SUCECEEES [N E’ROFESSKJNAL PHYSICAL EDUCATION AT THE
UNWERSITY OF {LENSES (CHICAGO)

 

TEE-5.5.5 1‘31“; THE DEQREE O? Lid. D.

MICHIGAN STATE UNEVERSH'Y’
SHELDON LEROY FORDHAM

1363

 

THESIS

This is to certify that the

thesis entitled

r r! r‘

A SCUDY Cr I113 EHLXLLIC LVQL k1" SLLLCLLD
FACECRS 'IO ACADEDHC SUCCESS IN PROFESSIONAL
PHYSICAL EDUCATION
AT "33' UNIVERSITY OF ILLINOIS (CHICAGO)

presented by

Sheldon LeRoy Fordham

has been accepted towards fulfillment
of the requirements for

71

bcho degree in Education

    

,////

/ Major professor

 

Date 9( 25¢ 63

0-169

LIBRARY

Michigan State
University

 

A STUDY OF THE RELATIONSHIP OF SELECTED FACTORS TO ACADEMIC
SUCCESS IN PROFESSIONAL PHYSICAL EDUCATION AT THE

UNIVERSITY OF ILLINOIS (CHICAGO)

by
Sheldon LeRoy Fordham

AN ABSTRACT OF A THESIS
Submitted to
Michigan State University

for the degree of

DOCTOR OF EDUCATION

Department of Health, Physical Education and Recreation

1963

ABSTRACT

A STUDY OF THE RELATIONSHIP OF SELECTED FACTORS TO
ACADEMIC SUCCESS IN PROFESSIONAL PHYSICAL EDUCATION
AT THE UNIVERSITY OF ILLINOIS (CHICAGO)

by Sheldon LeRoy Fordham

This study was designed to analyze thirty-four vari-
ables associated with academic success among freshman male stu-
dents in professional physical education at the Chicago Under-
graduate Division of the University of Illinois. The data
used were from the records of 12% students covering a period of
four years (1958-62). The variable used as the criterion for
academic success was grade point average.

The variables studied were commuting time, occupation
of father, education of father, education of mother, extra—cur-
ricular participation, high school rank, high school attended,
previous grade point average, grade point average by profes-
sion of father, grade point average by college, and test scores
on the School and College Ability Test, the Bild-Dutton Aca-
demic Interest Test, and the United States Army Five-Item Phys-

ical Fitness Test.

A separate IBM card was made up for each of the subjects,

into which was punched values for each of the variables. Part
of the calculations were done by electronic computer. Thus it

1

Sheldon LeRoy Fordham
2

was possible to deve10p the upper correlation matrix consist-
ing of correlations for each variable with the criterion and
the correlations for each variable with every other variable.
The calculation of Multiple R was completed by hand. A.mul-
tiple regression analysis was made of the data which resulted
in a prediction equation. Seventy-two students in the sample
entered the institution directly from high school and 52 stu-
dents entered from other colleges within the university.
These two groups of students were identified as "regular" and
"transfer" students respectively. The data obtained on the
two groups of students were analyzed separately.

High school rank had the highest correlation with the
criterion for regular students (r = .31). Previous grade point
average correlated .HS with the criterion for transfer students.
Four variables for both the regular and transfer students were
found to give as much predictive power as could be obtained.
The transfer student variables were (1) index for previous col-
lege, (2) previous grade point average, (3) interest in geol-
ogy as measured by the interest inventory test, and (u) high
school rank (Multiple R 8 .62). The best set of predictors
for regular students were (1) high school rank, (2) high school
attended index, (3) fitness test score, and (u) achievement
test score in social science (Multiple R = .HB). A few of the
zero order correlation coefficients for regular students were
as follows: verbal learning compared with social science
achievement, .59; mathematics achievement with natural science

achievement, .57; and verbal learning with vocabulary, .56.

Sheldon LeRoy Pordham
3
Education of father and education of mother correlated .97
for transfer students. Other correlations for transfer stu-
dents were: quantitative learning with mathematics achieve-
ment, .68; verbal learning with vocabulary, .67; and verbal
learning with social science achievement, .62. The analysis of
students who withdrew indicated that students who participate
in intercollegiate and intramural sports are more likely to
stay in school. The variables significant at the .05 level,
namely, high school rank, verbal learning, mathematics, vocabu-
lary, and grammar were in favor of the students who remained in
school. Extra-curricular participation correlated negatively
with the criterion for both groups of students. There was a
significant difference between the occupation of the father in
the two groups of students.
The general conclusions were that high school rank is
a valuable criterion in predicting college success, particu-
larly for incoming students directly from high school. However,
the previous college record provides the best information for
prognosis for students who have been enrolled in other col-
leges within the university. A formal testing program does
not provide ready or complete answers in predicting college
success. It is useful in making a contribution to the total

situation rather than to a specific case.

A STUDY OF THE RELATIONSHIP OF SELECTED FACTORS TO ACADEMIC
SUCCESS IN PROFESSIONAL PHYSICAL EDUCATION AT THE

UNIVERSITY OF ILLINOIS (CHICAGO)

by

Sheldon LeRoy Fordham

A THESIS
Submitted to
Michigan State University

for the degree of

DOCTOR OF EDUCATION

Department of Health, Physical Education and Recreation

1963

W'v

32/541”

PREFACE

This study was undertaken to determine the relative
value of several selected factors with respect to academic
success. The study sample included one hundred twenty-four
male students in the College of Physical Education at the Chi-
cago Undergraduate Division of the University of Illinois.

The study was designed to determine the factors rela-
ted to academic success and to develop a predictive equation
which might be used in the selection and retention of male
students majoring in physical education. Student withdrawals
were also studied to gain further insight into the problem.

Even though there have been several studies related in
one way or another to this one, the study samples were quite
different because of the nature of the institution. The re-
sults of these studies indicate a need for further investiga-
tion of the variables related to academic success in college.

The author is indebted to his wife, Margaret Fordham,
whose constant encouragement helped make this study possible.

The writer is particularly indebted to Dr. Wayne Van
Huss for his guidance and assistance in the organizational
phases of this work.

Acknowledgment is also made to James Creaser of the
University of Illinois (Chicago) for his help in the statisti-
cal analysis.

Sheldon LeRoy Fordham

ii

TABLE OF CONTENTS

Page
PREFACE . . . . . . . . . . . . . . . . . . . . . . . . ii
LIST OF TABLES . . . . . . . . . . . . . . . . . . . . v
Chapter
I. THE PROBLEM AND IMPORTANCE OF STUDY . . . . . . l
The Problem
Importance of the Study
Definition of Terms Used
II. REVIEW OF THE LITERATURE . . . . . . . . . . . u
III. METHODS OF RESEARCH . . . . . . . . . . . . . . 2M

Type of Study
Data
Type of Method Involved

IV. RESULTS: MULTIPLE CORRELATION AND MULTIPLE
REGRESSION O O O O O O O O O O O O O C O O O 33
V. RESULTS: CHOSEN VARIABLES . . . . . . . . . . . 38

Variables for Regular Students
Variables for Transfer Students

VI. RESULTS: OTHER VARIABLES . . . . . . . o . . . #2
VII. LOOKING AT STUDENT WITHDRAWALS . . . . . . . . 50

Original Sample for Study

College DrOp-Out Figures

National Study by United States Office
of Education

Questionnaire Study at Institution
Studied

iii

Chapter

VIII.

FINDINGS, CONCLUSIONS, AND RECOMMENDATIONS .

Significance of Variables between Regular
Students who Withdrew and Regular
Students who Remained in School
Disposition of Students in the Study

BIBLIOGRAPHY .

APPENDIX

0

iv

Page
. 57
. 62
. 67

Table
I.
II.
III.

IV.

VI.

VII.

VIII.

IX.

XI.

XII.
XIII.

XIV.

XV.

LIST OF TABLES
Intercorrelations: Transfer Students . . . . .
Intercorrelations: Regular Students . . . . .

Progressive Increase of Multiple R: Transfer
Students..........o.......

Progressive Increase of Multiple R: Regular
StUdentS O O O O O O O O O O C O C C O O O 0

Transfer Students: Means, Standard Deviations,
and Beta Weights . . . . . . . . . . . . . .

Regular Students: Means, Standard Deviations,
andBetaWeightS..o...........

Intercorrelations between Variables: Regular
and TranSfeI‘ StUdentS o o o e o e o e o e 0

Significance of Differences between Means:
Twenty Transfer and Regular Students . . . .

Significance of Differences between Means:
Twenty Transfer and Regular Students . . . .

Significance of Differences between Means:
TWenty Regular and Top Twenty Transfer
StUdentSoocooeoooeoooooeoo

Significance of Differences between Means:
Twenty Transfer and Low 'IWenty Transfer
StUdentS O O O I O O O I C O O O C O C O O 0

Reasons Why Students Withdrew . . . . . . . .

Significance of Variables between Regular
Students who Withdrew and Regular Students
Who Remained in SChCOl o o e e o e o o o e o

Disposition of Seventy-Two Regular Students .

Disposition of Fifty-Two Transfer Students . .

Top ‘

Page
3H

3H

35

35

35

36

an

H5

U6

H7

H8

52

53
55

56

CHAPTER I

THE PROBLEM AND IMPORTANCE OF STUDY

The problem was to determine the relationship of sev-
eral selected factors to academic success among freshman male
students in the College of Physical Education at the Chicago
Undergraduate Division of the University of Illinois.

The objectives of this study were (1) to develop cri-
teria which might prove useful in the selection of college
students majoring in physical education, (2) to develop a pre-
dictive equation based on the variables used in the study, and
(3) to determine some of the basic reasons why students with—
draw from the university.

In view of the expanding school population, the prob-
lem of the correct choice of a profession and the need for
some predictive criteria to use as a means of selection becomes
exceedingly more important. Students dropped for poor schol-
arship or for the various other reasons represent a consider-
able loss of human resources to society in addition to the
almost astronomical cost. We can no longer afford to ignore
or blandly dismiss the problem of both the entering and the
continuing student. We need to apply realistic research to his
problems and the end, which is the dropout, needs to be ana-
lyzed in terms of the means that caused the end. The nature

1

2
and extent of the college dropout problem may be completely
different in one school as compared to another. Therefore,
the problem needs to be studied in each school in order to
answer such questions as (1) What are dropouts like? (2) What
caused them to leave school? (3) How can we identify poten-
tial dropouts? (A) How may we evaluate possible remedies?

The limitations of this study were that (l) the data
are applicable only to the College of Physical Education at
the University of Illinois (Chicago), (2) the validity of some
of the instruments used in the comparisons may be open to
question, (3) the study is limited to the factors studied, and
(u) some of the students may return to the university after
withdrawal or academic failure.

Two definitions are necessary in order for the reader
to differentiate between the two groups of students in the
study sample. "Regular students" refers to those students who
entered the College of Physical Education immediately upon
completion of high school. "Transfer students" refers to those
students who transferred to the College of Physical Education
after completing one semester in one of the other three col-
leges within the university (Commerce, Engineering, or Liberal
Arts and Science).

The Chicago Undergraduate Division of the University
of Illinois is a two year institution offering courses in
four areas: Commerce, Engineering, Liberal Arts and Science,
and Physical Education. The school is co-educational, with

approximately forty-seven hundred students in full-time at-

3

tendance. The Colleges of Engineering and Liberal Arts each
have approximately two thousand students enrolled. The Col-
lege of Commerce has approximately five hundred students.

The College of Physical Education has an enrollment of approx-
imately one-hundred fifty students, consisting of one hundred
men and fifty women. This study was limited to the male stu-

dents in the College of Physical Education.

CHAPTER II
REVIEW OF THE LITERATURE

Considerable research has been done relative to the
droP-out problem on all levels of the educational system. How-
ever, the majority of the studies have been completed on the
elementary and secondary levels. Many of the investigations
also lack adequate statistical evaluation. For the sake of
reasonable brevity, only the more pertinent studies are re-
viewed in the following.

Nardelli, following his analysis of college freshman
drop-outs in 1959, made a plea for realistic research on drop-
out patterns and asked that educators not accept answers that
are too obvious.1 His analysis of the drop-out pattern showed
that much of it could be traced to naive interpretation and
manifestation of orientation programs. Further research was
recommended, especially to be directed toward determination of
the cause of dropouts.

Iffert, in 1956, studied a sample of 1n,ooo students

2

from 152 institutions. The survival record showed that the

probability of graduating from the institution of first regis-

 

1W. Nardelli, "An Analysis of Drop-Outs of Freshmen,"
Junior College Journal, XXIX (February, 1959), 322.

2R. Iffert, "The Student Retention and Withdrawal
Study," College and University, XXX (July, 1956), HOS-ull.

 

u

5
tration was in the same order as rank in high school graduat-
ing classes. The prospects of graduating were about twice
as good for students in the top fifth of their high school
class and about eight times as good for those as for the stu-
dents from the bottom fifth.) The percentage reaching univer-
sity graduation from the third and fourth quintiles of their
high school classes was about one-third of that for students
from the top fifth. The mortality rates showed that approxi-
mately sixty per cent of the students left the university of
original registration without a degree and forty-five per cent
did not graduate from any college or university during the
four year period.

Koelsche interviewed 180 former students at Indiana
University between the years l9u8 and 1952.3 He was primarily
concerned with isolating the characteristics of those persons
who enrolled as freshman students but who withdrew prior to
completion of graduation requirements. This study indicated
that low scholarship was not entirely dependent upon lack of
ability, but rather the result of a number of factors exerting
influence on the individual. The reasons listed by students
for withdrawal were a combination of lack of funds, low schol-
arship, contemplated marriage, ill health, and loss of inter-
est. Out of 2,091 new freshman enrollees at Indiana Univer-
sity-in September of l9u8, 1,173 (56.2 per cent) withdrew with-

out graduating. Twenty-five per cent of these drop-outs

 

3C. L. Koelsche, "A Study of the Student DroP-Out Prob-
lem at Indiana University," Journal of Educational Research,
XLIX (January, 1956), 357-365.

6
ranked in the highest one-fifty of their reSpective graduating
classes. The median for the group of drop-outs was located
in the middle fifth. Twenty-seven and eight-tenths per cent
of the fathers of drOp-outs were in proprietary and manager—
ical positions, and 15 per cent were skilled laborers. TWenty-
nine and four-tenths per cent of fathers of drOp-outs had at-
tended college, and 15 per cent had graduated.

Shuman summarized much of the research on college
drop-outs in 1956.u He stated that, although statistics on
drop-outs vary considerably, the consensus was that about 50
per cent of college and university freshmen fail to graduate.
It was his opinion that a considerable number of those who
drop out might, with pr0per pre-college guidance, discover
abilities to better qualify them for some other pursuits. He
envisioned an ideal college program as one that would select
students more carefully, orient them more fully, counsel them
more effectively, and, in event of withdrawal, interview them
and attempt to aid in their fuller development.

Patton, at Louisiana State University for the years
1953-1955, attempted to determine if twenty-one selected fac-

5 Students of

tors were significantly related to drop-outs.
low socio-economic status were more likely to drop out (P =

.05). A low score on the American Council Psychological Exam—

 

”R. B. Shuman, "College Drop-Outs: An Overview," Jour—
nal 2: Sociology, XXIX (April, 1956), 3u7-350.

5B. K. Patton, "A Study of DrOpouts from the Junior
Division of Louisiana State University" (unpublished Doctoral
dissertation, Louisiana State University, 1958).

 

7

ination and a low point credit ratio was also more prevalent
among drOp-outs.

Linns and Pitt, in evaluating staying power and rate
of progress of University of Wisconsin freshmen, found per-
centile rank in the high school graduating class and the
American Council Psychological Examination test scores were

6 Less than 56 per cent of the

related to college success.
original group of 1,99% students attained SOphomore standing
or higher during the second year after entrance. Less than
half of the original group was registered the eighth semester
after graduation. There was a positive relationship between
the number of semesters registered and (a) percentile rank in
the high school class with which graduated (r = .37“), (b)
percentile rank on the American Council Psychological Examina-
tion (1957 edition) (r = .288), and (c) first semester uni-
versity grade point average (r = .5ul); all correlation coef-
ficients for these variables were significant (P = .001).
Fults and Taylor studied 2,u62 college freshmen from
forty-four different high schools from the standpoint of stay—

7 In this Sample the high schools did not differ

ing power.
significantly. In this study the greatest number of droP-outs
came between the end of the freshman year and the end of the

soPhomore year.

 

6L. Linns and H. Pitt, "Staying Power and Rate of
Progress of University of Wisconsin Freshmen," College and
University, XXIX (October, 1953), 98.

7R. S. Fults and S. E . Taylor, "Staying Power of Col-
lege Students," National Association of Seconda Schools
Principals' Bulletin,’XLIII (OEtOBer,_I959), I03.

8

Michigan State University freshman drop-outs and non-
drop-outs were compared by Russell according to six selected
factors (occupation of father, high school rating scale,

Basic College Inventory of Attitudes and Beliefs, education of
parents, sex, and American Psychological Council Examination

Test scores). Significant differences (P = .01) between drOp-
outs and non-drOp-outs were found only in education of father
and education of parents.8

Quarles has suggested sub-freshman non-credit courses
and recommended that a final decision should be withheld on ad-
mission on conditional or unconditional status until the com-
pletion of an orientation program.9 He concluded that rank in
the high school graduating class was a sounder guide than the
pattern of courses followed. It was his Opinion that a college
with an effective guidance program could reduce the number of
withdrawals.

Marsh selected nine factors to equate with drop-outs
and non-drOp-outs.10 His conclusions indicated that the best
predictors of college success were evidences of actual achieve-
ment (high school rank, College Board Verbal Scores, and par-

ticipation in a reading improvement program). Although these

 

8J. W. Russell, "A Comparison of Michigan State College
First-Term Drop-outs and Non-Drop-outs According to Certain
Factors" (unpublished Doctoral thesis, Michigan State College,
1952).

9B. Quarles, "Student Separations from Schools," Asso-
ciation gf American Colleges Bulletin, XXXV (October, 19kg),

10R. E. Marsh, Jr., "An Analysis of Failure among Uni-
versity Freshmen" (unpublished Doctoral dissertation, Boston

University, 1959).

9
measures predicted freshman failure with a 70 per cent degree
of accuracy, a need for further study was indicated. Commut-
ing time, parental education, participation in extra-curricu-
lar activities, regular employment, and a fifth year in
secondary school were found to be unrelated to failure.

Yoshimo studied college drop-outs at the end of the
freshman year at the State College of Washington in Pullman.ll
In comparing 52 students who remained in school with the 45
who drOpped out, the high school grade point average and Amer-
ican Council Psychological Examination scores were signifi-
cantly higher for those remaining in school (t = ”.37, 5.19
respectively; P = .01). .There was also a significant differ-
ence (P = .01) between the students who dropped out for aca-
demic reasons and those who withdrew owing to other factors
(t = ”.56, high school grade point average; t = n.92, psycho-
logical examination). The significance was in favor of stu-
dents' withdrawing due to other factors. He emphasized the
need for guidance to place the academic setting of college life
and the social aspects in preper perspective.

Cumings, following his study of the causes of student

12 suggested that colleges

withdrawals at De Pauw University,
have an exit interview to help the drop-out crystallize his

reasons for leaving and to realize the importance of this deci-

 

11R. Yoshimo, "College Dr0pouts at End of Freshman
Year," Journal 2; Educational Sociology, XXXII (September,

195 8) ,uf'. "‘""'

12E. C. Cumings, "Causes of Student Withdrawals at
De Pauw University," School and Society, LXX (September, 19u9),
152-153.

 

10
sion. Such an exit interview eliminates the many general,
and usually hazy, reasons given by students who are allowed
to withdraw without going through a carefully worked-out pro-
cedure.

The United States Office of Education studied the re-
tention and withdrawal of college students.13 Their findings
indicated that a majority of the students who discontinued
their higher education program attributed their withdrawals to
factors identified with themselves rather than with the insti-
tutions they attended. Principal institutional complaints
centered on services, notably orientation and teaching, rather
than institutional facilities. This study was based on a
sampling of approximately 13,700 men and women enrolled in in-
stitutions of higher learning in the fall of 1950. It was
found that no more than sixty per cent of all students who
enter degree granting institutions receive degrees. The first
year of college was the most critical drop-out period. Two
hundred seventy-three students per thousand left school during
the first year in comparison with 283 per thousand during the
next three years.

Metder investigated male graduates and scholastic
drop-outs at Cortland Teachers College from the aspect of a

group of selected characteristics.1u Sample students were en-

 

132.3. Office of Education Bulletin, "Retention and
Withdrawal of cm 3%va 177 pp.

1|+S. M. Metzger, "A Study of Selected Characteristics
of Male Graduates and Scholastic DrOpouts of the 1951 Fresh-
man Class Entering the State University of New York Teachers
College at Cortland" (unpublished Doctoral dissertation, New
York University, 1959).

ll
rolled in physical education and elementary education curric-
ulums. Findings showed that 85 per cent of the school drop-
outs in elementary education achieved a high school grade
average below 77 per cent. Achievement of a satisfactory grade
point average during the first semester at this institution
was significantly related to a high school average of 77 per
cent or above.

Rogers interviewed 97 selected freshman students at
Appalachian State Teachers College in 1959.15 His recommenda-
tions leading to a reduction of drop-outs from that institu-
tion were:

1. Development of a testing and placement service for

in-coming students.

2. Strengthening of the present faculty advisory sys-

tem.

3. Restriction of students to the campus for the first

semester.

u. Initiation of an orientation course during the

first quarter of the freshman year.

5. Proper handling of student applications for finan-

cial aid.

6. Appropriate involvement of students through the

student council.

Fenelon investigated the secondary school drop-out

 

15L. L. Rogers, "Problem Analysis Study of Selected
Freshman Students at Appalachian State Teachers College" (un-
published Doctoral dissertation, University of Tennessee,
1959).

12

16 In this

problem at Port Washington, Wisconsin High School.
questionnaire study of 111 cases it was discovered that many
variances existed in this school situation. However, the
findings were conclusive enough to show definite relationships
between individual backgrounds and leaving school early. He
cited a need for study of students' backgrounds and of their
influences on leaving. No single reason was found for leaving
school early.

In a review of the literature on secondary school drop-
outs, Tesseneer and Tesseneer discovered that those pupils who
engage in at least one extra-curricular activity are much less

17 It was felt that

likely to drop out than those who do not.
a low income status might affect this factor.

Mowers investigated self-judgments and objective meas-
ures as they relate to first semester academic achievement in
a group of non-selected college students.18 Statistical anal-
ysis indicated that decile rank in high school and test scores
on the Ohio State Psychological Examination were equally ef-

fective in the prediction of first semester collegiate academic

success. No significant difference was found in the effective-

 

16W. J. Fenelon, "A Study of the Secondary School Drop-
out Problem at Port Washington, Wisconsin High School" (unpub-
lished Doctoral dissertation, Northwestern University, 1960).

17R. A. Tesseneer and L. M. Tesseneer, "Review of the
Literature on School Dropouts," National Association of Sec-

 

ondagy School Principals' Bulletin, XEII (May, 1956),’Iu3-Isa.
186. E. Mowers, "Self-Judgments and Objective Measures

as Related to First Semester Academic Achievement of Non-Se-
lected College Students," (unpublished Doctoral dissertation,
Pennsylvania State University, 1960).

l3
ness of one variable over the other. However, it was demon-
strated that the proper combination of the variables produced
a significantly superior prediction of first semester col—
legiate grade average over the prediction derived from either
variable taken alone.

Fullmer studied the success and perseverance of uni-
versity students at the University of Denver in 1956.19 The
sample included 1,028 students, and the study was designed to
test the hypothesis that students who change educational ob-
jectives during their college careers are poor risks for aca-
demic success and perseverance. Changing of educational ob-
jectives as expressed by a change in major subject was not an
indication of weakness on the part of the student. A higher
percentage (un.3 per cent) of students graduated from the Uni-
versity of Denver in the "changed“ group than did those in the
"not changed" group (31.3 per cent). These figures included
all the schools at the University of Denver (Arts and Science,
Business Administration, and Engineering). Most students re-
tained about the same grade point average following a change
of major subject (r = .98). The correlation between freshman
grade point average and final grade point average was .90 or
higher for each group.

Lynch, in an analysis of drop-outs in three public
junior colleges in Florida, found that 3k per cent of the study

sample withdrew during the first semester and 30 per cent

 

19E.W.Fu11mer, "Success and Perseverance of College
Students," Journal of Higher Education, XXVII (November,
1956), uns-

 

1a
of the study sample drOpped out before completing the two
year program.20~ Eighty-six per cent of the drop-outs in the
study were freshman students. Student mortality was caused
by a constellation of factors. It was recommended that junior
colleges consider the advisability of having each freshman
select, or be assigned to, a faculty advisor. It was the
investigator's Opinion that approximately 65 per cent of the
drop-outs could have been helped.

Lowe studied capable high school graduates of twenty
Illinois high schools.21 Of the 1,117 subjects in the sample,
5M2 were in college and 575 were not. The type and number of
extra-class activities engaged in while in high school was
significantly related to attendance in college (P = .05).
Student rank in high school was significantly related to at-
tendance at college (P = .05). The educational level attained
by both the father and mother was significantly related to
college attendance (P = .05).

Rank in high school graduating class was significantly
associated with social factors involved in academic mortality

in Slocum's study of 1,019 freshmen entering the State Col-

lege of Washington (Pullman) in the falls of 1951, 1952, and

 

20D. F. Lynch, "An Analysis of Dropouts in Selected
Public Junior Colleges of Florida" (unpublished Doctoral dis-
sertation, Pennsylvania State University, 1959).

21W. T. Lowe, "Factors Related to Attendance or Non-
Attendance by Capable Illinois High School Graduates" (unpub-
lished Doctoral dissertation, University of Illinois, 1961).

15

1953.22 This writer rated the importance of personal and so-
cial adjustment of the students second only to intellectual
development. He cited a need for an effort by faculty members
to communicate to students the existence of a genuine interest
in them as individuals. In this study, size of high school
was not related to academic survival or academic performance
(r = .06) at the college. The average percentile rank in high
school was 73 for enrolled students and 58 for drop-outs.
Median scores on the Ohio State Psychological Council Examina-
tion were 55 and #2 respectively (P = .01).

Cape studied characteristics of 302 drop-outs at Dil-

lard University over a ten year period.23 Education or occu-

pation of parents was not significantly related to academic
achievement. Forty-nine per cent of the drop-outs were doing
passing work at the time of their leaving the University.
Faunce, following a study of in-term drop-outs at
Michigan State University for a two year period (19u7-u9), con-
cluded that there should be a closer working relationship among
those personnel officers interested in the living situation,
academic advisory services, student health, and professional

counselors.2u Within the college community the potential

 

22W. L. Slocum, "Social Factors Involved in Academic
Mortality," College and University, XXXII (Fall, 1956), 53-55).

23W. J. Cape, "A Study of Selected Characteristics of
Drop-outs at Dillard University" (unpublished Doctoral dis-
sertation, Indiana University, 1958).

2“L. D. Faunce, "A Study of Within-Term Dropouts at
Michigan State University for the School Years 19u7-u9" (un-
published Doctoral dissertation, Michigan State University,
1952).

 

16
within-term drop-out failed to participate as much with others
in campus organizations. Poor academic work was acknowledged
by students only when some other reason was lacking.

In studying 1,5u7 students in the School of Education
at the University of Minnesota Landskov found that 36.8 per
cent dropped out during their freshman year, 13.2 per cent
during the second year, 7.u per cent during the third year, and

25 The author stressed

H.H per cent during the fourth year.
the need for careful counseling during the freshman year.

A study by Hanks at the University of Arkansas in 1950
compared several factors related to retention and withdrawal

of freshmen.26

Of the 1,902 freshman students included in this
study, 595, or 31.3 per cent, failed to enroll for their sec-
ond year of study. Freshman withdrawals, compared with those
who remained in school, had lower scholastic aptitude as meas-
ured by the American Psychological Council Examination, poor—
er mastery of the English language as measured by the Cooper-
ative English Test, and lower high school cumulative grade
point average.

Munger studied 891 freshman students, from the lower

third of their high school graduating classes, at the Univer-

sity of Toledo to determine whether a discernible pattern could

 

25N. L. Landskov, "Suggested Student Survival Tech-
niques Recorded at the University of Minnesota," College and
University, XXIII (June, 19u8), 235-236.

26C. J. Hanks, "A Comparative Study of Factors Related
to Retention and Withdrawal of Freshman Students at the Uni-
versity of Arkansas" (unpublished Doctoral dissertation, Uni-
versity of Arkansas, 195k).

 

17
be observed to account for the various lengths of residence of
college students.27 The biographical variables of sex, age,
veteran status, and religion were only slightly, if at all,
related to persistance. Significant relationships were found
between the means of the persistance groups for most of the
data on course work. Success in the first courses in English,
history, mathematics, and social science were significantly
related to persistance (P = .01). The higher the grade re-
ceived for these courses, the longer the student remained in
college.

Over 3,000 drop-outs at Los Angeles City College were

studied by Snyder.28

The profile of the drop-outs differed
little, except in academic ability, from those of the other
collegians. The most significant findings were in respect to
differing academic abilities of the two groups (withdrawals and
continuing). The withdrawal group consistently showed inferior
ability as measured by mental and reading tests. The mean for
the Thurstone Psychological Examination Scores of the with-
drawal group was 151.3, in contrast to 16u.9 for the college

as a whole (P = .01). The mean for the withdrawal group on the
Iowa Silent Reading Examination was 126.1, in contrast to

139.7 for the college as a whole (P = .01). There was no sig-

nificant difference in occupational level of the parents.

 

27P. F. Munger, "Factors Related to Persistance in Col-
lege of Students Who Were Admitted to the University of Toledo
from the Lower Third of Their Respective High School Classes"
(unpublished Doctoral dissertation, University of Michigan,
195% .

28L. Snyder, "Why D 0 They Leave?," Journal 2: Higher
Education, XI (January, 19u0),.26-32.

 

18
Johnson, in an analysis of drop-outs at a state col-
lege in Alabama, concluded that a high percentage of the rea-
sons for drop-outs could be corrected by a counseling program
which would assist the student in making academic, financial,
personal, and social adjustments, particularly during his first

semester at the college.29

Subjects for the study were a
representative sample of 1,9”8 students attending the college
as sophomores and a like sample of freshmen who did not return
after the freshman year. Subjects took the American Psycholog-
ical Council Examination, the Iowa Silent Reading Test, and the
California Test of Personality. Grade point averages were

also computed. The only significant difference between the men
who stayed and those who left was found in the grade point av-
erage, which was significant at the .01 level.

Heaton and Weedon, in discussing the failing student,
stated that scores on psychological examinations indicated that
mental ability as measured by such an examination is one impor-
tant factor, but not the only important factor, to be consid-
ered in the prediction of academic success or failure.30 Sub-
jects for the study were 916 students from Albion College,
Central State Teachers College, Michigan State College, and
Olivet College, who took the American Council Psychological Ex-

amination. The author took the national summary of scores for

four year colleges as a basis for classifying the students at

 

296. B. Johnson, "A Pr0posed Technique for the Analysis
of Dropouts at a State College," Journal 3: Educational 52?
search, XLVII (January, 195a), 38I-387.

30K. L. Heaton and V. Weedon, The Failing Student (The
University of Chicago Press, 19u0), p. 37.

 

 

19
Albion, Michigan State, and Olivet and the national summary
for teachers colleges as the basis for classifying students
at Central State Teachers College. The distribution for the
916 students showed that 9.7 per cent of the total were in the
highest one-fourth of freshmen, 21.8 per cent were in the
third fourth, 30.8 per cent were in the second fourth, and
37.8 per cent were in the lowest fourth. Approximately one-
third of the failing students were above the average of all
students in similar colleges on psychological scores, while
two-thirds were below average. The fact that 9.7 per cent, or
89 students of the total group of failing students, scored be-
.tween the 75th and 99th percentile would indicate that those
of superior ability as measured by the test are not uniformly
successful in academic efforts.

Crews compared 22 intellectual and non-intellectual
factors with college success for 326 graduates from the School
of Science at Oregon State College.31 The best predictive fac-
tor was the high school decile. (For students who had been en-
rolled in college one or more terms, the previous college rec-
ord provided the best information available for prognosis.
Selected correlations were as follows: high school decile vs.
success in the College of Arts and Letters, .u15, vs. over-all
college success, .uoz; credit transferred from other colleges
vs. over-all success, .692; first-term college grades vs. first

year grades, .812, and vs. over-all grades, .645.

 

316. T. Crews, "Selected Factors in Relation to College
Success for Science Majors at Oregon State College" (unpublished
Doctoral dissertation, Oregon State College, 1957).

20
Lunn studied the effectiveness of scholastic aptitude
and other selected variables in prediction of success in the
three sequential profesSional education courses of the teacher

education program at the University of Oklahoma.32

The great-
er the scholastic aptitude, the greater the achievement was
likely to be in reading comprehension, English, mathematics,
science, history, and grade point average for the first year
of college. Social class identification, occupation of parent,
and educational level of parents did not affect scholastic
aptitude.

A study at Purdue University by Palcios of 26a college
men and women dealt with an investigation of the validity or
predictive efficiency of a battery of tests selected by the Com-
mittee on Selection and Guidance at that institution.33 The
best single predictor of future academic success was the post
academic achievement as measured by the first year grade point
index (r = .532).

Kramer's study of 2,202 students at Rutgers University

dealt with students who enrolled as freshmen entering directly

from New Jersey public secondary sdhools in 1953, 195M, 1955,

 

32M. S. Lunn, Jr., "The Prediction of Success of Stu-
dents Enrolled in Professional Education Courses at the Univer-
sity of Oklahoma" (unpublished Doctoral dissertation, Univer-
sity of Oklahoma, 1961).

33d. R. Palacios, "A Validation Study of Selected
Tests for Possible Use in Admission to Professional Education
Sequences at Purdue University" (unpublished Doctoral disser-
tation, Purdue University, 1959).

21

and 1956.3” The effectiveness of reported rank in class as a
criterion for admission to Rutgers University was analyzed.
The relationship between reported high school ranks and cumu-
lative averages in college was computed and expressed by a
correlation (r = .393). The results of this study, in con-
trast to some of the earlier studies, showed better predictive
results could be secured if high school ranks were considered
separately for each of the schools.

Giusti investigated the relative importance of curric-
ulum experiences in high school for the prediction of academic
success in the College of Education at Pennsylvania State Uni-

35 Six variables were studied, namely: high school

versity.
index and achievement test scores in English, mathematics, his-
tory, science, and foreign language. Of the six predictive
variables, the high school index was the best predictor of col-
lege grade point average. High school index correlated .u7
with academic success. The combination of all high school sub-
ject fields and high school index yielded a multiple R of .usu.
Waller attempted to find some factors which might be
meaningful for predicting persistence at Trenton State Col-

36

lege. The sample consisted of 58 men and 199 women who per-

sisted to graduation and 72 men and 135 women who withdrew

 

3|lG. A. Kramer, "High School Class Rank and Academic
Performance in College" (unpublished Doctoral dissertation,
Rutgers University, 1959).

35J. P. Giusti, "The Prediction of Academic Success in
a College of Education Based on High school Curriculum Experi-
ences" (unpublished Doctoral dissertation, Pennsylvania State
University, 1962).

36C. Waller, "Predicting Persistence to Graduation at
Trenton State College" (unpublished Doctoral dissertation,
Columbia University, 1962).

22
before graduation from the entering class at Trenton State
College in 1957. Data studied included: curriculum in col-
lege, high school rank, grades in college, high school person-
ality rating, sex, father's occupation, educational level of
parents, geographical location of high school, extra-curricu-
lar activities, College Entrance Examination Board scores or
School and College Ability Test scores (verbal and quantita-
tive). A stepwise regression analysis was made for fourteen
factors known for #35 students of the total sample. First
semester grade average was the best single predictor for per-
sistance with a correlation of .HS. High school rank was the
second best predictor (r = .31). The objective data did not
reveal any factors which would give a high prediction, but
certain data showed a positive relationship to persistance.
These were high school rank, sex, verbal and quantitative test
scores, and the writing test. Socio-economic factors (partic-
ularly educational level of parents and financial resources)
showed some relationship (r = .19).

In summarizing the related research studied by the au-
thor, several generalizations may be made. In every study in-
volving high school rank it was found that this was an impor-
tant criterion for prediction of academic success in college.
Though some studies tended to minimize the importance of psy-
chological test scores as being useful in predicting college)
success, most of the research showed them to be a reliable pre-
dictor. Many investigators made strong pleas for effective

guidance, counseling, and orientation programs. There appeared

23
to be divided views on the importance of education and occupa-
tion of parents in the prediction of academic success. Most
of the research investigating the importance of participation in
extracurricular activities in both high school and college
showed that drop-outs tended to participate less. It was clear-
ly shown that the first year in college is the critical one in
terms of drop-out and withdrawal patterns. The figure of 50
per cent of all university freshmen who fail to graduate from
the institution of first registration was also shown to be
stable. In studies which tended to isolate the reasons for
students' withdrawal, a multiplicity of reasons were listed by
students, but no definite pattern was apparent. Sex, age, or
veteran status were not significantly related to persistence
in college in any of the literature reviewed.

It is clear that much work needs to be done before the
real causes of droP-outs are isolated. The need for develoPing
a more comprehensive and dependable predictive criterion of
college success becomes increasingly more important as college
enrollments continue to soar and financial support of univer-

sities becomes more critical.

CHAPTER III

METHODS OF RESEARCH

This study utilizes correlational procedures and
causal analysis to determine the factors related to academic
success in the university physical education curriculum. Sub-
jects for the study were 12a male students enrolled in the
College of Physical Education at the Chicago Undergraduate
Division of the University of Illinois from 1958 to 1962.
Seventy-two of the students (mean age, 18.27 years) entered
the university directly from high school; 52 students (mean
age, 18.69 years) transferred to the College of Physical Edu-
cation from other colleges within the university. The data
obtained on the groups of students were analyzed separately.

A total of 3H variables were studied. Grade point
average was used as the criterion for academic success. The
grade point average was based on the 5 point scale: 5.0 = A,
u.0 = B, 3.0 = C, 2.0 = D, and 1.0 = E or failure.

Commuting time represented the amount of time students
spent in minutes going to and from school. Approximately 30
per cent traveled by private automobile, while the remaining
70 per cent commuted via public transportation. There were
no housing facilities for students on the campus.

Occupation of father was categorized into one of seven

2”

25

areas; namely, clerical, laborer, managerial, professional,
sales, service occupations, and skilled laborer on the basis
of the Occupation Index.1

Education of father and mother was studied in terms
of number of years of formal education.

Each subject took the School and College Ability Test
(SCAT), which measures ability in verbal and quantitative
skills.2 SCAT yields a verbal score based on 60 items, a
quantitative score based on 50 items, and a total score based
on 110 items. Individual performance was interpreted by means
of converted scores and percentile bands.3

Test scores on the Essential High School Content Bat-
tery were also used.” This test is a comprehensive battery of
high school achievement tests covering mathematics, science,
and social studies. The basic premise underlying this battery
is that there is a common body of knowledge and skills which
it is reasonable to expect a high school graduate to possess.
The mathematics test measures broad understandings of general
mathematical concepts rather than specific "course work."

The science test probes the student's capacity for applying

his scientific knowledge to familiar and unfamiliar situations.

 

1Occupation Index, Dictionary of Occupational Titles,
Vol. 11, Occupational Cla551fications ,—2ndEd., MarEH, I999.

 

2Educational Testing Service, Collegg Ability Test
(Princeton, N.J., 1955), p. l.

31bido ’ p. 20

1‘P. Harry and W . N. Durost, Essential Hi h School
Content Battery (World Book C ompany,—New York, 1), p. 1.

 

26
The social studies test measures the student's graSp of those
concepts considered essential for the effective develOpment
of the world in which he lives. Each of the above listed
tests yielded a single score. Individual performance was in-
terpreted by means of a converted score and a percentile band.

All of the subjects took the Cooperative Reading Com-
prehension Tests.5 These tests are divided into tests of ex-
pression and tests of reading comprehension which includes
vocabulary, reading comprehension, reading Speed, and grammar.
The vocabulary test consists of 60 multiple-choice recognition
items. The reading test consists of 60 items based on se-
lected reading passages.. The grammar test measures several
aspects of correct English by testing grammatical usage, punc-
tuation, capitalization, and spelling. These tests are in-
terpreted using the Chicago Undergraduate Division norms for
entering freshmen.

The subjects also completed the Bild-Dutton Academic
Interest Test, which measures student interest in the areas of
commerce, engineering, mathematics, chemistry, geology, biol-
ogy, sociology, education, history, English, foreign language,

and fine arts.6

In this test the subject expressed his pref-
erences for studying different phases of college subject matter

or doing the work involved in various courses. The subject

 

5Educational Testing Service, The Cooperative Tests
(Princeton, N.J., 1951), p. 3.

6B. Bild and E. Dutton, Academic Interest Inventory
(Chicago: Publication of University of Illinois, Chicago Un-
dergraduate Division, 1951), p. l.

 

 

27
matter and activities were chosen from courses in various aca-
demic departments as cited above, with the h0pe that the
student's preferences might help him decide upon a major
field of specialization or might aid him in the selection of
specific courses.

Extra-curricular activity participation was another
variable studied, though it was confined to intramural sports
and intercollegiate athletics.

Student's rank in high school, both in terms of sten
score and percentile rank, was studied. Sten scores are
standard scores with a base, or range, of ten.

Other variables included grade point average by profes-
sion of father; grade point average by high school attended
index; grade point average by college from which the student
transferred within the University (i.e., Commerce, Engineer-
ing, Liberal Arts and Science); and previous cumulative grade
point average prior to transfer into the College of Physical
Education.

The Physical Fitness Test used was the Army three-item
test consisting of the 250 yard indoor shuttle run, timed sit
ups (two minutes), and floor push ups.7 These test items
measure strength, muscular endurance, cardio-respiratory endur-
ance, agility, and coordination. The same personnel adminis-
tered the test to all students in the study and the same items

were used in all tests. Scoring tables were used to determine

 

7Physical Training, Field Manual 21-20, Department of
the Army Field Manual, November, 1950, chap. xvii, p. 289-305.

 

28
each man's total score in the study sample.

Several variables were quantified; namely, high school
index, grade point average by college from which the student
transferred, and extra-curricular participation. The tech-
nique used for determining the grade point average by high
school attended was to investigate the academic record of stu-
dents who had previously attended the university from the high
schools attended by the students in the sample, some high
schools being better than others. The student from the better
high school usually tends to do better in college than one from
a less good high school. Thus, if it is known that a student
comes from a good high school for college attendance, it is
possible to predict a higher college grade-point average for
him. The relative standings of the high schools may be deter-
mined by finding out how well the student from each high school
does in college. Thus, the high school is evaluated on the
basis of performance of its students in college. After the high
school has been evaluated on the basis of one group of stu-
dents, it is possible to predict the level of achievement for
the next group of students. It is the average student from a
high school that will give the true picture of that high school's
worth. If one takes an average of all students that come from
one high school, it will be a biased sample, weighted too much
on the high or low end. Thus, it is necessary to pick a sample
of students from each high school that will be representative
of the whole class, with an equal number from the upper and

lower ends. In this study the author tried to balance the

29
students equidistant from the M9-l/2 percentile. If one stu-
dent was at the ”9th percentile, another student from the 50th
percentile was chosen. A student at the 38th percentile was
balanced by one from the Blst percentile. Since few students
come from below the 30th percentile, the sample was drawn
from students between the 30th and 70th percentiles. By aver-
aging the grade point average for these students, the writer
obtained, as accurately as possible, a picture of the "typi—
cal" student from each high school, and thus arrived at a quan-
titative value for that high school. The sample of entering
students which was used for this quantifying process was for
the years 1956 through 1958. This was not the same as the
sample for this study. However, this is of minor concern since
the high school values are quite constant.8

The grade point average by profession was quantified by
finding the mean of each of the professions represented. The
grade point average by college from which the student trans-
ferred was determined by computing the mean of the grade point
averages of the students from each college. (There were 15
students who transferred from the College of Commerce, in from
the College of Engineering, and 23 from the College of Liberal
Arts and Science.)

The extra-curricular participation variable was quanti-

fied by coding into three categories with #1 representing

 

8J. W. Creaser, Predicting Colle e Success from Equated
High School Ranks: A Cross Validated Stugy (Chicago: Puinca-
tion of the University or IIIinois Chicago Undergraduate Divi-
sion, 1962), 98 pp.

 

 

30
numeral award, #2 representing letter award, and #3 represent-
ing no participation.

The university student counseling bureau administered
the School and College Ability Test, the Essential High School
Content Battery, the COOperative Reading Tests, and the Bild-
Dutton Interest Inventory Test. These tests were given each
semester during Freshman Week. The results of these tests in
sten scores were punched on IBM cards. Commuting time and oc-
cupation of father were available on the reverse side of stu-
dent class schedules. Extra-curricular participation figures
were available in the departmental office. High school rank
and academic averages were obtained in the college office.

A separate IBM card was made up for each of the 12%
subjects, into which was punched the values for each of the an
variables. The means, standard deviations, and correlations
were completed, using an electronic computer.

Multiple R was calculated by hand, using the Jenkins

9

technique. The multiple correlation was arrived at by choos-

ing those variables which would seemingly have the most effect
on the Multiple R. This technique was adapted from Jenkins.10
The multiple regression equation, using beta weights,

means, and standard deviations, was used as described by Mc-

Nemar.11 These computations resulted in a predictive equation.

 

9W. L. Jenkins, "Quick Estimation of Multiple R," ng
ucational and Psychological Measurement, X (Summer, 1950),
No. 2, Bus-3E8.

10Ibid., p. 3u7.

11Q. McNemar, Psychological Statistics (New York: John
Wiley and Sons, 19u9), p. 182.

 

 

31
The fluctuations in a given series were seldom dependent upon
a single factor or cause.12
Since the multiple correlation coefficient is subject

to a positive bias, the magnitude of which depends upon the
degree to which (n) approaches (N), a correction for shrink-
age was made. The relative importance of the individual in-
dependent variables in a multiple correlation was determined

by using the beta coefficients.13

The beta coefficients are comparable measures which
indicate the increase in the dependent variable resulting
from an increase of one standard deviation in eaCh independ-

1H Once the beta coefficients are calculated,

ent variable.
one can (1) readily compute the betas needed in the raw score
form of the predicted equation, (2) determine the value of
the multiple correlation coefficient and the error of esti-
mate, and (3) ascertain the relative importance of the inde—
pendent variables as predictors or, if causation can be as-
sumed, as contributors to the variance of the dependent or
criterion variable.15
The reader is reminded that the multiple correlation

coefficient represents the maximum correlation to be expected

between the dependent variable and a linearly additive combi-

 

12H. Arkin and R. Colton, Statistical Methods (New
York: Barnes and Noble, Inc., 1961), p. 9%.

13

 

Q. McNemar, 92. cit., p. 186.

ll‘iH. Arkin and R. Colton, 22. cit., p. 96.

151bid., p. 97.

32

15 The extent of errors of

nation of independent variables.
prediction is indicated by the standard error of the esti-

mate.17

 

16Q. McNemar, 32. cit., p. 178.

17Ibid.

 

CHAPTER IV

RESULTS: MULTIPLE CORRELATION AND

MULTIPLE REGRESSION

The first variable chosen to be included in the mul-
tiple was the one which had the highest intercorrelation with
the criterion. After this variable was selected, the vari-
able which had a combination of the highest correlation with
the criterion and the lowest positive (or highest negative)
intercorrelation with the variables already included was chos-
en next, since it would add most to the total. Four variables
for both the transfer and regular students were found to give
as much predictive power as could be obtained. The variables
were not the same for the two groups of students.

The correlation matrices are shown in Tables I and 11.
Tables III and IV show the prOgressive increase of R with the
addition of variables. Tables V and VI present the means,
standard deviations, beta weights, and present contribution of
selected variables to the correlation and prediction equations.

Each variable that was chosen added considerably to
the multiple with the exception of high school rank for trans-
fer students. From the magnitude of R, it is clear that
previous college work is a better measure of success in anoth-
er college than any of the tests or high school measures.

33

TABLE I
INTERCORRELATIONS: TRANSFER STUDENTS

 

 

 

o P .x
u) m 5
o o a
rd T10 m
p at mzm o
a 0 sun -p .4
-a c) dis a o
:0 o H o
oaua tn 01> .5
>H o :3 z<c :a
storm 0 o o) co
crown -H -H4J o
523 > ’5 8 a
sea 2 8o o w
v< D-o 0-. O4 (.9 Cl:
Previous College .fl6 - .us .27 .33
Previous Grade
Point Average .HS - -.l3 .3u
Geology Interest .27 - .01
High School Rank . 33 -

 

 

 

 

 

 

R = .65, Gest. = .3 (grade point units)

 

 

 

TABLE II
INTERCORRELATIONS: REGULAR STUDENTS
m 8
v H c
: r4 r! m 044
-a o C) o wt:
cc) C) 0 El (Jo
Oma H nu ma
~4 o o o w o
$c0tm co awn m FI>
388 e.e$ 2 33
use: 3° 29:: t: '85
Uv< m m< h m<
High SCIIOO]. Rank .31 "' -016 010 010
High School At-
tended e 19 - - e 0 7 " e lll’
Fitness Test .23 - .25
Social Science
Achievement .29 -

 

 

 

 

 

 

R = .51, sest. = .3 (grade point units)

A

TABLE III

PROGRESSIVE INCREASE OF MULTIPLE R:
TRANSFER STUDENTS

Variable

Previous College Index .
Previous College Grade Point Average

Geology Interest . .
High School Rank . .

TABLE IV

Mul

PROGRESSIVE INCREASE OF MULTIPLE R:
REGULAR STUDENTS

35

tiple R

.ue
.57
.63
.65

 

 

 

Variable Multiple R
High SChOOl Rank 0 o o e o o o o o o e o o 031
Hi gh S ChOOl Attended 0 . Q . Q O O . O O . . 39
Fitness TeSt o o o e o o o o o o e o e o o 0.45
Social Science Achievement . . . . . . . . .51
TABLE V
TRANSFER STUDENTS: MEANS, STANDARD DEVIATIONS,
AND BETA WEIGHTS
. St. Beta Bet %
Variable Mean Dev. Wt. Wt.§ Contrib.
Grade Point Average
(Criterion). . . . . 3.u .H
Previous College Index 3.u .3 .26 .0676 20.98
Previous Grade Point
Average 0 O O O O O O 2.6 Du .36 .1296 .40.22
Geology Interest . . . 5.8 1.2 .31 .0961 29.82
High School Rank . . . 5.7 1.0 .17 .0289 8.98
.3222 I00.00

 

 

 

 

 

 

College Success (Grade Point Average)
.30 + .35 (Previous College Index)
+ .36 (Previous Grade Point Average)
+ .10 (Geology Interest) + .07 (High

School Rank)

36

TABLE VI

REGULAR STUDENTS: MEANS, STANDARD DEVIATIONS,
AND BETA WEIGHTS

 

 

St. Beta Bet %

Variable Mean Dev. Wt. Wt. Contrib.

 

Grade Point Average
(Criterion) . . .
High School Rank . . .

3. .9
5.
High SChOOl Index 0 o o 20
5.
l4

0

2 1.9 .33 .1089 92.02
5 .3 .27 .0729 28.12
21 .19 .0196 7.52

Fitness Test Score 5
Social Science
Achievement . . . . . .8 1.9 .29 .0579 22.39

.2535 100.00

 

 

 

 

 

 

 

College Success (Grade Point Average) = 1.080 + .096 (High
School Index) + .003 (Fitness Test) + .070 (Social Sci-
ence Achievement).

Since the Multiple R might be different if calculated
for another sample of students, it is important to determine
how much it might vary. Applying the shrinkage formula, the
probable lower limit of R obtained with a new sample is .62
for transfer students and .98 for regular students.1

The multiple regression equation for regular students
has weights of three decimal places, whereas two decimals were
considered adequate with transfer students, the reason for
this being the large values for the fitness test scores rela-
tive to the values of the other scores. For transfer students,
the means and standard deviations of all variables were rea-
sonably similar, and thus their weightings were not too dis-

crepant. The fitness test score, however, was twenty times as

great as the standard deviation of the high school index. The

 

lQ. McNemar, 22. cit., p. 186.

37

small weighting it receives eliminates the score altogether
(by rounding to zero), unless carried out to three decimals.
The regression equations present the prediction of

grade point average to .3 grade points for the transfer and

regular students. The formulas should be tested on future

students and other sources of predictive variance sought.

CHAPTER V

RESULTS: CHOSEN VARIABLES

This chapter presents a discussion of the variables
used in obtaining the multiple correlation. As was pointed
out in Chapterllh the variables chosen for the 72 regular
students were high school rank, index of high school attend-
ed, fitness test, and social science achievement. High school
rank had the highest coefficient of correlation with the cri-
terion (.31). These results add further support to the val-
ue of rank in high school as one of the best predictors of
academic success in college.

The sample of 72 students came from 91 different high
schools--93 from public high schools in Chicago, 18 from par-
ochial and private schools in Chicago, and 11 from suburban
schools. The caliber of education in these schools varied
considerably. For this reason the index was developed which
reflected the success of the graduates of the various high
schools.

The physical fitness test correlated (r) with the cri-
terion (r = .23 and r = .21 respectively for regular and
transfer students).

A comparison was made of the top twenty regular stu-
dents (grade point averages) with the low twenty regular stu-

38

39
dents. The mean fitness test scores were 60.5 for the former
and 59.6 for the latter. The difference was not statistically
significant. However, the top twenty transfer students scored
significantly higher on the fitness test than the t0p twenty
regular students (t = 9.95; P = .01 level). The low twenty
transfer students also made significantly higher scores on the
fitness test than the low twenty regular students (t = 9.95).
The top twenty transfer students also scored significantly
higher on the fitness test than the low twenty transfer stu-
dents (t = 3.96).

The problem of commuting added considerably to the
length of the school day for most students (mean: regular stu-
dents, 117.50 min.; transfer students, 117.69 min.). This loss
of time made the school day longer and may have had implica-
tions relative to fitness.

Social science was another variable used in the combi-
nation of those variables contributing most to the Multiple R.
An even higher correlation was expected because students major-
ing in Physical Education are for the most part socially-ori-
ented individuals with out-going personalities. Over 90 per
cent of the students in the sample had camping or scouting ex-
periences which would tend to make them more skilled in human
relationships even though they might not have had the actual
social studies curriculum background.

The four variables chosen for transfer students were
somewhat different than those for regular students, with the

exception of high school rank. The other three variables

90
which contributed most to the multiple correlation were pre-
vious college index, previous grade point average, and geology
interest. Correlations were highest between the criteria and
previous college index (r = .96) and previous grade point av-
erage (r = .95). High school rank (r = .33) and geology in-
terest (r = .27) followed in order.

The previous college index was quantified by determin-
ing the average grade (mean) for the students during their
first semester in the College of Physical Education according
to their previous college. The attrition of students who
transferred to the College of Physical Education from other
colleges within the university was exceptionally low. Only
six students of the total of 52 were dr0pped for poor scholar-
ship at the end of their first semester, a figure considerably
lower than the all-university figure of 18 per cent. The rea-
sons advanced for this low attrition rate are (1) added mo-
tivation of students in the new college, (2) prior experience
in university work and thus possible better study habits and
wiser utilization of time, (3) student-faculty advisory pro-
gram within the College of Physical Education, and (9) extra
maturity over regular students--approximately six months.

The correlation of high school rank and the criterion were al-
most identical among regular and transfer students. The grade
point average by previous college and grade point average dur-
ing the semester before transfer had high correlations with
the criterion (r = .96 and r = .95 respectively). This was

expected because these students had already survived the

91
first semester of study and also because these provide a meas-
ure of previous college success in the same university.

The correlation of geology interest with the criter-
ion (r = .27) is an interesting phenomenon since with the
regular students the correlation was only r = .01. It is
understandable that in the case of evaluation of interest bat-
tery information, students are most likely to express an in-
terest in areas in which they had already been accepted or
were contemplating entrance. There may be a reasonable expla-

nation, but it is not clear from these data.

CHAPTER VI

RESULTS: OTHER VARIABLES

Many variables that the author presumed would be sig-
nificant, following review of the literature at the beginning
of the study, proved to be of little value. Commuting time,
for example, which had a mean of 117.50 minutes for regular
students and 117.69 for transfer students, correlated poorly
with academic success.

Extracurricular participation (limited to intramural
and intercollegiate sports competition) was negatively corre-
lated with academic success (r = -.10 and r = -.20) for regu-
lar and transfer students respectively. This is in disagree-
ment with most of the previous work. The explanation of this
pheonomenon may be associated with the problem of commuting.
The time spent in participation in sports activities when added
to commuting time considerably reduces study time. The prob-
lems faced by students attending urban universities are much
different from those of students who live on or adjacent to
the campus, since time spent in commuting cannot be used pro—
ductively.

The occupation of the father was unrelated to academic
success (r = .00 for transfer students and r = .09 for regu-
lar students). The skilled laborer classification, which

92

93
dominated the study sample, represented 91 per cent of the
sample of regular students and 37 per cent of the sample of
transfer students. The service occupation comprised 22 per
cent of the regular students sample and 19 per cent of the
transfer students. Twenty-one per cent of the transfer stu-
dents' fathers were in managerial positions, and seven per
cent of the students in the regular group had fathers in this
category. In a statistical analysis of the occupation of the
fathers of regular and transfer students, it was found that a
significantly greater number of the fathers of the transfer
students came from the professional and managerial occupa-
tions (X2 = 8.051, P = less than .05). The occupations were
grouped for analysis into four classifications; namely, sales
and service, professional and managerial, laborer and skilled
laborer, and clerical.

The education of the father for both regular and trans-
fer students correlated .09 with the criterion. Education of
the mother correlated .10 for regular students and .05 for
transfer students.

There were several high intercorrelations between var-
iables for both groups of students. However, the variables
were not necessarily the same for regular and transfer stu-
dents. The highest in both categories were high school rank
by sten score and high school rank by percentile (r = .98 for
regular students and r = .97 for transfer students). These
should be high because they are measures of the same thing.

Table VII presents these higher intercorrelations be-

nu
tween variables for regular and transfer students. The dif-
ferences between the correlations were tested for significance
using the r to z transformation and testing with the "t"

test.1 None of the correlations were significantly different.

TABLE VII

INTERCORRELATIONS BETWEEN VARIABLES:
REGULAR AND TRANSFER STUDENTS

 

 

 

 

 

Regu- Trans- t
lar fer

High School Rank High School Rank

(sten score) (percentile) .98 .97 -l.15
Verbal Learning Soc.Sci.Achieve. .59 .62 - .27
Verbal Learning Vocabulary .56 .57 -1.02
Quantitative Mathematics

Learning Achievement .59 .68 -1.30

 

 

The twenty high and low (academically) regular and
transfer students were compared on the twenty-three variables
shown in Tables VIII and IX using the "t" test. Two variables
in each category (regular and transfer students) were found
to be significant (high school rank and fitness test score).
In each instance the scores of the transfer students were sig-

nificantly higher.

 

1H. M. Walker and J. Lev, Statistical Inference (New
York: Holt, Rinehart and Winston, 1953), p. 259.

 

95

TABLE VIII

SIGNIFICANCE OF DIFFERENCES BETWEEN MEANS: TOP TWENTY
TRANSFER AND REGULAR STUDENTS

 

 

 

Mean Mean
Variable (regular (transfer t
students) students)

High School Rank ($113) 51.7 63.8 9.30 (.01)
Verbal Learning (sten) 5.2 5.5 .19 (NS)
Quantitative Learning 9.6 6.3 1.95 (NS)
Mathematics Achievement 9.7 6.3 1.95 (NS)
Natural Science

Achievement 9.5 5.5 1.19 (NS)
Social Science

Achievement 5.1 5.6 .60 (NS)
Vocabulary 5.0 5.1 .11 (NS)
Reading Comprehension 9.9 9.9 .51 (NS)
Reading Speed 9.9 5.5 .67 (NS)
Grammar 5.2 5.2 .00 (NS)
Fitness Test Score 60.5 75.2 9.95 (.01)
Commerce Interest 9.6 5.0 .92 (NS)
Engineering Interest 9.5 5.6 1.09 (NS)
Mathematics Interest 3.9 5.3 1.62 (NS)
Chemistry Interest 9.0 5.5 1.79 (NS)
Geology Interest 9.9 6.2 1.97 (NS)
Biology Interest 5.9 5.8 .99 (NS)
Sociology Interest 6.0 9.8 1.30 (NS)
Education Interest 6.7 5.2 1.79 (NS)
History Interest 5.6 9.8 .85 (NS)
English Interest 9.9 5.2 .30 (NS)
Foreign Language

Interest 9.3 9.8 .62 (NS)
Fine Arts Interest 9.6 5.5 .95 (NS)

 

96

 

 

 

TABLE IX
SIGNIFICANCE OF DIFFERENCES BETWEEN MEANS: LOW TWENTY
TRANSFER AND REGULAR STUDENTS
Mean Mean
Variable (regular (transfer t
students) students)

High School Rank (%ile) 32.7 99.2 9.15 (.01)
Verbal Learning 9.9 5.9 1.19 (NS)
Quantitative Learning 9.9 5.7 .88 (NS)
Mathematics Achievement 9.8 5.7 1.08 (NS)
Natural Science

Achievement 9.3 5. 1.18 (NS)
Social Science

Achievement 9.8 5.0 .02 (NS)
Vocabulary 9.8 5.1 .92 (NS)
Reading Comprehension 9.6 9.9 .29 (NS)
Reading Speed 9.7 9.9 .22 (NS)
Grammar 5.1 9.5 .69 (NS)
Fitness Test Score 59.6 70.7 9.81 (.01)
Commerce Interest 9.9 6.0 1.71 (NS)
Engineering Interest 9.7 5.9 .86 (NS)
Mathematics Interest 9.1 9.9 .93 (NS)
Chemistry Interest 9.2 5.2 1.21 (NS)
Geology Interest 5.1 5.3 .10 (NS)
Biology Interest 5.3 5.0 .30 (NS)
Sociology Interest 5.8 5.6 .22 (NS)
Education Interest 6.0 5.6 .98 (NS)
History Interest 5.0 5.5 .99 (NS)
English Interest 5.0 5.6 .71 (NS)
Foreign Language

Interest 9.9 9.9 .55 (NS)
Fine Arts Interest 9.5 5.2 .88 (NS)

 

97
Table X presents a comparison between the twenty high
and low (grade point average) regular students. High school
rank was found to be significantly higher for the high aca-

demic group.

TABLE X

SIGNIFICANCE OF DIFFERENCE BETWEEN MEANS: TOP TWENTY
REGULAR AND LOW TWENTY REGULAR STUDENTS

 

 

 

Mean Mean
Variable (High (Low t
Regular) Regular)

High School Rank ($118) 51.7 32.7 2.25 (.05)
Verbal Learning 5.2 9.9 .96 (NS)
Quantitative Learning 9.6 9.9 .29 (NS)
Mathematics Achievement 9.7 9.8 .12 (NS)
Social Science

Achievement 5.1 9.8 .37 (NS)
Vocabulary 5.0 9.8 .23 (NS)
Reading Comprehension 9.9 9.6 .21 (NS)
Reading Speed 9.9 9.7 .29 (NS)
Grammar 5.2 5.1 .11 (NS)
Fitness Test Score 60.5 59.6 .28 (NS)
Commerce Interest 9.6 9.9 .23 (NS)
Engineering Interest 9.5 9.7 .25 (NS)
Mathematics Interest 3.9 9.1 .22 (NS)
Chemistry Interest 9.0 9.2 .23 (NS)
Geology Interest 9.9 5.1 .22 (NS)
Biology Interest 5.9 5.3 .10 (NS)
Sociology Interest 6.0 5.8 .22 (NS)
Education Interest 6.8 6.0 .83 (NS)
History Interest 5.6 5.0 .59 (NS)
English Interest 9.9 5.0 .11 (NS)
Foreign Language

Interest 9.3 9.9 .10 (NS)
Fine Arts Interest 9.6 9.5 .10 (NS)

 

The twenty high and low (grade point average) trans—
ger students were compared in Table XI. High school rank and
fitness test scores were found to be significantly higher for

the high ranking students academically.

98

TABLE XI

SIGNIFICANCE OF DIFFERENCE BETWEEN MEANS: TOP TWENTY
TRANSFER AND LOW TWENTY TRANSFER STUDENTS

 

 

 

Mean Mean
Variable (High (Low t
Transfer) Transfer)

High School Rank (%ile) 63.8 99.2 6.29 (.01)
Verbal Learning 5.5 5.9 .11 (NS)
Quantitative Learning 6.3 5.7 .66 (NS)
Mathematics Achievement 6.3 5.7 .69 (NS)
Natural Science

Achievement 5.5 5.9 .11 (NS)
Social Science

Achievement 5.6 5.0 .71 (NS)
Vocabulary 5.1 5.1 .00 (NS)
Reading Comprehension 9.9 9.8 .11 (NS)
Reading Speed 5.5 9.9 .68 (NS)
Grammar 5.2 9.5 .79 (NS)
Fitness Test Score 75.2 70.7 3.96 (.01)
Commerce Interest 9.2 6.0 1.33 (NS)
Engineering Interest 5.6 5.9 .23 (NS)
Mathematics Interest 5.3 9.9 .99 (NS)
Chemistry Interest 5.5 5.2 .23 (NS)
Geology Interest 6.2 5.3 1.01 (NS)
Biology Interest 5.8 5.0 .82 (NS)
Sociology Interest 9.8 5.6 .71 (NS)
Education Interest 5.2 5.6 .97 (NS)
History Interest 9.8 5.5 .75 (NS)
English Interest 5.2 5.6 .92 (NS)
Foreign Language

Interest 9.8 9.9 .01 (NS)
Fine Arts Interest 5.5 5.2 .33 (NS)

 

The results of the "t" test of significance substan-
tiate the importance of high school rank and fitness for both
groups of students (regular and transfer). The probability
of one out of twenty variables being significantly different
due to chance at the .05 level is one in twenty. With four
samples, the one that would be significant would normally be

different for each sample. The fact that the variables were

99
the same (high school rank and fitness test score) in each

case could hardly be due to sampling.

CHAPTER VII

LOOKING AT STUDENT WITHDRAWALS

The original study began with a total of 177 stu-
dents. However, during the course of the four-year period
(1958-1962) there were many students who withdrew before com-
pleting one semester. The reasons advanced by these students
for leaving school were many and varied, as is the case in
any withdrawal situation. W. P. Pillsbury, Dean of Students
at Knox College in Illinois, stated in an address to that
student body that parents, students, and educators are falling
down somewhere when half of our high school seniors in this
country who demonstrate their ambitions for higher education
by enrolling in colleges change their minds in midstream or
become discouraged or submit to outside attractions, and fail
to finish the college training which they so hOpefully began
as freshmen.l- It is obvious that college withdrawals cannot
be completed eliminated, but the author-~like many others in
education--be1ieves that they can be reduced.

In a nation-wide survey by the United States Depart-
ment of Health, Education and welfare, slightly less than 90

per cent of the freshman class remained at the institution of

 

1W. P. Pillsbury, "Address to Knox College Students,"
Chicago Tribune, July 12, 1962.

50

51
first enrollment to graduate four years later.2 Further data
from this study revealed the great majority of reasons given
for withdrawal referred to personal problems and deficiencies
rather than to deficiencies in the facilities of the institu-

3 This indicated the importance of greater attention to

tion.
services which can assist in solving these problems. This
study, which included a sample of 13,700 men and women, also
revealed some interesting figures on retention of interest in
subject-fields of interest.” Men whose initial interests were
in subject-fields of engineering, physical education, business
administration, and agriculture were least likely to change

to other subject-fields, and more changes were to related
fields than to unrelated fields.5 Student reports of reasons
for going to college, interests in subject-fields, and finan-
cial reasons, plus institutional reports of student standing

in high school graduating class and on college placement tests,
showed that many students enrolled in institutions of higher
education in which the prospects of completing their programs

6 A majority of students who discontinued

of study were poor.
their higher educational programs attributed their withdrawal
to factors identified with themselves rather than with the

institutions they attended.7

 

2United States Office of Education Bulletin, 22. cit.,
p. 100. “‘

3 - 9 .

Ibid., p. 106. Ibid., p. 109.

6

5Ibid., p. 109. Ibid., p. 109.

71bid., p. 109.

52

A questionnaire survey was conducted at the Univer-
sity of Illinois, Chicago, among 189 men students who with-
drew from the university between April 15, 1961 and June 1,
1962 (all colleges). The questionnaire used was one-half
page in length, requiring short answers and no signature.
Only six of the total respondents listed lack of teacher help
as the reason for withdrawal. Table XII presents a summary of
their reasons for withdrawal and the number of students in

each category.

TABLE XII
REASONS WHY STUDENTS WITHDRAW

Personal 0 e e o o o e o
Illness o e o e o e e e
Failing academically . .
Poor study habits . . .
Poor college choice . .
Military SerViCe e e o e o o
Transferring to another college 7
Lack of teacher help . . . . . . 6

TOtal o e o o e .18”

ace...
'0
CD

This survey substantiated the findings of the United States
Department of Health, Education, and Welfare cited above. It
should also alert all those persons who work directly or in-
directly with students to the many problems faced by the col-
lege student. The withdrawal problem is a myriad one and
needs to be continually studied and evaluated in the light of
all information that is available.

Table XIII presents a comparison of 22 variables be-
tween 22 regular students who withdrew and 72 students who re-

mained in school.

TABLE

XIII

53

SIGNIFICANCE OF VARIABLES BETWEEN REGULAR STUDENTS WHO

WITHDREW AND REGULAR STUDENTS WHO REMAINED IN SCHOOL

 

 

 

9 a lrt
m molo
F1 raazo
s a .5
3°: :92... 3
Variable n: 3 a: 3 c: 8 -.-c
pals samwa : .9
049a» 0+J a: -a
tHc:s 94:13 A .3
8'“ 88 .2 a
5 :15 5 5.3 94 0
0+le o+J vi h
ZmB sz Q a.
High School Rank (sten score) 9.38 5.19 .81 .05
Extra-curricular Participation 2.90 2.59 .36 .001
Occupation of Father 9.90 5.99 1.09 NS
Education of Father 10.33 10.75 .92 NS
Education of Mother 10.71 11.19 .93 NS
Verbal Learning 3.95 9.79 .89 .05
Quantitative Learning 9.09 9.88 .83 NS
Mathematics Achievement 3.85 9.78 .92 .05
Natural Science Achievement 3.38 9.39 1.00 NS
Social Science Achievement 9.09 9.82 .78 NS
Vocabulary 9.09 5.07 .98 .05
Reading Comprehension 3.76 9.72 .96 NS
Reading Speed 9.61 5.01 .90 NS
Grammar 3.76 9.92 1.16 .05
Grade Point Average by Pro-
fession 3.08 3.05 .03 NS
Grade Point Average by High
School 2.61 2.52 .11 NS
Commuting Time 98.57 117.50 18.93 .01
Commerce Interest 5.19 9.69 .50 NS
Engineering Interest 9.00 9.09 .09 NS
Mathematics Interest 3.19 9.08 .99 .05
Chemistry Interest 9.09 9.22 .18 NS
Geology Interest 9.33 9.98 .63 NS
Biology Interest 5.19 5.18 .09 NS
Sociology Interest 5.80 5.70 .10 NS
Education Interest 7.00 6.25 .75 .05
History Interest 5.82 5.70 .12 NS
English Interest 9.85 5.10 .25 NS
Foreign Language Interest 9.57 9.99 .13 NS
Fine Arts Interest 9.76 9.51 .25 NS

 

 

 

 

 

59

The one variable of the twenty-eight studied that was
significant at the .001 level (extracurricular participation)
indicated that students who participate in intercollegiate
and intramural Sports are more likely to stay in school.

These data were coded so that the lower value indicates great-
er participation. The students who withdrew might have been
more likely to remain, however, if there had been some group
or team attachment with the accompanying motivation that is
generally present. The socialization process that occurs in
extracurricular activities has a "holding power" for many stu-
dents and, despite the fact that students at this age are more
selective of their groups and idols, the urge to join and be

a member of a group is still strong in most instances.

The variables significant at the .05 level, namely,
high school rank, verbal learning, mathematics, vocabulary,
and grammar, are inter-related with the academic process and
are in the expected direction.

There was a considerable difference in commuting time
for students who withdrew and regular students who completed
the semester in favor of the withdrawing students. That is,
withdrawing students spent an average of 18.93 minutes less in
commuting time than the regular students. This was in the op-
posite direction than had been eXpected.

The two variables significant in the interest inven-
tory category were mathematics and education. The mathematics
differences were in favor of the regular student. The educa-

tion category of the interest inventory was significantly in

55
favor of the withdrawing student even though they did with-
draw. It is a paradox that they reflected significantly great-
er educational interest. This sheds little light on their
withdrawal, but rather intensifies the interest as to the
real reasons why they withdrew. Mathematics achievement was
also statistically significant.

The findings in this section of the study strengthen
the premise that high school rank is an important criterion in
predicting academic success for students who withdrew as was
the case for regular and transfer students in other sections
of the study. A follow-up of the present disposition of all
students in the study which appears in Tables XIV and XV sup-
ports the findings of the study of the United States Office of
Education that 60 per cent are still in school or graduated

and 90 per cent are no longer in school.8

TABLE XIV

DISPOSITION OF SEVENTY-TWO REGULAR STUDENTS

 

 

 

Present Status No. of Students % of Total
Graduated from a

Four Year School 7 .09
Still in School 37 .52
No Longer in School 28 .39

 

8Office of Education, 22. cit., p. 99.

56

TABLE XV
DISPOSITION OF FIFTY-TWO TRANSFER STUDENTS

 

 

Present Status No. of Students % of Total
Graduated from a

Four Year School 3 .09
Still in School 28 .59

No Longer in School 21 .92

CHAPTER VIII

FINDINGS, CONCLUSIONS, AND RECOMMENDATIONS

Findings

This study of 129 male students majoring in physical
education over a four year period (1958-1962) has shown the
predictive value of several variables as they relate to aca-
demic success. The four variables resulting in a Multiple R
of .62 for transfer students were previous college index, pre-
vious college grade point average, geology interest, and high
school rank. The four variables with the highest Multiple R
for regular students were high school rank, high school at-
tendance index, fitness test score, and social science achieve-
ment. The Multiple R for these variables was .98. The shrink-
age formula was applied to all computations. A prediction
equation was derived from the multiple regression analysis.
Despite the fact that students in the sample came from over 90
different high schools--both public and private--and repre-
sented many different socio-economic backgrounds, high school
rank was a reliable criterion in predicting academic success.
This variable was important in the study samples of both regu-
lar and transfer students.

The high correlations between variables other than
the criterion followed a logical pattern. Some of the high

57

58
correlations were verbal learning with social science achieve-
ment, verbal learning with vocabulary, and quantitative learn-
ing with mathematics achievement. The pattern of high cor-
relations between variables other than the criterion were
much the same for both regular and transfer students with the
exception of education of mother and father, which had an r =
.72 for transfer students.

The importance of physical fitness was illustrated by
the inclusion of this variable in the Multiple R for regular
students and the level of significance (.01) between the
means of test scores for both regular and transfer students.

Extracurricular participation correlated negatively
with the criterion for both groups of students. When regular
students who remained in school were compared with students
who withdrew, the regular students was found to be signifi-
cantly more active in extracurricular activities (P = .001).

Commuting time was negatively correlated with academic
success for transfer students.

Statistical analysis revealed that there was a signif-
icant difference between the occupation of the father in the
two groups of students (regular and transfer).

Students who transfer to the College of Physical Edu-
cation from other colleges within the university tend to do
well academically with a correspondingly low attrition rate.

Students who withdrew from the College of Physical
Education, when compared with regular students who remained in

school, tended to rank lower in their high school graduating

59
class, participated less in intramural and intercollegiate
sports, made lower scores on verbal learning tests (School
and College Ability Test), made lower scores on mathematics
achievement tests (Iowa High School Content Battery), made
lower scores on vocabulary and grammar tests (Cooperative
Reading Comprehension Test), and made lower scores on mathe-,
matics interest tests (Bild-Dutton Interest Inventory Test).
Students who withdrew tended to spend less time in commuting
and had higher scores on the education inventory test (Bild-
Dutton Inventory Test) than regular students who remained in
school.

Forty-eight (61 per cent) of the sample of 72 regular
students were in school or had graduated from a four-year in-
stitution at the conclusion of the study. Thirty-one of the
52 transfer students (68 per cent) had graduated from a four-
year institution or were still in school at the completion of

the 1961-62 academic year.

Conclusions

1. High school rank is a valuable criterion in predicting col-
lege success despite the fact that there may be an appre-
ciable difference in the level of academic programs of
high schools from which the students graduate.

2. A formal testing program does not provide ready or complete
answers in predicting college success. It is useful in
making a contribution to the total situation rather than

to a specific case.

3.

60
Male students who transfer to the College of Physical Ed-

ucation from other colleges within the University of
Illinois (Chicago) tend to do well academically provided
they have done satisfactory work prior to enrollment in
the new college.

Transfer students have significantly higher scores on the
physical fitness test and rank higher in their high school
graduating class than regular students.

Students who withdraw from the College of Physical Education
at the University of Illinois (Chicago) have significantly
lower test scores on verbal learning, mathematics achieve-
ment, vocabulary, grammar, and mathematics interest tests
than regular students who remain in school. Withdrawing
students also tended to spend less time in commuting and
less time participating in intramural and intercollegiate
sports when compared with regular students who remained in
school.

The time available for sports participation by the student
attending an urban university is somewhat reduced by the

commuting situation.

Recommendations
There is an acute and immediate need for further predictive
criteria for use in the selection of professional students
in the College of Physical Education at the University of
Illinois (Chicago).
The search should be continued for other variables besides

those used in this study for use in the selection process.

3.

6.

61
Because the first year of college is the most critical
drop-out period, every effort should be made to provide
counseling and advisory services, particularly by the col-
lege, to the first-year student. Faculty members should
arrange not only to be available for student consultation,
but should be interested enough in every student with
whom he comes in contact to make a contribution to the stu-
dent's total educational experience.
Because of the apparent effect of physical fitness on aca-
demic success, it is recommended that physical fitness
activities should be included to a greater extent in the
course content of the professional program in physical edu-
cation. Those activities which best contribute to the
components of physical fitness (endurance, strength, power,
agility, and coordination) should be emphasized in all phys-
ical education courses.
It is further recommended that a physical fitness test be
given to all incoming students into the professional physi-
cal education program prior to or shortly after registra-
tion and that these test scores be used constructively by
faculty members in the college.
The present advisory program in the College of Physical Ed-
ucation at the University of Illinois (Chicago) in which

each staff member serves as an advisor should be continued.

BIBLIOGRAPHY

A. BOOKS

American Association for Health, Physical Education and Recre-
ation. Research Methods Applied to Health, Physical
Education and Recreation. lst edT: I999.

 

 

Arkin, A., and R. Colton. Statistical Methods. New York:
Barnes and Noble, Inc., 1961.

 

Campbell, W. A Form Book for Thesis Writing. Boston: Hough-
ton MiffIin Co., 1939.

 

 

Edwards, A. L. Statistical Methods for the Behavioral Sci-
ences. New—York: Rinehart afid Co., 1959.

 

. Statistical Analysis. New York: Rinehart and Co.,
1959.

 

Freyer, H. C. Elements of Statistics. New York: John Wiley
and Sons, Inc., I959.

 

Good, C. V., A. S. Barr, and D. C. Scates. The Methodology 22
Educational Research. New York: D. Appleton-Century
Co., 1935.

 

 

Heaton, K. L., and V. Weedon. The Failin Student. Chicago,
111.: The University of_Chicago ress, 1990.

 

McNemar, Q. Ps cholo ical Statistics. New York: John Wiley
and Sons, I995.

 

Occupation Index. Dictiona 2: Occupational Titles, Vol. 11,
Occupational CIas31f1cations.5_2nd ed. Washington,
D.C.: Government Printing Office, 1999.

Physical Training, Field Manual 21-20. Department 22 the Army
Field Manual. November, 1950.

 

Walker, H. M., and J. Lev. Statistical Inference. New York:
Holt, Rinehart and Winston, 1953.

 

Yates, F., and R. Fisher. Statistical Tables. London: Oli-
ver and Boyd, 1999.

 

62

63
B. PERIODICAL LITERATURE
Cumings, E. C. "Causes of Student Withdrawals at DePauw Uni-

versity," School and Society, LXX (S eptember 3, 1999),
152-153.

 

Fullmer, D. W. "Success and Perseverance of College Students,"
Journal 2: Higher Education, XXVII (November, 1956),

 

Fults, R. 8., and S. E . Taylor. "Staying Power of College
Students," National Association of Seconda Schools

Principals' Bulletin, XLIII (OctEBer, 9 , TUB-I19.

Iffert, R. "The Student Retention and Withdrawal Study,"
College and University, XXX (July, 1955), 906-911.

 

 

Jenkins, W. L. "Quick Estimate of Multiple R," Educational
Psychology Measurement, X (Summer, 1950), 2.

 

 

Johnson, G. B., Jr. "A Proposed Technique for the Analysis of
Dropouts at a State College," Journal of Educational
Research, XLVII (January, 1959), 38I-3BV.

 

Koelsche, C. L. "A Study of the Student DrOp-out Problem at
Indiana University," Journal 22 Educational Research,
XLIX (January, 1956),-3373969.

 

Landskov, N. L. "Suggested Student Survival Techniques Recorded
at the University of Minnesota," College and Univer-
sity, XXIII (June, 1998), 235-236.

 

Linns, L. J., and H. Pitt. "Staying Power and Rate of Progress
of University of Wisconsin Freshmen," Collegg and Uni-
versit , XXIX (October, 1953), 98.

 

Nardelli, W. "An Analysis of DroP-Outs of Freshmen," Junior
Collegg Journal, XXIX (February, 1959), 322.

 

Quarles, B. "Student Separations from School," Association 22
American Colleges Bulletin, XXXV (October, 19997, 908.

 

 

Shuman, R. B. "College Dropouts: An Overview," Journal of E2-
ucational Psychology, XXIX (August, 1956), 397-350.

 

Slocum, W. L. "Social Factors Involved in Academic Mortal-
ity," College and University, XXXII (September, 1956),
53-59.

 

Snyder, L. "Why Do They Leave?," Journal 2: Higher Education,
XI (1990), 26-32. "“"‘

69

Tesseneer, R. A., and L. M. Tesseneer. "Review of the Liter-
ature on School DrOpouts," National Association of
Secondary School Principals Bu Iletin, XLII (May,—
19587: 191-153.

 

United States Office 2: Education Bulletin. "Retention and
Withdrawal of College Students," XV (1958), 1 - 109.

 

 

Yoshimo, R. "College Dropouts at End of Freshman Year,"
Journal of Educational Sociology, XXXII (September,
I958), 292.

 

C. UNPUBLISHED MATERIALS

Cape, W. J. "A Study of Selected Characteristics of the Drop-
outs at Dillard University." Unpublished Doctor of
Education dissertation, Indiana University, 1958.

Crews, G. T. "Selected Factors in Relation to College Success
for Science Majors at Oregon State College." Unpub-
lished Doctor of Education dissertation, Oregon State
College, 1957.

Faunce, L. D. "A Study of Within-Term Dropouts at Michigan
State University for the School Years, 1997-99." Un-
published Doctor of Education dissertation, Michigan
State University, 1952.

Fenelon, W . J. "A Study of the Secondary School Dropout Pat-
tern at Port Washington, Wisconsin HighSchool." Un-
published Doctor of Education dissertation, North-
western University, 1960.

Giusti, J. P. "The Prediction of Academic Success in a College
of Education Based on High School Curriculum Experi-
ence." Unpublished Doctor of Education dissertation,
Pennsylvania State University, 1962.

Hanks, C. J. "A Comparative Study of Factors Related to Reten-
tion and Withdrawal of Freshman Students at the Uni-
versity of Arkansas." Unpublished Doctor of Education
dissertation, University of Arkansas, 1959.

Kramer, G. A. "High School Class Rank and Academic Perform-
ance in College." Unpublished Doctor of Education
dissertation, Rutgers University, 1959.

Lowe, W. T. "Factors Related to Attendance or Non-Attendance
by Capable Illinois High School Graduates." Unpub-
lished Doctor of Education dissertation, University of
Illinois, 1961.

65

Lunn, M. S. "The Prediction of Success of Students Enrolled

in Professional Education Courses at the University
of Oklahoma." Unpublished Doctor of Education disser-
tation, University of Oklahoma, 1961.

Lynch, D. F. "An Analysis of Dropouts in Selected Public

Junior Colleges of Florida." Unpublished Doctor of
Education dissertation, Pennsylvania State University,
1959.

Marsh, R. E., Jr. "An Analysis of Failure among University

Freshmen." Unpublished Doctor of Education disserta-
tion, Boston University, 1959.

Metder, S. M. "A Study of Selected Characteristics of the

Mowers,

Munger,

Male Graduates and Scholastic Dropouts of the 1951
Freshman Class Entering the State University of New
York Teachers College at Cortland." Unpublished Doctor
of Education dissertation, New York University, 1959.

G. E. "Self-Judgments and Objective Measures as Relat-
ed to First Semester Academic Achievement of Non-Se-
lected College Students." Unpublished Doctor of Edu-
cation dissertation, Pennsylvania State University, 1959.

P. F. "Factors Related to Persistence in College of
Students Who were Admitted to the University of Toledo
from the Lower Third of Their Respective High School
Classes." Unpublished Doctor of Philosophy disserta-
tion, University of Michigan, 1959.

Palacios, J. R. "A Validation Study of Selected Tests for Pos-

Patton,

Rogers,

sible Use in Admission to Professional Education Se-
quences at Purdue University." Unpublished Doctor of
Education dissertation, Purdue University, 1959.

B. K. "A Study of Dropouts from the Junior Division of
Louisiana State University." Unpublished Doctor of
Philosophy dissertation, Louisiana State University,
1958.

L. L. "Problem Analysis Study of Selected Freshmen
Students at Appalacian State Teachers College." Unpub-
lished Doctor of Education dissertation, University of
Tennessee, 1959.

Russell, J. W. "A Comparison of Michigan State College First

Term Dropouts and Non-Dropouts According to Certain
Factors." Unpublished Doctor of PhilOSOphy disserta-
tion, Michigan State College, 1952.

66

Waller, C. "Predicting Persistence to Graduation at Trenton

State College." Unpublished Doctor of Education dis-
sertation, Columbia University, 1962.

D. NEWSPAPERS
Chicago Tribune, July 12, 1962.

 

E. PARTS OF SERIES

Bild, B., and E. Dutton. Academic Interest Invento . Chica-
go: Publication of the University of IIIihOls, Chica—
go Undergraduate Division, 1951, p. l.

 

Creaser, J. W. Predicting College Success from Equated High
School Ranks: A Cross Validated Stud‘. Chicago:
Ufiiversity 6f I1I1n01s, Chicago Unaergraduate Division
Publication, Student Counseling Service, February,
1962.

 

Educational Testing Service. College Ability Test. Princeton,
N.J.: 1955, p. 1. “'

 

. The Cooperative Tests. Princeton, N.J., 1951, p. 3.

Harry, F., and W. N. Durost. Essential High School Content
Battepy. New York: World Book Company, 1951, p. l.

 

APPENDIX

MEANS AND STANDARD DEVIATIONS: REGULAR STUDENTS

 

Standard
Variable Mean Deviation
Grade Point Average . . . . . . . 2.98 .36
Commuting time . . . . . . . . . 117.50 31.17
Extracurricular participation . . 2.59 .51
Occupation of father . . . . . . 5.99 2.19
Education of father . . . . . . . 10.75 2.38
Education of mother . . . . . . . 11.19 1.96
Verbal Learning . . . . . . . . . 9.99 1.56
Quantitative learning . . . . . . 9.88 1.56
Mathematics achievement . . . . . 9.78 1.92
Natural Science achievement . . . 9.39 1.96
Social Science achievement . . . 9.82 1.99
Vocabulary . . . . . . . . . . . 5.07 1.95
Reading comprehension . . . . . . 9.72 1.38
Reading speed . . . . . . . . . . 5.01 1.67
Grammar . . . . . . . . . . . . . 9.92 1.63
Fitness test . . . . . . . . . . 55.22 21.38
Commerce interest . . . . . . . . 9.69 1.93
Engineering interest . . . . . . 9.69 1.19
Mathematics interest . . . . . . 9.08 1.95
Chemistry interest . . . . . . . 9.22 1.32
Geology interest . . . . . . . . 9.96 1.73
Biology interest . . . . . . . . 5.18 1.61
Sociology interest . . . . . . . 5.71 1.56
Education interest . . . . . . . 6.25 1.55
History interest . . . . . . . . 5.31 1.75
English interest . . . . . . . . 5.10 1.51
Foreign language interest . . . . 9.99 1.61
Fine Arts interest . . . . . 9.51 1.99
High School Rank (Sten score) . 5.19 1.90
Grade Point Average (profession) 3.09 1.19
Grade Point Average (H. 8. index) 2.51 .27

High School Rank (percentile) .99 .20

'-

L

MEANS AND STANDARD DEVIATIONS: TRANSFER STUDENTS

 

 

Standard

Variable Mean Deviation
Grade Point Average . . . . . . . . 3.35 .90
Commuting time . . . . . . . . . . 117.69 27.33
Extracurricular participation . . . 2.25 .97
Occupation of father . . . . . . . 5.59 1.79
Education Of father 0 o o o o o o e 10.98 2.19
Education of mother . . . . . . . . 11.09 1.89
Verbal learning . . . . . . . . . . 5.33 1.28
Quantitative learning . . . . . . . 5.81 1.52
Mathematics achievement . . . . . . 5.96 1.21
Natural Science achievement . . . . 5.52 1.05
Social Science achievement . . . . 5.31 1.13
vocal-31.11am e o e o o e o e o e o o 5.21 1018
Reading comprehension . . . . . . . 9.90 1.19
Reading speed . . . . . . . . . . . 5.15 1.39
Grammar e o o o o e o o o o o o o o “.96 1033
Fitness test . . . . . . . . . . . 55.99 22.65
Commerce interest . . . . . . . . . 5.17 1.52
Engineering interest . . . . . . . 5.37 1.11
Mathematics interest . . . . . . . 9.99 1.20
Chemistry interest . . . . . . . . 5.19 1.28
Geology interest . . . . . . . . . 5.81 1.18
Biology interest . . . . . . . . . 5.38 1.96
Sociology interest . . . . . . . . 5.19 1.18
Education interest . . . . . . . . 5.27 1.16
History interest . . . . . . . . . 5.23 1.95
English interest . . . . . . . . . 5.33 1.57
Foreign language interest . . . . . 9.56 1.29
Fine Arts inteI‘eSt o e e o e e o o 5.15 1.35
High School Rank (Sten sc re) . . . 5.69 1.03
Grade Point Average (profession) . 3.11 1.16
Grade Point Average (H.S. index) . 2.50 .20
Grade Point Average (by college) . 3.39 .20
Grade Point Average (by previous

0011886) 0 o o o o e e o o o o e 206“ 038
High School Rank (percentile) . . . .59 .17

09416c

p...

bl".fll\h(‘l\~

2.
4:06 4'0
Hub-0&3
a 60
‘6 b
'605 a
4 b (.0
3000
inhuo ‘Ju
39.404 4
$45107 7p "
: a
l I,

O

l

as!

'4000000

0

ooooooo

'2666666

00000

7 0000
00

6656—

o
o o _
3600000
0000
000 o
o

‘oooo
ooooo
' 06666

0000
00000
00000
00000
0000

 

70

06766
2

U ~96

O 94

07002

077
4
09d56l

08355
0868
0901
08065
C7666
0
07409

0860
080641
94
0
08097
09031"
0 561
09612
61
0 9
O 95
075521
0 75
0 5
0881
O .
072
0762
8'
092 01
079461
08616
8800
07591
1
O -
08722

08010

06161

07559
8d

6

UDNéC~h.L}'b>-'C‘(JUW

\J
4
6

L
J

1

O N

UaCLDC‘CL

L>€

. - \ . . . .

O
O
O
O
O
O
O
O
O

O

COCOOC'OO

O

O

\

 

71

72

 

as.
ac.-
no.

8.
o~.o
mo.

Och-H. O‘cH
a...ao«
-8”

.

 

r

0A.- mac:
_ no-

8. flol

on.-.

ca.-.

ma.

2.- so:

no. ho.
no- 80

80
Horn 46.

.acH .ocH

cfi.l oo.~

an
.~.| pa.
H.‘ I- 3.

80! flNO

no.

 

go

n~ou no.
a“. ma.
ca. 00.

30
80

a..-
8.
OH.

3.

8.
en.
80

clue

I80

oo.a

acdom

I..4 Ct

 

73