.

Stinson, P.J. .A method for counseling engineering students.
Personnel and Guidance Joarnal, 1958, 37, 294-295.

Tatsuoka, M.M. Joint probability of membership and success
in a group: An index which combines the information
from discriminant and regression analyses as applied to
the guidance problem. Report No. 6, Harvard Studies in
Career Development. Naval Research Contract Nonr —
1866(31), 1957.

Thistlethwaite, D.L. College press ang_changes ln study
plans of talenteglstudents. National Merit Scholarship
Corporation, Mimeo., ca., 1960.

Tiedeman,D.V., & Bryan, J.G. Predictions of college field
of concentration. Harvard Educational Review, 1954, 24,
122-139.

Veldman, D.J., Peck, R.F., and Richek, H.G. Personality
correlates of the high school experiences of prospective
teachers. Personnel ang Ggidance Journal, 1968, 46,
473-477.

Voorhies, E.S. The use of discriminant and regression
analysis in the selection of an area of study in a state
college. lesertation Abstracts, 1966, 26(12), 7162-
7163.

154

Warnath, C.F., & Fordyce, H.R. Inventoried values of enter—
ing college freshmen. Personnel and Guidance Journal,
1961, 40, 277-281.

Warren, J.R. Self-concept, occupational role eXpectation,
and change in college major. Journal of Counseling
Psychology, 1961, 8, 164-169.

Werts, C.E. .Career changes ln college. NMSC Research
Report, Volume 2, No. 7. Evanston: National Merit
Scholarship Corporation, 1966.

 

 

APPENDICES

APPENDIX A

VARIABLES FROM THE PERSONAL INFORMATION INVENTORY

 

VARIABLE

1. Reading
preferences

2. High school
subjects

3. Self-rating

of abilities

APPENDIX A

VARIABLES FROM THE PERSONAL INFORMATION INVENTORY

3a.

PERSONAL INFORMATION mvspg'om ITEM

The type of book I prefer to read now
for pleasure

0. Novels

1. Technical books
2. Mystery books

 

3 Biography and history
4 Literary classics

 

For the same school subjects circle
the letter to indicate how well you
liked it.

a. Liked very much

b. Liked more than disliked

c. Disliked more than liked

d. Disliked very much

e. Never studied the subject

Art

Biology
Business
Chemistry
English
Foreign Langs.
Government
History

Home Economics
Mathematics
Music
Physical Educ.
Physics
Speech
Shopwork

DD

Hid
Hommummbwww
I

H
N
I

13.
14.
15.

”UNDWWNWNSDNIDIDW

UCTUCTUCYUCfUCYUCfUCfU
OOOOOOOOOOOOOOO
DanQaQanQaOaQanO-O-Qoflaflafl:
mommmmmmmmommmm

Compared to other entering students

here, I think my general capacity

for college work is

0. Superior 1. Above Average
(Upper 10%) (Upper quarter)

2. Average 3. Below Average
(Middle 50%) (Lower quarter)

4. Low (Lower 10%)

Compared to other entering students
here, I think my aptitude for solving
numerical reasoning problems is

0. Superior 1. Above Average

2. Average 3. Below Average

4. Low

Compared to other entering students
here, I think my aptitude for under-
standing and reasoning words is

0. Superior 1. Above Average

2. Average 3. Below Average

4. Low

Compared to other students entering
here, I think my reading skill is
0. Superior 1. Above Average

2. Average 3. Below Average

4. Low

The grade average I think I will be
able to Obtain at MSU is

0. A l. B to 8+
2. C+ to B- 3. C
4. C- or under

 

 

155

1.

2a.

U

3e.

SCALING CODE
For each type of book:

did not prefer
did prefer

0
II II

Categorized by subjects:
1. Fine and applied arts
and letters (1,5,6,

11,14)
2. Science and mathe-
matics (2,4,10,13)
3. Social studies (7,8)
. Non-academics (3,9,
12,15)

U)
0
DJ
...:
(D
Q.

4

0400‘!”

3
2
l ,e
Scored: Total points in

a category divided by the
number of subjects rated.

Scaled:

Superior
Above Average
Average

Below Average
Low

Hrou:&«n
II II II II II

Scaled:

A

B to 8+

C+ to B-

C

C— or under

Hrou:htn
II II II II

 

156

APPENDIX A--Continued

 

 

 

VARIABLE PERSONAL INFORMATION INVENTORY ITEM SCALING CODE
4. Selection of 4. In the following list select the one 4. For each group:
three occupa- group of occupations in which, on the
tional groups basis of interests and abilities, you 0 = not selected FE?
believe you fit best. Check it under 1 = selected as first,
first choice. Select the one group second, or third
of your second choice and mark it choice
under second choice. Mark a third
choice, too. Then mark under X any NOTE: The final sentence
group of the six remaining in which in the instructions for é
you feel sure you would fit. this item should have .
read, “Then mark under X 7
Occupations requiring special any group of the six 2,
artistic abilities, such as musician, remaining in which you %
actor, artist, interior decorator, feel sure you would NOT
designer, etc. fit." Due to a printing

error, the Ngg_was omit-
ted. This error was not
corrected during the
administration of the

Occupations involving work in the
physical sciences, such as engineer,
chemist, mathematician, physicist, etc.

Occupations involving work in the Inventory, which made it
biological sciences, such as zoologist, necessary to ignore all
botanist, nurse, physician, veterinar- occupational ratings
ian, physical therapist, etc. except the first three.

Occupations involving mechanical
and7or technical skills, such as farmer,
aviator, printer, forester, industrial
arts, science instruction, etc.

Occupations involving social ser-
vice activities, such as social worker,
youth organization leader, personnel
man, teacher, etc.

Occupations involving business
detail, such as accountant, business
statistician, cashier, banker, stenog-
rapher, office clerical worker, etc.

Occupations involving business con-
tacts with people, such as the various
fields of selling, promotional work,
politics, etc.

Occupations involving verbal or
linguistic work, such as lawyer, author,
newspaper man, advertising man, librarian,
etc.

Occupations involving executive
responsibilities such as office manager,
foreman, production manager, director,

 

etc.
5. Educational 5. Education (check one for each person) 5. 1 = Grade school
level of 0. Grade school 2 = Attended high school
fathera 1. Attended high school 3 = Completed high school
2. Completed high school 4 = Technical or business
3. Technical or business school school; College
4. College 5 = Post-college work
5. Post-college work
6. Educational 6. Same as #5 above. 6. Same as #5 above.
level 8f

mother

1157

APPENDIX A—-Continued

 

VARIABLE PERSONAL INFORMATION INVENTORY ITEM SCALING CODE
Occupational 7. Chief employment (check one for each 7. 1 = Manual worker;
level of adult) Service
fatherb 0. Manual worker--no special training 2 = Office, clerical,
required and sales
1. Skilled labor 3 = Skilled labor
2. Business owner 4 = Farm owner or
3. Farm owner or operator operator;
4. Executive or managerial Business owner
5. Office, clerical, and sales 5 = Executive or
6. Teacher (elementary or secondary) managerial
7. Professional 6 = Professional:
8. Service (store clerk, barber) Teacher
9. Homemaker
Employment 8. Same as #7 above. 8. O Homemaker only

 

 

 

 

of mother 1 Working outside
the home

Marital status 9. Parents' marital status (check one) 9. 3 = Married
of parents 0. Married 2 = Divorced or

1. Separated widowed and

2. Divorced remarried

3. Divorced, one or both remarried 1 = Separated,

4. Father widowed and unmarried divorced, or

5. Mother widowed and unmarried widowed and

6. Father widowed and remarried unmarried

7. Mother widowed and remarried
Number of 10. Family information 10. O = None
older Brothers Sisters 2 = One or two
siblings Age .. 4 = Three or more
Number of 11. Same as #10 above. 11. Same as #10 above.
younger
siblings
Sources of 12. Plans for financial support (rank 12. 4 = Entirely supported
financial if more than one) by family
support for O. Entirely supported by family 3 = Scholarship; Vaca-
college 1. Part-time work will be neces— tion & summer work

sary (___hrs. a week) 2 = Part-time work
2. Vacation and summer work will be necessary
3. Total self—support will be 1 = Total self-support
necessary will be necessary

4. Scholarship
Strength of 13. My feelings with regard to this 13. Same as PII item with
first major major as my best choice is the addition of:

5. Absolutely fixed choice

4. Reasonably certain l = No preference

3. Fairly certain

2. Somewhat doubtful

1. Very uncertain
Certainty of 14. My present vocational choice is 14. 3 = Very certain;
vocational . That this Fairly certain
choice choice is the one for which I really 2 = Uncertain

want to prepare I am 1 = Doubtful; Very

0. Very certain doubtful;

l. Fairly certain No choice

2. Uncertain

3. Doubtful

4. Very doubtful

 

158

APPENDIX A--Continued

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

VARIABLE PLRSONAL INFORMATION INVENTORY ITEM SCALING CODE

15. Factors 15. Rank in order the five most important 15a. Categorized by items:
influencing factors influencing your vocational 1. Family and significant
vocational choice others (O,l,2,4)
choice 0. Family suggestion or tradition 2. High school experiences

1. Friend's or teacher's advice (3,5,X)
2. The vocation of someone I admire 3. Personal needs and F31
or respect interests (6,7,8,9)
3. Suggested by study in school .
4. Counselor's suggestion b' icilggfik 1
5. Interest and aptitude testing 4 = Rank 2
6. A long personal interest in 3 _ R
the work _ ank 3
. . . . 2 = Rank 4 «
7. It 18 most profitable finanCially 1 = Rank 5
8. It is best suited to my abilities
9. It is most interesting intellec- c. Scored: Total points in
tually each category.
X. Ability in related school subjects
Y. Other .

16. Reasons for 16. Students usually have many good reasonsl6. Categorized by items:
coming to for coming to college. Rank in order 1. Educational and
college the five reasons that seem most impor- vocational needs

tant to you. You may add others or (0,1,8)
explain. 2. Social needs (3,4,5)
0. To get a liberal education 3. Family influence
1. To prepare for a vocation (6,7,9)
2. For the prestige of a college deg. 4. Status needs (2,X,Y)
3. To be with old school friends
4. To make friends and helpful Scaled and scored as #15
connections above.
5. For social enjoyment; "College
Life"
6. To please parents or friends
7. Family tradition
8. To learn more of certain subjects
9. It was the thing to do, foregone
conclusion; I never questioned why
X. Without college training, there is
less chance of getting a job
Y. Will enable me to make more money

 

aItem revised in 1964 edition.

bStudents classified father's occupation as falling into one of nine PII occupa-
tional groups. The North-Hatt classification system (Reiss, 1961; pp. 54—57), and the
Warner occupational scales (Warner, Meeker, & Belle, 1960; pp. 140-141), were employed
in coding the nine PII occupational groups into six occupational categories. Seventy-
nine of the ninety occupations used in the North-Hatt were distributed among the nine
PII groups. The nine PII groups were then combined into six categories on the basis of
mean North-Hatt scores and the six groups were ranked on the North—Hatt mean score.
These groups were also paired with comparable groups on the Warner occupational scale
and the resulting rank order compared with the North-Hatt ranks. Disagreements between
the two scales were decided in favor of the North-Hatt rank order. Results are summar—
ized below.

 

 

 

 

 

North—Hatt Warner
Dissertation

PII Occupational Groups Mean Score Range Rank Rating Rank Scale
1. Manual worker; Service 48.4 33-60 6 5-7 6 6
2. Office, clerical, & sales 66.5 62—68 5 2-5 4 5
3. Skilled labor 69.4 62-77 4 4-5 5 4
4. Farm owner or operator;

Business owner 73.5 62-82 3 2-4 3 3
5. Executive or managerial 81.2 67-93 2 l—3 1.5 2
6. Professional; Teacher 84.3 71-96 1 1-3 1.5 l

 

APPENDIX B

ORIGINAL AND FINAL FREQUENCY DISTRIBUTIONS

OF RESCALED VARIABLES

 

APPENDIX B

VARIABLES INCLUDED IN TABLES 1-2:

Ability/Achievement Measures
1. CQT Verbal
2. CQT Information
3. CQT Numerical
4. MSU English
5. MSU Reading

Reading Preferences

6. Novels

7. Technical books

8. Mystery boOks

9. Biography and History
10. Literary classics

High School Courses

11. Fine and applied Arts
and Letters

12. Science and Mathematics

13. Social studies

14. Non-academics

Self-Ratings

15. General capacity

16. Numerical reasoning

l7. Verbal reasoning

18. Reading skill

19. Anticipated college GPA

Occupational Interest Groups
20. Artistic

21. Physical science

22. Biological science

23. Mechanical and technical
24. Social service

25. Business detail

26. Business contact

27. Verbal and linguistic
28. Executive and managerial

APPENDIX B

Family and Background Experi-

ences

29. Educational level of RV
father

30. Educational level of
mother

31. Occupational level of
father

32. Employment of mother i~

 

33. Marital status of parents

34. Number of older siblings

35. Number of younger siblings

36. Size of high school
graduating class

37. High school in- or out-of—

 

 

state

38. Sources of financial
support

Educational and Vocational

Preferences

39. Strength of first major
choice

40. Certainty of vocational
choice

Factors influencing vocational

choice:

41. Family and significant
others

42. High school experience

43. Personal interests and
needs

Reasons for coming to college:

44. Educational/vocational
needs

45. Social needs

46. Family influence

47. Status needs

159

160

 

 

 

 

 

 

 

... ... mm mm 00A 00A RCA 50A NvA NvA m
... ... mew mew Ame Ame er er 5A0 FAQ w
... NmA NAh NAh Amm Amm Nmm Nmm mmm mmm m
... hmA mAA ¢AA mOA mm Nm mm mmA NmA N
mmv emA ... A ... e ... m ... 0A A
NAmA NANA ... ... ... ... ... ... ... ... 0
Achm .mAHo AmsAm .mAHo AmsAm .mAHo AmsAm .mAHo AmsAm .mAuo wAmom
0N mAnmAHm> mA wAQMAHm> mA wAQMAHm> SA mAnmAHm> 0A mAQMAHEV

AAA AAA ... ... ... ... ... ... ... ... m
mAh mAn mou mow wow wow ¢No eNo on on e
mNm mNm new new mNh mNn 0N0 on oem oem m

h m MAN mmA mmA mmA MON NON nmA hmA N
... A ... Nm SA NA NA AA 0 m A
... ... ... mm ... m ... A ... o 0
Achm .mAHO AmsAm .mAHO AmsAm .mAHO AmsAm .mAHO Achm .mAHO wAmom
mA oAQMAum> ¢A oAQmAHm> mA mAQMAHm> NA mAQMAHm> AA mAQMAHm>

 

 

mMAm<Hm4> QMA<Ummm m0 mZOHBDmHMBmHQ MUZWDOmmm AdZHm 92¢ AflZHmeO

 

possAusoollm NHszmm¢

161

 

 

... em ... meN ... ... ... ... ... ... o
em 0mm meN 0mm ... ... ... ... ... ... m
own va men 00A ... ... ... ... ... ... v
Ame Ame woe woe ... mo ... mNA ... mm m
omA mmA NON wON ... eoA ... eAN ... NmA N
em em 00A 00A hme mNN hum mmN mAm meN A
... ... ... ... momA momA mmAA mmAA heNA heNA o

AmsAm .mAuo AndAm .mAHo AmsAm .mAHO Achh .mAHo AmsAm .mAHO mAwom

om mAnmAum> mm «Annuum> mm «AnmAum> km mAnmAum> on «Anmeum>

 

 

 

coo fim coo Nmm .0. 0m ooo mNN coo 00¢ m
... SAN ... th ... MNA ... mmN ... owA N
0mg mmA ham wmA th mMA mmm mmA #05 mm A
thA mbNA men men moeA mowA bNAA hNAA AmOA A@OA 0
Achm .mAMO AmsAm .mAHO Achm .mAuo AmsAm .mAHO AmsAm .mAno mAmom
mN mAQmAHm> vN mAQMAnm> mN wAQMAHm> NN mAQMAum> AN mAQMAHm>

 

 

mmqm<Hm¢> QHA<Ommm m0 mZOHBDMHMBmHQ_NUZWdemm AfiZHh 92¢ A¢ZHUHMO

 

@mscoucoollm KHQZMAQC

162

 

55A
ONm
moo

NBA
Ame
mMN
@mA
Nmfi

AAv
wAN

AA¢
mAN

... N ...
... A ...
... m ...
... mA ...
... ON ...
mMN mm Am

... mMA ...
Amm mom mmm
... @mm ...
woo woo 0mm

MN
mm
CNN
mAm
0mm

#mm
Ah¢
mmm
NmN
mbA
omA

fimN
Ahfi
om

MMN
00A
NmN
mhA
Aw

mOA

OHNMfi‘mLOl‘CDON

 

Amuse .muuo
oe «Anmnum>

Hasan .meno
mm mAnmAHm>

AmaAm .mAHO
mm mAQMAHm>

Awash .mAHo
em mAnmAum>

Amman .muuo
Am mAnmAnm>

mAmum

 

mmfldem¢> GHANUmmm m0 mZOHBDmHMBmHQ.NUZmDOmmm AdZHh QZN A<ZHOHMO

 

wwscflucoollm XHQZMth

APPENDIX C

COMPUTATION OF INTRA- AND INTER-CLUSTER DISTANCES

FOR TWENTY-SEVEN GROUP CASE

163

 

 

a 35 m3. A mama 35
m mN.mN.¢N.MN «Na. moo.A o oem.m 8N
mN.eN.mN mmm. mmm.o m sAm.N mN
eN.mN Ame. .... A Ams.o «N.mN
N m.n.o mmm. Nom.o m mmm.N m
s.@ mm». .... A mmn.o s.0
0 mA.NA.AA.OA.¢ mom. mmm.o 0A Amo.m e
mA.NA.AA.OA New. sAm.o m emo.m AA
NA.NA.OA mus. mNm.o m ¢NN.N NA
MA.oA who. .... A osm.o NA.OA
m NN.NN.AN.ON.NA.@A.mA Rom. NeA.A AN .Nom.0N mA
NN.NN.AN.oN.mA.mA nmm. ANo.A mA mme.mA NN
NN.AN.ON.mA.mA emm. omm.o 0A mvm.m 0A
NN.AN.ON.mA Nos. mAm.o o Noo.¢ AN
RN.0N.NA ANm. mmo.o m Nmm.A 0N
RN.mA mam. .... A mmm.o NN.mA
m m.m.N.A mom. ooo.A m omA.m A
m.m.N NNN. Aem.o m omA.N N
m.m mme. .... A mme.o m.m
GOHumcmHme HTHMSHU AG\Q.WV C CH wmmmHUGH CC :3 HmumfiHU
mosmumAQ + mEAwB musmumAQ m on
mmmnm>< Q EA mummnocH mo UAOHusmU omnpd
Hmnfisz IAwusH AOnmz

 

 

mm024BmHQ QHOfiBZMUfiMBZH GZHmD
mm<0 mbomw Zm>mmIMBZMSB mcm mmUZ¢BmHQ MHBmDAUINMBZH m0 ZOHBdHDQEOU

U Xanmmmd

164

 

 

 

 

 

 

 

 

 

mm.A m0.A 0N.A NN.A 00.0 m0.A Am.A 5A.A ¢N.A 00.A 0N.A N¢.A N5.A 50.A 00.N 00.A Nm.A m5.A 5A.N N0.A A0.A 00.A 00.A N0.N 5A.N mA.N 0A

mm.A m5.A m0.N m0.A v0.A 05.A #0.A 05.A MN.A A0.A 00.A N0.A 5N.A 0N.A A0.A m0.A 00.A 00.A m5.A V0.A 00.A 50.A N0.A 0N.A m5.A A0.A 5A
AN.A 50.0 N0.0 0m.0 00.0 50.A 0A.A 0N.A 00.0 MA.A 00.A N0.A 05.A A0.A ¢¢.A 00.A 00.N 00.A A0.A 0A.N 0A.N 5A.N 0m.N 0m.N 5N
A0.0 NA.A 00.A 50.0 0v.A 00.A 00.A 0¢.A 0m.A mA.N 0N.N ¢M.N AM.N 00.A m0.N Am.N MA.N VA.N Am.N Nm.N 00.N N0.N A0.N NN
¢A.A N0.A 00.0 50.0 N0.0 0m.Afim¢.A m¢.A mm.A 0A.A 50.A NA.N 00.N 0A.N m5.A 00.A N¢.N 50.N 00.N 0N.N ¢M.N Nv.N m0.N 00.N AN
m0.0 00.0 00.0 V0.A 0N.A 50.0 00.0 50.A 00.A A0.A 00.A mm.A 50.A 0A.N @0.A m0.A 0A.N 0A.N VA.N mm.N 0m.N 0N

50.0 #0.0 00.A 0m.A 0N.A 00.A 0A.A 00.A 00.A 00.A 00.A 0m.A A0.A 0A.N N0.A 00.A A0.N MA.N 0A.N ov.N Nv.N 0A

00.A 5N.A Av.A 0N.A 5N.A A0.A 00.A 00.A N0.A 0m.A N0.A AA.N 05.A 05.A 00.A 0A.N mN.N 0¢.N A¢.N 0A

5N.A 0m.A 00.A Am.A N0.A 05.A m0.A N5.A N¢.A 0m.A V0.A m0.A 55.A 00.A 00.N 0A.N 0N.N AN.N 0A

N0.A m0.A 50.0 0m.A m0.A 00.A N0.A 0¢.A 0¢.A v0.A 00.N 50.A 0A.N NN.N 00.A vN.N AA.N 0N
0m.A Am.A 0N.A 00.A 55.0 mm.A 00.A 5m.A 00.A Am.A A¢.A 00.A 00.N ¢O.N 0m.N mv.N ¢A.N mm.N mv.N mN
N0.0 05.0 ¢V.A 05.A 0¢.A 00.A <¢.A AN.A 00.A 50.A N0.A 0A.N AN.N m0.N Am.N NN.N 6N
N0.A 00.A N5.A A0.A 50.A 00.A 0A.N MN.N 0A.N 0m.N 5¢.N vN.N 5m.N 5v.N MN

v0.A 0N.A 00.A m0.A 00.0 N0.A d0.A 00.A A0.A 5A.A mm.A Am.A V0.A 00.A ¢O.A 00.A 5N.A ¢A
00.0 0¢.A 0N.A AA.A mm.A «m.A 0N.A Nm.A 0m.A AN.A 00.A 0N.A MN.A 0

05.0 00.0 00.0 N0.A mv.A mm.A m0.A 05.A m0.A N0.A 55.A MA

00.0 00.0 e0.0 5N.A 0A.A V0.A 00.A mv.A 00.A 00.A NA

00.A 00.A 00.A m0.A 0A.A A0.0 00.0 50.A 00.0 v0.0 0N.A 0N.A 0v.A 0v.A 50.A AA
50.A Nm.A 0N.A 0v.A N0.A 00.A m5.A 00.A 0A

AN.A 0A.A 0m.A 0v.A 0N.A 5N.A 5N.A v

00.0 v0.0 00.0 50.A Av.A 00.A m

. . . . . . . v0.0 00.A A0.A A¢.A 50.A m
N0 A 00 A N0 N 5A N Nv A cm A 00 0 00.0 mv.A Am.A Am.A N
0N.A N0.0 AN.A A

m0.A 05.0 0

AA.N mm.A 5m.N 0N.N MN.A em.A N¢.A 05.0 5
00.0 0

0A 5A 5N NN AN 0N 0A 0A 0A 0N mN ¢N mN ¢A 0 NA NA AA 0A 6 m m N A 0 5 0
m 0 ...A m D U m <

 

 

 

 

 

 

 

 

 

 

mmofiemHQ QHOmBZmummBZH OZHmb mmdd @5000 ZN>N0IMBZME WEB mom MMUZANBWHD mMBmDAUImmBZH .mO ZOHB¢BD§OU

U XHQmem/w

APPENDIX D

COMPUTATION OF INTRA- AND INTER-CLUSTER DISTANCES

FOR FOURTEEN GROUP CASE

165

 

 

.0 eA.0 050. .... A 050.0 «A.0
.o mA.NA.AA.0A.¢ «an. Ame.m 0A smm.s 4
A05. A¢N.N 0 0om.¢ AA
NN5. N¢0.A m 00A.N .NA
0N0. .... A mN0.o 0A.0A
.0 0.5.0 500. 0Am.A m N00.N 5
000. .... A eem.o 0.0
.m m.m.N.A N05. 50m.N 0 00m.¢ A
500. ¢5¢.A m N50.A N
00¢. .... A 00¢.o 0.0
GOAumsmAmwQ HmumsAu AC\QWV s GA mmmmuusH Adv A00 HmumsAo
mosmumAn + maums mosmumAa m on
m0mum>¢ a GA ommmuosH mo vonusmO ©m©©<
Hmnfisz lumusH Hons:

 

 

0HOZ¢BWHQ QHomBZMUmmBZH DZHOD M040
mbomw meemDom mmB mom mmUZ<BmHQ MHBmDAUINMBZH ho ZOHBdHDmSOU

Q XHQmem<

166

 

 

 

 

 

am.o No.0 No.0 AA.A em.o No.0 00.A em.A mo.A NA.A so.A m¢.A AN.A 4A a
sm.o om.A OA.A eo.A «N.A sm.o mN.A oN.A mN.A mo.A Am.o mo.A Nm.o m .
ma.o sm.o No.0 em.o we.A vN.A mo.A mo.A eo.A mk.A NA.A mA
ms.o ms.o mn.o mN.A NA.A Nm.A N¢.A e¢.A Nm.A Nm.A NA
00.A Ns.o om.o mm.o om.o mA.A mA.A we.A me.A sm.A AA .0
Nm.o mN.A mN.A om.A mm.A om.A NN.A mm.A OA
ms.o mo.A mo.A NN.A mA.A AA.A oA.A NA.A e
om.o es.o mm.o mm.A mm.A em.A m
. . es.o no.0 mm.A mm.A mo.A m
mm A NN A mh.o sm.A oN.A om.A N .m
on.o NA.A mk.o NA.A A
mm.o em.o m
mo.A Am.A mm.A No.0 5 .a
mm.o o
A m .NA NA AA 0A a m m N A m a m
cm .0 .m .a

 

 

 

 

 

 

mmUZNBmHQ QHOMBZMUEWBZH GZHOD
MWNU mDOmw ZNMBMDOh HEB mom mWUZNBmHQ mmBmDAUImmBZH mO.ZOHE¢HDmZOU

cmSGAuCOOIIQ NHQZEAQd

APPENDIX E

COMPUTATION OF INTRA- .AND INTER-CLUSTER DISTANCES

FOR THIRTEEN GROUP CASE

167

 

 

.H 5A .... .... .. .... 5A
.0 0A .... .... .. .... 0A
.0 0N.mN.¢N.mN 005. 500. 0 005.¢ 0N
005. N05. 0 0NA.N ¢N
0N0. .... A 0N0.0 mN.MN
.m NN.AN.0A 0A5. 005. m 00A.N 0A
¢N0. .... A ¢N0.o NN.AN
.0 5N.0N.0A.mA N¢0. 000. 0 A00.m mA
05¢. 00¢. m 0N¢.A 0N
m0¢. .... A m0¢.0 5N.0A
soAumsmAme HmumsAU AG\QNV a GA mmmmHUsH Adv ADV Hmumsdo
mosmumAQ + m5nt mosmumAQ m on
m0mnm>¢ Q sA mummHUcA mo 0Aouusmo 00000
Hmnfisz lumusH Hohmz

 

 

H0¢U m50m0 meHmHmB HEB mom mMUZNBmHQ MMB0DAUINMBZH m0 ZOHBNBDmZOU

0mUZ¢BmHD QHOMBZMUMHBZH OZHOD

m XHQmemd

168

 

 

 

 

 

 

A¢.A 0N.A 00.A 0¢.A 00.A 05.A 05.A A5.A ¢0.A A0.A 00.A ¢0.A 5A .H
A¢.A 0A.A N0.A 0A.A ¢M.A 0A.A 0A.A MM.A A0.0 ¢0.0 00.0 0A.A 0A .3
N0.0 00.0 00.0 5M.A mN.A MN.A 00.0 00.0 50.A 5A.A 0N
00.A AM.A N0.0 N0.0 AM.A AN.A AA.A 50.A N0.A A0.A 0A.A 0N 0
00.0 0A.A 0A.A 00.A A0.0 A5.0 05.0 A0.0 ¢N .
00.0 50.A M0.0 M0.A N0.0 N0.0 00.0 ¢A.A MN
N0.0 05.0 N0.A 00.0 00.0 ¢N.A NN
05.A NN.A 0A.A N5.0 00.0 00.0 N0.0 0A.A AN .0
A5.0 ¢0.0 ¢0.0 00.0 00.0 0A
5¢.0 0¢.0 00.0 5N
. . . . 0¢.0 55.0 ON
A0 A 00 0 00 0 00 0 ¢0.0 55.0 0A .0
0A
5A 0A 0N mN ¢N MN NN AN 0A 5N 0N 0A 0A
.H .m .0 .m .m

 

 

 

 

 

 

 

meZNHmHQ QHOMBZmumMBZH OZHmD
mmdv abomw meBMHmB HEB mom mmUZ<BmHQ MMBmDAUImmBZH m0 ZOHBdBDmZOU

 

UmchUGOUIIm NHQmem4

APPENDIX F

GROUP MEANS AND STANDARD DEVIATIONS FOR TWENTY-SEVEN GROUPS

AND FOR . ALL GROUPS ON FORTY-SEVEN VARIABLES

1£59

 

 

 

 

 

 

 

 

 

0N.0 00.0 0N.0 50.0 04.0 05.0 00.5 00.Nm 00.0 00.0N 05.0 40.50 00.0 00.00 00.AA 40.40 A0009
0A.0 No.0 00.0 00.0 50.0 40.0 5N.0 0N.Am A0.0 4A.0N 0A.0 55.0N 50.0 04.54 40.NA 00.N0 .5N
04.0 0N.0 0N.0 00.0 04.0 00.0 00.0 00.00 00.0 40.0N 00.0 00.00 00.0 50.N0 00.0 04.A0 .0N
00.0 00.0 5A.0 00.0 04.0 A5.0 00.5 55.40 4N.0 00.0N 00.0 A0.0A 04.5 00.00 04.0 00.A0 .mN
NN.0 00.0 00.0 00.0 04.0 05.0 04.0 0A.4m No.0 0N.5N 00.0 N0.Nm 05.5 00.40 A4.0 45.00 .4N
0A.0 00.0 0N.0 40.0 04.0 40.0 5A.5 AN.0m 00.0 N0.0N 0A.0A 00.Nm 00.0 00.N0 A0.0 5A.m0 .mN
0A.0 00.0 00.0 00.0 00.0 50.0 04.0 00.0N AA.0 50.0N 00.0 0N.5N 00.5 N5.04 00.AA 00.A0 .NN
5A.0 00.0 5A.0 00.0 00.0 00.0 05.0 50.0N 00.4 40.0N 50.0 NA.0N 05.0 00.44 00.0 05.A0 .AN
0N.0 00.0 00.0 00.0 A4.0 05.0 50.0 Nm.Nm 0A.4 A0.0N 00.0 05.Am 0A.5 00.54 00.0A 0A.00 .0N
0A.0 A0.0 50.0 000.0 00.0 N0.0 A0.0 00.00 00.0 04.0N N0.0 0N.00 00.5 0A.54 00.0A 50.00 .0A
0A.0 00.0 0A.0 00.0 A4.0 05.0 00.4 40.Nm 0N.0 00.0N 00.0 00.Nm 00.5 A0.54 00.0 AA.Am .0A
NN.0 00.0 00.0 00.0 04.0 00.0 40.0 00.50 05.4 00.0m A0.N 00.04 50.5 NN.00 05.NA A0.00 .5A
4N.0 00.0 00.0 00.0 00.0 N0.0 N0.4 00.0N 40.4 A4.0N NN.5 00.0N 00.0 00.04 50.0 00.40 .0A
0A.0 00.0 0A.0 00.0 00.0 40.0 05.5 00.AA 00.0 00.4N 00.0 00.50 A5.5 40.04 AA.AA N0.mm .0A
0N.0 00.0 00.0 0A.0 04.0 00.0 4A.5 Nm.0m 0N.5 00.05 A0.0 00.04 45.0 N5.00 0A.0A 00.00 .4A
0N.0 50.0 0N.0 00.0 00.0 00.0 00.0 00.AA 04.0 04.0N 05.0 A0.5m 00.5 00.50 00.AA 00.50 .0A
5A.0 00.0 Nm.0 NA.0 00.0 00.0 00.0 0N.00 4N.0 05.0N N4.0 05.00 AN.5 40.50 A0.AA 0A.00 .NA
0A.0 No.0 0N.0 00.0 N4.0 05.0 0A.5 40.Nm 04.0 NN.4N 00.0 50.50 0A.5 00.40 AN.0A 00.00 .AA
0N.0 40.0 4N.0 00.0 00.0 40.0 05.0 4N.0m 00.0 00.0N 00.0 00.00 50.0 N5.00 00.AA 0N.00 .0A
0N.0 00.0 04.0 4N.0 04.0 00.0 0A.0 00.50 40.0 AA.0N 00.4 00.04 00.0 0N.00 00.4A N4.50 .0
05.0 00.0 5N.0 00.0 04.0 05.0 00.0 00.00 00.0 4N.0N 0A.0 04.N4 NA.0 40.00 54.AA 00.40 .0
00.0 4A.0 04.0 0N.0 00.0 40.0 0N.0 m0.Nm 00.0 00.0N 0m.N 00.04 A0.5 00.00 04.0A 40.00 .5
A0.0 AA.0 00.0 0A.0 04.0 N0.0 00.0 00.50 00.0 Nm.0N 50.0 00.04 A0.0 00.A0 05.0A 00.00 .0
5N.0 00.0 4N.0 00.0 04.0 00.0 00.0 N0.0N A0.0 0N.NN 00.5 00.00 05.0 4A.Am 4A.0 40.54 .0
00.0 00.0 4N.0 00.0 00.0 00.0 A5.5 54.00 5N.0 Nm.5N 55.0 4N.04 40.0 00.00 00.0 40.00 .4
NA.0 NA.0 0N.0 50.0 04.0 00.0 00.0 A0.5N AN.0 n4.NN 50.0 44.00 NA.5 00.N0 N0.0A 55.54 .m
00.0 0A.0 0N.0 50.0 04.0 N0.0 00.5 00.00 04.0 50.0N 55.0 00.A4 00.5 00.40 0N.AA 40.04 .N
A0.0 AA.0 0A.0 0A.0 04.0 00.0 N0.0 N5.0N 0N.0 04.0N No.0 5A.N4 55.5 00.00 A0.0A 50.54 .A
.Q.m 200: .Q.m 2002 .Q.m C00: .Q.m C002 .Q.m C002 oQ.m C602 .AA-W G602 .Q.m G002 QQOHU
sown:
5 4
mmADMAHm>

 

 

0mdm¢Hm¢> zm>mmlfifimom 20 003000 Add mom 024 mmbomw zw>mmlNBZNZB mom mZOHB¢H>mQ QM<QZ<BM 92¢ mdez Abomo

h XHQZEhAd

170

 

00.0

 

 

 

 

 

00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 40.0 00.0 0muoa
00.0 00.0 04.0 00.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 40.0 00.0 00.0 .00
00.0 04.0 00.0 00.0 00.0 00.0 40.0 00.0 40.0 00.0 00.0 40.0 04.0 00.0 00.0 00.0 .00
00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 40.0 04.0 00.0 04.0 00.0 00.0 00.0 .00
40.0 00.0 00.0 04.0 40.0 00.0 04.0 00.0 00.0 00.0 40.0 00.0 04.0 00.0 04.0 00.0 .40
00.0 00.0 00.0 00.0 40.0 00.0 00.0 04.0 00.0 00.0 04.0 00.0 00.0 04.0 00.0 00.0 .00
00.0 40.0 44.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 04.0 40.0 04.0 00.0 00.0 00.0 .00
00.0 00.0 04.0 40.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 .00
04.0 00.0 44.0 00.0 00.0 40.0 00.0 00.0 00.0 00.0 00.0 04.0 04.0 40.0 40.0 40.0 .00
00.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .00
00.0 00.0 04.0 00.0 00.0 00.0 00.0 40.0 00.0 00.0 00.0 40.0 04.0 00.0 00.0 00.0 .00
00.0 00.4 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 40.0 04.0 04.0 00.0 00.0 00.0 .00
00.0 00.0 04.0 00.0 00.0 00.0 04.0 00.0 40.0 40.0 00.0 00.0 04.0 40.0 00.0 00.0 .00
00.0 00.0 00.0 04.0 00.0 00.0 40.0 00.0 00.0 00.0 04.0 40.0 40.0 00.0 00.0 00.0 .00
00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 04.0 40.0 00.0 00.0 .40
00.0 04.0 00.0 00.0 00.0 00.0 44.0 40.0 00.0 00.0 00.0 00.0 40.0 40.0 04.0 00.0 .00
00.0 00.0 00.0 00.0 00.0 00.0 40.0 00.0 40.0 00.0 00.0 00.0 00.0 00.0 04.0 00.0 .00
00.0 00.0 00.0 40.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 40.0 00.0 00.0 00.0 00.0 .00
40.0 00.0 00.0 00.0 00.0 00.0 04.0 00.0 40.0 00.0 00.0 00.0 04.0 00.0 04.0 00.0 .00
00.0 04.4 40.0 00.4 00.0 00.0 00.0 00.0 04.0 00.0 00.0 40.0 00.0 00.0 00.0 00.0 .0
00.0 00.0 00.0 00.0 40.0 00.0 00.0 40.0 00.0 40.0 00.0 00.0 44.0 00.0 00.0 00.0 .0
00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .0
00.0 00.4 00.0 00.0 00.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .0
00.0 00.0 04.0 00.0 00.0 04.0 00.0 04.0 00.0 40.0 00.0 00.0 00.0 40.0 04.0 00.0 .0
00.0 40.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 40.0 00.0 04.0 00.0 .4
00.0 04.0 04.0 00.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .0
00.0 00.0 00.0 04.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .0
00.0 00.0 04.0 00.0 00.0 04.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 .0
.o.m :00: .0.0 :00: .0.0 :00: .0.0 :00: .o.m :00: .0.0 :00: .Q.m :00: .o.m :00: macaw
0000:
00 00 40 00 00 00 00
00030000>

 

 

 

GODCHuCOUIIm xHszmmd

171

 

 

 

 

 

 

 

 

 

 

00.0 40.0 00.0 00.0 04.0 00.0 04.0 04.0 44.0 00.0 00.0 00.0 00.0 04.0 00.0 04.0 00:08
00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 44.0 00.0 00.0 00.0 40.0 00.0 00.0 00.0 .00
00.0 00.0 00.0 00.0 04.0 00.0 04.0 00.0 04.0 00.0 40.0 00.0 00.0 00.0 00.0 00.0 .00
00.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 40.0 00.0 00.0 00.0 .00
00.0 00.0 00.0 00.0 04.0 00.0 04.0 00.0 04.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 .40
40.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 04.0 40.0 00.0 00.0 00.0 40.0 00.0 .00
04.0 40.0 00.0 00.0 44.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 04.0 00.0 04.0 00.0 .00
04.0 00.0 00.0 00.0 04.0 00.0 40.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 44.0 .00
00.0 40.0 00.0 00.0 04.0 04.0 04.0 00.0 04.0 00.0 40.0 04.0 00.0 00.0 00.0 00.0 .00
00.0 00.0 00.0 40.0 04.0 00.0 00.0 00.0 04.0 04.0 00.0 00.0 00.0 40.0 00.0 00.0 .00
04.0 00.0 00.0 00.0 00.0 40.0 44.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 40.0 .00
00.0 40.0 00.0 00.0 00.0 04.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 04.0 00.0 04.0 .00
04.0 00.0 40.0 00.0 04.0 00.0 40.0 00.0 04.0 00.0 40.0 40.0 40.0 00.0 00.0 04.0 .00
00. 00.0 00.0 00.0 04.0 00.0 40.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .00
04.0 00.0 00.0 00.0 00.0 00.0 04.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 40.0 00.0 .40
04.0 04.0 00.0 00.0 04.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .00
04.0 40.0 00.0 00.0 04.0 00.0 00.0 44.0 00.0 00.0 00.0 00.0 40.0 00.0 00.0 40.0 .00
44.0 00.0 04.0 00.0 04.0 04.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 04.0 00.0 04.0 .00
00.0 00.0 00.0 00.0 04.0 00.0 44.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .00
04.0 00.0 04.0 00.0 04.0 04.0 00.0 00.0 00.0 00.0 40.0 00.0 00.0 00.0 00.0 00.0 .0
04.0 40.0 04.0 00.0 40.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 04.0 40.0 00.0 .0
04.0 00.0 04.0 00.0 04.0 40.0 00.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 .0
00.0 00.0 04.0 00.0 04.0 04.0 00.0 00.0 00.0 00.0 44.0 04.0 00.0 04.0 00.0 04.0 .0
44.0 00.0 04.0 40.0 04.0 00.0 04.0 04.0 00.0 00.0 04.0 00.0 40.0 00.0 00.0 00.0 .0
04.0 40.0 00.0 00.0 00.0 00.0 04.0 00.0 40.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .4
04.0 00.0 04.0 40.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 .0
04.0 00.0 44.0 00.0 40.0 40.0 00.0 00.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 40.0 .0
40.0 00.0 00.0 00.0 04.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 00.0 40.0 40.0 00.0 .0
.o.m :00: .Q.m :mm: .a.w :00: .Q.m :mw: .Q.m :00: .Q.m :00: .Q.m :00: .Q.m :mm: @5000
mono:
40 00 00 00 00 00 00 00
000:00um>

 

 

©w5:0u:00llm XHQmem<

172

 

 

 

 

 

 

00.0 0m.0 «m.« 00.0 00.0 50.0 mo.« h0.m 00.0 0«.0 50.0 «0.0 00.0 0«.0 00.0 0«.0 «muoe
00.0 50.0 hm.« «0.0 05.0 mo.m 00.0 on.m ««.0 00.0 00.0 0«.0 50.0 «m.0 50.0 «0.0 .n«
00.0 50.0 0m.« 00.0 mn.0 «0.0 00.« 0n.m 0«.0 «0.0 00.0 00.0 «m.0 ««.0 00.0 0«.0 .0«
00.0 0«.0 00.« 0«.0 05.0 00.0 00.« 00.0 0«.0 00. 00.0 00.0 00.0 00.0 00.0 0«.0 .m«
00.0 «0.0 h0.« «m.0 «0.0 00.0 m«.« 0m.m >«.0 00.0 00.0 00.0 00.0 «0.0 00.0 0«.0 .0«
00.0 00.0 00.« 0«.0 00.0 m0.m m«.« 00.0 «0.0 ««.0 00.0 00.0 00.0 0«.0 00.0 0«.0 .mm
00.0 «0.0 «m.« «0.0 05.0 0o.m 00.0 m«.m 0m.0 0«.0 «0.0 m«.0 00.0 0«.0 00.0 ««.0 .««
00.0 0«.0 «0.« 00.0 00.0 0«.m 00.« 00.0 0«.0 00.0 00.0 «0.0 h0.0 00.0 m0.0 0«.0 .««
00.0 00.0 m0.« 00.m 00.0 00.0 00. m0.m mm. 00.0 00.0 >«.0 00.0 0«.0 00.0 h«.0 .0«
00.0 mm.0 n0.« 00.0 m0.0 00.0 00.0 00.m 0«.0 00.0 00.0 0m.0 00.0 0«.0 00.0 hm.0 .0«
00.0 «0.0 m0.« 00.0 00.0 0m.« 0«.« 00.m n«.0 00.0 0m.0 0«.0 n«.0 00.0 hm.0 0«.0 .0«
00.0 0«.0 hm.« «m.0 00.0 0«.0 «0.0 m0.m ««.0 00.0 00.0 >«.0 00.0 0«.0 00.0 0«.0 .n«
50.0 «m.0 «m.« 00.0 «0.0 ««.m n«.« h0.m 00.0 0«.0 00.0 0«.0 00.0 mm.0 00.0 mm.0 .0«
00.0 00.0 00.« mm.m 0n.0 0«.m 50.0 0m.m h«.0 00.0 00.0 0«.0 50.0 00.0 «0.0 ««.0 .m«
00.0 00.0 m0.« 0«.0 00.0 0«.m 00.0 00.0 00.0 0«.0 00.0 50.0 50.0 mm.0 0«.0 00.0 .0«
00.0 mm.0 0m.« 0«.0 05.0 h0.m ««.« om.m 00.0 00.0 00.0 00.0 00.0 00.0 00.0 0«.0 .m«
00.0 «0.0 ««.« 00.0 00.0 «0.m mo.« 00.0 00.0 50.0 00.0 00.0 00.0 «0.0 «m.0 ««.0 .««
h0.0 mm.0 00.« «0.0 0n.0 n0.m 00.« 00.m 00.0 00.0 00.0 00.0 00.0 «0.0 «0.0 ««.0 .««
00.0 00.0 m@.« 05.m «0.0 00.m 00.0 00.m 00.0 «0.0 00. 00.0 00.0 00.0 00.0 0«.0 .0«
00.0 0m.0 mn.« 05.0 «0.0 0«.m «0.« 00.0 00.0 0«.0 50.0 h«.0 «0.0 ««.0 00.0 00.0 .0
n0.0 «m.0 mm.« 00.0 00.0 0«.0 0«.« 0m.m 00.0 00.0 h«.0 00.0 «m.0 ««.0 hm.0 0«.0 .0
00.0 0«.0 00.« 00.0 «0.0 0«.m 0«.« 0«.0 00.0 00.0 «0.0 ««.0 0«.0 00.0 00.0 0«.0 .0
00.0 0«.0 m0.« 00.m 00.0 0«.m «0.« 00.0 50.0 «0.0 0«.0 h0.0 0«.0 00.0 «0.0 ««.0 .0
«0.0 0«.0 0m.« ««.0 00.0 ««.m 0«.« 50.0 00.0 00.0 «0.0 ««.0 00.0 «0.0 00.0 «0.0 .m
«0.0 0«.0 «m.« 00.0 «0.0 mm.m m0.« 00.0 50.0 00.0 00.0 0«.0 00.0 50.0 «0.0 0«.0 .0
00.0 00.0 0«.« 00.0 05.0 00.0 «0.0 00.0 00.0 00.0 00.0 «0.0 00.0 0m.0 00.0 50.0 .m
00.0 0m.0 «0.« «0.m 05.0 00.m «0.« 00.0 00.0 00.0 00.0 0«.0 «0.0 0«.0 «0.0 0h.0 .«
00.0 00.0 mm.« «0.m 00.0 00.m 00.« 0«.0 00.0 00.0 «m.0 ««.0 0m.0 0«.0 m0.0 0«.0 .«
.0.0 :00: .0.0 :00: .0.0 :00: .0.0 :00: .0.0 :00: .0.0 :00: .0.0 :00: .0.0 :00: 050:0
uoflm:
«m «m 0m 0« 0« >« 0« m«
m0«n0«u0>

 

 

vmscflucoollm XHQmem¢

173

 

 

 

 

 

«0.0 0m.« 00.0 «0.« «o.« 00.« «0.0 00.0 00.0« 0«.0N mm.« 00.« o«.« m«.« 00.0 00.0 «0u09
00.0 00.« 00.0 mo.« 00.0 0«.0 00.0 00.0 00.0« 00.0N «m.« ««.« m«.« 00.« 00.0 00.« .0«
00.0 m0.« 00.0 00.« 00.« 00.« 00.0 00.0 0m.m« 00.0N mm.« 0m.« mm.« «m.« mm.0 00.~ .0N
00.0 00.« 00.0 00.« 00.« «0.N 00.0 00.0 00.0« 00.00 o0.« 0«.N 0«.« «m.« 00.0 mm.~ .mm
00.0 00.« «0.0 00.« 00.0 00.~ 00.0 ~0.0 00.0« 00.0« 0«.« 00.« 0«.« «0.0 00.0 00.~ .0N
00.0 00.« 00.0 mo.« 00.0 00.« 00.0 «0.0 00.0« 00.0« N0.« «0.« m«.« 00.0 00.0 mm.« .mm
00.0 00.« 00.0 00.« 00.0 0«.0 0«.0 «0.0 00.0« «0.0« om.« 00.« 0o.« m~.« 0«.0 00.~ .«N
00.0 00.N 00.0 00.« 00.« 00.« 0«.0 00.0 00.0« 00.00 «N.« N«.« 0«.« «m.« 00.0 00.N .«N
00.0 00.0 00.0 00.« 0o.« 00.« 0m.o 00.0 00.0« 00.0N 0«.« 00.« 0«.« 00.« 00.0 00.« .0«
00.0 00.« «0.0 m0.« 00.« 00.N «0.0 00.0 00.0« m«.~m mm.« 00.« 0«.« 0«.« 00.0 00.0 .0«
00.0 «0.« «0.0 00.0 mo.« 00.« 0«.0 «0.0 00.0« 00.00 om.« 00.« 0«.« 0«.« 00.0 00.« .0«
00.0 00.« «0.0 m0.« «0.0 00.0 00.0 «0.0 00.0« 00.m« 0«.« 00.H 00.« 00.« 00.0 00.N .0«
00.0 mm.« 00.0 00.« 00.« 00.« 00.0 «0.0 00.0« «0.0m 0m.« 0N.« 00.0 0«.« 00.0 00.N .0«
00.0 «0.~ 00.0 00.« 00.« «0.« 00.0 00.0 00.m« 00.0« 00.« 00.« 0«.« 00.« 00.0 00.« .m«
00.0 «0.« 00.0 0m.« 00.« 00.« 00.0 «0.0 om.«« 00.00 0«.« 00.« 0«.« 0o.« 00.0 00.« .0«
00.0 0m.« «0.0 mm.« 0«.« 0m.« 00.0 00.0 00.m« 00.00 0«.« 0«.« oo.« 0«.« 00.0 00.~ .m«
00.0 00.« 00.0 00.« «0.0 no.0 00.0 00.0 0«.m« 00.00 mm.« 00.« 0«.« 0~.« 00.0 «0.N .N«
«0.0 mm.« 00.0 mm.« 00.« 00.« 00.0 00.0 «m.m« 00.m« 0«.« mo.« 0«.« m0.« «0.0 00.« .««
00.0 0m.~ «0.0 00.« 00.« om.« 00.0 00.0 00.0« 00.00 om.« 00.« 00.« 00.0 00.0 «0.« .o«
00.0 m0.« 00.0 00.« 00.0 00.« 00.0 00.0 om.0« 00.0N 0«.« 0m.« 0~.« 0N.« 00.0 00.« .0
«0.0 «m.« m0.0 00.« 00.0 00.0 00.0 00.0 0«.0« 00.m« 0«.« 00.« o«.« 00.« «0.0 «0.« .0
0«.0 00.« «0.0 00.« 00.0 00.« 00.0 00.0 00.m« 00.m« m0.« 00.« 0«.« «m.« 00.0 00.« .0
00.0 «0.« 00.0 m0.« 00.0 00.« 00.0 «0.0 0«.0« 0«.0N 00.« «o.« mm.« 00.« 00.0 00.« .0
00.0 m0.« 00.0 «0.« 00.0 00.0 00.0 00.0 00.0« m0.0« 0«.« 0~.« 00.0 «0.« 00.0 00.0 .m
00.0 00.0 «0.0 00.« 00.0 00.« 00.0 00.0 00.0« mm.0~ 00.0 «0.« 0«.« 0«.« 00.0 00.« .0
00.0 «m.« 00.0 00.« 00.0 00.0 00.0 00.0 00.0« 0«.mm 0«.« 0«.« 0«.« 0«.« 00.0 00.« .m
00.0 00.« «0.0 «0.« 00.0 «0.« «0.0 00.0 00.0« 00.0« mm.« 00.« m«.« 00.0 00.0 «0.« .«
00.0 mo.« 00.0 0o.« 00.0 00.« 00.0 00.0 «0.0« 0N.0« 0«.« 0«.« N«.« «0.« 00.0 «0.N .«
.0.0 :00: .Q.m :00: .Q.m :00: .a.m :00: .n.m :00: .n.m :00: .o.m :00: .Q.m :00: macho
uOn0:
00 0m mm 00 00 mm 00 mm
m0«n0«u0>

 

 

UmDCHHCOUIIm XHQZmEmd

174

 

 

 

 

 

 

 

 

00.« 00.0 00.« 00.0 00.« 0«.« 00.« 0«.0 0«.0 00.0 00.« «0.0 00.« 00.« «muoa
0«.« 00.« 00.« 0«.« 00.« «0.« 0«.« 00.0 0«.0 0«.0 00.« 0«.« 00.0 00.0 .0«
0«.« 00.« «0.« ««.« 00.« 00.« 0«.« 00.0« 00.« 0«.0 00.« 00.0 «0.« 00.« .0«
00.« 0«.« 00.0 00.0 «0.« 00.« 00.« ««.0« «0.m no.0 00.« 00.0 00.« 00.« .0«
0«.« 00.« 00.« 0m.0 00.« 00.« ««.« «0.0 00.0 00.0 00.0 0«.0 00.« «0.« .0«
«0.« mm.« 00.« «0.0 «0.« 00.« ««.« 0«.0« 0«.m 00.0 00.« 00.0 00.« 00.« .0«
««.« «0.« 00.« 00.0 00.« 00.« 0«.« 0m.0 0«.0 0«.0 «0.0 00.0 0«.m 00.« .««
m0.« 00.« 00.« 00.0 «0.« ««.« «0.« 0«.0« 00.« 0«.0 0«.« 0«.0 «0.« 0«.« .««
«0.« 00.« 00.« 0«.« «0.« 00.« m0.« 00.0 «0.« 00.0 «0.« 0m.m 00.« 00.« .0«
0«.« mo.m 00.« «0.0 00.« 00.« 00.« 0«.0 00.m 0«.0 m0.« 0«.0 00.« 00.0 .0«
«0.« 00.0 00.« «0.0 «0.« ««.« 0m.« 00.0 00.0 00.0 «m.« 0«.0 00.« 00.« .0«
«0.« 00.« «0.0 0m.0 00.« «0.« 00.« «m.0« «0.« 00.0 00.« 00.0 00.« 00.« .0«
0«.« 00.« 0«.« 00.0 00.« 00.« 00.« 00.0« 00.« «0.0 0«.« 00.0 0«.0 00.« .0«
00.« 00.« «0.« 00.0 «0.« 00.« 00.« 00.0 00.0 00.0 00.« 00.0 00.« 0«.0 .0«
0«.« «0.0 0«.« m0.0 00.« 0«.« 00.« 00.0 00.« «0.0 00.« 00.0 0«.« 00.« .0«
0m.« 00.0 00.0 00.0 0«.« «0.0 00.« 00.0 m0.m 0«.0 00.0 «0.0 0«.« 00.« .0«
«0.« 00.0 00.« 00.« «0.« 00.« «0.0 0«.0 «0.0 00.0 00.« 0«.0 «0.« 00.« .««
0«.« 00.0 00.« 00.0 0«.« 00.0 00.« 00.0 00.0 00.0 00.« 00.0 00.« ««.m .««
0«.« 00.0 0«.« 00.0 00.« 00.« 00.« «0.0 00.0 00.0 00.« 00.0 mm.« «0.« .0«
0m.« 00.m 00.« 00.0 00.« 00.« «0.« 0«.0 0«.0 00.0 00.« 00.0 «0.« 00.« .0
00.« 00.0 ««.« «0.0 «0.« 00.0 0«.« 00.0« «0.« m0.0 «0.« 0«.0 «0.« 00.« .0
«m.« 00.0 00.« 00.0 ««.« 00.0 00.« 00.0 00.« 0«.0 0«.0 00.0 «0.« ««.« .0
00.« 0«.0 00.« «0.0 00.« ««.« 00.« 00.0 00.« 0«.0 «0.« 0«.0 00.« 0«.« .0
00.« 00.0 00.« mm.0 0«.« 00.0 mo.m 00.0 00.« 0«.0 00.« «0.0 00.« 00.« .m
00.« 00.0 m«.« 00.0 0m.« 00.0 mm.« «0.0 0«.m «0.0 0«.m 00.0 00.« 00.« .0
m«.« 00.0 00.« 00.0 00.« «0.0 00.« 00.0 00.0 «0.0 00.« 0«.0 «0.0 00.0 .m
0«.« 00.0 ««.« 00.0 m«.« 00.0 00.« 00.0 ««.m 00.0 00.« 0«.0 00.« 0«.m .«
0«.« 0«.0 0«.« 00.0 ««.« «0.0 00.« 0«.0 00.0 «0.0 0«.0 «0.0 0m.m 00.« .«
.0.0 :00: .0.0 :00: .0.0 :00: .:.0 :00: .Q.m :00: .:.0 :00: .:.m :00: 050:0
“000:
00 00 00 00 m0
00«90«u0>

 

 

©05CHUCOUIIW XHQZMde

"7'11’7111411'1'11111111