INFORMATION TO USERS

This dissertation was produced from a m icrofilm copy o f the original document.
While the most advanced technological means to photograph and reproduce this
document have been used, the quality is heavily dependent upon the quality of
the original submitted.
The following explanation o f techniques is provided to help you understand
markings or patterns which may appear on this reproduction.
1.

The sign or "target" fo r pages apparently lacking from the document
photographed is "Missing Page(s)". If it was possible to obtain the
missing page(s) or section, they are spliced into the film along w ith
adjacent pages. This may have necessitated cutting thru an image and
duplicating adjacent pages to insure you complete continuity.

2.

When an image on the film is obliterated w ith a large round black
mark, it is an indication th a t the photographer suspected that the
copy

may have moved during exposure and thus cause a blurred

image. You will find a good image of the page in the adjacent frame.
3.

When a map, drawing or chart, etc., was part of the material being
p h o to g rap h ed

the

photographer

followed

a definite

method

in

"sectioning" the material. It is customary to begin photoing at the
upper left hand corner of a large sheet and to continue photoing from
left to

right

in equal sections w ith a small overlap. If necessary,

sectioning is continued again — beginning below the first row and
continuing on until complete.
4.

The m ajority o f users indicate that the textual content is o f greatest
value,

however, a somewhat higher quality reproduction could be

made from
dissertation.

"photographs" if essential to the understanding of the
Silver

prints

of

"photographs"

may

be

ordered

at

additional charge by writing the Order Department, giving the catalog
number, title, author and specific pages you wish reproduced.

University Microfilms
300 North Zeeb Road
Ann Arbor, Michigan 48106
A Xerox Education Company

72-30,050
SNOWDEN, Raymond, 1940A CRITICAL ANALYSIS AND APPRAISAL OF THE
MARTIN LUTHER KING DEVELOPMENTAL PROGRAM
FOR MARGINAL STUDENTS AT WESTERN MICHIGAN
UNIVERSITY.
Michigan State University, Ph.D., 1972
Education, higher

University Microfilms, A XEROX Com pany, Ann Arbor, Michigan

© C o p y r i g h t by
RAYMOND SNOWDEN
1972

A CRITICAL ANALYSIS AND APPRAISAL OF THE MARTIN LUTHER KING
DEVELOPMENTAL PROGRAM FOR MARGINAL STUDENTS AT
WESTERN MICHIGAN UNIVERSITY
By
Raymond Snowden

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
College of Education
19 72

P L E A S E NOTE:
Some pages may have
i n d i s t i n c t print.
F i l m e d as r ec eive d.
University Microfilms, A Xerox Education Company

ABSTRACT

A CRITICAL ANALYSIS AND APPRAISAL OF THE MARTIN LUTHER KING
DEVELOPMENTAL PROGRAM FOR MARGINAL STUDENTS AT
WESTERN MICHIGAN UNIVERSITY
By
Raymond Snowden

Purpose of the Study
This study attempts to analyze the effectiveness of
the Martin Luther King Developmental Program at Western
Michigan University relative to its purpose which is pre­
paring high-risk high school graduates to perform success­
fully in curriculums at Western Michigan University.
Methods and Techniques
A multivariate analysis of variance, a one sample
t-test, and the Scheffe'Post hoc comparisons were the
statistical techniques applied.
Three groups of students were selected as the basis
for this study.

The sample for the developmental group

(Group I) consisted of the entire group of sixty-two Martin
Luther King students.

The subjects were selected to par­

ticipate in the program in June, 1969, by the Program Director.

Raymond Snowden
The mean high school grade point average of the group was
1.9 8 and the mean ACT score was 13.

Twenty-nine females

and thirty-three males were admitted and became the basis
of this study.

The group participated in a six-week resident

Summer Orientation Program where developmental courses,
counseling and financial assistance was provided.

This

group was enrolled as full-time students in the Fall of
19 70, and were given additional counseling, tutoring and
financial assistance.
The sample for the control group (Control Group II)
consisted of twenty-nine females and thirty-three males.
The subjects were randomly selected from 177 students who
were admitted as freshmen at the beginning of the Fall,
19 70 semester.

All of the students of the control group

were admitted as regular students without benefit of a
special program.

The mean high school grade point average

of this control group was 2.17 and the mean ACT score was
18.
The sample for the control group (Control Group III)
consisted of twenty-nine randomly selected females with
high school grade point averages of exactly 2.50 and fifteen
males with high school grade point averages of exactly 2.20.
The mean high school grade point average for the group was
2.39 and the average ACT score was 18.

All of the students

in this control group were admitted as beginning freshmen

Raymond Snowden
at the beginning .of the Fall semester, 1970, without benefit
of any special program.
Findings
This study was designed to evaluate the effectiveness
of the developmental program at Western Michigan University
relative to its purpose of preparing high-risk students to
perform successfully in curriculums at Western Michigan
University.

The results of this study support the contention

that such a program can benefit high-risk students in insti­
tutions of higher learning.
From the data and subsequent findings of this study
the following conclusion regarding the effectiveness of the
Martin Luther King Developmental Program is drawn.
The Martin Luther King Developmental Program is
successful in preparing high-risk high school
graduates to perform successfully in curriculums
as Western Michigan University.
Relative to this conclusion the major findings were:
1.

The college grade point average of the Develop­

mental Group was significantly higher than the grades two
control groups.
2.

The developmental students achieved higher grades

than were predicted by the ACT program.
3.

The Developmental Group not only met the 2.00

minimum requirement for good standing but significantly
exceeded the 2.00 with an earned grade point average of
2.56.

Raymond Snowden
4.

The Developmental Group was able to maintain a

level of academic persistence equal to that of the regularly
admitted students in Control Groups II and III.

In addition,

while 19 per cent of all freshmen dropped out of school
during their first year, only 16 per cent of the students
in the Developmental Group dropped out.

TO CHARLENE
This thesis is affectionately dedicated to you for your
unfailing encouragement, ready assistance,
and willing sacrifices during
this investigation.
To my son, Darrl, who somehow understood why the television
had to be turned off during his favorite
programs.

ii

ACKNOWLEDGMENTS

The author wishes to express his sincere thanks to
the numerous people who assisted, advised and encouraged him
throughout the duration of this study.
I wish to express my sincere appreciation to my
thesis committee chairman, Dr. Max Raines, for his interest
and encouragement while directing this research and through­
out my doctoral degree program.

I would like to acknowledge,

also, the assistance of my committee members, Dr. Gordon
Aldridge, Dr. Dale Alam, and Dr. Normal Bell.
A special debt is owed to the director of the Martin
Luther King Program, Mr. Roger Pulliam, and his assistant,
Mr. Nelson Jackson, for allowing me to study the program
and to Dr. Russell Seibert, Dr. Russell Gabier, Mrs.
Juanita Hennings and Mr. Dave Morris for their assistance
in the collection of data.
ing is greatly appreciated:

Further, the help of the follow­
Mr. Robert Wilson, Research

Consultant, whose statistical skill and time was foremost
in the completion of this study, Dr. Howard Teitelbaum and
Dr. Larry Lezotte for their assistance during the earlier
stages of this research.
Special gratitude is extended to my lifelong friend
and colleague, Dr. William Pickard and Mrs, Ruby Donaldson,
iii

both of whom have been a continuous source of encouragement
throughout my professional career.
Finally, I wish to express my deep gratitude to my
mother and father, William and Lula Snowden.

Their con­

tinued encouragement and faith in my ability has been an
impetus for the successful completion of my educational
goals.

TABLE OF CONTENTS
Page
A C K N O W L E D G M E N T S .......................................... i:Li
LIST OF TABLES

....................................vii

LIST OF F I G U R E S ......................................... viii
LIST OF A P P E N D I C E S ........................................ ix
Chapter
I.

INTRODUCTION

.....................................

1

Background
....................................
3
The P r o b l e m ....................................
6
Purpose of the S t u d y ..........................
7
General Statement of Hypotheses
. . . .
8
Design ........................................
9
Delimitations
..............................
11
D e f i n i t i o n s ..................................... 11
O v e r v i e w ..........................
12
II.

REVIEW OF THE L I T E R A T U R E ............................14
Interest in Developmental Education . . . .
14
Studies Relating to Academic Achievement
. .
31
Developmental Programs Currently in
E x i s t e n c e ........................................ 43
S u m m a r y ........................................... 50

III.

DESIGN OF THE S T U D Y .................................. 52
Definition of the P o p u l a t i o n .................... 52
A d m i s s i o n ........................................ 54
General Requirements
.......................
54
Honor P o i n t s ..................................... 55
..............................
55
Grading System
Grade Point A v e r a g e ........................... 56
Scholarship Standards .......................
56
Nature of the Sample for Developmental
Group I ........................................... 57

v

Chapter

Page
Library Science ..............................
59
S p e e c h ........................................... 60
Study C e n t e r .....................................60
Physical Education and Recreation . . . .
60
Nature of the Sample for Control Group II .
62
Nature of the Sample for Control Group III
.
62
M e a s u r e s ........................................... 65
Data C o l l e c t e d ................................. 65
American College Testing Program (ACT)
. .
68
Cumulative GPAC and Enrollment Status
. .
69
Testable Hypotheses
..........................
70
M e t h o d o l o g y ........................................ 74
S u m m a r y ........................................... 75

IV.

ANALYSIS OF D A T A .....................................78
Hypotheses for First Design
................
78
Hypotheses for Second Design
................
79
Exploratory Analysis ..........................
80
First D e s i g n .....................................82
Summary of First Design
....................
90
Second Design
.................................
91
Summary of Second Design ....................
96
S u m m a r y ....................................... 9 8

V.

SUMMARY AND C O N C L U S I O N S ..........................101
S u m m a r y ....................................... 101
First D e s i g n ................................... 104
Second Design
............................
10 7
Findings and Discussion .......................
109
C o n c l u s i o n s ....................................... 115
Implications for Future Research .............
118

BIBLIOGRAPHY
APPENDIX

...........................................

119
125

LIST OF TABLES

Table

Page

3.1.

Grading S y s t e m ..................................... 55

3.2.

Summary of the Characteristics of the
Population:
GPA^s and S e x ........................ 63

3.3.

Summary of High School Grade Point Averages of
All Beginning Students Entering the
University in 19 70

64

3.4.

Summary of the Characteristics of the Sample:
GPAhs and S e x ..................................... 66

3.5.

Data C o l l e c t e d ..................................... 67

3.6.

ACT Standard Scores for All 19 70 Beginning
Freshmen at WMU Test Subparts Means (X)

.

.

69

4.1.

Correlation Matrix with Covariable Included .

.

81

4.2.

Correlation Matrix with Covariable Eliminated

.

81

4.3.

Cell Frequencies for First D e s i g n ................. 83

4.4.

Summary of Univariate Analysis— Group Effect

.

85

4.5.

Scheffe'Post hoc Comparisons for Differences

.

86

4.6.

Summary of Univariate Analysis— Sex Effect

.

89

4.7.

Summary of First Design Group Means

4.8.

Cell Frequencies for Second D e s i g n ..............9 3

4.9.

Summary of Univariate Analysis Treatment Group
E f f e c t ........................................... 95

4.10.

Summary of Univariate Analysis Response Group
E f f e c t ........................................... 95

4.11.

Summary of Second Design Group Means

vii

.

. . . .

. . . .

92

97

LIST OF FIGURES

Figure

Page

3.1.

First D e s i g n ........................................ 72

3.2.

Second Design ....................................

viii

73

LIST OF APPENDICES
Appendix

Page

1.

Summary of C o r r e l a t i o n s .......................... 126

2.

General Studies Program

3.

Academic Load

4.

Martin Luther King Summer Session Academic
C a l e n d a r .......................................... 129

5.

Major and Minor Areas ofEnrollment

6.

Cost and R e s o u r c e s ................................. 131

...................

12 7

................................. 12 8

. . . .

130

CHAPTER I

INTRODUCTION
Historically, blacks and other minority group members
have been disproportionately represented in institutions of
higher learning.'*'

These groups have, by and large, been

poorly prepared in our high schools and later refused admis­
sion to colleges and universities.
The 19 60s began a new era toward the education of
minority group members in this country; an era that witnessed
more blacks than ever before entering into institutions of
higher education.

Many of these "new students" were able,

without any assistance from the institutions, to reach their
educational goal.

And at the same time, many of these

newly admitted students were met by total frustration and
failure as they soon found that what they thought was an
open door to education was really a revolving door.

Many

of the marginal or high-risk students found themselves in
no credit remedial programs that not only failed to teach
them anything but also frustrated and discouraged them from
continuing their education.

Still other students found

themselves wandering aimlessly as they attempted to wade
^■Robert A. Green, Racial Crisis in American Education
(Chicago:
Follett Educational Corporation, 1969), pp. 33-37.

1

2
their ways through program schedules that had been rubber
stamped by their counselors.

Many of these students were

misdirected and uncertain of their goals.

They were not

aware of their own resources and had never learned proper
2
study habits.
Following the year 1967, many colleges and uni­
versities began to take a second look at their programs
for minority students and found them woefully inadequate.
Black studies programs were developed at several universi3
ties, more remedial programs were added at others, and
fewer still created comprehensive developmental programs
in an attempt to meet the total needs of the "new students."
A review of the literature suggest that, by and large, these
programs were poorly conceived and lacking of scientific
research to support their existence or to measure their
effectiveness.
Fortunately, not all developmental programs fall
into the above category.

Many programs have been well

thought out and are doing a seemingly good job in providing
innovative and sound educational opportunities for highrisk students.
2 .

William Moore J r . , Against the O d d s : The High Risk
Student in the Community College (San Francisco, California:
Jossey Bass, Inc., 1970.
3
Armstead D. Robinson, Craig C. Foster, and Donald H.
Oglilive, eds., Black Studies in the University (New York:
Bantam Books, 1969).

3
One such program is the Martin Luther King Develop­
mental Program at Western Michigan University to which this
research is directed.

The purpose of this study is to

analyze and appraise the program relative to its purpose
of helping low achieving high school graduates perform
successfully in college.
Background
The Martin Luther King Program at Western Michigan
University is a developmental program for marginal students.
The Program began in September of 196 8, as a result of a
five-year grant from the Kellogg Foundation.
purposes of the program are:

The stated

(1) to provide educational

opportunities for disadvantaged students by recruitment
efforts;

(2) to force the re-examination of admission

criteria for these students;

(3) to encourage marginal

students to pursue higher education by providing financial
assistance, counseling and remedial help at the university
level.
The typical Martin Luther King
is a black American student.

(MLK) participant

He is no stranger to failure

and to the fear that continued failure engenders.

The

student is typically from lower socio-economic groups, is
notably deficient in academic strengths, and may fall into
one or all of the following categories:
1.

Real ability not identifiable by grades or test
scores.

4
2.

Individual potential not developed due to
inadequate schooling.

3.

Lack of referents who have the desire and
ability to assist the individual in his
educational career.

To effect the aforementioned three-fold purpose of
the program the following projects were instituted:
1.

To Provide Educational Opportunities for
Disadvantaged Students.

In order to accomplish this end visits were made,
by the program director and his staff, to six southwestern
Michigan High schools to inform the students of the edu­
cational opportunities available to them at Western Michigan
University (WMU).

In addition, arrangements were made for

375 of these students to visit the campus of WMU.

While

on campus, these students met with university students,
faculty and administrative staff.
2.

To Force a Re-examination of Admission Criteria.

Operating on the contention that the background,
environment, and aspirations of the educationally dis­
advantaged black student does not fit the mold of the
average white aspirant to higher education, new admission
criteria were developed.
school grade point average

To this end, the existing high
(GPA) requirement of 2.20 for

males and 2.50 for females was removed for the MLK students.
3.

To Encourage Marginal Students to Pursue Higher
Education by Providing Financial Assistance,
Counseling and Remedial H elp.

5
For each of the students accepted into the program
a financial package was put together consisting of funds
from one or all of the following:

Economic Opportunity

Grants, Work Study, or National Defense Loans.
The counseling team consisted of one half-time
counselor, two graduate assistants and thirty junior and
senior level student aides.

The student aides had the

primary responsibility for tutoring the students and for
making visits to the respective instructors to determine
the progress of the student.

In addition, remedial pro­

grams were instituted to supplement and assist the students
with their normal classroom assignments.

One of the more

promising remedial programs was that of the Study Center.
The Study Center was located in the campus library.

Each

MLK student attended the Study Center two nights per week.
The Center was staffed by graduate students, tutors and the
Study Center Director.
To further encourage and assist marginal students
in their efforts to attain higher education, a Summer
Orientation Program was instituted.

Attendance at the

summer orientation program is required of each student
planning to enter the following Fall semester.

Students

enrolled in the Summer session are registered for a full
academic load of credit and non-credit courses.

The ac­

credited courses consist of Race and Culture (4 hours credit)
and Adult Reading

(2 hours credit).

The non-accredited

6
courses are Speech and Library Science.
are also available.

Counseling services

The Summer Session staff is made up of

the Project Director and his assistant, a counselor, faculty
members and student aides.

The primary responsibility of

the student aide is to serve as tutor-counselors in the
residence halls and to offer needed assistance to faculty
members.
The Problem
This study attempts to analyze the effectiveness of
the Martin Luther King Developmental Program at Western
Michigan University relative to its purpose which is pre­
paring high-risk high school graduates to perform success­
fully in curriculums at Western Michigan University.

The

study is based on the analysis of three groups of students:
1.

Developmental Group I whose participants were

admitted to the University at the beginning of the Summer
Session of 19 69, as members of the Martin Luther King
Developmental Program; 2.

Control Group II whose members

were enrolled as freshmen at the beginning of the Fall
Semester, 19 70, and whose high school grade point average
(GPA^g) were 2.49 or lower females and 2.19 or below for
males;

3.

Control Group III whose members were enrolled

as freshmen at the beginning of the Fall Semester, 19 70
and whose high school grade point averages were 2.50 for
females and 2.20 for males.

7
Academic success or failure, and correspondingly,
program success, of Developmental Group I will be determined
on:

(1) the basis of the members' ability to obtain a 2.00

(4.0 scale) college grade point average (GPA ) at the end of
c
one year's attendance;

(2) a comparison of the level of

persistence of Group I with Group II and Group III.
Purpose of the Study
The purpose of the study is to analyze and appraise
the Martin Luther King Developmental Program for Marginal
Students at Western Michigan University relative to helping
low-achieving high school graduates perform successfully at
the college level.

More specifically, this study attempts

to:
1.

Describe the Developmental Program as it is now

operating at Western Michigan University.
2.

Compare the rate of persistence of those students

who entered the MLK Program during the Summer of 1969, with
a group of regularly admitted students who entered during
the Fall of 19 70, and whose high school grade point averages
were below the reported 2.50 and 2.20 minimum standard
required for regular admission.
3.

Compare the rate of persistence of those students

who entered the MLK program during the Summer of 19 69, with a
group of regularly admitted students who entered during the
Fall of 19 70, and whose GPA,
were at the 2.50 and 2.20
hs
minimum standard required for regular admission.

8
4.

Compare the college grade point averages of

those students who entered the MLK Program during the
Summer of 1969, with the college grade point average of a
group who entered during the Fall of 19 70, and whose GPAhg
were below the reported 2.50 and 2.20 minimum standard
required for regular admission.
5.

Compare the grade point averages of those

students who entered the MLK Program during the Summer of
1969, with the grade point averages of a group who entered
during the Fall of 19 70, and whose GPAhg at the 2.50 and
2.20 minimum standard required for regular admission.
6.

Compare the predicted grade point average

(GPA1 ), derived from the American College Testing Program
ir

(ACT), with the college GPAc of each of the three groups.
7.

Identify, clarify and interpret the data as

they might relate to future developmental programs.
Generally, then, the results of this study can be
useful in determining whether such a program is a practical
means of providing further education for high school gradu­
ates who were a "high-risk" for higher education because of
a lack of basic skills rather than a lack of intellectual
ability.
General Statement of Hypotheses
The primary purpose of this study is to determine
the success of the single structured developmental program

9
at Western Michigan University in preparing high-risk high
school graduates to perforin successfully at Western Michigan
University.

In order to accomplish the stated goals of this

study it is necessary to examine the three groups in terms
of group difference and sex differences as well as inter­
action between groups and sexes.
Relative to these goals the major hypotheses are:
H^:

The mean of Developmental Group I will be equal
to the mean of Control Group II and Control
Group III on GPAC , drop-out rate, and on the
discrepancy between the predicted GPA and the
actual GPA.
The mean of the males will be equal to the mean
of the females on GPAC , drop-out rate and on the
discrepancy between the predicted GPA and the
actual GPA.

: The mean of each of the three groups will be
equal to the mean of each of the two genders
on GPAC , drop-out rate and on the discrepancy
between the predicted GPA and the actual GPA.
Design
All data was obtained from the records office at
Western Michigan University.

Demographic information such

as age, and sex has been compiled for each group.

In

addition, the standard scores on the ACT test were compiled,
were available, for each member in each of the three groups.
The sample for Group I consists of all of the 62
students

(29 females and 33 males) who participated in the

Summer Orientation Program in 1969, and who enrolled in the
Fall.

Grade point averages were compiled for the Summer,

Fall and Winter terms.

10
WMU reported the minimum high school GPA required
for regular admission to be 2.20 male and 2.50 female.
However, a number of students were given regular admission
status during the Fall of 19 70, with GPAhg of 2.19
and below and 2.49

(male)

(female) and below.

Therefore, the membership of Control Group II was
derived from a stratified random sample of 29 females
having GPAhg ranging from 1.86 to 2.49 and 33 males having
GPA^s ranging from 1.72 to 2.19 and who were enrolled as
freshmen during the Fall term, 19 70, without benefit of
any special program.

(The mean GPA^g of all beginning

freshmen in Fall, 19 70, was 2.83.)
Control Group III consist of a random sample of
29 females and 15 males

(only 15 males had GPA^g of exactly

2.20) who were enrolled as freshmen during the Fall term,
1970, without benefit of any special program.
At the end of the academic year the cumulative
grade point averages for each student in each group was
tabulated and analyzed to determine the effect of the
Developmental Program on the academic success of Develop­
mental Group I.

The number of students in each group who

successfully completed one academic year was also tabulated
and analyzed and, finally, a comparison was made between
the predicted GPA and the actual college GPAc at tne end
of one year's attendance.

11
Since all three groups have GPA^g substantially
below the mean for all freshmen, regression toward the
mean was anticipated.

To assure that any difference in

college GPAc reflected a true difference, consideration
was given to using multivariate analysis of covariance
(MANOCVA).
Thus, the proposed analysis consist of a 3 x 2 x 3
block design using group membership and sex as the independ­
ent variables and multivariate analysis of covariance with
high school

as covariables and drop-out r ate, college

GPAc and the discrepancy between predicted GPA and actual
GPA as dependent variables.
Delimitations
This investigation was limited to one hundred and
sixty-eight students who were beginning freshmen in 1970,
at Western Michigan University.

Regularly admitted students,

with high school grade point averages below admission
standards, constituted one control group and regularly
admitted students, with high school grade point averages
at the admission standard level, made up the second control
group.

The experimental group consisted of marginal stu­

dents placed in a developmental program.
Definitions
Drop-out.— Any student who was enrolled at the
beginning of Summer or Fall semesters, 1970, but was not
enrolled at the beginning of Fall semester, 1971.

12
Regular Admitted Student.— Any student admitted to
WMU at the beginning of Fall semester, 19 70, with a high
school GPA of 2.50 or below, who was not a member of any
special developmental program and was given regular admis­
sion status by the university.
One Year's Attendance.— Any student who completes
the Fall and Winter semesters, 19 70, is said to have
completed one year's attendance.
Persistence.— Any student who was enrolled as a
beginning freshman in June, 1969, or Fall, 19 70, completed
Fall and Winter semesters and who was also enrolled in
Fall, 1971, is said to have persisted.
Overview
In this chapter the problems which generated the
study are discussed along with the background and description
of the program to be evaluated.

The problem was stated as

an attempt to analyze the effectiveness of the Martin Luther
King Developmental Program at Western Michigan University
relative to its purpose which is preparing high-risk high
school graduates to perform successfully in curriculums at
Western Michigan University.

The purpose has been dis­

cussed, the major hypotheses stated in research form, and
the research design has been described.

The completion of

this thesis will be presented in four additional chapters.
In the next chapter, Chapter II, a review of the
literature relating to developmental programs and marginal

13

high school students is presented.

The focus will be con­

fined to the current and most pertinent aspects of those
studies.
The sampling techniques, measurement procedures,
research design and other measures used in this study is
presented in Chapter III.

The analysis procedures and

hypotheses are also described.
In Chapter IV the analysis of the data is presented,
the hypotheses are examined and findings discussed in
relation to the data.
The entire study is summarized, conclusions are
drawn, and implications for future research are discussed
in Chapter V.

CHAPTER II

REVIEW OF THE LITERATURE
In this chapter the review of the literature is
presented.

The chapter is divided into three sections.

In the first section the interest in developmental education
is discussed.

In the second part a review of some of the

studies related to academic achievement is presented, and
in the third section a description of four currently
existing developmental programs is presented.
Interest in Developmental Education
Many colleges and universities have instituted
special programs for disadvantaged or minority students.
A discussion on developmental programs should not begin
without a clear understanding of what is meant by such
terms as disadvantaged, high-risk, marginal or other terms
usually intended to refer to poorly prepared high school
graduates.

William Moore uses such terms as high-risk,

marginal, educationally disadvantaged, and academically
unsuccessful, interchangeably to refer to students whose
poor high school grades, economic plight, low standard
test scores, and race, cultural or class distinctions
place them at a disadvantage with the majority of students
14

15
entering into college.^

Moore contends that while many

of these students appear to have little prognosis for
success, many of them possess intangible qualities.

These

intangible qualities of creativity, personality and tenacity
9
counteract the customary indicators of academic prowess.
In fact, he says, the students considered disadvantaged
or high-risk on one campus would be considered intellectually elite on other campuses.

3

Disadvantagement is defined

by many colleges within the context of their own student
bodies; "that is, disadvantaged students are those whose
educational and economic background is considered markedly
4
inferior to that of their regular students."
The term
high-risk is defined by Williams as the financially poor
student who, in terms of traditional predictive criteria,
5
has a poor chance of succeeding at a particular college.
William's definition of high-risk appears a little too
narrow to give the full picture of the type of student
generally referred to as high-risk.

Moore, on the other

hand, seems to have a better grasp of the high-risk student.
'*'Moore, op. cit. , p. 5.
^Ibid., p. 6.
3
Robert L. Williams, "What Are We Learning From
Current Programs for Disadvantaged Students," Journal of
Higher Education (April, 1969), pp. 274-275.
^Ibid., p. 275.
^Ibid.

16
For the sake of this discussion all of the terms used above
will be used interchangeably to refer to those students
whose low prognosis for success is conteracted by a sincere
desire to learn.
Recent reports indicate that probably more than 50
per cent of the institutions of higher education have special
g

programs for high-risk students.

Most of these programs,

according to Egerton are little more than token efforts and
with the exception of predominantly Negro colleges, private
schools have evidenced more involvement than have public
institutions.

7

The extent to which these programs constitute

a viable approach to higher learning is a subject of contro­
versy.

A recent study by McDaniel and McKee suggest that,

for the most part, colleges and universities are not
responsive to the needs of minority students as they relate
O

to higher learning.

From a survey of 2,764 predominantly

white colleges and universities, the authors found that
57 per cent of the 1,393 respondents stated that they had
open admissions.

Twenty-five per cent of the respondents

reported that they had adjusted their admission criteria
to admit more minority students.

Slightly over 50 per cent

g

John Egerton, Higher Education For "High Risk"
(Atlanta:
Southern Education Foundation, 1968),

Students
p. 59.

^Ibid., p. 13.
O

Ruben R. McDaniel and James W. McKee, An Evaluation
of Higher Education's Response to Black Students (Indiana
University: September, 1971).

17
of the institutions reported the existence of an academic
help program.

In an attempt to meet the cultural needs of

the minority students, one half of the institutions reported
the presence of some form of ethnic studies program.

Forty-

five per cent of the institutions reported the existence of
training programs geared toward the improvement of race
relations.

However, the majority of the programs were

geared only toward the counseling staff leaving the faculty
members almost totally uninvolved.

Activities geared towards

the active recruitment of minority faculty members was
reported by 44 per cent of the respondents.

Other results

of this study showed that 8 per cent of the colleges were
making efforts to provide residential patterns which
promoted good race relations and 25 per cent were providing
financial aid programs.

One fifth of the institutions

reported the use of institutional funds to support special
programs and less than 30 per cent of the institutions had
developed policy statements concerning race which might act
9
as guides for institutional behavior.
McDaniel's and McKee's study appears less saddening
when one reviews the study by Cash involving seventeen
colleges and universities in the state of Georgia.

Cash

found that the majority of the predominantly white institu­
tions provided no academic support programs for high-risk
students beyond admitting them on a probationary basis when
9Ibid.

18
they failed to meet the unadjusted high school grade point
average and SAT requirements.

On the other hand, Trent

writes that the impatience of today's youth in their un­
willingness to accept traditional promises and platitudes
have spurred many universities to re-evaluate their policies
and to provide more educational opportunities to disadvantaged
students.

He continues that although circumstances such as

lack of background, knowledge, resources or pressure to
take immediate action have hindered the proper development
of many support programs, some universities have designed
programs which are beginning to serve the educationally
disadvantaged student.

Trent cautions, however, that ill-

designed programs are often degrading and exclusionary and
fail to achieve the goals of the student or the university.
Community colleges appear to have taken the lead in
addressing themselves to meeting the higher learning needs
of minority students, as evidenced by Edmund Gleazer's
comments.
The Association [American Association of Junior Colleges]
recognizes that poverty and prejudice are barriers to
opportunity for millions of American and thus impedi­
ments that restrict and threaten national progress.
AAJC also believes that education, and particularly
two year colleges, must help lead the assault on these
"'"^Carol D. Cash, Educationally Inferior Students:
Getting In and Out of College (University of Missouri:
M a y , 1970).
11William T. Trent, College Compensatory Programs
for Disadvantaged Students, Report No. 3 (Washington, D . C . :
Eric Clearinghouse, Sept. 1970).

19
barriers— an assault which is only in its formative
stages.
This assault must close not only the gap
in educational opportunities, but cultural and
economic gaps as well, and the two year colleges
can and should play a leading role in overcoming
all of these gaps.12
Dorothy Knoell concluded from her study of five
cities that, while community colleges are doing well in
attracting minority students to their institutions, the
problem appears to be of insuring successful performance
after admission.

13

In terms of insuring success, Gordon Morgan has
made several points worth mentioning in detail.

First,

traditional teaching methods are not effective for the
poorly prepared student.

Innovative techniques are

essential to educating these students.

Second, the

student from the ghetto is less concerned with academic
success than with improving living conditions and oppor­
tunities of people living in the slums.

Third, these

students feel that colleges do not want them, view them
and their culture disparagingly, and have no intention
of being truly relevant to the needs of the black corn14
munity.
Gordon concludes . . . it appears
that colleges need to rethink and re-orient their
activities, teaching practices and expectations in
12

Edmund J. Gleazer, cited in Dorothy Knoell, Black
Student Potential (Washington, D . : American Association
of Junior Colleges, 1970), p. 11.
13Ibid., p. 77.
14

Gordon D. Morgan, The Ghetto College Student
City, Iowa:
American College Testing Program, 19 70).

(Iowa

20
the light of the sociology of economic deprivation
of our inner city students.
In order to allow the
ghetto student a chance to succeed, some toleration
of departures from middle class values must be
exercised.
The ghetto student is a complex,
complicated person, admitting his existence forces
the college to try to understand not only what the
student wants for himself but what the college
really wants for him.
Colleges can no longer
avoid social decisions and social action.
In
answering the questions, 'Should this college
educate ghetto youth?, and if so how?,' the
college is making social decisions and taking
important social action.15
Additional indications that traditional teaching methods
are not effective has been given by Irene Tinker.

She

comments that traditional counseling and remedial programs
have not been successful because they are often warmed-up
high school courses.

She continues that courses geared

toward enabling a student to study on his own or to evaluate
himself will not help a student who lacks motivation.

By

the same token, repeating a grammar course from which he
learned little in high school will not motivate a student.
Among Tinker's suggestions for improving the education of
disadvantaged students are:

reducing the minimum load

requirements, disregarding poor grades during the first
year, continuing scholarships when the student is on
probation and peer counseling.

16

The success of any educational program designed to
meet the needs of the high-risk student is dependent in the
15Ibid. , p. 56.

16

Irene Tinker, "The Underprepared College Student,"
American Education (November, 19 70).

21
final analysis, on its bringing to maturity the potential
of the educationally disadvantaged students who have not
had previous success in academic life and preparing them
for success in institutions of higher learning.
One of the more comprehensive studies on disadvantaged
students has been completed by Robert Williams.

In his

report high-risk students were characterized as having lower
standardized test scores than regular students.
College Testing Scores

American

(ACT) in the vicinity of 13 or 14

and Scholastic Aptitude Test Scores of 700-800 are cited
as norms for disadvantaged students.

Several colleges were

reported to accept only those students with outstanding
academic records and limited finances.

Few institutions

were reported to have been actively recruiting impoverished
ghetto students with major academic weaknesses.

Criteria

most often employed in the selection of disadvantaged
students included:

(1) tangible evidence of ability, i.e.

high school grades or promising standardized test scores;
(2) willingness to accept personal responsibility for
success or failures;
motivation;

(3) high self concept;

(5) ability to think creatively;

to set realistic goals;

(4) strong
(6) ability

(7) prior success in any activity

that required sustained effort.

Data upon which selection

was made were typically obtained through subjective evalua­
tions accruing from personal interviews and on the recommendations of high school counselors, ministers, teachers, etc.
17 .
Williams, op. cit., p. 276.

17

22

Nearly all of the programs in William's study had
some provision for financial assistance.

The money came

from such sources as Educational Opportunity Grants
National Defense Education Act Loans
study programs and local gifts.

(EOG),

(NDEA), federal work

The Ford and Rockefeller

Foundations were reported as major contributors.

18

The primary academic objective of most programs
is the development of communication skills.

Many programs

include compensatory study in standard English during a
bridge program in the summer before beginning as a freshman
or on weekends during the regular academic year.

An at­

tempt is made to teach standard English in a manner that
does not conflict with the student's cultural dialect.
The student is taught how his dialect functions as a
legitimate language, and standard English is taught as a
second language."^
Several approaches have been utilized to maximize
the academic success of disadvantaged students.
use has been made of individualized instructions.

Considerable
Much use

has been made of tutoring in virtually all academic areas.
Tutors include instructors, graduate students, undergraduate
students, and junior and senior level disadvantaged students.
18Ibid.
19

Leslie Berger, College Now For Ghetto Youth (City
University of New York:
1968) , p. 8T

23
While many of the tutors are work study students, several
serve on a volunteer basis.
Programmed instruction is another type of individual­
ized instruction currently being used to supplement classroom
instruction.

Goodrich sees a great future in programmed

learning and criticizes the heavy use of remedial courses.
He writes that the use of programmed learning may turn out
to be a major part of the answer to corrective education
m

institutions of higher learning.

20

Moore,

21

Johnson,

22

and others also see great value in programmed texts and
feel that they have unusual potential for helping low
achievers.

Programmed texts are used in teaching freshmen

communications courses at Palm Beach Junior College,
Florida.

At Bakersfield College, California, a teaching

machine is used to provide for immediate feedback.

In an

attempt to evaluate programmed teaching, three plans were
used in various sections of the course.

One section was

taught by the machine, Auto Tutor Mark II; another section
was taught by the conventional lecture; and the third by
a combination of the machine and the lecture.

Achievement

20

Andrew L. Goodrich, Community Services For The
'New Student' At Inner City Community Colleges (Michigan
State, Kellogg Community Services Leadership Program, 1970),
p. 12.
91
*

22

Moore, op. cit.

Lamar B. Johnson, Islands of Innovation Expanding:
Changes In The Community College (Beverly Hills, California;
Glencoe Press, 1969), pp. 73-87.

24
test results indicated that the students taught by the
combination method scored almost one standard deviation
above the other two groups.

There was no significant

difference in achievement between the groups taught by
the other two methods.

The conclusion was that programmed

instruction was as effective as traditional learning.

23

Much use has also been made of programmed learning
centers.

Forest Park Community College, St. Louis; Macomb

County Community College, Michigan; and Oakland Community
College, Michigan are notable examples.

At Forest Park

all students who score low on the placement test are
required to enroll in the General Curriculum.

There they

are assigned to the learning laboratory for six hours per
week.

The student's work in the laboratory is planned

with the assistance of a counselor.

At the end of each

unit of study the student is required to pass a test before
he can proceed to the next unit.

24

The Programmed Learning

Center at Macomb County Community College is similar to the
Learning Resource Center at Oakland Community College.

At

both colleges the students take diagnostic achievement test
to aid in selecting appropriate learning programs.

These

examinations, followed by test, after the completion of
the program are used to evaluate the student's learning.
An important feature of the Macomb Center is that the
23Ibid., p. 75
24

Moore, op. cit . , pp. 184-198.

25

learning center staff studies the course outlines and then
provides instructors with lists of programs which may be
helpful m

teaching a particular course.

25

Johnson cautions

against three possible shortcomings in the use of programmed
instruction:
to develop;

(1) Programs can be costly and time consuming
(2) Problems in motivating some students through

the use of self-paced programs; and

(3) The tendency for

instructors to use programs in teaching dull and monotonous
material.
Other policies currently being used by colleges and
universities in their endeavors to serve the disadvantaged
student include lighter class loads and liberal probationary
policies.

A liberal probationary policy often includes

allowing the student to repeat courses many times and to
take several quarters to raise their grade point average
to the acceptable level.

Other colleges and universities

were reported to use placement examinations to determine
whether a student should enroll in a regular course or a
remedial course.
non credit.

27

The remedial course may be credit or

Di Russo and Aven found remedial English

to be of little value in their study of fifty-four college
freshmen in Missouri.

Their study attempted to determine

the effect of remedial English on the student's proficiency
Johnson, op. cit.
26Ibid.
27

Williams, op. c i t . , p. 278.

in written English.

A random sample of fifty-four freshmen

were selected from several hundred who had taken the
Missouri College English Test and who scored below the
40th percentile.

The experimental group took a remedial

English course and the control group did not.

At the

conclusion of the courses both groups took another form of
the test.

No significant difference was found between the

two groups.

28

Taylor, on the other hand, found signifi­

cantly high results with his study on tutorial service.
His study included thirty-one engineering students matched
on school and college performance.
in English, Math and Physics.

One group was tutored

The grades of the students

involved in the tutorial program were significantly higher
than the non-tutored students.

29

The overuse and the

ineffectiveness of traditional remedial courses is dis­
cussed in detail by G o o d r i c h ^ and T i n k e r . ^
Much attention has been given to the affective
side of high risk students as to the academic.

In

efforts to bridge the gap between high school and college
many colleges and universities make efforts to soften the
28

L. DiRusso and S. D. Aven, "Does Remedial English
Provide Help for College Freshmen," California Journal of
Educational Research, Vol. 22 (1971), pp. 5-8.
29

R. G. Taylor, "Tutorial Service and Academic
Success," Journal of Educational Research, Vol. 62 (1968),
pp. 195-1971
:
30
31

Goodrich, op. cit.
Tinker, op. cit.

r-

27
transition.

Some institutions bring the prospective stu­

dents to the campus for several days or weeks during the
summer months.

The visit

affords the student an oppor­

tunity to develop a more realistic conception of college
t *
32
life.

Personal counseling is often the backbone of many
special programs.

Several programs insist that the new

students have weekly sessions with their counselors.
arrangements allow
get out of hand.

Such

problems to be dealt with before they
Often the number of counseling sessions

are reduced as the student demonstrates adequate personal
and academic progress.

The role of the counselor is to

bridge the gap between students and the academic establish­
ment, to provide a personal orientation to college life,
to advise in course selection, to facilitate formulation
of career goals, to assist students in overcoming poor
study habits, and to aid them in achieving a sense of their
own identity.

In addition to counseling, special courses

such as Negro History, Poverty, Urban Conditions and Civil
Liberties have been utilized to assist the students in
better understanding themselves and their role in society.
Some schools have found that small group discussions on
social issues often evoke more candid articulation of
student feelings than counseling sessions.
32

Williams, op. cit., p. 278.

33Ibid., p. 479.

33

28
Most of the reporting programs attempt to provide
role models for disadvantaged students.

A few program

staffs include full-time members from disadvantaged back­
grounds.

Other schools have cooperative work study programs

with industry, whereby students alternate between quarters
with school and work.

Such experience affords the student

an opportunity to see members of his own minority group
function at respectable positions.

34

In terms of the effectiveness of the various pro­
grams, Williams states that in most instances the drop-out
rate for the high-risk students has been no higher than
for regular students.

In addition, the grades of the dis­

advantaged students have generally been much higher than
predicted grades resulting from high school grades and
standardized test scores.

In evaluating the programs he

mentions that it is extremely difficult to determine
exactly what factors are responsible for the success or
failure of the students because the programs were not
experimentally designed to permit empirical assessment.
Consequently, evaluation is based more on the personal
testimony of the program staff than on empirical evidence.
Williams concludes his report with several cogent
recommendations worth repeating:
^ I b i d . , p. 279 .
~^Ibid. , p. 281.

(1) Living conditions in

35

29
the ghetto are often adversely related to academic develop­
ment.

In these cases, unless students are physically removed

from these socially destructive circumstances, they have
little chance of success in college;

(2) Special program

staff, administrators, and tutors should not expect dramatic
success initially, as high risk students frequently get off
to slow starts academically;

(3) If remedial courses are

included in special programs they should carry college
credit since, initially, credit may be more important to
the student than academic skills in a particular subject.
Immediate success tends to enhance later success and a
program filled with non-credit remedial classes is not
likely to motivate the disadvantaged student;

(4) Remedial

assistance through the use of programmed instruction
machines, which permit the students to move along at their
own pace, are often less frustrating and anxiety arousing
than the traditional classroom;

(5) Teaching relationships

with high risk students must be highly personal.

The

instruction must convey concern for the student as a person;
(6) Intensive personal counseling, on a voluntary level,
seems to be an important affective feature in disadvantaged
programs;

(7) Individual tutoring by advanced high risk

students can be a major source of academic support for
the student receiving the tutoring and can provide af­
fective support for advanced high risk student;

(8) Non­

disadvantaged students should be encouraged to participate,

30
along with the disadvantaged, in special courses relating
to minority groups;

(9) Disadvantaged students should have

a role in the development of the special program;

(10) To

achieve success, any developmental program must have the
full support of the top administrators.

Administrative

practices regarding rigid entrance requirements, financial
support, mass instruction, impersonal facuity-student
relationships must also be modified;

(11) Faculty entrenched

in middle class values must be retrained to communicate with
non-middle class, disadvantaged students.

Frequently,

instructors contend that high risk students are academically
irredeemable and that academic standards are threatened by
the presence of disadvantaged students.

As a result,

without the full support of administrators who make the
decisions regarding money, staff, and facilities a program
for disadvantaged students can be doomed before it has had
a chance to r i p e n . ^
Thus far, the review has been directed toward the
interest in developmental education with a summary of the
nature and kinds of services available to disadvantaged
students.

The next part will be concerned with a review

of some of the services dealing with academic achievement.
The third part of this review will be directed toward a
description of some of the existing developmental programs
at the community college level.
36Ibid., pp. 280-285.

31
t

Studies Relating to Academic Achievement
Much has been written in the recent years about
the importance of standardized test scores and their value
in predicting academic success.

Most colleges and universi­

ties have formulas to assist them in deciding which appli­
cants will be accepted.

Nearly all of these formulas take

into account Scholastic Aptitude Test (SAT) or American
College Testing Service
school grades.

(ACT), high school rank and high

Fortunately, many colleges and universities

have began to take a second look at this form of admission
criteria when reviewing disadvantaged applicants.

The

purpose of this section of the review is to present some
of the most recent findings about the use of standardized
tests as a major element in admission criteria.
As recent as 19 70, Stanley advocated the undiscriminate use of standardized test.

He concludes from his

review of the literature that aptitude test scores and high
school grades predict college grades of disadvantaged
students about as well as they do for other students.

He

suggests, therefore, that admission to selective colleges
should be based substantially on test scores and high
school grades regardless of ethnic or socio-economic background.

37

What Stanley apparently overlooks in this con­

clusion is the importance of motivation and that with the
37

Julian C. Stanley, "Predicting College Success of
Educationally Disadvantaged Students" (John Hopkins Uni­
versity, September, 19 70).

32
possible exception of the Achiever Personality Scale

38

there is no objective instrument upon which to measure
the aptitude of the disadvantaged student.

Stanley also

fails to note the impact of what Jacobson, Rosenthal
and Brookover
other.

40

39

refer to as the influence of the significant

Arthur Cherdack reached different conclusions from

his study of 200 disadvantaged, minority freshmen enrolled
at the University of California.

Cherdack found that high

school grades was the best predictor for both white and
minority freshmen and that the SAT verbal score was a more
consistent positive predictor for white freshmen than
minority f r e s h m e n . ^
In a sample of 477 males and 82 7 females from
predominantly black and predominantly white colleges,
Borgen found striking results.

All of the subjects took

the National Merit Scholarship Qualifying Test (NMSQT).
38
Doris M. Miller and Patricia O'Connor, "Achiever
Personality and Academic Success Among Disadvantaged College
Students," Journal of Social Issues, Vol. 25, No. 3 (1967),
p. 105.
39

Robert Rosenthal and Lenore Jacobson, Pygmalion
In The Classroom (New York:
Rinehart and Winston, Inc.,
1968) .
40

Wilbur Brookover and Edsel Erickson, Society,
Schools and Learning (Boston: Allyn and Bacon, 1969).
41

Arthur Cherdack, Predictive Validity of the Scho­
lastic Aptitude Test For Disadvantaged College Students
Enrolled in a Special Education Program Final Report
(University of California, April, 1971).

33
An inverse relationship was reported between average college
grades and averaged NMSQT scores.

In general, the students

in the largely black colleges had lower NMSQT scores but
higher freshmen grades while the students in the primarily
white colleges had higher NMSQT scores and lower freshmen
^
42
grades.
A study was conducted by Morgan in 196 8, to deter­
mine what part SAT scores played in the acceptance of a
special group of students at Kutztown

(Pa.) State College

and what predictive value the scores had in forecasting
achievement in college.

The study points out the point­

lessness of relying too heavily on standardized tests scores
in attempting to predict academic success.

In 1961, fifty

high risk freshmen, representing 10 per cent of the incoming
freshman class, were admitted to Kutztown State College.
The group of high risk students had SAT verbal scores below
400 which was one standard deviation below the mean of 500.
The high risk students were selected over other high risk
students on the basis of:
(2)

(1) A sound high school record;

A high SAT Math score; or (3) Having impressed the

interviewer.

A follow-up report showed that 36 of the

original students received their B.A. degrees

(72 per cent)

while only 65 per cent of the non-high risk students were
42

Fred H. Borgen, Differential Expectations?
Pre­
dicting Grades for Black Students in Five Types of Colleges
(Evanston, 1 1 1 : National Merit Scholarship Corp. , 19 71).

34

graduated.

Of these 36 high risk graduates, 23 received

their degrees on schedule in May, 1965; eight needed an
additional semester to meet graduation requirements and
five were graduated ahead of schedule.
average

The mean grade point

(GPA) for the high risk graduates was 2.40 and 2.57

for the non-risk graduates.

There was no correlation

between risk graduate's GPA and SAT Math score and a
slight correlation
rank.

(r - .2 8) between GPA and high school

Morgan concluded from his study that SAT scores

can be deceptive as a valid indicator of success or failure.
To have rejected those fifty high risk students on the basis
of their SAT verbal scores would have been an injustice to
at least thirty-six students.

43

In general, studies relating high school grade point
average with college grades reveal correlations between .40
and .60.

General scholastic aptitude test such as the ACE

make similar predictions.

44

The value of the Millers Analogies Test,

(MAT)

which is currently used by several of the universities
offering a Ph.D., as a predictor of academic success,
has been called into question in the recent years.

The

MAT is most noted for its low correlation with grade point
43

Morgan, op. c it., pp. 20 3, 20 7.

^4Leona Tyler, The Psychology of Human Differences
(New York:
Appleton-Century-Crafts, 1966).

35
average, according to Hyman.

45

Fricke defends the low

correlation of the MAT on the basis that one should not
expect good discriminative efficiency since it is used in
the selection process.

He points out that the low cor-

relation may only be an indication of successful screening.

46

Previous studies relating MAT score with grades have re­
ported correlations from as low as .16 to as high as .68.4^
The inference to be drawn from a study conducted
by Cleary of black and white college students in three
integrated colleges is that the SAT is not biased against
black students.

In other words, college grades for blacks

were not underestimated when the verbal of Math SAT scores
were used as predictors.

48

While Cleary's study is probably the best controlled
research of its kind to d at e , the evidence concerning test
bias is not conclusive.

The implications of her results

45
'S. R. Hyman, "The Miller Analogies Test and Uni­
versity of Pittsburgh Ph.D.s in Psychology," American
Psychologist, Vol. 12 (1957), pp. 35-36.
46

B. G. Fricke, "Prediction, Selection, Morality,
and Quality Control," College and University, Vol. 32 (1956),
pp. 34-52.
47

E. E. Cureton, Louise W. Cureton and Ruth Bishop,
"Prediction of Success in Graduate Study of Psychology at
the University of Tennessee," American Psychologist, Vol. 4
(1949), pp. 361-362.
48

T. A. Cleary, Test Bias: Validity of the SAT for
Negro and White Students~ln Integrated Colleges (Princeton,
New Jersey:
Educational Testing Service, Research and
Development Report 65-66, No. 18, 1966).

36
are so weighty that additional similar research is required.
Clark and Plotkin concluded from their study of over 1500
black students at integrated colleges that SAT scores cannot
be used to predict academic success for black students as
they can for whites.

They found that, while the median

score on the SAT is lower for blacks than whites, a higher
proportion of blacks than whites who begin college obtain
degrees.^
If there is disagreement to this extent over the
value of standardized testing, is there any reason to
believe that these questionable instruments should be used
to determine the fate of poor w h i t e s , blacks and other
minority disadvantaged students?
American educational institutions reflect racial
and class distinctions, even though those responsible often
claim that they are seeking to eradicate such distinctions.
The public schools remain, for the most part, segregated
by race, and many of the black schools as well as other
minority schools have the poorest facilities, the highest
teacher turnover, and the least support.

51

The existence

of these distinctions explains the fact that the quality
49

Miller and O'Connor, op. cit., p. 104.

50

K. B. Clark and L. Plotkin, The Negro Student at
Integrated Colleges (New York:
National Scholarship Service
and Fund for Negro Students, 1963).
"^Howe, Clark, Allen, et al. Racism and American
Education:
A Dialogue and Agenda For Action (New York:
Harper & R o w , 1970).

49

37

of education provided for American children differs ac­
cording to their racial, economic, and social class status.
Therefore, they assert, children of lower-class status
generally perform academically at a level lower than children
of middle-and upper-class status.

Such distinctions not

only include flagrant damage to black children as well as
other minority youth but also insidious and subtle damage
to the whrte youth.

52

These inequalities must be removed

from our educational system, and the causes of these
inadequalities can only be removed if the old traditional
rules are changed.
Furthermore, there is little question in anyone's
mind that the average disadvantaged black high school
student compares unfavorably with the average middle class
white student on test of academic ability.

Consequently,

the blind application of criteria in college admissions
will result, naturally, in proportionately fewer blacks
than whites being admitted to the more selective institu­
tions .
A point frequently overlooked in the use of high
school grades and aptitude tests for predicting achievement
in college is that students from different environments
score differently, that disadvantaged students score dif­
ferently, and that motivation has a pronounced effect on the
results obtained.
52Ibid., p. 152.

38

Prediction studies have generally neglected the nature
of the sample— the characteristics of the institution and the
characteristics of the student.
gests

53

Research literature sug-

several factors that might be related to predicta­

bility, including range of talent and personality character­
istics.

In a preliminary study of institutional factors

that influence predictability,

54

the following variables

were found to account for 42 per cent of the variation in
predictability:
freshmen class;
body;

(1) Range of talent;

(2) Size of new

(3) Ability level of freshmen student

(4) Socio-economic background and motivation; and

(5) Motivation.
Another commonly overlooked point in the use of
high school grades and aptitude test scores for prediction
of achievement in college is that such predictions are
subject to a considerable amount of error.

Not all of the

most promising students succeed; neither do all of the
least promising ones fail.
Ernest Lake states candidly his position on the
use of standardized test.
The multivarious requirements for testing students
continue to compound the clerical and administrative
duties of our counseling departments.
These
53

Leo A. M u n d a y , "Factors Influencing the Prediction
of College Grades," Paper presented at Annual Meeting of
National Council on Measurement in Education, Chicago,
Illinois, February 10, 1968.
54
(New York:

Warren W. Willingham, Free-Access Higher Education
College Entrance Exam Board, 1970).

39
requirements have become more odious than the essay
contests sponsored by prestige seeking organizations.
We're establishing an elite, selected by not too
reliable testing practices comparable to the '400'
of society. . . . This is the time of year when school
administrators and high school counselors compare notes
on their record of wins and losses in the admissions
race.55
Numerous studies have shown that test involving
self-concept of academic ability are often more important
in predicting success than are intelligence tests them­
selves, and aptitude tests are frequently not valid pre­
dictions of academic success because they depend too much
on an abundant environment and on whether a student has
had enough experience to be test-proficient rather than
test-blind.^
One of the most encouraging studies relative to the
use of criteria other than standardized test to predict
academic success has been conducted by

Miller

and O'Connor

They argue that whether or not findings prove the SAT to
be a valid and unbiased predictor of academic performance
for blacks there is another very important consideration.
It is the fact that scores obtained on the SAT typically
account for less than twenty-five per cent of the variation
in college g r a d e s . ^
55

Ernest G. Lake, "The Case Against External Stand­
ardized Tests:
They Create an Elite," The Nation's Schools
70 (August, 1962), p. 51.
Robert L. Green, "The Black Quest for Higher Edu­
cation:
An Admission Dilemma," The Personnel and Guidance
Journal, 47 (May 1969), pp. 906-908.
57

Miller and O'Connor, op. cit . , p. 105.

40
In an attempt to explore the possible value of
motivation as a predictor they examined the Achiever
Personality scale, constructed by Benno G. Fricke, as one
of the scales comprising the Opinion Attitude Interest
Survey

(OAIS).

The OAIS is described in the test hand­

book by Fricke as follows:
This scale measures personality attributes associated
with the traditional criterion of academic success,
grades.
Students who score high on the Achiever
Personality scale tend to realize their potential
ability and/or achieve high grade point averages in
college.
It predicts college grades about as well
as the typical academic ability test.
Furthermore,
and this is important, scores from the Achiever
Personality scale do not correlate with scores from
the ability test; that is, this scale measures some­
thing important in academic success not measured by
the ability test.
In short, the Achiever Personality
scale is a good indicator of academic motivation and
conscientiousness.5 8
The results of two studies concerning the use of
Achiever Personality scale are reported.

Subjects in the

initial longitudinal study are the Opportunity Award
Program students who entered the University of Michigan
as freshmen in 1964 and 1965.
considered disadvantaged.
and 59 men.

All of the subjects were

Data are reported for 70 women

Academic success was determined by the subjects'

ability to obtain a 2.00 grade point average
end of the sophomore year.

(GPA) at the

The other measure of success

was the actual continuation as juniors for the 1965 entrants
58
B. G. Fricke, Opinion Attitude and Interest Survey
Handbook (Ann Arbor, Michigan: Michigan Oais Testing Program, HT65) .

41
or seniors for 19 64 entrants.

The second study attempted

to replicate the finds of the original study.
consisted of 46 women and 43 males.

The subjects

The single measure of

success was a 2.00 GPA in the freshman year.
In both studies the independent variables w e r e :
(1) SAT scores;

(2) High school precentile rank (HSPR);

(3) Scores on the Achiever Personality scale
(4) An index of high school quality.

(Ach P ) ; and

The chi-square test

was the method of analysis for both studies.

All tests

were two-tailed with one degree of freedom and the Fisher
Exact Probability Test was used when the expected fre­
quency

in any cell was less than
The results of the

first

.10.
study showed that high

school rank was not associated with grades for males or
females.

The correlation between high school rank and

high school quality was -.29 for females and -.36 for
males,

both significant at P<.05.
For the men in the

first

study nosignificant

relationship was found between high school quality and
grades.

SAT scores were found to be an ineffective

predictor of success.

Of those men with high SAT score

57 per cent achieved passing grades, but so did fifty-five
per cent of those with low SAT scores.

For the men with

low SAT scores, a high score on the AchP scale was sig­
nificantly related to academic success

(P = .04).

AchP

was not related to academic success for men with high SAT

42
scores.

For men with low SAT scores AchP was significantly

related, to continuation as junior and seniors P = .05.
For the women in the first study no significant
relationship was found between high school quality, SAT
scores and grades.

A significant relationship was found

for women between AchP and grades, Pc.001.

The relation­

ship between AchP and the eligibility of the females to
continue as upper-classmen was reported as significant at
P<.001.
The results from the second study showed that for
neither males nor females was the SAT score, high school
quality or AchP significantly related to freshmen grades.
Thus, the second study failed to replicate the findings of
the first.

The possible explanation for this failure may

have been related to the change in counseling practice
over the two studies according to Miller and O'Connor.

59

It seems that the study reported above may have a
great deal of potential as researchers continue in their
efforts to perfect a tool to use in place of the culturally
biased standardized test.

Further research is certainly

warranted in this area.
The final part of this review will be directed
toward a description of some of the developmental programs
currently in operation at institutions of higher learning.
59

Miller and O'Connor, op. cit.

43
The programs described will be limited to four which seem
to show promise in developmental education.
Developmental Programs Currently in Existence
One of the most recent and detailed reviews of
developmental programs at the community college level has
been conducted by the Southern Regional Education Board.
Through its Institute for Higher Educational Opportunity
it is currently engaged in a three-year project to increase
opportunity for black students through the community
college.

The reviews which follow have been taken from

the 19 70 Project Report.^®
In 1969, Gulf Coast Junior College designed the
"Self Concept Institute."

The program is based upon the

assumption that disadvantaged students who are exposed to
experiences which help them understand themselves and the
forces which contribute to their success or failure will
develop behavior patterns conducive to their academic
success.
Disadvantaged students are selected through inter­
view with the Institute's staff.

The students participate

in a two-week Self-Concept Institute on the college campus.
The students are racially mixed and are given $15.00 per
week for room and board.

Each student is assigned a room

on or near campus and a roommate.
60

The Black Community and the Community College— A
Project Report (Atlanta, Georgia;
Southern Regional Edu­
cation Bo a r d , 1970).

44
Activities for the two-week period include class­
room exercises, group guidance, social activities, field
trips and guest lecturers.

Evaluation of the Institute is

based on evaluations submitted by the instructors and stu­
dents regarding the value of the experiences.

In addition,

the Tennessee Self Concept Scale (counseling form) is
administered to the students on the first day of the
institute and again on the final day.

Changes in the

self-concept profiles of the students are considered
indicative of change in se l f - e s t e e m . ^

At this writing

data w e r e not available on the outcome of the education.
Spartanburg Junior College's program is based on
the following assumptions:
1.

The key to academic success is in the discovery
and concentration on individual strength.

2.

No two students are academically alike, there­
fore, the gap of academic deficiency must take
into account individualism.

3.

Testing and guidance should complement academic
work undertaken.

4.

Individual adjustment and maturity can indicate
academic success.

5.

Personalized instruction, using the latest
technology, is the key to an intensified learning
experience.

^ I b i d . , p. 17.

45
A program built on those assumptions was conducted
for eight weeks during the Summer of 1968, for 64 students.
In addition to cultural, reading groups, independent study
each student was involved in 320 hours of classroom study.
The classroom study included 110 hours of English, 80 hours
of Math,

80 hours of Reading Lab, and 50 hours of Group

Dynamics seminar.
The overall achievement progress for each student
was of 1.4 grades.

Fifty-six of the 46 students were accepted

at the college the following September and 6 returned to
high school level work.

ft

?

An important factor in the success of the program
is the absence of any stigma being placed on the students
who participated in the program.

The freedom from stigma

contrast with situations of embarrassment noted by many
students enrolled in similar programs during the regular
term.

63
The Basic Studies Program at Tarrant County Junior

College has some very promising innovations.

The one year

program consists of six areas of study in a unit called
"Who am I."

The areas of study are natural science,

humanities, communications, reading improvement, social
science and personality foundation.
6 2 T, .j

Ibid.

63Ibid., p. 19.

In Natural Science

46
the student studies his physical and biological self.

In

Humanities he studies religious and philosophical beliefs
as they relate to his values and attitudes.

Communications

centers around improving methods of communications with
others.

In Reading Improvement the student analyzes and

improves basic reading habits.

Social Science deals with

the socialization process'and the student's relationship
to others.

The student's own personality development is

studied m

Personality Foundations.

64

The program operates in an integrated teaching
approach in which a team of six instructors provide general
education for 100 students.

Approximately twenty students

are placed in each section and attend their classes as a
unit.

Five sections are assigned to a team of six instruc­

tors.

Each instructor covers one of the six areas of

program.

the

The teams meet weekly to plan curriculum and to

discuss individual learning problems of their students.
This approach affords the students an opportunity to
establish strong peer relationships and provides the
faculty a chance to get to know the students on a personal
basis.

Upon the termination of the one year Basic Studies

Program the

student is placed in one of the following areas

1.University Parallel Transfer
2.

Program

Technical or Occupational Associate Degree
Program
/r a

Success Breeds Success, Basic Studies, 1970-71
Report (Fort Worth, Te x a s : Tarrant County Junior College
District, 1971), p. 5.

47
3.

Non-degree Vocational Program

4.

On-the-Job Training Program.

65

The key to the success of the Basic Studies Program
is the faculty.

In general, the faculty members are people

who:
1.

See themselves and function as counselors as well
as teachers

2.

Who are more interested in individual progress
than with predetermined content that must be
covered in a course

3.

Who can communicate with students, who have a sense
of humor and can poke fun at students and themselves

4.

Who can become human development facilitators—
people who are open to experience, democratic,
understanding, caring, non-judgmental, selfactualizing

5.

Teachers who will establish behavioral objectives
for their courses, who will communicate these
objectives to their students and be overjoyed to
find that they have all 'A's' in a class rather
than the normal curve

6.

Finally, teachers who realize that each student is
a gifted person and that student motivation comes
about as a result of success and will let success
build upon success.66
The thrust of the program is toward a more humanistic

approach to higher learning.

One primary goal is to help the

students progress toward a realistic and positive self-image.
The results of a recent study show that of the 274 students
who were enrolled during the 19 70-19 71 academic year,
^ Ibid. , p. 9.
^ I b i d . , p. 10.

81 per

48
cent were in good academic standing at the end of the Spring
semester.

The average grade point average of the unit was

2.43 and only 22 per cent received averages of 1.99 and
below.^
Gains in the student's self-concept are measured
by the Tennessee Self-Concept Test.

The test has eight

sub-scores including general identity, physical self,
personal self, family self and an overall self analysis
composite score.

The composite score was used to measure

any change in self concept.

From a sample of 132 students

who received pre, middle and post test a 17 percentile gain
was reported, which is statistically significant.

68

It

should be noted that while no control group was used in
this study, the results seem very promising.

The most

important factors related to the apparent overall success
of the program seem to be related to the generalized
humanistic approach to education.
Miami-Dade Junior College has developed a compen­
satory educational program at its North campus.

The

students are enrolled in special courses designed to meet
the needs of the disadvantaged.

The students receive full

credit for the courses as the objectives for special courses
in English, Social Science and Natural Science are the same
^ I b i d . , pp. 27-28.
^ I b i d . , pp. 35-36.

49

as the objectives for the standard college-level courses.
The courses are not watered-down preparatory courses.
Remedial learning takes place through additional hours in
classrooms and laboratories, through individualized instruc­
tion and tutoring and through diagnostic testing.

The team

approach is utilized in that each group is assigned to a
team consisting of a reading specialist, a counselor and
instructors of English, Social Science and Natural Science.
Preliminary results indicate that the drop-out rate for
these students is lower than for students in other kinds
of remedial programs.

In addition, the concentration

counseling approach has resulted in a higher and more
realistic level of aspiration for the special program
69
students.
The preceeding selection of developmental programs
is by no means intended to be a complete representation of
practices.

It was intended only as a brief overview of

what appears to be successful at four selected community
colleges.

For a detailed discussion of compensatory pro­

grams, the reader is directed to Gordon,^® T r e n t , ^
69

The Black Community and the Community College,
op. cit., p. 25.
70
71

Gordon and Wilkerson, op. cit.
Trent, op. cit.

50
Southern Regional Education Boards 19 70 Project Report,
Goodrich,

73

Johnson,

74

and Moore.

72

75

Summary
The review of related research has been presented
in three sections:

Interest in Developmental Education;

Studies Relating to Academic Achievement and the presenta­
tion of Currently Existing Programs.
The disadvantaged student has been defined as the
financially poor student whose low prognosis for success
is conteracted by a sincere desire to learn.
For the most part a great deal of emphasis has been
placed upon the importance of expanding the opportunity for
disadvantaged students to obtain an education.

Community

colleges appear to have taken the lead in addressing them­
selves to meeting the higher learning needs of disadvantaged
students.

Most writers agree that traditional teaching

methods are ineffective with disadvantaged students and
much emphasis has been placed upon program learning,
tutoring and special counseling.
72

The Black Community and the Community College,

op. cit.
73

Goodrich, op. cit.

^ J o h n s o n , op. cit.
75

M o o r e , op. cit.

51
While the evidence on the value of standardized
test is conflicting, most of the writers seem to agree that
self-concept of academic ability is often more important
in predicting success than standardized test.

Aptitude

test are frequently not valid predictors of academic success
because they depend too much on environmental factors.
While the evidence is inconclusive, one of the most promising
measures of academic success appears to be the Achiever
Personality scale developed by Frick.

The results of the

use of this test has been presented.
Four seemingly promising programs were presented
in the last section of the review.

The components of a

high risk program may be summarized as having, in varying
degrees, flexible admission criteria, special recruitment
procedures, innovative teaching approaches, provisions for
financial aid, provision of a pre-college experience, and
personal and academic counseling.

CHAPTER III

DESIGN OF THE STUDY
Definition of the Population
Institutions of higher learning vary in size, loca­
tion and curriculum offerings.

Consequently, no institution

may be considered as a typical university.

For this reason

it is necessary to describe Western Michigan University and
its student body as this is the population from which the
samples for this study were drawn.
Western Michigan University is a four year institu­
tion located in central Kalamazoo, Michigan.

Among

Michigan's institutions of higher education, Western ranks
fourth in number of students, diversity, complexity and
level of programs.

Western was created in 1904, with 117

students and three faculty members, for the purpose of
educating teachers.

Enrollment in the Fall of 19 71, was

21,846, and the faculty totaled 1,173.
The original concept of the school has been expanded
and Western is now considered a multi-purpose university.
The first bachelor's degrees were authorized in 1819.
Graduate work was offered in 19 39, and doctoral programs
have been offered since 19 66.

There are forty-three

52

53
departments in the colleges of Arts and Science.

Bachelor's

degrees are offered in Arts, Business Administration, Music,
Science and Industrial Engineering.
leading to a master's degree.

There are 55 programs

Advanced graduate degrees

are Specialist in Arts; Specialist in Education; Doctor
of Philosophy in Chemistry, Mathematics, Science, Education
and Sociology.
The present enrollment of Western is more than
triple the 6,857 enrollment of 1957, the year Western was
given university status.

In the last ten years the enroll­

ment has more than doubled the 1961 enrollment of 9,545.
Of the present 21,846 enrollment; 11,389 of the students
are from southwestern Michigan;

8,206 are from southeastern

Michigan; 531 are from northern lower Michigan; 211 are
from the Upper Peninsula; 1,2 40 are from other states; and
269 are from foreign countries.
Western operates year-round programs with Fall and
Winter semesters, Spring and Summer sessions.
sessions combined equal one semester.

The two

Continuous full­

time attendance at all semesters and sessions would permit
completing a bachelor's degree program in two years and
eight months.
In its efforts to recognize students who have been
denied an education because of their cultural, economic and
education environments, Western has initiated the Martin
Luther King Junior Program.

The program involves talent

54
search, recruitment, orientation and support for students
of marginal high school academic records.

Continuous

counseling and tutoring services are provided.

In addition,

the university is the sponsoring agency for an Upward Bound
project, a program designed to motivate certain high school
students to continue their education beyond high school.1
Admission
Western Michigan University currently, considers
a satisfactory high school or transfer record as the most
reliable single factor used to predict college success.
The college preparatory subjects are given maximum weight
in the admission.

In 19 70, the minimum high school grade

point average required for admission was 2.20 for males
and 2.50 for females.

However, this requirement was

changed in 19 71, to a 2.00 for all students.
The American College Testing Program (ACT) is
required of all new freshmen.

Applicants with marginal

records are required to submit test results before a final
decision is reached relative to their admission status.
General Requirements
(1) Students are required to complete a major with
a minimum of 24 hours and a minor with a minimum of 15
hours.

(2) Each student must complete 40 hours of work

W e s t e r n Michigan University Undergraduate Catalogue,
1969-1970.

55
in the General Studies Program or in approved alternatives
(see Appendix).

(3) A minimum grade point average of 2.00

must be attained in any major or minor presented for gradua­
tion .
Honor Points
The number of honor points earned in a course is
the number of semester hours credit given by the course
multiplied by the number of honor points per hour of credit
corresponding to the letter grade received, as shown in
Table 3.1.

For example, a grade of B in a four hour course

gives 4 x 3 or 12 honor points.
Grading System
The unit of credit is the semester hour.

Grades

are indicated by letters, to each of which is assigned a
certain value in honor points per hour of credit as shown
in Table 3.1.
TABLE 3.1.— Grading system.
Grade

A
B
C
D
E
I
CR
NC

Significance

Excellent
Good
Fair
Passing
Failure
Incomplete
Credit
No Credit

Honor Points
per hour Credit
4
3
2
1
0
0
0

56
Grade Point Average
A grade point average is obtained by dividing the
total number of honor points earned by the total number of
semester hours of work for which the student is officially
enrolled during any period.

For example, a total of 32

honor points earned in a semester by students enrolled
for 16 hours of work gives a grade point average of 32/16
or 2.00 for that semester.
Scholarship Standards
A student must earn an overall grade point average
of at least 2.00 to satsify degree requirements.

The

policy operates as follows:
1.
2.

3.

4.
5.

6.

Good Standing
A student is in good standing whenever his
overall grade point average is 2.00 or above.
Warning
Whenever the grade point average for any
enrollment period is less than 2.00, but the
overall grade point average is 2.00 or above,
the student is warned.
Probation
The student will be placed on probation
whenever his overall grade point average falls
below 2.00.
Probation Removed
Whenever the conditions of good standing are
restored, probation will be removed.
Continued probation
If the overall grade point average increases,
although still below 2.00, the student may be
continued on probation for one additional enroll­
ment period.
Dismissal
The student who fails to increase his overall
grade point average at the end of an enrollment
period of probation, or whose overall grade point

57
average fails to reach 2.00 at the end of one
enrollment period of continued probation, will
be asked to withdraw from t h e u n i v e r s i t y .2
Nature of the Sample for Developmental
Group I
From the population described above, three groups
of students were selected as the basis for this study.

The

participants of developmental group I consisted of the
entire group of Martin Luther King Jr. students.

The

students were selected to participate in the program by
the program director on the basis of recommendations from
non-school related persons, from financial need, and
personal interviews.

Only those students were selected

who demonstrated financial need and academic promise but
who had low grades and/or low ACT scores.

Twenty nine

females and thirty three males were admitted.

All of

these students had been recruited during the previous
Winter and Spring sessions from six high schools:

Albion,

Battle Creek Central, Benton Harbor, Covert, Kalamazoo
Central, and Muskegon Heights.

The recruitment procedure

involved person-to-person contact with every student who
indicated a desire and interest in attending college.
These visits were made by the program director, his
assistant and selected Western Michigan University students
who had previously attended the high schools.
2Ibid.

Factual

58
information about Western Michigan University and admission
applications were disseminated to interested students.
High school records were also reviewed.

Moreover, recruit­

ment procedures involved personal interviews with students,
teachers and other non-school related people such as parents,
community leaders, ministers and alumni to obtain information
about the students that was not readily available via the
routine institutional channels.

This latter approach

provides additional insight and complementary knowledge
about the student which assisted the program director in
reaching a decision as to whether or not a particular
student should be accepted.

Interested students were

invited to spend a week-end on Western Michigan University's
campus.

While on campus the students were exposed to such

university resources as the library; industrial engineering
and technology; music; student center and other curricular
programs that would have a motivational appeal.
After the selection process was complete sixty-two
students were accepted and financial packages were developed
for each student demonstrating need in the amount of
approximately $450.00

(see Appendix).

All students were

housed in the campus dormitory, two to a room, and were
served three meals a day.

Each of the sixty-two students

attended a summer orientation session lasting from June 21,
1969 through August 31, 1969.

Each student took a total of

59
six hours credit; two hours of Adult Reading and four
hours of Race and Culture.

Adult Reading was provided

by the Psycho-Educational Clinic and emphasized instruc­
tional and developmental procedures for improving the
students'

reading skills.

Race and Culture was offered

by the School of General Studies.

The purpose of this

course was to introduce the student to the tools necessary
for gaining and testing knowledge in the Social Science.
The emphasis of the course was on; Race and Culture
contacts; social stratification; structure of the minority
group community; race and status; history of minority groups;
the nature of prejudice; patterns of discrimination and
its cause; race relations; and current problems related
to race and culture.

Both of the above courses met five

times a week during the entire summer session.
In addition to the above credit courses the students
also took non-credit courses.

These courses included:

library, Science, Speech, Study Center, and Physical
Education.
Library Science
This course acquainted the students with the
library.

The course met for two hours one day per week

throughout the summer orientation.

The students were

given special instructions on the physical arrangements
of the library and assignments geared toward learning basic
reference tools and improving library skills.

60
Speech
This course met two nights each week for two hours
each night.

The course was geared toward discussion groups

and an introduction to radio and television media.
activities included:

The

involvement in the campus radio

studio, films; discussion groups; art, music, and other
creative student projects; field trips and role playing
to emphasize self-exploration.
Study Center
The primary purpose of the Study Center is to
provide social and academic support for the developmental
program students.

The center was located in the library

and remedial and tutorial assistance was provided by
Senior Student Aides.

During the Summer, the students

attended the study daily for two and one half hours each
day and during the Fall and Winter semester the students
attended two days per week for two and one half hours
each day.
Physical Education and Recreation
All of the Developmental Group members participated
in Physical Education for two hours each day, five days
per week during the Summer orientation.

The purpose of

the course is to provide physical fitness and recreational
activities to meet the mental, social and leisure needs of
the students.

Instruction and supervision was provided in

61
swimming, gymnastics, tennis, golf, dancing, softball, and
basketball.
In addition to the educational programs described
above, cultural enrichment programs were provided by means
of weekend trips, speakers, and entertainment.

An all-day

Saturday trip to Chicago included visits to the University
of Chicago, Operation Breadbasket, and the Museum of
National History.

A similar trip was made to Detroit and

included a visit to the Oakland University Campus.

Speakers

were obtained from Western Michigan University's Religion
Department, Admissions Office and Student Services.
During the Fall and Winter semesters the students
received counseling and guidance in the selection of their
courses.

Tutorial programs were provided for those stu­

dents who had difficulty with their academic assignments.
In addition, all of the students spent a minimum of two
days per w e e k , two hours each of the two d a y s , in the
Study Center under the supervision of a student counselor
aide.

The academic progress of each student was closely

scrutinized by student-aid counselors who were assigned to
groups of eight students.

Problematic situations which

included financial, academic, social and personal problems,
were discussed with the student counselor aides.
Students in need of financial assistance were
provided with financial packages of approximately $1,800
(see Appendix).

62
Nature of the Sample for Control
Group II
The sample for Control Group II was selected by
using a table of random numbers.

Twenty nine females were

selected from a population of ninety-eight females and
thirty-three males were selected from a population of
seventy-nine males.
average

The mean high school grade point

(GPA^g ) of the group from which the females were

selected ranged from 1.86 to 2.49 with a mean of 2.30.
The GPAhs of the group from which the males were selected
ranged from 1.72 to 2.19 with a mean of 2.03.

The popula­

tion from which the sample was selected were admitted as
beginning freshmen during the Fall semester 19 70, without
benefit of any special program or counseling.
Nature of the Sample for Control
Group III
The population from which the membership of Control
Group III was selected consisted of fifty-two females having
GPAhs of exactly 2.50 and fifteen males having a GPA^s of
exactly 2.20.

Twenty-nine females were selected, using a

table of random numbers, from the group of fifty-two females
and the entire population of fifteen males with GPA^g of
exactly 2.20 were selected.

The population from which

Control Group III was selected were admitted as beginning
freshmen during the Fall semester 1970, without benefit
of any special program or counseling.

TABLE 3.2.— Summary of the Characteristics of the Population:
Group

GPAhs anc^ Sex.

N
Females

GPAhs
Range

GPAhs
Mean

N
Males

GPAhs
Range

GPAhs
Mean

N
Total

Dev. I

29

1.41-1.63

2.03

33

1.38-2.35

1.93

62

Cont. II

98

1.84-2.49

2.27

79

1.32-2.19

2.03

177

Cont.III

52

2.50

15

2.20

67

64
Table 3.3 shows the high school grade point averages
of all beginning freshmen at Western Michigan University in
19 70.

Foreign, guest and unclassified students are not

included in this table since these groups present records
not typical of other freshmen entering the University.

The

mean high school grade point average for the group of females
was 2.9 8 and the mean high school grade point average for the
group of males was 2.65.

The mean high school grade point

average for all students who were beginning freshmen in
19 70 was 2.83.
TABLE 3.3.— Summary of High School Grade Point Averages of
All Beginning Students Entering the
University in 19 70.a
GPAhs

Number
Male

Number
Female

Total

Per Cent In
Each Interval

3.50-4.00

68

278

346

10. 4%

3.00-3.49

235

586

821

24.6%

2.50-2.99

612

854

1466

43.9%

2.00-2.49

532

122

654

19.6%

0.00-1.99

36

15

51

1.5%

No G.P.A.

16

76

92

TOTAL

1499

1931

3430

Mean GPAfts

2.65

2.98

2. 83

aGabier, Russell, "Freshmen-Transfer Profile 19 70,"
Office of Admissions, Western Michigan University, p. 8.

65

The high school grade point averages of the sample
by group and by sex is illustrated in Table 3.4.

The mean

GPA^s of the females in Developmental Group I is 2.0 3 and
the range is 1.41 to 2.63.

The mean G P A ^

of the group of

62 Developmental students is 1.9 8.
The 29

females in Control Group II

GPA^sof 2.30 and a range

of 1.86 to 2.49.

have a mean
The 33 males

in Control Group II have a range of 1.72 to 2.19 and a
mean of 2.05.

The mean

of the 62 Control Group II

members is 2.17.
The 29
GPA,
hs

of 2.50.

females in Control Group III have a mean
Since all the females have the same GPA, ,
hs

there is no range.

The mean GPAhg for the males is 2.20,

with no range as all members have the same GPA^g .
GPA^

The mean

of the 44 members in Control Group III is 2.39.
Measures

Data Collected
The data listed in Table 3.5 was collected, when
available, on each student in each group.
The Summer GPA was collected for Developmental
Group I only, since Control Group II and Control Group III
did not enroll until the following Fall.

For statistical

analysis, the Summer GPA and Summer hours passed was not
used in computing the cumulative GPA or the cumulative

TABLE 3.4.— Summary of the Characteristics of the Sample:

Group

GPAhs and Sex.

N
Females

GPAhs
Range

^PAhs
Mean

N
Males

GPAhs
Range

GPAhs
Mean

Group
Mean

N
Total

Dev. I

29

1.41-2.63

2.03

33

1.38-2.35

1.93

1.98

62

Cont. II

29

1.86-2.49

2.30

33

1.72-2.19

2.05

2.17

62

2.20

2.39

44

Cont.III

29

2.50

15

67
TABLE 3.5.— Data Collected.
Data Collected
Sex
Month--Year of Birth
ACT Scores:

English, Math, Social Science, Natural Science
Comp. Score

Hours Passed in Summer
Grade Point Average in Summer
Hours Attempted in Fall
Hours Passed in Fall
Grade Point Average in Fall
Hours Attempted in Winter
Hours Passed in Winter
Grade Point Average in Winter
Total Hours Attempted Fall + Winter
Total Hours Passed Fall + Winter
Cumulative Grade Point Average Fall + Winter
High iSchool Grade Point Average

(GPAC )

(GPA^g)

Predicted GPA
Enrollment status as of Fall, 19 71 (Drop-out Rate)

hours earned because of the possible influence of group
interaction.

With this exception, the cumulative grade

point average up to and including the final semester of each
students attendance was used for all three groups.
All data were collected from the individual student's
records at the Office of Admissions and Records at Western
Michigan University.

American College Testing Program scores

and predicted GPA was not available for nineteen students in
Developmental Group I and for seven students in Control

68
Group II.

Complete data were available for all of the

members of Control Group III.
American College Testing Program (ACT)
Two types of measures are used:

"measures of

academic potential and measures of college achievement.
Several measures of academic potential termed 'standard
ACT predictors' are used for all colleges."

3

They included

four ACT test standard scores and recent high school grades
in English, Mathematics, Social Studies and Natural Science.
The ACT Battery consist of four test:
Social Studies and Natural Science.
standard score of from 1 to 36.

English, Mathematics,
Each test has a possible

"The statistical technique

of multiple regression analysis considers the correlations
of each test score with college GPA and with each other;
it produces an optimum set of weights to be applied to each
4
test score."
The multiple correlation of the four test
are then correlated with the high school grades in English,
Mathematics, Social Studies and Natural Science to produce
a predicted college grade point average.
The predicted grade point average GPA'pwas compared
with the actual GPA

c

at the end of two semesters to deter-

mine if there was any descrepancy on this measure.

The

"^Donald P. Hout and Leo A. Munday, Interpretive
Guide to ACT Research Service for Higher Education (Iowa
City, Iowa:
The American College Testing Program, 1969),
p. 11.
4Ibid. , p . 40.

69
multiple correlation of the four test and high school grades
with college GPAC is .523.

The standard error of estimate

for five variable prediction is .517.
5
for predicted GPA^ is .622.

The standard deviation

The means for each of the ACT test subparts, by sex,
are shown in Table 3.6.
TABLE 3.6.— ACT Standard Scores for All 19 70 Beginning
Freshmen at WMU Test Subpartsa Means (X).
N

English
X

Math
X

Soc.Sci.
X

N a t .S c i .
X

Comp.
X

M

1374

19.0

23.6

22.5

23.5

22.3

F

1759

20.6

20. 7

20. 8

21.4

21.0

Total

3133

19 .9

22.0

21.6

22.3

21.6

aGabier, Russell, "Freshmen-Transfer Profile 1970,"
Office of Admissions, Western Michigan University, p. 10.
The ACT test was administered by the testing depart­
ment at Western Michigan University to Developmental Group I
in June, 1969 and to Control Groups II and III in September,
19 70.

The results of these test are shown in Chapter IV.

Cumulative GPAc and Enrollment Status
The cumulative GPAC was used to compare the GPAc
of Development Group I with the GPAC of Control Group II
5

"Summary Analysis, 1969 Standard Research Service
of the American College Testing Program" (Iowa City, Iowa:
The American College Testing Program, 1969).

70
and the GPAc of Control Group III.

The effectiveness of

Developmental Program was determined, in part, by this
measure.
The enrollment status is a measure used to deter­
mine the number of drop-outs in each group.

Any student

in any of the three groups who was not enrolled at the
beginning of the Fall Semester, 19 71, is considered a
drop-out.

The rate of persistence is determined by compar­

ing the percentages of drop-outs in Developmental Group I
with the percentages of Drop-outs in Control Group II and
Control Group III.
Testable Hypotheses
The three major hypotheses stated in Chapter I are
restated below in testable form.

Each of the following

hypotheses were tested using three repeated measures:
(1) Grade point average;

(2) Drop-out rate; and (3) The

discrepency between the actual GPAc and predicted GPA'p.
1.0.0 Null Hypothesis:
There is no significant difference
between groups on GPAC drop-out rate, and the discrepancy
between actual and predicted GPA.
1.1.0:

There is no significant difference between
groups on GPAC .

1.2.0:

There is no significant difference between
groups on drop-out rate.

1.3.0:

There is no significant difference between
groups on the discrepancy between actual
and predicted GPA.

71
2.0.0 Null Hypothesis:
There is no significant difference
between genders on GPAC , drop-out rate, and the discrepancy
between actual and predicted GPA.
2.1.0:

There is no significant difference between
genders on GPAC .

2.2.0:

There is no significant difference between
genders on drop-out rate.

2.3.0:

There is no significant difference between
genders on the discrepancy between the actual
and predicted GPA.

3.0.0 Null Hypothesis:
There is no significant interaction
between groups and genders on GPAC , drop-out rate and the
discrepancy between the actual and predicted GPA.
3.1.0:

There is no significant interaction between
groups and genders on GPAC .

3.2.0:

There is no significant interaction between
groups and genders on drop-out rate.

3.3.0:

There is no significant interaction between
groups and genders on the discrepancy between
actual and predicted GPA.

Two designs were used in this study:

the first

design, Figure 3.1, excluded all subjects with missing data
and attempted to answer three basic questions:

(1) Is

there any significant difference between Developmental
Group I, Control Group II and Control Group III on the
college GPAc ? on the drop-out rate; and on the discrepancy
between the attained college GPAc and the ACT predicted
college GPA ;

(2) Is there any significant difference

between the genders on college G P A c ; on the drop-out rate;
and on the discrepancy between the attained college GPA
and the ACT predicted college GPA ;

(3) Is there any

c

72

Males Females

Developmental
n = 22
Group I

n = 21

Control
Group II

n = 28

n = 27

Control
Group III

n = 15

n = 29

Figure 3.1.--First Design
significant interaction between groups and genders on college
GPAc ; on drop-out rate; and on the discrepancy between the
attained college GPA
3 x 2 x 3

and the ACT predicted GPA.

The

block design illustrated in Figure 3.1 was utilized

in an attempt to answer these questions.
The diagram in Figure 3.1 illustrated that the
subjects are nested with the Group-gender interaction.
this design the three groups

In

(Development Group I, Control

Group II and Control Group III)

are measured by two genders

(male and female), by groups and gender interaction by three
repeated measures.
The second design, Figure 3.2, was developed to
determine whether the results were biased by the lack of

73

R1

R2
No
Missing
Missing
Data
Data
Developmental
Group I

Control
Group II

N = 19 N = 43

N = 7

N = 55

R^ = Response Group 1
R 2 = Response Group 2

Figure 3.2.— Second Design.
predicted GPA's on the 19 subjects in Developmental Group I
and on 7 subjects in Control Group II.

Since none of the

subjects in Control Group III had missing data, Group III
was not included in this design.
Three basic questions were explored in the second
design.

(1) Is there any significant difference between

Developmental Group I and Control Group II on GPA
Drop-out rate?

and

(2) Is there any significant difference

between Response Group I (missing data group)

and Response

Group II (no missing data group) on GPAc and drop-out rate?
(3) Is there any significant interaction between treatment
groups and response groups on GPAc and drop-out rate?

74

The following hypothesis were tested to answer the
above questions:
4.0.0 Null Hypothesis:
There is no significant difference
between treatment groups on GPAC and drop-out rate.
4.1.0:

There is no significant difference between
treatment groups on GPAC .

4.2.0:

There is no significant difference between
treatment groups on drop-out rate.

5.0.0 Null Hypothesis:
There is no significant difference
between response groups on GPAC and drop-out rate.
5.1.0:

There is no significant difference between
response groups on GPAC .

5.2.0:

There is no significant difference between
response groups on drop-out rate.

6.0.0 Null Hypothesis:
There is no significant interaction
between treatment groups and response groups on GPAC and
drop-out r ate.
6.1.0:

There is no significant interaction between
treatment groups and response groups on GPAC .

6.2.0:

There is no significant interaction between
treatment groups and response groups on drop­
out rate.
Methodology

Hypotheses 1.1.0, 2.0.0 and 3.0.0 were tested through
the use of the 3 x 2

block design illustrated in Figure 3.1.

The dependent variables in this analysis are drop-out r a t e ,
college GPAC , and the discrepancy between actual GPA and
predicted GPA.

The independent variables are sex and group

membership.
Hypotheses 4.0.0, 5.0.0 and 6.0.0 were tested through
the use of the 2 x 2

block design shown in Figure 3.2.

Since

75
this design was used to determine if the results were biased
by the lack of predicted GPA^ on some of the subjects,
discrepant GPA

variable was eliminated.

C~P

the

The two remaining

dependent variables for the second design are drop-out rate
and college G P A c .

The independent variables are treatment

groups and response groups.
The design of this study is descriptive in nature
and consists of the two block designs described above.

The

proposed analysis will consist of the Multivariate Analysis
of Covariance

(MANOCVA) using high school GPA^S as covariables

If, through exploration, high school GPA^g is discovered not
to be correlated with the dependent variables, the analysis
will consist of the Multivariate Analysis of Variance (MANOVA)
Summary
Three groups of students were selected as the basis
for this study.

The subjects of Developmental Group I

consisted of the entire group of Martin Luther King Jr.
students.

The subjects were selected to participate in

the program in June, 19 69, by the Program Director on the
basis of recommendations from non-school related persons.
Only those students were selected who appeared motivated
to succeed, had low grades and were in need of financial
assistance.

Twenty-nine females and thirty-three males

were admitted and became the basis of this study.

The mean

high school grade point average of this group was 1.98.

76
The group participated in a six-week Summer Orientation
Program where counseling and special developmental services
were available.

During the Fall of 1970, they were enrolled

as a full-time student.

Follow-up counseling and financial

assistance was provided during the Fall and Winter semesters.
The population from which the sample was selected
for Control Group II consisted of ninety-eight females and
seventy-nine males who were admitted as beginning freshmen
during the Fall, 19 70 semester, without benefit of any
special program or counseling.

From this population twenty-

nine females and thirty-three males were randomly selected.
The mean high school grade point average for Control Group II
was 2.17.
The population from which the subjects of Control
Group III was selected consisted of fifty-two females and
fifteen males.

Twenty-nine females were selected who had

high school grade point averages of exactly 2.50.

Fifteen

males were selected with high school grade point averages
of exactly 2.20.

All of the subjects in Control Group III

were admitted as beginning freshmen during the Fall semester
of 1970, without benefit of any special program or coun­
seling.
Two designs were used in this study.

The first

design attempted to measure the difference between the
three groups on three dependent variables; college GPAc ,
drop-out rate and the discrepancy between actual and

77
predicted GPA.

The second design was used to test for

bias in the first design.
The proposed analysis for this study will consist
of a multivariate analysis of covariance
high school

as covariables.

(MANOCVA) using

If, through exploration,

high school GpA^g is found not to be correlated with the
dependent variables,

the analysis will consist of a multi­

variate analysis of variance

(MANOVA).

CHAPTER IV

ANALYSIS OF DATA
In this chapter an analysis of the data is presented
along with the testable hypotheses.
into four main sections:

The chapter is divided

testable hypotheses, exploratory

analysis, first design, second design.

In the first design

all subjects with missing data have excluded from the three
groups.

The second design was utilized to determine whether

the results in the first design were biased by the absence
of the missing data subjects and includes all of the subjects
in Groups I and II.

Group III was not included in the

second since none of its subjects had missing data.
Throughout this chapter several grade point averages
(GPA) are referred to and are designated as follows:

High

school GPA (GPA^s ), Actual College GPA (GPAc ) , ACT predicted
GPA (GPA1 ) and the discrepancy between the actual college
lr

GPA and the ACT predicted GPA (GPA _ ).
c p

The discrepancy

(GPAc_p) is determined by substracting the actual college
GPA from the ACT predicted GPA.
Hypotheses for First Design
1.0.0 Null Hypothesis:
There is no significant difference
between groups on GPAC drop-out rate, and the discrepancy
between actual and predicted GPA.
78

79
1.1.0:

There is no significant difference between
groups on GPAC .

1.2.0:

There is no significant difference between
groups on drop-out rate.

2.3.0:

There is no significant difference between
groups on the discrepancy between actual
and predicted GPA.

2.0.0 Null Hypothesis:
There is no significant difference
between genders on GPAC , drop-out rate, and the discrepancy
between actual and predicted GPA.
2.1.0:

There is no significant difference between
genders on GPAC .

2.2.0:

There is no significant difference between
genders on drop-out rate.

2.3.0:

There is no significant difference between
genders on the discrepancy between the actual
and predicted GPA.

3.0.0 Null Hypothesis:
There is no significant interaction
between groups and genders on GPAC , drop-out rate and the
discrepancy between the actual and predicted GPA.
3.1.0:

There is no significant interaction between
groups and genders on GPAC .

3.2.0:

There is no significant interaction between
groups and genders on drop-out rate.

3.3.0:

There is no significant interaction between
groups and genders on the discrepancy between
actual and predicted GPA.
Hypotheses for Second Design

4.0.0 Null Hypothesis:
There is no significant difference
between treatment groups on GPAC and drop-out rate.
4.1.0:

There is no significant difference between
treatment groups on GPAC .

80

4.2.0:

There is no significant difference between
treatment groups on drop-out r a t e .

5.0.0 Null Hypothesis:
There is no significant difference
between response groups on GPAC and drop-out rate.
5.1.0:

There is no significant difference between
response groups on GPAC .

5.2.0:

There is no significant difference between
response groups on drop-out rate.

6.0.0 Null Hypothesis:
There is no significant interaction
between treatment groups and response groups on GPAC and
drop-out rate.
6.1.0:

There is no significant interaction between
treatment groups and response groups on GPAC .

6.2.0:

There is no significant interaction between
treatment groups and response groups on drop­
out rate.
Exploratory Analysis

Before testing the hypotheses, an exploratory analysis
was made to determine the strength of high school GPA as
covariables.

The pooled within cell correlations of each

of the dependent variables are shown in Table 4.1.

A cor­

relation of .17 was found between actual college GPA and
high school GPA; a correlation of -.0 3 was found between
drop-out rate and high school grades; and a correlation of
.12 was found between the discrepancy between actual and
predicted college GPA and high school grades.

The .92

correlation between college GPAC and the discrepant GPA
would naturally be high because GPA

c

c-p

is a part of the formula,

TABLE 4.1.— Correlation Matrix with Covariable Included.
Variable

sd

College
GPAC

Drop-out
Rate

College GPAC

.59

Drop-out Rate

.42

.29

Discrepant GPAc_p

.57

.92

.28

.17

-.03

High School GPAhs

.29

Discrepant
GPAc_p
.92

.17

.28

-.03
.12

.12

TABLE 4.2.— Correlation Matrix: with Covariable Eliminated.
Variable

sd

College
GPAC

College GPAC

.58

Drop-out Rate

.58

.30

Discrepant GPAc_p

.43

.92

Drop-out
Rate
.30

Descrepant
GPAc_p
.92
.29

.29

High School

82
The correlation matrix with the variance due to the
covariable (GPA^g ) eliminated is shown in Table 4.2.

An

examination of the results showed that no significant
change had occurred as a result of introducing high school
GPA as a covariable.

The canonical correlation between

GPA^s and the other measures was 0.2124 with high school
GPA accounting for 1.50 per cent of the variation in the
dependent variables.
Since high school GPA was not found to be a suitable
covariable it was not introduced.
First Design
Multivariate analysis of variance was used to test
the hypotheses.

The three groups were divided into six

cells using sex and group membership as cell determinants.
The subjects with missing ACT scores, and correspondingly
predicted GPAs were dropped from this design.

The cell

frequencies are presented in Table 4.3.
In order to maintain an appropriate error rate the
alpha level for the experiment was partitioned not only
principally between two major sections of Design I and
Design II, but also within each design to account for post
hoes where appropriate.
out at P .10 .

All findings are being reported

83

TABLE 4.3.— Cell Frequencies for First Design.
Group

Male

Female

Development I

22

21

43

Control II

28

27

55

Control III

15

29

44

Total

65

77

142

Total

The first hypothesis tested is stated below.
1.0.0 Null Hypothesiss
There is no significant difference
between groups on GPAC , drop-out rate and the discrepancy
between the actual and predicted GPA.
The multivariate analysis of variance of means for
the three groups produced an F— ratio of 9.6479 with six
degrees of freedom for the hypothesis and 268 for the error
term,

(P<.0001).

A significant multivariate F — ratio is

interpreted as meaning that the three groups differ on at
least one of the dependent variables.

The decision was to

reject the null and a univariate analysis of variance was
subsequently conducted to determine upon which variable
or variables the three groups differed.

The following

hypotheses were tested:
1.1.0:

There is no significant difference between
groups on GPAC .

1.2.0:

There is no significant difference between
groups on drop-out rat e .

1.3.0:

There is no significant difference between
groups on the discrepancy between the actual
and predicted GPA.

84
A summary of the univariate analysis of the hypothe­
ses is presented in Table 4.4.

The overall alpha level for

the multivariate analysis was set at .05.

Since the multi­

variate analysis was rejected, the .05 alpha was split into
equal portions for each univariate.

Thus, each univariate

was tested at the .0167 level of confidence.
Hypothesis 1.1.0 on the actual college grade point
average was not rejected, P<.0 380.

The failure to reject

this hypothesis indicates that the three groups did not
differ on the actual college grade point average.

The

failure to reject Hypothesis 1.2.0 on drop-out rate indi­
cates that the three groups did not differ on that variable,
P<.4759.

Hypothesis 1.3.0 rejected P<.0001 indicating that

at least one of the groups had a higher discrepancy between
the predicted and actual GPA.
Further study was conducted to determine in what
ways the three groups differed on this measure.

The fol­

lowing hypothesis was explored:
1.3.1:

There is no significant difference between
Group I and Group II on the GPA discrepancy.

1.3.2:

There is no significant difference between
Group I and Group III on the GPA discrepancy.

1.3.4:

There is no significant difference between
Group I and the average of Groups II and III
on the GPA discrepancy.

Scheffe'Post hoc analysis was used to determine
which of the three groups differed on the GPA discrepancy

TABLE 4.4.— Summary of Univariate Analysis— Group Effect.

Variable

Degrees of Freedom
------------------Hypothesis
Error

Between Mean
Squared

Ratio

P

Decision

College GPAC

2

136

1.1569

3.3503

.0380 Not rejected

Drop-out Rate

2

136

.1344

.7467

.4759 Not rejected

Discrepant GPAc_p

2

136

4.3827

12.9631

.0001 Rejected

86

variable.

The results are shown on Table 4.5.

The post

hoes were tested at the .01 alpha level, with two degrees
of freedom for the hypothesis and 120 for the error, F=4.79.
TABLE 4.5.--Scheffe'Post hoc Comparisons for Differences.
A

Contrast

Y

Scheffe1 Gap

Decision

A

A

1.3.3

547

.366

Rejected

x3

546

.386

Rejected

2 = X3

-.001

.364

Not rejected

.3295

Rejected

x

II

V 1.3.2

Xi = X2
1
—1
IX

V 1.3.1

x2 +

x

3

/\

¥

X • J • (z

.5465

Xl =
2

Since the confidence interval for the contrast between
the two control groups, Group II and Group III, crossed zero,
the Null Hypothesis for this contrast was not rejected.
The rejection of the other post hoes indicates that
the ACT predictor under predicted college grades more for
Developmental Group I than it did for Control Groups II or
III

(see Table 4.7c).

Since it had been shown that Group I

did in fact differ from the other two groups, a subsequent
test was to be made to determine whether Group I actually
did earn better than a 2.00 grade point average.

The analysis

was made on the following hypothesis using the one-tailed,
one sample t-test.

87

1.3.5:

The mean grade point average of Developmental
Group I will be less than or equal to 2_.00 at
the end of one year's attendance. Ho: x < 2.00

Its alternate was:
1.3.6:

The mean grade point average of Developmental
Group I will be greater than 2.00 at the end
of one year's attendance.
Ho: X > 2.00

The mean GPA of the Group was calculated and measured against
2.00

criterion measure using the one sample t-test.

The

decision was to reject the Null Hypothesis and accept the
alternate since the mean GPAc was greater than 2.00, t=7.0 89 8
significant at the .005 level.

The mean college GPA of each

of the three groups are shown in Table 4.7.
The next hypothesis was tested to determine the
effect of the sex differences on the three dependent
variables.
2.0.0 Null Hypothesis:
There is no significant difference
between genders on GPAC , drop-out rate, and the discrepancy
between actual and predicted GPA.
The multivariate analysis of variance of means for
the two genders produced an F--ratio of 3.2021 with three
degrees of freedom for the hypothesis and 134 for the error
term, P<.0255.

The decision was to reject the null and

conduct a univariate analysis on each of the three dependent
variables using the following hypothesis.
2.1.0:

There is no significant difference between
genders on GPAC .

2.2.0:

There is no significant difference between
genders on drop-out rate.

88
2.3.0:

There is no significant difference between
genders on the discrepancy between actual and
predicted GPA.

A summary of the univariate analysis on the three
hypotheses is presented in Table 4.6.
reject the hypothesis if P<.0167.

The decision was to

The rejection of Hypothe­

sis 2.1.0 indicates that the females had a significantly
higher grade point average than the males

(see Table 4.7a).

The rejection of Hypothesis 2.3.0 indicates that the ACT
predictor under predicted college grades more for the
females than it did for the males

(see Table 4.7c).

The

failure to reject Hypothesis 2.2.0 indicates that there
was no difference between the males and the females on
drop-out rate.
The next hypothesis was analyzed to determine if
there was any interaction between the three groups and
males and females on the three measures.
3.0.0 Null Hypothesis:
There is no significant interaction
between groups and genders on GPAC , drop-out rate and the
discrepancy between the actual and predicted GPA.

The multivariate test for interaction produced an
F— ratio of 1.0114 with six degrees of freedom for the
hypothesis and 268 for the error term, P<.4182.

The

hypothesis was not rejected since the probability was
>.05.

The failure to reject this hypothesis indicates

that there was no significant interaction between the
three groups and the males and females on any of the three
measures.

TABLE 4.6.— Summary of Univariate Analysis— Sex Effect.
Degrees of Freedom
Variable

Hypothesis

Error

Between Mean
Squared

Ratio

College GPAC

1

136

2.9503

8.5472

Drop-out Rate

1

136

.0009

.0048

Discrepant GPAc_p

1

136

2.8155

8.3277

Decision
.0041 Rejected
.9448 Not rejected
.0046 Rejected

90
Summary of First Design
In this design all of the subjects with missing
data were excluded.

An exploratory analysis was made to

determine the strength of high school grade point average
as covariables.

Since GPA^g accounted for only 1.50 per

cent of the variation in the dependent variables it was
not introduced.

Instead, the hypotheses were tested using

multivariate analysis of variance.
No significant difference was found between the
three groups on the actual college grade point average
(GAPC ) or on the drop-out rate.

A significant difference

between the groups was found on the discrepancy factor
(GPA

c-p

) and the Scheffe Post hoc analysis was used to
■*

determine in what ways the groups differed.

The results

of the Scheffe analysis showed that the ACT predictor under
predicted more for Developmental Group I than for the other
two groups.

Having determined that Developmental Group I

did have a significantly higher GPA discrepancy, a test
was made to see if the actual college grade point average
of the group exceeded 2.00.

The results of the one sample

t-test showed the actual college grade point average of
Developmental Group I to be greater than 2.00, P .005.
The results of the test for sex differences on the
three measures showed that the females earned a significantly
higher grade point average than the males.

No significant

91
difference was found between the males and females on drop­
out rate.

A significant difference was found on the GPA

discrepancy variable.

c-p

The results indicated that the ACT

predictor under predicted more for the females than for
the m a l e s .
The multivariate test for interaction produced no
significant differences between the groups and genders on
any of the three measures.
A summary of the means of the three groups on six
measures is presented in Table 4.7.
on percentages.

Table 4.7b is computed

Eighty-two per cent of the males in

Developmental Group I and seventy-nine per cent of the males
in Control Group II were enrolled in school at the end of
one year.
Second Design
This design was used to determine if the results of
the first design were biased by the exclusion of the missing
data subjects.
As in the first design, the analysis for this design
consisted of the multivariate analysis of variance.
findings are reported tested at the .10 alpha level.

All
The

subjects in this design were divided into four cells using
treatment group membership and response group membership
as cell determinants.

The two treatment groups were

Developmental Group I and Control Group II.

Control

TABLE 4.7.— Summary of First Design Group Means.
Male
Treatment
Group

Female

Total

(a) College GPAC

Male

Female

(b) Persistency

Total
(%)

Male

Female

Total

(c) (GPA Discrepancy

Dev. I

2. 37
(22)a

2.76
(21)

2.56
(43)

.82
(22)

.86
(21)

.84
(43)

+ .35
(22)

+ .76
(21)

.550b
(43)

Control
II

2.18
(28)

2.33
(27)

2.25
(55)

.79
(28)

.74
(27)

.77
(55)

-.09
(28)

+ .10
(27)

.003
(55)

Control
III

2.14
(15)

2.48
(29)

2.36
(44)

.73
(15)

.72
(29)

.72
(44)

-.16
(15)

+ .09
(29)

.004
(44)

Total

2.23
(65)

2.50b
(77)

2.37
(142)

.79
(65)

.77
(77)

.78
(142)

.042
(65)

.276b
(77)

.169
(142)

(d) ACT Predicted GPA

(e)Composite ACT Score

(f)Cumulative Hours
Passed

Dev. I

2.02
(22)

1.99
(21)

2.00
(43)

14
(22)

13
(21)

13
(43)

22
(22)

23
(21)

22
(43)

Control
II

2.27
(28)

2.23
(27)

2.29
(55)

20
(28)

16
(27)

18
(55)

24
(28)

23
(27)

23
(55)

Control
III

2.30
(15)

2.39
(29)

2.35
(44)

18
(15)

18
(29)

18
(44)

27
(15)

24
(29)

25
(44)

Total

2.19
(65)

2.25
(77)

2.22
(142)

17
(65)

16
(77)

16
(142)

24
(65)

23
(77)

23
(142)

aThe number in the parentheses represent the number of subjects in each
cell.
^Significant at p < .05.

93
Group III was eliminated from this analysis because none of
its members had missing ACT scores.

The two response groups

were R.^ = missing data group and R 2 = no missing data group.
The two dependent variables are actual college grade point
average (GPA ) and drop-out rate.
c

The cell frequencies are

shown in Table 4.8.
TABLE 4.8.— Cell Frequencies for Second Design.
Treatment Group

Missing Data
R1

No Missing Data
r2

Developmental I

19

43

62

7

55

62

26

98

12 4

Control II
Total

Total

4.0.0 Null Hypothesis:
There is no significant difference
between treatment groups on GPAC and drop-out rate.
The multivariate analysis of variance of means for
the two groups produced an F — ratio of 4.5538 with two
degrees of freedom for the hypothesis and 119 for the error
term, P<.0125.

The decision was to reject the null and

proceed with the univariate analysis on the following
hypotheses.
4.1.0:

There is no significant difference between
treatment groups on GPAC .

4.2.0:

There is no significant difference between
treatment groups on drop-out r a t e .

I ^

94

The hypotheses were tested at the .0 25 alpha level.
A summary of the univariate analysis for the treatment group
effect is shown in Table 4.9.
4.1.0

The rejection of Hypothesis

on GPA c indicates that Developmental Group I had a

significantly higher actual college grade point average than
Control Group II.

The failure to reject Hypothesis 4.2.0

on drop-out rate indicates no significant difference between
the Developmental Group I and Control Group II.
The next hypothesis tested the difference between
the missing data group and the no missing data group on
actual college grade point average

(GPAc ) and drop-out rate.

5.0.0 Null Hypothesis:
There is no significant difference
between response groups on GPAC and drop-out rate.
The multivariate analysis for this hypothesis
produced an F— ratio of 9.6 401 with two degrees of freedom
for the hypothesis and 119 for the error term, P<.0002.
The decision was to reject the null since P<.05 and the
univariate analysis was conducted using the following
hypotheses.
5.1.0:

There is no significant difference between
response groups on GPAC .

5.2.0:

There is no significant difference between
response groups on drop-out rat e .

The hypotheses were tested at the .025 alpha level.
A summary of the univariate analysis for the response group
effect is shown in Table 4.10.

TABLE 4.9.— Summary of Univariate Analysis Treatment Group Effect.
Degrees of Freedom
Variable

Hypothesis

Error

Between Mean
Squared

College GPAC

1

20

3.2362

Drop-out Rate

1

20

.889

Ratio
9.080
.4936

P

Decision

.0032

Rejected

.4836

Not rejected

TABLE 4.10.--Summary of Univariate Analysis Response Group Effect.

Variable

Degrees of Freedom
------------------Hypothesis
Error

Between Mean
Squared

Ratio

P

Decision

College GPAC

1

120

6.4713

18.4713

.0001

Rejected

Drop-out Rate

1

120

1.0054

5.5837

.0198

Rejected

96

The rejection of Hypothesis 5.1.0 indicates that
the no missing data group (Response Group II) earned a
significantly higher grade point average than the missing
data group

(Response Group I), see Table 4.11a for mean

comparison.
The rejection of Hypothesis 5.2.0 shows that the
missing data group (Response Group I) had a significantly
higher drop-out rate than the no missing data group
(Response Group II), see Table 4.11b.
The final hypothesis tested for interaction between
Developmental Group I and Control Group II and the missing
data/no missing data groups on the measures of GPA

c

and

drop-out rate.
6.0.0 Null Hypothesis:
There is no significant interaction
between treatment groups and response groups on GPAC and
drop-out rate.
The multivariate analysis for this hypothesis had
a F— ratio of .6131 with two degrees of freedom for the
hypothesis and 119 for the error term, P<.5434.

The hypothe

sis was not rejected since the probability was greater than
the .05 alpha level.

The failure to reject this hypothesis

indicates that there is no significant interaction between
Group I, Group II and the missing data/no missing data group
Summary of Second Design
The second design was used to test for bias in the
first design.

The analysis consisted of multivariate

TABLE 4.11.— Summary of Second Design Group Means.
Missing
Data R-^
Treatment
Group

Complete
Data R 2

Total

(a) College GPAC

Dev. I

2.28
(19)a

2.56
(43)

2.47
(62)

Control
II

1.28
(7)

2.25
(55)

Total

2.01
(26)

2.38
(98)

Missing
Data Ri

Complete
Data R 2

(b) Persistency
b

Total

(%)

Missing
Data R^

Complete
Data R 2

Total

(c) GPA Discrepancy

.68
(19)

.83
(43)

.78
(62)

.56
(43)

.56
(43)

2.14
(62)

.42
(7)

.76
(55)

.72
(62)

.04
(55)

.04
(43)

2.30
(124)

.61
(26)

.79b
(98)

.75
(124)

.26
(98)

.26
(98)

(d) ACT Predicted GPA

(e) Composite ACT Score

(f)Cumulative Hrs. Passed

Dev. I

2.01
(43)

2.01
(43)

13
(43)

13
(43)

23
(19)

22
(43)

22
(62)

Control
II

2.25
(55)

2.25
(43)

18
(55)

18
(55)

14
(7)

24
(55)

22
(62)

Total

2.14
(98)

2.14
(98)

16
(98)

16
(98)

20
(26)

23
(98)

22
(124)

aThe number in parentheses represent the number of subjects in each cell.
Significant at p < .05

98

analysis of variance.

Developmental Group I had a sig­

nificantly higher college grade point average than Control
II.

No significant difference was found between the two

groups on drop-out rate.

The no missing data group (R2)

earned a significantly higher GPA
group

(R1) .

c

than the missing data

The no missing data group also had a signifi­

cantly lower drop-out rate than the missing data group
(R-^) .

No significant interaction was found between the

two treatment groups and the two response groups.
The findings of the Second Design are consistent
with the findings of the First Design on drop-out rate when
the groups are compared.

However, the findings are incon­

sistent with the findings of the First Design on actual
college grade point average

(GPA ) when the groups are
c

compared.
Summary
In this chapter data from Developmental Group I ,
Control Group II and Control Group III were presented and
analyzed relative to the major hypothesis and post hoes.
A multivariate analysis of variance, a one sample
t-test, and the Scheffe'post hoc comparisons were the
statistical techniques applied.

The exploratory analysis

indicated that high school grade point average was not a
suitable covariable.

High school grades accounted for

99
1.50 per cent of the variation in the dependent variables.
A correlation of .17 was found between college grades and
high school grades; a correlation of -.0 3 was found between
drop-out rate and high school grades; and a correlation of
.12 was found between high school grades and the predictor
discrepancy.
The mean grade point average of Developmental Group
I was calculated and measured against the 2.00 criterion
measure using the one sample t-test.

Developmental Group I

was found to have done significantly better than 2.00,
t=7.089 8, significant at the .00 5 level.

The ACT predictor

under predicted college grades more for Developmental
Group I than it did for either of the other two groups.
No significant difference was determined between the three
groups on the attained college grade point average or on
the drop-out rate in the first design.
In the first design, the females had a significantly
higher college grade point average than the males.

The ACT

predictor significantly under predicted college grades more
for the females than it did for the males.
findings were significant at the .05 level.

Both of these
There was no

significant difference between the males and females on
drop-out rate.

No significant interaction was found between

the groups and genders on the dependent variables.

100
In the second design, Developmental Group I had a
significantly higher grade point average than Control
Group II, which was significant at P<.0032.

No significant

difference was found between Developmental Group I and
Control Group II on drop-out rate.

The group with no missing

data earned a higher grade point average than the missing
data group, P<.0001.

The missing data group had a signifi­

cantly higher drop-out rate than the no missing data group,
P<.019 8.

No significant interaction was determined between

the treatment groups and response groups on the two dependent
variables of college grade point average and drop-out rate.
The findings of the Second Design were inconsistent
with the findings of the First Design on actual college
grade point average

(GPAc ). Possible bias is indicated

since Developmental Group I earned a significantly higher
GPAc than Control Group II in the Second Design, but not
in the First.
In the next chapter conclusions will be drawn along
with a discussion and implications for future research.

CHAPTER V

SUMMARY AND CONCLUSIONS
This final chapter is divided into the following
four sections:

Summary, discussion of the findings, con­

clusions and implications for future research.

The dis­

cussion of the problem, the design and the methodology of
the study is presented in the first part.

In the second

part the major findings are presented along with a dis­
cussion of each.

Conclusions are drawn in the third part

and the implications for future research are stated in the
fourth part.
Summary
Four year colleges, universities and community
colleges have become more and more concerned with the large
number of high-risk students who are currently seeking
admission into their respective schools.

Such institutions

are faced with the problem of developing and implementing
innovative courses and curriculums to meet the special
needs of these disadvantaged high school graduates.
Several colleges and universities are attempting
to meet these special needs through the use of one or all

101

102

of the following:

reduced admission, criteria, special

recruitment efforts, innovative teaching approaches, pro­
visions for financial aid, provision of a summer bridge
program and personal and academic counseling.

The Martin

Luther King Program at Western Michigan University has all
of the above components.

In this study, an attempt was

made to analyze and appraise the Martin Luther King
Developmental Program relative to its purpose of helping
high-risk students perform successfully at Western
Michigan University.
Three groups of students were selected as the basis
for this study.

The sample for the Developmental Group

(Group I) consisted of the entire group of sixty-two Martin
Luther King students.

The subjects were selected to par­

ticipate in the program in June, 19 69, by the Program
Director.

The mean high school grade point average of the

group was 1.9 8 and the mean ACT score was 13.

Twenty-nine

females and thirty-three males were admitted and became
the basis of this study.

The group participated in a six-

week resident Summer Orientation Program where developmental
courses, counseling and financial assistance was provided.
This group was enrolled as full-time students in the Fall
of 1970, and were given additional counseling, tutoring
and financial assistance.
The sample for the control group

(Control Group II)

consisted of twenty-nine females and thirty-three males.

103

The subjects were randomly selected from 177 students who
were admitted as freshmen at the beginning of the Fall,
19 70 semester.

All of the students of the control group

were admitted as regular students without benefit of a
special program.

The mean high school grade point average

of this control group was 2.17 and the mean ACT score was
18.
The sample for the control group

(Control Group III)

consisted of twenty-nine randomly selected females with
high school grade point averages of exactly 2.50 and fifteen
males with high school grade point averages of exactly 2.20.
The mean high school grade point average for the group was
2.39 and the average ACT score was 18.

All of the students

in this control group were admitted as beginning freshmen
at the beginning of the Fall semester, 19 70, without benefit
of any special program.
The dependent variables in this study were actual
college grade point average

(GPAc ) , drop-out rate and the

discrepancy between the actual and the ACT predicted grade
point average

(GPA

c p

).

The independent variables were

group membership and sex.
Two designs were used in this study.

The first

design attempted to measure the difference between the three
groups on the dependent variables with the missing data
subjects excluded.

The second design was utilized to

104
determine whether the results in the first design were
biased by the exclusion of the missing data subjects.
An exploratory analysis was made to determine the
strength of high school grade point averages as a covariable.
High school grade point average was not introduced as a
covariable since it accounted for only 1.50 per cent of the
variation in the dependent variables.

Since high school

grades was not a suitable covariable the hypotheses were
tested using multivariate analysis of variance.
The effectiveness of the Martin Luther King Program
was measured through the use of the following hypothesis.
First Design
1.0.0 Null Hypothesis:
There is no significant difference
between groups on GPAC , drop-out rate, and the discrepancy
between actual and predicted GPA.
1.1.0:

There is no significant difference between
groups on GPAC .

1.2.0:

There is no significant difference between
groups on drop-out rate.

1.3.0:

There is no significant difference between
groups on the discrepancy between actual and
predicted GPA.

2.0.0 Null Hypothesis:
There is no significant difference
between genders on GPAC , drop-out rate, and the discrepancy
between actual and predicted GPA.
2.1.0:

There is no significant difference between
genders on GPAC .

2.2.0:

There is no significant difference between
genders on drop-out rate.

2.3.0:

There is no significant difference between
genders on the discrepancy between the actual
and predicted GPA.

105
3.0.0 Null Hypothesis:
There is no s.i gn i.I i(\m I in t.er.ie t i<>n
between groups and genders on G P A C , drop-out rate and the
discrepancy between the actual and predicted GPA.
3.1.0:

There is no significant interaction between
groups and genders on GPAC .

3.2.0:

There is no significant interaction between
groups and genders on drop-out rate.

3.3.0:

There is no significant interaction between
groups and genders on the discrepancy between
actual and predicted GPA.

The first hypothesis

(1.0.0) was tested using the

multivariate analysis of variance to determine if there was
any difference between the three groups on the three
dependent variables.

The rejection of that hypothesis

led to the subsequent univariate analysis on Hypotheses
1.1.0, 1.2.0 and 1.3.0 to determine upon which variable
or variables the three groups differed.

The rejection of

Hypothesis 1.3.0 led to the testing for group difference
on the discrepant GPA variable.

The Scheffe Post hos

analysis was used to determine which of the three groups
differed on the GPA discrepancy variable.

The following

hypotheses were tested:
1.3.1:

There is no significant difference between
Group I and Group II on the GPA discrepancy.

1.3.2:

There is no significant difference between
Group I and Group III on the GPA discrepancy.

1.3.3:

There is no significant difference between
Group II and Group III on the GPA discrepancy.

1.3.4:

There is no significant difference between
Group I and the average of Groups II and III
on the GPA discrepancy.

106

Hypothesis 1.3.3 on the GPA discrepancy was not
rejected.

Hypotheses 1.3.1, 1.3.2, and 1.3.4 were rejected

indicating that the ACT predictor under-predicted more for
the Developmental Group than for the two control groups.
Having determined that Developmental Group I differed from
the other two groups on the GPA discrepancy variable, a
subsequent test was made to determine if Developmental
Group I actually exceeded the 2.00 minimum GPAc requirement.
An analysis was made of the following hypotheses using the
one-tailed, one sample t-test:
1.3.5a :

The mean grade point average of Developmental
Group I will be less than or equal to 2.00
at the end of one year's attendance.
Ho:
X £ 2.00.

Its alternate w a s :
1.3.5k:

The mean grade point average of Developmental
Group I will be greater than 2.0JJ at the end
of one year's attendance.
Ho: X > 2.00.

The Null Hypothesis was rejected indicating that the mean
college grade point average of Developmental Group I was
significantly above the minimum college grade point average
of 2.00.
The second major hypothesis

(2.0.0) was tested

using the multivariate analysis of variance to determine
if there was any difference between the males and females
on the three dependent variables.

The rejection of the

second major hypothesis led to the utilization of the
univariate analysis to determine upon which variable or

107

variables the males and females scored differently.

The

Hypothesis on GPAc (2.1.0) was rejected which showed that
the females had a significantly higher college grade point
average than the males.
sis on drop-out rate

The failure to reject the hypothe­

(2.2.0) indicated that the males and

females did not differ on this variable.
on the discrepant grade point average

The Hypothesis

(2.3.0) was rejected

which showed that the ACT predictor under predicted more
for the females than the males.
The third major hypothesis was tested using the
multivariate analysis of variance to determine if any inter­
action existed between the three groups and the males and
females.

This hypothesis was not rejected indicating that

no interaction existed.
Second Design
The second design tested for bias in the first
design.

The two dependent variables in the second design

were college GPA (GPAc ) and drop-out rate.

The independent

variables were treatment group membership and missing data/
no missing data group membership.

The following hypotheses

were tested.
4.0.0 Null Hypothesis:
There is no significant difference
between treatment groups on GPAC and drop-out rate.
4.1.0:

There is no significant difference between
treatment groups on GPAC .

10 8
4.2.0:

There is no significant difference between
treatment groups on drop-out r a t e .

5.0.0 Null Hypothesis:
There is no significant difference
between response groups on GPAC and drop-out rate.
5.1.0:

There is no significant difference between
response groups on GPAC .

5.2.0:

There is no significant difference between
response groups on drop-out rate.

6.0.0 Null Hypothesis:
There is no significant interaction
between treatment groups and response groups on GPAC and
drop-out rate.
6.1.0:

There is no significant interaction between
treatment groups and response groups on GPAC .

6.2.0:

There is no significant interaction between
treatment groups and response groups on
drop-out rate.

Hypothesis 4.0.0 was tested using the multivariate
analysis of variance to determine if there was any difference
between Developmental Group I and Control Group II on actual
college grade point average and persistence.

The rejection

of the hypothesis led to the use of the univariate analysis
to determine which of the two groups scored differently on
which dependent variable.

Hypothesis 4.1.0 on GPAc was

rejected which indicated that Developmental Group I had a
significantly higher GPAc than Control Group II.

The

Hypothesis on drop-out rate (4.2.0) was not rejected
indicating no difference on drop-out rate.
Hypothesis 5.0.0 was tested using the multivariate
analysis of variance of means to determine if there was
any difference between the missing data group and the no

109

missing data group on actual college grade point average
and persistence.

The rejection of the hypothesis led to

the use of the univariate analysis to determine whether
the missing data group or the no missing data group scored
higher on actual college GPAc and on the drop-out rate.
The Hypothesis on GPA

(5.1.0) was rejected indicating

that the group with no missing data had a higher GPAc than
the group with missing data.

The rejection of Hypothesis

5.2.0 on drop-out rate showed that the missing data group
had a higher drop-out rate than the no missing data group.
Hypothesis 6.0.0 was tested using the multivariate
analysis of variance to determine if any interaction
existed between Developmental Group I, Control Group II
and the missing data/no missing data groups.

This Hypothe­

sis was not rejected which revealed that no interaction
existed.
Findings and Discussion
The findings relative to the three major hypotheses
in the first design are as follows:
1.

Students in Developmental Group I, Control
Group II, and Control Group III achieved
equally well during their first year of
enrollment at Western Michigan University.

This finding strongly suggests that the Developmental
Program was successful.

The students in the Developmental

Group had an average high school grade point average of 1.99

110

and an average ACT score of 13.

This group was able to

perform academically as well as a second group of students
who had a mean high school grade point average of 2.17
and a mean ACT score of 18.

Still further, this finding

shows that the Developmental Group of students

(who would

not have been admitted under regular admission criteria)
did as well as the control group who met the minimum
standards.

The mean grade point average for the third

group was 2.39 and the average ACT score was 18.

The

average college grade point average for each of the three
groups is shown in Table 5.1.

TABLE 5.1.— Summary of Means Group Effect.
n

GPA
°

Persistency
Ratea

GPA
Discrepancy

43

2.56

.84%

.550b

II

55

2.25

.77%

.003

Control Group III

44

2.36

.72%

.004

Developmental Group I
Control Group

aPercentage of students enrolled at end of year.
Significant at P<.05.

Ill

2.

Students in Developmental Group I, Control
Group II and Control Group III persisted in
the university equally well during their
first year of enrollment at Western Michigan
University.

The second finding also implies that the program was
of value to the Developmental Group.

This group was able

to persist in the university as well as the control group
members.
Nineteen per cent of all freshmen drop out at
Western Michigan before the end of their first year."'"
Twenty-eight per cent of the students dropped out before
the end of the first year in Group III, twenty-three per
cent dropped out in Group II and only sixteen per cent
dropped out in Group I.
Group
it does

While this does not imply that

I did significantly better than the other

two groups,

suggest that they did as well.
3.

There is a significantly greater discrepancy
between the predicted and actual college
GPAc_p in Group I than in either of the
other two groups with the ACT predictor
under predicting in all cases but minimal
in Group II and Group III.

This finding is supported by the literature

and

suggest that the traditional standardized test are poor
indicators of success with disadvantaged students.

In

this study the disadvantaged group achieved over one-half
grade higher than the prediction criteria indicated.
^"Admissions Office, Western Michigan University,
1970.

112
4.

The cumulative college grade point average
of Developmental Group I was significantly
higher than the 2.00 minimum academic
requirement.

This finding shows that the Developmental Group
not only met the 2.00 minimum requirement for good academic
standing, moreover, the group significantly exceeded the
2.00 grade point average with a GPAc of 2.56

(t-7.089

p < .005).
TABLE 5.2.--Summary of Means Sex Effect.
Persistency
Rate

GPA
Discrepancy

n

gpac

Male

65

2.23

.79%

.042

Female

77

2. 50a

.77%

.276a

aSignificant at P<.05.
5.

The females in Developmental Group I, Control
Group II and Control Group III achieved
significantly higher cumulative college grade
point averages than the males in Developmental
Group I, Control Group II and Control Group
III.

The implications of this finding is that females
tend to perform somewhat higher than males.
generally supported by the literature.

This fact is

A summary of

college grade point averages is shown in Table 5.2.
6.

The ACT predictor significantly under
predicted more for the females in Develop­
mental Group I, Control Group II and
Control Group III than for the males in
Developmental Group I, Control Group II
and Control Group III.

113
This finding suggests that the ACT predictor in
addition to being a poor indicator of college success for
disadvantaged students appears also to be a poor indicator
for females at Western Michigan University.
The findings relative to the hypothesis in the
second design are as follows:
1.

The students in Developmental Group I
achieved significantly higher cumulative
college grade point averages than the
students in Control Group II during their
first year of enrollment at Western Michigan
University.

The implications of this finding is that absence
of themissing data
to bias the

subjects in the first design tended

results in favor of the control groups.

The

finding of a significant difference in this design seems
to be related to the extremely low grade point averages of
the seven subjects with missing data in Group II who were
included in the second design.

The inclusion of this group

tended to lower the GPAc of Group II more than the 19
subjects with missing data lowered the GPAc of Group I.
However, since this study attempted to examine the
entire membership of each of the three groups, the inclusion
of all subjects was essential.

The findings in this design

clearly indicate that Group I earned a significantly higher
grade point average than the control group.

The mean grade

point averages of both groups are shown in Table 5.3.

114

TABLE 5.3.— Summary of Means— Group Effect.
Group

n

Persistency Rate

GPAC

Developmental Group I

62

2. 47a

.78

Control Group II

62

2.14

.72

S i g n i f i c a n t at P<.05.
2.

The students in Developmental Group I and
Control Group II persisted in the university
equally well during their first year of
enrollment at Western Michigan University.

This finding is in accordance with the findings of
the first design and again shows that Group I did as well
as the two control groups in remaining enrolled in the
university.
3.

The students in Developmental Group I and
Control Group II with missing data have lower
college grade point averages (GPAC ) and higher
drop-out rates than students who did not have
missing data.

While the evidence clearly supports this finding
there is no available data to explain this phenomenon.
One can only speculate that the students who did not take
the ACT test also tended to be less sure of themselves and
their ability to succeed in college.

The data suggest that

they may have been less promising academically than their
counterparts who took the test.

If this is so, it becomes

somewhat understandable why they tended to have lower grade
point averages and higher drop-out rates.

115
Conclusions
This study was designed to evaluate the effective­
ness of the developmental program at Western Michigan Uni­
versity relative to its purpose of preparing high-risk
students to perform successfully in curriculums at Western
Michigan University.

The results of this study support

the contention that such a program can benefit high-risk
students in institutions of higher learning.
From the data and subsequent findings of this study
the following conclusion regarding the effectiveness of the
Martin Luther King Developmental Program is drawn.
The Martin Luther King Developmental Program is
successful in preparing high-risk high school
graduates to perform successfully in curriculums
at Western Michigan University.
The MLK program was established with the objective
of helping disadvantaged students maintain a successful
grade point average while attending the university.

The

data shows that the college grade point average of the
Developmental Group was significantly higher than the two
control groups and that the developmental students achieved
higher grades than were predicted by the ACT program.
The finding in the first design illustrates that
the Developmental Group not only met the 2.00 minimum
requirement for good academic standing but significantly
exceeded the 2.00 with an earned grade point average of
2.56.

In addition, the Developmental Group was able to

116
maintain a level of academic persistence equal to that of
the regularly admitted students in Control Groups II and
III.

Finally, while 19 per cent of all freshmen dropped

out of school during their first year, only 16 per cent
of the students in the Developmental Group dropped out.
Thus, the evidence clearly points out that when
high-risk students at Western Michigan University are
given supportive services such as a summer bridge program,
financial assistance, tutoring and peer counseling, they
have as good a chance of succeeding in college as regularly
admitted students.

Moreover, as indicated in the second

design, high-risk students who participate in developmental
programs can actually achieve at a higher rate than students
who meet the minimum standards for regular admission.
One of the more significant aspects of this study
is that the Developmental program tended to undermine the
traditional predictors of success and in a sense the highrisk students actually became "over-achievers" as their
grade points exceeded the traditional predictors by an
excess of one half of one grade point.

In addition, it

can be concluded that while the data produced no significant
difference between the three groups on persistence, the
program had the effect of keeping the high-risk students
in school at a rate equal to regularly admitted students.
This finding is in keeping with the literature.

In most

117
instances, the academic mortality rate for high-risk students
has been no higher than for regular students when developmental assistance is provided.

2

Can it be assumed thatthis

low drop-out rate is a result of the special program?
Current evidence from institutions with equivalent control
groups provide an affirmative answer:

essentially, that

the drop-out rate is much higher for the control group than
for those m

the high-risk programs.

3

Furthermore, the

grades of students in high-risk programs have generally
been much higher than they would have been predicted from
traditional prediction equations.
The Hawthorne Effect has undoubtably played a major
role in the success of the program.

This tendency for

students to achieve higher when they know they are being
observed and given special attention should not be viewed
negatively, but should be seen as an integral part of
Developmental Programs.

If one can increase the chances

of a student's success in school by merely giving him special
attention, then the Hawthorne Effect seems a small price to
pay for the potential high pay-off resulting from its con­
scious use.
2

See Leslie Berger, College Now for Ghetto Youth
(New York:
City University of New York, 19 68), p. 15.
3
See Robert L. Williams, "What Are We Currently
Learning From Current Programs for Disadvantaged Students,"
Journal of Higher Education (April, 1969), p. 280.

118
Implications for Future Research
1.

A follow-up study of students who dropped out of

the program before the end of the year was not made.

Further

research could be designed to study those students who
dropped out of similar programs to determine what positive
effect the program has on their future academic plans.
2.

This study has reported on the results of the

freshman year of the Developmental students.

Further

research could be conducted to determine how many of these
students will actually graduate from the university.
3.

Further research could be done using the

Tennessee Self-Concept Test to determine if programs such
as the one in this study has any impact on improving the
self-concept of high-risk students.
4.

Several studies have emphasized the importance

of instructors and special counselors in the effectiveness
of developmental programs.

Further research is needed in

this area.
5.

Further research is needed relative to the use

of criteria other than standardized test to predict academic
success with high-risk students.

BIBLIOGRAPHY

119

BIBLIOGRAPHY

Alsop, J.
"Ghetto Education."
1967, pp. 18-23.

The New Republic, November 18,

Archie, Jackson Marguerite.
"Observations on Minority Group
Education." The Journal of the Association of
College Admissions Counselors, 13 (June, 1968),
23-25.
Berg, E. G . , and Axtell, D. Programs for Disadvantaged
Students in the California Community Colleges.
Oakland, California:
Peralta Junior College
District, 1968.
Berger, Leslie.
College New For Ghetto Youth.
City University o£ New York, 19 68.

New York:

Black, Millard H.
"Programs for Disadvantaged Youth."
Report of the Conference on Reading. University
of Pittsburgh, School of Education (19 68), 139048.
Black Community and the Community College--A Project Report,
Th e . Atlanta, Georgia:
Southern Regional Education
Board, 19 70.
Blumenfeld, Warren S.
"Selecting Talented Negro Students:
Nominations vs. Test Performance."
NMSC Research
Reports, 5 (November 6, 1971).
Borgen, Fred H.
"Able Black Americans in College:
Entry
and Freshman Experiences." NMSC Research Reports,
6 (November 2, 19 70).
________ . Differential Expectations? Predicting Grades
for Black Students in Five Types of Colleges.
Evanston, Illinois: National Merit Scholarship
Corporation, 19 71.
Brookover, Wilbur, and Edsel, Erickson.
Society, Schools
and Learning. Boston:
Allyn and B a c o n , 1969.
Brown, Frederick G . , and Bubo i s , Thomas E.
"Correlates of
Academic Success For High-Ability Freshman Men."
Personnel and Guidance Journal, 42 (February, 1964),
603-7.
120

121
Bruck, M . , and Bodwin, R. F.
"The Relationship Between
Self-Concept and Absence of Scholastic UnderAchievement."
Journal of Clinical Psychology, 18
(1962), 181-2.
Cash, Carol D. Educationally Inferior Students:
Getting
In And Out of College.
University of Missouri:
May, 19 70.
Cass, James.
"Can the University Survive the Black
Challenge?"
Saturday Review (June 21, 1969),
68-84.
Cherdack, Arthur.
Predictive Validity of the Scholastic
Aptitude Test for Disadvantaged College Students
Enrolled m a Special Education Program Final
Report. University of California: April, 1971.
Clark, Burton R. The Open Door College— A Case Study. The
Carnegie Series in American Education.
New York:
McGraw-Hill, 1960.
Clark, Kenneth B.
"A Charade of Power: Black Students at
White Colleges."
The Antioch Review, 19 (Summer,
1969), 145-8.
________ . "Alternate Public School Systems."
Harvard
Educational Review, 38 (Winter, 1968) , 100-13.
Dark Ghetto:
Dilemmas of Social Power.
Harper & R o w , 1965.
_________. ed. Racism and American Education.
Harper & Row, 19 70.

New York:

New York:

Clark, K. B . , and -Plotkin, L.
The Negro Student at Integrated
Colleges. New Y o r k : National Scholarship Service
and Fund for Negro Students, 196 3.
Cleary, T. A.
Test Bias: Validity of the SAT for Negro and
White Students xn Integrated Colleges. Princeton,
New Jersey:
Educational Testing Service, Research
and Development Report, 18 (1966), 65-66.
Clift, Virgil A.
"Further Considerations in the Education
of the Disadvantaged."
The Educational F o r u m , 34
(January, 1970), 223-28.
"Higher Education of Minority Groups in the United
States."
The Journal of Negro Education, 38 (Summer,
1969), 291-302.

122
Drossland, Fred E.
"Heavey Burdens at the Narrow Gate:
The American Negro and Higher Education."
The
Journal of the Association of College Admissions
Counselors, 13 (March, 1968), 7-40.
________ . Minority Access To College. A Ford Foundation
Report.
New Y o r k : Schocken B o o k s , 19 71.
Cureton, E. E . ; Cureton, Louise W . ; and Bishop, Ruth.
"Prediction of Success in Graduate Study of
Psychology at the University of Tennessee."
American Psychologist, 4 (1949), 361-362.
Daniel, Walter G.
"Needed:
A Re-Examination of Plans for
Disadvantaged Negro Youth."
The Journal of Negro
Education, 35 (Summer, 1966) , 199-203.
DiRusso, L . , and Aven, S. D.
"Does Remedial English Provide
Help for College Freshmen."
California Journal of
Educational Research, 22 (19 71),5-8.
Dudley, David A.
"College Entrance Tests and The Dis­
advantaged Student."
The Journal of the National
Association of College Admissions Counselors, 13
(February, 19 69), 13-15.
Egerton, John.
Higher Education for 'High Risk' Students.
Atlanta: Southern Education Foundation, 1968.
Fricke, B. G. Opinion Attitude and Interest Survey Handbook
Ann Arbor, Michigan:
Michigan Oais Testing Program,
1965.
"Prediction, Selection, Morality and Quality
Control."
College and University, 32 (1956), 34-52.
Goodrich, Andrew L.
Community Services for the 'New Student
At Inner City Community Colleges"
Michigan State,
Kellogg Community Services Leadership Program,
1970.
Green, Robert L.
Chicago:

Racial Crisis in American Education.
Follett Educational Corporation, 19*69.

"The Black Quest for Higher Education:
An
Admission Dilemma."
The Personnel and Guidance
Journal, 47 (May, 1969) , 906-908.
Howe, Clark, Allen, et al.
Racism and American Education:
A Dialogue and Agenda for Action"
New York:
Harper
& R o w , 1970.

123
Hyman, S. R.
"The Miller Analogies Test and University of
Pittsburgh Ph.D.s in Psychology."
American
Psychologist, 12 (1957), 35-36.
Johnson, Lamar B.
Islands of Innovation Expanding;
Changes
in the Community College. Beverly Hills, California:
Glencoe Press, 1969.
Knoell, Dorothy.
Black Student Potential. Washington,
D.C.: American Association of Junior Colleges,
19 70.
Lake,

Ernest G."The Case Against External
Standardized
Tests:
They Create an Elite."
The Nation's Schools,
70 (August, 1962), 51.

McDaniel, Ruben R . , and McKee, James W.
An Evaluation of
Higher Education's Response to Black Students.
Indiana University:
September, 1971.
Miller, Doris M . , and O'Connor, Patricia.
"Achiever
Personality and Academic Success Among Disadvantaged
College Students."
Journal of Social Issues, 25
(1967), 105.
Moore, William Jr. Against the Odds:
The High Risk Student
in the Community College. San Francisco, California:
Jossey Bass, Inc., 1970.
Munday, Leo A. Factors Influencing the Prediction of College
Grades. Paper represented at Annual Meeting of
National Council on Measurement in Education.
Chicago, Illinois: February 10, 196 8.
Morgan, Gordon D. The Ghetto College Student. Iowa City,
Iowa: American College Testing Program, 19 70.
Robinson, Armstead D . ; Foster, Craig C . ; and Oglilive,
Donald H . , Editors.
Black Studies in the University.
New York:
Bantam Books, 1969.
Rosenthal, Robert, and Jacobson, Lenore.
Pygmalion in the
Classroom.
New York Rinehart and Winston, Inc.,
1968.
Stanley, Julian D.
Predicting College Success of Educationally
Disadvantaged Students.
John Hopkins University,
September, 19 ?0.

124
Success Breeds Success, Basic Studies, 19 70-71 Report.
Fort W o r t h , Texas: Tarrant County Junior College
District, 1971.
Taylor, R. G.
"Tutorial Service and Academic Success."
Journal of Educational Research, 61 (1968) , 195-197.
Tinker, Irene.
"The Underprepared College Student."
American Education (November, 19 70).
Trent, William R. College Compensatory Programs for Dis­
advantaged Students, Report No. 3 . Washington,
D . C . : Eric Clearinghouse, September, 19 70.
Tyler, Leona.
York:

The Psychology of Human Differences.
Appleton-Century-Crafts, 1966.

New

Williams, Robert L.
"What Are We Learning from Current
Programs for Disadvantaged Students."
Journal of
Higher Education (April, 1969), 274-275.
Willingham, Warren W.
Free-Access Higher Education.
York:
College Entrance Exam Board, 1970.

New

APPENDICES

APPENDIX 1

SUMMARY OF CORRELATIONS
sd

ACT

CUM HP

GPA'p

GPAC

D 0 R

GPAc_p

.049

.724

.155

-.005

-.134

.112

.629

.442

.590

.224

.013

-.174

.286

.920

ACT

4.85

CUM HP

6.17

.049

GPA'p

.23

.724

.112

GPAC

.58

.155

.629

.224

D 0 R

.42

-.005

.442

.013

.286

GPAc_p

.58

-.134

.590

-.174

.920

LEGEND:

.284
.284

Comprehensive American Testing Program score (ACT); cumulative hours
passed (CUM H P ) ; Predicted GPA (GPA'p); College GPA (GPAC ); Drop out
Rate (D 0 R ) ; discrepancy between predicted and College GPA (GPAc_p);
standard deviation (sd).

APPENDIX 2
GENERAL STUDIES PROGRAM*
I.

Freshman-Sophomore Level
Both
College Writing 1 1 6
4
Freshman Reading 1 4 0
2
Either
Early Western Civilization 1 0 0 ..................4
(Greeks to 17th Century)
or
Modern Western Civilization 1 0 1 ................. 4
(17th Century to Present)
One course from the following:
Man and Society 202
4
Social Bases for Human Behavior 203
. . . .4
Comparative Social Institutions 204
. . . .4
Race and Culture 205.............................. 4
Two courses from the following:
Physical Geography 105 or Geological
Science 1 1 2 ..................................... 4
Biological Science 1 0 7 ........................... 4
Physical Science 10 8
........................... 4
Aims and Achievements of Science 110
. . . .4
Either
Arts and Ideas 222
4
or
Non-Western Arts and Ideas
224
4

hrs.
hrs.
hrs.
hrs.
hrs.
hrs.
hrs.
hrs.
hrs.
hrs.
hrs.
hrs.
hrs.
hrs.

k

This is the General Studies Program required of all
Students at Western Michigan University.

127

APPENDIX 3

ACADEMIC LOAD
SUMMER SESSION
4--- Race & Culture
2_
6

Adult Reading
Academic Hours

FALL SEMESTER
4--- College Writing
4--- Arts & Ideas
3--- Speech
1--- Physical Education
—

13

--- Techniques of learning and adjustment
Academic Hours

WINTER SEMESTER
2--- Freshman Reading
4--- Aims & Achievements of Science
3--- History
3--- Introductory

(Sociology, Business etc.)

1^-- Physical Education
13
Academic Hours

These Hours Represent a typical academic load for the Martin
Luther King Student.

smsum

APPENDIX 4
MARTIN LUTHER KING SUMMER SESSION ACADEMIC CALENDAR
JUNE 21 - AUGUST 14

129

6:00 a.m.

Breakfast
6:45-7:15

Breakfast
6:45-7:15

Breakfast
6:45-7:15

Breakfast
6:45-7:15

Breakfast
6:45-7:15

7:00 a.m.

Race&Cult.
7:20-9:00

Race&Cult.
7:20-9 :00

Race&Cult.
7:20-9:00

Race&Cult.
7:29-9 :00

Race&Cult.
7:20-9:00

9 :00 a.m.

Race&Cult.
9:20-11:00

Race&Cult.
9:20-11:00

Race&Cult.
9:20-11:00

Race&Cult.
9 :20-ll:00

Race&Cult.
9 :20-11:00

Adult Read.
9 :20-11:00

Adult Read.
9 :20-11:00

Adult Read.
9 :20-11:00

Adult Read.
9 :20-11:00

Adult Read.
9 :20-11:00

12 :00 noon

Lunch
12:00-12:20

Lunch
12 :00-12:20

Lunch
12 :00-12:20

Lunch
12 :00-12 :20

Lunch
12:00-12:20

1:00 p.m.

Study Cent.
1:00-3:30

Study Cent.
1:00-3:30

Study Cent.
1:00-3:30

Study Cent.
1:00-3:30

Study Cent.
1:00-3:30

Library Sci.
1:00-3:30

Library Sci.
1:00-3:30

Library Sci.
1:00-3:30

Library Sci.
1:00-3:30

Library Sci.
1:00-3:30

3:00 p.m.

Phys.Ed.&Rec.
1:30-5 :30

Phys.Ed.&Rec.
1:30-5:30

Phys.Ed.&Rec.
1:30-5 :30

Phys.Ed.&Rec.
1:30-5:30

Phys.Ed.&Rec
1:30-5:30

5:00 p.m.

Dinner
5:30-6:00

Dinner
5:30-6 :00

Dinner
5:30-6:00

Dinner
5:30-6:00

Dinner
5:30-6:00

7:00 p.m.

Special
Programs

Special
Programs

Special
Programs

Special
Programs

Special
Programs

Art
Broad./T.V.

Music

Art
Groad./T.V.

Speech

Speech

APPENDIX 5

MAJOR AND MINOR AREAS OF ENROLLMENT
AREA
Business Administration/Education etc.

19 70 MLK STUDENTS
8

Sociology/Social Work

14

Liberal Arts

14

English

0

Engineering Technology

4

Art

0

Chemistry

0

Distributed Education

0

Industrial Arts

3

Occupational Therophyl

0

Speech Pathology

1

Music

2

Pre-Medicine

1

Pre-Nursing

2

Undecided
Special Education

11
2

APPENDIX 6
COST AND RESOURCES
Below is a tabulation of the average expenses for
Attendance at Western.

Martin Luther King Student in

Fall/Winter Semester
COSTS

RESOURCES

TUITION/FEES

430.00 PERSONAL SAVINGS

BOOKS/SUPPLIES

100.00 SAVINGS-SUMMER

ROOM/BOARD
PERSONAL EXPENSE

$ 200.00

1,100.00 PARENTS
270.00 VET. BENEFITS
$1 ,806.00
W.M.U. SCHOLARSHIP
N.D. STUDENT LOAN
EDU. OPPOR. GRANT
WORK STUDY
TOTAL

$ 100.00

$

250.00

$

250.00

$ 300.00
$ 900.00
$ 900.00

$ 600.00
$ 600.60

$1 ,000.00

$ 300.00
?1,800.00 $1,800.00 $1,800.00

Summer Session
COSTS
TUITION/FEES
BOOKS/SUPPLIES
ROOM/BOARD
PERSONAL EXPENSE

100.00 WORK-STUDY
30.00 PERSONAL-SUMMER

$

100.00

$

100.00

$

100.00

250.00 SAVINGS-SUMMER
70.00 PARENTS
VET. BENEFITS
W.M.U. SCHOLARSHIP
N.D. STUDENT LOAN
EDU. OPPOR. GRANT
WORK STUDY
TOTAL

$ 200.00

$ 220.00
$ 230.00

$ 250.00

$

150.00

$ 450.00

$ 450.00

?

450.00