A STUDY OF THE ACADEMIC PROGRESS
OF STUDENTS ADMTTBD TO THE UNIVERSITY
OF TEXAS AT AUSTIN UNDER THE
PROVISIONAL ADMISSION PROGRAM 1972-1973

Dissertation for the Degree of PhD.

mamm STATE UNIVERSITY

DOUGLAS FORREST BL'CONKBY

.1972

Ill W H l \Nlll \\ WW\lllllllllllll

3 1293 10414 6984

This is to certify that the

thesis entitled

A Study of the Academic Performance of Students
Admitted to the University of Texas at Austin
Under the Provisional Admission Program.

presented by

Douglas F. McConkey

has been accepted towards fulfillment
of the requirements for

Ph.D.

 

degreein Higher Education

 

 

Van C. Johnson

 

Major professor

Nov. 1974

 

Date

0-7539

\--.

v‘..‘

\
“~.

 

.u‘ ‘

‘A.

‘vu.

5“

"I.
H.‘

ox
\\ x
. “P

fr
,Q/\, ABSTRACT

F

l KY :'

\v/‘ A STUDY OF THE ACADEMIC PROGRESS
OF STUDENTS ADMITTED TO THE UNIVERSITY
OF TEXAS AT AUSTIN UNDER THE
PROVISIONAL ADMISSION PROGRAM
1972-1973

BY

Douglas Forrest McConkey

Problem

While college enrollments show signs of leveling, there
continues to be a significant demand for the services offered
by institutions of higher education. And, with more institu-
tions to choose from, the student is faced with important
but complex decisions; similarly, colleges and universities
must continue to address themselves to the development of
effective methods for selecting from prospective students
those who show the highest probability of success in a given
academic setting.

The objective of this study is to gain insight into
some of the factors that influence the academic progress of
marginal students admitted to a moderately selective public
university. This insight can help the university in develOp-
ing better programs to assist such students; it can also be

of benefit in developing more useful means of identifying

Douglas Forrest McConkey

those students who have an optimal chance for success at the

University.

Significance

 

With some colleges under pressure to recruit more stu-
dents, there is a noticeable trend away from helping students
find the right institution, to convincing them to select a
particular one. The resultant opening of some institutions
to previously non-admissable students poses serious ethical
questions to the conscientious admissions officer.

Studies such as this one must be continued, and the
results used, in order to help students and institutions
determine the most propitious match for each. The focus must
be away from coercion and toward helping the student find an

institution suitable for his or her abilities and interests.

Population

 

This thesis examines the progress of four groups of
Provisional Admission students at the University of Texas
at Austin identified by their first semester enrollment.
For comparison purposes, four groups of regular students
admitted to the university in the same semesters are also
studied. Subgroups based on sex and ethnic background are

also identified for the Provisional and regular samples.

Douglas Forrest McConkey

Method of Analysis

 

Each of the Provisional and regular samples are ana-
lyzed according to their high school and SAT performance,
and the groups compared accordingly. The college performance
is measured by first semester and total grade point averages,
as well as a test of the students' persistence in remaining
in school. Sex and ethnic identification are used as indep-
endent variables.

Analysis of variance with post-hoc comparisons is used
to examine the factors that seem to effect college performance.

Finally, multiple regression equations are developed for
several measures of ability used to predict first semester

and total grade point averages.

Findings and Conclusions

 

Provisional students enter the university with markedly
inferior records when compared to regular students. Provis-V
ional males present different qualities in the high school
and SAT performances than do Provisional females, but the
college performance of each subgroup is about the same.

Although regular students achieve clearly superior
college grades, there is not a large difference in the ten-
dency of regular and Provisional students to remain enrolled

in school.

I]

Douglas Forrest McConkey

For predicting performance it is clear that high school
grades and rank, as well as SAT scores, have varying levels
of influence on the college grades of regular and Provisional
students.

Continued development of counseling and support services
for Provisional students is implied by the findings. The
program can reasonably be maintained, but to do so requires
that the university pay attention to the special needs of the
marginally admissable student. Thetxnnxnmsnature of predic;
tive techniques is reaffirmed; consequently, the university
must be willing to continue lending support to the signifi-
cant number of students who will encounter academic diffi-

culty.

A STUDY OF THE ACADEMIC PROGRESS
OF STUDENTS ADMITTED TO THE UNIVERSITY
OF TEXAS AT AUSTIN UNDER THE
PROVISIONAL ADMISSION PROGRAM

1972-1973

BY

Douglas Forrest McConkey

A DISSERTATION

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

DOCTOR OF PHILOSOPHY

Department of Administration and Higher Education

1975

ACKNOWLEDGEMENTS

I would like to give my sincere thanks to Dr. Van
Johnson, guidance committee chairman, for his encouragement
and support. Also, I am in debt to Dr. Paul Kelley and Ms.
Jody Slover of the University of Texas for their advice and
assistance. The other members of my committee, Dr. Max
Raines, Dr. William Sweetland, and Dr. Albert Blum, provided
additional help.

I am especially grateful to Dr. David Hershey for the
support he has given.

The dedication of this document, and my degree, is
given to my parents. Their constant faith and the pride
they've had from this accomplishment, make it especially
meaningful for me. This is a way of saying thank you for

all they have done over the years.

ii

TABLE OF CONTENTS

Chapter

I.

II.

III.

IV.

INTRODUCTION AND STATEMENT OF

Introduction. . . .
The Problem . . . . .
Theory. . . . . . . .
Significance. . . .

Definition of Terms .

REVIEW OF THE LITERATURE.

DESIGN. . . . . . . .
Population. . . . .
Method of Analysis. .
Data. . . . . . . . .

Null Hypotheses . . .

RESULTS . . . . . . .

THE

Demographic Data for Samples.

PROBLEM

Mean High School and College Performance.

Tests of Significance of Difference

Analysis of Variance.

Post-Hoc Comparisons.

Degree of Persistence .

iii

Page

ll

14

27
27
28
33

34

36
36
38
54
73
82

100

TABLE OF CONTENTS (continued)

Chapter Page
IV. Correlation Coefficients. . . . . . . . . . . lll
Multiple Regression Analyses. . . . . . . . . 123

V. SUMMARY AND CONCLUSIONS . . . . . . . . . . . . 137
Summary . . . . . . . . . . . . . . . . . . . 137
Conclusions . . . . . . . . . . . . . . . . . 148
Implications for Research . . . . . . . . . . 152
FOOTNOTES . . . . . . . . . . . . . . . . . . . . . . 156
BIBLIOGRAPHY. . . . . . . . . . . . . . . . . . . . . l6l

iv

Table

1.

4a.

4b.

4C.

4d.

4e.

4f.

Composition

LIST OF TABLES

of regular and Provisional

samples by sex. . . . . . . . . . . . . . . .

Composition of regular and Provisional
samples by ethnic background. . . . . . . . .

Mean scores
performance
students. .

on high school, test and college
for regular and Provisional

Frequency distribution of high school rank

for regular

and Provisional students. . . . .

Frequency distribution of SAT-Verbal for
regular and Provisional students. . . . . . .

Frequency distribution of SAT-Math for

regular and

Provisional students. . . . . . .

Frequency distribution of SAT-Total for
regular and Provisional students. . . . . . .

Frequency distribution of first semester

grade point
Provisional

average for regular and
students. . . . . . . . . . . . .

Frequency distribution of total grade point

average for

Mean scores
performance
students by

Mean scores
performance
students by

Comparison of regular and Provisional students
using 2 distribution, one-tailed test a = .05;

regular and Provisional students.

on high school, test, and college
for regular and Provisional
sex 0 O O O O O O O O O O O C O O

on high school, test, and college
for regular and Provisional
ethnic background . . . . . . . .

critical value = 11.645 . . . . . . . . . . .

Page

36

37

39

43

44

45

46

47

48

50

51

57

{.1

Table

8.

10.

11.

11a.

12.

13.

14.
15.
16.
17.

18.

19.

LIST OF TABLES (continued)

Comparison of male regular and male
Provisional students using 2 distribution,
one-tailed test; a =

$1.645.

.05,

critical value =

Comparison of female regular and female
Provisional students using 2 distribution,
one-tailed test; a =

11.645.

.05,

critical value =

Comparison of Mexican-American regular and
Mexican-American Provisional students using
t distribution, one-tailed test . . . . . .

Comparison of male and female Provisional
students using 2 distribution, one-tailed
05, critical value = 11.645. . .

test; a =

t-tests for grade point averages in English,
Science/Math and Social Science . . . . . .

Comparison of male regular and female regular

students using 2 distribution, one-tailed test;
a = .05, critical value = $1.645. . . . . . . .

Comparison of white Provisionals and Mexican-
American Provisionals using 2 distribution,
one-tailed test; a =

11.645.

Analysis
Analysis
Analysis
Analysis

Analysis

of
of
of
of

of

variance
variance
variance
variance

variance

point average . . .

.05,

for
for
for
for

for

critical value =

SAT-Verbal . . . .
SAT-Math . . . . .
rank in class. . .
first semester GPA

total college grade

Analysis of variance for total hours completed.

vi

Page

58

59

60

64

66

69

71
73
75
76
77

78
79

ll

.“

'5'

he.

(I!

Table

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

LIST OF TABLES (continued)

Analysis of variance for total grade points
earned I O O I C O C O O O I O O C O O C O O O O 0

Mean SAT, high school rank, and college grade
point averages for regular and Provisional
students by first semester enrolled - males,
females and combined. . . . . . . . . . . . . . .

Comparison of SAT-Verbal scores for regular
and Provisional students by first semester
enrOIled. O O O O O O O O O O O O O O I O O O O 0

Comparison of SAT-Math scores for regular and
Provisional students by first semester
enrOIJ-ed. O O O O I I O O O O O O O I O O O O O 0

Comparison of high school rank for regular and
Provisional students by first semester enrolled .

Comparison of first semester grade point average
for regular and Provisional students by first
semester enrolled . . . . . . . . . . . . . . . .

Comparison of total grade point average for
regular and Provisional students by first
semester enrolled . . . . . ... . . . . . . . . .

Current enrollment of regular and Provisional
students controlling for sex and first semester
enrOI-led. I O O O O O I O O O I O O C O O I O O 0

Chi square contingency table for persistence
in enrollment of all Provisional versus all
regular students. . . . . . . . . . . . . . . . .

Chi square contingency table for persistence
in enrollment of male Provisional versus male
regular students. . . . . . . . . . . . . . . . .

Chi square contingency table for persistence

in enrollment for female Provisional versus
female regular students . . . . . . . . . . . . .

vii

Page

80

83

86

88

90

92

94

102

103

104

104

Table

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

LIST OF TABLES (continued)

Chi square contingency table for persistence
in enrollment of male Provisionals versus
female Provisionals . . . . . . . . . . . . .

Chi square contingency table for persistence
in enrollment of male regular versus female
regular students . . . . . . . . . . . . . .

Chi square contingency tables for regular and
Provisional students by first semester
enrOJ-led O O O O I O I O O O I O O O O C O C

Current enrollment of regular and Provisional
students controlling for ethnic background
and first semester enrolled . . . . . . . . .

Chi square contingency table for white versus
Mexican-American Provisional students . . . .

Chi square contingency table for white versus
Mexican-American regular students . . . . . .

Correlation coefficients for regular students'

high school rank, SAT scores, and college
grade point averages . . . . . . . . . . . .

Correlation coefficients for Provisional
students' high school rank, SAT scores, and
college grade point averages . . . . . . . .

Correlation coefficients for male Provisional
students' high school rank, SAT scores, and
college grade point averages . . . . . . . .

Correlation coefficients for female Provisional

students' high school rank, SAT scores, and
college grade point averages . . . . . . . .

Correlation coefficients for male regular

students' high school rank, SAT scores, and
college grade point averages . . . . . . . .

viii

Page

105

106

107

109

110

111

112

114

117

118

120

Isl

LIST OF TABLES (continued)

Table Page

42. Correlation coefficients for female regular
students' high school rank, SAT scores, and
college grade point averages . . . . . . . . . . 121

43. Model 1: Multiple regression using SAT-
verbal, SAT-Math, and high school academic
grade point average to predict first
semester grade point average . . . . . . . . . . 124

44. Model 2: Multiple regression using SAT-Verbal
SAT-Math and high school academic grade point
average to predict total grade point average. . . 124

45. Model 3: Multiple regression using SAT-Total
and academic grade point average to predict
first semester grade point average. . . . . . . . 125

46. Model 4: Multiple regression using SAT-Total
and academic grade point average to predict
total grade point average . . . . . . . . . . . . 125

47. Model 5: Multiple regression using SAT-Verbal
SAT-Math, and high school overall grade point
average to predict first semester grade point
average . . . . . . . . . . . . . . . . . . . . . 125

48. Model 6: Multiple regression using SAT-Verbal,
SAT-Math and high school overall grade point
average to predict total grade point average. . . 126

49. Model 7: Multiple regression using SAT-Total
and high school overall grade point average to
predict first semester grade point average. . . . 126

50. Model 8: Multiple regression using SAT-Total
and high school overall grade point average
to predict total grade point average . . . . . . 127

51. Model 9: Multiple regression using SAT-Verbal,

SAT-Math, and high school rank to predict
first semester grade point average . . . . . . . 127

ix

{II

Table

52.

53.

54.

LIST OF TABLES (continued)

Model 10: Multiple regression using SAT-
Verbal, SAT-Math and high school rank to
predict total grade point average . . . . .

Model 11: Multiple regression using SAT-Total
and high school rank to predict first semester
grade point average . . . . . . . . . . . . .

Model 12: Multiple regression using SAT-Total
and high school rank to predict total grade
point average . . . . . . . . . . . . . . .

Page

128

128

131

Ill

C H A P T E R I

INTRODUCTION AND STATEMENT OF THE PROBLEM

Introduction

 

"With colleges and universities becoming more
accessible to youth from all segments of our society, it is
apparent that indices of probable future success are in—
valuable in assisting admissions officers and counselors at
their tasks. More importantly, educators are becoming
increasingly aware that the validity of existing methods of
selecting students must be scrutinized carefully. Because
our society's demand for technological and intellectual
competence places the stakes quite high, no student should
be unjustly barred from gaining entrance to an inStitution
of higher education."1 Thomas and Stanley have set the
framework for this study.

Though the vicissitudes of the employment picture
have brought under some question the market value of a
college degree, and while some colleges are now experiencing
at least a temporary leveling in enrollment, there is still
a significant demand for higher education. The Chronicle of

Higher Education (October 1, 1973)2 reported that Fall 1973

"KC
9-“

.
v. 2"
V‘seco

eegn‘

viii:

~.\-y|
—

v iv“
.

44. U

bu

rf

enrollment at institutions of higher education was greater
than ever before. And, while the Carnegie Commission has
revised its projections for the 1970's downward,3 there
remains the expectation that college enrollments will be
one-third higher in 1980 than they were in 1970. Conse-
quently, we may expect to see a continued demand for the
services offered by colleges and universities, particularly
at higher quality institutions with selective admission
policies. ‘We may also expect that those students who have
historically been barred from higher education (marginally
qualified students, minorities, and educationally and
socially disadvantaged) will continue to demand that
colleges and universities become more accessible and respon-
sive to their needs.

To effectively respond to such students, a university
must have current information about their probable degree of
academic success. The university must know which factors in
the baCkgrounds of these students tend to be strong indica-
tors of success in college, and which factors tend to
indicate probable failure. This study addresses these
questions as they relate to one moderately selective, large,
public university. The overall objective is to make the
university better able to respond to its constituency, I

either by encouraging the enrollment of certain types of

marginal students, or through providing better evidence as
to why such students should find alternate routes to

higher education.

The Problem

 

The central objective of this study is to examine
the academic performance of four groups of students admitted
under the Provisional Admission Program at The University of
Texas at Austin.

The University can be classified as a moderately
selective public university, according to Astin's defini-
tion.4

In Astin's groupings of various institutions,
according to the mean score on the Scholastic Aptitude
Test of each student body, The University of Texas at Austin
would be placed at roughly the 60th percentile; that is,
about 40% of the institutions in Astin's study would have
higher mean SAT scores than does the University.

As a moderately selective institution, the Uni-
versity bars some residents of the State of Texas from
regular admission because they do not meet the following
admission requirements:

Top 10% of graduating class, admitted regardless

of SAT

Top half of graduating class, admitted with SAT
total of 800
Bottom half of graduating class, admitted with

SAT total of 1000

The Provisional Admission Program was initiated in
the 1960's in response to public demand that students who
were denied admission be given an Opportunity to attempt
courses at the University in an effort to prove their
ability to succeed. Thus, a Texas resident who graduates
from high school and who has no prior college experience,
may begin at The University of Texas at Austin under the
Provisional Admission Program, even though the student does
not meet the regular admission requirements. The student
admitted under this program must enroll in either a Spring
or Summer semester. Twelve semester credit hours (four
courses) of course work are taken for the first semester
(six hours in each of two six—week sessions when the student
begins in the Summer semester). The student must maintain
an overall 2.0 grade point average during the first semester
(on a 4.0 scale), and pass every course to be eligible for
continuation as a regularly admitted student in the subse-
guent semester. The areas from which the student may
select courses while under the Provisional program are

English, natural sciences, mathematics, and foreign

n.“,,.

u
.1
(h

E

cal!!!
‘ .
U‘VV

H 9

‘qefl
v

:VIO‘I

.
e_Vc
ii. I
b

332:

‘A;
D'vl.

Q
‘5
«V

«k—
zi-

“on

*5.

.4

languages. Courses in other academic disciplines, such as
social sciences, fine arts, education, etc., normally
available to regularly admitted freshmen, are only avail-
able to Provisional students after the first semester is
successfully completed with at least a 2.0 (C) average.
This study examines the high school records and
Scholastic Aptitude Test scores of Provisional students,
particularly as they relate to subsequent academic perfor—
mance in college. Particular attention is directed toward
the relative value of various types of predictors of
performance. Rank in class, overall and academic high
school grade point averages, SAT Math and Verbal scores,
and various combinations of these criteria are studied as
they relate to college performance. Of interest also is a
comparison of college academic performance of Provisional
students to that of regularly admitted students. Sex and
ethnic background are also used as independent variables.
The overall pattern of academic success in college is
examined, to ascertain those academic disciplines in which
Provisional students seem to meet with the most success.
The objective of the research is to gain a better under-
standing of the type of academic background and tested
ability necessary for successful completion of the Provi-
sional Admission Program and subsequent academic success at

the University. In this way, the research demonstrates the

“I

 

most helpful criteria upon which to base decisions regarding
the admission of similarly defined "marginal" students in
the future.

A corollary to the above analyses is the comparison
of the performance of two separate groups of Provisional
students admitted for the Summer 1972 and Summer 1973
sessions. The Summer 1973 Provisional students were given
a more in-depth orientation to the University than were the
Summer 1972 students. In particular, services such as
tutoring, counseling, academic advising, and reading and
study skills development programs were introduced. Students
were urged to take advantage of these services, and there
was a more systematic attempt to inform Provisional students
of these various types of academic support services avail—
able on campus. The frequency of use of these services
cannot be compared for the two groups in question because of
lack of available data; however, any noticeable differences
in the degree of success experienced by the two groups of
students are of interest. Whether such differences could
reasonably be attributed to the benefits of the support

programs offered is a related concern.

Theory

The generally accepted notion in the area of

academic prediction is that prior academic performance

(grades and/or rank in class), used in conjunction with a
sound test of scholastic aptitude, affords the best avail-
able indication of the likelihood of success in college.5
The success with which a student completes a high school
curriculum indicates his or her likelihood of success in
similar tasks in the future, particularly a college curric-
ulum. School performance is thus an indication of a
person's ability, but also shows something about the
motivation of the student. Ability testing may also be an
important factor in assessing a student's likelihood of
success, since some minimal level of aptitude for college
study is necessary regardless of the level of motivation.

There has been considerable discussion over the use
of past performance and aptitude tests for restricting
admission to colleges and universities. It has been obvious
that our prediction methods have been imperfect, particularly
since relatively large percentages (50%—70%) of students who
begin college fail to graduate.6 Clearly, not all students
for whom we predict academic success do, in fact, succeed.
In this respect, an argument can be made in favor of the
abandonment or significant alteration of our present methods
for determining who is to gain access to our colleges and
universities.

Some efforts have been made to develop new criteria

‘Upon which to base such decisions. Lavin reports on the

l
.uau

...\u

I
ha,»-

.5].

hr-

r11

developments in the areas of personality measures and socio-
logical determinants.7 However, the findings of such
research have been inconclusive, and certainly not depend-
able enough so that one could base admissions decisions on
them.

As a result, colleges and universities continue to
formulate admission policies relating to past performance
and tested aptitude. The pressing need is to refine these
measures. Research that assesses the conditions at a
particular institution is very valuable. By examining its
own students' performance, an institution can more effective-
ly develop admission policies that serve the interests of
the student and community. Perhaps more refined methods of
predicting college performance would show that marginal
students at some institutions should be treated differently
from other students in the admission process. For example,
Scott reports that we should not weigh the test scores as
heavily for marginal students (his were minority students
only).8 Hardie and Anderson9 suggest that we should put
more importance on balanced SAT scores (math and verbal
scores about equal) for marginal students than we do for
higher achieving students.

Evidence such as this implies that further study is
laeeded to (1) improve the information on which admission

Ciecisions are based, and (2) account for local conditions

that may influence a student's performance at a particular
institution. So far, results of research have lead to the
use of past academic performance and tested ability with a
moderate degree of confidence. Refining these tools so
they are more relevant to specific kinds of students and
institutions is one area in which continuing research is
needed.

Therefore, the focus of this study is an in—depth
look at our traditional criteria (tested aptitude and past
performance) as they relate to one type of student and one
type of institution (in particular-~marginal students in a
moderately selective public university). It is important to
further refine the various types of intellective predictors
for marginal students to permit better policy decisions
regarding the admission and support of such students. Then
an investigation of the usefulness of more subjective and
illusive non—intellective measures would be helpful. Most
fundamentally, it is essential to determine the amount of
success that marginal students realize, then assess the
advisability of opening the institution to such students in

the future.

Significance
Many institutions of higher education are eXperi—

encing financial difficulties due to a leveling or, in some

I'll

10

cases, a decrease in enrollments. Some institutions are
finding themselves overextended because of large scale
building programs in the 1950's and 1960's that were based
on projected enrollment growths that have not materialized.
As a result, there is an obvious trend toward perceiving the
admission process as recruitment of students rather than
a counseling and educational service. Attention is turning
away from helping students find the right institution, to
convincing them to choose a particular one. The opening of
some colleges and universities to previously non—admissible
students poses important questions as to the likelihood of
success of these students. Consequently, serious ethical
considerations in the admission process must be faced.
Studies such as the one made here must be continued
and the implications of the findings heeded. The admission
officer should be of central importance in the college going
process, because that person presumably has access to the
important research data relating to college success. Now he
is also at the center of an important ethical issue. The
needs of the institutions for financial support must be
balanced against those of the student for sound advice
regarding his or her educational goals. The assumption must
be that the student is of primary importance; if this
principle is accepted, there is an implied obligation to

develop better means of identifying those students who stand

 

n.
cf)

,n; H
Ho!
J

ku-
hu-

r1

11

to profit from higher education at each of the many and
varied types of institutions. The research data may be
used by the admission officer to counsel students into
attending a certain college or university if that is
appropriate. It must also be used to counsel the student
toward other institutions if that is what appears best for
a given individual.

This study, then, is important for the tools it can
give us to help students and professional educators make
the most appropriate decisions concerning the post secondary
educational and vocational Opportunities Open to students.
The accompanying concern is, of course, the extent to which

such tools are used for the benefit of students.

Definition of Terms

Academic Courses: (sec. ed.) Pertaining to the

 

fields of English, foreign languages, history, economics,

mathematics, and science.10

Credit Hour: The unit by which an institution may

 

measure its course work. The number of credit hours assigned
to a course is usually defined by the number of hours per
week in class and the number of weeks in the session.11

Credit Hour! Semester: A credit hour based on the

 

semester calendar which is usually divided into terms of

11 weeks.12

12

Differential Prediction: Prediction of differences

 

in achievement levels attained in pairs Of school subjects
or fields of endeavor, as contrasted with general prediction
of success in a single activity or in several activities.13

Grade Point Average: A measure of average scho—

 

lastic success in all school subjects taken by a student
during a certain term or semester, or accumulated for several
terms or semesters; obtained by dividing grade points by
hours of course work taken, when course marks are weighted

by some such system as the following to Obtain grade points:
A-40, B—30, C—20, D—lO, F—O; most commonly used at the
college level.14

Marginal Student: A marginal student is one who

 

is close to meeting the admission requirements Of a particu-
lar institution; a marginal student at one institution may
not be classified as such at another institution because Of
variances in admission requirements.

PersistenCe: A term used to indicate a student's

 

or group of students'tendency to remain enrolled at the
University. For the purpose of this study, a student is
said to have "persisted" if he/she was enrolled in the
Spring 1974 semester.

Rank in Class (Class Standing): The relative

numeric position of a student in a defined p0pulation of

13

students based upon grade point averages or some other

measure of scholastic achievement.15

Selective Admission: Exercise of discretionary

 

powers of selection by an institution Of higher education.
Selection is based on merit as determined by the college
rather than on order of application and is limited to those
who are judged to have a reasonable probability of success

in their chosen program Of studies. Academic achievement,
recommendations, reports on character, personal qualities,
and entrance examinations are important elements in the
selective admission procedure. Personal interviews are some-
times required. The policy is sometimes referred to as

restrictive admission.16

C H A P T E R I I
REVIEW OF THE LITERATURE

Research on the prediction of academic success in
college is voluminous. Results from such research, however,
have been somewhat limited in scope. Most studies report
that college academic achievement is tied most closely to
high school performance and tested academic aptitude.17

Various studies have looked at specific aspects of
grades and test scores as they are used in predicting college
success. Loeb18 found that academic grade point average
(baSed only on English, math, science, social studies, and
foreign language courses) in high school predicts only as
well as overall grade average. She saw no need to separate
academic courses from other courses when evaluating a
student's credentials. Gelso and Klock19 had substantially
the same report; however, their students were junior college
entrants and the generalization to four-year colleges may
not be warranted. In an earlier study, Hills and Klock20
found that academic and total grade point average in high,

school predicted equally well for students within each of

several types of higher education institutions. The study

14

HA

15

covered students entering junior colleges and fourvyear
public and private colleges. Each of the schools under
consideration used tests in conjunction with high school
grades to predict success. The results may thus not be
applicable to institutions that rely solely on grades.
Arvey21 found that students whose grades were in-
' consistent over time were less predictable in terms of
future performance than were students whose performance was
fairly constant. He also noted that consistency in math,
natural sciences, and total grades was particularly important
for the sake Of making reliable predictions. Consistency in
English and social sciences did not appear to be as crucial.
0n the other hand, Reyes and Clarke22 foundno difference
in the predictability Of future performance of inconsistent
versus consistent high school students. "It would appear
that counselors should use considerable caution in attri-
buting significance to variability manifested by a student's
academic record."23 Loeb24 concurred with this. She said
that consistency in various curricular areas should not be
used as a factor in predicting futUre academic performance;

the matter, however, deserved further study in her view.

Elton25 found that students whose grades improved in'

the senior year of high school did not perform significantly»

better in college than did those students whose grades

declined or remained constant.

l1

16

Hence, for a student to be accepted or
rejected on the basis of an improvement or

decline in his senior year grade average is

a clinical decision. As such, it may be

justified by many different rationalizations,

but it cannot be justified as rational.26
Loeb27 also concurred with this finding that improvement in
senior grades was not significant, but again she felt the
matter needed further study.

Horst28 conducted studies in which he attempted to
predict performance in specific curricular areas. He showed
that there was some benefit in using past performance in
verbal or quantitative tasks to predict future success in
each area respectively. This "differential prediction"
showed promise when used in conjunction with prediction of

29 reported, however, that studies

overall performance. Lavin
by Cronbach, Berdie, and Eells refuted such findings and that
the extra effort involved in gathering and analyzing the
necessary data may not be worthwhile, given the small
improvement in predictive ability.

The American College Testing Corporation and the
College Entrance Examination Board have reported the
effectiveness of their respective tests in predicting
academic performance in college.

In a typical report, the ACT scores were shown to
have a median multiple correlation of .465 with overall

college grade point average. When used in conjunction with

high school grades, the ACT was shown to correlate even

I]

17

higher (about .58) with overall GPA in college.30 Each of
the subtests correlated as follows with college grade point

average in the respective curricular areas:

English .43
Mathematics .42
Social Studies .47
Natural Sciences .47

The median correlation of high school and college

grade point average in each of these areas was very

similar:
EngliSh .44
Mathematics .43
Social Studies .46
Natural Sciences .49

The combination Of high school grades and ACT scores

in each curricular area correlated most closely with college

performance in each area:31
English .52
Mathematics .51
Social Studies .55
Natural Sciences .59

In 1971, the College Entrance Examination Board
reported the following correlations between college grade

point average and the SAT-Verbal and SAT-Math, for men and

women:32

COMBINED
MEN WOMEN (Men and WOmen)

SAT-Verbal with total
college grade point average .33 .41 .39

SAT-Math with total
college grade point average .30 .36 .33

18

These results were for students in liberal arts curricula
in college, but similar findings were reported for students
in other academic areas.

Interestingly, there has been evidence that the
validity of the SAT tends to be lower in more highly
selective colleges and universities.33 This seems to be due
to the reduction Of heterogeneity of ability in such
institutions. Generally, the lower the standard deviation
Of SAT scores in a given institution (i.e., the more
restricted the range), the lower will be the test's
predictive validity.

When high school record, as indicated by rank in
class, was used with the SAT in a multiple correlation,
the predictive validity rose to .62.34 This indicates that
a much stronger predictor was Obtained when SAT and school
performance were combined and both taken into account.

Astin35 confirmed that tests are useful predictors
but noted little difference between the effectiveness of
using a composite or total score and using the various sub—
scores available on either the SAT or ACT,

...in general, the Scholastic Aptitude Test,

National Merit Scholarship Qualifying Test, and

American College Test are interchangeable for the

purpose of predicting the student's overall

freshman GPA in college; use of scores on the
separate subtests of these batteries, instead of

the simple composite scores, adds little predictive
value.

I“

19

Munday36

confirmed this interchangeability, noting that the
tests did not necessarily measure the same things, but can
serve the same predictive purpose; he felt that it is
unnecessary to look at more than one test when evaluating
a particular student.
Other efforts have gone into assessing the useful-
ness of aptitude tests as predictors of college success.
Stanley37 has supported the use Of tests as
predictors of academic success. He favored their use in
college admissions because they lent a degree of Objectivity
to an otherwise more subjective process. As early as 1959,

the American Council on Education reported that "

...tests
have repeatedly proven their usefulness as predictors of
success in the most common academic areas."38

On the other hand, Juola reported to the American
Educational Research Association in 1964 his reservations
concerning the use of ability tests administered at the time
of admission to college. He maintained that such tests lose
validity when used to predict performance in semesters after
the first. (This study accounted for the expected decrease
in predictive validity due to the restriction in range of
grade point averages for students still in school in later
terms.) Juola found also that the students' performance in

college immediately preceeding the semester or term under

study is a better predictor of future college performance

m

20

than is test information. He Opted for research devoted to
more complete analysis of previous academic achievement
measures for use in selection, placement, and guidance.39
Recent concern over test bias against minorities
and disadvantaged students has led to a good deal of
research in that area. Stanley and Porter40 found the SAT
to predict performance of students at black colleges as
well as it predicted for white students at white colleges.
They did not, however, measure its validity for minority

41 found

students at integrated colleges. Thomas and Stanley
tests to be better predictors than grades for students at
black colleges; this was not the case at predominately white
colleges. Cleary42 had the same results as the above studies
in her look at blacks in black colleges and whites in white
colleges. She, too, expressed concern over the lack of data
for minority students at integrated colleges.

Studies done by the American College Testing
Corporation have confirmed these results. In fact, there
is evidence that the ACT may overpredict college grades for
minority or disadvantaged students. It has also been shown
that the ACT battery may underpredict college grades for
women and Older students.43

A limitation of findings like those above is that

they deal exclusively with previous academic accomplishment

and test scores. Lavin has reported on personality and

21

socio-economic measures as they relate to academic perfor—
mance. Students with such personality characteristics as
better study habits, favorable attitude toward school, more
impulse control (and resultant ability to delay gratifica—
‘ tion), more independence, and less anxiety in test taking
situations, tended to perform better in college.44 However,
Lavin went on to state: "Many of the relationships just
described are tenuous at best. . .and. . .cannot be used
confidently for practical purposes, such as for college
admissions."45
Lavin also reported on studies that use a multi-
variate approach to look at the interrelationships among
several variables. These studies were summarized and the
following general statements made: (a) social maturity was
positively related to academic achievement, (b) emotional
stability was characteristic of higher achievers, (c) high
achievement motivation and endurance were positively related
to performance, (d) intellectual curiosity and flexibility
were typical of students who have higher performance levels,
(e) students who were more docile, passive, and orderly
performed better in some social contexts, but (f) indepen-
dence and low need for affiliation characterized high
performers in other contexts. Lavin said that the obvious
h.46

ambiguity of the latter findings deserves further researc

He also went on to say that:

22

(We cannot) be very confident about the
state of knowledge regarding the relationship
between personality characteristics and
academic performance. In most cases these
relationships are weak and, as we have seen,
the findings are Often inconsistent.

Essentially, we think that the literature

represents a somewhat disappointing picture.47

Lavin also summarized the findings concerning the
relationship Of sociological factors and academic perfor—
mance. Socio—economic status was positively related to
performance, except at the higher SES levels where the
relationship is inverse. A Princeton University study48
showed public high school graduates perform better in
college than private school graduates when ability is
controlled. Lavin attributed this, in part, to motivational
factors; high SES students (private high school graduates
in this case) need not worry about the level of performance
in college as long as they graduate; that is, their social
and economic success in life is assured because of family
background and inherited advantages. However, lower SES
students have more motivation to achieve in order to climb
the social or economic ladder. Also, private school
graduates may encounter difficulty with a college environ-
ment that is less structured than that of the private
secondary school.

In comparing males and females, Lavin reported

findings which showed that females performed better

23

academically in college than males.49 This confirmed the
earlier finding of the American College Testing Corporation.

Religious preference and rural-urban background
have also been studied but with inconclusive results. TOO
many intervening variables remain uncontrolled to make the
findings in these areas useful.50

High school size51 may relate to college perfor—
mance, but there has been little study in this area. The
conclusion drawn has been that any observed differences
probably related to the adequacy of the facilities,
teacher salaries, and the like. Further study in this
area has been indicated, since a variety of underlying
variables (economic and social status, the values and
goals of rural youth versus those of urban youth, etc.) are
involved in the consideration of high school size.

With respect to this question of high school size
and its importance in making predictions of college
performance, Creaser52 favored the development of separate
prediction tables for each high school. When the school
is too small to make this practical, we should group schools
of similar size for the purpose Of predicting the college
success Of the students. Hoyt53 found high school rank to
overpredict for students from very small schools, though
he warned against generalizing his findings beyond his own

institution.

24

Just as with personality measures, so far the
findings on sociological determinants of success are nOt
conclusive enough to have a direct impact on the question
of determining access to higher education. As a result,
the develOpment of a useful battery of non-intellective
measures, or factors that could reasonably be Used in making
admission decisions, seems relatively remote. For the
present the primary measures used in admissions Offices
continue to relate directly to previous academic performance
and tested ability.

At the same time, there is a belief that access to
universities can and should be broadened and that the several
criteria in use for determining who has that access need
careful scrutiny. Astin's findings lead him to the conclu-
sion that "segregating students into separate institutions
on the basis of their academic ability may not really benefit
either the bright or the dull students."54 As a proponent
of open admissions, Astin said that the traditional use of
predictive tools (tests and grades) is now inapprOpriate.

He is interested in measuring a student's progress on a
basis other than grades (for example the degree to which a
student improves as a measure of the worth of college to
the individual). Some Of the poorest students, Astin said,
stand to profit the most if given an opportunity for higher

education even though their grades may not prove to be high.

25

(For example he reported that disadvantaged students in
highly selective colleges may not achieve grades as high
as regular students, but the drop—out rates for both groups

55 Huber56 concurred in this. He

are roughly similar.)
thinks we ought to look more at change and growth in an
individual and admit students other than just the academi-
cally elite to our selective admission institutions.

Others have also written about the issue of access
to higher education, and, in particular, to universities.
Most of the recent effort has related to the admission of
minority and disadvantaged students, many of whom are defined
as "marginal" in the sense that they do not meet the tradi-
tional criteria for admission used by the more selective
universities.

Several researchers57 have supported the concept of
opening our more selective institutions to marginal students
on the theory that they can succeed at nearly the same level
as regularly admitted students, particularly when given
supportive services and adequate financial assistance. Such
special assistance, of course, is expensive. Consequently,
the crucial dilemma that relates to this issue of more open
access is that of the most effective use Of our institutions'
limited resources.

Astin58 described the "elitest" view of higher

education as favoring the most investment Of effort and

26

money in those students who have the best chance of academic
success. He depicted the "social welfare" approach as one
which brings up to standard those who have been traditionally
poor achievers in the past; this necessitates an investment
of resources opposite to that in the "elitest" model.
Institutions of higher education are now facing
times of limited resources. The setting of priorities is
therefore mandatory. Whether a moderately selective
institution, which has traditionally invested most of its
resources in only the most able students, can or should
continue to admit students of marginal ability is an open
question. To answer this question of access for marginal
students we must determine the degree of success that such
students achieve; similarly, it is necessary to know which
of these students tend to succeed so that apprOpriate
programmatic decisions can be made in the admission and
supportive services areas. This study is an attempt to help
the institution develop policies that set priorities for the
investment of available resources. It will assist the
institution in determining its prOper role within the
parameters set by Astin's typology of elitest and egalitar—

ian institutions.

C H A P T E R I I I

DESIGN

Population

 

Part of the population under study consists of four
groups of Provisional Admission students admitted to The
University of Texas at Austin for the Spring 1972, Summer
1972, Spring 1973, and Summer 1973 semesters. These were
students who applied for regular admission as first time in
college freshmen and, upon learning of their failure to meet
the minimum admission requirements, requested admission
under the Provisional Admission Program.

The number Of students in each group is as follows:

Spring 1972 . . . . . . . . 28
Summer 1972 . . . . . . . . 204
Spring 1973 . . . . . . . . 45
Summer 1973 . . . . . . . . 208

It should be noted that relatively few freshmen
students enroll for the first time in the Spring semester
because most high school students presently graduate in
June, at the end of the academic year. Since all Texas high

schools are implementing the more flexible academic calendar

of the quarter system, an increasing number Of students will

27

28

be able to graduate from high school early. It is, there—
fore, expected that more students may wish to enter the
University in January in the future. The performance of
students in the Spring semester Provisional program is thus
an important concern.

In addition to the Provisional students studied,
four other random samples are examined. Each sample con-
sists of 100 students admitted for the Spring 1972, Summer
1972, Spring 1973, and Summer 1973 semesters. These
students met all the normal freshman admission requirements
for the University. They were randomly selected for this
study from alphabetical listings of students who entered
for each Of the semesters under study. Each sample was

selected from a pOpulation of several hundred students.

Method Of Analysis

For purposes of this study, the groups are numbered

as follows:

Group 1 . . Provisional Students-Spring 1972
Group 2 . . Regular Students—Spring 1972
Group 3 . . Provisional Students-Summer 1972
Group 4 . . Regular Students—Summer 1972
Group 5 . . Provisional Students—Spring 1973
Group 6 . . Regular Students-Spring 1973
Group 7 . . Provisional Students—Summer 1973
Group 8 . . Regular Students—Summer 1973

The following areas of study are of primary concern.

29

(1) Comparison of the academic performance in
college of Provisional students with that of regularly
admitted students. This is accomplished by three-way
analysis of variance with a 2 by 4 by 2 construction, the
sources of variance being sex, first semester of enrollment,
and mode of admission (i.e. Provisional or regular). The
means for first semester in college grade point average,
total hours earned, total grade points earned, and total
grade point average are compared for groups 1 and 2, groups
3 and 4, groups 5 and 6, groups 7 and 8 respectively. Post-
hoc comparisons using Scheffe's techinque are performed to
gain insight into specific variations shown in the ANOVA.

(2) Comparison of the college academic performance
of students entering the Provisional program in the Spring
semester with that of Summer Provisional students. The
principle criteria are mean first semester in college grade
point average and mean total college grade point average.
Groups 1 and 5 and groups 3 and 7 are combined respectively
for the purpose of the analysis of variance and post-hoc
comparisons. Also, groups 1 and 3 and groups 5 and 7 are
compared individually.

The quality of the high school record as measured
by rank in class and the Scholastic Aptitude Test scores
(Math, Verbal, and Total) are compared for the Spring and

Summer 1972 and Spring and Summer 1973 groups to determine

30

any differences that may exist in the quality of students
entering in the Spring Versus those entering in the Summer.
By using the analysis of variance with the post—hoc compari-
son Scheffe' technique, the effect of such differences on
the college performance is assessed.

(3) Analysis of the performance of Summer 1972
Provisional students with that of Summer 1973 Provisional
students. As discussed, there was considerably more
attention given to academic support services for the Summer
1973 group; any improvement in performance that cannot be
accounted for by examining possible differences in high
school preparation and SAT scores, may be attributable to
on-campus experiences with support services. The principle
criteria used to measure college performance are mean first
semester in college grade point average and mean overall
college grade point average. The analysis of variance with
post-hoc comparisons is used to measure the extent of
differences.

(4) Comparison of the performance of men and women
in the Provisional Admission Program. A t—test of difference
of means is performed to look at mean first semester hours,
grade points, and grade point average, as well as mean total
semester hours, grade points, and grade point average for
men and women in the Provisional Admission Program. Also,

the performance of men and women admitted on a regular

31

basis is compared to that of Provisional men and women
respectively. Male and female regular students are compared
with each other, as are male and female Provisional students.
The t-test with the 2 distribution is used in each case.

(5) Determination Of the most valid predictive
measures to be used for Provisional students. The following
independent variables are examined singly to determine their
correlation with first semester college grade point average
and overall college grade point average:

Academic grade point average in high school
(Provisional Students only)

Overall grade point average in high school
(Provisional Students only)

SAT—Verbal

SAT-Math

SAT-Total

Rank in class

Also, the following combinations of independent
variables are used with the dependent variables of first
semester in college grade point average and overall college
grade point average to determine the most useful multiple
regression equations for purposes of predicting college

performance. In each model, the multiple r is computed.

1. Academic grade point average in high
school, plus SAT-Math and SAT—Verbal

2. Academic grade point average in high
school, plus SAT total

3. Overall grade point average in high
school, plus SAT—Math and SAT—Verbal

32
4. Overall grade point average in high
school, plus SAT—total

5. Rank in class, plus SAT-Math and
SAT-Verbal

6. Rank in class, plus SAT—total

(6) Comparison of the performance of two ethnic
groups, white and Mexican Americans, in the Provisional and
regular student samples for all the semesters under study.
Single tailed t-tests are employed to determine whether any
significant differences exist between whites and Mexican—
Americans within both the regular and Provisional groups.
Also, comparisons of regular and Provisional whites are
made, along with comparisons of regular and Provisional
Mexican—Americans. Mean first semester hours, grade points
and grade point average, and mean total hours, grade points,
and grade point average are the dependent variables.

A study of Black students is not feasible because
very few Blacks were identified in each group, with only 13
in the total Provisional sample.

(7) Examination of the relative persistence in
enrollment of Provisional and regular students. For this
purpose, the proportions of students from each sample that
are still enrolled in the Spring 1974 semester are examined.
A chi-square test is used to test the significance of any

differences shown between regular and Provisional students.

Data

33

The following data is utilized for Provisional

students and the randomly selected regular admission stu— a

dents:

10.
11.

12.

13.

14.

15.

Overall high school grade point average
(Provisional students only)

Academic high school grade point average
(Provisional students only)

SAT—Verbal

SAT—Math

SAT—Total

Rank in high school graduating class
First semester hours earned

First Semester grade points

First semester in college grade point
average

Total hours earned

Total grade points

Total college grade point average for
work completed at The University of
Texas at Austin

Grade point average in college courses
categorized by the following disciplines:
_English, science and math, and social
science

Sex

Ethnic background, categorized by white
and Mexican—American

34
Null Hypotheses

The following null hypothses are presented as a
basis for the study. These are not meant to be exhaustive;
data other than that which is relevant to these hypotheses
is presented, but these areas of examination form the
essence of the most important aspects of the study.

Hol: There is no significant difference between
the first semester in college grade point average of
students admitted under the Provisional Admission Program
and that of students admitted on a regular basis. (H1:
Regular student mean greater than Provisional student mean)

H02: There is no significant difference between
the first semester in college grade point average of
Provisional students offered extensive support services from
that of students not Offered the same services. (H1: Summer
1973 Provisional student mean greater than Summer 1972
Provisional student mean)

H03: There is no significant difference between
the first semester in college grade point average of students
admitted to the Summer Provisional Program and that of
students admitted to the Spring Provisional Program. (H1:
Summer Provisional students' mean greater than Spring
Provisional students' mean)

H04: There is no signficant difference between the

first semester in college grade point average of males in

35

Provisional Admission Program from that 0: females in the
same program. (H1: Male Provisional mean greater than
Female Provisional mean)

H05: There is no significant difference between
the first semester in college grade point average Of white
students in the Provisional Admission Program from that Of
Mexican-American students in the same program. (H1: White
Provisional mean greater than Mexican—American Provisional
mean)

H06: There is no significant difference in the
proportion of students from the Provisional sample enrolled
in Spring 1974 semester from the proportion Of regular
students enrolled in the same semester. (H1: Proportion of
regular students enrolled greater than proportion of

Provisional students enrolled)

C Ii A.I? T I3 R I \7

RESULTS

DEMOGRAPHIC DATA FOR SAMPLES

 

In comparing the entire group Of Provisional students
with that of regularly admitted students, Table 1 shows

similar percentages of men and women.

TABLE l.--Composition of regular and Provisional samples by

 

 

 

 

SEX
Spring Summer Spring Summer
1972 1972 1973 1973 Total
2.1 :11 Hi 21 A 3;

35.1212;

Men 58 58 4O 40 48 48 45 45 191 47.8

Women £42 32.69. 2.2.2.2. 22a 29.9.2212.

Total 100 100 100 100 400
Provisional

Men 11 39 88 43 23 51 101 49 223 45.9

Women Jlél 1_1__6_fl .2312 E151 2221.;

Total 28 204 45 208 485

 

36

37

These percentages contrast somewhat to the makeup
of the undergraduate student population for the entire
University, where approximately 60% Of the students are
men.59

Spring 1972 is the only semester in which there is
a marked difference between the regular and Provisional
groups using sex as a criteria.

Table 2 reveals that the ethnic composition of
the regular and Provisional groups is also roughly similar.
Using ethnic identity as a criteria, both groups reflect

the ethnic composition of the University as a whole.60

TABLE 2.--Composition of regular and Provisional samples
by ethnic background

 

 

 

Spring Summer Spring Summer
1972 1972 1973 1973 Total
2. E. 2. E. 2. E. 2. E. 2. .1
Regular

White 89 89 88 88 92 92 96 96 365 91.3
Black 1 1 3 3 o - 'I 1 5 1.3
Mexican-
American 9 9 5 5 8 8 2 2 24 6.0
Other 1 l __4_ 4 O - l l 6 1.4

 

Total 100 100 100 100 400

 

38

TABLE 2.--Continued

 

 

 

 

Spring Summer Spring Summer
1972 1972 1973 1973 Total

Provisional

White 27 96 179 88 39 87 179 86 424 87.4

Black 0 - 6 3 l 2 6 6 13 2.7

Mexican-

American 1 4 l9 9 5 11 23 ll 48 9.9

Other _0_; _9_; _9_:_ _9_; __9.__'_

Total 28 204 45 208 485

 

Due to the small number of Black students in either
sample (5 regular and 13 Provisional Black students),
drawing conclusions from their performance would be extremely
tenuous. Therefore, the only ethnic groups studied are white

and Mexican-American.

MEAN HIGH SCHOOL AND COLLEGE PERFORMANCE

 

Table 3 is used to demonstrate the respective mean
high school, test, and college performance of all regular
and Provisional students in the samples. The fact that
the number of regular students varies somewhat for the items

in question is due to a lack of available data in certain

39

instances. That is, some of the 400 students in the total
random sample have incomplete records. This is not seen to
affect the outcome since the percentage of such students is
small, and the occurrence of missing data is randomly
distributed across the sample.

The measure of high school rank used in this table
and throughout the study is a percentile ranking, arrived
at by dividing the students' rank in high school class by
the number of students in the class. In cases where the
school only categorizes students as to their quartile
rank (first, second, third, or fourth), the midpoint for

percentile ranking is used according to the following:

First quarter 12.5 percentile

Second quarter 37.5 percentile

62.5 percentile

Third quarter

87.5 percentile

Fourth quarter

TABLE 3.-—Mean scores on high school, test and college
performance for regular and Provisional students

 

 

 

 

REGULAR PROVI S IONAL
Standard Standard
n Mean Deviation n Mean Deviation
High School Rank 363 73.7 19.1 485 43.7 20.9
SAT—Verbal 380 513 92.9 485 385 62.3
SAT-Math 380 530 90.3 485 409 68.0

SAT-Total 380 1043 159.9 485 793 105.3

 

40

TABLE 3.--Continued

 

 

 

 

REGULAR PROVIS IONAL
Standard Standard
n Mean Deviation n Mean Deviation
First Semester
Hours Completed 400 9.9 3.9 485 11.3 2.5
First Semester
Grade Points 400 25.8 13.9 485 22.7 9.0
First Semester
Grade Point
Average 400 2.52 .97 485 1.94 .71
Total Semester
Hours Completed 400 35.9 21.7 485 30.2 16.5
Total Grade
Points 400 96.0 71.9 485 62.5 42.6
Total Grade
Point Average 400 2.48 .85 485 1.92 .68

 

There is slightly less variance in the SAT-Verbal,
SAT-Math, and SAT-Total scores for Provisional students when
compared with that of regular students. This is also true for
the several measures of college performance used. The variance
of the high school rank for Provisional students is very close
to the variance of high school rank for regular students.

The fact that Provisional students by definition are

those applicants who do not meet the normal admissior1

41

requirements, probably explains some of the restriction in
range of their SAT scores. Since all Provisional student
SAT scores fall below those required for entrance as

regular students, they would tend to be distributed slightly
more narrowly than scores for regular students.

For the measures displayed in Tables 4a-4f, it is
apparent that scores for both regular and Provisional
students approximate a normal distribution. For each
variable displayed, the frequency distributions show
Provisional students to lie at the lower end of the scale,
with regular students' scores centered at the upper range.

Comparison of respective percentile rankings reveals
substantial differences between regular and Provisional
students. For example, approximately 92% of the regular
students graduated in the upper half of the high school
class. Only 32% of the Provisional students did so. Using
the same guide, 73% of the regular students scored over
450 on the SAT-Verbal while only 17% of the Provisional
students did so. Eighty-three percent of regular students
made better than 450 on the SAT-Math, but only 30% of the
Provisional students were able to do that well.

These types of differences are also reflected in
the college performance measured in terms of first semester
and total grade point averages. Seventeen percent of the

regular students made less than a "C" average (2.0 on a 4.0

42

scale with A=4.0, B=3.0, C=2.0, etc.) during the first
semester. Thirty-six percent of the Provisional students
earned less than a "C" average. Twenty-four percent of
the Provisional students made a 2.5 grade point average
or better during the first semester, but fully 69% of the
regular students did so.

The distributions for total grade point average
reveal similar achievement. Twenty-two percent of the
regular students maintained less than a 2.0 grade point
average, but 48% of the Provisional students achieved
below a "C" average on all coursework taken. Twenty-nine
percent of the regular students were able to make at least
a "B" average overall, but only 4% of the Provisional
students did so.

In addition to comparing the overall performance
of regular and Provisional students, it is helpful to
analyze the mean scores on the above measures by controlling
for sex and ethnic background. Table 5 presents the data
relating to the performance of regular and Provisional
men and women.

Females in each sample performed somewhat better
than males on the measures of college performance. Females
had lower SAT-Math, SAT-Verbal and SAT-Total scores, but

their rank in the high school graduating class was higher

43
than that for males. This also is true within both the

regular and Provisional groups.

TABLE 4a.--Frequency distribution of high school rank for
regular and Provisional students

 

 

REGULAR PROVISIONAL

High School Fre- Per- Per- Fre- Per- Per-
Rank quency centage centile quency centage centile
0-9 1 - l 18 4 2
10-19 4 1 1 42 9 8
20-29 8 2 2 62 13 19
30-39 9 2 5 102 21 36
40-49 12 3 8 102 21 57
50-59 39 ll 15 45 9 72
60-69 66 18 29 43 9 81
70-79 42 12 44 37 8 89
80-89 120 33 66 31 6 96

90-99 62 17 91 3 - 99

 

44

TABLE 4b.--Frequency distribution of SAT-Verbal for regular
and Provisional students

 

 

REGULAR PROVISIONAL

 

SAT-Verbal Fre- Per- Per— Fre- Per- Per
quency centage centile quency centage centile

 

200-249 1 - 1 1 - 1
250-299 1 - l 26 5 3
300-349 8 2 2 107 22 17
350-399 28 7 6 152 31 43
400-449 52 14 17 117 24 71
450-499 80 21 34 58 12 89
500-549 74 19 54 18 4 97
550-599 63 17 72 6 1 99
600-649 41 11 86 - — -
650-699 20 5 94 - - -
700-749 10 3 98 - — -

750-799 2 - 99 - - -

 

45

TABLE 4c.--Frequency distribution of SAT-Math for regular
and Provisional students

 

 

REGULAR PROVISIONAL

 

SAT-Math Fre- Per- Per- Fre- Per- Per-
quency centage centile quency centage centile

 

200-249 - - - - - -
250-299 1 - l 12 2 1
300-349 3 1 83 17 11
350-399 15 4 3 118 24 32
400-449 48 13 11 125 26 57
450-499 75 20 27 94 19 79
500-549 78 21 48 39 8 93
550-599 67 18 67 13 3 98
600-649 54 14 83 l - 99
650-699 22 6 93 - — -
700-749 15 4 97 — — -

750-799 2 - 99 - - -

 

46

TABLE 4d.--Frequency distribution of SAT-Total for regular
and Provisional students

 

 

REGULAR PROVISIONAL

 

SAT-Total Fre- Per- Per- Fre- Per- Per-
quency centage centile quency centage centile

 

550-599 - - - 8 l 1
600-649 1 - l 29 6 5
650-699 3 - l 48 10 13
700-749 1 - l 75 15 25
750-799 8 2 2 111 23 44
800-849 29 8 7 52 ll 61
850-899 33 9 15 67 14 74
900-949 41 ll 25 52 11 86
950-999 41 11 36 43 9 96
1000-1049 44 12 47 - — _
1050-1099 44 12 59 - - -
1100-1149 27 7 68 - — _
1150-1199 37 10 76 - - _
1200-1249 30 8 85 — — -
1250-1299 20 5 92 - - _
1300-1349 11 3 96 - — _
1350-1399 5 l 98 — _ _
1400-1449 3 - 99 - — -
1450-1499 2 - 99 _ - -

1500-1549 1 - 99 - - -

 

47

TABLE 4e.--Frequency distribution of first semester grade
point average for regular and Provisional students

 

 

 

 

REGULAR PROVIS IONAL
First Semester
Grade Point Fre— Per- Per- Fre- Per- Per-
Average quency centage centile quency centage centile
0-.24 l7 4 2 l6 3 2
.25-.49 2 - 5 4 - 4
.50-.74 8 2 6 9 2 5
.75-.99 3 - 7 12 2 7
1.00-1.24 9 2 9 29 6 ll
1.25-1.49 10 3 ll 18 4 l6
1.50-1.74 15 4 14 44 9 23
1.75-1.99 7 2 17 49 ' 10 32
2.00-2.24 55 14 25 101 21 48
2.25-2.49 32 8 36 8O 16 66
2.50-2.74 51 13 47 63 13 81
2.75-2.99 29 7 57 3O 6 91
3.00-3.24 67 17 69 19 4 96
3.25-3.49 20 5 8O 5 1 98
3.50-3.74 25 6 86 5 l 99
3.75-3.99 14 4 91 - - 99

4.00 29 7 96 1 - 99

 

48

TABLE 4f.--Frequency distribution of total grade point
average for regular and Provisional students

 

 

REGULAR PROVISIONAL

 

Total Grade Fre- Per- Per- Fre- Per- Per-
Point Average quency centage centile quency centage centile

 

0-.24 10 3 1 15 3 2
.25-.49 3 - 3 3 - 3
.50-.74 5 1 4 7 l 4
.75-.99 8 2 6 l4 3 7

1.00-1.24 6 2 7 33 7 ll
1.25-1.49 13 3 10 27 6 18
1.50-1.74 15 4 13 62 13 27
1.75-1.99 27 7 19 71 15 ' 41
2.00-2.24 56 12 28 89 18 57
2.25-2.49 46 ll 39 73 15 74
2.50-2.74 52 13 51 49 10 86
2.75-2.99 54 14 65 21 4 94
3.00-3.24 42 ll 77 15 3 97
3.25-3.49 32 8 86 5 l 99
3.40-3.74 21 5 93 - - 99
3.75-3.99 10 3 97 l - 99

4.00 7 2 99 - - 99

 

49

The magnitude of the differences shown between
males and females in both samples is generally the same.
For example, female regulars had a mean SAT-Verbal that
was eighteen points below that for male regulars. Female
Provisionals scored four points lower than male Provisional
students. In each case, the SAT-Math scores showed more
marked differences. Similarly, female regulars ranked
about seven percentage points higher than male regulars
in the high school class. Female Provisionals ranked
about nine percentage points higher than male Provisionals.
A similar pattern is demonstrated in the respective first
semester and total grade point averages. Females performed
slightly better than males on both measures for the regular
and Provisional samples. 2

Table 6 presents data relative to the performance
of white and Mexican-American regular and Provisional
students. Mexican-American Provisionals had a considerably
higher rank in the graduating class from high school than
their white counterparts; however, the SAT scores were
substantially below those for white Provisional students.
The same pattern exists for the regular student sample,

but the differences are much less distinct.

50

 

om.a mom sm.a mmm mm.m mow mv.m Had mmmum>¢ ucaom
mango Hmuoe

m.mO NON O.mm MNN v.vm mON m.hm HmH mHCNom OOMHU HmuOB

m.am NON m.mN MNN H.mm mON 0.0m Hma Omumamfioo muaom
umummemm Hmuoa

Om.H NON mO.H MNN va.N mON OHO.N Had wmmuw>¢ unflom
mpmuo Hmummfimm umuwm

H.mN NON N.NN MNN O.¢N mON N.hN Hma museum mpmuw
Hmummamm umuflm

m.HH mom m.HH mmm v.m mom v.oa Had emumamaoo mason
umummfiom umuflm

 

 

 

mhh NON «HO MNN OHOH NON HOOH Oha HmuOBlamm
mmm NON ONV MNN Oom NON mmm mud £DMZIB¢O
mmm NON hwm MNN vOm NON NNm OPH HMQH0>IB¢m
b.5v NON mm MNN m.Oh VOA H.Oh OOH xcmm Hoonom swam
N G N C N C um G
mquzmm mquz WMdemm mmdmz
A<ZOHmH>Omm quowmm

 

 

xmm he musmospm Hmsoflmw>oum can “masmmu
H0O cosmEHOMHmm mmmaaoo can .ummu .Hoonom now: so mmuoom :mmZI|.m mqmda

51

 

 

Nm.a mv mm.H va I I I I ucmEmOMHm Ooocm>nm

usozuflz mmmnm>m

ucwom wcmuu Hmuoa

mN.N wv mm.H vmv mO.N mm nv.N HOm momnm>< ucflom

mpmuu Hmuoe

O.mm mv m.mm qu h.mNH mN m.Nm HOm mucaom mpmuu Hmuos

N.mm mv H.mN vmv O.mv MN m.vm HOm Omumameou musom

HmummEmm Hmuoa

mh.a mv mm.a vmv mv.N NN Hm.N OOM mmmnm>¢ unwom

momma HmummEmm umufim

m.mH mv mm va N.om NN m.mN OOm mucflom mOmHO

Hmumwamm umuflm

H.HH mv m.HH vmv 0.0H NN m.m OOm Owumamsou mHsom

nmummewm umHHm

New mv Now qu NNQH NN mvoa hem HmuoaIemm

mmm we Nae va mam NN Hmm mam suszeam

mmm mv 0mm va mom NN mam hem Hmnnw>Ismm

m.>m mv b.Hv vmv m.Ob ma m.mh mmm xcmm Hooaom nmwm
x a x c x c x :

zmuHmmzm mBHmz zmonmzd mBHmz
IZdUmeS IZdonmz
AdonmH>omm maqoomm

 

 

Ocsoumxomn cannum ma mucmpsum Hmcowmw>oum can unasmwu
How mosmEHOOHmm mmmaaoo Ocm .ummu .Hoonom now: so mmuoom cmeII.O mqmfla

52

A detailed analysis of the apparent differences
between the college performance of white and Mexican-
American students in both the regular and Provisional
samples is presented below. However, a brief explanation
of some findings is appropriate here.

The University has a policy of granting credit
and grades for successful completion of certain advanced
standing examinations; inclusion of such grades in the
computation of the total grade point average for students
tends to inflate this measure. This is because students
generally choose to accept the grade only if it is an
"A" or "B". Consequently, the grades in work completed
only on campus tend to be somewhat lower than grades
earned through advanced standing examinations.

This phenomenon seems to affect the relative overall
grade point averages of regular white and Mexican-American
students to somewhat different degrees. This is apparent
because the difference noted in first semester grade point
average (2.49 for Mexican-Americans and 2.51 for whites)
is less than the difference in the respective total grade
point average (2.68 versus 2.47). The advanced standing
examination grades are computed only in the total grade
point average, and the effect is that they improve Mexican-
American student grades more than they improve those of

white students in the regular student sample.

53

For Provisional students there are similar but
much more pronounced differences. While white students
did better than Mexican-Americans during the first semester,
the opposite is true when total grade point average,
including advanced standing grades, is used to measure
overall success. When the advanced standing grades are
not included in the computation of total grade average,
the white students again show superior performance. In-
spection of individual student records reveals that
Mexican-American students were very often able to do quite
well on the Spanish placement examination. This affects
the total grade point average significantly, giving an
inflated picture of the success of the Mexican-American
in courses at the University. Conversely, white Pro-
visional students were able to obtain advanced standing
grades in only a few courses and thus the total grade
point average was virtually unaffected.

The more consistent manner in which to judge over-
all performance is therefore to delete advanced standing
grades from the consideration. This is particularly true
in the case of comparing the relative performance of
Mexican-American and white Provisional students since
the total grade point average of the former is so markedly
affected by the Spanish advanced standing examination

results.

54
TESTS OF SIGNIFICANCE OF DIFFERENCE

Having determinedtflmanature of some of the dif-
ferences in performance of the regular and Provisional
samples, using sex and ethnic background as controlling
factors, the next step is to test for the statistical
significance of any observed differences. The standard
t-test for large and small samples is used to determine
the significance of differences between various sample
means. Chao61 states that, for large samples (n greater
than 30) with known variances, it is permissable to "apply
the normal distribution for inferences about the difference
between two means regardless of the nature of the population
distribution." Consequently, except for the analyses
involving Mexican—American regular students (n equals 22),
it is permissable to test for differences using the 2
distribution in the standard normal table. A one-tailed
test is chosen and the critical value of the test statistic
is thus 1 1.645 for a = .05, depending on the hypotheses
tested in each comparison for the variables under study.

In the case of small samples, Chao62

suggests use
of the t-distribution since sample variance can be quite
different from population variance. It is assumed that

the two samples are independent. Table 10 below depicts

55

the data for comparison of the Mexican-American samples
where the small sample theory is used.

Table 7 presents the relevant data for the regular
and Provisional students taken as a whole. Inspection
of the 2 values reveals highly significant differences
between the two samples. For each measure examined, with
the exception of first semester hours completed, regular
students performed dramatically better than Provisional
students. In each instance, the differences shown would
be significant at the a = .01 level of significance (critical
value = 2.33). (The finding that Provisional students
completed more hours the first semester than did regular
students reflects the fact that Provisional students are
required to take twelve semester hours the first semester,
whereas some regular students elect to take fewer than that
number, particularly during the summer session. Therefore,
this variable proves to be an unreliable indication of
the relative success of regular and Provisional students.)

In reference to the first null hypothesis presented
above, we can very confidently reject the suggestion that
there is no significant difference between the first
semester grade point average of regular and Provisional
students. The difference shown of .58 is significant even

at the .01 level.

56

These contrasts between regular and Provisional
students are also evident when the data in Tables 8 and
9 is examined. Male regular students clearly out-performed
male Provisionals;each of the differences would be sig-
nificant if alpha were set equal to .01. The only measure
(except first semester hours completed) on which female
regulars were not clearly superior to female Provisionals
is total grade points. On all other criteria, the dif-
ferences shown are significant at the a = .05 level.

The size of the differences between male regular
and Provisional students is the same general magnitude
as the differences for females. Correspondingly, these
approximate the overall differences between all regular
and Provisional students.

Table 10 shows that Mexican-American students in
the regular sample performed significantly better on the
SAT-Verbal, SAT-Math, and SAT-Total than did their Pro-
visional sample counterparts. The differences shown are
significant at the d = .05 level. However, when high
school rank and college performance for each group is
compared, there is less of a disparity between regular and
Provisional Mexican-Americans. In fact, in each of the
measures--high school rank and first semester and total
hours completed, grade points, and grade point average--

the Provisional Mexican-Americans show no statistically

57

 

mm» mo.oa mm. as. mm.a mmv ms. mv.m oov womum>¢ ucaom
wOmuw HmOOB

mm» HN.O m.mm m.OHOH m.NO mmv O.Hham 0.0m oov mucflom

mpmuw Hmuoa
mm» Nm.v h.m O.m5N N.Om mmv O.Hhv m.mm OOV OOHOHQEOU mhsom

kumosmm Hmuoe
mm» 50.0 mm. Om. vm.H mmv vm. Nm.N OOQ mmmum>£

unwom OOMHO
uwummfimm umufim

 

 

 

 

mm» mO.m H.m 0.00 h.NN mmv O.NmH O.mN OOO mucflom OOMHD

umuwwewm umuwm
mm» OH.OI v.HI mN.O m.HH mmv v.mH m.m . OOv OODOHQEOU muflom

“mummsmm umuflm
mm» h.HN 0.0m m.mmv h.mv mmv O.mOm h.Mh MOm xcmm Hoonom flmflm
mow mm.ON OmN O.thHH mom mmv N.HhmmN mvoa 0mm HMHOBIBmm
mm» mh.HN HNH m.mHOv mow mmq H.vaO 0mm 0mm SDMZIadm
mow H.MN ONH 0.0mOm mmm mwv h.ONOO mam OOm HMQH0>I84O
unmofl N NxIHx mm Nx N: an fix a: camoufluo

Imacmam .I .I m .I
qaonmH>omm maqoomm
va.H H n msam> HMUHDHHO “mo. n a ummu cmawmuImno

.GOHusQflHumHO N mcwms mucmOsum Hmsowmfi>oum cam amasmwu mo QOmflHmmEOUII.h mqm<a

58

 

mm» h0.0 mm. mHNm. hO.H MNN mm. N¢.N HmH wmmnm>d unwom
momnw Hmuoa

mm» HN.O m.mm N.vmha O.mm MNN m.HNmm m.hm HmH muCfiom
mpmuu HMDOB
mm» Hm.m m.h m.mmN m.ON MNN h.OOm 0.0m HOH Omumamfioo muoom

umummfimm Hmuoe

mmw O0.0 NO. Novm. OO.H MNN OO.H mam.N HmH wmmum>¢
unflom mcmuu
HmummEmm umuflm

 

 

 

 

 

mm? 00.? O.m m.mm N.NN MNN 0.5HN N.hN HmH mfiflflom OUMHO

nwummfimm umuflm
WON wN.NI om.l N.h N.HH MNN m.hH V.OH Hod UquHQEOU MHDOE

HmpmmEmm umuflm
WOW w.vH H.Hm m.mmm O.mm MNN 0.0%? H.Oh OOH Mcmm HOOQUW Smflm
mm? mv.mH NON ®.vmmHH va MNN N.m©mmm HmOH th HOHOBIBdm
mm? mh.mH OMH m.mva va MNN m.mmmm mmm th Sumzlfidm
mmw O¢.hH mma m.mvmm hmm MNN O.vomh NNm whH HMQHO>IB¢m
ucmofl n. NxIHx” Nm Nx N: mm ax Ha maumufluu

Imacmam I I m I I
AdZOHmH>Omm MQQZ M¢QDUHM fiddz
meO.H H u wsam> Hmowuwuo .mo. n a “pump OmawmuImso .coausnfluumflc

N mcams mucwcsum ascofimfi>oum mama paw amasmmu wane mo nomwummfiooII.m mqmda

 

59

 

mm» vq.m om. ov. om.a mom om. mm.m mom mmmum>a acaoa

momuo Hmuoa

mm» mm.m m.mm m.mHmH m.mo mom m.mamw 4.4m mom muaaom

mamuw Hmuoe

mm» om.m m.m o.mmm m.am mom m.mmm H.mm mom amumamaoo mason

kummEmm ANDOB

mm» om.s am. as. mm.a mom Hm. amm.m mom mmmum>¢

ucflom mOmuo

Hmummamm umuflm

oz mv.H m.H o.mn H.mm mom o.moa 6.4m mom mucflom mango

Hmumwfiwm umufim

mm» 54.6- m.HI v.m m.HH mom «.ma v.m mom cmumadsoo muaom

HmpmmEmm umuwm

was mm.oH m.mm m.saa p.54 mom H.msm m.oa «ma xcmm Hoonom roam

mm» mo.mH mmm 0.5mom map New m.masvm oaoa mom Hmpoeueam

mm» mm.oH mam H.mnwm mam mom m.samo mom mom nuszeam

mm» He.ma HNH o.moav mmm mom m.mmmm «om mom Hmnum>ueam

unmow N NxIH mm NC Hm fix a: mflnmuwuu
Imacoam I. I. m .I

 

 

AdZOHmH>Omm m4¢zmm Mddbomm ma<zmm

 

 

mam.a H u msam> Hmoauauu .mo. u a tummy omaamuumao .aoausnauumao
N mewms mucmwsum Hmsoflmfi>onm mHmEmw can HmHsmmH mHmEmm mo :omflHmQEOUII.m mqmfia

 

60

 

oz mo NoH. «.4 «.mmv ooooameoo musom

Houmwfimm Hmuoa

oz mo saw. ms. mo. oomuo><

Ocaom momHO

umummfimm umuflm

oz mo mom. e.oH o.os muoaoo moons

H0#mm5@m #mHHh

oz mo omH.Iom.I No.4 ooooamsoo musom

HGUMOEOW #mHHrm

oz mo Nmm. o.ma m.nmm o.sm we 6.65m m.oa ma scam Hoozom roam

mo» mo oNN.N omm m.ommn was we H.mssmm mmoa mm Hmuoeueam

mo» mo ms.a oma n.6mmm mom mo H.momm mam mm zuszomm

66% mo Ho.m oma o.mmmm mmm me m.ovaH mom mm Honuo>ueam

pcmofl mo u NxIHx mm Nx N: mm fix an camoufluo
.333 I I I I

 

..HdZOH mH>Omm maqowmm
ZdUHmMZmIZQUHxMZ ZdUHMMZdIEUHxMZ

 

 

 

 

umou OmawmuImao .cowusnfluumwo u mcwms mucocsum HMGOHmfl>OHm
:mowuwfidIcmoflxmz cam Awasmmu cmomeEmIcmowxmz mo QOmemmEOUII.oa mamme

61

zao>aoooomou moo.a .moo.a .oso.fi on Hoooo ma mo .mo .mo u so now oofio> Hmofiuauo
Aummu Omawmqusov mo. u o

N I Na + H: u Eocomum mo mmmummo

 

oz mO Nom. mv. mmm. mN.N Ow vmm. OO.N MN mmmnm>m ucflom
mOmnu Hopoa

02 0O Omv. H.mm O.mvmm O.mm Ov 0.0mhb h.ONH MN mucwom
mpmum Hmuoa

 

 

 

 

pcmow mo u NXIHx Nm Nx N: an fix a: mwuouwnu
Ioaomam .I .I m .I m .I
AdZOHmH>Omm mfidDUmm
ZflUHMMZflIZdUHXMZ ZfiUHMMEfiIZflUHXNZ

 

 

UmscflusooII.oa mqmds

62

significant difference in achievement from regular Mexican-
Americans. However, there are some moderate differences
shown; of particular note are the differences in first
semester and total grade point average. The lack of
significant differences in these cases is partly due to

the necessity of utilizing the more conservative small
sample theory in approaching the analysis. The very

real differences shown in Table 10 should not be overlooked
in the overall study of Provisional admission students.

The relative performance of males and females in
the Provisional samples indicates that the null hypothesis
H05:"There is no significant difference between the first
semester in college grade point average of men_in the
Provisional Admission Program from that of women in the
same program" is not disproven. Table 11 demonstrates
the values of z for the differences between men and women
in the Provisional sample. Men in the sample presented
superior scores on the SAT-Math and SAT-Total measures,
each difference being significant at the a = .05 level.

The SAT-Verbal scores were very close, and no statistically
significant difference between men and women is shown.

On each of the measures of high school performance,
women proved to be superior to men in the Provisional sample.

The criteria of rank in class and academic and overall grade

63

point average indicate that females presented better high
school grades than did males. These differences are
significant at the a = .05 level.

When college grades are considered, females per-
formed consistently better than males, but only the
differences in total grade points and grade point average
in English are significant at a = .05 level. With the
critical value of the test statistic equal to 1.645, and
z = 1.63 for the difference between male and female total
hours completed, this criteria is very close to showing a
significant difference. SimilarLy,if the value of
a = .1 is used, there appears a significant difference
between male and female first semester and total grade
point average. (For a = .l in a single-tailed test, the
critical value is 1.282 for large samples.)

Table lla shows the results of t-tests for male
and female Provisional students when grade point averages
in the three curricular groupings are compared. In each
case, the 2 value is clearly above the critical value for
significance at the a = .05 level; therefore, both male
and female Provisional students performed significantly
better in English courses than they did in the other

academic areas under consideration.

64

 

 

OZ MH.HI m.l O.mb H.MN NON m.mm N.NN MNN mUCHOQ OOOHU

Hmumoamm umHHm

OZ HO.I VH.I V.m m.HH NON N.h N.HH MNN mmpOHQEOU mHDOZ

kummawm umhfim

mmfi mm.VI OH.I MOmH. Oh.N NON mmmH. Nm.N MNN OOMHO>¢ #CHOQ

opmuu Hamnm>o

Hooaom :64:

www mH.mI NN.I OGHN. mv.N NON HONN. MN.N MNN OOOHO>< #CHOA

mOmuO anmomom

Hoozom seam

mm» OO.VI h.ml m.hv¢ F.5v NON m.mmm 0.0m MNN xcmm HOOQUW Qmflm

OOV OH.V mm O.hmOm mhh NON O.meHH VHm MNN HMflOBIBdm

mmw mh.m mm H.mhmm mmm NON m.mva va MNN Suwzlfidm

OZ HF. v O.mOHv mmm NON m.mvmm hmm MNN HOQHO>IB¢O

unmofl N NxIHx Nm Nx Nc mm fix an maumufinu
Imacmam I I m I I

 

quzOHmH>Omm m4<2mm

 

mAdZOHmH>Omm mgfiz

 

 

msam> HMOHDNHU

.3. u

364 H u

mummy OOHflmuImso .GONUDAHHuwfiO

N unflms musoosum HMQOflma>OHm mamawm Ono mama mo somHHmmEOOII.HH mqmée

65

 

 

02 AN. No.I mmom. vh.a HON HOOm. Nh.H NOH Jocoflom Hafioom

IImmmum>¢ ucflom momuw

oz om.HI NH.I moms. Oh.a mmN NNmn. NO.H NHN sum: was mocmflom

IImmmno>< unfiom mOmuO

mm» mo.NI OH.I vONO. hH.N omN Ova. HO.N MNN gmflamcm

IImmmum>4 unflom oomuw

oz Hm.HI mo.I ow. Om.a NON mHNm. 50.H MNN mmmno>m ucflom

momuw Hmuoe

mm» on.HI m.OI m.mHOH m.mO NON N.vwsa o.mm mNN macaom

mOmuO Hmuoe

oz mO.HI v.NI O.mON m.am NON m.mmN m.mN MNN Omuoamfioo musom

kumoEwm Hmuoa

oz vm.HI mo.I Ov. Om.a NON Novm. mO.H mNN mmmuo>¢

unflom momma

kummEmm umnflm

unmow N NxIHx Nx N: Mm fix a: maumufluu
.383 I I I I

 

mAdZOHmH>Omm quzmm

 

mA¢ZOHmH>Omm HANS

 

 

OmscwucooII.HH mammfi

66

TABLE llarmwrtests for grade point averages in English,
Science/Math and Social Science.

 

 

 

 

 

 

 

GPA—English GPA-Science/Math Difference .E
Male 2.01 1.62 .39 4.73
Female 2.17 1.76 .41 5.56
GPA-English GPA—Social Science Difference ‘5
Male 2.01 1.72 .29 3.23
Female 2.17 1.74 .43 5.33

 

Means, variances, and respective test statistics
are given for male and female regular students in Table 12.
Male regulars scored significantly better on each segment
of the SAT, but women presented a higher rank in the
graduating class (academic and overall high school grade
point averages were not available for regular students.)
Males and females had very similar academic records in
college, particularly when performance is expressed in
terms of first semester and total grade point average.
Women completed slightly fewer semester hours overall,
but maintained a slightly higher total grade point average.
These differences do not approach significance.

The pattern of differences for regular males and

females can be compared to the pattern of such differences

67

for Provisional students. Males in both samples tended

to present higher entrance test scores, while females
presented superior high school records as measured by rank
in class. Both groups demonstrate similar performances

by males and females once the students enrolled in the
University. Only the total grade points for Provisional
students and the first semester hours and grade points for
regular students show significant differences between
males and females. The more meaningful measures of first
semester and total grade point average do not show sig—
nigicant differences in either the regular or Provisional
samples for males and females.

It is also useful to examine the significance of
the differences between the performance of white and
Mexican-American students within the Provisional sample.

Table 13 shows the results of the t-test when
applied to Provisional white and Mexican-American students.
It is evident that white Provisionals achieved significantly
higher scores on the SAT Verbal, Math and Total measures.
Conversely, Mexican-Americans presented significantly
better high school records when judged in terms of rank in
class, and academic and total grade point averages.

Using first semester in college grade point average
as a criteria, white Provisionals performed at a higher

level than Mexican-American Provisionals. The difference

68

is significant at the a = .05 level. This difference
is also reflected in the first semester hours completed
and grade points earned.

Mexican-American students earned higher over-
all grade point averages than did white students in the
Provisional sample. As suggested above, however, this
measure is clouded by the presence of grades earned through
advanced standing examinations. When this confounding
variable is taken out, the grades of white students are
not significantly better than those of Mexican-American
students.

White Provisionals made significantly higher
grade point averages than Mexican-Americans in both
English and science/math related courses. These dif-
ferences are significant at the a = .05 level. There is
no significant difference shown when grade point average
in social science courses is examined and compared for

each group.

69

 

 

 

 

 

 

OZ Nb. h.H O.NOm H.mm OON h.OOm m.Om HOH OOUOHQEOU mhsom
Hmummsmm Hmuos
OZ OH.I HO.I HO. Nm.N OON OO.H Hm.N HOH OOMHO>¢
DGHom mpmnw
“mummfiom umHHm
mww OO.H O.N 0.00H O.vN OON O.NHN N.nN HOH “CHOA OOMHO
kummfiwm uwHHm
mmw mm.N O.H O.NH v.O OON m.hH v.OH HOH OOOOHQEOU mhnom
umpmmfiom umuHm
mm» Om.MI O.OI H.NhN 0.0h VOH 0.0v¢ H.Oh OOH xcmm HOOEOO SOHm
mmw mv.v Hh m.MOBVN OHOH NON N.OOOMN HOOH OOH HmpOBIBdm
mm» Hm.m Nm O.NHOO OOm NON m.MMNm mmm th QumEIBdm
mmw OO.H OH N.OMNO vOm NON O.¢OOB NNm OOH HMQHO>IB¢O
HCNOH N NxIHx mm mm N: mm Hx H: NHkuHHU
4.83m I I I
mfibwmm magnum ”HEDGE mg
364 H u
09Hm> HmowuHHo .mo. u a “pump OmHHmuImco .GOHOSQHHumHO

N OCHms mucmcsum HMHDOOH meEmm Osm HMHsOmH mama mo somHHmmEooII.NH mqm<a

70

 

 

oz 5H.HI oH.I oo. mm.m mom mm. Nv.m Had woouo>< oofloa

mOmHO HMDOB

oz he. v.m m.mama 6.46 mom m.ammm m.eo Hod moaaoo

momuo Hmuoe

unmoH N NxIHx mm Nx Na mm Hx H: .MHHmuHHO
423m I I I I

 

mSDUmm magma“

 

gDOmE mg

 

 

OmscHuGOUII.NH mqmfla

 

71

mmm OO.H N.N 0.0h 0.0H Ow 0.0h O.NN va UGHOA OOMHU
kumosmm umHHm

oz me. m. «6.4 H.HH ma Hm.o m.HH ems oouoamsoo muses
Hmumofimm umHHm

mo» mo.NI mH.I oe. ms.~ ma OH. oo.~ vmv omouo>¢ ucfiom
oomuo Hmuoe

mo» om.~I HN.I mm. vm.~ ma mm. mm.m one omouo>< uoaoo
oOmHO owEoOmod
Hoonom roam

 

WOW h.vl N.NHI m.hmm O.NO mv O.HOV h.Hv VNV xcmm HOOQOO gmflm
mmm mv.v 0O m.ommw th mv m.mHMOH Now VNV HmuOBIBQm
mmw hm.N MN B.Ommm Omm mv m.thv NHV VNV SHMZIfiflm
mm» mm.v hm V.N®MN mmm O? V.Hmmm omm va HOQHO>IB¢W
DCMOH N NxIHx mm Nx H: mm Hx H: MHHmuHHU
Imacmflm I. I. .I
WQ@ZOHWH>Omm WH¢ZOHWH>Omm MBHI3

ZdUHMMZdIZdUmeE

 

 

. 3o; H u
05Hm> HMUHHHHO .mo. n o «ummu OmHHmuImco .aowusnwuumwc N OGHms
mHMGOHmH>OHm QMOHHOENIGMOmeE cam mHmsome>oum muHc3 mo comHHmmEOUII.m4 mamas

 

72

oz Om. mo. mO. on.H Om NO. mn.H OHm mocmHom HMHoom
IImOmum>< usflom mOMMO

mow mO.N Hm. hm. mv.H Ov mm. vn.H Oov numz\mocmflom
IIwOmum>m UCHom mwmuo

mm» mO.H OH. mm. vO.H Ov mv. NH.N OOm anHOcm
IImOmum>< ucHom mOmuO.

oz 50. no. 5v. NO.H Ov mv. OO.H va ApcmsmomHm
Omocm>Om ozv mmmnm>m

pCHom OOMHO Hmuoe

mow mO.mI Nv.I mm. ON.N Ov mv. OO.H VNv mOmHm>< UCHom
wOMHO Hmuoe

mm» OO.vI h.OMI O.Hth O.mO mv O.thH O.mm va mUCHom
mOmnw Hmuoe

mm» vN.mI H.OHI m.mmv N.Om Ow m.hVN H.ON va OmumHmEOU mndom
Hmummamm Hmuoe

mos sm.H Om. me. us.H we mo. om.a «No omouo><
usaoo moons
Hmpmemm umnwm

 

unmoH N NxIHx Nm Nx H: mm Hx Hc mwumuwuo
Imaamam I._I m I. .I
mH¢ZOHmH>Omm mHQZOHmH>Omm mBng

240HmmE¢IZ¢UmeE

 

 

OmscHuGOOII.mH mqm<e

73

ANALYSIS OF VARIANCE

 

The three way analysis of variance performed for

the regular and Provisional samples examines their re-

lative performance on the following measures:

SAT-Verbal,

SAT-Math, high school rank, first semester in college

grade point average, total college grade point average,

total semester hours earned, and total grade points

earned. Tables 14-20 present the results of each of the

analyses in question.

The sources of variance are as follows:

Sex

UJ>'

First Semester Enrolled (Spring 1972,

Summer 1972, Spring 1973, Summer 1973)

C)
"

TABLE l4.-—Ana1ysis of variance for SAT-Verbal

Mode of Admission (Provisional or regular)

 

 

 

PROBA-

SOURCE DF SS MS F BILITY

A 1 42505.29 42505.29 7.51 .0063

B 3 849995.36 283331.79 50.03 .0000

C l 2095317.79 2095317.79 369.97 .0000

Interaction Effects

AB 3 13565.61 4521.87 .7984 .4949

AC 1 1912.16 1912.16 .3376 .5614

 

74

TABLE 14.--Continued

 

 

 

PROBA-

SOURCE DF SS MS F BILITY
BC 3 383852.74 127950.91 22.59 .0000
ABC 3 35887.35 11962.45 2.11 .0972

within 850 4813976.29 5663.50

 

The analysis of variance involves the use of unequal
cell sizes. Also, since some of the data items for several
students are missing or incomplete, certain regular students
are omitted from the analysis. The data for Provisional
students is complete.

Table 14 presents the ANOVA results for the variable
SAT-Verbal. Highly significant differences are shown using
sex, first semester enrolled, and mode of admission as
sources of variation. The interaction effects reveal that
only when the latter two variables are used do we find

significant differences among groups.

75

TABLE 15.--Ana1ysis of variance for SAT-Math

 

 

 

PROBA-

SOURCE DF SS MS F BILITY

A l 261993.73 261993.73 45.66 .0000

B 3 631139.42 210379.81 36.67 .0000

C 1 1975013.90 l975013.90 344.23 .0000
Interaction Effects

AB 3 46493.46 15497.82 2.70 .0446

AC 1 5421.13 5421.13 .9449 v.33l3

BC 3 527025.44 175675.15 30.62 .0000

ABC 3 85505.91 28501.97 4.97 .0020

within 850 4876861.43 5737.48

 

Table 15 demonstrates the ANOVA results for the
variable SAT-Math. Again, significant differences are
shown using sex, first semester enrolled, and mode of
admission as sources of variation. As with SAT-Verbal, the
interaction effect using variables B and C reveals signif-
icant differences. Unlike the SAT-Verbal score, SAT-Math
also shows significant differences for the interaction of
sex and first semester enrolled, as well as for all three
factors combined.

The third variable examined is the students' rank

in class. Table 16 is used to demonstrate the analysis of

variance in this case.

76

Significant differences are found

for each of the sources of variance taken separately. In

addition, each interaction affect is significant with the

exception of sex by mode of admission.

Table 17 presents the data for first semester in

college grade point average.

The variables sex and mode

of admission are shown to have a significant effect on

this variable when taken separately.

The variable first

semester enrolled does not show significant differences.

TABLE l6.--Ana1ysis of variance for rank in class

 

 

 

PROBA-

SOURCE DF SS MS F BILITY

A 1 13473.25 13473.25 35.55 .0000

B 3 15554.27 5184.76 13.68 .0000

C l 112008.32 112008.32 295.55 .0000

Interaction Effects

AB 3 8987. 64 2995.88 7.90 .0000

AC 1 881.78 881.78 2.33 .1275

BC 3 26667.61 8889.20 23.46 .0000

ABC 3 3487.08 1162.36 3.07 .0273
within 846 320619.91 378.98

 

77

TABLE 17.--Analysis of variance for first semester GPA

 

 

 

PROBA-

SOURCE DF SS MS F BILITY

A l 3.82 3.82 7.02 .0082

B 3 3.03 1.01 1.86 .1353

C l 57.40 57.40 105.37 .0000

Interaction Effects

AB 3 8.00 2.67 4.90 .0022

AC 1 .0011 .0011 .002 .9646

BC 3 21.79 7.26 13.34 .0000

ABC 3 .6432 .2144 .39 .7576

within 836 455.38 .5447

 

The interaction effect between sex and first semester
enrolled is significant at the a = .05 level. Similarly,
the interaction of first semester enrolled and mode of
admission produces significant differences. The other
interactions examined do not have a significant effect on
the variable first semester grade point average.

Table 18 reflects the ANOVA for variable five, total
college grade point average. The results are the same as
those for first semester grade point average except that

in the present case, first semester enrolled is a significant

78

source of variance whereas it was not for first semester
grade point average. Sex and mode of admission also cause
differences to occur in total grade point average. The inter-
actions of first semester enrolled with sex and mode of
admission, respectively, also cause significant differences.
The interaction of sex with mode of admission, and the inter-
action of all three sources of variance taken together, do
not cause significant differences.

TABLE 18.--Analysis of variance for total college grade
point average

 

 

 

SOURCE DF SS MS F PROBA-

BILITY

A 1 3.69 3.69 7.61 .0059

B 3 5.34 1.78 3.67 .0120

C l 46.82 46.82 96.72 .0000

Interaction Effects

AB 3 6.69 2.23 4.61 .0033

AC 1 .1697 .1697 .3506 .5540

BC 3 14.32 4.77 9.86 .0000

ABC 3 1.24 .4149 .8570 .4630
within 839 406.18 .4841

 

79

An analysis of variance is also performed for total
semester hours earned and total grade points accumulated.
Tables 19 and 20 show the results for each analysis

respectively.

TABLE l9.--Analysis of variance for total hours completed

 

 

 

PROBA-

SOURCE DF SS MS F BI LI TY

A 1 185.08 185.08 .7486 .3872

B 3 68587.40 22862.47 92.47 .0000

C 1 7224.07 7224.07 29.22 .0000
Interaction Effects

AB 3 744.46 248.15 1.00 .3905

AC 1 1011.93 1011.93 4.09 .0434

BC 3 8513.89 2837.96 11.48 .0000

ABC 3 2585.08 861.69 3.49 .0155

within 839 207429.41 247.23

 

80

TABLE 20.--Analysis of variance for total grade points

 

 

 

earned

PROBA-

SOURCE DF SS MS F BILITY
A 1 94392.61 94392.61 36.63 .0000
B 3 276708.56 92236.19 35.79 .0000

C 1 99050.74 99050.74 38.43 .0000

Interaction Effects

AB 3 275740.67 91913.56 35.67 .0000
AC 1 98278.08 98278.08 38.13 .0000
BC 3 27874.72 9291.57 3.61 .0131
ABC 3 44323.51 14774.50 5.73 .0007

within 839 2162216.67 2577.14

 

In the case of semester hours completed, only first
semester enrolled and mode of admission are sources of
variance that produce significant differences. The inter-
action effect of these two variables produces significant
differences, as does the interaction of sex with mode of
admission. The interaction effect of all three sources
of variance taken together also is significant at
a = .05.

The ANOVA for total grade points earned shows highly

significant effects produced by each source of variance

73m. i
.nterc

”'1‘ '1
:23“; I .3

n:
(I)
H

when
with

.‘c
61216

.
£3381

lmCr
‘

L

81

taken separately, as well as for each of the interactions.
Each effect is significant at a = .05 and all but the
interaction of first semester enrolled with mode of
admission would be significant at a = .01.

To summarize the above findings relating to the
ANOVA tables, highly significant effects are found when-
ever mode of admission is a source of variance. Also,
when the interaction effects of mode of admission taken
with first semester enrolled are studied, significant
differences are found for each of the seven variables
under consideration.

In addition, sex appears to be a source of variance
with some impact. Taken alone, sex produces significant
effects in six of the seven variables studied, the lone
exception being total semester hours completed. When
used in the interaction analysis, sex and first semester
enrolled produce significant effects for five of the seven
variables; sex with mode of admission has somewhat less
impact, significantly affecting only the two variables
total hours and total grade points earned.

Taken alone, first semester enrolled is a signif-
icant source of variance in six of the seven variables,
the only exception being first semester grade point average.

It is evident from the above that there are some

important factors at work within the three sources of

acc
tot

icar

abo
and

the

82

variance studied. These factors significantly affect the
students' college performance. At the same time, it is
evident that variances in the students' high school and
SAT records can be tied to their sex, first semester

enrolled, and mode of admission.

POST-HOC COMPARISONS

 

In order to gain more specific insight into the
nature of the differences revealed already, it is helpful
to employ the technique of post-hoc comparisons. To
accomplish this, Scheffe's technique is used in relation
to the analysis of variance tables given above. A signif-
icance level of a = .05 is used, and the analysis is done
accommodating the unequal sample sizes. The variables of
central concern are SAT-Verbal, SAT-Math, high school rank,
first semester grade point average, and total grade
point average.

Table 21 displays the mean values for each of the
above measures of performance as demonstrated by regular
and Provisional students who entered the University during
the semesters in question.

The results of the test using Scheffe's technique
for SAT-Verbal are shown in Table 22. Significant dif-
ferences are found between groups when regular and Pro-

visional students are compared. This corresponds with

833

 

hO.N

MH.N

N.hv

mov

0mm

Hv.N

hm.N

Om.N

mv.N

Omv

OO.N

mn.N

OO.H

OO.H

Nm.N

mm.N

OO.N

hO.N

ov.H

Nm.H

vv.N

Nv.N

OO.N

mb.N

VO.H

OO.N

VO.N

OO.N

OO.H

mO.H

vm.N

hO.N

vm.H

HO.H

nv.N

mv.N

ONm

OO.N

OO.H

mov

Nm.N

mm.N

Ov.H

Nm.H

0.0m

va

NHv

Nv.N

mm.N

H.OO

nOm

mmmuo>< ucHom
womuu Hmuoe

mmmum>¢ uCHom
momuo uwummEmm umuHm

xcom Hoozom Boa:

BuozIaam

Hoouo>Ie<m
mammmmmmmmm

mmmum>< ucflom
mcmuu Huuoa

monum>¢ USHom
mpmuo umummemm umuHm

xcom Hooaom new:
zoozIeam
Hmouo>Ie<m

m<ADUmm

 

zuom

oHNEmm

mez

MBOH muzzam

zoom

meEwm

OHMZ

MNOH Dszmm

zoom

mHmEmm

mHmz

NhOH mmZZDm

room

meEmm

OHM!

NhOH Oszmm

 

 

noumeON umuflm an mucmpsum HmconH>oum 0cm umHsomu

pmcHnEoo can mmHmEou

.mmHuE I OmHHoucm
new mommuo>m ucHoa momuo mowHHoo ocm .xcou Hoonum cow: .ecm cmszI.HN mqmde

84

the previous finding that regular students as a whole
scored significantly better than Provisional students.

At the same time, each of the Provisional groups
appears to have roughly the same scores on SAT-Verbal.

In no case do the differences noted appear to approach
the level of significance. No discernable pattern is
noted with respect to the possible differences existing
between Spring and Summer entrants, nor between entrants
in the Spring and Summer of one year in contrast to the
other. The same can be asserted for regular students.
No significant differences are found between any of the
pairs of regular groups studied.

As with SAT-Verbal, there is a highly significant
difference between regular and Provisional students on
the SAT-Math score for each of the semesters studied.
Table 23 shows the relevant data for SAT-Math; it can
be seen that no particular pattern exists when comparing
groups of Provisional students. The differences in scores
are not significant, as was the case for the SAT-Verbal.
Regular students also demonstrated some differences but
they are not statistically significant.

Table 24 shows that essentially the same phenomenon
exists for the measure of high school rank in class as
that reported for SAT-Verbal and SAT-Math. That is, in

each semester, regular students performed significantly

85

better than Provisional students. Provisional students
presented essentially equal rank in class in each of the
semesters; this also is true for the regular student
samples.- In addition, no perceptible pattern seems to
operate with regard to the small differences noted within
the Provisional and regular samples for each semester.

In general, the evidence is strong that Provisional
students presented basically the same quality of high
school record and entrance test scores in each of the
semesters in which they started. Regular students
demonstrated the same tendency.

In examining the performance in college of regular
and Provisional students, controlling for first semester
enrolled, several points can be made. Tables 25 and 26
present the data resulting from the use of Scheffe's
technique for first semester and total grade point aver-
ages respectively.

When Provisional students are compared controlling
for the first semester enrolled, no significant differences
in first semester grade point average appear except when
the Spring 1973 and Summer 1973 Provisional groups are
examined. The difference of .57 in the first semester

grade point average is significant at the a = .05 level.

ru.» —I. H hymn :5... It..%.& Ur...C.&.I,u. 1— "..Ir ..r i Ni; "V. I+Crt~t: NU.
JZZC‘£LPCHL 3:0 U54355h HOH ICROUE HEELS>I5<$ Q0 COWflHQQEOUIIINm u§m<fi

 

86

mos Nv.oo mm.a A mHHI omm boo .mom ms. nmm >ouo me. New

 

66% nm.vm mm.H a mmaI mmm ham .mom ms. sum .>oum mm. 55m
oz nm.mm mo.H a mm Ame on .oom ms. sum .mom mm. “mm
oz mN.Hw mm.H s H mmm omm .mom me. saw .mom ms. mom
oz mm.ov mm.a s 0 6mm omm .mom mu. new .mom ms. Ham
oz mo.mm mm.H 5 mm was mmm .mom ms. sow .mom mm. 55m
oz oo.mo mm.H s m omm mom .>ouo ms. sow .>ouo ms. Nam
oz Am.mm mo.H A om 5mm sow .>ouo ms. sow .>ouo ms. Ham
oz am.oo mm.H n ma mmm moo .>ono me. New .>oum NA. Ham
oz om.sm mm.H s e omm pom .>oum ms. sum .>oum NA. sum

unmoH B m HIx fimIfim mm fim Nx Hx

Iuacmam
monomo

 

 

OOHHoucm HmummEmm umHHm an mucmosum
HNGOHmH>oum can HmHsOmH How mmnoom Hmnum>IB¢m mo CONHHMQEOUII.NN mnmfla

87

NC HG

 

 

 

 

A.w +.vaz|7. n mm mo. n o

mOO u Eovwmnm mo mmmHOmQ
mm.m u a OAHIxO 7. n m om.moom n ouoswm coo:
oz om.mm mm.H s on mom Omm .mom sow Ham .mom New Haa
oz om.mm mm.H n ma mmm mmm .>oum sum Haa .>oum Ham Ham
mos mm.oa mm.H s mmHI omm mom .oom ms. Ham .>ouo me. Nam
mo» ms.mm mm.a s noHI new omm .amm ms. sum .>oum ms. saw
unmoH B m HIx NxIHx Nx Hx Nx Hx

Iwacmam
monomu

 

 

©ODGHHCOUII.NN mqmdfi

 

88

 

mow No.4m mm.H AoHI mam mow .moz mA. sow .>oum NA. sum
mo» No.06 mm.H NmHI Aqm mas .oom NA. new .>oum NA. Nam
mo» mm.om mm.H HNHI Hmm oaa .oom NA. sum .>oum NA. sum
oz mo.mm mm.N A NH mam mmm .mom NA. sum .mom NA. pom
oz om.ao mm.a A OH Hmm Aqm .mom NA. sum .mom NA. Nam
oz ON.HN mm.H A NH mmm Aom .oom mA. Ham .moz NA. mam
oz vm.mm mm.a A ma on Hmm .mom mA. sum .mom NA. sum
oz mm.ma mm.a A «I moo mos .>ouo NA. sow .>ouo MA. New
oz mm.om mm.H A m 0H4 mas .>ouo NA. sow .>ono NA. Adm
oz mN.Ao mm.H A OH moo was .>oum mA. umm .>ouo NA. Nam
oz am.AN mm.H A H mos ofiv .>onm NA. sum .>ouo NA. sow
#COUH .H. M HIvH NxIHN Nun Hun Nun Hun
Imaaoam
mmsomu

 

 

OmHHoucm HmummEmm umuwm an mucmosum
HocOHmH>onm 0cm umHsOmu How mwuoom nuszedm mo somwummEOUII.mN mqmde

89

 

 

IHI + M m2 H mm
mo. n a
mOO u Eoowmum mo mmmHOmQ
mm.m u e OAHIxV7. u m mv.AmAm n ouosam coo:
oz mO.mN OO.H A OH MNm va .Omm Esm HH< .Omm Hmm HHm
oz mv.mm OO.H A o Oov mow .>0Nm 85m HHm .>oum Hmm HHO
mow mN.om OO.H A OMHI mmm mov .Omm MA. Hmm .>0Hm mA. Hmm
unmofi B m HIx NxIHx Nx Hx Nx Hx
48on I I I I

mmDQmU

 

 

OmssHHGOUII.MN mqmfle

90

 

 

mo» 4A.N mm.H o.ANI N.NA N.Ao .mom NA. sum .>ouo NA. sow
mos NO.NH mo.H N.NNI m.OA N.AN .moz NA. “mm .>oum NA. “mm
mm» NN.N mo.H v.6NI A.NA N.Nv .mom NA. saw .>ouo NA. sum
oz NH.oH NN.H N.NI N.NA A.oA .mom NA. saw .moz NA. Nam
oz Ao.oH mo.H N.NI A.NA m.oA .mom NA. 25m .mom NA. Nam
oz mm.OH NN.H N.I A.OA N.OA .mom NA. “mm .mom NA. Nam
oz ON.oH mo.H N.N N.NA A.NA .mom NA. sum .mom NA. sum
oz HN.HH NN.H N.NI N.Av N.AN .>ono NA. saw .>ouo NA. Nam
oz No.4H NN.H H.NI N.Nv N.AN .>oHo NA. sow .>ouo NA. Nam
oz NN.AH mm.H o.I m.AN N.AN .>ouo NA. mam .>ouo NA. “mm
oz No.A NN.H N.NI N.Aa N.Nv .>ono NA. saw .>ouo NA. sum
unmoH a m wMIfim NM NM Non Hx
IOHoon
monomo

 

 

OmHHouam Hmumemm umHHm An mucmvsum
HmconH>onm was HMHSOOH How xcmu Hoonom nOHn mo somHHmmEOUII.vN wands

91

N
I:
H

 

 

.m m I mm
H + H mgz .. .
No I o

HOO n Eocmwum mo mmmHOmO

mm.m u A NAHIxcvx u m NN.NAN u muooom com:
oz ON.A mm.H A N.NI N.NA o.OA .mom sow HHN .mom Adm HHN
oz NH.N mm.H A N.AI N.Nv N.AN .>onm saw HHN .>on Adm HHN
mo» HN.NH No.H A N.NNI A.oA N.AN .mom NA. “mm .>onm NA. Nam

unmoH B m HIx NxIHx Nx Hx Nx Hx
Imaooam I I I I
monomo

 

OmscfluGOUII.vN mHmda

 

92

mm.» mm. OO.H A vv.l AmN MH.N .Omm MA. 55m (NORA MA. 85m

 

mm» mm. mm.H A NN.I No.N HO.H .mom NA. umm .>oum NA. “mm
mm» 6N. mm.H A NA.I NA.N Am.H .mom NA. sum .>oum NA. sum
oz mN. NN.H A No.I AN.N NN.N .oom NA. saw .oom NA. Nam
oz oo. NN.H A ON.I NA.N No.N .oom NA. saw .mom NA. “mm
oz 04. NN.H A OH.I NN.N Ne.N .oom NA. NON .mom NA. Nam
oz AN. NN.H A NH. AN.N NA.N .moz NA. sum .mom NA. sum
mo» Na. NN.H A AN.I NH.N ON.H .>oum NA. sum .>onm NA. HON
oz mm. NN.H A ON.I Am.H HO.H .>onm NA. saw .>ono NA. Adm
oz oo. NN.H A No. ON.H Ho.H .>oum NA. HON .>oum NA. Adm
oz AN. NN.H A OH.I NH.N AN.H .>oNo NA. saw .>oum NA. sow
OomoN A m HIN mmIHm mm Hm Nx Hx
INHamNN
mmoomo

 

 

OOHHOHGO HOumOEOm umuHm an mucOOsum HOGOHmH>OHm cam
HMHOOOH NON Ommnm>m ucHom OOOHO HOumOEOm umHHm mo cOmHHmmEOOII.mN mqmde

93

N: H:

 

 

M+Mmz H mm
No. I o
Hmm u EOOOOHO mo mOOHOOQ
mm.m n A NAHINO7. n m Aevm. u oposom zoo:
oz NN. mm.H A OH.I OO.N om.N .oom sum HHN .mom New HHN
mo» NN. NN.H A Ao.I No.N NN.H .>oum sum NHN .>oum NON HHN
mo» av. NN.H A NN.I NN.N ON.H .mmm NA. Nam .>ouo NA. “mm
ucmOH B m HIx NxIHx Nx Hx Nx Hx
INHcmHm I. I. .I ..
mmsomo

 

 

OODGHOGOOII.ON mamme

 

94

mm.» Hm. OO.H A vm.I Hv.N AO.N .Omm MA. 55m (Ohm MA. Edm

 

mo» ON. NN.H A NO.I AO.N OO.H .mmm NA. udm .>oum NA. Ham
mo» NN. NN.H A NO.I oO.N NN.H .mmm NA. saw .>oum NA. sum
oz ON. NN.H A HH. Hv.N NN.N .mmm NA. sum .mmm NA. “mm
oz NN. No.H A NH.I OO.N AO.N .mom NA. saw .mmm NA. Adm
oz NN. NN.H A No.I NN.N AO.N .mom NA. NON .mmm NA. Nam
oz AN. NN.H A NH. HO.N OO.N .mmm NA. sum .mmz NA. sum
oz Na. NN.H A NN.I Ao.N OO.H .>oum NA. sow .>oum NA. Nam
oz NN. No.H A No. NN.H OO.H .>o»m NA. sow .>on NA. new
oz NO. NN.H A OH. OO.H OO.H .>oum NA. Nam .>oum NA. Nam
oz NN. NN.H A NH.I Ao.N NN.H .>oum NA. sow .>oum NA. sow
HomoH a O HIx wmIHm mm Hm Nx Hx
IOHomHm
. mosomu

 

 

OOHHOHGO HOumOEOm umuHm NO musmusum HOGOHmH>OHm
can HOHSOOH MOO OOOHO>O ucHom OOOHO Hmuou mo GOOHHOQEOUII.ON mammB

95

 

 

I._IH.+IHIm "mm
No. I o
«mm M EOOOOHm Ho mOOHOOO
OO.N u A OHHIHL< u m Heme. u mumsvm ommz
oz ON. No.H A No. NO.N OO.N .mmm sow HHa .mmz HON HHN
oz NN. NN.H A ON.I oo.N «A.H .>oua saw HHN .>oum Adm HHN
mm» Oa. OO.H A «N.I NN.N OO.H .mmm NA. umm .>oum NA. Nam
ucmOH B m HIx NxIHx Nx Hx Nx Hx

IOHcmHm
mmsomo

 

 

OOSGHUGOOII.ON mqm<a

96

Further study shows that the difference in first
semester grade point average between the Spring 1972 and
Summer 1972 Provisional students was .36; this reflects
a substantial difference, but it is not significant with
a = .05. The size and direction of the difference is
consistant with that found in the 1973 Provisional groups.

Therefore, the following null hypothesis is tested
and the results show it to be disproved for the 1973
Provisional groups and upheld for the 1972 Provisional
groups.

H03: "There is no significant difference be-
tween the first semester in college grade point average
of students admitted to the Summer Provisional program
and that of students admitted to the Spring Provisional
Program."

When both Spring groups are pooled, and both
Summer groups pooled, a significant difference appears
and the null hypothesis is disproved. Table 25 demon-
strates these findings, with a difference of .47 in the
first semester grade point average of Spring versus
Summer Provisional students significant at the a = .05
level.

The lack of significant difference between the
Summer 1972 and Summer 1973 Provisional groups supports

the following null hypothesis.

97

H02: "There is no significance between the first
semester in college grade point average of Provisional
students offered extensive support services from that of
students not offered the same services." It should be
noted, however, that the Summer 1973 group showed a
modestinpmovement(.16) over the Summer 1972 group in
first semester grade point average. This occurred despite
the apparent similarity in the high school and test
records presented by each group.

When each Spring Provisional group is compared,
the difference of .05 in grade point average is small;
this difference is not close to the level required for
statistical significance.

The regular student samples for each semester
earned essentially equal first semester grade point aver-
ages. Unlike the Provisional samples, regular students
did not have a markedly lower performance in Spring
semester than they did in Summer session. The differences
shown of .30 and .02 between Spring and Summer 1972 and
Spring and Summer 1973 respectively are not significant
at the a = .05 level. When the Spring groups are pooled
and compared with the combined Summer groups there is no
significant difference shown in first semester grade

point average.

98

The data of Table 25 reaffirms the prior analysis
which indicated strong differences between the college
performance of regular and Provisional students. However,
for the 1973 Summer groups the Scheffe's test shows a
smaller difference between Provisional and regular students.
The Summer 1973 Provisional students maintained a mean
first semester grade point average that was .44 lower than
that for regular students. With T in this instance being
.33, however, there is an indication of less marked
difference between regular and Provisional students in
Summer 1973 than for the other semesters in question.

When considering total grade point average, the
findings are similar to those for first semester grades.
Strong differences between regular and Provisional students
persist.

For each of the semesters in question, the dif-
ferences in total grade point average are significant
at the a = .05 level. The Summer 1973 groups come closest
to having non-statistically significant results, with
the measured difference of .34 comparable to the T value
of .31. The difference for the Spring 1972 groups is
much larger (.63), but the T value of .56 makes that
difference also just over the level of significance. It
is therefore evident that regular students performed

better than Provisional students when measured by total

99

grade point average, but the magnitude of this differ-
ence was somewhat variable across semesters.

The comparisons of total grade point average for
Provisional students who entered in each of the four
semesters in question reveal no significant differences.
However, it is evident that the Summer 1973 group ob-
tained higher grades than did other Provisional groups,
particularly when contrasted with the Spring 1973
entrants.

It should also be noted that the tendency shown
above for Summer Provisionals to have higher first
semester grade point averages than Spring Provisionals
is sustained when total grade point average is examined.
Taken separately, the 1972 Summer and Spring Provisionals
differ by .08 gpa; the 1973 groups differ by .39. When
the Spring groups are combined and compared to the pooled
Summer Provisionals, the overall difference in total
grade point average is .26, with the Summer group again
showing superiority. These differences, however, in
contrast to those for first semester grade point average,
are not statistically significant with a = .05.

When total college grade point averages are
compared for each semester's group of regular students,
we find somewhat the same pattern operating as in the

case of the Provisional students. That is, some

100

difference is found in college performance for regular
students from each semester, but none of the differences
approaches significance.

The slightly superior first semester performance
for Summer regular students compared to Spring regulars
is reversed for the total grade point average of the 1973
groups. In this case the Spring entrants performed some-
what better overall, though the difference is not close
to significance. This means that the tendency for
Summer Provisionals to perform better than Spring Pro-
visionals on total college grade point average is not
the case for regular students. The tendency in the

Provisional group seems to be unique to that sample.

DEGREE OF PERSISTENCE

 

Having examined in some depth the relative per-
formance of regular and Provisional students, as measured
by first semester and total grade point average, it is
also helpful to test the students' persistence in en-
rollment.

This variable can be examined while controlling
for sex, ethnic background, and the semester in which

the student first enrolled.

101

Table 27 presents the basic data reflecting the
relative persistence of regular and Provisional students
accounting for the sex and first semester of enrollment.

For purposes of the study, students enrolled
in the Spring 1974 semester at the University of Texas
at Austin are counted as "currently enrolled." This
explains the obviously higher proportion of students
currently enrolled who began in the Summer 1973 semester,
and leads to the conclusion that persistence cannot be
used to compare groups that started in different semesters,
since those who began most recently would naturally have
a higher tendency to be currently enrolled.

That being the case, the most noteworthy findings
are those which compare regular and Provisional students
within semester and as a whole.

For instance, of all the Provisional students who
began at the University during the four semesters in
question, 59% were enrolled in the Spring 1974 semester.
This percentage compares to 66% of the regular students
currently enrolled.

Table 28 demonstrates the chi square test that
is used to measure this difference between regular

student and Provisional student persistence.

102

 

 

 

 

 

 

 

NOO OOO OON OOH NO OOH OON OOH ON OOH Hmuoe
HO OOH ON ANH ON HA NH NH NN ON NN NN HO OO NN NN OO OH NN NN oz
ON AON OO NON OO ANH NO NO AO HN OO OO ON ONH NO NO NN O NO NO OON
mfiom
NON OON AOH NN NN NN OHH OO AH NO HOOoA
OO OOH NN OA ON ON OH O NN A NN NH HO OO AN NN NO HH NO ON oz
OO ONH AO ONH OO OO OO OO OO NH NA ON ON OO NO ON NN O ON OH OON
.mmmmm
NNN HOH HOH NO NN OO OO OO HH ON Hmuoe
NO OO NN AO NN NN NH O OA AH OO OH HO ON NN NH NA O ON ON oz
ON ONH NO ONH AO OO AO ON ON O OO ON ON NN OO AN AN N ON ON OON
mmm
w S w S w G O c O G O S w S w C O c w G
.>OOO .OOO .>OOO .OOO .>OOO .OOO .>OOO .OOO .>OOO .OOO
.mmmmm NAOH OOzzOO NAOH ozHOOO NAOH Omzzsm NAOH quOON

 

 

OOHHOHGO HOumOEOm umHHw cam xmm How OQHHHOHOCOO
mpcmosum HOGOHmH>oum new HOHDOOH mo OGOEHHOHGO OQOHHSUII.AN mquB

103

TABLE 28.--Chi square contingency table for persistence in
enrollment of all Provisional versus all regular

 

 

 

 

students
PROVIS IONAL REGULAR TOTAL
Enrolled 287 363 550
Not Enrolled 198 137 33
Total 485 400 885
Chi square = 4.028 Degrees of Freedom = l
a = .05 Critical Value of chi square = 3 841

The chi square contingency table shows that the
value separating the upper five percent of the distribution
with 1 degree of freedom is 3.851. Since the chi square
valuecflfTable 28 is 4.028, there is a significant dif-
ference shown.

It is apparent that the proportion of regular
students enrolled in Spring 1974 is significantly higher
than the proportion of Provisional students enrolled in
that semester.

The relative persistence of regular and Provisional
males is also examined. Table 29 presents the chi square

test for this comparison.

TABLE

"I“?-

104

TABLE 29.--Chi square contingency table for persistence in
enrollment of male Provisional versus male
regular students

 

 

 

 

MALE MALE
PROVISIONAL REGULAR TOTAL
Enrolled 129 124 253
Not Enrolled 94 67 161
Total 223 191 414
Chi square = 2.166 Degrees of Freedom = l
a = .05 Critical value of chi square = 3.841

Male regular students do not show a significantly
higher level of persistence than do male Provisional
students, since the chi square value is less than 3.841.

Female persistence is noted in Table 30.

TABLE 30.--Chi square contingency table for persistence in

enrollment of female Provisional versus female
regular students

 

 

 

 

FEMALE FEMALE
PROVISIONAL REGULAR Total
Enrolled 158 139 297
Not Enrolled 104 70 174
Total 262 209 471
Chi square = 1.920 Degrees of Freedom = l
a = .05 Critical value of chi square = 3.841

the CI
Show I

do fe:

persi

to do

visic

”‘WV .-

ihbut

the ]

admi:

preSI

105

Since the chi square value of 1.92 is less than
the critical value for a = .05, female regular students
show no significantly greater level of persistence than
do female Provisional students.

Finally, it is useful to compare the relative
persistence of male and female Provisional students, and
to do the same for regular males and females.

Table 31 presents the chi square test for Pro-
visional male and female students.

TABLE 31.--Chi square contingency table for persistence

in enrollment of male Provisionals versus
female Provisionals

 

 

 

 

MALE FEMALE TOTAL
Enrolled 129 158 287
Not Enrolled 94 104 198
Total 223 262 485
Chi square = .301 Degrees of Freedom = l
a = .05 Critical value of chi square = 3.841

Clearly no significant difference exists between
the persistence of males and females in the Provisional
admission program.

The comparable data for regular students is

presented in Table 32.

PEI

to

106

TABLE 32.--Chi square contingency table for persistence
in enrollment of male regular versus female
regular students

 

 

 

 

MALE FEMALE
REGULAR REGULAR TOTAL
Enrolled 124 139 263
Not Enrolled 67 70 137
Total 191 209 400
Chi square = .111 Degrees of Freedom = l
a = .05 Critical value of chi square = 3.841

As in the case of Provisional students, there
appears no significant difference in the persistence of
male regulars when compared to that of female regular
students.

It is also instructive to examine the degree of
persistence for regular and Provisional students according
to the semester in which they first enrolled.

Table 33 presents the relevant data in this regard.

107

TABLE 33.--Chi square contingency tables for regular and
Provisional students by first semester enrolled

 

 

REGULAR PROVISIONAL TOTAL

 

Spring 1972

 

Enrolled 45 9 54
Not Enrolled 55 19 74
Total 100 28 128

Chi square = 1.483

Summer 1972

 

Enrolled 65 120 185
Not Enrolled 35 84 119
Total 100 204 304

Chi square = 1.075

Spring 1973

 

Enrolled 68 21 89
Not Enrolled 32 24 56
Total 100 45 145

Chi square = 5.958

Summer 1973

 

Enrolled 85 137 222
Not Enrolled 15 71 86
Total 100 208 308

Chi square = 12.286

 

(For all semesters, 0 = .05, DF = 1, Critical value of chi
square = 3.841)

Amerj

Pers:

aCCOL

Whit
Ame r
purp

8am;

and

108

Significant differences in persistence are shown
between regular and Provisional students who entered in
the Spring and Summer of 1973. However, for both Spring
and Summer of 1972, there appears no significant difference
in the tendency to persist.

These findings, in conjunction with those above
relating to the overall persistence of regular and Pro-
visional males and females, support the contention that
at least in certain instances there is not a great deal of
difference in the ability of regular and Provisional students
to stay in school.

The relative persistence of white and Mexican-
American students is also of considerable interest.

Table 34 shows the frequency and percentage of
persistence for white and Mexican-American students
according to their first semester enrolled.

It is most helpful to examine the persistence of
white Provisional students relative to that of Mexican-
American Provisional students. Also, for comparison
purposes, the persistence within the regular student
sample should be noted.

There is not a wide discrepancy between white
and Mexican-American student persistence. Sixty percent

of the white Provisional students were enrolled Spring 1974,

MUSHHOIHCM. IIHNJUnuUPHHmwmw ...HmHflIH HOS—m. HOCHHOIHNIOVHUMOQ UNHOHCUMO HON UCHHHOIHUCOU
ECHO. HGHDWUH “Ho UCMOEHHIOLHCQ UCNWIHIHHJUIIIIVMI muHMNANpfi

VOIHHHIVFUNOJEN HHWCCIHMHIM>OIHPH

109

 

Ov VN

Ov NN om NH

vm ON Om NH

ONO mOm

0v HAH vm MNH

OO mmN OO NON

MN N

Om O cm H

HO vH Om H

OAH OO

Om OO OH OH

OO OHH mO NO

OO m mN N

0v N mA O

Om NO

Hm ON mm on

Ov OH AO NO

OH O
Av O Ow N

mm OH OO m

OAH OO

Hv MA Om Nm

Om OOH vO Om

H

OOH H

O O

AN

AO OH

mm O

O Hmuoe

OA A oz

NN N OON
mmwmmmmm
IzOOmez

OO HOuoe

NN AO oz

AO NO NO»

MBHmZ

 

.>Omm .wmm

H4809

w E w G

.>Omm .Omm

 

NAOH mmEZDO

w c w :

.>Omm .Umm

 

NAOH OZHmmO

w s w a

.>Omm .Omm

 

NAOH MOZZDO

& C

.>Omm

.Umm

 

NAOH UZHmmm

 

 

OOHHOHGO uOumOEOm umHHm cam ccsoumxomn OHssuO How OsHHHOHpcOO
mucmpsum HOGOHOH>OHm Osm HOHDOOH mo UOOEHHOHQO ucOHHOUII.Om mqmda

While
There
where

Iexica

TABLE

Persi

Provi

and M

110

while 54% of the Mexican-American students were enrolled.
There is a somewhat wider difference for regular students
where 66% of the white students persisted and 50% of the
Mexican-Americans did so.

Table 35 examines these differences in more detail.

TABLE 35.--Chi square contingency table for white versus
Mexican-American Provisional students

 

 

 

 

WHITE MEXICAN -AMERICAN TOTAL
Enrolled 263 26 289
Not Enrolled 171 22 193
Total 434 48 482
Chi square = .007 Degrees of Freedom = l
a = .05 Critical value of chi square = 3.841

It is evident that no significant difference in
persistence is shown between white and Mexican-American
Provisional students.

Table 36 depicts the relative persistence of white

and Mexican-American regular students.

£\“’ L7
La

eac
enr
of

‘I

Ame

CO

111

TABLE 36.--Chi square contingency table for white versus
Mexican-American regular students

 

 

 

 

WHITE MEXI CAN -AMERICAN TOTAL
Enrolled 242 12 254
Not Enrolled lg}_ 12 135
Total 365 24 389
Chi square = 2.641 Degrees of Freedom = 1
a = .05 Critical value of chi square = 3.841

Although more marked than the difference between
white and Mexican-American Provisionals, the contrast
between these ethnic groups for regular students is not
statistically significant.

The low number of Mexican-Americans enrolled in
each semester makes useful analysis using first semester
enrolled as a control factor impossible. However, the lack
of any significant difference between white and Mexican-

American students taken as a whole is of major importance.

CORRELATION COEFFICIENTS

So far, this study has presented a thorough analysis
of the entrance test scores, high school performance, and

college grades of Provisional and regular students. It

112

is also useful to analyze the most relevant factors to
be used in trying to predict college performance of each
of these groups of students. In doing so, it will be
possible to gain some insight into the interrelationships
of the variables examined.

Table 37 presents the correlations obtained between
several variables for regular students taken as a whole.
TABLE 37.--Correlation coefficients for regular students'

high school rank, SAT scores, and college grade
point averages

 

 

 

HIGH SCHOOL FIRST TOTAL
RANK SAT-V SAT-M SAT-T GPA GPA

High School Rank - .202 .196 .225 .339 .326
SAT-Verbal .202 - .548 .879 .379 .316
SAT-Math .196 .548 - .877 .303 .292
SAT-Total .225 .879 .877 - .380 .338
First Semester
Grade Point Average .339 .379 .303 .380 - .761
Total Grade
Point Average .326 .316 .292 .338 .761 -

 

The data reveals that for regular students the

high school rank and various measures of SAT performance

have roughly equal degrees of correlation with first

113

semester and total college grade point average. Each

of these correlations is only of moderate strength, with
the range being from a low of .292 for the r between
SAT-Math and total college grade point average, to a
high of r = .379 between SAT-Verbal and first semester
college grade point average.

A slight relationship is shown between high school
rank and SAT scores. The values of .202, .196, and .225
are given as the correlation between high school rank and
SAT-Verbal, SAT-Math, and SAT-Total respectively. These
values are lower than the correlations between SAT scores
and first semester and total college grade point average.

Table 38 presents the same set of data as above
but for Provisional students. In addition, data for the
students' academic and overall high school grade point
averages is presented.

The relationship between high school performance
as measured by rank in class and first semester grade point
average is clearly very low. The correlation improves to
.268 for total grade point average, but this does not
approach a strong positive relationship.

The relationship between the two measures of high
school grade point average and college grade point average
is somewhat stronger, though not very high. The cor-

relations range from .241 between academic grade average

03 @UQHO UOOHHOU UGO NOQHOUO PQW ~¥CUH

MUOMH$>M UGO
OUCUOUflMMUOU COfluQHOHHCUIIIQM WQQ<5

Acozum Smflz .mUCGUDUW ngoflmfl>0Hm HON

114

 

I ONA. HNO. OHO. ANO., OON. ONN. HHN. OOONO>< ucHom
OOOHO Hmuos

VOA. I OOH. OOO. MOH. OOH. OAN. HVN. OOMHO>¢
ucHom OOOHO
HOumOEOm umHHm

HMO. OOH. I mOA. OAA. ONm.I MAH.I AMH.I HMHOBIBOO
OHO. OOO. mOA. I ONN. OvN.I OOH.I OHH.I SHMZIBOO
ANO. MOH. vAA. ONN. I OON.I OOH.I OOO.I HMOHO>IB¢O
OON. mOH. ONm.I OON.I vON.I I I I Hamm

HoonoO OOHO

ONN. OAN. MAH.I mOH.I OOH.I I I OOO. mmmum>< ucflom
OOOHO HHOHO>O
Hoonom OOHO

HHN. HON. ANH.I OHH.I OOO.I I OOO. I OOOHO>N ucHom
OOOHO OHEOOmod
Hoozom OOH:

 

«m6 <m0 <m0
mmw OMBOMZOO H<BOB OBOE .Hmmmm> MZOM HHdmm>O UHZWQ<U<
HOBOB EmmHm IBdm IBOO Iafim HOOmUm mem HOOmUm mem HOOmUm SOHO

 

 

OOOOHO>O ucHom OOOHO OOOHHOO 0cm .mmnoom BOO .xsmn
Hoonom OOH: .mucmosum HmconH>OHm How mucOHonmmoo coHumHOHHOUII.Om mqmda

no 1
mea:
betI

and

115

and first semester grades, to .328 between overall high
school grade average and total college grade average.

There appears little difference in the correlations
of academic grade average and overall high school grade
average with first semester and total college grade point
average. Using all high school grades in computing the
students' average actually produces a slightly higher
correlation than using just the academic (college pre-
paratory) courses.

The data for the various SAT scores shows virtually
no relationship between them and performance in college as
measured by grade point average. The correlations range
between .018 and .109 for the SAT scores and first semester
and total grade point averages.

An unusual finding is that high school rank has a
negative correlation with SAT-Verbal, SAT-Math, and SAT-
Total scores. This is confirmed when the correlations
between high school academic and overall grade averages
and SAT scores are examined. For academic grade average
these correlations range from -.098 for SAT-Verbal, to
-.137 for SAT-Total. Overall high school grade average
correlates from -,105 with SAT+Math to -.l73 with SAT-
Total.

The findings for Provisional students are thus

similar to those for regular students, but show

116

several unique qualities. For example, the relationship
between SAT scores and college grades is weaker for
Provisional students than it is for regular students.

In addition, the correlations between various measures

of high school performance and college grade point averages
are weaker for Provisional students than they are for
regular students.

The property that is common to the correlation
study for both regular and Provisional students is that
none of the high school and test measures seem to show
strong relationships with college grades.

First semester in college grade point average
shows by far the strongest relationship with total college
grade average of any measure under study. The correlation
of .743 for Provisional students very closely approximates
that of .761 for regular students. It is evident that
the relationship between each samples' initial perfor-
mance in college and the total college grade point average
is moderately strong.

Having analyzed the regular and Provisional samples
as a whole, it is helpful to determine the respective
correlation coefficients for men and women within each
group. Tables 39 and 40 present the data for male and
female Provisional students respectively. Once again,
academic and overall high school grade point averages are

included in the analysis for Provisionals.

IIIIIIIIIIIIII||IIIII|||I

EUHwQHOKVG UCHOQ @wufihmw @WUHHOU @2un smmeOUmw BANW \XCQH
doszgfi :3q2 .QJCEUSdfi H520Hfifi>0h£ @HCE HOH NUcmflUflUWUCU COflUfiHWHHOUlIImm WJQQE

 

117

I OHO. mmo.l OMO. HOH.I Omm. OOm. vmm. mmmum>€ ucfiom
mpmnw Hmuoa

OHO. I OOH. me. ONO. hmH. mHm. mom. mmmum>€
ucHom mpmuo
Hmpmmemm umHHm

mmo.I OOH. I mmm. mmh. OOm.I mNH.I NNH.I Hmuoaledm
Omo. mvH. mmm. I ovm. NNN.I ONO.I Nvo.l nuMZIfldm
HOH.I ONO. mmh. ovm. I OON.I va.I OOH.I Hmnum>lfidm
ONO. hmH. OOm.I NNN.I OON.I I I I xcmm

Hoonom gmflm

mam. mam. mmH.I omo.I va.I I I Hum. mmmum>¢ pcflom
mpmuu HHmum>o
Hoonom nmflm

vmm. NON. NNH.I NvO.I OOH.I I HNO. I wOMHm>¢ HCHom
dunno UHEmpmo<
Hoosom roam

 

dmw flmw A4908 $542 Admmm> MZ¢M flaw Afldmm>0 flmw UHENQ<U¢
AdBOB MNBmmEmm lfiflm IB<m IB¢m AOOmUm mUHm HOOmUm mem Hoomum EOHE
9mmHh

 

 

mmOMHm>m ucHom mpmum mOmHHoo can .mmuoom Bfim .xcmu
Hoocom nOHn .mucwpsum HMGOHmH>0Hm mHmE Hem mucmHonumoo GOHHMHmHHOUII.mm mqmde

$356H3>m uCHOQ @UQHW 0$®HHOU Ucm .mGMOUm E<w .XCMN
Atufiznnuma 23A; ..m..dC.¢Hus.uI HECCfiTuflx/AUIHQ .UIHHJEGH Hon.“ WUCQflUflMWflaOU :OflUmemVIH.IHOIIUIIIIIOV MNWHQANIH.

118

 

I ens. mmo. nae. omfl. pom. mmm. mmm. mmmum>< ucflom
momma Hmuoe

Ohm. I mHH. MHO. HOH. MOH. OmN. MHN. mmmuw>¢
HCHom mpmuu
umummsmm umHHm

OOO. mHH. I Own. OOO. OON.I OMH.I HBO.I HmuOBIfimm
5H0. MHO. ONN. I hmm. va.I VOO.I OOO.I numZIfiém
ONH. HOH. OOO. NNN. I OmN.I OMH.I va.I HmQHm>IBmm
NON. mOH. OON.I va.I OmN.I I I I xcmm

Hoosom roam

mmm. mmm. mma.I woo.I OMH.I I I Hmm. mmmum>< pqaom
mpmuw HHmuw>O
Hoocom roam

mom. mam. duo.I coo.I «vo.I I Hmm. I mmmum>¢ ucflom
mpmuo UHEmUmom
Hoonom nmflm

 

dmw «m0 AfiBOB mBmS A<mmm> M2¢m 4&0 HAflmm>O flaw UHZNQéUfl
HdBOB MMBmmzmm Ifidm lfiflm IBdm HOOEUm mDHE AOOEUm EUHm AOOmUm SUHE
BWfiHh

 

 

mommum>m ucHom mpmum mOmHHoo can .mmuoom Bfim .xcmu
Hoozom cmwn .muampsum HmconH>oum mHmEmm Hem mucmHonmmoo :oHumHmHHOUII.ov mqmde

&C

CO

in
wi
Pr
V6
in

me

a;
mi

DC

119

The correlation coefficients for males and females
do not vary greatly from those for the Provisional sample
taken as a whole. Similarly, there is not a great amount
of difference in the respective r's of Tables 39 and 40.
However, several factors should be mentioned which indicate
some variation between males and females in terms of the
correlations under study.

The high school performance measured by grade
point average and rank in class seems to correlate some-
what better with college grades for males than it does
for females. As an example, the r for high school overall
grade point average and total college grade average is
.399 for males. For females, the same r is .255. Similarly,
academic grades in high school correlate .354 with total
college grade average for males, but only .262 for females.

The situation for SAT-Verbal is reversed. In this
instance, there is a somewhat stronger positive relation
with college grades for females than for males. Female
Provisionals show an r of .161 and .126 between SAT-
Verbal and first semester and total college grade point
averages respectively. The corresponding r values for
males are .202 and -.101.

When SAT-Math and SAT-Total are examined, there
appears to be similar correlations with college grades for
males and females. Both these measures have practically

no relationship to college grades for either sex.

SEITXE

col]

clos

TAB?

II

Big]
SAT

SAT

Fir
Gra

Tot
Poi

120

It should be stressed again that almost all the
correlations examined above are relatively weak. Any
of the variations between males and females should be
considered with that as perspective.

Tables 41 and 42 present the corresponding data
for male and female regular students respectively. It is
once again clear that the correlation between high school
rank in class and college gradesi¢3relatively low. For
males, the r for rank in class correlated with first
semester grade point average is .330. For rank and total
college grade point average the r is .289. These are quite
close to the values found for male Provisionals.

TABLE 41.--Correlation coefficients for male regular

students' high school rank, SAT scores, and
college grade point averages

 

 

HIGH SCHOOL FIRST TOTAL

RANK SAT-V SAT-M SAT-T GPA GPA

High School Rank - .117 .131 .142 .330 .289

SAT-Verbal .117 - .540 .867 .310 .260

SAT-Math .131 .540 - .888 .250 .264

SAT-Total . 142 . 867 . 888 - . 316 . 299
First Semester

Grade Point Average .330 .310 .248 .316 - .783

Total Grade
Point Average .289 .260 .264 .299 .783 -

 

121

Table 42.--Correlation coefficients for female regular
students' high school rank, SAT scores, and
college grade point averages

 

 

 

HIGH SCHOOL FIRST TOTAL
RANK SAT-V SAT—M SAT-T GPA GPA
High School Rank - .327 .395 .404 .367 .372
SAT-Verbal .327 - .546 .893 .439 .374
SAT-Math .395 .546 - .857 .371 .360
SAT-Total .404 .893 .857 - .448 .402
First Semester
Grade Point Average .367 .439 .371 .448 - . .740
Total Grade
Point Average .372 .374 .360 .402 .740 -

 

Rank in class proves to be slightly more closely tied
to college grades for female regular students. The r's for
first semester and total grade point average are .367 and
.372 respectively. This tendency contrasts with that of the
female Provisional group which shows lower correlations be-
tween high school record and college grades than the male
Provisionals. Once again, however, the differences noted are
quite small, and drawing any conclusions from them would be
tenuous at best.

In relation to males, female regular students show
distinctly more positive relationships between SAT-Verbal,

SAT-Math, and SAT-Total and the measures of college perfor-

with

for I

high
and

high
cor:
poi:

and

to '
the
col
Gas
is
do

CO:

gr.

rl

122

mance. The SAT-Verbal and SAT-Total show the strongest
correlation with first semester grade point average. This
is true for both males and females, but the latter have r's
of .439 and .448 for first semester grade average correlated
with SAT-Verbal and SAT-Total respectively. These values
are higher than the corresponding r values of .310 and .316
for male regulars.

In general, then, female regular students show a
higher positive r value than male regulars for high school
and test records correlated with college grades. Analyzing
high school rank and the three measures of SAT scores, and
correlating each with first semester and total college grade
point average, one obtains eight values for r for both males
and females. In each case, the r is higher for female
regular students than for male regulars. This does not prove
to be the case in the Provisional sample. In that instance,
the correlations of high school records and SAT scores with
college grade point averages are higher for males in seven
cases, but higher for females in five cases. Therefore, it
is apparent that males and females in the Provisional sample
do not evidence a distinctive pattern when the r's for male
correlations of high school and test information with college
grades are compared to the r's for females.

Finally, regular males and females show very similar

r's for first semester grade point average correlated with

123

total grade point average. These values, .783 and .740
respectively, are quite close to the corresponding values in
the Provisional sample of males and females, which are .716

and .770 respectively.

MULTIPLE REGRESSION ANALYSES

 

In addition to examining the single correlations
just covered, this study presents multiple regression
analyses using various combinations of high school records,
test results and college grade point averages. This is done
to develop some better means of predicting performance than
the use of high school rank, grade point average, and SAT
scores independently from one another.

Twelve distinct models are developed in order to
determine the most beneficial method of predicting college
success. Also, multiple correlations are computed for each
model. The models are applied to Provisional and regular
students as a whole, and to males and females within each
sample.

Tables 43-54 present the models along with the
respective raw weights and unit vector weights. In addition,
the multiple r is shown for each group within the models.

It is initially useful to examine the various

commonalities and differences in the data that are revealed

124

for Provisional versus regular students. In this way it is

possible to gain further insight into the ways in which

Provisional students tend to behave differently from regular

students.

TABLE 43.--Mode1 1: Multiple regression using SAT-Verbal,

SAT-Math, and high school academic grade point
average to predict first semester grade point
average

 

 

MULTIPLE r SAT-VERBAL SAT-MATH ACADEMIC GPA k

 

Male Prov.

Female Prov.

All Prov.

.309 .0003 .0014 .0039 .3878
.273 .0017 -.0001 .0030 .6901
.277 .0011 .0006 .0035 .5049

 

TABLE 44.--Model 2: Multiple regression using SAT-Verbal,

SAT-Math and high school academic grade point
average to predict total grade point average

 

 

MULTIPLE r SAT-VERBAL SAT-MATH ACADEMIC GPA k

 

Male Prov.

Female Prov.

All Prov.

.362 -.0007 .0006 .0049 .8312
.296 .0012 .0000 .0033 .6954
.317 .0004 .0003 .0041 .7090

 

Male
Fame

All

TAE

Fen

A11

125

TABLE 45.-~Model 3: Multiple regression using SAT-Total and
academic grade point average to predict first
semester grade point average

 

 

MULTIPLE r SAT-TOTAL ACADEMIC GPA k

 

Male Prov. .302 .0009 .0040 .3343
Female Prov. .250 .0008 .0030 .6363
All Prov. .277 .0008 .0035 .4999

 

TABLE 46.—-Model 4: Multiple regression using SAT-Total and
academic grade point average to predict total
grade point average

 

 

MULTIPLE r SAT-TOTAL ACADEMIC GPA k

 

Male Prov. .354 .0001 .0050 .7687
Female Prov. .288 .0007 .0034 .6306
All Prov. .317 .0004 .0041 .6900

 

TABLE 47.--Model 5: Multiple regression using SAT-Verbal,
SAT-Math and high school overall grade point aver-
age to predict first semester grade point average

 

HIGH SCHOOL
MULTIPLE r SAT-VERBAL SAT-MATH OVERALL GPA k

 

Male Prov. .352 .0005 .0013 .0054 -.1534
Female Prov. .308 .0019 -.0001 .0043 .1807

All Prov. .317 .0013 .0006 .0048 -.0246

 

126

TABLE 48.--Mode1 6: Multiple regression using SAT-Verbal,
SAT-Math and high school overall grade point
average to predict total grade point average

 

 

HIGH SCHOOL

 

MULTIPLE r SAT-VERBAL SAT-MATH OVERALL GPA k
Male Prov. .403 -.0005 .0005 .0065 .2391
Female Prov. .301 .0014 -.0000 .0041 .3331
All Prov. .337 .0006 .0003 .0052 .2338

 

TABLE 49.--Model 7: Multiple regression using SAT-Total and
high school overall grade point average to
predict first semester grade point average

 

 

HIGH SCHOOL

 

MULTIPLE r SAT-TOTAL OVERALL GPA k
Male Prov. .349 .0010 .0055 -.2072
Female Prov. .281 .0010 .0042 .1407

All Prov. .315 .0010 .0048 -.O277

 

127

TABLE 50.--Mode1 8: Multiple regression using SAT-Total
and high school overall grade point average to
predict total grade point average

 

 

HIGH SCHOOL

 

MULTIPLE r SAT-TOTAL OVERALL GPA k
Male Prov. .399 .0001 .0066 .1713
Female Prov. .288 .0008 .0041 .2703
All Prov. .338 .0005 .0052 .2122

 

TABLE 51.--Model 9: Multiple regression using SAT-Verbal,
SAT-Math, and high school rank to predict first
semester grade point average

 

 

HIGH SCHOOL

 

MULTIPLE r SAT-VERBAL SAT-MATH RANK k
Male Prov. .243 .0003 .0017 .0068 .8187
Female Prov. .276 .0022 .0001 .0068 .7712
All Prov. .240 .0015 .0009 .0069 .7652
Male Reg. .434 .0025 .0009 .0117 .0140
Female Reg. .508 .0026 .0012 .0098 -.1063

A11 Reg. .472 .0026 .0010 .0110 -.0592

 

H

TA

II

128

TABLE 52.-—Model 10: Multiple regression using SAT-Verbal,
SAT-Math and high school rank to predict total
grade point average

 

 

HIGH SCHOOL

 

MULTIPLE r SAT-VERBAL SAT-MATH RANK k
Male Prov. .347 -.0004 .0012 .0115 1.1198
Female Prov. .290 .0018 .0003 .0074 .8113
All Prov. 295 .0009 .0007 .0092 .9262
Male Reg. .388 .0014 .0013 .0088 .4004
Female Reg. .473 .0017 .0013 .0096 .2952
A11 Reg. .430 .0016 .0012 .0096 .3821

 

TABLE 53.--M0de1 11:

Multiple regression using SAT-Total
and high school rank to predict first semester
grade point average

 

 

 

MULTIPLE r SAT-TOTAL HIGH SCHOOL RANK k
Male Prov. .228 .0011 .0070 .7667
Female Prov. .246 .0012 .0066 .7276
A11 Prov. .239 .0012 .0070 .7633
Male Reg. .428 .0016 .0117 .0341
Female Reg. .492 .0019 .0098 -.0909
A11 Reg. .460 .0017 .0110 -.0364

 

etude

rang
stud
some
is,
and

the

Ver
the
fem
Thi

mod

of

129

Tables 51-54 reveal that the multiple r for regular

students shows a modest but consistently higher value
than the corresponding value for Provisional students.
Regular students as a whole show an r value of .472,
.430, .460, and .424 for models 9-12 respectively. The
corresponding values for all Provisional students are
.240, .295, .239, and .296. As previously mentioned,
part of this phenomenon is due to the restriction in
the range of SAT scores in the Provisional sample. How-
'ever, the fact that there is no more restriction in the
range of high school rank percentiles for Provisional
students than for regular students would tend to lend
some significance to the multiple r's reported. That
is, some but not all of the difference in r between regular
and Provisional student samples may be attributable to
the restriction in range of SAT scores.

Within each group, males and females obtained
very nearly equal r values. In model 9, for instance,
there is a difference of .033 and .074 between males and
females in the Provisional and regular samples respectively.
This magnitude of difference is seen to carry over to
models 10, 11, and 12 as well.

Another method of looking at the data in models
9, 10, 11, and 12 is with respect to the relative predictability

of first semester and total college grade point average.

Prm
whet
same

for

Stu
'I'he
the
re\
f0]

is

ma

Ve
th
in
1’ E

ti.

130

Provisional students produce a higher multiple correlation
when we try to predict total grade average than when first
semester grade point average is predicted. The is true
for males and females when models 9 and 10 are compared.
It is also true when models 11 and 12 are compared. The
multiple r values in models 10 and 12 are consistently
higher than those in models 9 and 11 for Provisional males
and females as well as all Provisional students together.

The opposite of this phenomenon is true for regular
students. In this sample the correlation is strongest
when first semester grades are predicted. Comparison of
the r values in models 9 and 10 for regular students
reveals these differences to be quite smallbut consistent
for males, females, and the group as a whole. The same
is true when models 11 and 12 are compared for regular
students.

The SAT score is to be used in a different
manner for Provisional students than for regular students.

Models 9 and 10 show that the weighting on SAT-
Verbal is to be substantially less for Provisional males
than for Provisional females. This should be considered
in conjunction with the earlier finding of a lower cor-
relation between SAT-Verbal and college grades for males
than for females in the Provisional samples. The reverse

is true for SAT-Math. Here the score is to be weighted

’54

2‘3

r61“

131

much stronger for males than for females. For example,
model 9 shows a raw weight of .0017 for SAT-Math for
males, but only .0001 for females in the Provisional
sample.

TABLE 54.--Model 12: Multiple regression using SAT-Total

and high school rank to predict total grade
point average

 

 

 

MULTIPLE r SAT-TOTAL HIGH SCHOOL RANK k
Male Prov. .334 .0005 .0117 1.0601
Female Prov. .277 .0011 .0074 .7495
All Prov. .296 .0008 .0093 .9074
Male Reg. .388 .0014 .0088 .4020
Female Reg. .463 .0014 .0099 .3522
All Reg. .424 .0013 .0097 .4141

 

Regular students do not show the same tendency. For
that sample, SAT-Verbal and SAT-Math are to be weighted
essentially the same. The raw weights for SAT-Verbal in
models 9 and 10 are .0025 and .0014 for male regulars. The
corresponding values for females are .0026 and .0017. For
SAT-Math the values of .0009 and .0013 for males compare
to respective values of .0012 and .0013 for females.

Analysis of the raw weights for high school rank

reveals somewhat less clear distinctions. In predicting first

132

semester grade point average, very nearly the same weight
is attached to rank in class for Provisional males and
females. Model 9 shows equal weights of .0068 for both
groups; model 11 shows weights of .0070 and .0066 for
males and females respectively. Prediction of total
grade point average shows almost equal weights on high
school rank for male and female Provisionals. Models 10
and 12 show values of .0115 and .0117 for males, and
.0074 in each case for females.

Regular students exhibited the same consistency
between males and females for the respective weights on
high school rank. For example, model 9 shows weights
on rank in class of .0117 and .0098 for males and females;
model 10 indicates weights of .0088 and .0096. Models
11 and 12 reveal very similar findings.

For predicting total grade point average, the
weight put onto high school rank is about the same for
all Provisional and all regular students. In this regard
the weights for Provisionals are .0092 from model 10 and
.0093 from model 12. The corresponding values for regular
students are .0096 and .0097 respectively.

In the case of first semester grade point average,
the weight on high school rank is very nearly the same for

Provisional and regular students. Models 9 and 11 reveal

133

weights of .0069 and .0070 for Provisional students, and
values of .0110 are found in each model for regular
students.

The final aspect of the study to be reported is
the multiple regression analyses performed for Provisional
students using academic and overall high school grade
point averages. Tables 43-50 present the data relating
to these equations.

It is evident that the multiple r's in these
tables are roughly comparable to those found when high
school rank is used as a predictor. The values mostly
range between .20 and .35 with the exceptions being minor
deviations.

The evidence is that the multiple r when overall
high school grade point average is used exceeds that in
the analyses using high school academic average. For
instance, Tables 47-50 show r values of .317, .337, .315,
and .338 for all Provisionals using overall grade average.
These values are consistently higher than those found
for all Provisionals in Tables 43-46: .277, .317, .277,
and .317 respectively. The differences are not large in
magnitude but the consistency of the direction is of

interest.

134

Taking the Provisional students as a whole, it
is also evident that total college grade point average is
somewhat more predictable than first semester average.
Model 1 shows a multiple r of .277 while model 2 reveals
one of .317. This magnitude of difference is approximated
when models 3 and 4, 5 and 6, and 7 and 8 are compared
respectively. While the differences are clearly extremely
small, the consistency of the pattern is worth noting.
In each instance the multiple r for predicting total
grades exceeds that for predicting first semester grades.

Females exhibit slightly lower multiple r's than
do males in each model. In model 1, for instance, the
values are .309 and .273 for males and females respectively;
model 2 shows values of .362 and .296 for males and
females. This pattern is continued across the other
models in which high school grades are used. While
these differences are not large, again the consistency
of their direction is worthy of comment.

Models 1 - 8 reflect a finding noted earlier when
high school rank was used in the multiple regression.
That is, SAT-Verbal score carries a considerably smaller
raw weight in the equation for males than for females.
Models 1, 2, 5, and 6 show raw weights for males to be
.0003, -.0007, .0005, and -.0005 respectively. For

females the weights are .0017, .0012, .0019, and .0014 in

135

each model. Not only are there lower values for males,
in two cases they are actually negative. This is a
further indication that SAT-Verbal is a weak predictor
of college performance for Provisional males.

The same pattern described previously with respect
to SAT-Math applies when high school grade point average
is used rather than high school rank. That is, the raw
weights for female Provisional students are consistently
less than those for males. Models 1, 2, 5, and 6 reflect
values for this weight to be .0014, .0006, .0013, and
.0005 for males. The corresponding weights for females
are -.0001, .0000, -.0001, and -.0000. It is clear that
SAT-Math is to be weighted considerably less for females
than for males within the Provisional sample when high
school average is used. Also, models 1, 2, 5, and 6 show
that SAT-Math carries much less weight than SAT-Verbal
for female Provisional students.

One other pattern that appears is that high school
overall grade point average carries somewhat more weight
than academic grade point average. This can be seen by
examining Tables 43 with 47 and 44 with 48. In the first
two, SAT-Verbal and SAT-Math are used with academic grade
point average and overall grade point average respectively
to predict first semester grade point average. There is

a raw weight of .0035 in Table 43 for academic grade point

136

average using all Provisional students; the comparable
value in Table 47 is .0048. Tables 44 and 48 show predic-
tion equations for total college grade point average.
Table 44 shows a weight of .0041 for academic average, but
Table 48 shows .0052 as the weight for overall average.
These same kinds of differences appear in an examination
of Tables 45 with 49 and 46 with 50.

It is evident from the above that when academic
and overall high school grade point averages are used
with the same measures of SAT score to predict college
grades, overall average is slightly more influential than

academic average in the prediction.

C H A P T E R V
SUMMARY AND CONCLUSIONS

SUMMARY

Higher education in the United States has become
more accessible to students over the past decade. The
development of community colleges, vocational-technical
schools, and the expansion of four-year colleges and.
universities have helped to make it so. Since students now
have more opportunities in post secondary education from
which to choose, it is incumbent upon educators to help
them identify and select the most appropriate option.

Part of this decision making process for many young
people relates to the selection of a college or university.
In turn, the institutions across the country are faced with
decisions which involve the selection of students for A
admission. The task of helping these two parties-—students
and universities--identify the most propitious match for
each, has been a most difficult one in the past. This,
of course, is in large measure due to a lack of predict-
ability in human behavior. It also, unfortunately, is due

137

Ar

Un:

su}

st

Pr

mc

81

Cu

138

to the dearth of adequate measurement techniques now being
used to predict success in college.

The goal of this study was to add to the body of
knowledge relating to the admission of students to college.
A major strength of its focus was the relevance for The
University of Texas at Austin. The data presented added
substantially to knowledge about the success rate of.
students admitted to the University under its special
Provisional Admission Program.

The data relate to the degree of success for
marginal students admitted under this program to what is a
moderately selective state University. The quality of the
students' preparation for college, and the quality of
their academic work in college, was measured and compared
to that of normally admissable students.

From the analysis came a better understanding of
several factors that contribute to students' success or
failure. Thus, a second strength of the study was in the
breadth of its analysis. The students' performance was
measured according to a variety of criteria. This enabled
the drawing of conclusions from several sets of data, each
providing a somewhat different type of analysis.

Specifically, the students' high school records and
entrance test scores were compared to detect any observable

differences, particularly between the Provisional and

’t'1

139

regular student samples. Sex and ethnic background, as well
as the semester of first enrollment, were used as independent
variables in these comparisons.

Similarly, the college performance was analyzed for
each of the sample groups, again using sex, mode of admis-
sion, first semester of enrollment, and ethnic background
as independent variables. The use of t~tests, three—way
analysis of variance, and post—hoc comparisons served to
provide the necessary data in this regard.

In addition to the college performance measured by
first semester and total grade point averages, the study
also examined the relative persistence in enrollment for
the various groups under consideration. Again, sex, ethnic
background, first semester of enrollment, and mode of
admission were the independent variables used to identify
these groups. The chi—square test of significance was used
to detect any observable differences among the groups.

Prediction of performance was one of the other key
factors of the study. Various types of predictors were
used, singly and in combination, and their relative values
were studied. The goal was to reach some conclusion as to

the most effective means of identifying those students who

"flew—a __ N...
am Hmfluww‘h ..' "n, ..vh~ - ..- -I-‘ __,,.-_.-.— ’"-

are likely to succeed in  the_ future.
nun—M ”M1”, ‘__ ‘_,...

The most notable findings of these analyses can

be summarized as follows:

 

In

140

(l) The student composition with regard to sex and
ethnic background was about the same for the samples under
study as for the University as a whole. There was some
variation in the proportion of men to women noted in the
samples compared to the general student pOpulation, but this
was not seen to limit the value of the findings. The ethnic
background of the samples very closely approximated that of
the general student body in both the regular and Provisional
groups.

(2) Analysis of the high school records, entrance
test scores, and college grade point averages revealed that
students in the regular admit group performed significantly
better in all respects than did Provisional students. (The
only exception was for the variable, "hours earned—~first
semester" which was not as useful a measure of performance
as the others because Provisional students were required to
enroll for at least 12 semester hours as part of the pro-
visions of the program; no such requirement applied to
regular students.) This proved to be true when the respec-
tive means were compared; it also was evident when the
relative distributions of the scores were compared for each
group.

In addition, males in the regular group performed
better on all measures than did male Provisional students;

the female regular students, in turn, did better than female

141

Provisionals on all measures of ability and performance.

The same proved true when white and Mexican-American regular
students were compared to their respective ethnic groups in
the Provisional sample.

(3) No significant difference was noted between the
college performance of male Provisional students and that of
female Provisionals. However, there were significant
differences found in the quality of the entrance records for
male and female Provisionals. Males tended to present
higher SAT scores than females; females tended to have a
higher rank in the graduating class than did males. These
differences apparently evened out with respect to their
effect on students' performance in college.

The same phenomenon occurred in the regular sample.
Males and females achieved the same grades in college
although presenting different levels of high school per—
formance and SAT scores (males had higher SAT scores and
females a higher rank in class).

(4) White students presented superior scores on the
SAT-Math and SAT—Verbal to those of Mexican-American stu-
dents within both the regular and Provisional samples.
Mexican-Americans achieved a higher class rank in high
school than did white students within both samples. The
first semester college grade point average for white students

was significantly higher. When grades in advanced standing

142

examinations were not counted, there was no significant
difference in total grade point average, but when they were
counted, Mexican—American Provisional students had signifi-
cantly higher college grade point averages than white
Provisionals. In the regular sample, white students
achieved higher grades in college with or without advanced
standing grades included.

(5) Both male anui female Provisional students
performed significantly better in English courses than they
did in science/math or social science courses. Although one
might expect that this would be the case since the students
may have had most familiarity with the English curriculum
when entering college, the low grades evidenced in social
science are of particular interest. Since these courses
could not be taken during the period in which the students
were on Provisional status, only those students who success-
fully completed the first semester actually enrolled in the
social science curriculum. That is, these courses were
taken only by students removed from Provisional status. One
might expect these more successful Provisional students to
cause the mean grade point average in social science courses
to be higher than in the other curricular areas, but this
did not prove to be the case.

(6) The analysis of variance produced evidence of

interaction effects operating when sex, mode of admission,

Vd

de

pf

65

143

and first semester of enrollment were used as sources of
variance. These factors proved to be related to the stu-
dents' high school record, entrance test scores, and college
performance.

(7) High school rank, SAT-Verbal, and SAT-Math were
essentially the same for the groups of Provisional students
entering during each of the four semesters in question. The
same was the case for the regular student sample. When both
Spring Provisional groups were combined and compared to both
Summer Provisional groups, no significant differences were
found in the quality of the high school record and SAT
scores. The same was true for regular students.

(8) Summer Provisional students achieved signifi-
cantly higher first semester grade point averages than did
the Spring Provisional students when the Summer 1972 and
1973 groups were pooled and compared to the combined Spring
1972 and 1973 groups.

This did not prove to be the case for the regular
students. Entering during the Spring or Summer for these
students did not affect their first semester grade point
average.

(9) The total grade point averages of the pooled
Summer and Spring Provisional groups showed some variance in
the performance level, although the differences were not

statistically significant. The tendency for Summer

In

144

Provisionals to do better than Spring Provisionals the first
semester was thus somewhat abated when total grade point
average was considered. However, the differences in per-
formance on this measure should not be discounted since they
were comparable in magnitude to those noted for first
semester grade point average.

Again, for total grade point average, regular stu-
dents did not show the same tendency as Provisional students.
There appeared no significant difference in the total grade
point average of Summer regular students when compared to
that of Spring regular students.

(10) The Summer 1973 Provisional group achieved a
somewhat higher first semester grade point average than did
the Summer 1972 group, but this difference was not signifi-
cant with a = .05. The difference was large enough that it
is of some interest, however. Since the Summer 1973 group
had essentially the same quality of high school record and
SAT scores as the 1972 group, the 1973 students' superior
college grade point average may have been due in part to
experiences the students encountered after enrollment at
the University.

(11) With respect to the students' ability to remain
in college, there appeared no significant difference between
males and females in the Provisional sample. Both tended

to persist to about the same degree. The same was true when

In

In

145

white and Mexican—American students were compared in the
Provisional group.

Regular students evidenced the same tendency. Sex
and ethnic background did not appear to affect the students'
ability to persist in college.

(12) Sixty-six percent of the regular students were
currently enrolled at the time of the study, compared with
59% of the Provisional students. This difference proved to
be statistically significant at the a = .05 confidence level.

When male Provisionals were compared with male
regulars, there was not a significant difference with
respect to persistence. The same was true for females from
each group. This finding coupled with the statistically
significant but still small difference in persistence for the
regular and Provisional students taken as a whole, was in
sharp contrast with the evidence presented on college grade
point averages. The highly significant differences found on
those measures were not found when the measure of performance
was that of ability to remain in school.

(13) For both the Provisional and regular samples,
high school record and SAT scores showed relatively low
correlations with college grade point averages. Some of
this phenomenon may be attributable to the restriction in

the range of high school rank and SAT scores for each sample.

146

The Provisional students' high school rank and
academic and overall grade point averages showed small but
positive correlations with college grades. SAT—Math, SAT-
Verbal, and SAT-Total appeared to have no relationship at
all with college grade point averages for Provisional
students. For regular students, the relationship between
SAT scores and college grades was about the same as that
for high school rank and college grades; a small to moderate
but positive correlation was found in each case.

The correlation between high school rank and
college grades was thus higher by a small amount for regular
students than for Provisional students.

(14) For Provisional males the correlations for
high school rank and grade point average with college
grades were higher than those for Provisional females. The
SAT-Verbal correlated better with college grades for females
than for males in the Provisional group. The SAT-Math and
SAT-Total produced about equal correlations with college
grades for male and female Provisionals.

In the regular sample, the correlations for high
school rank and SAT scores with college grade point averages
were higher for females in all cases.

(15) Regular students were found to be somewhat

more predictable than Provisional students when multiple

147

correlations were developed for high school and test data
with college grade point average. The multiple r values for
the regular sample were consistently higher than for the
Provisional sample.

(16) Male Provisional students were about as
predictable as female Provisionals, having produced similar
multiple r values in the correlation study. The same was
true in comparing males and females in the regular student
sample.

(17) It was possible to predict total college
grade point average more accurately than first semester
grade point average in the Provisional sample. The
opposite phenomenon occurred when predictions for the
regular sample were attempted. In that instance, it was
possible to predict first semester grade point average more
accurately, though the differences were more noteworthy for
the consistency than for the amount.

(18) SAT-Verbal was considerably less potent in in-
fluencing male Provisional college grades than it was for
female Provisional students. SAT-Math, on the other hand,
was weighted more for males than females in the Provisional
group. For regular students, the indication was that the
math and verbal scores should be weighted equally for both

males and females.

148

(19) Rank in class was equally important for male
and female Provisional students in affecting first semester
and overall grade point averages. The same was true for
regular males and females.

(20) High school rank was about as effective as
high school academic and overall grade point averages in
predicting the performance in college of Provisional stu—
dents. Overall grade point average in high school was
slightly but consistently more effective in such predictions

than was academic grade point average.

CONCLUSIONS

 

It has been shown that students admitted under the
Provisional Admission Program at The University of Texas at
Austin do not achieve grades equal to those of regularly
admitted students. The grades they do obtain are, in many
cases, below those required for normal progress at the
University.

It has also been shown, however, that there are
many Provisional students who do remain enrolled at the
University. And, although regular students stay in school
in greater percentages, the persistence of Provisional
students is somewhat encouraging. Many students who would

otherwise not have been able to attend the University as

149

freshmen find the Provisional program to be a successful
avenue of access.

It is, however, important to make several related
points. First, although many of the Provisional students
remain at the University, it is clear from the research that
they are often only barely achieving acceptable grades. It
can be presumed that this level of performance requires
considerable effort and motivation on their part. Further-
more, such stndents could probably continue to benefit from
supportive services such as tutoring, reading skills, and
counseling beyond the first semester of enrollment. The
importance of continued monitoring of Provisional student
performance and the maintenance of support services should
thus be stressed. The mere fact that many of the students
persist at the University should not negate their needs for
support and encouragement.

In addition, since many of the Provisional students
do not succeed at the University, it is important that
counselors be available to discuss with the students their
educational and vocational alternatives after experiencing
this initial failure. This could also apply to some of the
students who remain at the University, but who are only
barely in good standing. It is quite possible that attending

another institution could prove beneficial to such students

150

in that success there would not necessitate the total com-
mitment of effort and time that success at this University
often requires.

This study has reaffirmed the fact that prediction
of academic success is not an easy matter. Depending on
the specific subgroup under study here, the correlations
between prior performance in school or on the SAT and
college performance were from zero to moderate. In no
instance were the relationships shown to be strongly
positive. There were clearly some non-academic variables in
operation that were not specifically identified. All this
leads to the conclusion alluded to above--our lack of
success in predicting academic performance can be tied to
the unpredictability of human behavior and the dearth of
useful tools.

The study described here has shown that the tools
we presently have should be used with different emphasis for
various types of students. Men, for example, behave
differently from women, and should be treated somewhat
differently from them in the process of predicting academic
performance. Similarly, it is apparent that Provisional
(marginal) students' college performance is related to
different factors in their academic background than is the
performance of regular students. The evidence also indicates

that our present predictors of college performance should

151

only be used as guides. Basing inflexible policy decisions
on them may be convenient, but hardly justifiable for
educational reasons. Also, in trying to assess an individ-
ual student's chances of success, our measurement techniques
must be used with a considerable degree of care; for,
although our usual predictive measures may not indicate it,
many students like those admitted under the Provisional
program can succeed at a moderately selective University.
They do not, as a rule, achieve higher than average grades.
Many leave the University after a short time. But many
others do succeed to the extent of staying in school for

an extended period.

If the University, as an institutional policy,
perceives the Provisional program as a contract with the
student, then it has an obligation to fully inform the
student of the requirements of the program. It is also
important that the student appreciate the implications of
the evidence that relates to the past level of success of
students in the program. Further, the University has an
obligation to provide support and encouragement to the
Provisional students.

If these responsibilities are met, the level of
success shown by the students in this study implies that
continuation of the program is warranted. However, as al-

ready stated, the relatively lower level of performance,

152

and somewhat higher tendency to drop—out, imply that
specific attention must be given to the many Provisional
students who encounter academic difficulty. Again, the
University assumes a responsibility for all the Provisional
students due to the act of granting their admission in

the first place. Such responsibility is best expressed in‘
terms of support services for the students to facilitate
their academic success, or to help them find satisfactory

alternatives to attending the University.

IMPLICATIONS FOR RESEARCH

The most natural follow-up to this study would be
an attitudinal survey of Provisional and regular students
to ascertain the relationship of such factors as levels of
Inotivation and biographical data to students' success at
tflie University. Although sex and ethnic background were
Slflbject to extensive analysis in the present study, other
factors such as family income, size of high school class,
arua occupational goals might prove to be interesting areas
tC> study. Also, the student's attitudes toward school, their
need to achieve, and other attitudinal factors may prove
Iffilevant to academic success at the University. Of interest
also would be any variations detected between the impact of
Sucni factors on the grades of Provisional students from their

iHHPact on regular student grades. The fact that academic

153

factors in the prediction models studied here had varying
effects on regular and Provisional grades, would lead one
to think that non-intellective factors may show the same
variations. \

The degree to which these non-intellective factors
could be used in the admission process, particularly in the
decision of whether to admit or not, is problemmatic. Such
information would help, however, to the extent that the
relevant items are identified in the Provisional student
samples. Knowing that a particular group possesses a
certain trait that has historically been a strong indication
of academic difficulty, could help the faculty and staff in
anticipating potential problems. Program modifications,
especially with respect to support services, could then
be implemented to respond to the anticipated needs of a
particular group.

Another major area of concern that future research
could address would be a study of several types of supportive
programs and their impact on Provisional student performance.
The degree to which tutoring, counseling, reading skills
courses, and other programs can help such students is not
presently known for the particular program at The University
of Texas at Austin. If the University is to assume the
responsibility to which this paper has already alluded, then

it is important to research the impact of the various kinds

154

of support services offered. Such research could result in
policy decisions to implement the most effective programs
to maximize the success of Provisional students.

A third recommended focus for future research is
to further examine the apparent differences in performance
between Spring and Summer Provisional students. It would
be very instructive to find the factors that operate to
cause the Spring entrants to perform more poorly than the
Summer students. With such information, some programmatic
efforts could be made to improve the performance of the
Spring students.

Finally, it would be very helpful to ascertain the
attitudes of Provisional students toward the program itself.
Presumably, some differences would be found between success-
ful and unsuccessful students. More importantly, it would
be good to learn about the students' reactions in general,
and to solicit their suggestions for making improvements
in the program. Having experienced the pressure and
frustration along with the satisfaction and pride, the
students would be able to provide valuable insight into the
relative merits of the Provisional Admission Program. Per-
haps that is a good place to stop; it is important to focus
directly on the reactions of the young people who subject

themselves and who are subjected to the rigors of the

155

University. Their thoughts and feelings should be a central
part of our assessment of programs such as Provisional

admission at The University of Texas at Austin.

FOOTNOTES

FOOTNOTES

1Charles T. Thomas and Julian C. Stanley, "Effect-
iveness of High School Grades for Predicting College Grades

 

 

of Black Students: A Review and Discussion." Journal of
Educational Measurement, Vol. 6, No. 2 (Winter, 1969),
p. 1.

2The Chronicle of Higher Education, Vol. VIII,
No. 2 (October, 1973), p. 1.

31bid., p. 6.

4Alexander W. Astin, "Open Admissions and Programs
for the Disadvantaged," Journal of Higher Education, Vol.
XLI, No. 8 (November, 1971), p. 632.

I, 5Krista Ludenia, "Academic Prediction and the
Marginal Student," (Missouri University, May, 1970)
ED 045-040. ,

.z’r‘

 

2* David E. Lavin, The Prediction of Academic Per-
formance (New York, New York: Russell Sage Foundation,
1965).

 

Julian C. Stanley, "Predicting College Success of
the Educationally Disadvantaged," Science, Vol. 171, No.
3972 (February 19, 1972). pp. 640-647.

Benjamin S. Bloom and Frank R. Peters, The Use
of Academic Prediction Scales (The Free Press of Glencoe,
Inc., 1961).

 

Jane Loeb, "High School Performance as Predictive
of College Performance," NASSP Bulletin, Vol. 56, No. 365
(September, 1972), pp. 19-26.

 

Thomas and Stanley, op. cit., pp. 203-215.

John Paraskevopoulos and L. F. Robinson, "Compari-
son of Regression Equations Predicting College Performance
from High School Record and Admissions Test Scores for
Males and Females," College and University, Vol. 45, No.

2 (Winter, 1970). PP. 211-216.

 

156

157

6William H. Huber, "Channeling Students for Greater
Retention," College and University, Vol. 47 (Fall, 1971),
p. 19.

 

7Lavin, gp. cit., pp. 109-132.

8Robert A. Scott, "Challenges Presented by Open
Enrollment," Paper presented to American Association of
Collegiate Registrars and Admissions Officers (October 3,
1970), ED 044-745.

9Virginia S. Hardie and John R. Anderson, "College
Students--A Revealing Comprehensive Seven-Year Study of
1106 University Freshmen Attrition, Graduation, and Follow-
up," (Clemson University, August, 1969), ED 055-300

10Carter V. Good, Dictionary of Education (New York:
McGraw-Hill Book Company, Inc., 1945), p. 3.

 

11Definitions of Student Personnel Terms in Higher
Education, U. S. Department of Health, Education and Welfare
(Washington: U. S. Government Printing Office, 1968),
p. 15.

 

 

lzIbid.
l3Good, gp. cit., p. 433.
l4Ibid., p. 53.

15Definitions of Student Personnel Terms in Higher
Education, op. cit., p. 49.

 

 

16ipig., p. 3.

l7Ludenia, igg. git.

Lavin, igg. git.

Stanley, igg. git.

Bloom and Peters, igg. git.
Loeb, igg. git.

Thomas and Stanley, igg git.

Paraskevopoulos and Robinson, loc. cit.

18Loeb, gp. cit., p. 21.

158

19Charles Gelso and Joseph A. Klock, "A Comparison
of the Effectiveness of Basing Predictive Average Grades
on Three Different Methods of Computing High School
Averages," College and University, Vol. 42, No. 3 (Spring,
1967), p. 344.

 

20John R. Hills and Joseph A. Klock, "Total vs.
Academic High School Averages in College Grade Prediction,"
College and University, Vol. 41, No. 2 (Winter, 1966), pp.
231-232.

 

21Richard D. Arvey, Intra Individual Variability as
a Moderator of College Predictability, Paper read at Mid—
Western Psychological Association (Chicago, May, 1969).

 

 

22Richard Reyes and Robert B. Clarke, "Consistency
as a Factor in Predicting Grades," Personnel and Guidance
Journal (September, 1968), pp. 50-55.

 

23Ibid., p. 54.
24Loeb, loc. cit.

25Charles F. Elton, "The High School Grade Average:
Where is a Difference Different?" College and University,
Vol. 41, No. 2 (Winter, 1966) p. 185.

26

 

Ibid., p. 189.

27Loeb, loc. cit.

 

28Paul Horst, "Differential Prediction in College
Admissions," College Board Review, Vol. 33(1957), pp. 19-23.

 

29Lavin, gp. cit., pp. 53-54.

30"Highlights of the ACT Technical Report," American
College Testing Program (Iowa City, 1973), p. 18.

311bid., p. 19.

32William H. Angott, The College Board Admissions
Testing Program: A Technical Report on Research and Develop-
ment Activities Relating to the Scholastic Aptitude Test
and Achievement Tests, College Entrance Examination Board
(New York, 1971).

 

 

 

33Ibid., p. 129.

159

34Ibid., p. 217.

35Alexander Astin, PredictingiAcademic Performance
in Collegg, (New York: American Council on Education, The
Free Press, 1971), p. 11.

 

 

36Leo Munday, "Correlations Between ACT and Other
Predictors of Academic Success in College," College and
University, Vol. 44 (Fall, 1968), p. 74.

 

 

37Stanley, gp. cit., p. 641.

38College Testing--A Guide to Practices and Programs,
(Washington, D.C.: American Council on Education, 1959).

 

39Arvo Juola, "Freshmen Level Ability Tests Versus
Cumulative Grades in the Prediction of Successive Terms
Performance in College," Paper presented to the Annual
Meeting of the American Educational Record Association
(1964).

40Julian C. Stanley and Andrew C. Porter, "Correla-
tions of Scholastic Aptitude Test Scores with College Grades
for Negroes versus Whites," Journal of Educational Measure-
ment, V01. 4, No. 4 (Winter, 1967), p. 200.

 

41Thomas and Stanley, gp. cit., p. 213.

42T. Anne Cleary, "Test Bias: Prediction of Grades
of Negro and White Students in Integrated Colleges,"
Journal of Educational Measurement, Vol. 6, No. 4 (Winter,
1969), p. 213.

 

43"Highlights of the Act Technical Report," loc. cit.

 

44Lavin, 10c. cit.

 

45ipig., p. 101.
46igig., p. 109.
47ipig., p. 111.
48igig., p. 126.
49igig., p. 129.
soigig., p. 132.
Slipig., p. 133.

160

52James W. Creaser, "Predicting College Success
from Equated High School Ranks: A Cross—Validated Study,"
College and University, Vol. 41, No. 1 (Fall, 1965), p. 97.

53D. P. Hoyt, "Size of High School and College
Grades," Personnel and Guidance Journal, Vol. 37 (April,
1959), p. 57.

 

54Astin, "Open Admissions and Programs for the Dis-
advantaged," p. 634.

55Astin, Predicting Academic Performance in College.

 

56Huber, gp. cit., p. 21.
57Scott, loc. cit.

S8Astin, "Open Admissions and Programs for the Dis-
advantaged," p. 630.

59Official Enrollment Report, (Austin, Texas: Office
of Institutional Studies, The University of Texas at Austin,
September 30, 1973).

 

60Data obtained from the Office of Institutional
Studies, The University of Texas at Austin, Fall, 1973.
(of total enrollment, 1% Black, 5% Mexican-American, 90%
White, 4% foreign students and "other".)

61Lincoln L. Chao, Statistics: Methods and Analyses
(New York: McGraw-Hill, Inc., 1969), p. 260.

 

6zIbid., pp. 261-262.

 

BIBLIOGRAPHY

 

1.

2.

3.

4.

5.

6.

7.

8.

9.

BIBLIOGRAPHY

Angott, William H. The College Board Admissions Testing
Program: A Technical Report on Research and Develop-
ment Activities Relating to the Scholastic Aptitude
and Achievement Tests. College Entrance Examination
Board, New York, 1971.

 

 

 

 

Arvey, Richard D. "Intra-Individual Variability as a
Moderator of College Predictability," Paper read
at Midwestern Psychological Association, Chicago,
May, 1969.

Astin, Alexander W. "Open Admission and Programs for .//
the Disadvantaged." Journal of Higher Education, XLII,
No. 8, (November, 1971), 629-647.

 

Astin, Alexander W. "Predicting Academic Performance in
College." American Council on Education. New York: C//“
The Free Press, 1971.

 

Astin, Alexander W. "Student Persistence: Some Stay, 9,/
Some Don't--Why?" American Association of Collegiate
Registrars and Admissions Officers 59th Annual
Meeting, Los Angeles, April 19, 1973.

Birnbaum, Robert. "Class Rank Weighting." College and
University. 41, No. 1 (Fall, 1965), 101-106.

 

 

Blanchfield, W. C. "College Dropout Identification--A
Case Study." Utica College, ED 058-605; 1971.

Bloom, Benjamin S. and Peters, Frank R. The Usesgof
Academic Prediction Scales, The Free Press of Glencoe,
Inc., 1961.

 

 

Bogue, E. G. "Application of a Minimum Loss Decision
Strategy in the Selection of Cutoff Points in College
and University Admissions." College and University,
43, No. 2 (Winter, 1968) 131-142.

161

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

162

Boreys, Jerry H. "The Validity of American College Test
for Discerning Potential Academic Achievement Levels--
Ethnic and Sex Groups." The Journal of Educational
Research, 65 (September, 1971), 3-6.

 

 

Brown, William F. and Holtzman, Wayne H. "A Study--
Attitude Questionnaire for Predicting Academic I,/
Success." Journal of Educational Psychology, 46, I.
No. 2 (February, 1955), 77-84.

 

Chao, Lincoln L. Statistics: Methods and Anaiysis.
New York: McGraw-Hill, Inc., 1969.

 

Chronicle of Higher Education, The. Vol. VIII, No. 2,

 

(October, 1973).

Cleary, Anne T. "Test Bias: Prediction of Grades of
Negro and White Students in Integrated Colleges."
Journal of Educational Measurement, 5, No. 2
TSummer, 1968), 115-124.

 

College Testing--A Guide to Practices and Programs.
American Council on Education, Washington, D.C. 1959.

 

Creaser, James W. "Predicting College Success From
Equated High School Ranks: A Cross Validation Study."
College and University, 41, No. 1 (Fall, 1965), 4-9.

 

Davis, J. A. and Temp, G. "Is the SAT Biased Against
Black Students?" College Board Review, No. 81,
(Fall, 1971), 4-9.

 

Davis, Richard H. "Compensating Education--A Calculated
Cop-Out." An Address to the 1970 Convention of
Wisconsin Association of Collegiate Registrar's and
Admissions Officers, ED 047-060, October 16, 1970.

Definitions of Student Personnel Terms in Higher Educa-
tion. U. S. Department of Health, Education, and
Welfare, U. S. Government Printing Office, Washington,
1968.

 

 

Dickason, Donald G. "Predicting the Success of Freshmen
Engineers." Personnel and Guidance Journal, 47,
No. 10, (June, 1969).

 

Donnan, Hugh. "Personality Factors Related to College
Achievement and Attrition." Journal of College
Student Personnel, 9, No. 2, (March, 1968), 116-119.

 

 

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

163

Elton, Charles F. "The High School Grade Average: When
is a Difference Different?" College and University,
41, No. 2, (Winter, 1966), 185-189.

 

Ferrin, Richard 1., Jones, Richard W., and Tremble,
Cesar M. Access to College for Mexican-Americans in
the Southwest, Higher Education Surveys Report, No.
6, College Entrance Examination Board, July 1972.

 

 

Ferrin, Richard I. "Barriers to Universal Higher Educa-
tion." College Entrance Examination Board--Access
Research Office, Palo Alto, California, March, 1970.

 

 

Gelso, Charles and Klock, Joseph. "A Comparison of the
Effectiveness of Basing Predictive Average Grades on
Three Different Methods of Computing High School
Averages." College and University, 42, No. 3 (Spring,
1967), 337-344.

 

Gelso, Charles and Klock, Joseph. "The 'Academic' vs.
'Total' High School Average: A Re-Examination."
Journal of Educational Measurement, 4, No. 2 (Summer,
1967), 59-63.

 

Glass, Gene V. and Stanley, Julian C. Statistical
Methods in Education and Psychology. New Jersey:
Prentice-Hall, Inc., 1970.

 

 

Good, Carter V. Dictionary of Education. New York:
McGraw-Hill, Inc., 1945.

 

Green, Robert L. "Money, Motivation, and Academic
Achievement." Phi Delta Kappan, L, No. 4 (December,
1968), 228-230.

 

Green, Robert L. "The Black Quest for Higher Education:
An Admissions Dilemma." Personnel and Guidance
Journal, 47, No. 9, (May, 1969), 905-911.

 

Gunne, Michael G. and Leslie, Larry L. "Exceptional
Graduate Admissions at the Pennsylvania State
University." Center for the Study of Higher
Education, ED-061-889, March, 1972.

Hardie, Virginia S. and Anderson, John R. "College
Students--A Revealing Comprehensive Seven-Year Study
of 1106 University Freshmen--Attrition, Graduation,
and Follow-Up." Clemson University, ED-055-300,
August, 1969.

33.

34.

35.

36.

37.

38.

39.

40.

‘41.

42.

43.

164

Hershey, David E. "An Analysis of the Academic
Achievement and Progress of Students Admitted to
Michigan State University Through the Summer Test
Admit Program 1965—1969." Unpublished dissertation,
Michigan State University, 1970.

Highlights of the ACT Technical Report. American
College Testing Program, Iowa City, 1973.

 

Hills, John R. and Gladney, Marilyn B. "Predicting
Grades from Below Chance Test Scores." Journal
of Education Measurement, 5, No. 1, (Spring, 1968),
45-54.

 

Hills, John R. and Klock, Joseph. "'Total' vs.
'Academic' High School Averages in College Grade

Prediction." College and University, 41, No. 2
(Winter, 1966), 231-232.

 

Hilton, Thomas L. and Meyers, Albert E. "Personal
Background, Experience, and School Achievement:
An Investigation of the Contribution of Questionnaire
Data to Academic Prediction." Journal of Educational
Measurement, 4, No. 2 (Summer, 1967).

 

Horst, Paul. "Differential Prediction in College
Admissions." College Board Review, 33 (1957),
19-23.

 

Hoyt, Donald P. "Generalized Academic Prediction in
Four Year Colleges." Personnel and Guidance
Journal, (October, 1968).

 

Hoyt, Donald P. "Size of High School and College
Grades." Personnel and Guidance Journal, 37
(April, 1959) 549-573.

 

Huber, William. "Channeling Students for Greater
Retention." College and University, 47 (Fall, 1971)
19-29.

 

Hyman, Seymour. "The Open Admissions Freshmen at
City University of New York." Opgn Admissions:
The Pros and Cons--A Symposium. Council for
Basic Education, Washington, D. C., 1971.

 

 

Ivey, Allen E., Peterson, Floyd E., and Trebbe, E.
Stewart. "The Personality Record as a Prediction
of College Attrition: A Discriminant Analysis."
College and University, 41, No. 2 (Winter, 1966),
199-205.

 

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

165

Juola, Arvo. "Freshmen Level Ability Tests versus
Cumulative Grades in the Prediction of Successive
Terms Performance in College." Paper presented
to Annual Meeting of American Educational Research
Association, 1964.

Kristol, Irving. "Open Admissions and the Politics
of the Absurd." Open Admissions: The Pros and
Cons--A Symposium. Council for Basic Education,
Washington, D. C., 1971.

 

 

Lavin, David E. The Prediction of Academic Success.
Russell Sage Foundation, New York, 1965.

 

Loeb, Jane. "High School Performance as Predictive
of College Performance." NASSP Bulletin, 56,
No. 365 (September, 1972) 19-26.

 

Ludenia, Krista. "Academic Prediction and the Marginal
Student." Missouri University, Ed-045-040,
May, 1970.

Mayhew, Lewis B. "Jottings." Change Magazine, 5,
No. 1 (February, 1973) 62-64.

 

Morgan, Lewis B. "The 'Calculated Risks' - A Study
of Success." College and University, 43, No. 2
(Winter, 1968), 203-206.

 

Munday, Leo. "Correlations Between ACT and Other
Predictors of Academic Success in College."
College and University, 44 (Fall, 1968) 67-76.

 

Official Enrollment Report, Fall 1973. Office of
Institutional Studies, The University of Texas
at Austin, September 30, 1973.

Paraskevopoulos, John and Robinson, L. F. "Comparison
of Regression Equation Predicting College Per-
formance from High School Record and Admissions
Test Scores for Males and Females." College and
University, 45, No. 2 (Winter, 1970) 211-216.

 

 

Pesqueira, Richard E. "Mexican-American Student
(Staying) Power in College." College Board Review,

 

55.

56.

57.

58.

59.

60.

61.

62.

63.

64.

166

 

 

Reyes, Richard and Clarke, Robert B. "Consistency
as a Factor in Predicting Grades." Personnel and
Guidance Journal, 47, No. 1 (September, 1968)
50-55.

Richards, James M. and Lutz, Sandra W. "Predicting
Student Accomplishment in College From the ACT
Assessment." Journal of Education Measurement, 5,

 

No. 1 (Spring, 1968) 17-31.

Scott, Robert A. "Challenges Presented by Open
Enrollment." Paper presented to National Conference
of American Association of Collegiate Registrars
and Admissions Officers, October 3, 1970, ED-044-745.

Sedlacek, W. E., Brooks, G. C. Jr., and Mindus, L. A.
"Black and Other Minority Admission to Large
Universities: Three Year National Trends."
Research Report No. 3-72, Cultural Study Center,
University of Maryland, 1972.

Seyfort, Warren C. "The Facts in the Case--What
Colleges Do and Think About Rank-In—Class." NASSP
Bulletin, 56, No. 365 (September, 1972) 27-40.

 

Smith, Mortimer. "Varieties of Open Admissions."
Open Admissions: The Pros and Cons--A Symposium.
Council for Basic Education, Washington, D. C.,
1971.

 

Sommerville, Bill. "Can Selective Colleges Accommodate
the Disadvantaged? Berkley Says 'Yes'." College
Board Review, No. 65 (Fall, 1967) 5-10.

 

Stanley, Julian C. and Porter, Andrew C. "Correlation
of Scholastic Aptitude Test Score with College
Grades for Negroes Versus Whites.“ Journal of
Education Measurement, 4, No. 4 (Winter, 1967)
199-218.

 

 

Stanley, Julian C. "Predicting College Success of
the Educationally Disadvantaged." Science, 171,
No. 3972 (Feburary 19, 1971) 640-647.

Temp, George. "Test Bias: Validity of the SAT for
Blacks and Whites in Thirteen Integrated Institutions."
College Board Research Bulletin. RB-7l-2,

January, 1971.

65.

66.

67.

68.

69.

70.

167

Thomas, Charles L. and Stanley, Julian C. "Effective-
ness of High School Grades for Predicting College
Grades of Black Students: A Review and Discussion."
Journal of Educational Measurement, 6, No. 4
(Winter, 1969) 203-215.

 

Weiss, Kenneth P. "A Multi-Factor Admission Pre-
dictive System." College and University, 45,
No. 2 (Winter, 1970) 203-210.

 

Whitney, Douglas R. and Boyd, Norlin. "Limiting
Effect of Predictive Validity on the Expected
Accuracy of Admissions Decisions." College and
University, 46, No. 3 (Spring, 1971) 180-190.

 

 

Williams, Robert L. "What Are We Learning From
Current Programs for Disadvantaged Students?"
Journal of Higher Education, 39, No. 4 (April,
1968) 274-285.

 

Willingham, Warren W. Free Access Higher Education.
New York: College Entrance Examination Board,
1970.

 

Witzke, Judith. "A Review of Selected Research
Related to College Admissions Criteria for Educa-
tionally Disadvantaged Students." Research Bulletin
of the Measurement and Evaluation Center, The
University of Texas at Austin, December 23, 1971.

IIIIIIIIIIIhi“