AN EVALUATLON OF DIFFERENTEAL
PREDIICTLON OF ACADEMIC W FOR
STUDENTS AT WASHENGTON WATE-
UNN‘ERSITY

Thesis To? Hm Degree of Ed. D.
STTCHLGM‘VL STETE Uv‘ 1'

Gimme H. Bagley
19:66

 

fiié/ r;

Jrcr’:
,. , _-

   
 
  

“as," ‘HH‘L _ _.

LIBRARY Lg

chhz‘gan S rate
Umversity

LLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLLL

9051

 

This is to certify that the

thesis entitled

m Evaluation of Differential Prediction of
lea-:10: tic Success for Stueents at Jashington
St9 ue Universi ‘7;

presented by
Clarence iiimm Ba ‘18.)”

has been accepted towards fulfillment
of the requirements for

1561.1) degree inJQJIEL-llinij, Personnel
Luervice 5,; education-9.1 P53}.

fl/ML m

Major professfl

Date J une 2'? , 19536

0-169

 

 

 

 

ABSTRACT

AN EVALUATION OF DIFFERENTIAL PREDICTION
OF ACADEMIC SUCCESS FOR STUDENTS AT
WASHINGTON STATE UNIVERSITY

by Clarence H. Bagley

The purpose of the study was to evaluate for stu-
dents at Washington State University the grade predictions
from the Washington Pre-College Testing Program. The state-
wide testing program uses a multiple regression approach to
the prediction of college grades in academic areas from
high school grade averages and scores on aptitude and
achievement tests. The evaluation was concerned with
(l) the accuracy of the predictions in predicting achieved
grades at Washington State University, (2) the determination
of existing hierarchies for predicted and achieved grades
for the two normative groups at the University of Washington
and Washington State University, and (3) a comparison between
the predicted grades from the Washington Pre-College Testing
Program and similarly developed predicted grades from
normative data at Washington State University, when compared
with accuracy of predicting achieved grades at Washington
State Universityo

The data in the study were in the form of punched

cards or on computer magnetic tape. Existing computer

Clarence H. Bagley

programs frOm the testing program, specially developed tabu-
lating programs, and research programs were extensively used
in the study. Normative groups for deriving predictions and
for comparison purposes were the 1956-1957 freshmen at the
University of Washington and the 1958-1960 freshmen at
Washington State University. The cross-validity group for
comparison of the predictions from the two normative groups
was the 1961 freshmen at Washington State University.

The data for the first and second part of the study
were developed from regular statistical programs using
past operational data. The grade predictions for each
student were in 36 criterion or academic areas (chemistry,
speech, zoology, etc.), common to Washington State University
and previously designated areas of the Washington Pre-College
Testing Program. The predictions for both groups used the
Iterative Predictor Selection Program for the IBM 709
computer. The Program was developed at the University of
Washington and followed the Horst technique for multiple
differential prediction. The Program determined for each
criterion area a corrected multiple correlation and predictor
beta weights.

There were no statistically significant differences
among three levels of achieved grades when compared with
the predicted grades of the present Washington Pre-College
Testing Program. The present predictions can be used to
predict cumulative freshmen grades as well as cumulative
freshmen—sophomore grades or cumulative freshmen-senior

grades.

Clarence H. Bagley

A rank-correlation coefficient of .88 between the
predicted grades for the normative groups of the two insti-
tutions showed a definite hierarchy. A rank-correlation
coefficient of .57 between the achieved grades showed a
greater variation exists in achieved than in the predicted
grades. A comparison of the ranking of the predicted grades
with the ranking of the achieved grades for the 1961 fresh-
men at Washington State University showed a rank-correlation
coefficient of .57. The hierarchy among the predicted and
achieved grades reflects the concept of differential pre-
diction and demonstrates the differences in grading practices
of departments within the university.

The predictions derived from multiple iteration
procedures of the criteria of achieved grades of students
at Washington State University did not improve the corre-
lations to achieved grades over the predicted grades from
the Washington Pre—College Testing Program except for the
criterion areas of architecture, art, biology, engineering,
forestry, geology, journalism, pharmacy, and physics. The
prediction formulas now used in the state-wide testing pro-
gram need not be changed for students at Washington State

University except as noted.

AN EVALUATION OF DIFFERENTIAL PREDICTION
OF ACADEMIC SUCCESS FOR STUDENTS AT

WASHINGTON STATE UNIVERSITY

BY

f(

- or
Clarence HF Bagley

A THESIS

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

DOCTOR OF EDUCATION

Department of Counseling,Personnel Services,
and Educational Psychology

1966

 

© CLARENCE HIRUM BAGLEY 1967

All Rights Reserved

ple
spe
fo:

st'

Mr
pr
to
£0

My

ACKNOWLEDGMENTS

Many persons have made contributions toward the com-
pletion of this thesis. The author would acknowledge his
special appreciation to his major adviser, Dr. Buford Stefflre,
for his guidance and helpfulness in the preparation of this
study and his patience and encouragement throughout the
graduate program.

A word of thanks is also given to Mr. James Thummell,
Mr. Richard Baird, and Mr. Richard Wick for the computer
programming and operations at Washington State University:
to colleagues in the Washington Pre-College Testing Program
for the use of records and computer programs; to Mrs. Cathy
Myers for typing the rough drafts of the statistical tables:
and to many others who helped with the many small details
of the study.

Finally, the author must recognize his indebtedness
and gratitude to his wife and children who have had to
patiently postpone shared time and events while the thesis

was being completed.

ii

TABLE OF CONTENTS

Chapter Page
I O TIE PROBLEM O O O O O O O O O O O O O O O O O 1
Introduction 1

Need for the Study 6

Purpose of the Study 10

Limitations of the Study 11

Overview of Remainder of Thesis 13

II. REVIEW OF RELATED LITERATURE . . . . . . . . . 14

Multiple Prediction 16
Multiple Differential Prediction of

College Grades 17
Statistical Models for Multiple

Prediction 21
Multiple Differential Prediction at

the University of Washington 26
Washington Pre-College Testing Program 28
Summary 31

III. SOURCES OF DATA AND METHOD OF PROCEDURE . . . 33

Sources of Data 33
Criterion Areas 37
Procedure for Study 39

Computation of Intercorrelation and
validity Coefficient Matrices 43
Summary 46
IV. EVALUATION OF DATA . . . . . . . . . . . . . . 47
V. SUMMARY AND CONCLUSIONS . . . . . . . . . . . 135
Summary 135
Conclusions 139
Recommendations 141
BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . 143
APPENDICES . . . . . . . . . . . . . . . . . . . . . . 149

iii

Table

LIST OF TABLES

Page

Identification of the predictor variables

constituting the test scores and high

school grade averages in this study . . 35
Number of students constituting each of the

original criteria groups . . . . . . . . . . 38
Means and standard deviations of predictor

variables for 1958-1960 freshmen at

Washington State University and 1954-

1955 freshmen at University of

Washington . . . . . . . . . . . . . . . . . 49
A comparison of differences between W.P.C.T.P.

predicted grade-point averages and

achieved grade-point averages for the

1961 freshmen at Washington State

University . . . . . . . . . . . . . . . . . 52
Correlations, Z-transformations of correlations,

means, standard deviations between W.P.C.T.P.

predicted and W.S.U. achieved grade-point

averages for 1961 freshmen at Washington

State University . . . . . . . . . . . . . . 71

Comparison of zero-order correlation coefficients,
means, and standard deviations between
W.P.C.T.P. predicted and achieved grades
for the 1955-1956 freshmen at University of
Washington and 1961 freshmen at Washington
State University . . . . . . . . . . . . . .

Rank-order correlations of predicted grades for
the 1961 freshmen at Washington State
University and 1955-1956 freshmen at
University of Washington . . . . . . . . . .

Order of selection of predictor and cumulative
squared adjusted multiple correlations for
each of the differential predictor measures
as given in the "Iterative Predictor
Selection Program” for each of thirty-six
criteria of academic success at Washington
State University . . . . . . . . . . . . .

iv

79

84

88

Table

9. Multiple regression Beta weights for 1958-
1960 differential predictions of thirty-
six criteria of academic success at
Washington State University . . . . . . .

10. Multiple regression predictor weights for
1958-1960 differential predictions of
thirty-six criteria of academic success
at Washington State University . . . . . .

11. Comparison of simple correlation coefficients,
Z-transformation of correlations and E
values of significance for W.P.C.T.P.
predicted grades and Washington State
University predicted grades when cor—
related with achieved grades for 1961
freshmen at Washington State University .

Page
. 125
. 129
. 133

LIST OF FIGURES

Figure Page
1. Data sheet from Washington Pre-College
Testing Program . . . . . . . . . . . . . . 34
2. IBM cards from Washington Pre-College
Testing Program . . . . . . . . . . . . . . 41

vi

CHAPTER I

THE PROBLEM

Introduction

A continually increasing college-age population and
an ever-rising number of students are subjecting American
colleges and universities to unprecedented enrollment
pressures. Modern technology and a rapidly changing world
are demanding greater numbers of quality graduates in every
field. Mbre students and parents are seeing higher educa-
tion as necessary for achieving success in a complex society.
But with limited money and physical facilities, college
administrators are faced with the formidable task of
satisfying the quantitative and qualitative demands of
an education-minded populace.

From 1900 to 1963, college enrollments increased
from 250,000 to 4,100,000,1 and the total number of students
is expected to increase even more rapidly in the future.

The question is, then: "How can the present limited resources

and facilities of colleges and universities provide for

the forecast increase in students?" These limitations are

 

lStatistical Abstracts of the United States, 84th
annual edition (Washington: U.S. Government Printing Office,
1963), P. 114.

already causing colleges and universities to be more selec-
tive in their enrollment procedures. This, in turn, is
creating more pressure on the confused college students of
the future. Admission procedures vary greatly among
institutions, and it is imperative that the successful
student be given proper guidance before and after entering
college.

Differences in admission requirements and the great
variation in quality and content of the curriculum demon-
strate the uniqueness of each institution's educational
environment. These differences, coupled with the range of
student abilities and the diversity of high school academic
training, bewilder students. Consequently, institutions of
higher education must, collectively and individually, deter-
mine the best policy for admissions and, more important, must
strive to create the best academic climate for each student.
f‘ The selection of students is the result of many
variables, not all of which are easily measured. Motivation
and personality variables are not sufficiently understood,
and progress toward their understanding suffers from inade-
quate research design and lack of empirical data. 0ther
variables such as faculty size and classroom facilities may
be more easily measured, yet their impact on learning effective-
ness and subsequent grades is difficult to ascertain. The
methods of measuring the student's accomplishment in college

L are, therefore, often invalid and unreliable.

Faculty and student services personnel in institutions
of higher education must obtain the best possible estimate
of student potential and student limitations in order to
provide effective guidance and counseling programs for
today's students. Institutions must develop a more reliable
method for selecting applicants, a more efficient test for
evaluating the qualities of college environment, and a
greater proficiency in determining the educational skills
developed by each student. In 1954 Del- Wolfe made the
following statement, which is particularly applicable to
higher education today.

Improved programs of student guidance, founded

upon better manpower information and better methods
of appraising an individual's aptitudes will en-
able more young people to make choices which are
best for them, and for the nation; and thus con-
stitute an important element in a total effort to
secure better use of the nation's intellectual
resources.

The total guidance program, which includes not only
academic advising and professional counseling, but also
testing and prediction, is an important source of informa-
tion for determining student needs. An admissions officer
considers carefully the past academic records and test
information gathered on each student. However, admission
to an institution of higher learning should be considered

as only the first step in a student's educational experience.

During his university years the student is confronted with

 

lDeal Wolfe, America's Resources of Specialized
Talent (New York: Harper & Brother, 1954), p. 280.

 

a myriad of academic and social problems. If he is to profit
from his college experience, more guidance and testing are
often necessary. A student may switch majors several times
before he finds one suitable to his talents: a change in
universities is sometimes in order: or there is a re—defining
of educational goals. Some students will benefit from one
year of university work, although they may not complete the
undergraduate course work. In some cases, a student may be
directed primarily to graduation or to a post-baccalaureate
education. To meet these varied situations effectively,
counseling, testing, and prediction are important to the
student. The process of prediction becomes a day-to-day
reality of research and selection on the educational,
military, or industrial scene, where the emphasis is on
the best utilization of human resources.
r. The success of the student in a college cannot and
should not be predicted wholly by the process of aptitude
and achievement testing, nor can it be interpreted other
than in the context of the individual and his environment.
Hewever, test data play a significant part in any selection
\xor admissions program involving prediction.

There are widely divergent opinions concerning the
use of prediction and of psychological testing in the areas
of guidance and counseling. One historical VieWpoint was
a complete rejection of predictions and testing, as in
individual therapy as presented in Rogers' earlier books, in

contrast to the View of the dedicated trait—and—factor

person of the early 1940's.1 Advocates of the moderate
position, which combines the use of psychometric data with
other aspects of the counseling or advising program, state:
"Tests are valuable in the extent to which they improve the
accuracy of inescapable judgments."2 Thus, the controversy
as to whether or not psychological tests can be used for
prediction involves consideration of the tests and the pre—
diction criteria.

Inasmuch as it involves error of measurement and
probability, prediction in individual guidance practices
should be in accordance with statistical rationale.

Tyler summarizes the use of tests for prediction
work by stating:

It is because all existing aptitude tests make
these errors in prediction that reputable psycholo-
gists in counseling positions refuse to let final
decisions as to what individuals should do with their
lives rest on tests alone. Since limitations vary
from test to test, the task of drawing valid con-
clusions from a combination of several of them
presents complex problems.

Thus, modern statistical theory, data processing

machines and computers, and trained and informed guidance

and counseling staffs have contributed to the increased use

of prediction and testing in the academic setting.

 

lCarl Rogers, Client Centered Therapy (New York:
Houghton—Mifflin Company, 1951), p. 220.

2John G. Darley and Gordon V. Anderson, "The Functions
of Measurement in Counseling," E. R. Lindquest (ed.), Educa-
tional Measurement (Washington, D.C.: American Council on
Education, 1951), p. 76.

3Leona E. Tyler, The Psychology of Human Differences
(Boston: Appleton-Century Crofts, Incorporated, 1956), p. 133.

Need for the Study

The year 1960 marked a significant change in the use
of testing data for placement, admissions, and academic
prediction of grades at Washington State University. A
change in the evaluation for admission of high school
seniors and the formulation of a different and distinctive
freshman testing program created an awareness of the need
for more research in the area of prediction of academic
success. Increased concern by the university resulted in a
new admissions policy which used test data and high school
records and thereby made a more systematic program of testing
necessary.

With the initial use of the Washington Pre-College
Testing_Program as a required prerequisite for admission
to Washington State University in 1960, the former fresh-
man orientation testing program was eliminated in favor of
the more feasible pre-college testing program. The Washington
Pre-College Testing Program is a state-wide program supported
by and involving 25 of the 29 colleges and universities and
all the high schools in the state. The program uses as its
governing board the Council on High School—College Relations.
The primary function of the program is to provide test data
for counseling and guidance of the high school senior as
well as for the college freshmen. Evolving from the former
University of Washington Differential Guidance Program, it
uses multiple regression analysis for predicting college

grades in various academic areas.

The transition in 1960, from the former program
under the University of Washington to that of a state-wide
testing program, exposed for many a need to evaluate the
testing program. It was necessary to examine the accuracy
of the test data for prediction, placement, and counseling.
It was also necessary to determine the cross-validity of
the statistical procedures used in the new state-wide test-
ing program, as well as to compare the differences in the
normative groups.

The diversity of the educational programs offered
by institutions of higher learning in Washington and the
increasing quality and quantity of the student body challeng-
ed the effectiveness of using one set of testing norms for
all students and all colleges in the state. Multiple pre-
dictions of college grades from test scores and high school
grades needed more specific study.

Other state institutions have raised their high
school grade-point entrance requirements from 2.0 to 2.5.
This change was necessitated by the overcrowding of school
facilities and by the realization that students with less
than a 2.5 high school grade-point average have a limited
chance of succeeding at these institutions.1 State colleges
are studying the wisdom of current admission standards in

an effort to determine the probable effect of such changes.

 

1Ad Hoc Committee to Study the Freshman Year, "A
Progress Report,” University Admission Policy, University
of Washington, Seattle, February 25, 1960.

The high school senior, as well as the college
freshman, is interested in the collegiate freshman year
and its subsequent academic demands. For many students
the first year of college is, academically, a "make or
break"situation. For other students it represents the only
college training that they will receive. Is one year of
college sufficient for some, or must all students be
encouraged to graduate? Can a student lead a successful
life in our present society without a college diploma?

The increasing emphasis upon the problem of ascertaining
the probable success of a student in college is, therefore,
desirable in a quality testing program.

At Washington State University about 30 percent of
the entering freshmen leave school before the beginning of
their second year, and many more approach their sophomore
year with low grades. Those interested in a more effective
counseling procedure and appropriate admissions criteria
are concerned with grade prediction in different academic
areas, as well as a general level of achievement by students.

The faculty and staff of Washington State University
believes that a student should be admitted to the university
if he has a reasonable chance of success. The term
"reasonable chance of success" is interpreted to mean a
completion of the freshman year with a grade—point average
of 2.0 or higher, with 4.0 equaling A. The use of prediction
data should apply especially to the freshman year, since

this is the crucial time for determining a student's academic

ability. The freshman year is critical to the student who
is faced with the task of determining the specific major
most suited to him, and of selecting the necessary classes
and instructors. The former testing program at Washington
State University did not include differential prediction

of academic success. The prediction data of the present
Washington Pre-College Testing Program, using as a criterion
the accumulative four years of collegiate grades in the
academic areas, may not accurately predict freshman grades
at Washington State University. Therefore, a careful
appraisal of a program for the prediction of academic grades
would be useful in the evaluation of the student potential
necessary to achieve during the freshman year as well as

the remaining academic career. High school students, their
parents, and counselors would benefit from such a study.
Predictions that could be applied to the freshman year would
serve as an intermediate step toward the final decision of

a major during and after the freshman year.

The differences found in the test scores of the
present group of students, in relationship to their academic
work at Washington State University, and the 1955 freshman
group at the University of Washington who are used in the
normative data for the Washington Pre-College Testing Program,
may be significant or negligible for the purpose of this
study. No valid evidence has previously been shown to prove

any hypothesis of differences or similarities.

10

Purpose of the Study

It has been previously stated that the Washington
Pre-College Testing Program was developed through research
conducted at the University of Washington. Initially, all
the necessary data for the program were based on University
of Washington students. The actual collegiate grades
earned by students at the University of Washington were
used with multiple regression analysis to derive the pre—
dicted grades in the Washington Pre-College Testing Program.

The question pertinent to this study is whether a
prediction from the Washington Pre-College Testing Program
for a certain grade-point average in a definite major with-
in the college curriculum relates specifically to the student
taking course work at Washington State University. Are the
multiple regression formulas, beta weights, and multiple
R's now used in the computation of predictions with the
state-wide testing program also applicable to another univer—
sity's population?

Specifically:

1. As measured by high school grades and testing data
on the Washington'Pre-College Testing Program, are there
wide differences in students' aptitudes and achievements
between students at Washington State University and students
in the University of Washington's normative group?

2. How accurate are the current predicted grades of the
Washington Pre—College Testing Program for students at

Washington State University at various levels of progress,

ll

i.e., freshman, sophomore, or senior? Can the grade pre-
dictions be used to predict grades equally well at all three
levels?

3. Is there a hierarchy of predicted and achieved
grades from the Washington Pre-College Testing Program and
what similarity does this hierarchy have to the predicted
and achieved grades for students at Washington State Univer-
sity?

4. Are the present predictor equations used in the
Washington Pre-College Testing Program valid for Washington
State University students, or should predictor equations
be based on the grades achieved by students at Washington

State University?

Limitations of the Study

This study does not directly reflect the part that
motivation and interest play in academic achievement. Recog-
nizing the effect of these two variables and the difficulty
of their measurement, the study is restricted to an examin-
ation of those predictor variables now used in the Washington
Pre-College Testing Program. The variables of motivation
and interest are reflected in high school achievement and
thus are indirectly considered in the prediction formulas.
Certain areas, such as music and art, may require skills or
abilities not measured by the Washington Pre-College Testing
Program, but again the study is restricted to the use of

presented predictor variables.

12

A second limitation of the study is the sample sizes
for the predictor correlations and the criterion correlations.
The elective nature of the curriculum results in a disparity
in the number of type of students in the various criterion
areas. Ideally, the criterion groups should be students
with identical programs selected as random samples of a
total group. However, the selection of course work by the
students in the samples is not random, and thus small samples
are produced where an unknown and, therefore, immeasurable
amount of sampling error may arise. The resulting sample
should, therefore, have separate symmetric matrices for
each regression formula, since there is a difference in the
size of the criterion sample.

The study assumes that the criterion groups were
representative of the total group, as does the underlying
rationale for the Washington Pre—College Testing Program.

The names of the criterion areas, reflecting the
appropriate academic titles, such as history, chemistry,
etc., do not convey all the essential differences in course
content, grading patterns as affected by the instructor, or
possible differences in required attributes necessary to
aittain a particular grade. The variations within a single
criterion area, related to the content, grades, and attitude
of the students, are largely inferred.

The criterion areas are named to correspond with
those names given in catalogs which are issued by the two

universities involved in the study.

13

Overview of Remainder of Thesis

In Chapter I the need for the study was presented
and the specific questions regarding the validity of the
predicted grades for the Washington Pre-College Testing
Program were stated. In Chapter II the review of related
literature is presented. The sources of data, procedure
for the computer processing, and outlined procedure for the
study are given in Chapter III. The evaluation and inter-
pretation of data regarding the four questions given in
Chapter I are presented in Chapter IV. The summary and

conclusions for the study are presented in Chapter V.

CHAPTER II

REVIEW OF RELATED LITERATURE

The continually increasing college-age population,
and consequently higher enrollment figures for institutions
of higher education, has emphasized the need for a more
selective admissions policy and for accurate prediction of
academic success for each individual. The increased use of
psychological testing and other prediction data has resulted
in numerous studies of the validity of testing instruments
and methodology and the subsequent accuracy of the resulting
predictions. The desire for better and more diversified
testing programs with a reduction of duplication has prompted
the public, as well as the educator, to demand a more syste-
matic program of testing.

The use of tests in the selection of applicants for
admission and the prediction of academic success, defined
in terms of college grades, has been the most explored topic
in educational-psychological research. Segal had summarized
the findings of 23 studies before 1933.1 Garrett, in his

1949 review, covering nearly two decades, mentioned

 

lDavid Segal, Prediction of Success in College.
United States Department of Interior, Office of Education,
Bulletin No. 18, Washington. Government Printing Office,
1934.

14

15

approximately 194 studies.l Fishman reported 580 studies
in the years 1950-1958.2 Travers, who cited more than 200
prediction studies in his review, concluded that high school
grades are the best single predictor of college success.3
Data from a summary by Fishman demonstrated that the classifi-
cation of the various studies was limited primarily to a
global prediction of either a semester grade-point average
or the freshman-year grade-point average.4

These summaries of studies illustrate (1) that the
progress toward improved prediction has been slight despite
the many studies which have been made, (2) that most of the
present studies follow the pattern of past studies, that is,
a global prediction of grades from intellectual factors,
and (3) that the development of better predictors and criteria
must be concerned with measuring different factors with

Lfi‘different data.

 

lHarley F. Garrett, "A Review and Interpretation
of Investigations of Factors Related to Scholastic Success
in Colleges of Arts and Sciences and Teachers Colleges,"
Journal of Experimental Education, 18:91-131, December,
1949.

2Joshua A. Fishman, "unsolved Criterion Problems
in the Selection of College Students," Harvard Educational
Review, 28:320-29, Fall, 1958.

 

3Robert M. W. Travers, "The Prediction of Achievement,’
School and Society, 70:293, November, 1959.

 

4Joshua A. Fishman and Ann K. Pasanella, "College
Admission-Selection Studies," Review of Educational Research,
30:298-310, 1960.

 

16

Multiple Prediction

Although earlier studies attempted to predict
academic success through the use of a single variable (usual—
ly standardized tests) a review of studies related to the
prediction of academic achievement in college indicates that
there is a distinct superiority in multi-variable prediction
in comparison with prediction by the use of a single factor.
More recently the emphasis in educational guidance and pre—
diction has been on the use of a combination of variables
in a carefully integrated battery.l

Bruce summarizes the past research by stating:

Since the early twenties well over 1,000
studies have been made in an attempt to better
understand and cope with the problems of univer-
sity admissions and failures. About 90 percent
of these studies used one variable and calculated
zero order coefficients or correlations to deter-
mine evidence of predictive value of these variables.
Approximately 5 percent of the studies combined two
variables and computed multiple coefficients of
correlation. Some increase in the multiple co-
efficient or correlation was achieved by investi—
gators using three-variables combinations but
only some twenty of such studies have been com-
pleted. About eight studies attempted four or more
variables with limited success, but rarely does
anyone attempt as many as eight independent
variables.2

 

1Benjamin S. Bloom and Frank R. Peters, The Use
(of Academic Prediction Scales (New York: The Free Press of
Glencoe, Inc., 1961), p. 37.

2William J. Bruce, "The Contribution of Eleven
variables to the Prognosis of Academic Success in Eight
Areas at the University of Washington" (unpublished
Doctor's dissertation, University of Washington, Seattle,
1953), p. 12.

‘ 17

In 1943 Cosand summarized studies of multiple pre-
dictors which showed a range of .53 to .83 with a median
of .63. He believed that the multiple predictors with these
correlations pointed out the advantage of several rather
than a single predictor.l

Durflinger reported in 1953 that the multiple
correlations, found in summarizing studies, were between
.60 and .70 and concluded that the use of several predictors
results in the highest multiple R's.2

Harris reported the results of combining variables
in a review of significant studies which showed the multiple
approach superior to a single predictor.3

Segal summarized: ”It will be noted that coefficients
using a combination of items are higher than those given

for single predictive items as given previously."4

Multiple Differential Prediction
of College Grades

Predictors in multiple correlation studies of college

grades were primarily intellective and used aptitude or

 

1Joseph P. Cosand, "Admissions Criteria: A Report
to the California Committee for the Study of Education,"
(Zollege and University, 28:338-364. April, 1953.

2G. W. Durflinger, "The Prediction of College Success:
A Summary of Recent Finding," The American Association of
College Registrars Journal, XIV (October, 1943), 68—78.

3Daniel Harris, "Factors Affecting College Grades:
A Review of the Literature, 1930-37," Psychological Bulletin,
37:125-26, March, 1940.

 

4Segal, op. cit., p. 127.

18

achievement tests in combination with high school grades or
rank in class. Non—intellectual predictors have been used
recently, but these have not produced significantly higher
multiple correlations.1 The highest multiple correlations
were obtained in the southwest and western colleges, where
selective procedures were either so new, or so restricted
by statute, that applicant talent was not restricted in
ranges.2 In contrast, the selection procedures studies in
other areas resulted in lower multiple correlations. In
certain testing programs, where the range of test scores
and high school grades was restricted due to a rigid selec-
tion system, the size of the multiple correlation coefficient
was considerably reduced.3

The acceptance of multiple correlation techniques
has led to the problem of criteria used in prediction and
selection studies. With the advent of high—speed data pro-
cessing equipment and computer technology, the researcher
can obtain more comparative information concerning grade—
point averages. thwithstanding the criticism of the grade-

point as a criterion of academic success, it still serves

 

lJoshua A. Fishman, op. cit., p. 346.

2John Spencer Carlson and Victor Milstein, "The
Relation of Certain Aspects of High School Performance to
Academic Success in College," College and University,
33:185-92, Winter, 1958.

3John L. Holland, "The Prediction of Scholastic
Success for a High Aptitude Sample," School and Society,
86:290-293, June, 1958.

 

 

19

as the best evidence available.1 The use of a grade-
point average as a criterion for prediction reduces the
problem to a question of what criteria to study and what
prediction model to use.

When predicting college success as represented by
grades in various subjects, one should make a specific
prediction of success in the academic areas in accordance
with accepted curriculum separations and empirical research.
Differences in curriculums and institutions point to the
importance of differential prediction of success within
each university as well as among colleges and universities.
Differential prediction of success in different academic
areas, based on experimental work, and given to the student
in data form, is a valid procedure. Certainly, each specific
or general ability of each student should be studied care—
fully in order to provide the maximum information concerning
each individual's potential in a wide range of subjects.

The reason for using differential prediction is to
maximize the potential of the individual in the choice of
curriculum as well as to facilitate selection of the best
possible candidates. Also, the multi-dimensional character
of students, colleges, and curriculums requires a more care-

ful and systematic approach to selection than is possible

 

lPaul Heist and Harold Webster, "Differential
Characteristics of Student Bodies--Implications for Selection
and Study of Undergraduates in Conference on Selection and
Educational Differentiation," Selection and Educational
Differentiation, p. 91—106, Berkeley, California, The Center
for the Study of Higher Education, 1960.

 

20

with a single criterion. The prediction of success in
college by using the criteria of college grades is one
step to a more helpful program. The prediction of college
grades will be dependent upon the individual variation that
exists in the student and the college.

The question then becomes whether criterion should
be based on over-all success for each student or whether
it should be based on success in each individual course
area. One practical difference between the global approach
and a more sophisticated differential prediction is the
economy involved in the necessary time and money needed to
summarize grades within the academic setting.

Over 95 percent of the studies located by Fishman l/,

were of the global type in methods and goals.1 The criteria

~5- ’——_—.——-

 

for these studies were measurements represented by a total
grageflprgduct, primarily thewfreshman-year average or the
first-semester grade average. The separation of grades
by academic-year or subject—matter area was not attempted
and the problem of grade prediction or expectancy was not
pursued. Many of the prediction studies have been made
using essentially the same methods, thus resulting in a
standardization of criteria.)

Some current research is attempting to expand the
global prediction to more definitive subject-matter areas.
Crawford and Burnham reported differences between correlations

of test scores and freshman marks in various subjects.

 

lFishman, op. cit., p. 341.

21

Correlations range from + .57 to - .01 for two freshman
populations. The SAT verbal score correlates with the
average grades in English and history + .49 and with
physics grades + .40. When achievement tests were used
instead of grades, the correlations with aptitude tests
tended to be higher, but the pattern of differences re-
mained the same.1 Stone reviewed the predictions of college
grades in broadly defined curriculum areas by using test
scores and high school grades. The underlying rationale
of this study, then,Was to determine if the criterion

of college grades would lend itself to differential pre-

. . . . 2
diction along curriculum lines.

Statistical Models for Multiple Prediction

The type of differential prediction used in this
study is multiple regression or correlation analysis. In
contrast, there are other models of differential prediction.
Multiple discriminant analysis has been extensively investi-

gated at Harvard by Rulon3 and Tiedeman.4 This approach

 

1A. B. Crawford and P. S. Burnham, Forecasting
College Achievement (New Haven: Yale University Press, 1946),
291 pp.

 

 

2J. B. Stone, "Differential Prediction of Academic
Success at Brigham Young University," The Journal of Applied
Psychology, 38:109-110, April, 1954.

 

3P. J. Rulon, "Distinction between Discriminant and
Regression Analyses and a Geometric Interpretation of the
Discriminant Function," Harvard Educational Review, 21:80—90,
Spring, 1951.

 

4David V. Tiedeman, "The Multiple Discriminant
Function—-A Symposium," Harvard Educational Review, 21:
167-186, Spring, 1951.

 

22

involves the study of differences in groups defined a ppiori
for those Variables held in common. The discriminate function
in this approach maximizes the ratio of the variance among
groups in relation to the variance within the groups.
These workers feel that the discriminant function is a pro-
mising method for comparing an individual's scores with
those of the groups that he is considering joining. Tyler
favors the type of measurement which seeks to characterize
the individual's customary pattern of choice rather than
the test score.1 French disagrees. He notes that the dis-
criminant method tells only which group one is similar to,
when most persons want to know the degree of success or
satisfaction that can be expressed within the group.2
Thus, while helpful, the discriminant function does not make
as thorough a differential prediction as the multiple re-
gression approach to test score analysis.

Pattern analysis and joint regression with a dis—
criminant function, as seen in the work of Fricke and
Tatsuoka, are variations attempting to bridge the gap be-

I

tween discriminant function and multiple regression analysis.

 

1L. E. Tyler, "Toward A Workable Psychology of Indi-
viduality," American Psychologist, 14:75-81, 1959.

2J. W. French, The ngic of and Assumptions Under—
lying Differential Testing. Proceedings 1955 Invitational
Conference on Testing Problems. Princeton, New Jersey,
Educational Testing Service, pp° 40-48.

 

3Benno G. Fricke, "A Coded Profile Method for Pre—
dicting Achievement," Educational and Psychological Measure-
ment, 17:98-104, Spring, 1957.

4Maurice M. TatsuOka, "Joining Probability of Member-
ship and Success in a Group: An Index Which Combines the
Information From Discriminant and Regression Analysis as

23

Cronbach and Gleser presented a review of more simplified
techniques in combining pattern analysis and profile analysis.1
waever, these techniques have not found wide_acceptance

in testing or research programs.

The use of the multiple regression model for dif-
ferential prediction makes it necessary to clarify the terms
comparative prediction and differential ppegigtion as they
are used in the literature, and in the terms multiple
absolute prediction and multiple differential prediction.
Differential prediction attempts to foresee a difference
between the success one individual will have on two criteria,
while comparative prediction endeavors to establish the
absolute levels of each single criterion. Tucker suggested
the term comparative prediction while Mollenkopf introduces
the problem of differential prediction.2 The computations
for these two techniques are similar: for a given battery of
tests there is no essential difference between the two since
.the predicted differences are equal to the difference in the
predictions. However, the two techniques differ in their
selection of predictor variables, since a test yielding

high correlations with each of two criteria contributed

 

Applied to the Guidance Problem," Harvard Studies in Career
Development, No. 6, Harvard Graduate School of Education,
October, 1957. (Mimeographed.)

1L. J. Cronbach and G. C. Gleser, "Assessing Similar—
ity Between Profiles," Psychological Bulletin, 50:456-473.
1953.

 

2W. G. Mollenkopf, "Some Aspects of the Problem

of Differential Prediction," Educational and Psychological
Measurement, 12:39—44, 1952.

 

24

little to the prediction of differences between the criteria.1
Still, neither differential nor comparative prediction will
provide effective discrimination if the criterion areas are
highly correlated. This high correlation between criterion
areas makes it exceedingly difficult to differentiate between
certain academic subject-matter areas within the university's
curriculum. Wesman and Bennett write:

The tests which survive attempts to predict
criterion differences directly are naturally enough
those which correlate with those differences . . . .
A scholastic aptitude test may be one of our best
predictors of success in courses in a liberal arts
college: but because that aptitude is very important
to success in all courses taken by the freshmen, it
will receive little or no weight in the prediction
of differences in course grades. Success in each
course may depend to a large extent on that aptitude
measured by the test, while predictable differences
in success may be the product of other characteristics
or traits. Tests of these other characteristics or
traits will receive greatest weight in the direct
prediction of differences.

Michael states that for differential prediction:

The problem posed was to select a specified num-
ber of predictors from several available ones that
would yield simultaneously the most nearly accurate
prediction of differences between scores on all pos-
sible pairs of criterion variables within a given
set. For multiple absolute prediction, an attempt
was made to select a given number of predictors
such that the degree of accuracy with which all of
the criterion variables are predicted will be at a

 

1Paul R. Dressel, A Report on Differential Prediction
and Placement in Colleges and Universities (New York: College
Entrance Examination Board, 1959), 17 pp. (Mimeographed.)

2A. G. Wesman and G. K. Bennett, "Problems of Dif-
ferential Prediction," Educational and Psychological Measure-
ment, 11:265-272, 1951.

 

 

25

maximum irrespective of the extent to which the
chosen battery differentiates among the various
criterion measures.

Herst has discussed the effectiveness of multiple
absolute prediction, which has a goal of yielding the
highest possible correlations with several criteria.2’3
In contrast, multiple differential prediction has the goal
of yielding the greatest possible differentiation between
criteria.4 The operational implementation of both methods
was reported. Zeigler, Bernreuter, and Ford have somewhat
similar goals but apply individual procedures to obtain their
results.5 The results of these four methods of analyses
can justify reducing the terms comparative prediction,
multiple absolute prediction, multiple differential pre-
diction and differential prediction to two terms, multiple

differential and multiple absolute. Multiple prediction

is, then, a choice between two distinct approaches,

 

1W. B. Michael, "Development of Statistical Methods
Especially Useful in Test Construction and Evaluation,"
Review of Educational Research, 29:89-109, 1959.

2Paul Horst, "Differential Prediction in College
Admissions,” College Board Review, 33:19-23, Fall, 1957°

3Paul Horst, "A Technique for the Development of a
Multiple Absolute Prediction Battery," Psyphological Monographs,
Vol. 69, No. 5, Whole No. 390, 1955.

4Paul Horst, "A Technique for the Development of a
Differential Prediction," Psychological Monographs: General
and Applied, Vol. 68, No. 9, Whole No. 380, Washington,
D.C.: American Psychological Association, 1954, p. 31.

 

 

 

5Martin L. Zeigler, Robert G. Bernreuter, and
Donald H. Ford, "A New Profile for Interpreting Academic
Abilities,“ Educational and Psychological Measurement,
18:583-88, Autumn, 1958.

26

differential and absolute. The literature refers to over-
lapping terms given to both methods and does not clearly
differentiate between the approaches. However, for this
study, multiple prediction refers to differential pre-
diction between criteria as was initially theorized by
Hbrst but not realized operationally in the prediction
formula found in the Washington Pre—College Testing Program.

Multiple Differential Prediction at the
University of Washington

In 1930 Brammell advocated an approach to multiple
differential prediction by recommending a combination of
criteria in predicting student success. This initial study,
which was written under the direction of August Dvorak,
was the first of many studies to investigate multiple
variables.1 Blair and Salyer broadened the consideration

2'3 The

of entrance requirements and criteria of success.
Angell, Langton, Meyer, and Pettit investigations in 1950

initiated a serious approach to multiple prediction of

 

1P. R. Brammell, "A Study of Entrance Requirements

at the University of Washington" (unpublished Doctor's
Dissertation, University of Washington, Seattle, 1930).

2Glenn M. Blair, "The Prediction of Freshmen Success
in the University of Washington” (unpublished Master's
thesis, University of Washington, Seattle, 1931).

3Rufus C. Salyer, "An Investigation in the Pre—
diction of Success in the School of Engineering at the
University of Washington" (unpublished Master's thesis,
University of Washington, Seattle, 1931).

27

criterion areas using multi—variable factors.1 These in-
vestigators proposed to evaluate individual success in 26
university subject-matter areas by using multiple regres—
sion equations, following a simplification of the Yule
method of partial correlation as developed by Dvorak.2
Using this Yule method of correlation and the then new IBM
650 computer, which increased the efficiency factor even
more, these investigators reduced the formidable task of
multiple prediction calculations. Hurst developed his
technique to be used with the computer so that the formidable
problem of calculations could be done in a few minutes.3
The progress made in the reduction of calculation time and
the expansion of work toward a greater range in criterion
areas was complementary to the development of theoretical
models, computational methods, and statistical procedures,
as well as the use of normative groups for valid studies.
The first statistical rationale reported by Herst

in 1950 utilized a technique suitable for the prediction of

a single criteria.4 Two alternate methods were then proposed

 

1M. A. Angell, R. c. Langton, G. A. Meyer and M. A.
Pettit, "An Evaluation of General and Specific Admission
Requirements at the University of Washington" (unpublished
Doctpr's Dissertation, University of Washington, Seattle,
1950 .

2G. Udny Yule, "On the Theory of Correlation,"
Journal of the Royal Statistical Socieyy, London, 60:835-838,
December, 1897.

3Paul Herst, "A Technique for the Development of a
Differential Prediction Battery," Psychological Monographs,
loc. cit.

 

4Paul Horst and Stevenson Smith, "The Discrimination
of Two Racial Samples," Psychometrika, 15:271—289, September,
1950.

 

28

to permit the selection of a single battery of predictors
from a much larger initial battery, meanwhile maximizing
the effectiveness of specific predictions within each
criteria.1 The first rationale was multiple absolute pre-
diction, designed with a sub—set of the original predictors
which.wou1d yield the "best” prediction of each criterion.2
The second rationale, multiple differential prediction, was
used to select a sub-set of the original prediction battery
which would yield optimum predictions of differences in
achievement for all possible pairs of criteria. The pur-
pose of multiple differential prediction was to predict

the area or areas in which the student would be most success-
ful without regard to his potential level of achievement.3
The basic research in the development of these two models

has been conducted at the University of Washington.

Washington Pre-College Testing Program

The original prediction battery selected for the
Washington Grade Prediction Program and, subsequently, the
Washington Pre—College Testing Program, was based on the

academic performance of the 1949 freshmen who entered the

 

1William M. Meredith, "Cumulative Calculations of
Regression Constants," Multiple Prediction Studies, Office
of Naval Research Contract Nonr-477 (08), Paul Horst,
principle investigator (unpublished report, The University
of Washington, Seattle, June, 1956), p. 1-2. (Mimeographed.)

2Paul Horst, "A Technique for the Development of a
Multiple Absolute Prediction Battery," Psychological Mono-
graphs, Vol. 69, No. 5, Whole No. 390, 1955.

Horst, op. cit. passim.

 

29

University of Washington. This battery was "selected"
from a larger battery by the Horst's Differential Pre-
dictor Selection Technique.1 The sample criteria were grade-
point averages for the 1949 freshmen for the eleven quarters
of study ending with the winter quarter of 1953. Grades
were analyzed in 32 separate subject areas—-anthropology,
architecture, art, biology, botany, business administration,
chemistry, drama, economics, education, engineering,
English, Far Eastern studies, foreign languages, forestry,
geography, geology, history, home economics, journalism,
mathematics, music, nursing, pharmacy, philosophy, physical'
education, physics, political science, psychology, sociology,
speech, zoology--and in terms of an all-University average.
Fifteen of the original sixteen selected predictor
variables used in the differential model in 1953 were also
used in the present (1957) battery, with only the Guilford-
Zimmerman test of numerical operations eliminated. Mills
reviewed the history and development of the differential
prediction batteries used at the University of Washington.2
Recent studies have continued to investigate multiple

variable differential prediction. Long,

 

lHorst, "A Technique for the Development of a
Differential Prediction Battery," Psychological Monograms,
loc. cit.

 

2D. F. Mills, "An Interative Selection of Variables
for Predicting Certain Criteria of Academic Success at the
University of Washington" (unpublished Doctor's Dissertation,
University of Washington, Seattle, 1957).

3James R. Long, "Academic Forecasting in the Technical—
Vocational High School Subjects at West Seattle High School,"
(unpublished Doctor's Dissertation, University of Washington,
Seattle, 1957).

30

Franks,l Mattrick,2 and Horst3 suggest that differential
predictions of levels of future scholastic achievement can
be developed. The state-wide adoption of the Washington
Pre-College Testing Program in 1960 was an application of
these studies. Two studies by Reas and Lounsberry, have
since initiated a comparison of predictions of grade achieve—
ment between two institutions.

Reas concluded:

1. On the basis of predictor data for entering
freshmen at the University of Washington and enter-
ing freshmen at Seattle-University, the two groups
were markedly similar in mean high school grades
and mean test scores. On the basis of mean pre-
dicted grades in various college subjects, the
two groups exhibited small differences.

2. The correlations between predicted grades
and achieved grades of students in corresponding
departments in the two universities were very
similar. . . . When the difference between stu-
dents' predicted and achieved grade-point averages
in corresponding areas at the two universities
were compiled for each tenth of a grade-point
difference and plotted, the compilations were so
similar and the plotted curves so overlapped each
other that without appropriate labels, identifica-
tion would have been impossible.

 

lDean K. Franks, ”A Study of the Success of West
Seattle High School Students in Languate Arts, Foreign
Language, Social Studies, and Music and Art" (unpublished
Doctor's Dissertation, University of Washington, Seattle,
1958).

2William E. Mattrick, "A Study of the Contributions
of Twelve Variables to Prediction of Academic Success in
Five Ninth Grade Subjects" (unpublished Doctor's Disserta-
tion, University of Washington, Seattle, 1958).

3Horst, "Relationship between Preadmission Variables
and Success in College," Office of Naval Research Contract
Nonr—477 (08), and Public Health Research Grant M—743 (c3),
Paul Horst, Principle investigator (unpublished report, The
University of Washington, Seattle, June, 1959). (Mimeo-
graphed.)

31

3. When the multiple correlation coefficients,
predictor beta weights, and regression equations for
Seattle University student data were compared with
those developed on the University of Washington
student data and currently used in the state-wide
program, it was found that the similarities out—
weighed the differences. . . .1

Lounesberry found that a close relationship existed

between the educational and grading standards at Western
Washington State College and the University of Washington,
and that the predictions from both schools were similarly
accurate. He concluded that a group of predictors developed

from data at one institution can be used with comparable

. . . 2
accuracy at a second institution.

Summary

A review of studies related to prediction of academic
achievement in college indicated that there is a distinct
superiority in the use of multi-variable predictors com—
pared with the use of a single factor. The value of some
r's for prediction was between .35 and .83, with a mean
correlation value of r = .50.

Predictions were made using the accumulated grade-

point average and the freshman year grade-point average as

 

lHerbert D. Reas, "A Follow-Up Study of the Washington
Pre-College Differential Guidance Test at Seattle University"
(unpublished Doctor's Dissertation, University of Washington,
Seattle, 1962).

2James Rodney Lounesberry, "An Evaluation of the
Accuracy of the Differential Prediction Test Battery in
Predicting Grades for Students at Western Washington State
College” (unpublished Doctor's Dissertation, University of
WaShington, Seattle, 1962).

32

the criteria. There were few studies related to grade
predictions in Specific academic areas. Until the develop-
ment of computer techniques for minimizing calculations and
processing large amounts of data, the clerical work proved
to be a formidable task.

A review was made of the development of statistical
theories for use in multiple prediction and multiple dif-
ferential prediction at the University of Washington. The
application of the differential prediction model was made
to the statistical theory underlying the development of the

Washington Pre-College Testing Program.

CHAPTER III
SOURCES OF DATA AND METHOD OF PROCEDURE

Sources of Data

Students used in this study were the 1958-1961
freshman class members at Washington State University for
whom a complete and valid record of entrance battery test
scores of the Washington Pre-College Testing Program was
available. These students had been awarded at least one
or more grades in regular classes at Washington State
University. All students were regularly enrolled and were
resident students of the state. All were high school stu-
dents immediately before entering college.

A sample copy of the data sheet from the test battery,
which was given to the students, is presented in Figure 1.
The data consist of raw test scores and high school grades
as well as predicted college grade-point averages. The
identification of the variables used in the state-wide test
battery is presented in Table 1. Fifteen of these variables
were used for prediction in the Washington Pre—College
Testing Program.

The criterion variables include grade—point averages
earned by the student in each subject-matter area during

his undergraduate tenure at Washington State University.

33

34

 

     

   

 

   

 

 

 

 

 

l—AdOD 3931103
'lVflNVW NI OEION 'IVIDleO 3931103 OJ. 831N130 HO ‘IIVW

' L
l .1 1901114001! 0 . ‘ N BDNZDS . a I" L
L E z ‘IVDIIVMDIH O ”mun" ! N m 1mm V N
a“ F L
n Drums ,.
L -< omnosvn H d 3'5"" “ ”fl uunm T ' F
, Z w name in ,_ g E
‘ «n O 7‘
<1 5 1 N JDNZDS u
L . . . I
Hill
.. 1 m
L '3 0 g m nsnvuunor ~ g "om” 31 9 M L
2 .1
.J _ '5 an
I LL, 2 E H SDIIONODEQ O m unvm )nun v 0‘ l
L -- nuns to 0 no u. N m " 1 N <
u, u woos N . .—
0 i _ . o ., o <
z AHOISIH _ m mango-nae . m
L é a r: o L
.0 3 0 inc? F V‘ Aso‘lon n 3 «033 new N M Lu
L u gmaygu H "‘ N NOIllalnN V m g l
I
L 0 Hanson "‘ o Aomonlon-u : o <
0 ands 0* — M m o
n omavad N w _
L L— "'" AIISIUOI * 3 I: ”mot.” 3 Z 8 L
L " |n Q A901OIOIAJN I
0 unit :_' <1
w usual H c: szsvnsuvi -.n szevnaun o 0\ -‘ L
L N 3 q ,_ 0 ' n alumna n m S
H . 1 .—
F V 3 I: a) 2
L 2 sumaas ,0 3 m lsv: an _ .fi : n :1) E
o |
W
mi w «Museum; ,1 ~’\| Jumvann ,. O E
i: 'x -‘ WW! - vfi nsneu: M m u.
'4 a: O yawn V G" A a
I "II '— O ail-Inna!!! I N N saovnsnn ‘ 4'
o— u u - I‘ wolssna "' m Ill L
L D B O H Ln 3
on — m
w nun-nu m m uouvana: ._ :fi Anouousv n m z j
u
L °‘ _ 3 = 0
CL 0 N u I
9‘ l‘ 3 samouoa: 2 (Q ouuunoaov I, m z 3 U
L sumosvza F o 3 : “I.
O m \0 n O 3.. ° E L
L H g a: vuvlu a ,n mongol ,., m 3;;
‘° ’ *3 ‘L o a :3: 2 L
‘2
L O :3 AllSIIJND a f“ M33345 2'. m .3“ " o
< _, salmon: ¢ 1‘: a 'c‘.‘ S .— L
. H < <
L >3 ., v..." . fi _ m :550
D r u ssamsnl 0“ ”mom“ '9 m o o n z
0 z u In!
L O. n g nuns O 55 E
o O H "’ ' _
5 “mm": ¢ 5 Q unno- 0 fl AoowouaAsa 9, m S E E 3 L
J
I J "5 _; Q 3 a. L
O U I) 0 g
I O JDN3DS O 3, ”01°.- “ m Drums 0. m a! 3
L 5 “macs d’ ~q 1v3unoa N f“. 5 g g
( ‘
O h u w
I Luv c 0‘ smsma " (h T I: l
O V
L L’ aovneun O .., N N N V N "
I NDIZIOJ 4‘ N 3 m o g 2
N ‘6' 2 ‘° um . m : ~ °° . 3:
m ‘ onluauv N I "‘ CD — ~ ~
«I g 1 co m m m
H In Hun o o o
L é I > m 1190104 N m maoso‘mu “’ H :i:
0 g ‘ own" 3‘ "' m _ - -
F. 1 ‘ r~ vs a
l ‘0 nsnsu; 0 =2 ”"3" ~ wvnuma '“ h T T a L
O m §g .n L, 39:30: 1'1 an N :3
L . . L
. _L

Figure 1. Data sheet from Washington Pre-College Testing
Program.

Table l.

35

Identification of the predictor variables con-
stituting the test scores and high school grade
averages employed in this study.

 

 

Variable

 

High school English grade-point average: the
average of grades earned in courses such as
English, journalism. and speech.

High school mathematics grade-point average:

the average of grades earned in algebra, business
arithmetic, geometry, trigonometry and advanced
mathematics courses.

High school foreign language grade-point average:
the average of grades earned in courses such as
French, German, Latin and Spanish.

High school social science grade-point average:
the average of grades earned in civics, eonomics,
geography, psychology, sociology. United States
history, and world history.

High school natural science grade-point average:
the average of grades earned in biology. chemistry.
physics, physiology, and in some cases, health
education.

High school electives (non-academic) grade-point
average: the average of grades earned in subject
areas such as art, business administration. home
economics, music, and manual arts. This average
does not include grades from courses such as
physical education. driver training. study hall.
stage crew, and other courses of non-scholastic
content.

Guilford-Zimmerman Aptitude Survey. Form A, Part I,
Verbal Comprehension: primarily a vocabulary test
which requires an understanding of word and con-
cept meanings.

Guilford-Zimmerman Aptitude Survey, Form A,

Part VII, Mechanical Knowledge: a survey of
knowledge of the functions of tools commonly used
in the home, for repairing automobiles, or in one
of the trades such as carpentry or plumbing.

Educational Testing Services Cooperative English
Test, Form OM, Part I, English Usage: a test of
the ability to recognize and correct errors in
grammar and diction, punctuation, capitalization
and sentence structure.

Table l.

36

Continued.

 

 

 

variable

 

10

ll

12

l3

14

15

l6

17

18

ETS Cooperative English Test, Form OM, Part II,
Spelling: a test of the ability to recognize the
misspelled word in groups of five words.

ACE Cooperative General Achievement Test, Form
X, Section III, Mathematics, Part I, Terms and
Concepts: requires understanding of ideas and
detection of logical errors in quantitative and
spatial concepts.

ACE Cooperative General Achievement Test, Form X,
Section I, Social Studies, Part II, Comprehension
and Interpretation: requires the interpretation
of readings, charts, and tables as well as the
definition of terms and concepts relevant to
different social science areas. It tests the
ability to understand the central thought and
important details in a passage, to draw inferences
from the passage, and to appraise it critically
in order to detect contradictions, bias, and
irrelevant data.

ACE Psychological Examination, 1948 Edition,
Quantitative Reasoning Score: includes problems
involving number series, figure:analogies, and
arithmetic.

Age: chronological age.
Sex: a designation of either male or female.

ACE Psychological Examination, 1948 Edition,
Linguistic Score: primarily a test of verbal
(rather than "school learned") abilities which
consist of subtests involving verbal analogies,
vocabulary completion types of items, and same-
opposites.

Cooperative English Test, Text C2, Reading Speed,
Higher Level, Form Z: a test to answer questions
directly or indirectly related to reading of
paragraphs.

Cooperative English Test, Test C2, Reading Com-
prehension, Higher Level. Form Z: a check of
accuracy of reading paragraphs and interpreting
questions concerning such paragraphs.

 

37

These grade-point averages are summarized at three levels--
cumulative freshman, cumulative freshman and sophomore. and
total grade-point average. These data were obtained from

records in the Office of the Registrar.

Criterion Areas

The preparation of the computer tape record of the
criterion area grades used in this study followed a summari-
zation procedure as given in a computer program developed
at the University of Washington and modified for use-at
Washington State University.1 The program summarized a
grade-point average for each student in a specified criterion
area, or subject-matter area, and indicated the number of
credit hours taken. The grade-point averages were calcu-
lated regardless of a student's credit hours, so that a
3.0 grade point for 2.0 credit hours was recorded in the
same way as a 3.0 grade point for 35.0 credit hours. A
cumulative grade-point average was calculated, at the end
of the freshman year, for the freshman and sophomore years,
and at the end of the senior year. The classification of
the curriculum into appropriate criterion areas and the
number of students in each is shown in Table 2-

The grades were recorded and averaged for Washington

State University students with the freshman of 1958-1960

 

lGil Atkinson, "An IBM 709 Grade Summarization Pro-

gram, IBM Type 709 Program Library Report, Seattle, Division
of Counseling & Testing Services, 34 pp (ditto).

38

Table 2. Number of students constituting each of the original
criteria groups.

 

 

 

Subject Area ngés Subject Area ngés
All University 2293 Geography 572
Accounting 292 Geology 589
Anthropology 862 History 961
Architecture 74 Journalism 101
Art 608 **Home Economics 320
Bacteriology 499 **Nutrition 164
Biology 543 Mathematics 1246
Botany 402 Music 472
Business Administration 365 **Nursing 58
Chemistry 1042 Philosophy 545
Economics 742 Pharmacy 30
Drama 133 Physics 233
Education 647 Political Science 644
*Engineering 427 Psychology 1516
English Composition 391 Radio & TV 59
English Literature 879 Sociology 1541
Languages 344 Speech 657
*Forestry 64 Zoology 749

 

*Males Only.

**Fema1es Only.

39

as Group No. l and the 1961 freshmen as Group No. 2. The
study teminated with the spring semester of 1964. Thus.
for each group of students, except the 1961 freshmen, all
four years of academic study were recorded. For the 1961
freshmen only three years of records were available for use.
Previous studies at the University of Washington reported
little differences between achieved grades which had been
summarized at the junior or the senior level.1
The criterion areas defined are those common to
both the University of Washington, as developed and used
in the Washington Pre-College Testing Program, and to
Washington State University. Reasonable care was taken
to determine that matching criterion areas were identical.
The procedure for summarization of the grade point
averages in the study followed that used at the University
of Washington and in the Washington Pre-College Testing
Program, since the purpose of the study was to assess the
validity of the grade predictions made from the criterion
areas at Washington State University and those at the

University of Washington.

Procedure for Study

The data used in this study were either processed

on IBM cards or processed by computers using magnetic tape,

 

lPaul Horst, Differential Prediction of Academic

Success. Technical Report. Office of Naval Research Contract
Nonr-477 (08), University of Washington Division of Counseling
and Testing Services, May, 1959.

40

thereby collating and arranging the data economically for
computer processing. Each student's predictor data, his

raw test scores and high school grades, and his predicted
college grade points, as presented on the Washington Pre-
College Testing Program data sheet, needed only minor changes
to be suitable for the study. These data were already on
IBM cards and had been used previously to print the data
sheet sent to the student. Three cards were taken from the
files for each student's data: the entrance card, prediction
card No. l, and prediction card No. 2. These cards are
presented in Figure 2. A special mark sense card. with the
student's Washington State University identification number,
which had already been punched, was then marked by clerical
help with the state-wide testing program identification
number, processed by the IBM reproducer punch, then collated
with the entrance and prediction cards. The mark sense

card then contained the identification number for each of
the two systems to be used in securing data, the state-
wide testing system for the predictor variables and pre—
dicted grades, and the university system for obtaining the
achieved college grade.

The placing of the garde-card records on computer
magnetic tape was initiated in 1962 when the Office of
Institutional Research implemented a university system of
records in which the entire grade records for all students
were placed on tape each semester. These records began

with the 1958 fall semester. The records were arranged by

41

 

05250056

EE

III ooooofllooomanoo oocoooon ognomooomoo

"I H1L“ mt.

uuuocn
III

IIIIII
22 2222222
3333333333
4444444444
555555 555

CRAFT

_- . £957 new:
23 25 17 23
wmwgma

 

LI...

muq

It: ,
I

IflIZII LAST NAME

IILI HI

2L
33133; 333
4

332

 

 

3i

5 KIL

KA

‘ '45? II“ film II".

20
. .36.

SCHOOL (R

IIIIIIIIIIIIJI
22222222222222 2"

3333333L33
44444”*6’44444444444
555:.55555555 55555 55L55
5 55555 5555555555 55555555555

5
4
I
4

28

_L

 

55

555

7777777777777777777777777777777777
55555 5555555L:555555555555555555555

l2 3 a s o 1 I 0 10111 1213141: 1111 u 11232122232423.5121an“:

H999 9999599999 99L9999

 

23

37

II

55'
55555555
7771

“M

IIHI
2723
22!”.

L
3533

 

4444

BSL

£115: '

FCHpDL

39';

”.3

0
55

s.
I
2

'—

”:44
.I
30;

3'?

”:9...-

3

4444

 

5555

7777
5555
99999999;

31311353313133: 1101.141

5555

 

77
55

00

 

55
99

 

 

29L22YE 21

w moMn'ocMOM

oL'nu-cho:
I IL I
31:32 33 34 35 T'-

1111111151.» :21 .151 3. .- Irma-111211114.

52

ELL;

MIN FIB

 

26
42

 

 

 

 

 

 

77
47

40

 

77
40

 

25151515
11111
45 46
22
33

44

aLno

IILIII II

 

'

l

2N L
33353
4444
5555
5455:5555
777L777?
55555559

93993999

2
3
4

 

 

 

33
44L4
5555
55
77

33

 

J5

 

 

 

55

 

 

 

 

55
gm

 

 

 

 

 

 

 

99 999 959 59
n

47'

.5

55
99

nfln.

0;”
..
2
3

1'
55
655
as
as

 

 

 

 

06250556

555 55555MWBH555555 55555555 535M059 NOMfiNODO
"MIL WJI

NUMBER
uII

IIIIII
22 22222222
3333333333WJ3L33'
4444444 444
555555 5555

   

CRAFT
_ LAST NAME
1672

{11:23, 1

RES. LAST NAME

.L'“ ISI' .32!)
11L1 at
L I
. .

5
I

313 333332
111414111441111
' 5-55555 555

 

5 55555 ssssstssssc 5555553
7777777777777H7777777777777

52345575

W855555885588585

 

 

5I5lII2IJI4I$I$I7

54.0.5 203530

215IIII mial
”5 5 I5 [5‘47 IOI

522222222222

 

 

I
L

SCHOOI-‘ YR

IIIIIIIIIIIIJI

q

22 42
333313
1111.311

55 ”5

5555 5
7777L77

LII!
III
I

2

a:

N

 

W

MN:

 

L

99 L959 9999999999 99L9fl99999

1513252l22232I232527752333 ”3211333 .3535

     

77 26 27

‘. "IIEEI. I InlISII

I557

II
4

III
3

 

2
33

 

 

 

222
3333

11L_t:sv vea_1_u_.:1..tm1.

5'51

I
5

 

 

 

 

               

111-1 "fathom. mntm

20

557

 

' 27

- 1111mm ".2

.23

MUN"

59
u ' 3111

-..— ,

II'N.’

 

 

 

1:31" 1:1.

 

27

mm:

 

   

11'151

GIL‘MLI .1; 3.1
II‘I
El 22

292
3333

 

 

 

 

 

a"Woo .‘p 105053
' ”-L‘:

11‘ IIII
17hs!.3;.o

F .
I

I ”L2 212
1
assaaazahsaaL 33
L

murmur
LII II

II IIZ‘

r

1%
3

2

”III“
MIIS

" .
4

I

 

.DL

 

 

s (-

 

 

 

33

 

 

 

L
.4444. 4W4444y44
5

4444L4444
”fl555555
553555

I

5 .5fl5*55 L
LL 77WTL777J
L

 

r:

V

H55 .55

U

 

L
L .....
L5 5i5%555
5655555
7
8

 

 

 

 

p

55550

 

 

7J 7

 

 

 

 

I
71

‘I‘IL
" at

58L35fl858 BCJTI 5135

:L .1 L

858885i85L88

 

 

 

a5
“7351 '54“

99999999"7

QSIQSIG 99
47484541137: 7.3.11.5

'.
:4 u»!. u.

"as..§999&1

.111:ta.r...11c:17.

23

33

3‘

74175 7.‘ SI .‘3

Oflfl

'03; m.

III

2
2
33
444
555
555
7

N

 

 

Figure 2.

Program.

IBM cards from Washington Pre-Colle

ge Testing

 

 

 

 

42

University identification number and followed a standard
blocking system whereby one tape contained all grades for
one semester. The grades were recorded following the standard
format of the actual IBM card containing the recorded grades.

A computer program for the IBM 1401 was used to
place the grade records for those students included in the
study on a master tape which accurately recorded all the
data.1 Records that did not match or were incorrectly identi-
fied were corrected and entered later on the master tape
file.

In this study Group No. l, which is composed of
students entering Washington State University in the fall of
1958, 1959 and 1960, was used as the initial sample from
which multiple regression data was formulated, using the
grades from the Washington State University criterion areas.
Group 2 contains the predictor data for students entering
Washington State University in the fall of 1961 and was
used as the cross-validity sample to check the prediction
efficiency of the grade predictions from the Washington Pre-
College Testing Program and secondly, the grade predictions
made from data derived from Group 1.

The 36 criterion areas used in the study were re-
tained even though 5 areas registered below 100 individual

students. Walley reported that a greater amount of variance

 

1James E. Thummell, An IBM 1401 Computer Program for
Grade Summarization at Washington State University.

43

in multiple correlation problems is accounted for as the
number of cases is increased up to 100, with gains then
becoming more gradual.1

In computing grade-point averages, each credit hour
of A was valued at four grade points, each credit hour of
B at three grade points, C at two, D at one, and E at zero.
Incompletes (I) and withdrawals (W) were not recorded directly
and thus were not included in determination of the grade-
point average. Grades in military science and physical
education activity courses were excluded in computing the
all-university grade-point average.

Computation of Intercorrelation and
Validity Coefficient Matrices

Computation of intercorrelations and validity co-
efficients was made by using the standard programs for the
IBM 709 computer.2 The symmetric predictor intercorrelation
matrix was made from tape records of the entrance data cards
of the 1958-1960 Group 1 freshmen. There were 18 predictor
variables in this operation. The non-symmetric criterion
correlation coefficient matrices were calculated by using
SHARE program 215 on the IBM 709. This program gave the
means, standard deviations, and variances for all 18 variables

as well as the zero order correlation coefficients between

 

lDonivan Walley, "Factors that Influence the Selection

of Predictor Variable in Multiple Regression," College and
University, 1963, 39:72-76,

 

 

2IBM 709 Correlation Matric Program, Washington State
University,

44

the predictors and grade-point averages in the 36 criterion
areas.1

The Horst multiple regression, or iteration method,
was used for the computation of the corrected multiple
correlation coefficients (R's) between the predictors and
the criteria of students' achieved grades for the criterion
areas at Washington State University.2 The Horst method
simultaneously computed the beta weights (B's) of the
predictors which consisted of successive approximations of
the contribution of each independent predictor variable to
the dependent criterion variable. The iteration procedure
selected in sequence the predictors, which, when combined,
yielded the largest corrected multiple correlation (Rc)
with the criterion. The iteration method selected the
highest value in the predictor-criterion vector, which was
used to multiply each element in the corresponding vector
of the predictor intercorrelation matrix. The products
obtained were subtracted from their corresponding correlations
in the predictor-criterion vector in order to produce a

second criterion of residual vector. The same method was

then repeated or iterated, with the residual criterion vector.

 

1IBM 709 Correlation Program for SHARE 215.

2Richard C. Sorensen and August Dvorak, "An IBM
Type 709/7090 Program to Select Predictors, Calculate
Multiple Correlations, and Determine Linear Regression
Equations," IBM Type 709 Program. Library Report (Seattle:
Division of Counseling & Testing Services, n.d.), 10 pp.
(Ditto.)

45

This iteration procedure was continued until the
residual vector values for all predictors reached the pre-
determined limit of .0010. After each iterative cycle, or
repetition, the method provided the following: the squared
multiple-correlation coefficient (R2), a result of the cumu-
lative sum of the square of the high values on the predictor-
criteria vectors; the corresponding corrected squared multiple
correlation coefficients (Rcz); the corrected multiple
correlation coefficient (Rc); the beta coefficient (B's);
and the b-weights corresponding to each of the B's.

A spuriously high multiple correlation might have
occurred since the iteration technique used in the study
selected the highest value in the predictor—criterion vector.
The multiple correlation of this predictor-criterion vector
is always higher than the initial value of the high multiple
correlation. Therefore, a small sample could lead to the
spuriously high correlation for some of the other criterion
areas.

The predictor b-weights determined from the Horst
iteration procedure on the Group 1 1958-1960 Washington
State University freshmen were substituted in the present
IBM 709 computer program used in the Washington Pre—College
Testing Program. New predicted grade-point averages were
computed for the 1961 Group 2 Washington State University
freshmen in the 36 criterion areas. A comparison was then
made of the grade predictions in the 36 criterion areas, by

the two different prediction methods, data based on grades

46

obtained by students at University of Washington and by
students at Washington State University. The comparison of
the two prediction methods was made on the 1961 Group 2
Washington State University entering freshmen. Thus, for
each freshman enrolled, the comparison involved the actual
achieved grade-point average, with the predicted grade point
based upon normative data at University of Washington and
with the predicted grade point based upon normative data at

Washington State University.

Summary

The sources of data for this study were the test
records and achieved grades for the 1958—1961 entering fresh—
men at Washington State University. The data were primarily
used with computerized procedures and programs, special
programs were modified for the grade summarization procedure
and the iteration procedure.

The procedure for the study were calculated to
(Sevelop, by use of the Horst iteration method, a set of
multiple predictions using the achieved grades and test
data for Washington State University students and comparing
these predictions with those of the Washington Pre-College

Testing Program.

Chapter IV will present the evaluation and interpre-

tation of this data.

CHAPTER IV

EVALUATION'AND INTERPRETATION OF DATA

The purpose of this study was to evaluate for stu-
dents entering Washington State University the grade pre-
dictions from the Washington Pre-College Testing Program.
These grade predictions were based upon normative data
from students at another university. This chapter is con-
cerned with four questions as pertaining to Washington State
University students.

1. As measured by high school grades and testing data
on the Washington Pre-College Testing Program are there
differences in student aptitudes and achievements between
students in a Washington State University normative group
and students in the University of Washington's normative
group?

2. How accurate are the current predicted grades of the
Washington Pre-College Testing Program for students at
Washington State University at various levels of student
academic progress--freshman, sophomore, or senior level?
Can the grade prediction be used to predict grades at all
three levels equally well?

3. Is there a hierarchy in the subject matter areas as
represented by the predicted and achieved grades from the

Washington Pre-College Testing Program and what similarity

47

48

does this hierarchy have to the predicted and achieved
grades for students at Washington State University?

4. Are the present predictor equations used in the
Washington Pre-College Testing Program valid for Washington
State University students or should there be developed pre-
dictor equations based on the grades achieved by students
at Washington State University?

1. As measured by high school grades and testing data
on the Washington Pre-College Testing Program are there
differences in student aptitudes and achievements between
students in a Washington State University normative group
and students in the University of Washington's normative
group?

At Washington State University there were 2,276
freshmen in the 1958-1960 group--48% women and 52%.men.

The 1954-1955 University of Washington group was compOsed

of 5,531 students--35.7%.women and 64.3% men. The means and
standard deviations for high school grades and test scores
are shown in Table 3. The 1958-1960 freshmen at Washington
State University had higher mean grades in four high school
subject areas and more restricted standard deviations in

all six subject areas.

The Washington State University normative group
tested higher than the University of Washington normative
group in English Usage (mean difference of 12.01), spelling
(mean difference of 1.56), mathematics (mean difference of
1.73), social studies (mean difference of 0.20), and A.C.E.-Q
(mean difference of 3.10). The University of Washington

group tested higher on the Guilford-Zimmerman Mechanical

Knowledge (mean difference of 2.90) and was higher in age

Table 3.

49

Means and standard deviations of predictor variables

for 1958-1960 freshmen at Washington State Univer-
sity and 1954-1955 freshmen at University of

Washington.

 

 

 

 

Predictor variables 1.2:." “is? .233. 3:3: 3:22.
1. H. S. English 2.89 .65 2.84 .69 .05
2. H. S. Mathematics 2.68 .74 2.64 .77 .04
3. H. S. Foreign Language 2.70 .80 2.66 .82 .04
4. H. S. Social Science 2.97 .65 2.90 .70 .07
5. H. S. Natural Science 2.80 .69 2.81 .70 .01
6. H. S. Electives 3.11 .55 3.11 .55 .00
7. Guilford—Z verbal 25.06 10.05 25.88 11.10 —.82
8. Guilford-Z Mech. Kn. 14.22 12.42 17.12 13.42 -3.90
9. English verbal 99.45 26.03 87.44 28.37 12.01

10. Spelling 17.86 9.70 16.30 9.75 1.56
11. Mathematics 22.13 9.31 20.40 9.32 1.73
12. Social Studies 17.21 6.62 17.01 7.05 .20
13. A.C.E. - O 47.14 9.43 43.94 10.62 -3.20
14. Age 18.29 .63 18.67 2.03 -.38
15. Sex .48 .48 .36 .48 .12
16. A.C.E. - L 67.24 13.54 65.29 14.13 1.95
17. Coop. Reading Speed 24.91 9.54 26.13 9.67 -1.22
18. Coop. Reading Level 16.35 5.13 18.41 6.03 -2.06
Total N = 2,276 5,531

50

(mean difference of .38 years).

The larger differences in Guilford-Zimmerman Mechani-
cal Knowledge, English Usage, and A.C.E.-Q may be explained
by the larger number of men in the University of Washington
group.1 The University of Washington group, with a mean
age of 18.67, was older than the Washington State University
group whose mean age was 18.29.

An inspection of the high school grades and test
scores of the two normative groups showed the two groups
were similar and comparable for the purpose of this study:
i.e., the purpose of the use of test scores and high school
grade averages for comparative use within the counseling
and guidance functions. A statistical comparison was not
used for the small differences, such differences being ir—
relevant to any comparative interpretation drawn from test
scores and grades and used by the personnel in the guidance

setting.

2. How accurate are the current predicted grades of the
Washington Pre-College Testing Program for students at
Washington State University at various levels of student
academic progress--freshman, sophomore, or senior level?

Can the grade predictions be used to predict achieved grades
at all three levels equally well?

The grade summarization program for the IBM 709

computer summarized three sets of achieved grades--cumu1ative

 

lLouise B. Heathers, Robert Kintneo, Thomas D.
Langen, and Susan Bjork, "Comparison of Male and Female Stu-
dents in the 1961 Entering Freshman Class of the University,"
part of DCT Project 0961-100 (Seattle, Washington: Division
of Counseling and Testing Services, University of Washington).
(Dittoed .)

51

freshman (01), cumulative freshman and sophomore (02), and
cumulative freshman, sophomore and junior (03) averages

for the 1961 Washington State university freshmen.l The
achieved grades were summarized in grade-pohat averages and
used subsequently as equals, regardless of the calculated
credit hours contained in each number. For each student
the differences were calculated between the Washington Pre-
College Testing Program predicted grade-point average and
the achieved grade point at each of the three levels. The
derivation of the frequencies, the percentages of the
errors of the prediction, and the cumulative percentages
were calculated for the 36 criterion areas using Program
0083 for the IBM 1401.2 The cumulative percentages are
presented in Table 4 for each of the three methods of
calculation.

The summary of absolute differences, expressed as
cumulative percentages between the predicted and the three
levels of achieved grades, cumulative freshmen (01), cumu-
lative freshmen and sophomore (02), cumulative freshmen-
junior (03), illustrated a similarity between percentages
for the 36 criterion areas. Statistically differences at
the .05 level were found for some absolute differences for

Bacteriology, Home Economics, Political Science, and Zoology.

These differences were primarily between the cumulative

 

lAtkinson, op. cit.

2James Thummel, "An IBM 1401 Program to Summarize
Differences in Percentages of Grade—Point Averages."
Washington State University. (Dittoed.)

52

 

 

 

Table 4. A comparison of differences between W.P.C.T.P.
predicted grade—point averages and achieved grade—
point averages for the 1961 freshmen at W.S.U.
Differences given in cumulative percentages for
three levels of achievement, freshmen (01),
freshmen and sophomores (02), and juniors and
seniors (03).
All-University Accounting
5b5°IUte 01 02 03 01 02 03
Difference
0.0 8.90 9.03 9.24 3.70 3.51 5.13
0.1 25.62 25.39 24.80 7.40 14.06 14.37
0.2 40.72 40.40 39.23 14.80 21.09 22.24
0.3 53.29 53.36 52.92 25.91 28.12 33.19
0.4 63.24 64.05 63.69 29.61 35.65 39.69
0.5 71.18 71.86 72.06 40.72 44.69 49.96
0.6 79.21 78.80 79.43 44.42 51.72 58.52
0.7 84.53 84.73 84.92 70.34 59.25 65.02
0.8 88.85 88.96 89.10 81.45 65.78 71.18
0.9 91.73 91.97 91.89 81.45 70.80 75.97
1.0 93.78 94.54 94.24 85.15 72.32 80.42
1.1 95.83 95.80 95.94 88.85 80.84 83.84
1.2 96.87 97.02 96.94 96.25 83.85 87.26
1.3 97.87 97.84 97.81 96.25 85.35 89.31
1.4 98.43 98.40 98.42 96.25 89.37 90.67
1.5 98.73 98.70 98.76 99.95 90.87 93.06
1.6 99.21 99.18 99.23 99.95 91.87 93.40
1.7 99.25 99.26 99.31 99.95 92.37 93.74
1.8 99.42 99.43 99.48 99.95 92.37 94.42
1.9 99.59 99.60 99.65 99.95 95.38 96.13
2.0 99.63 99.64 99.69 99.95 96.88 96.81
2.1 99.67 99.68 99.73 99.95 97.88 97.83
2.2 99.80 99.81 99.86 99.95 97.88 98.85
2.3 99.84 99.85 99.86 99.95 98.88 98.85
2.4 99.88 99.89 99.90 99.95 99.38 99.53
2.5 99.88 99.89 99.90 99.95 99.88 99.53
2.6 99.88 99.89 99.90 99.95 99.88 99.53
2.7 99.88 99.89 99.90 99.95 99.88 99.87
2.8 99.88 99.89 99.90 99.95 99.88 99.87
2.9 99.88 99.89 99.90 99.95 99.88 99.87
3.0 99.88 99.89 99.90 99.95 99.88 99.87

 

53

 

 

 

 

Table 4. Continued
Anthropology Architecture
AbSOIute 01 02 03 01 02 03
Difference
5.67 5.92 5.56 0.00 1.42 1.35

0 1 14.27 15.44 16.00 8.33 8.56 6.75
0.2 23.39 24.32 25.51 19.99 18.56 16.20
0.3 29.93 31.65 33.28 26.65 25.70 24.30
0.4 36.64 38.85 40.35 34.98 35.70 33.75
0.5 45.59 47.34 49.63 43.31 44.27 41.85
0.6 52.64 54.67 56.47 48.31 47.12 47.25
0.7 58.66 60.71 62.15 53.31 54.26 52.65
0.8 66.92 68.56 69.69 54.97 59.97 60.75
0.9 72.77 74.09 75.25 56.63 61.39 63.45
1.0 77.76 78.46 80.00 58.29 65.67 67.50
1.1 81.37 82.06 83.48 64.95 75.67 75.60
1.2 85.50 85.53 86.96 69.95 78.52 78.30
1.3 87.90 87.84 89.04 76.61 82.80 82.35
1.4 89.62 89.89 91.01 79.94 84.22 83.70
1.5 92.20 92.20 92.98 81.60 85.64 85.05
1.6 92.71 93.35 94.14 86.60 91.35 90.45
1.7 94.08 94.50 95.06 88.26 91.35 90.45
1.8 95.80 95.91 96.33 89.92 94.20 93.15
1.9 96.83 97.06 97.25 91.58 94.20 93.15
2.0 97.51 98.21 98.17 93.24 95.62 94.50
2.1 97.85 98.33 98.28 93.24 95.62 95.85
2.2 98.36 98.84 98.86 94.90 97.04 97.20
2.3 99.04 99.35 99.32 96.56 98.46 98.55
2.4 99.38 99.60 99.66 98.22 99.88 99.90
2.5 99.72 99.72 99.77 98.22 99.88 99.90
2.6 99.72 99.72 99.77 99.88 99.88 99.90
2.7 99.89 99.84 99.88 99.88 99.88 99.90
2.8 99.89 99.84 99.88 99.88 99.88 99.90
2.9 99.89 99.84 99.88 99.88 99.88 99.90
3.0 99.89 99.84 99.88 99.88 99.88 99.90

 

54

 

 

 

 

Table 4. Continued.
Art Bacteriology

Absolute

Difference 01 02 03 01 02 03
0.0 3.39 3.13 3.78 3.40 3.66 5.01
0.1 7.82 10.31 10.19 10.78 12.03 13.62
0.2 14.08 17.49 17.92 19.30 20.93 24.04
0.3 22.95 26.32 26.80 26.11 28.78 31.45
0.4 35.22 37.18 37.49 35.76 40.29 41.47
0.5 48.01 49.88 50.48 40.30 47.09 48.68**
0.6 60.80 62.77 62.98 46.55 55.72 57.69**
0.7 69.93 71.42 71.53 51.66 60.69 62.90**
0.8 75.67 76.94 76.79 57.91 66.44 69.11**
0.9 79.58 81.17 81.06 62.45 70.36 73.51**
1.0 81.92 83.93 84.18 68.13 76.11 79.32**
1.1 83.48 85.40 86.31 72.10 79.51 82.72**
1.2 85.30 87.05 87.95 75.50 83.17 85.92**
1.3 88.17 89.62 90.08 81.18 87.35 89.32
1.4 91.04 92.19 92.71 84.58 89.96 91.52
1.5 92.86 93.66 94.02 85.71 91.79 92.52
1.6 94.68 95.50 95.82 87.41 93.09 93.92
1.7 94.94 95.68 95.98 89.11 94.92 95.32
1.8 95.46 96.23 96.63 91.38 96.75 96.32
1.9 95.46 96.41 96.79 91.94 96.75 96.52
2.0 95.72 96.59 96.95 94.21 97.01 96.72
2.1 95.98 96.77 97.11 97.05 98.58 98.12
2.2 96.50 97.13 97.43 98.18 98.84 98.52
2.3 97.02 97.49 97.75 98.74 99.10 98.92
2.4 97.80 98.22 98.40 98.74 99.10 99.32
2.5 98.58 98.95 99.05 99.30 99.36 99.52
2.6 99.62 99.68 99.70 99.86 99.88 99.92
2.7 99.62 99.68 99.70 99.86 99.88 99.92
2.8 99.62 99.68 99.70 99.86 99.88 99.92
2.9 99.88 99.86 99.86 99.86 99.88 99.92
3.0 99.88 99.86 99.86 99.86 99.88 99.92

 

**Significant at

.05 level.

55

 

 

 

 

'Table 4. Continued.
Biology Botany
5b3°1Ute 01 02 03 01 02 03
Difference
0.0 3.78 2.94 2.76 2.27 2.67 2.23
0.1 16.71 14.30 13.62 9.54 10.97 9.69
0.2 24.59 22.08 21.53 14.54 16.60 16.90
0.3 33.73 31.34 31.29 22.26 24.01 24.36
0.4 42.56 39.34 39.57 29.53 31.42 31.82
0.5 53.60 49.86 49.69 37.25 38.54 38.28
0.6 58.64 55.33 55.39 45.43 46.55 44.99
0.7 65.26 62.27 62.20 51.79 53.07 51.20
0.8 71.88 69.00 68.64 57.69 58.70 56.42
0.9 76.92 74.68 74.90 62.69 63.74 61.89
1.0 80.39 79.73 80.42 65.41 67.30 65.62
1.1 86.06 84.15 84.83 70.86 72.64 70.84
1.2 88.89 87.51 88.32 75.40 77.68 75.56
1.3 89.83 88.56 89.60 79.49 82.13 81.03
1.4 91.72 90.45 90.70 82.67 86.58 85.25
1.5 92.66 92.13 92.17 86.30 ‘89.25 88.48
1.6 94.23 93.81 93.45 89.02 91.32 89.97
1.7 95.80 95.91 96.02 92.20 93.99 92.70
1.8 97.06 97.17 97.12 94.47 95.77 95.18
1.9 97.37 97.80 97.67 95.37 96.36 95.67
2.0 98.00 98.64 98.40 95.37 96.36 95.91
2.1 98.63 98.85 98.58 96.27 97.25 97.15
2.2 98.94 99.06 98.94 96.72 97.54 97.39
2.3 99.25 99.48 99.30 98.08 98.13 97.88
2.4 99.25 99.48 99.30 98.08 98.42 98.37
2.5 99.25 99.69 99.48 98.53 98.71 98.61
2.6 99.56 99.90 99.84 98.53 99.00 98.85
2.7 99.56 99.90 99.84 98.98 99.29 99.34
2.8 99.56 99.90 99.84 98.98 99.29 99.34
2.9 99.56 99.90 99.84 99.43 99.58 99.58
3.0 99.87 99.90 99.84 99.88 99.87 99.82

 

56

 

 

 

 

Table 4. Continued.
Business Administration Chemistry

5bs°lUte 01 02 03 01 02 03

Difference
0.0 2.88 5.11 6.57 4.22 4.27 4.60
0.1 14.41 13.77 19.72 15.22 15.09 15.54
0.2 28.83 24.39 30.40 23.66 25.21 25.52
0.3 37.48 35.41 40.81 33.32 34.44 34.73
0.4 44.21 45.25 50.67 41.87 43.27 43.46
0.5 49.97 54.69 59.16 51.42 51.81 51.90
0.6 55.73 60.59 64.36 58.42 59.25 60.24
0.7 62.46 66.88 69.56 64.86 65.60 66.47
0.8 70.15 72.78 76.13 71.08 71.35 72.61
0.9 74.95 78.29 80.78 75.19 75.71 76.83
1.0 78.79 83.40 86.53 79.07 80.47 81.05
1.1 78.79 84.97 87.89 81.95 83.64 84.60
1.2 83.59 88.90 91.17 85.28 86.81 87.38
1.3 88.39 91.65 93.08 87.83 89.39 90.16
1.4 88.39 93.61 94.72 90.83 91.87 91.98
1.5 89.35 94.79 96.08 92.38 93.26 93.32
1.6 91.27 96.36 96.90 94.49 95.24 95.23
1.7 92.23 97.54 97.72 95.60 96.33 96.28
1.8 92.23 97.93 97.99 96.26 96.82 96.85
1.9 93.19 98.32 98.26 97.26 97.71 97.71
2.0 93.19 98.32 98.26 98.14 98.40 98.47
2.1 96.07 98.32 98.53 98.80 98.89 98.94
2.2 97.99 99.10 98.80 99.02 98.98 99.03
2.3 98.95 99.49 99.07 99.02 99.07 99.12
2.4 98.95 99.49 99.07 99.46 99.36 99.40
2.5 99.91 99.88 99.34 99.57 99.45 99.49
2.6 99.91 99.88 99.61 99.57 99.45 99.49
2.7 99.91 99.88 99.88 99.68 99.54 99.58
2.8 99.91 99.88 99.88 99.90 99.73 99.77
2.9 99.91 99.88 99.88 99.90 99.73 99.77
3.0 99.91 99.88 99.88 99.90 99.82 99.86

 

57

 

 

 

 

Table 4. Continued.
Economics Drama

5bS°1ute 01 02 03 01 02 03

Difference
0.0 6.06 5.21 5.25 2.08 2.94 5.26
0.1 24.24 17.26 17.10 14.58 11.76 13.53
0.2 30.30 30.20 30.44 22.91 24.50 24.80
0.3 33.33 39.73 38.52 27.07 38.22 36.07
0.4 33.33 47.46 47.28 33.32 48.02 44.34
0.5 36.36 53.75 54.69 33.32 53.90 54.86
0.6 39.39 61.66 62.23 47.90 62.72 63.13
0.7 51.51 67.05 67.35 66.65 76.44 75.16
0.8 60.60 72.44 73.95 72.90 82.32 81.17
,0.9 66.66 76.75 78.53 74.98 83.30 84.92
1.0 78.78 81.60 82.70 79.14 86.24 86.42
1.1 90.90 85.55 86.87 83.30 88.20 89.42
1.2 90.90 87.16 89.29 95.80 93.10 93.93
1.3 90.90 88.77 91.04 95.80 94.08 95.43
1.4 93.93 92.00 92.92 95.80 96.04 96.18
1.5 93.93 93.61 94.26 95.80 96.04 96.18
1.6 93.93 94.86 95.33 95.80 97.02 96.93
1.7 93.93 95.75 96.27 95.80 98.98 97.68
1.8 93.93 96.46 97.07 95.80 98.98 98.43
1.9 93.93 96.81 97.33 97.88 99.96 99.18
2.0 96.96 97.34 97.73 97.88 99.96 99.18
2.1 96.96 98.59 98.67 97.88 99.96 99.18
2.2 96.96 99.48 99.47 99.96 99.96 99.18
2.3 99.99 99.48 99.60 99.96 99.96 99.18
2.4 99.99 99.48 99.73 99.96 99.96 99.18
2.5 99.99 99.65 99.86 99.96 99.96 99.18
2.6 99.99 99.82 99.86 99.96 99.96 99.18
2.7 99.99 99.82 99.86 99.96 99.96 99.18
2.8 99.99 99.82 99.86 99.96 99.96 99.18
2.9 99.99 99.82 99.86 99.96 99.96 99.18
3.0 99.99 99.82 99.86 99.96 99.96 99.93

 

58

 

 

 

 

Table 4. Continued.
Education Engineering

5b3°lute 01 02 03 01 02 03

Difference
0.0 5.22 5.06 5.10 5.72 4.18 4.44
0.1 16.65 18.98 19.16 14.47 15.17 14.74
0.2 27.10 31.81 31.98 19.52 21.71 21.06
0.3 39.19 43.20 46.19** 25.91 28.77 26.91
0.4 48.01 49.89 54.38 33.99 35.57 35.10
0.5 58.79 60.55 63.96 38.36 41.32 41.18
0.6 72.18 69.22 71.99 45.43 46.29 47.03
0.7 78.06 74.28 76.16 49.80 53.61 52.41
0.8 82.63 78.61 80.17 55.86 59.10 56.85
0.9 87.20 82.04 84.18 62.25 64.85 61.29
1.0 91.77 86.19 87.27 65.95 68.25 65.03
1.1 93.73 89.44 89.89 68.30 72.96 70.18
1.2 94.71 91.97 92.36 71.66 75.05 71.81
1.3 95.69 93.77 94.06 75.02 78.19 74.15
1.4 97.32 95.57 95.91 77.04 80.02 76.49
1.5 97.32 95.75 96.06 80.07 80.80 76.72
1.6 98.30 97.01 97.45 83.10 82.89 78.12
1.7 98.30 97.19 97.60 85.12 83.41 78.58
1.8 98.62 97.91 98.06 88.15 85.24 80.21
1.9 98.94 98.27 98.21 90.84 87.07 82.08
2.0 98.94 98.45 98.36 92.86 88.11 83.25
2.1 98.94 98.63 98.51 93.87 88.37 83.48
2.2 99.26 99.17 99.12 94.54 88.63 83.71
2.3 99.26 99.35 99.27 95.21 89.15 84.17
2.4 99.58 99.53 99.42 95.54 89.41 84.40
2.5 99.58 99.71 99.57 95.54 89.41 84.40
2.6 99.58 99.71 99.57 95.54 89.41 84.40
2.7 99.58 99.71 99.57 95.87 89.67 84.63
2.8 99.90 99.89 99.87 95.87 89.67 84.63
2.9 99.90 99.89 99.87 95.87 89.67 84.63
3.0 99.90 99.89 99.87 99.91 99.87 99.85

 

**Significant at .05 level.

59

 

 

 

 

Table 4. Continued.
English Composition English Literature

9b3°1ute 01 02 03 01 02 03

Difference
0.0 3.82 4.85 6.20 7.83 7.84
0.1 15.73 14.31 18.17 18.64 20.58
0.2 28.07 28.12 31.91 33.09 34.80
0.3 38.28 39.11 40.55 42.00 45.49
0.4 43.81 46.01 46.09 49.70 51.51**
0.5 48.91 52.91 55.18 58.34 60.49
0.6 56.99 61.60 63.60 66.31 68.90
0.7 64.64 66.97 70.47 72.93 75.49
0.8 74.85 74.89 78.67 79.55 82.08
0.9 79.10 79.23 82.88 84.14 86.28
1.0 83.35 83.32 86.87 89.14 90.71
1.1 87.17 86.13 91.08 92.11 92.98
1.2 90.99 90.22 93.74 94.13 94.91
1.3 91.84 92.26 95.07 95.48 96.16
1.4 94.39 94.05 95.95 96.15 96.95
1.5 94.81 95.07 97.28 97.09 97.74
1.6 96.08 96.34 97.94 98.30 98.53
1.7 96.50 97.36 97.94 98.70 98.98
1.8 96.50 97.61 98.16 98.70 98.98
1.9 98.20 98.88 98.82 99.24 99.20
2.0 98.20 98.88 99.48 99.51 99.42
2.1 98.20 99.13 99.70 99.78 99.64
2.2 98.62 99.13 99.92 99.91 99.86
2.3 99.04 99.38 99.92 99.91 99.86
2.4 99.46 99.63 99.92 99.91 99.86
2.5 99.46 99.63 99.92 99.91 99.86
2.6 99.46 99.63 99.92 99.91 99.86
2.7 99.88 99.88 99.92 99.91 99.86
2.8 99.88 99.88 99.92 99.91 99.86
2.9 99.88 99.88 99.92 99.91 99.86
3.0 99.88 99.88 99.92 99.91 99.86

 

**Significant at the .05 level.

60

 

 

 

 

Table 4. Continued.
Languages Forestry

Absolute

Difference 01 02 03 01 02 03
0.0 4.60 4.85 4.94 6.52 6.77 7.81
0.1 12.54 12.61 12.78 15.21 18.63 17.18
0.2 18.39 20.37 22.37 23.90 27.10 24.99
0.3 28.01 29.75 29.34 32.59 32.18 32.80
0.4 38.05 39.45 38.64 39.11 40.65 42.17
0.5 46.41 47.54 46.48 43.45 50.81 49.98
0.6 52.26 55.63 53.74 43.45 52.50 51.54
0.7 59.37 61.45 49.84 49.97 59.27 57.79
0.8 63.13 65.33 63.61 52.14 62.65 60.91
0.9 68.15 70.83 68.84 58.66 71.12 73.41
1.0 72.75 76.00 73.78 69.52 79.59 81.22
1.1 76.51 79.23 77.55 71.69 82.97 84.34
1.2 78.60 81.17 80.74 73.86 84.66 87.46
1.3 81.11 83.43 83.93 73.86 84.66 87.46
1.4 85.29 86.98 87.12 76.03 86.35 89.02
1.5 87.38 89.24 88.86 78.20 88.04 89.02
1.6 89.47 90.85 90.31 82.54 88.04 89.02
1.7 90.30 91.82 90.89 84.71 91.42 90.58
1.8 91.55 92.79 92.92 89.05 93.11 92.14
1.9 93.22 94.08 94.37 95.57 96.49 96.82
2.0 93.63 94.08 94.95 99.91 98.18 98.38
2.1 96.14 96.66 96.98 99.91 98.18 98.38
2.2 97.39 97.63 97.85 99.91 98.08 98.38
2.3 97.80 97.95 98.14 99.91 98.18 98.38
2.4 98.21 98.27 98.43 99.91 98.18 98.38
2.5 99.04 98.91 99.30 99.91 98.18 98.38
2.6 99.45 98.91 99.30 99.91 98.18 98.38
2.7 99.45 98.91 99.30 99.91 98.18 98.38
2.8 99.45 98.91 99.30 99.91 98.18 98.38
2.9 99.45 98.91 99.30 99.91 98.18 98.38
3.0 99.86 99.88 99.88 99.91 99.87 99.94

 

61

 

 

 

 

Table 4. Continued.
Geography Geology

5bS°IUte 01 02 03 01 02 03

Difference
0.0 3.73 4.04 3.67 4.00 5.17 5.09
0.1 17.43 17.19 17.30 11.33 13.03 13.74
0.2 27.08 28.52 29.01 23.33 26.48 27.15
0.3 35.17 37.22 37.05 35.33 36.62 37.16
0.4 44.51 46.73 46.49 43.33 44.90 45.30
0.5 53.54 55.63 55.93 51.33 52.14 52.60
0.6 60.39 62.31 62.22 57.66 58.55 59.05
0.7 68.48 69.39 69.03 63.66 64.55 65.16
0.8 74.71 75.05 74.97 67.66 69.72 70.08
0.9 79.38 79.90 79.69 73.99 75.51 76.19
1.0 83.42 84.55 84.23 80.32 81.10 81.79
1.1 88.09 88.39 88.25 83.98 84.82 85.86
1.2 92.13 91.62 91.39 89.31 89.16 89.76
1.3 94.31 93.84 93.31 91.31 91.85 91.62
1.4 96.17 95.86 95.58 92.31 92.67 92.46
1.5 97.41 96.66 96.80 94.31 94.32 94.49
1.6 98.03 97.46 97.67 94.97 95.14 95.16
1.7 98.34 98.26 98.36 95.97 96.38 96.51
1.8 98.34 98.26 98.36 96.97 97.00 97.01
1.9 98.65 98.66 98.53 97.63 97.62 97.51
2.0 98.96 99.06 99.05 98.29 98.24 98.18
2.1 99.27 99.26 99.22 98.62 98.65 98.68
2.2 99.27 99.46 99.39 99.28 99.27 99.18
2.3 99.58 99.66 99.56 99.28 99.27 99.18
2.4 99.89 99.86 99.73 99.94 99.68 99.68
2.5 99.89 99.86 99.73 99.94 99.68 99.68
2.6 99.89 99.86 99.73 99.94 99.88 99.84
2.7 99.89 99.86 99.73 99.94 99.88 99.84
2.8 99.89 99.86 99.90 99.94 99.88 99.84
2.9 99.89 99.86 99.90 99.94 99.88 99.84
3.0 99.89 99.86 99.90 99.94 99.88 99.84

 

62

 

 

 

 

Table 4. Continued.
History Journalism

5b5°IUte 01 02 03 01 02 03

Difference
0.0 5.19 5.30 6.45 6.81 2.77 4.95
0.1 18.54 17.24 18.62 11.35 6.93 11.88
0.2 29.48 29.66 30.58 27.25 19.43 24.75
0.3 40.05 41.84 42.75 31.79 34.70 36.63
0.4 50.43 52.93 51.69 43.15 43.03 45.54
0.5 55.81 59.68 58.87 49.96 48.58 51.48
0.6 62.30 66.43 66.77 59.05 56.91 57.42
0.7 68.23 71.85 72.90 65.86 63.85 64.35
0.8 73.42 76.55 76.64 70.40 73.57 73.26
0.9 76.75 80.41 81.01 72.67 74.95 78.21
1.0 80.64 84.51 84.65 81.76 79.11 81.18
1.1 84.16 87.52 87.66 88.57 84.66 85.14
1.2 86.76 90.29 90.67 88.57 88.82 88.11
1.3 90.27 92.46 92.85 88.57 90.20 91.08
1.4 92.12 94.75 94.82 90.84 90.20 92.07
1.5 94.16 96.07 95.86 90.84 90.20 92.07
1.6 94.53 96.31 96.06 95.38 92.97 94.05
1.7 96.75 ~97.39 97.10 95.38 94.35 95.04
1.8 97.30 97.75 97.62 97.65 95.73 96.03
1.9 97.85 97.99 97.82 97.65 95.73 97.02
2.0 98.22 98.35 98.44 99.92 97.11 98.01
2.1 98.40 98.71 98.96 99.92 97.11 98.01
2.2 98.95 99.07 99.06 99.92 97.11 98.01
2.3 99.32 99.31 99.26 99.92 97.11 98.01
2.4 99.50 99.79 99.67 99.92 97.11 98.01
2.5 99.50 99.79 99.67 99.92 97.11 98.01
2.6 99.87 99.91 99.77 99.92 97.11 98.01
2.7 99.87 99.91 99.77 99.92 97.11 98.01
2.8 99.87 99.91 99.87 99.92 98.49 99.00
2.9 99.87 99.91 99.87 99.92 98.49 99.00
3.0 99.87 99.91 99.87 99.92 99.87 99.99

 

63

 

 

 

 

 

Table 4. Continued.
Heme Economics Mathematics

Absolute

Difference 01 02 03 01 02 03
0.0 2.94 5.78 5.62 3.78 4.35 3.77
0.1 11.76 14.28 14.99 12.92 13.13 12.83
0.2 20.58 23.46 23.42 22.15 22.77 22.54
0.3 31.98 35.36 38.42 30.73 31.89 32.33
0.4 40.74 43.18 48.42** 39.50 41.01 41.47
0.5 51.24 51.34 55.29 45.77 47.75 48.05
0.6 57.12 57.80 61.22 52.51 55.08 55.11
0.7 65.10 65.96 68.09 58.14 60.79 61.45
0.8 70.56 71.06 71.52 64.23 66.59 67.30
0.9 76.02 76.84 76.83 69.67 71.96 72.51
1.0 79.80 79.22 79.01 73.73 75.28 76.20
1.1 82.74 82.62 82.44 77.70 79.03 79.73
1.2 84.42 85.00 84.31 82.22 83.46 83.58
1.3 87.78 87.38 86.49 85.63 87.04 87.43
1.4 89.88 90.10 88.99 88.67 89.77 89.99
1.5 91.56 91.80 90.86 90.97 91.73 91.83
1.6 92.40 92.48 91.48 92.81 93.60 93.67
1.7 94.50 94.18 93.04 94.84 95.13 95.19
1.8 95.76 95.20 94.29 96.13 96.40 96.71
1.9 96.18 95.54 94.60 97.42 97.76 97.99
2.0 97.02 96.22 95.22 98.52 98.61 98.87
2.1 97.02 96.22 95.22 98.79 97.78 98.95
2.2 97.02 96.22 95.22 99.06 99.12 99.19
2.3 98.28 97.24 96.15 99.24 99.37 99.43
2.4 98.28 97.24 96.15 99.33 99.45 99.51
2.5 98.28 97.24 96.15 99.42 99.53 99.59
2.6 98.70 97.58 96.46 99.42 99.53 99.59
2.7 98.70 97.58 96.46 99.60 99.70 99.75
2.8 98.70 97.58 96.46 99.87 99.87 99.91
2.9 98.70 97.58 96.46 99.87 99.87 99.91
3.0 99.96 99.96 99.89 99.87 99.87 99.91

**Significant at the .05 level.

64

 

 

 

 

Table 4. Continued.
Music Nutrition

Absolute

Difference 01 02 03 01 02 03
0.0 4.49 3.85 5.72 3.50 2.58 3.65
0.1 18.81 18.06 19.70 11.39 11.61 11.57
0.2 30.04 28.18 29.65 15.77 16.77 17.05
0.3 39.30 36.13 36.85 24.54 25.15 25.58
0.4 44.07 39.74 41.72 29.80 34.18 35.33
0.5 48.56 43.59 46.38 33.30 37.40 39.59
0.6 53.61 47.44 49.76 42.07 48.36 52.39
0.7 57.54 52.98 54.63 45.57 54.16 57.87
0.8 60.91 58.28 60.56 49.95 57.38 61.52
0.9 66.80 52.02 68.18 52.58 60.60 65.17
1.0 75.78 73.93 76.44 56.08 65.11 68.82
1.1 84.76 83.56 84.91 63.97 71.56 73.69
1.2 90.09 88.13 89.78 66.60 74.14 76.12
1.3 94.02 92.46 93.59 68.35 76.07 77.94
1.4 94.30 93.90 94.86 70.10 77.36 78.54
1.5 95.42 95.10 96.13 75.36 79.29 79.75
1.6 96.54 96.06 96.55 76.23 79.93 80.35
1.7 96.82 96.30 96.76 76.23 79.93 80.35
1.8 97.66 97.02 97.60 76.23 79.93 80.35
1.9 97.94 97.50 98.02 76.23 79.93 80.35
2.0 98.50 98.22 98.44 76.23 79.93 80.35
2.1 98.78 98.70 98.86 76.23 79.93 80.35
2.2 98.78 98.70 98.86 77.10 80.57 80.95
2.3 98.78 98.70 98.86 77.10 80.57 80.95
2.4 99.06 98.94 99.07 77.10 80.57 80.95
2.5 99.06 98.94 99.07 77.10 80.57 80.95
2.6 99.06 99.18 99.28 77.10 80.57 80.95
2.7 99.06 99.18 99.28 77.10 80.57 80.95
2.8 99.06 99.18 99.28 77.10 80.57 80.95
2.9 99.06 99.18 99.28 77.10 80.57 80.95
3.0 99.90 99.90 99.91 99.90 99.92 99.85

 

65

Table 4. Continued.

 

 

 

 

Nursing Pharmacy
Abs°lute 01 02 03 01 02 03
Difference
0.0 7.31 6.00 6.89 4.00 10.00
0.1 14.62 12.00 15.51 8.00 16.66
0.2 24.37 22.00 24.13 12.00 29.99
0.3 36.56 34.00 36.19 20.00 46.65
0.4 51.19 46.00 48.25 28.00 53.31
0.5 60.94 54.00 55.14 32.00 56.64
0.6 60.94 64.00 58.58 32.00 59.97
0.7 68.25 70.00 65.47 36.00 66.63
0.8 73.12 80.00 70.64 44.00 73.29
0.9 77.99 86.00 79.26 56.00 76.62
1.0 82.86 88.00 84.43 60.00 79.95
1.1 85.29 92.00 89.60 76.00 83.28
1.2 90.16 96.00 93.04 80.00 83.28
1.3 95.03 98.00 96.48 80.00 83.28
1.4 97.46 98.00 98.20 84.00 83.28
1.5 97.46 98.00 98.20 88.00 83.28
1.6 97.46 98.00 98.20 92.00 89.94
1.7 97.46 98.00 98.20 96.00 89.94
1.8 97.46 98.00 98.20 96.00 89.94
1.9 97.46 98.00 98.20 96.00 96.60
2.0 97.46 98.00 98.20 100.00 96.60
2.1 97.46 98.00 98.20 100.00 96.60
2.2 97.46 98.00 98.20 100.00 96.60
2.3 97.46 98.00 98.20 100.00 96.60
2.4 97.46 98.00 98.20 100.00 96.60
2.5 97.46 98.00 98.20 100.00 99.93
2.6 97.46 98.00 98.20 100.00 99.93
2.7 97.46 98.00 98.20 100.00 99.93
2.8 99.89 100.00 99.92 100.00 99.93
2.9 99.89 100.00 99.92 100.00 99.93
3.0 99.89 100.00 99.92 100.00 99.93

 

66

Table 4. Continued.

 

 

 

Philosophy Physics
AbS°1ute 01 02 03 01 02 03
Difference
0.0 4.74 4.58 4.58 2.94 4.29
0.1 11.51 13.09 14.12 12.74 15.01
0.2 18.62 22.47 23.84 21.07 23.59
0.3 24.38 30.33 31.54 34.30 34.31
0.4 29.12 38.19 39.42 41.16 42.89
0.5 35.56 44.74 46.02 50.47 52.76
0.6 41.32 52.81 52.62 58.31 60.05
0.7 46.40 57.39 57.02 63.70 65.62
0.8 53.51 63.94 64.72 67.62 67.76
0.9 59.61 68.96 69.67 70.56 72.05
1.0 64.35 73.10 73.89 75.46 77.20
1.1 69.09 77.03 78.66 78.89 79.77
1.2 73.15 80.30 81.77 80.36 84.06
1.3 75.86 83.35 84.52 82.32 86.20
1.4 79.24 85.53 86.53 84.77 88.77
1.5 83.64 88.58 89.46 86.24 90.05
1.6 84.99 89.67 90.92 89.18 92.62
1.7 87.02 91.19 92.75 90.16 92.62
1.8 88.37 92.06 93.48 90.16 93.04
1.9 90.74 93.58 94.76 91.63 94.32
2.0 92.77 95.10 96.04 93.10 95.17
2.1 94.46 96.84 97.32 96.53 97.74
2.2 95.81 97.71 97.87 97.02 98.16
2.3 96.48 97.92 98.23 98.49 99.01
2.4 97.49 98.57 98.78 98.98 99.01
2.5 98.50 99.22 99.33 98.98 99.01
2.6 98.83 99.43 99.51 99.47 99.43
2.7 99.50 99.86 99.87 99.96 99.85
2.8 99.83 99.86 99.87 99.96 99.85
2.9 99.83 99.86 99.87 99.96 99.85
3.0 99.83 99.86 99.87 99.96 99.85

 

67

 

 

 

 

 

Table 4. Continued.
Political Science Psychology

§b8°lute 01 02 03 01 02 03

Difference
0.0 3.61 4.63 5.74 3.27 4.17 4.15
0.1 12.49 14.47 16.29 12.42 13.78 13.91
0.2 21.37 25.47 27.78 21.01 22.62 '22.81
0.3 28.93 32.99 36.32 29.14 31.46 31.71
0.4 37.48 41.29 44.54 39.51 41.00 41.14
0.5 44.05 50.74 51.52 48.10 48.86 49.18
0.6 48.98 57.49** 58.19** 55.67 56.79 56.76
0.7 54.90 63.08** 64.40** 61.65 63.26 63.29
0.8 58.84 67.32** 69.36** 68.09 69.66 69.55
0.9 64.43 72.72** 74.32** 78.79 74.60 74.49
1.0 69.03 77.93** 78.97** 77.90 78.36 78.11
1.1 72.97 81.01** 82.54** 82.29 82.60 82.66
1.2 77.24 83.90 85.17** 86.58 86.56 86.81
1.3 80.52 86.21 87.65** 89.66 89.62 90.10
1.4 82.49 87.56 89.04** 92.08 92.26 92.60
1.5 85.77 89.87 91.36 93.29 93.65 93.72
1.6 87.41 91.02 92.75 95.62 95.46 .95.43
1.7 89.38 92.56 93.99 96.64 96.78 96.61
1.8 91.35 93.91 95.07 97.66 97.68 97.59
1.9 92.99 95.26 96.31 98.78 98.58 98.51
2.0 94.96 96.61 97.39 99.15 98.99 98.77
2.1 96.27 97.38 98.16 99.43 99.47 99.16
2.2 96.92 97.95 98.47 99.71 99.67 99.48
2.3 97.90 98.52 98.93 99.89 99.67 99.61
2.4 98.55 98.90 99.24 99.89 99.73 99.67
2.5 98.55 98.90 99.24 99.89 99.86 99.86
2.6 98.87 99.28 99.39 99.89 99.86 99.86
2.7 99.85 99.85 99.85 99.89 99.86 99.86
2.8 99.85 99.85 99.85 99.89 99.86 99.86
2.9 99.85 99.85 99.85 99.89 99.86 99.86
3.0 99.85 99.85 99.85 99.89 99.86 99.86

**Significant at the .05 level.

68

 

 

 

 

Table 4. Continued.
Radio-TV Sociology

Absolute

Difference 01 02 03 01 02 03
0.0 5.88 7.40 3.38 5.35 5.39 5.71
0.1 17.64 18.51 11.85 16.80 17.42 17.58
0.2 23.52 27.76 22.01 28.34 29.52 29.58
0.3 26.46 37.01 33.87 38.89 40.99 40.87
0.4 38.22 48.12 44.03 46.80 49.35 50.21
0.5 49.98 57.37 57.58 56.44 58.96 59.42
0.6 55.86 59.22 59.27 64.93 66.49 66.88
0.7 70.56 70.33 71.13 70.12 71.88 72.20
0.8 73.50 77.73 77.90 75.23 76.65 76.67
0.9 82.32 81.43 82.98 79.26 80.3I 80.43
1.0 85.26 83.28 84.67 83.38 84.94 85.03
1.1 85.26 83.28 84.67 86.84 87.77 87.82
1.2 85.26 86.98 88.05 89.23 90.53 90.80
1.3 85.26 90.68 89.74 91.20 92.60 92.87
1.4 91.14 96.23 94.82 92.60 94.12 94.55
1.5 97.02 99.93 98.20 93.91 95.22 95.65
1.6 97.02 99.93 98.20 95.39 96.46 96.81
1.7 97.02 99.93 98.20 96.79 .97.63 97.78
1.8 97.02 99.93 98.20 97.20 97.97 98.16
1.9 97.02 99.93 98.20 98.10 98.66 98.74
2.0 97.02 99.93 98.20 98.84 99.14 99.19
2.1 97.02 99.93 98.20 99.25 99.41 99.51
2.2 99.96 99.93 98.20 99.33 99.41 99.51
2.3 99.96 99.93 98.20 99.41 99.47 99.57
2.4 99.96 99.93 98.20 99.65 99.67 99.76
2.5 99.96 99.93 98.20 99.73 99.73 99.82
2.6 99.96 99.93 98.20 99.81 99.79 99.82
2.7 99.96 99.93 98.20 99.81 99.79 99.82
2.8 99.96 99.93 98.20 99.81 99.79 99.82
2.9 99.96 99.93 98.20 99.89 99.85 99.88
3.0 99.96 99.93 99.89 99.89 99.85 99.88

 

69

 

 

 

 

 

- Table 4. Continued.
Speech Zoology
5b5°lute 01 02 03 01 02 03
Difference
0.0 6.22 5.86 5.47 3.28 4.56 4.53
0.1 20.38 20.17 20.38 9.84 12.32 13.34
0.2 31.32 31.89 33.01 16.40 20.69 22.28
0.3 42.47 42.75 43.81 24.71 29.82 30.29
0.4 52.34 52.57 54.00 32.80 39.10 39.50**
0.5 61.13 60.84 61.15 42.20 46.40 46.84
0.6 67.99 68.25 67.69 48.54 53.24 53.78
0.7 73.99 75.14 73.47 53.79 59.17 58.71
0.8 80.64 81.00 79.86 60.79 64.95 64.58
0.9 84.93 85.82 84.57 65.60 69.51 69.92
1.0 89.22 89.78 89.13 70.85 75.44 75.66
1.1 91.58 92.53 91.41 75.22 79.09 78.86
1.2 94.15 94.59 93.69 79.59 83.04 84.06
1.3 96.08 96.48 95.66 83.52 85.62 86.59
1.4 96.72 97.34 96.42 87.67 89.88 90.19
1.5 96.93 97.51 97.02 91.17 93.07 93.66
1.6 97.57 98.02 97.78 92.92 94.89 95.92
1.7 98.42 98.70 98.23 94.67 95.95 96.85
1.8 98.42 98.87 98.38 96.63 97.31 97.78
1.9 98.63 98.87 98.38 97.06 97.61 97.91
2.0 98.84 99.04 98.68 97.27 97.91 98.17
2.1 98.84 99.04 98.68 98.14 98.67 98.70
2.2 98.84 99.04 98.83 99.01 99.27 99.23
2.3 99.69 99.72 99.59 99.01 99.27 99.23
2.4 99.90 99.89 99.74 99.44 99.57 99.63
2.5 99.90 99.89 99.74 99.65 99.72 99.76
2.6 99.90 99.89 99.74 99.86 99.87 99.89
2.7 99.90 99.89 '99.74 99.86 99.87 99.89
2.8 99.90 99.89 99.74 99.86 99.87 99.89
2.9 99.90 99.89 99.89 99.86 99.87 99.89
3.0 99.90 99.89 99.89 99.86 99.87 99.89
**Significant at the .05 level.

70

freshmen and cumulative freshmen-junior levels.1

The correlations and their Z-transformations be-
tween the Washington Pre—College Testing Program predicted
grades and the three levels of Washington State University
student achieved grades are given in Table 5. There were
no statistically significant differences between the three
levels of achieved grades when correlated with the predicted
grades from the Washington Pre-College Testing Program

except All-University at the cumulative freshmen and the

 

cumulative freshmen-junior levels, Art at the cumulative
freshmen and cumulative freshmen and sophomore levels and
the cumulative freshmen and cumulative freshmen-junior
levels, English Composition at the cumulative freshmen and
sophomore and cumulative freshmen-junior levels, and
Bacteriology at the cumulative freshmen and cumulative
freshmen and sophomore levels and the cumulative freshmen
and cumulative freshmen-junior levels.

The prediction most often used for student assess-

ment, the All—University prediction, indicated very similar

 

absolute differences at the three levels of achievement.
Only a slight difference in percentages-—79.21 — 78.80 -
79.43—-was recorded for the three levels of achievement at
the .6 grade-point level of absolute difference. The corre—
lation of the predicted grade with the cumulative freshman

level of achieved grade was significantly higher than with

 

1Appendix B.

71

 

 

 

Table 5. Correlations, Z—transformations of correlations,
means, standard deviations, between W.P.C.T.P.
predicted and W.S.U. achieved grade-point averages
at the cumulative freshmen level for 1961 W.S.U.
freshmen.

Criterion Z- Mean Mean S.D. S.D.
N r values Ach. Pred. Ach. Pred.

All—University 2,274 .6682 .807** 2.16 2.26 .688 .454

Accounting 27 .2372 .242 2.41 2.50 .740 .332

Anthropology 544 .5866 .672 2.14 2.49 .825 .431

Architecture 55 .0910 .091 2.68 2.49 .889 .242

Art 367 .4317 .461* 2.23 2.59 .707 .197

Bacteriology 150 .5030 .555** 1.88 2.22 .802 .377

Biology 277 .6193 .723 2.11 2.02 .860 .703

Botany 189 .4780 .520 2.04 2.14 .920 .780

Business Admin. 98 .3813 .400 2.24 1.97 .854 .414

Chemistry 819 .5634 .637 2.18 2.14 .863 .521

Economics 32 .6790 .827 2.61 2.80 1.000 .370

Drama 48 .3523 .368 2.72 2.71 .868 .454

Education 302 .5188 .575 2.38 2.57 .680 .301

Engineering 261 .4532 .490 2.51 2.22 .914 .235

English

Composition
English
Literature 445 .4460 .470 2.47 2.33 .728 .359

Forestry 41 .1950 .196 2.31 2.20 .927 .260

Geography 312 .4620 .501 2.14 2.07 .769 .395

Geology 288 .4127 .438 2.09 2.27 .778 .338

History 513 .4708 .511 2.19 2.31 .769 .342

Journalism 44 .3055 .316 2.79 2.89 .853 .311

Home Economics 230 .4790 .522 2.30 2.68 .769 .301

Languages 217 .4746 .517 2.56 2.72 .893 .510

Mathematics 1,018 .4426 .476 2.23 2.17 .891 .565

Music 352 .0693 .069 3.07 2.98 .780 .189

Nursing 40 .6634 .801 2.67 2.68 .839 .277

Nutrition 87 .4840 .530 2.85 2.64 .897 .492

Pharmacy

Philosophy 261 .4840 .530 1.99 2.44 .881 .401

Physics

Political

Science 271 .4748 .516 2.11 2.41 .869 .454

Psychology 1,033 .5802 .663 2.37 2.11 .922 .543

Radio & TV 34 .3605 .378 3.11 2.72 .731 .428

Sociology 1,165 .6066 .694 2.21 2.31 .860 .524

Speech 460 .5094 .652 2.50 2.47 .732 .403

Zoology 416 .4942 .538 2.19 2.15 .943 .459

 

*Significant at the .01 level.
**Significant at the .05 level.

72

 

 

 

 

 

Table 5. Continued. Correlations, Z-transformations of
correlations, means, standard deviations, between
W.P.C.T.P. predicted and W.S.U. achieved grade-
point averages at the cumulative freshmen and
sophomore level for 1961 W.S.U. freshmen.

Criterion Z- Mean Mean S.D. S.D.
Area N r values Ach. Pred. Ach. Pred.

All-University 2,275 .6625 .795 2.14 2.26 .676 .454

Accounting 184 .4271 .456 2.06 2.33 .798 .386

Anthropology 736 .5806 .671 2.21 2.51 .856 .449

Architecture 66 .0708 .071 2.69 2.43 .813 .250

Art 525 .3294 .343* 2.31 2.60 .698 .205

Bacteriology 358 .4562 .492** 2.12 2.27 .859 .388

Biology 436 .5395 .604 2.16 2.09 .889 .729

Botany 307 .4643 .504 2.19 2.22 .942 .733

Business Admin. 248 .3459 .361 2.22 2.20 .768 .450

Chemistry 926 .5381 .601 2.13 2.11 .832 .531

Economics 534 .4964 .545 2.29 2.29 .825 .392

Drama 102 .2613 .268 2.81 2.70 .663 .438

Education 544 .4922 .539 2.41 2.55 .737 .312

Engineering 326 .4125 .438 2.35 2.22 .847 .232

English

Composition 231 .6316 .742* 2.69 2.42 .718 .419

English

Literature 734 .4209 .448 2.48 2.33 .696 .358

Forestry 54 .2405 .245 2.35 2.20 .755 .252

Geography 482 .5197 .576 2.29 2.11 .780 .410

Geology 469 .3698 .388 2.16 2.27 .777 .343

History 801 .5194 .576 2.23 2.36 .759 .372

Journalism 70 .2279 .232 2.73 2.84 .783 .305

Home Economics 282 .4888 .534 2.33 2.68 .762 .306

Languages 286 .4973 .540 2.49 2.73 .863 .512

Mathematics 111 .4561 .492 2.20 2.17 .884 .558

Music 410 .0194 .019 3.10 2.97 .800 .192

Nursing 49 .6021 .693 2.70 2.70 .829 .277

Nutrition 124 .5126 .554 2.67 2.66 .851 .488

Pharmacy 25 .5099 .550 3.12 2.61 .600 .282

Philosophy 423 .4593 .495 2.11 2.44 .830 .393

Physics 182 .3648 .386 2.30 2.29 .776 .289

Political

Science 480 .5126 .567 2.09 2.39 .792 .449

Psychology 1,389 .5529 .622 2.37 2.13 .900 .542

Radio & TV 54 .3806 .406 3.01 2.71 .700 .389

Sociology 1,400 .6010 .690 2.24 2.33 .827 .530

Speech 574 .4697 .510 2.51 2.49 .701 .393

Zoology 609 .4885 .533 2.17 2.15 .892 .465

*Significant at the .01 level.
**Significant at the .05 level.

73

 

 

 

 

 

Table 5. Continued. Correlations, Z-transformations of
correlations, means, standard deviations, between
W.P.C.T.P. predicted and W.S.U. achieved grade-
point averages at the cumulative freshman--junior
level for 1961 W.S.U. freshmen.

Criterion Z- Mean Mean S.D. S.D.
Area N r values Ach. Pred. Ach. Pred.

All-University 2,293 .6549 .744** 2.16 2.26 .678 .454

Accounting 292 .4521 .488 2.00 2.32 .709 .394

Anthropology 862 .5856 .672 2.23 2.50 .851 .452

Architecture 74 .3663 .382 2.72 2.35 .801 .395

Art 608 .3560 .272 2.34 2.60 .713 .207

Bacteriology 499 .4590 .496 2.15 2.27 .824 .413

Biology 543 .5644 .639 2.19 2.09 .903 .735

Botany 402 .4335 .464 2.23 2.21 .970 .730

Business Admin. 365 .4128 .440 2.20 2.25 .714 .455

Chemistry 1,042 .5309 .591 2.12 2.10 .817 .535

Economics 742 .4920 .539 2.24 2.28 .804 .388

Drama 133 .3197 .331 2.83 2.66 .632 .484

Education 647 .5058 .550 2.48 2.53 .727 .312

Engineering 427 .3780 .397 2.34 2.22 .819 .235

English

Composition 391 .3929 .420* 2.68 2.39 .707 .449

English

Literature 879 .4556 .491 2.49 2.33 .676 .364

Forestry 64 .2178 .221 2.38 2.19 .734 .264

Geography 572 .5096 .563 2.31 2.11 .765 .418

Geology 589 .3897 .410 2.17 2.27 .775 .350

History 961 .5089 .561 2.23 2.33 .746 .347

Journalism 101 .2373 .241 2.75 2.83 .777 .302

HOme Economics 320 .4920 .539 2.37 2.69 .760 .307

Languages 344 .5090 .550 2.37 2.69 .864 .526

Mathematics 1,246 .4599 .497 2.21 2.17 .883 .554

Music 472 .0524 .052 3.14 2.97 .771 .188

Nursing 58 .5891 .671 2.78 2.71 .809 .272

Nutrition 164 .5279 .586 2.66 2.65 .809 .471

Pharmacy 30 -.l607 .161 2.78 2.26 .638 .338

Philosophy 545 .4489 .483 2.17 2.43 .849 .392

Physics 233 .3431 .357 2.28 2.27 .773 .292

Political

Science 644 .4968 .545 2.08 2.38 .764 .448

Psychology 1,516 .5418 .605 2.37 2.14 .893 .940

Radio & TV 59 .4537 .490 3.01 2.69 .698 .404

Sociology 1,541 .5953 .688 2.26 2.35 .818 .532

Speech 657 .4623 .501 2.53 2.48 .719 .396

Zoology 749 .4674 .507 2.14 2.13 .880 .467

*Significant at the .01 level.
**Significant at the .05 level.

74

 

 

 

 

 

Table 5. Continued. Significance of differences (expressed
as £_values) of correlations between W.P.C.T.P.
predicted and W.S.U. achieved grade-point averages
at three levels for 1961 freshmen at Washington
State University.

Ol-Freshmen Ol-Freshmen 02-Sophomores
02-Sophomores 03-Seniors 03-Seniors

All-University ' .40 2.17** 1.70

Accounting .87 .35 1.17

Anthropology .01 .01 .01

Architecture .11 1.55 1.71

Art 6.40* 2.82* 1.22

Bacteriology 2.00** l.99** .01

Biology 1.58 1.12 .51

Botany .55 .68 .52

Business Admin. .10 .10 .96

Chemistry .75 .99 .21

Economics 1.48 1.52 .10

Drama .55 .21 .12

Education .50 .35 .02

Engineering .63 1.15 .53

English Composition .01 3.98*

English Literature '.40 .38 .82

Forestry .23 .12 .13

Geography 1.05 .89 .19

Geology .67 .39 .32

History 1.20 .94 .33

Journalism .44 .42 .50

Home Economics .13 .19 .06

Languages .25 .36 .01

Mathematics .34 .48 .01

Music .70 .24 .48

NUrsing .49 .60 .12

NUtrition .22 .52 .31

Pharmacy .01 1.44

Philosophy .46 .68 .01

Physics .01 .29

Political Science .70 .41 .36

Psychology 1.00 1.41 .45

Radio & TV .12 .50 .56

Sociology .01 .01 .01

Speech .83 1.01 .16

Zoology .01 .57 .50

*Significant at the .01 level.
**Significant at the .05 level.

75

the cumulative freshman-junior level of achieved grade
(I3 < .05).

Accounting cannot be considered a freshman course
because the majority of students taking the course are
beyond their freshman year. The small number of students
concerned invalidates the use of the prediction at the
freshman level. The predicted grade had a higher correlation
at the cumulative freshman-junior level than the cumulative
freshman and sophomore level, although the difference is
not statistically significant.

The correlation for Architecture at the cumulative
freshman and cumulative freshman and sophomore levels is so
low that the use of the predicted grade has little meaning
for predicted purposes. At the cumulative freshman-junior
level the correlation is .39, a substantial increase over
the correlation at the cumulative freshman and cumulative
freshman and sophomore levels, but there is little increase
in the number of students. Apparently for the same students
the achieved grades received at the junior or senior level
correlate better with the predicted grades than do the
achieved grades at the cumulative freshman and cumulative
freshman and sophomore levels.

Art showed a significantly higher correlation at
the cumulative freshman level than at either the cumulative
freshman and sophomore or cumulative freshman-junior levels
(p < .01). The freshman grade in art may be more predictable

than the grades in later art courses which depend more on

76

skills not measured by the predictors in the Washington Pre-
College Testing Program. The absolute differences between
the predicted and the three levels of achieved grades in
Table 4 for art reflect a similarity in percentages between
the three levels of achieved grades.

Though the differences are not statistically signifi-
cant, Biology, Botany, Chemistry, Engineering, Geology,
Nursing, Philosophy, Psychology, Speech and Zoology showed
higher correlations at the cumulative freshman level than
the cumulative freshman and sophomore or cumulative freshman—
junior levels. In Table 4 for these criterion areas, the
cumulative percentages of absolute differences varies among
the three levels of achieved grades, a highzrficumulative
percentage appearing more often at the cumulative freshman-
junior level than at the cumulative freshman or cumulative

freshman and sophomore levels.

For Business Administration, Geography, History,

 

Languages, Nutrition and Radio and TV, the correlation between

 

predicted and achieved grades showed a trend toward being
higher at the cumulative freshman-junior level than the
cumulative freshman level. The trend was possibly a result
of the restriction of the number of students in the summari-
zation at the freshman or cumulative freshman level for

Radio and TV only. Economics and Drama showed a higher

 

 

correlation at the cumulative freshman level but again the
number of students at that level is small, influencing the

magnitude of the correlation.

77

In Table 5 for Anthrppology, Education, English
Literature, Forestry, Home Economics, Mathematics, Physics,

Political Science and Sociology the correlations of the pre-

 

dicted grades with the three levels of achieved grades showed
little variation from one level to another.

Bacteriology had a significantly higher correlation
at the cumulative freshman level than at the cumulative
freshman and sophomore level or cumulative freshman—junior
levels (p < .05). The large lecture class provided grading
patterns more consistent with the predicted grades than did
the smaller laboratory orientated classes at the cumulative
freshman-junior levels. English Composition had a signi-
cantly higher correlation at the cumulative freshman and
sophomore level than at the cumulative freshman-junior level
(p < .05). No freshmen students take course work in this
area. Music showed a higher correlation at the cumulative
freshman than the cumulative freshman and sophomore or
cumulative freshmen-junior levels but the correlation at any
level is low, precluding its use for any comparative purpose.
Pharmacy, with no students taking course work at the freshman
level, showed a higher correlation at the cumulative
freshman and sophomore than the cumulative freshman—
junior level, the latter correlation being the only neg-
ative value registered in this study. The addition of
grades during the junior and senior year did not
correlate with the grade predictions as did the grades

received at the cumulative freshman and sophomore

78

level. For the four criterion areas discussed in this para—
graph the reduction in correlation between the achieved
grades and predicted grades may be a function of the re-
stricted grades received by students during the junior and
senior years.

The currently used predicted grades of the Washington
Pre-College correlated equally well with achieved grades
summarized at three levels, cumulative freshman, cumulative
freshman and sophomore, cumulative freshman-junior except
for the areas of All-University, Art, Bacteriology and
English Composition. The latter four correlated higher with

achieved grades at the cumulative freshman level.

3. Is there a hierarchy in the subject matter areas
as represented by predicted and achieved grades from the
Washington Pre-College Testing Program and what similarity
does the hierarchy have with the predicted and achieved
grades for students at Washington State University?

Two student classes, the 1961 Washington State
University freshmen and the University of Washington fresh-
men of 1955-1956, were used to determine whether (1) the
Washington Pre—College Testing Program grade predictions
for the two groups were similar in their ranking, (2) whether
the ranking of the achieved grades was similar, and (3) for
Washington State University, whether the ranking of achieved
with predicted grades was similar. A SHARE 966 Program for
the IBM 709 computer was used to develop means and standard
deviations for the predicted and achieved grades for the
1961 Washington State University freshmen, as given in

Table 6. Data for the University of Washington group were

79

om. Nb. Um>wflco< mm.m om.m Dm>mflnom

 

 

 

 

 

 

 

m4. m4. cmuoflomum mm.m m~.~ amuoflomum mm. H4. momm mom m .cH566 .msm
mo.s so. 6m>mhno¢ mm.a mm.m 6m>mflno<

as. ms. nmuoflcmum mo.m Hm.m amuogwwum mm. ms. mum Nos m samuom
mo.s om. nm>mflau< 4H.m ms.m 6m>oflno<

on. «A. 6¢p0snmum AH.~ mo.m nmuoflomum 66. mm. Nam mwm s smoHon
«a. mm. 6m>mflau¢ sv.m ms.m mm>mflao<

os. 44. 6¢p0fltmum om.m nm.m mmuugcmum mm. 64. mum mas o smoHoHumuomm
me. up. wm>mflao¢ mm.m 4m.m 6m>mfiao<

om. ms. ambusomum m6.m oo.m amuoflomum 6m. 6m. moss moo m “.4
mm. as. nw>mflao< wv.m ms.m 6m>wflao<

mm. mm. cmpoflumum mm.m mm.m omuoflnmnm 6m. 6m. vms vs 4 musuomuflcoua
Hm. mm. nm>msno< 4m.m mm.m 6m>wflao<

m6. m4. umpoflomum sv.m om.m wmuoflnmum Hm. mm. mmmm New m smosomouauca
so. as. 6m>mflao< HH.~ oo.m om>mflno<

m4. ow. embosvmum Hm.~ mm.m omuoflnmum m4. m4. mmoa mom m mcflucsooom
ms. m6. 6m>mfinum Hm.m oH.N 6m>msao¢

64. sq. amuosomum mm.m sm.m mmuoflvmum so. 66. ssmm mmmm H suflmum>flcouss<
2: am: 3: pm: as pm; 3: 5m: mmu< mmusoo
mQOHDMH>mQ pumpcmum com: M Z

.Homs

smficmwum mcflumucm xuflmnm>HCD mumym cowmcanmmz tam .ommalmmma smegmmuw @cflumusm
CODmCHnmmB mo >uamum>ficb wnu How mmpmum ©m>mH£om paw Umuuflpmum .m.B.0.m.3 cmm3um£
mcoauma>m© pumpcmum pcm .mcmmE .mustUHmMQOU coaumamuuoo Hothououmm mo somHHMQEOU .o magma

80

m5.
mm.

mm.
wm.

Hm.
mm.

mm.
«v.

an.
wm.

mm.
mg.

om.
ow.

mm.
vm.

mm.
mm.

mm.
mm.

mm.
0%.

>0.
mm.

on.
om.

fin.
gm.

mm.
mm.

on.
Hfi.

mm.
mm.

mo.
mm.

on.
mq.

Hm.
mm.

m5.
mm.

mm.
mfl.

om.
mm.

mm.
gm.

Um>mflnufl
cwuoflpmum

6m>mflaoa
pmuoflpmnm

©m>mflno¢
Umuoflcwnm

6m>mhno<
pmuoflpmum

©w>mflno<
pmuoflpmum

pm>wfl£o¢
Umuoflpmum

©m>mH£U¢
Umuoflpmum

©m>mflsofl
pwuoflpmum

om>mgao<
pwuoflpmnm

6m>mflno<
tmuoflpmum

pm>ma£om
pmuoflpmum

Um>mH£U<
pmpoflomum

oh.m
®F.N
om.N
wv.m
ON.N
om.N
NN.N
hN.N

mh.N
mm.N

mm.m
0H.N
m¢.m
mm.m
mo.m
mm.m
mm.N
NN.N
m¢.N
mm.m
mm.m
©®.N
¢N.N
mN.N
NH.N
OH.N

6m>mflaom
omuoflcmum

©m>wflnom
Uwuowpmum

pw>mflno<
pwuoflpwnm

pw>wH£o¢
Umuoflpmum

pm>mflso<
meOHUme

©m>mfl£o<
nmuoflpmum

6m>mflno<
pmuofinmum

6m>mflzo<
Umpoflpmum

Um>mflno¢
pmuowpmum

6m>mflnum
emuofluoum

pw>mH£o¢
pmuoflpmum

om>mago¢
pmuoflpmum

Hm.

mw.

afl.

mv.

Nw.

am.

we.

Nv.

0v.

mg.

mg.

mm.

mm.

Hm.

mm.

Hm.

mm.

0%.

mm.

mm.

Hm.

Nm.

mg.

mm.

mflm

ovoa

wmoa

Howa

N©H

mmma

mmww

mmma

mom

womw

Homm

momm

HOH

Hem

mmm

th

we

mom

Hmm

hmw

bvo

MMH

Nwh

Nvoa

Hm

ON

ma

ma

ha

0H

ma

wa

ma

NH

Ha

OH

EmHHMCHSOh

>uoumam

wmoHomo

hammumomo

wuummuom

.psq gmflamcm

.QEOU .mcm

massmmcamcm

coHumospm

manna

mUflEOCOUm

>HumHEo£U

81

 

 

 

 

 

 

 

 

 

 

 

 

mm. on. pm>mH£U¢ om.m mo.m Um>mH£U¢

og. vv. meUprum ov.m mm.m Umuoflcmum mg. om. omh wwo Hm .Hom HMUHDHHom
hm. hm. Um>mH£U¢ 0H.N mm.N ©m>mH£U¢

am. am. UwUUHUmHm ©N.N hm.m meUHpmHm av. gm. omma mmm om moflmmnm
Hm. mm. ©m>mH£U¢ mv.m hH.N Um>mH£o¢

Nv. mm. pmuoflpwnm om.m mv.m Umuoflpmum mg. mg. mam mwm mm >£Q0moaflnm
mm. mm. Um>mflno< om.H wh.m U0>®H£U¢

ow. mm. pwuoflpmum ¢O.N ©N.N pmuUflpmum gm. ©H.I OOH om mm momfiumzm
mu. om. Um>mflno< ¢©.N ®©.N Um>mflno<

vs. ow. pwuoflpmum ®©.N mo.m Umuoflpwum mm. mm. mma flea hm COHDHHDDZ
mo. Hm. ©w>mH£U¢ mh.m mh.m Um>mH£U¢

mm. mm. pmuoflpmum hh.m HB.N Umuoflpmum mm. mm. mam mm mm mcflmnsz
mo. on. Um>®H£U¢ mm.m vH.m ©m>mflno¢

ma. ma. UmuUHUmnm oo.m hm.m UwDUHpmhm hm. mo. mam va mm Dams:
HO.H mm. Um>mflno¢ mm.a HN.N Uw>mH£U¢

mm. mm. pmuoflpmum HH.N ha.m Umuoflpmum mm. ow. 55mm owma gm npmz
NH.H vm. ©m>mH£U< Hm.m hm.m Um>®flnod

00. mm. pwuoflpmum mm.m m©.m pmuUHUmHm mm. om. omh va mm mmmmsmcmq
vb. on. Um>®H£U< Hm.m hm.m Um>mH£U<

mm. mm. pmuoapmnm mh.m m©.N Umuoflpmum ow. mg. mvb omm mm .coom wEom
3: DmB 25 D93 25 DmB SD 395 omu< mmHDOU
mCOHpmH>®Q pumpcmcm com:

.UQDCHDCOU .o manmfi

82

 

oo.H mm.

6m>mflgo¢
cmuoflcmum

6m>mflno<
twuofltmum

6m>mflao<
Umuoflvmum

Um>manofi
pmuoflpwum

6m>manom
pmuuapmum

ho.N
wH.N
m®.N
N@.N
mN.N
«m.N
b©.N
mm.N
mN.N
¢N.N

fiH.N
ma.m
mm.N
m¢.N
0N.N
mm.N
Ho.m
m©.N
mm.N
MH.N

©m>mflno¢
pmuoflpmnm

pw>mflnufl
chUflpmHm

pm>ma£o¢
Umuoflpwum

©m>mH£U<
Uwuoflpmum

Um>mflnom
pouuanmum

meHOON

>moHOHUOm

>9 A Oawmm

>00a0£0>mm

83

taken from a published report.1 Table 7 shows the ranking
of predicted grades, the achieved grades, and the predicted-
achieved grades. A rank-correlation coefficient of .89 was
found for the predicted grades.

The rank-correlation coefficient of .89 showed a
definite hierarchy for the means of the predicted grades
which was common to both student groups. Two criterion

areas, Economics with means of 2.28 and 2.67 and Romance

 

Languages with means of 2.69 and 2.36, accounted for much
of the variation in the ranking of the predicted grades.
The rank-correlation coefficient for the achieved
grades was .57, a greater variation in internal ranking
than the predicted grades.
The lower correlation was primarily a result of

differences in ranking for seven criterion areas, Bacteriology,

 

Botany, English Composition, Home Economics, Pharmagy,

 

Philosophy, and Political Science. There were no trends or

 

 

direction to show whether one group had mean grades higher
or lower than the other group.

The rank-correlation coefficient for the means of
the predicted grades and the means of the achieved grades
for the Washington State University sample was .57. The

criterion areas of Accounting, Anthrgpology, Forestry,

 

Geography, Pharmacy, Philosophy, Political Science, and

 

 

l“Validity Coefficients for 1955-1956 Weights."

Duplicated Report, University of Washington Division of
Counseling and Testing Services, July, 1959.

84

Table 7. Rank-order correlations of predicted grades for
the 1961 Washington State University freshmen, and
University of Washington 1955-1956 freshmen.

 

 

 

 

 

 

 

 

Mean Rank

Criterion Area W.S.U. U.W. W.S.U. U.W. Diff. (Diff)2
All-University 2.27 2.28 14 11 3 9.00
Accounting 2.31 2.31 17 14.5 2.5 6.25
Anthropology 2.50 2.44 26 23.5 2.5 6.25
Architecture 2.35 2.33 20.5 17 3.5 12.25
Art 2.60 2.65 28 27 1 1.00
Bacteriology 2.27 2.30 14 12.5 1.5 2.25
Biology 2.09 2.17 1 6 5 25.00
Botany 2.21 2.02 8 1 7 49.00
Business Admin. 2.25 2.33 10 17 7 49.00
Chemistry 2.10 2.09 2 3 1 1.00
*Economics 2.28 2.67 16 29 13 269.00
Drama 2.66 2.73 30 30.5 .5 .25
Education 2.53 2.73 27 30.5 2.5 6.25
Engineering 2.22 2.31 9 14.5 5.5 30.25
English Comp. 2.39 2.33 23 17 6 36.00
English Lit. 2.33 2.40 18.5 21.5 3 9.00

Forestry 2.19 2.21 7 7 0 0
Geography 2.11 2.27 3 10 7 49.00
Geology 2.27 2.30 13 12.5 1.5 2.25
History 2.33 2.44 18.5 23.5 5 25.00
Journalism 2.83 2.76 35 32 3 9.00
HOme Economics 2.69 2.78 32 34 2 4.00
*Languages 2.69 2.36 32 20 12 244.00
Mathematics 2.17 2.11 6 4 2 4.00

Music 2.97 3.00 36 36 0 0
Nursing 2.71 2.77 34 33 1 1.00
Nutrition 2.65 2.66 29 28 l 1.00
Pharmacy 2.26 2.04 11 2 9 81.00
Philosophy 2.43 2.50 24 25 l 1.00
Physics 2.27 2.26 13 9 4 4.00
Political Sci. 2.38 2.40 22 21.5 .5 .25
Psychology 2.13 2.24 4.5 8 3.5 12.25
Radio & TV 2.69 2.83 32 35 3 9.00
Sociology 2.35 2.34 20.5 19 1.5 2.25
Speech 2.48 2.62 25 26 l 1.00
Zoology 2.13 2.14 4.5 5 .5 .25
962.00

r1 = 1 - 6 (962.00) = r1 = 1 _ 962.00
36 (1295) 6 (1295)
r1 = 1 - 3§2”OO = r1 = 1 — .12 = .888

Table 7.

Continued.

 

85

Rank-order correlations of achieved

grades for the 1961 Washington State University

freshmen,

freshmen.

and University of Washington 1955-1956

 

 

 

 

 

 

Mean Rank

Criterion Area W.S.U. U.W. W.S.U. U.W. Diff. (Diff)2
All—University 2.16 2.21 6 12 6 36.00
Accounting 2.00 2.11 1 6 5 30.00
Anthropology 2.23 2.24 13 15 2 4.00
Architecture 2.72 2.44 30 24 6 36.00
Art 2.34 2.55 19 26 7 49.00
*Bacteriology 2.15 2.47 5 25 20 400.00
Biology 2.19 2.14 9 7.5 1.5 2.25
*Botany 2.23 1.93 13 1 12 144.00
Business Admin. 2.20 2.33 10 20 10 100.00
Chemistry' 2.12 2.00 3 4 1 1.00
Economics 2.24 2.14 15 7.5 7.5 56.25
Drama 2.73 2.81 31 34.5 3.5 12.25
Education 2.48 2.57 25 27 2 4.00

Engineering 2.35 2.33 20 20 0 0
*English Comp. 2.68 2.19 29 10 19 361.00
English Lit. 2.49 2.32 26 19 7 49.00
Forestry 2.38 2.23 23 14 9 81.00
Geography 2.31 2.22 18 13 5 25.00
Geology 2.17 2.20 7.5 11 3.5 12.25
History 2.23 2.36 13 21.5 8.5 72.25

Journalism 2.75 2.76 32 32 0 0
*Hbme Economics 2.37 2.81 21.5 33.5 12 144.00
Languages 2.37 2.31 21.5 18 3.5 12.25
Mathematics 2.21 1.98 11 3 8 64.00
Music 3.14 2.99 36 35 1 1.00
Nursing 2.78 2.75 33.5 31 2.5 6.25
Nutrition 2.66 2.64 28 29 1 1.00
*Pharmacy 2.78 1.96 33.5 2 31.5 992.25
*Philosophy 2.17 2.43 7.5 23 15.5 240.25
Physics 2.28 2.16 17 9 8 64.00
*Political Sci. 2.08 2.36 2 21.5 19.5 380.25
Psychology 2.39 2.25 24 16 8 64.00
Radio & TV 3.01 2.67 35 30 5 25.00
Sociology 2.26 2.28 16 17 l 1.00
Speech 2.53 2.63 27 28 l 1.00
Zoology 2.14 2.07 4 5 1 1.00
3,373.90

r1 = 1 — 6 (3373.90) r1 = 1 — 3373.90
36 (1295) 7770.00
r1 = 1 ~ .43 r1 .57

 

Table 7.

Continued.

 

86

Rank-order correlations of achieved

and predicted grades for the 1961 Washington
State University freshmen.

 

 

 

 

 

 

 

Mean Rank

Criterion Area Pred. Ach. Pred. Ach. Diff. (Diff)2
All-University 2.27 2.16 14 6 8 64.00
*Accounting 2.31 2.00 17 1 16 256.00
*Anthropology 2.50 2.23 26 13 13 169.00
Architecture 2.35 2.72 20.5 30 9.5 90.25
Art 2.60 2.34 28 19 9 81.00
Bacteriology 2.27 2.15 14 5 9 81.00
Biology 2.09 2.19 1 9 8 64.00
Botany 2.21 2.23 8 13 5 25.00

Business Admin. 2.25 2.20 10 10 0 0
Chemistry 2.10 2.12 2 3 1 1.00
Economics 2.28 2.24 16 15 1 1.00
Drama 2.66 2.83 30 34 4 16.00
Education 2.53 2.48 27 25 2 4.00
Engineering 2.22 2.35 9 21.5 8.5 72.25
English Comp. 2.39 2.68 23 29 6 36.00
English Lit. 2.33 2.49 18.5 26 7.5 56.25
*Forestry 2.19 2.38 7 23 16 256.00
*Geography 2.11 2.31 3 18 15 225.00
Geology 2.27 2.17 13 7.5 5.5 32.25
History 2.33 2.23 18.5 13 5.5 32.25
Journalism 2.83 2.75 38 31 7 49.00
HOme Economics 2.69 2.37 32 21.5 10.5 110.25
Languages 2.69 2.37 32 21.5 10.5 110.25
.Mathematics 2.17 2.21 6 ll 5 25.00

Music 2.97 3.14 36 36 0 0
Nursing 2.71 2.78 34 32.5 1.5 2.25
Nutrition 2.65 2.66 29 28 1 1.00
*Pharmacy 2.26 2.78 11 32.5 21.5 462.25
*Philosophy 2.43 2.17 24 7.5 16.5 272.25
Physics 2.27 2.28 13 17 4 16.00
*Political Sci. 2.38 2.08 22 2 20 400.00
*Psychology 2.13 2.39 4.5 24 19.5 380.25
Radio & TV 2.69 3.01 32 35 3 9.00
Sociology 2.35 2.26 20.5 16 3.5 12.25
Speech 2.48 2.53 25 27 2 4.00
Zoology 2.13 2.14 4.5 4 .5 .25
3,416.55

r1 _ 1 - 6 (3416-55) 1 _ 1 - 3416.55
36 (1295) — 7770.00
r1 = l - .43 = r1 = .57

87

Psychology showed the greatest variation in their rankings.
Again there were no trends or direction as to whether the
predicted or achieved grades for one group were higher or

lower than for the other group.

4. Are the present predictor equations used in the
Washington Pre-College Testing Program valid for Washington
State University students or should there be developed pre-
dictor equations based on the grades achieved by students
at Washington State University?

This, the principal part of the present study, was
designed to determine the validity of the predicted grades
of the Washington Pre-College Testing Program for students
at Washington State University. A comparison was made of
the multiple regression prediction equations used in the
Washington Pre-College Testing Program with the multiple
regression prediction equations developed by using achieved
grades at Washington State University as the criterion
variables. Table A in the Appendix presents the symmetric
correlation matrix for the 18 predictor variables and a
non—symmetric correlation matrix for the 18 predictors and
36 criterion areas of achieved grade summaries for the 1958-
1960 freshmen at Washington State University. A multiple
regression program was used, using the Washington State
University matrices as input, following the Horst iteration
of a single criterion. Table 8 is the iteration order of
selection of predictors in each of the 36 criterion areas
and the cumulative squared multiple correlations at each
successive iteration. For each criterion area the predictors

are shown as selected and at what iteration step. The new

88

 

 

 

 

Table 8. Order of selection of predictor and cumulative
squared adjusted multiple correlations for each
of the differential predictor measures as given
in the "Iterative Predictor Selection Program"
for each of thirty—six criteria of academic
success at Washington State University.

Criterion -- A11 University

Order of Cumulative Squared

Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 1 .543800 .54350964
2 2 12 .200524 .57907788
3 3 6 .083881 .58487399
4 5 5 .119242 .60719971
5 7 2 .083719 .61795592
6 9 18 .068412 .62369416
7 11 4 .047127 .62646857
8 13 9 .040984 .62927384
9 14 10 -.028985 .62970377
10 15 8 -.033965 .63022728
11 16 7 .031978 .63080583
12 l9 14 .029596 .63321219
13 20 3 .020933 .63332590
14 33 11 .016508 .63736877

89

Table 8. Continued.

 

 

 

 

Criterion —- Accounting
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
1 l 2 .400375 .39752322
2 2 11 .162081 .42679004
3 4 5 .086437 .44054891
4 5 6 -.067288 .44323643
5 6 15 .069757 .44630671
6 7 .10 -.065393 .44846324
7 8 i 13 .070016 .45153063
8 10 3 .050717 .45367554
9 11 8 .044939 ‘ .45292850
10 13 9 -.048677 .45669355

ll l4 17 .051443 .45716441

90

Table 8. Continued.

 

 

 

 

 

Criterion -- Anthropology

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 l 5 .406312 .40529174

2 ,2 12 .225424 .46297771

3 3 15 -.128469 .47970925

4 4 2 .073366 .48451932

5 j 6 3 .090708 .50209240

6 8 7 .074880 .51035780

7 9 5 .044552 .51159144

8 11 17 .041423 .51596111

9 15 8 -.022223 .52022403

10 18 11 " -.018880 .52062798

ll l9 16 .021375 .52028985

91

Table 8. Continued.

 

 

 

 

 

Criterion -- Architectural Engineering

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 11 .404629 .38936404

2 2 14 -.350277 .51519942

3 3 6 .238554 .56028817

4 4 7 -.242096 .60472323

5 5 4 .206013 .63374901

6 6 10 -.119005 .63729180

7 7 9 .107577 .63977052

8 8 15 -.110158 .64281743

9 9 13 -.093038 .64284153

10 11 5 -.093800 .64762837

11 13 3 —.O79563 .65054823

92

Table 8. Continued.

 

 

 

 

 

Criterion -- Art

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 l 5 .340438 .33844057

2 2 16 .175038 .37936026

3 3 15 -.097642 .39009634

4 4 13 .066458 .39409734

5 6 4 .081417 .40638774

6 7 14 .054940 .40854255

7 9 6 .060668 .41467071

8 10 11 -.036478 .41474707

9 12 10 | —.036184 .41576301

10 13 7 .031140 .41538573

11 14 18 -.032605 .41512136

12 15 3 .028449 .41454426

13 17 1 .028589 .41521761

14 18 9 —.029625 .41472313

15 23 12 .016272 .41597682

16 30 8 .015311 .41369939

17 31 2 —.015195 .41220487

18 63 17 .003643 .41309649

93

Table 8. Continued.

 

 

 

 

 

Criterion -- Bacteriology

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 5 .421030 .41959044

2 2 2 .122890 .43587729

3 3 15 -.085390 .44285209

4 4 7 .078124 .44840823

5 6 3 .091800 .46588144

6 8 .4 .053995 .47140933

7 9 14 -.043390 .47220236

8 10 13 -.045444 .47319149

9 11 17 .032099 .47304209

10 12 12 -.040723 .47359447

ll l3 16 .033408 .47356617

94

Table 8. Continued.

 

 

 

 

 

Criterion -- Biology

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection NUmber Selected to Beta Correlation

1 1 4 .417988 .41616949

2 2 12 .157774 .44346217

3 3 2 .101468 .45335339

4 5 5 .097116 .47709004

5 6 1 -.084665 .48313474

6 8 8 .057468 .48800746

7 9 9 .029421 .48734207

8 13 17 .022501 .49011051

9 15 3 .029362 .49029619

10 16 13 -.024438 .48935480

11 20 6 -.Ol7884 .48952340

95

Table 8. Continued.

 

 

 

 

 

Criterion -- Botany
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
l 1 2 .428503 .42677099
2 2 11 .177337 .46066171
3 4 5 .124324 .48397329
4 6 7 .076764 .94550371
5 7 15 -.072179 .49940164
6 8 6 .063066 .50204417
7 10 4 .040997 .50457297
8 ll 13 -.030466 .50414828
9 12 ~ 14 -.029079 .05363830
10 14 8 -.032963 .50313167

11 18 1 .022391 .50432664

96

Table 8. Continued.

 

 

 

 

 

Criterion -- Business Administration

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 4 .300948 .29757854

2 2 13 .156670 .33344004

3 ‘ 3 12 .066859 .33719841

4 5 4 .069442 .34804916

5 6 9 -.058407 .35017370

6 7 7 .047306 .35060293

7 9 3 .054758 .35398923

8 10 10 -.037693 .35294836

9 11 2 .030061 .35141366

10 13 1 .032188 .35258957

ll 14 18 -.029747 .35101791

97

Table 8. Continued.

 

 

 

Criterion -- Chemistry

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 5 .439731 .43886931

2 2 11 .213676 .48743261

3 3 8 -.083552 .49375605

4 4 12 .079526 .49940522

5 6 2 .081651 .51542231

6 8 3 .045892 .51958619

7 9 15 -.045465 .52089929

8 14 10 .020494 .52492080

9 15 17 -.024374 .52481446

10 20 14 -.018121 .52654640

11 21 16 -.011748 .52600376

12 26 4 .013307 .52627366

13 27 1 -.011381 .52572121

14 44 7 -.003518 .52605747

15 70 18 .001540 .52558574

16 9O 9 -.000582 .52492425

98

Table 8. Continued.

 

 

 

 

Criterion -- Drama
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
1 1 1 .246807 .23476953
2 2 10 -.086641 .23842772
3 3 11 .090829 .24341080
4 4 13 —.111537 .25700213
5 5 9 .102386 .26630462
6 7 6 .060656 .26836891
7 11 15 .043949 .27394244
8 12 5 -.044037 .26660503
9 14 16 —.035972 .25982723
10 15 3 .033160 .25009455

ll 16 2 -.O39729 .24089900

99

Table 8. Continued.

 

 

 

 

Criterion -- Economics
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
1 l 5 .357997 .35651239
2 2 12 .208269 .41172431
3 4 17 .090715 .42609394
4 5 3 .046278 .42744029
5 7 4 .054680 .43412621
6 9 11 .052370 .43866357
7 10 15 .042789 .43963516
8 12 8 .037989 .44219756
- 9 l4 2 .032802 .44378062
10 15 16 —.041149 .44455067

ll l7 14 -.026457 .44540717

100

Table 8. Continued.

 

 

 

 

Criterion -- Education
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
l 1 1 .465104 .46397778
2 2 12 .154892 .48813867
3 3 10 -.097651 .49679755
4 4 5 .093799 .50461327
5 6 17 .085633 .51492804
6 7 15 .066696 .51830342
7 9 4 .070396 .52556349
8 13 13 .031340 .53156631
9 16 2 .030784 .53377187
10 18 18 .029616 .53440336

11 19 16 —.02603O .53413358

101

Table 8. Continued.

 

 

 

 

 

Criterion —- Engineering
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
1 l 6 .297920 .29453956
2 2 15 .157898 .33133601
3 3 5 .117978 .34904056
4 5 ' 8 .072953 .36623237
5 6 9 .082653 .37291751
6 7 11 -.076542 .37821294
7 8 2 .060599 .38055156
8 9 10 -.069950 .38450633
9 11 7 -.049275 .38753793
10 12 12 .049754 .38828391

11 15 3 .033382 .39226837

102

Table 8. Continued.

 

 

 

 

 

Criterion -- English Composition
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection NUmber Selected to Beta Correlation
1 1 1 .401834 .39969428
2 2 14 .090699 .40779135
3 3 10 .088673 .41527244
4 4 6 .055630 .41695879
5 6 17 .059524 .42477129
6 7 13 -.049444 .42567413
7 8 2 .050508 .42670525
8 10 3 .037851 .42867203
9 ' 11 8 -.029671 .42627499
10 13 5 .036054 .42714912

ll 16 15 -.Ol9338 .42713969

103

Table 8. Continued.

 

 

 

 

Criterion -- English Literature
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
1 1 1 .370076 .36887041
2 2 12 .166641 .40373215
3 3 15 ~.086626 .41189969
4 4 10 —.048115 .41366144
5 5 16 .065578 .41782403
6 6 6 .041865 .41890527
7 8 4 .048063 .42347828
8 9 13 -.045048 .42488154
9 10 2 .034476 .42528658
10 12 9 .042520 .42908883
11 14 17 .032988 .43138982
12 16 7 .035965 .43395641

13 17 11 -.027522 .43384755

104

Table 8. Continued.

 

 

 

 

Criterion -- Foods and Nutrition
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
1 l 1 .433131 .42855409
2 2 12 .153318 .45101226
3 3 10 -.079275 .45361894
4 4 5 .099191 .46027529
5 6 6 .083631 .47022087
6 7 13 .052497 .46907644
7 8 16 -.072300 .47064377
8 9 17 .064431 .47103481
9 11 3 .062258 .47427278
10 15 11 .035786 .47604737

11 ' 16 18 -.027l73 .47266780

105

Table 8. Continued.

 

 

 

 

 

Criterion -- Forestry

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

l l 1 .541687 .52947175

2 2 16 .244756 .57352227

3 3 7 -.l45392 .58170509

4 4 14 -.131486 .58655975

5 6 3 -.127416 .59948212

6 7 9 .097821 .59753283

7 8 11 -.121497 .60047052

8 9 8 .079898 .58993882

9 10 12 —.074084 .58266866

10 11 18 .098543 .57948195

11 12 6 .067661 .57044333

106

 

Table 8. Continued.

 

 

 

 

Criterion —- Geography

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

l l 4 .355227 .35322759

2 2 12 .189989 .39945189

3 3 8 .099808 .40961568

4 4 9 -.056481 .41170531

5 5 17 .068244 .41557682

6 7 1 .059786 .42166279

7 9 16 .036081 .42424842

8 13 11 .036241 .42930075

9 22 15 .021372 .43368591

10 23 3 .014707 .43223546

11 28 10 -.Ol4004 .43179864

107

Table 8. Continued.

 

 

 

 

 

Criterion -- Geology

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

l 1 11 .372391 .37070156

2 2 5 .157697 .40137444

3 3 10 -.079484 .40765521

4 4 4 .076520 .41332706

5 6 18 .071417 .42182422

6 8 2 .050074 .42557684

7 9 3 -.053447 .42753387

8 10 16 .045914 .42860506

9 12 13 .044954 .43265752

10 13 9 -.037797 .43293112

11 16 14 -.037846 .43681534

12 17 7 .032126 .43663134

108

Table 8. Continued.

 

 

 

Criterion -- History

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 12 .388764 .38768669

2 2 4 .231090 .45052270

3 4 5 .089346 .46630211

4 5 14 -.059578 .46928449

5 6 10 -.047732 .47086291

6 7 17 .068654 .47505373

7 9 16 .044050 .48000898

8 11 2 .036126 .48205299

9 14 13 -.031722 .48385040

10 15 1 .021566 .48354191

11 18 15 -.027922 .48482036

109

Table 8. Continued.

 

 

 

 

Criterion -- Home Economics

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 l 5 .424794 .42235729

2 2 6 .146922 .44497853

3 3 11 .119842 .45871171

4 5 1 .105687 .48165176

5 6 10 -.079585 .48619890

6 7 7 .097604 .49405987

7 9 3 .062701 .50216646

8 11 9 .057407 .50646402

9 13 12 .040837 .50851431

10 l4 16 -.042501 .50843824

11 15 17 .038527 .50803582

12 18 18 -.031802 .50951017

13 19 8 .034269 .50472387

14 21 2 —.033013 .50426900

15 22 4 .027649 .50277779

16 27 14 .020850 .50363987

17 29 13 .018870 .50210542

18 38 15 -.018759 .50277317

110

Table 8. Continued.

 

 

 

 

Criterion -- Journalism

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 3 .404787 .39700413

2 2 17 .199091 .43762073

3 3 . 8 .143534 .45260813

4 4 6 .124386 .46283419

5 5. 11 -.115689 .47070002

6 6 15 -.094437 .47363853

7 7 9 .094143 .47654683

8 9 7 .063383 .48273512

9 10 18 -.072553 .48170517

10 13 12 .065601 .48514450

11 l4 14 .045974 .48053910

111

Table 8. Continued.

 

 

 

 

 

Criterion -- Language

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

l l 3 .397947 .39513963

2 2 9 .154683 .42184208

3 3 8 -.120060 .43515046

4 5 12 .105638 .45630578

5 7 15 -.033587 .45625684

6 8 2 .029598 .45456147

7 10 10 .035410 .45541570

8 14 4 -.036002 .45754044

9 15 14 -.025338 .45557407

10 16 18 .022209 .45341775

11 19 11 -.01606O .45181151

112

Table 8. Continued.

 

 

 

 

 

Criterion -- Mathematics
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
1 l 2 .383695 .38273409
2 2 8 -.182626 .42320703
3 3 11 .109898 .43643726
4 4 15 .081834 .44325054
5 6 5 .067657 .45145568
6 10 18 .038077 .45988353
7 12 10 -.023799 .46034233
8 13 1 .025912 .46030504
9 14 9 -.023470 .46013557
10 15 13 .020257 .45981161
11 18 17 -.016188 .46031372
12 22 16 -.015769 .46062940
13 28 14 —.009578 .46093982
14 30 7 .010111 .46036660
15 40 12 .007454 .46045269
16 48 6 .005011 .45996642

17 58 3 .003578 .45932654

113

Table 8. Continued.

 

 

 

 

Criterion -- Music

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 3 .154475 .14978446

2 2 13 .083085 .16711499

3 3 6 .044110 .16868997

4 4 15 —.053330 .17288855

5 5 17 -.045627 .17472730

6 6 7 .061891 .18154901

7 7 12 -.049312 .18434817

8 8 18 .047079 .18648566

9 11 2 -.039127 .19625052

10 12 4 .037362 .19620919

11 15 5 -.027406 .19916067

12 18 1 .024993 .20082625

13 19 9 -.025298 .19886424

14 22 14 -.021134 .19885284

15 24 16 —.016850 .19683372

16 37 10 .013025 .20079640

17 38 8 .012150 .18879987

18 58 11 -.OO4575 .18801481

114

Table 8. Continued.

 

 

 

 

 

Criterion -- Nursing

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 4 .454165 .43596708

2 2 18 .190925 .45952698

3 3 8 .135375 .45635761

4 4 12 -.140051 .45878057

5 5 10 .141866 .46193448

6 6 7 -.115377 .45669876

7 7 2 .121966 .45316974

8 8 1 -.107225 .44482497

9 9 9 .093429 .43185537

10 10 16 -.073999 .41266993

11 12 13 -.058320 .39812897

115

Table 8. Continued.

 

 

 

 

Criterion -- Pharmacy

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

l 1 14 -.424632 .39226401

2 2 1 .316860 .48245247

3 3 15 .161702 .48492921

4 4 2 -.168728 .49028217

5 5 11 .182563 .50174059

6 6 16 -.103314 .48590287

7 7 18 .154236 .48579359

8 10 8 .089476 .48424373

9 14 10 .050532 .47840752

10 15 3 -.07l944 .44698327

11 18 7 .045804 .41970519

116

Table 8. Continued.

 

 

 

 

 

Criterion -- Philosophy

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 5 .379832 .37788121

2 2 18 .205496 .42856318

3 4 12 .079599 .43881320

4 5 8 -.067889 .44192544

5 6 1 —.031150 .44129912

6 7 9 .044111 .44178758

7 9 11 .027001 .44196150

8 11 4 .028505 .44204358

9 13 7 .023981 .44167249

10 15 3 .019476 .44073573

11 l7 17 .017747 .43969443

117

 

 

 

 

 

Table 8. Continued.
Criterion -- Physics
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
l 1 5 .334599 .32909263
2 2 9 .123991 .34658063
3 3 17 -.122952 .36300476
4 4 16 .117341 .37703218
5 5 6 -.116487 .39040615
6 6 8 .069993 .39204830
7 10 13 .068032 .41129471
8 11 7 .032838 .40832335
9 13 12 .041969 .40975450
10 14 10 -.028875 .40642568
11 15 3 .027570 .40294438
12 16 2 -.034l69 .39993838
14 26 15 .021662 .39886986
15 29 4 -.014510 .39519874
16 32 11 .013723 .39133085
17 49 1 .008165 .38847767
18 51 14 .006255 .38369904

118

Table 8. Continued.

 

 

 

 

 

Criterion -- Political Science

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 l 4 .434103 .43276989

2 2 12 0199680 .47551351

3 4 2 .109670 .49244645

4 6 5 .080624 .50219073

5 8 18 .068923 .50872286

6 9 8 -.053772 .50989114

7 10 9 -.041341 .51042815

8 11 16 .050132 .51176257

9 13 7 .026282 .51253666

10 17 15 .029088 .51542827

11 18 17 .019865 .51467387

119

Table 8. Continued.

 

 

 

 

 

Criterion -- Psychology

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 1 2 .401157 .40043091

2 2 12 .269003 .48189551

3 4 4 .116362 .50296156

4 6 7 .085041 .51425799

5 7 15 —.053626 .51656026

6 8 5 .052563 .51874625

7 - 10 18 .042705 .52416722

8 13 10 -.035031 .52844809

9 14 14 -.025376 .52858161

10 15 6 -.022952 .52860373

11 16 3 .025035 .52872113

12 18 17 .024301 .52922079

13 21 8 -.Ol4564 .52920029

120

Table 8. Continued.

 

 

 

 

 

Criterion -- Radio & TV
Order of _' Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
1 1 12 .233164 .18828889
2 2 2 -.224400 .26111037
3 3 3 .260343 .34753428
4 5 6 .202375 .41973748
5 6 8 .167408 .43513450
6 7 13 -.159563 .44741742
7 8 10 .159283 .45979562
8 9 14 -.126950 .46063796
9 10 9 -.132600 .46344391
10 11 11 .160209 .47702806
11 12 15 .166496 .49354734
12 15 16 -.125227 .53353426

13 16 18 .107529 .53042644

121

Table 8. Continued.

 

 

 

 

 

Criterion -- Speech

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

1 l ' 1 .400286 .39903669

2 2 10 -.201328 .44590480

3 3 7 .242497 .50683349

4 5 9 .180641 .55200262

5 9 6 .057896 .57297605

6 10 18 .033233 .57323543

7 12 3 .045040 .57602455

8 14 15 .043071 .57782816

9 15 13 .026992 .57776346

10 17 14 -.019893 .57815708

11 18 4 .018091 .57774079

12 24 17 -.016577 .57879584

13 29 11 .011011 .57881104

14 34 2 -.007419 .57839946

15 38 12 —.006283 .57789951

16 41 8 .005966 .57634125

17 54 16 -.002876 .57584268

122

Table 8. Continued.

 

 

 

 

 

Criterion -- Sociology

Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple

Selection Number Selected to Beta Correlation

l 1 5 .458901 .45835336

2 2 12 .276629 .53498065

3 4 7 .134511 .56058103

4 6 2 .067913 .56674068

5 8 17 .060527 .57173360

6 9 10 -.054211 .57391402

7 10 18 .031180 .57438748

8 12 11 .032819 .57622112

9 13 14 .032128 .57674707

10 15 6 .042452 .57865546

11 16 9 .022286 .57871565

12 18 5 .025296 .57975272

13 19 13 -.027242 .58002703

14 20 15 -.025357 .58021533

15 21 8 -.O3l934 .58012041

 

123

 

 

 

 

Table 8. Continued.
Criterion -- Zoology
Order of Cumulative Squared
Predictor Iteration Predictor Accretion Adjusted Multiple
Selection Number Selected to Beta Correlation
1 1 5 .421329 .41989766
2 2 11 .156245 .44675539
3 3 3 .088524 .45418728
4 5 2 .078134 .47223729
5 7 7 .060267 .47769335
6 8 14 .035089 .47780192
7 10 1 -.042425 .47975972
8 11 8 -.035969 .47917130
9 12 16 .025318 .47856113
10 14 15 -.033174 .47917056
11 17 9 .022055 .48027775

124

predictor beta weights (B's) are found in Table 9. The

raw score predictor weights (b's), a product of the iteration
procedure and used in developing new predictions of grade
points, are found in Table 10.

Table 11 is the product moment correlations and
their E-values for the Z-transformation of the correlation
when a comparison of the two predicted grade points was
made with the achieved grade-point average. The criterion
area All—University showed no statistically significant
difference between the two predicted grade points in their
correlation with the achieved grade-point average. No

significant differences were found for Accounting, Anthropology,

 

 

Bacteriology. Business Administration, Chemistry, Economics,

 

 

Education, English Composition, English Literature, Geography,

 

History, HOme Economics, Language, Mathematics, Music,

Nutrition, Philosophy, Political Science, Psychology, Radio &

 

3y, Sociology, Speech, and Zoology.

Statistically significant differences were found
between the predicted grade points from the Washington State
University sample and the predicted grade points from the
Washington Pre-College Testing Program for the criterion

areas of Architecture, Art, Biology, Botany, Drama, Engineering,

 

 

 

 

Forestry, Geology, Journalism, Nursing, Pharmagy: and Physics.

 

Three areas, Biology, Drama, and Nursing have a higher cor-

 

relation for the predicted grade point from the Washington

Pre-College Testing Program.

 

 

 

 

 

Table 9. Multiple regression beta weights for 1958—1960
differential predictions of thirty—six criteria
of academic success at the Washington State
University.

Predictor Measures

m
m
m o
m c
m .4 H m .4 a m
m o o o o 0-8
0 0 0+4 0 m 0 U
c c-u : a) 3:0

4: 0.: u m L1C m U
0 03m m E uaoim mrq
-H m -H s m
H .C.‘ .—| .C: .C.‘ .C‘. (D O1 ..c‘. H
. . m Uim m+J mxac mt)
Criterion '2 E .‘i 5': g 5‘: :2 .3 :13 :3
No. Measures 5 1 2 3 4

1. A11 Univ. 2031 .1022 .0952 .0475 .1067

2. Accounting 356 .1710 .0965

3. Anthro. 907 .0928 .1035 .1264

4. Arch. Eng. 69 .3799

5. Art 581 .1360

6. Bact. 681 .1597 .1149 .0810

7. Biology 545 -.l759 .1575 .0489 .2471

8. Botany 512 .1611 .0292

9. Bus. Admin. 450 .0605 .0759 .0699 .0634

10. Chemistry 1029 .1315 .0614

11. Economics 790 .0328 .0526 .1095
12. Drama 140 .2021 .0603

13. Education 646 .1803 .0369 .1595
14. Engineer. 446 .1015

15. Eng. Comp. 451 .1608 .0763 .0458

16. Eng. Lit. 846 .1665 .0808 .1013
17. Forestry 53 .4755 -.1118

18. Geography 611 .1641 .0230 .1708
19. Geology 627 .1669 —.0707 .0839
20. History 1008 .0671 .0635 .1245
21. Journalism 125 .2855

22. Home Econ. 395 .1734 .0929

23. Language 332 .1064 .2326 .0568
24. Math. 1118 .0619 .1814

25. Music 682 .0748 .1106 .0739
26. Nursing 51 -.2411 .1984 .4686
27. Nutrition 188 .2100 .1089

28. Pharmacy 30 .4831 .3656 -.1423

29. Philosophy 531 -.1205 .0376 .0718
30. Physics 245

31. Pol. Sci. 635 .1329 .1599
32. Psychology 1300 .1507 .0408 .1241
33. Radio/TV 51 -.6660 .3605

34. Sociology 1403 .0662 .2144
35. Speech 766 .2649 .0899

36. Zoology 676 —.1050 .1613 .1822

126

 

 

Predictor Measures

 

 

 

H H m
0 O H O
0 0m to H
.c: .ca) 0 m u
UHQ) U> - o-v-i r: c: (U
(GNU Ul-u-l NH NC.‘ m «H E
H: JJ m as Ha) H (D
.250) .130 ID I.C‘. HOW H .C‘.
3‘33 .31.“: .3 .8 2‘3 8. “
3120: 31m (D{> (DE! L's-1D a) g
5 6 7 8 9 10 11
.1305 .1218 .0715 -.0534 .0683 -.0651
.1386 —.0912 .1271 -.0919 -.0858 .2619
.1221 .0976 -.0265 -.0397
.1679 .1942 —.3575 .3455 -.1829 -.3563
.1704 .0730 .0955 —.0549 —.0721
.1937 ..0966
.1662 .0606 .0674
.1871 .0813 .0821 -.0869 .1875
.0992 .0882 -.1121 -.0437
.2339 -.1761 .0309 .1710
.1146 .1126 .1022
.0707 .0648 .1550 -.1849 .2459
.1106 -.1476
.1279 .1472 -.1094 .1955 .1788 -.0775 -.1638
.0517 .0948 -.0555 .0841
.0617, .0544 . .0916 -.1263
-.3075 .1267 .2237 —.1982
.0682 -.1931 .0825
.0939 -.0781 -.1192 .1929
.0811 -.1094
.1994 .1055 ..1452 .1787 -.2599
.1331 .1338 .1402 .0989 —.2013 .1286
-.2278 .0791 .0680
.0927 —.1819 -.0524 —.0311 .2095
.0822 .1169
-.1397 .1710 .1484 .2308
.1233 .0946 —.0792 .0282
.2125 .0998 .3554
.2623 .0320 —.0827 .0677 .0328
.3148 —.1214 .0835 .0629 .2271 —.0437
.1439 .0419 -.0550 -.1155
.1197 -.0316 .1179 -.0553
.2393 .4257 -.3055 .2836 .4799
.0479 .0518 .1632 .0628 —.0835 .0312
.0665 3346 .2646 —.53()3

 

127

 

 

 

 

 

Table 9. Continued.
Predictor Measures
m
m
m
(U C)1
U m l
r40 .
m MH m
o «4U -
o: O m x
M 0+1 . m m
m com < < m
'2
No. Measures 2 12 13 14 15
1. All Univ. 2031 .1117 .0296
2. Accounting 356 .0905 .2617
3. Anthro. 907 .1337 -.2036
4. Arch. Eng. 69 —.1057 —.3348 -.l966
5. Art 581 .1443 .0515 .0997
6. Bact. 681 .0765 —.0700 -.0489 -.1310
7. Biology 545 .1056 -.0382
8. Botany 512 -.0388 -.0269 -.l737
9. Bus. Admin. 450 .0725 .1347
10. Chemistry 1029 .1367 -.Ol93 -.1359
11. Economics 790 .1394 .1530
12. Drama 140 -.1990 .0805
13. Education 646 .1009 .0443 .1294
14. Engineer. 446 .0833 .2249
15. Eng. Comp. 451 -.0829 .0982
16. Eng. Lit. 846 .0733 —.0823 -.0798
17. Forestry 53 .1583 -.1252
18. Geography 611 .1309 .0301
19. Geology 627 .0568 -.0385
20. History 1008 .1934 -.0466 —.0579
21. Journalism 125 .0782 -.1989
22. Home Econ. 395 .0685
23. Language 332 .1158 -.0246 -.0635
24. Math. 1118 .0276 .1294
25. Music 682 .0814 .1153 —.0616
26. Nursing 51 .1524
27. Nutrition 188 .1517 .0573
28. Pharmacy 30 -.3939 .4586
29. Philosophy 531 .0836
30. Physics 245 .0527 .0810
31. P01. Sci. 635 .1410 .0416
32. Psychology 1300 .1507 -.0253 .0703
33. Radio/TV 51 .3296 -.2287 -.1445 .5014
34. Sociology 1403 .1422 -.0489 .0299 -.0421
35. Speech 766 .0327 -.0184 .0657
36. Zoology 676 .0366 -.0722

128

 

 

Predictor Measures

 

Coop - Speed
Coop - Level

A.C.Eo—L

 

H
O‘
5...:
\l
|._a
m

 

.0773
.0825
.0925
.0782
.0468
.0506
-.0810 .1245
-.0472
.0851 .0447
.1067
.0720 .0610
.4959 .1372
.0673 .1013
.0977 .0772
.0888 .1236
.1778 -.1096
-.1007
.0287
.0414
-.2034 .1518
-.1642 .2957
-.1598 .1267
—.3308 .2002
.1427
.1276 -.3481
.0889 .0888
.0404 .0410
.2200
.1067 .0214
.0537

.0710

129

 

 

 

 

 

Table 10. Multiple regression predictor weights for 1958-
1960 differential predictions of thirty-six
criteria of academic success at the Washington
State University.

Predictor Weights
Predictor Measures
m
(D H HID H H
m o o o o o
m o o-a o m o
U c :44 .c m s m
a): U m ()c m U m
m m m m E mtmm mrau
O -H m H s ruc
..CH £2.12 .CCDU‘ .21-HG)
H Dim OLD mxac mcra
. . w .. 99 -~93 as:
Criterion E tub: m 5
No. Measures a 1 2 3 4
1. All Univ. 2031 .0108 .0088 .0040 .0113
2. Accounting 356 .0179 .0092
3. Anthro. 907 .0102 .0104 .0158
4. Arch. Eng. 69 .0434
5. Art 581 .0145
6. Bact. 681 .0199 .0132 .0115
7. Biology 545 -.0240 .0189 .0053 .0338
8. Botany 512 .0213 .0044
9. Bus. Admin. 450 .0071 .0078 .0066 .0074

10. Chemistry 1029 .0154 .0066

11. Economics 790 .0034 .0051 .0132

12. Drama 140 .0209 .0050

13. Education 646 .0201 .0036 .0178

14. Engineer. 446 .0103

15. Eng. Comp. 451 .0178 .0074 .0040

16. Eng. Lit. 846 .0189 .0080 .0115

17. Forestry 53 .0662 .0125

18. Geography 611 .0181 .0020 .0189

19. Geology 627 .0189 .0074 .0108

20. History 1008 .0074 .0062 .0138

21. Journalism 125 .0279

22. Home Econ. 395 .0199 .0086

23. Language 332 .1064 .2326 -.0568

24. Math. 1118 .0082 .0210

25. Music 682 -.0748 .1106 .0739

26. Nursing 51 —.0291 .0210 .0566

27. Nutrition 188 .0254 .0106

28. Pharmacy 30 .0565 -.0375 .0134

29. Philosophy 531 .0158 .0040 .0095

30. Physics 245

31. P01. Sci. 635 .0147 .0201

32. Psychology 1300 .0180 .0044 .0169

33. Radio/TV 51 -.0674 .0334

34. Sociology 1403 .0072 .0265

35. Speech 766 .0296 .0080

36. Zoology 676 -.1050 .1613 .1823

.130

 

 

Predictor Weights

Predictor Measures

 

 

 

m

U

r:

m m
H-H F: U
00 O H H
OU) Om m 4.1
1:: £0) 0 Di m
UF4 O > . o-H .c s E
mtu mua Nr4 N a m -H w

H u m m -H 0 F! .c
42:5 £0 IQ IS .401 H +3
2:: .92 .3: .8 9: a 9
2:2 mid o:> EDS 93D m

5 6 7 8 9 10 11
.0129 .0149 .0049 -.0030 .0018 -.0046
.0155 -.0127 .0079 -.0027 -.0068 .0219
.0142 .0079 —.0041 . ' . - .0034
.0179 .0258 .0264 .0099 -.0140 .0284
.0169 .0090 .0065 —.0039 .0054
..0258 .0089
.0213 .0043 .0023
.0265 .0143 .0080 —.0069 .0198
.0109 .0062 -.0033 -.0034
.0293 -.0124 .0028 .0160
.0129 .0071 .0086
-.0069 .0078 .0040 -.0128 .0178
.0115 -.0110
.0141 .0198 .0082 .0118 .0052 -.0060 .0132
.0055 .0122 -.0032 .0062

.0081 .0040 .0026 -.OO96
.0277 .0092 .0078 .0193
.0039 -.0053 .0063
.0114 -.0025 -.0103 .0175
.0084 -.0081
.0281 .0083 .0092 .0054 .0220
.0143 .0179 .0104 .0028 .0155 .0105
-.2278 .0792 .0680
.0115 —.0126 -.0017 -.0027 .0194
.0821 .1169
—.0109 .0108 .0045 .0187
.0139 .0133 ° -.0064 .0023
.0130 .0078 .0290
.0322 .0027 —.0057 .0022 .0030
.0372 —.0178 .0068 .0042 .0072 -.0037
.0169 .0034 —.0036 -.0036
.0153 —.0050 .0104 —.0050
.0321 .0257 -.0088 .0219 .0387
.0056 .0074 .0131 .0019 —.0069 .0027
.0086 .0169 .0074 —.0398
2165 .0540 -.0643 .0660

 

131

 

 

 

 

 

Table 10. Continued.
Predictor Measures
m
m
U)
r0 0
U m l
H m .
LH row-l E13
0 'r-I'U .
o 5 O m
H OJJ . m
m mcn c a
Criterion g
No. Measures E 12 13 14
1. All Univ. 2031 .0116 .0317
2. Accounting 356 .0074
3. Anthro. 907 .0164
4. Arch. Eng. 69 -.0083 -.3871
5. Art 581 .0105 .0554
6. Bact. 681 -.0107 -.0068 -.0705
7. Biology 545 .0142 -.0036
8. Botany 512 -.0040 -.0412
9. Bus. Admin. 450 .0083 .0109
10. Chemistry 1029 .0180 -.0262
11. Economics 790 .0165
12. Drama 140 -.0142
13. Education 646 .0111 .0034
14. Engineer. 446 .0094
15. Eng. Comp. 451 -.0063 .1100
16. Eng. Lit. 846 .0081 -.0064
17. Forestry 53 .0216 -.l764
18. Geography 611 .0142
19. Geology 627 .0051 -.0505
20. History 1008 .0211 —.0036 -.0649
21. Journalism 125 .0093
22. Home Econ. 395 .0077
23. Language 332 .1158 -.0247
24. Math 1118 .0025
25. Music 682 -.0815 .1153
26. Nursing 51 —.0181
27. Nutrition 188 .0180 .0048
28. Pharmacy 30 -.4674
29. Philosophy 531 .0108
30. Physics 245 .0066 .0071
31. Pol. Sci. 635 .0175
32. Psychology 1300 .0202 -.0350
33. Radio/TV 51 .0387 .0374 -.0182
34. Sociology 1403 .0173 -.0042 .0376
35. Speech 766 .0025 -.0209
36. Zoology 676 .0366 .0722

132

 

 

Predictor Measures

 

 

 

H
J
. I I .
m p H
- Q.w o.m
x U o m o >
(D - 00.: 00)
U) a: Om 0.4
Regression
15 l6 17 18 Constants
.0103 -.68488
.4077 .61744
-.3311 .0070 .57207
.2921 7.55245
-.1381 .0047 -.49290
-.2430 .0076 1.71129
.0043 .23202
-.3420 .91035
.75202
-.2372 -.0046 .90485
.2403 .0046 .0102 .71889
.1087 -.0023 2.17504
.1885 .0064 .0063 .53078
.3392 .68321
.0080 -.62547
-.1l80 .0039 .0047 .85946
.0331 .0242 1.99790
.0435 .0035 .0076 .88507
.0061 .0127 1.39189
.0047 .0093 1.83797
-.3145 .0147 —.0169 .50397
-.0055 .41265
-.0646, .0287 1.69675
.2228 .0069 .65697
—.0616 -.2034 .1517 1.97221
-.0095 .0452 .86913
-.0092 .0104 .49910
.6993 —.0186 .0297 10.04656
.0238 .69455
.0078 —.0300 .61195
.0682 .0054 .0142 .17572
—.1250 .0037 .0070 .96284
—.1691 .7536 —.0122 5.62997
-.0677 .0090 .0033 -.47507
.0957 .0076 .81135
.0710 -.46000

Table 11.

133

Comparison of simple correlation coefficients, Z-

transformation of correlations and E values of
significance for W.P.C.T.P. predicted grades
and Washington State University predicted grades

when correlated with achieved grades for 1961
freshmen at Washington State University.

 

 

 

Z Z t
Criterion Area W.S.U. Trans. W.P.C.T.P. Trans. Value
All-University ..65 .738 .66 .744 th Sig.
Accounting .43 .406 .45 .409 th Sig.
Anthropology .59 .671 .59 .671 NOt Sig.
Architecture .57 .648 .36 .377 .01
Art .42 .448 .36 .377 .05
Bacteriology .42 .448 .46 .491 Not Sig.
Biology ..51 .563 .56 .633 .05
Botany .59 .671 .43 .460 .01
Business Admin. .42 .448 .41 .436 Not Sig.
Chemistry .52 .576 .53 .590 Not Sig.
Economics .47 .510 .49 .536 Not Sig.
Drama .22 .224 .32 .332 .01
Education .49 .536 .51 .563 Not Sig.
Engineering .44 .472 .38 .400 .05
English Comp. .41 .436 .39 .412 Not Sig.
English Lit. .50 .549 .46 .497 Not Sig.
Forestry .40 .424 .22 .224 .01
Geography .51 .563 .51 .563 Not Sig.
Geology .46 .497 .39 .412 .05
History .50 .549 .51 .563 Not Sig.
Journalism .42 .448 .23 .234 .01
Home Economics .47 .510 .49 .536 Not Sig.
Language .47 .510 .50 .549 Not Sig.
Mathematics .48 .522 .46 .497 Not Sig.
Music .13 .131 .05 .050 NOt Sig.
Nursing .46 .497 .59 .671 Not Sig.
Nutrition .56 .637 .53 .590 Not Sig.
Pharmacy .38 .400 -.16 .161 .01
Philosophy .41 .436 .45 .485 NOt Sig.
Physics .40 .430 .34 .354 .05
Political Sci. .46 .497 .50 .549 th Sig.
Psychology .52 .575 .54 .604 Not Sig.
Radio & TV .49 .536 .45 .485 Not Sig.
Sociology .60 .693 .59 .671 Not Sig.
Speech .48 .522 .46 .497 Not Sig.
Zoology .48 .522 .46 .497 Not Sig.

 

134

The criterion areas Architecture, Art, Biology,
Forestry, Journalism, Nursing and Pharmacy are special areas
at Washington State University inasmuch as they are separate
schools with small enrollments or somewhat different curricu-
lum than similarly named departments at the University of
Washington. The differences in Bptany, Drama, Engineering,
Geology, and Physics are not so readily explained.

In summary the grade predictions used in the Washington
Pre-College Testing Program correlate with achieved grades
of students at Washington State University as well as did
the grade prediction developed using normative data from
Washington State University except in the prediction areas
of Architecture, Art, Biplogy, Botany, Drama, Engineering,

Forestry, Geology, Journalism, Nursing, Pharmacy and Physics.

CHAPTER V

SUMMARY AND CONCLUSIONS

Summary

The increasing number of students desiring the oppor-
tunity of higher education has not only increased demands on
institutions of higher education but has created problems
in the assessment of students and institutions. One primary
assessment problem is admissions requirements and subsequent
student academic performance in the classroom. Faculty and
student personnel workers must obtain the best possible
estimate of the student's academic potential and limitations
and must use this information in working with the student.
Assessment of the individual student and the total student
group must be integrated with a comprehensive understanding
of the institution to provide for an effective utilization
of the resources of higher education and the nation.

The total student personnel program of an institution
of higher education includes the use of psychological test-
ing for the study of student aptitudes and achievements.
Recognizing that these tests are only part of a total pro-
gram, college student services personnel and faculty are
able to plan effectively a total and valid program. The

wide variety of available tests, the varying techniques for

135

136

presentation of data on students, and the existing dif—
ferences among students and between institutions present a
bewildering problem in subsequent use of test data for assess-

ment and admissions.

Purpose

The advent of the Washington Pre-College Testing
Program led to an evaluation of the state—wide testing pro-
gram.at Washington State University and the use of grade
predictions from the state-wide program. The Washington
Pre-College Testing Program was developed at the University
of Washington and had as its primary basis the use of pre-
dicted grades in collegiate academic areas as indices of
academic performance. The predicted grades are formulated
from normative data at the University of Washington.

The basic question of this study is whether a pre-
dicted college grade from the Washington Pre-College Testing
Program for a certain academic area is applicable to students
at Washington State University. Are the multiple regression
formulas, predictor beta weights, and the basic normative
data applicable for the best prediction of students' grades
at Washington State University? Specifically four questions
with related hypotheses were asked:

1. Students' aptitudes and achievements, as measured

by high school grades and testing data on the

Washington Pre-College Testing Program, were

inspected to determine what similarities or dif-

ferences existed between the two normative groups
in this study.

137

2. How accurate are the current predicted grades of
the Washington Pre-College Testing Program for
students at Washington State University at various
levels of progress, i.e., freshman, sophomore,
or senior? Can the grade predictions be used to
predict grades equally well at all three levels?

3. Is there a hierarchy in the subject matter areas
as represented by the predicted and achieved grades
from the Washington Pre-College Testing Program
and what similarity does this hierarchy have to the
predicted and achieved grades for students at
Washington State University?

4. A null hypothesis was made in each of the 36
criterion areas between grade predictions from the
Washington Pre-College Testing Program and grade
prediction derived from normative data at Washington
State University when correlated with the achieved
grades for students at Washington State University
at the cumulative freshman-junior level. The
differences in prediction would be such that no

new prediction formulas were necessary for pre-
diction studies at Washington State University.

Procedure

The data used in the study were available on punched
cards for the test scores and high school grades and on
computer magnetic tape for the college grades. The 1958-1960
freshmen at Washington State University, for whom a complete
and valid record of data from the Washington Pre-College
Testing Program was available, were used as a normative group
to derive the correlation matrix for the predictor variables
and the predictor-criterion matrix. The 1961 freshmen group
at Washington State University were used for certain cor-
relation data in the study.

The two matrices were used with the Horst multiple
regression or iteration program for the computation of the
corrected multiple correlation coefficients (R's) between

the predictors and the criteria of student achieved grades

138

at Washington State University. The predictor weights from
the computation of the multiple R's were substituted into

the IBM 709 computer program for the present Washington Pre-
College Testing Program and new predictions were developed.
The new predictions were compared with the predictions from
the Washington Pre-College Testing Program as to correlations
with achieved grades. The 1961 freshmen at Washington State
University were used as a cross-validity sample: the criteria
of comparison were the achieved grades of the 1961 freshmen

through the academic years of 1961-1964.

Findings

 

The two normative groups in the study were comparable
and similar for the purpose of the study. Inspection of
student test scores and high school grades showed marked
similarity for the two groups; the few differences noted
were attributable to the differing ratio of male to female
students in the two universities.

The grade predictions from the Washington Pre-College
Testing Program, when correlated with the achieved grades
at the cumulative freshman, cumulative freshman and sophomore
and cumulative freshman-junior levels, showed no significant
differences except that there was a higher correlation at

the cumulative freshman level for All-University, Art, and

 

Bacteriology.

 

A rank correlation coefficient of .88 between the
predicted grades for the normative groups of the two insti-

tutions showed that a hierarchy exists when ranking the

139

predicted grades from high to low. The two criterion areas

of English Composition and Economics accounted for most of

 

the deviation from complete correlation. _A rank—correlation
of .57 between the achieved grades of the two normative
groups showed a greater variation exists in achieved than

in predicted grades. A comparison of the ranking of the
predicted grades with the ranking of the achieved grades

for the 1961 freshmen at Washington State University showed
a rank-correlation of .57.

The predictions derived from multiple iteration pro-
cedures of the criteria of grades at Washington State
University did not improve the correlations with achieved
grades over the predicted grades from the Washington Pre-
College Testing Program except for the criterion areas of
Architecture, Art, Biology, Engineering, Forestry, Geology,
Journalism, Pharmagy, and Physics. The prediction formulas
now used in the present Washington Pre-College need not be

changed except in the cases noted.

Conclusions

It is not necessary to develop for Washington State
University students multiple regression formulas different
from those used in the Washington Pre-College Testing
Program except for certain criterion areas. These criterion
areas are for the most part special curriculum areas or
programs at Washington State University. Empirical research

has shown that similarly named criterion areas at different

140

institutions may or may not be predicted by the same
statistically derived formula and that the application of
the grade predictions from the Washington Pre-College
Testing Program to other institutions and student groups
must be validated by research.

The grade predictions from the Washington Pre-College
Testing Program can be used to predict cumulative freshman
grades as well as the previously recommended total cumula-
tive grades. The general acceptance of a null hypothesis
for differences among the three levels of achieved grades
(cumulative freshman, cumulative freshman and sophomore, and
cumulative freshman-junior), when correlated with the pre-
dicted grades, makes it possible to use the cumulative fresh-
man grades as a criteria for the prediction procedure rather
than using the criteria of the cumulative four year grades.
The resulting saving in time and the more immediate vali-
dation procedure for introduction of new testing data or
criteria make such a procedure a recommendation in future
normative studies.

The hierarchy among the predicted and achieved
grades in the Washington Pre-College Testing Program reflects
the concept of differential prediction and the differences
in grading practices of departments within the university.
A clear perception of such a hierarchy can be of help in
the understanding of differential prediction and the pre—
dicted grades from the Washington Pre-College Testing Program.

The similarity of the student groups from Washington State

141

University and University of Washington when compared on

the predictor variables in the Washington Pre-College
Testing Program, and the similarity of the grading practices
as compared on the predicted grades, demonstrate the student
populations and grading practices are much the same and,
therefore, the prediction formulas can be used equally

well for both groups.

Recommendations

1. The predicted course grades in the Washington Pre-
College Testing Program should be redefined in other than
the established criterion areas. Summarization by varying
levels of course work or curriculum within a criterion area,
rather than by all courses within the criterion area, might
more sharply define sampling and grading differences.

2. Decisions as to further reductions in the number
of prediction areas should be made on the basis of empirical
analysis of the similarity or prediction statistics and on
practical considerations from counselors and research workers.
The present use of criterion areas from only one university
and the grouping of courses into criterion areas as defined
in that university's bulletin ignores other curriculum or
programs.

3. Predictions should be made for general areas,
primarily for general courses taken during the freshman
year. For example, many freshman students are interested

in social science as a general area but lack the information

142

necessary to differentiate between such specific fields
as sociology or psychology.
4. The predictions in the criterion areas of Music,

Art, Drama, and Radio and Television should be dropped from

 

the testing program because of the low multiple correlation
in prediction and the difficulty in prediction in these
areas from the types of tests in the present test battery.
The continued use of predictions in these areas can lead
only to further misunderstanding and erroneous conclusions
concerning prediction validity.

5. The prediction formulas used in the Washington Pre-
College Testing Program should be based upon one year achieve-
ment grades rather than four year extent of achievements.
Certain exceptions may be necessary such as criterion areas
where the necessary grades are not available until the

second or third year.

BIBLIOGRAPHY

Ad Hoc Committee to Study the Freshman Year, "A Progress
Report,” University Admission Policy, University
of Washington, Seattle, February 25, 1960.

Angell, M. A., Langton, R. C., Meyer, G. A. and Pettit, M. A.
"An Evaluation of General and Specific Admission
Requirements at the University of Washington."
Unpublished Doctor's Dissertation, University of
Washington, Seattle, 1950.

An IBM 709 Grade Prediction Program, Library Report.
Seattle: Division of Counseling & Testing Services.
n.d. 30 pp. (Ditto.)

Atkinson, Gil. "An IBM 709 Grade Summarization Program,"
IBM Type 709 Program Library Report. Seattle:
Division of Counseling & Testing Services, 34 pp.
(Ditto.)

Blair, Glenn M. "The Prediction of Freshmen Success in the
University of Washington." Unpublished Master's
thesis, University of Washington, Seattle, 1931.

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

Brammell, P. R. ”A Study of Entrance Requirements at the
University of Washington." Unpublished Doctor's
Dissertation, University of Washington, Seattle, 1930.

Bruce, William J. "The Contribution of Eleven variables
to the Prognosis of Academic Success in Eight Areas
at the University of Washington.” Unpublished
Doctor's Dissertation, University of Washington,
Seattle, 1953.

Carlson, John Spencer and Milstein, Victor. "The Relation
of Certain Aspects of High School Performance to
Academic Success in College,” College and University,
33:185-92 (Winter, 1958).

 

Cosand, Joseph P. "Admissions Criteria: A Report to the
California Committee for the Study of Education,"
College and University, 28:338—364 (April, 1953).

 

143

144

Crawford, A. B. and Burnham, P. S. Forecasting College
Achievement. New Haven: Yale University Press,
1946.

 

Cronbach L. J. and Gleser, G. C. "Assessing Similarity
between Profiles," Psychological Bulletin, 50:456-473
(1953).

Darley, John G. and Anderson, Gordon v. "The Functions of
Measurement in Counseling," E. R. Lindquist (Ed.),
Educational Measurement. Washington, D.C.: American
Council on Education, 1951.

 

Dressel, Paul R. A Report on Differential Prediction and
Placement in Colleges and Universities. New York:
College Entrance Examination Board, 1959, 17 pp.
(Mimeographed.)

Durlinger, G. W. "The Prediction of College Success: A
Summary of Recent Finding,“ The American Association
of College Registrars Journal, 14:68-78 (OCtober,,,v

1943).

Fishman, Joshua A. "Unsolved Criterion Problems in the
Selection of College Students,” Harvard Educational
Review, 28:320-29 (Fall, 1958).

Fishman, Joshua A. and Pasanella, Ann K. "College Admission-‘1
Selection Studies," Review of Educational Research,
30:298-310 (1960).

Franks, Dean K. "A Study of the Success of West Seattle
High School Students in Language Arts, Foreign
Language, Social Studies, and Music and Art."
Unpublished Doctor's Dissertation, University of
Washington, Seattle, 1958.

French, J. W. The Logic of and Assumptions Underlying
Differential Testing. Proceedings 1955 Invitational
Conference on Testing Problems. Princeton, New
Jersey: Educational Testing Service.

 

Fricke, Benno G. "A Coded Profile Method for Predicting
Achievement," Educational and Psychological Measure-
ment, 17:98-104 (Spring, 1957).

 

Garrett, Harley F. "A Review and Interpretation of
Investigations of Factors Related to Scholastic
Success in Colleges of Arts and Sciences and
Teachers Colleges," Journal of Experimental Education,
18:91-131 (December, 1949).

 

Harris, Daniel, ”Factors Affecting College Grades: A
Review of the Literature, 1930-37," Psychological
Bulletin, 37:125-26 (March, 1940).

 

145

Heathers, Louise 8., Kintneo, Robert, Langen, Thomas D.
and Bjork, Susan. "Comparison of Male and Female
Students in the 1961 Entering Freshman Class of the
University," part of DCT Project 0961-100.
Seattle, Washington: Division of Counseling and
Testing Services, University of Washington.
(Dittoed.)

Heist, Paul and Webster, Harold. "Differential Character-
istics of Student Bodies--Implications for Selection
and Study of Undergraduates in Conference on Selection
and Educational Differentiation," Selection and
Educational Differentiation. Berkeley, California:
The Center for the Study of Higher Education, 1960.

Holland, John L. "The Prediction of Scholastic Success for
a High Aptitude Sample," School and Society, 86:
290-293 (June, 1958).

Horst, Paul. "A Technique for the Development of a
Differential Prediction," Psychological Monographs:
General and Applied, vol. 68, No. 9, Whole No. 380.
Washington, D. C.: American Psychological Association,
1954.

. "A Technique for the Development of a
Multiple Absolute Prediction Battery," Psychological
Monographs, vol. 69, No. 5, Whole No. 390 (1955).

 

 

Horst, Paul and Smith, Stevenson. "The Discrimination of
Two Racial Samples," Psyohometrika, 15:271—289
(September, 1950).

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

. ”Differential Prediction of Academic
Success," Technical Report, Office of Naval Research
Contract Nonr-477 (08), University of Washington
Division of Counseling and Testing Services
(May, 1959).

 

. "A Technique for the Development of a
Multiple Absolute Prediction Battery," Psychological
Monographs, V01. 69, No. 5, Whole No. 390 (1955).

 

 

IBM 709 Correlation Program for SHARE 215.

IBM 709 Correlation Matrix Program, Washington State
University.

Long, James R. "Academic Forecasting in the Technical-
Vocational High School Subjects at West Seattle
High School." Unpublished Doctor's Dissertation,
University of Washington, Seattle, 1957.

146

Lounesberry, James Rodney. "An Evaluation of the Accuracy
of the Differential Prediction Test Battery in
Predicting Grades for Student at Western Washington
State College.“ unpublished Doctor's Dissertation,
University of Washington, Seattle, 1962.

Mattrick, William E. "A Study of the Contributions of
Twelve Variables to Prediction of Academic Success
in Five Ninth Grade Subjects." Unpublished Doctor's
Dissertation, University of Washington, Seattle,
1958.

Meredith, William M. "Cumulative Calculations of Regression
Constants," Multiple Prediction Studies, Office of
Naval Research Contract Nonr-477 (08), Paul Horst
principle investigator. Unpublished report of the
University of Washington, Seattle, June, 1956.
(Mimeographed.)

Mollenkopf, W. G. "Some Aspects of the Problem of Differen-
tial Prediction," Educational and Psychological
Measurement, 12:39044 (1952).

Michael, W. B. "Development of Statistical Methods
Especially Useful in Test Construction and Evaluation,‘
Review of Educational Research, 29:89-109 (1959).

 

Mills, D. F. "An Interative Selection of variables for
Predicting Certain Criteria of Academic Success at
the University of Washington." Unpublished Doctor's
Dissertation, University of Washington, Seattle,
1957.

"Relationship between Preadmission variables and Success in
College," Office of Naval Research Contract Nonr-477
(08), and Public Health Research Grant Mr743 (c3),
Paul Horst, principal investigator. Unpublished
report, The University of Washington, Seattle, June,
1959. (Mimeographed.)

Reas, Herbert D. "A Follow-Up Study of the Washington
Pre-College Differential Guidance Test at Seattle
University." Unpublished Doctor's Dissertation,

University of Washington, Seattle, 1962.

Rogers, Carl. Client Centered Therapy. New York: Houghton-
mifflin Company, 1951.

Rulon, P. J. "Distinction between Discriminant and Regression
Analyses and a Geometric Interpretation of the
Discriminant Function," Harvard Educational Review,
21:89-90 (Spring, 1951).

 

147

Salyer, Rufus C. "An Investigation in the Prediction
of Success in the School of Engineering at the
University of Washington." Unpublished Master's
thesis, University of Washington, Seattle, 1931.

Segal, David. Prediction of Success in College. United
States Department of Interior, Office of Education,
Bulletin No. 18, Washington Government Printing
Office, 1934.

Sorensen, Richard C. and Dvorak, August. "An IBM Type
709/7090 Program to Select Predictors, Calculate
Multiple Correlations, and Determine Linear Regres-
sion Equations," IBM Type 709 Program. Library
Report. Seattle: Division of Counseling & Testing
Services, n.d. (Dittoed.)

Statistical Abstracts of the United States, 84th Annual
Edition. Washington, D.C.: U.S. Government
Printing Office, 1963.

Stone, J. B. "Differential Prediction of Academic Success
at Brigham YOung University," The Journal of
Applied Psychology, 38:109-110 (April, 1954).

Tatsuoka, Maurice M. "Joining Probability of Membership
and Success in a Group: An Index Which Combines
the Information from Discriminant and Regression
Analysis as Applied to the Guidance Problem,“ Harvard
Studies in Career Development, No. 6, Harvard
Graduate School of Education, October, 1957.
(Mimeographed.)

 

Thummel, James. "An IBM 1401 Program to Summarize
Differences in Percentages of Grade Point Averages,"
Washington State University. (Dittoed.)

. An IBM 1401 Computer Program for Grade
Summarization at Washington State University.

 

Travers, Robert M. W. "The Prediction of Achievement,"
School and Society, 70:293 (November, 1959).

 

Tyler, L. E. "Toward A Workable Psychology of Individuality,’
American Psychologist, 14:75-81 (1959).

 

Tyler, Leona E. The Psychology of Human Differences.
Boston: Appleton-Century-Crofts, Incorporated,
1956.

 

Tiedeman, David V. "The Multiple Discriminant Function-—A
Symposium," Harvard Educational Review, 21:167-186
(Spring, 1951).

 

148

"Validity Coefficients for 1955-1956 Weights," Duplicated
Report, University of Washington Division of
Counseling and Testing Services, July, 1959.

Walley, Donivan. "Factors that Influence the Selection of
Predictor Variable in Multiple Regression," College
and University, 39:72-76 (1963).

Wesman, A. G. and Bennett, G. K. ”Problems of Differential
Prediction," Educational and Psychological Measurement,
11:265-272 (1951).

Wolfe, Del. America's Resources of Specialized Talent.
New York: Harper & Brother, 1954.

Yule, G. Udny. "On the Theory of Correlation," Journal of
the Royal Statistical Society, London, 60:835-838
(December, 1897).

Zeigler, Martin L., Bernreuter, Robert G., and Ford, Donald
H. "A New Profile for Interpreting Academic
Abilities," Educational and Psychological Measurement,
18:583-88 (Autumn, 1958).

AP PENDICE S

Hw>mq .pmmm .mooo .mH
.ommmm .pmmm .moou .04
.4 I .m.o.¢ .04

xmm .04

004 .44

.0 I .m.0.¢ .MH

.50m .00m .0000 .NH
:00: .0000 .44

.44000 .z.o .0000 .04

 

oooo.H mommD .2.0 .Qoou .m
0004.- 0000.4 .3ocx .0002 .0.0 .0
0000. 4400. 0000.4 40440> .0.0 .0
000m. 0004.I 4004. 0000.4 400 .00040 .0.0 00
0004. 0040.: 00mm. 4004. 0000.4 006 .400 .002 .0.0 .0
4004. 4004.I 000m. 4004. 0000. 0000.4 <06 .400 .000 .0.0 .4
0000. 0404.: 4040. 0004. 0000. 0040. 0000.4 000 .0004 .000 .0.0 .0
0004. 0440. 4000. 4004. 0000. 0000. 0000. 0000.4 000 .0002 .0.0 .0
4000. 0000.: 0044. 0400. 4000. 0400. 0000. 4400. 0000.4 400 0044000 .0.0 .4
00000 .0002 .num> 406 «mu <00 0mm <06 «mo

so .0.o .0.o .40040 .400.2 .400.0 .0004 .0002 .000

0 0 0 0 0 4 .400 0 4
m

 

 

mH x wH I mwoo mefidz ZOHBflHmmMOU UHmBmzzwm
MBHmmm>HZD m9¢9m ZOBUZHmmflz I Zmzmmmmm oomHlmmmH <B<O WUZ¢mBZm

fl XHQmem<

150

151

 

0400. 40>04 .0000 .0ooo .04
0004. 0004. .00000 .0000 .0000 .04
0000. 0000. 0000.4 .4 : .0.o.< .04
0000. 0000.: 4000. 0000.4 000 .04
0000. 0000.: 4000.: 0000.: 0000.4 000 .44
4400. 0004. 0400. 0400.: 0000.: 0000.4 .0 I .0.o.0 .04
0000. 0000. 4400. 0400.: 0000.: 0004. 0000.4 0040040 .000 .0000 .04
0044. 4004. 4044. 0044.: 0000.: 0000. 0404. 0000.4 0002 .0ooo .44
0000. 0004. 0040. 0000. 0000.: 0000. 0040. 0040. 0000.4 .44000 .z.o .0000 .04
0004. 0040. 0400. 4000. 0404.: 0000. 0004. 0000. 0000. 00000 .s.o .0000 .0
0004. 0000. 4400. 0000.: 0004. 4000. 4004. 0044. 0040.: .3000 .0002 .0.0 .0
0400. 4000. 0440. 0000. 0040.: 0000. 0400. 44000 0040. 40440> .0.0 .0
4404. 00040 4440. 4040. 0000.: 0404. 0404. 4404. 4400. 000 .00040 .0.0 .0
0400. 0000. 0000. 0004. 0000.: 0000. 0000. 4000. 0000. 000 .400 .002 .0.0 .0
4000. 0004. 0004. 0040. 0000.: 0040. 0000. 4400. 0000. 000 .400 .000 .0.0 .4
4000. 0040. 0400. 0400. 0000.: 0000. 0000° 0000. 0004. 000 .0004 .000 .0.0 .0
4040. 0440. 0000. 0400. 0400.: 0004. 4000. 0000. 0000. 000 .0002 .0.0 .0
0000. 4004. 0004. 0000. 0000.: 0000. 0000. 0440. 4040. 000 0044000 .0.0 .4

40>04 .00000 4 o .0.0 .0000 .44000

.0000 .0000 000 000 004 004 .0000 .0000 so

.4000 .mooo

04 04 04 04 44 04 04 44 04

 

 

 

.pmsoHucoo .4 xHUcmmm<

 

152

0000. 0440. 0004. 0004. 0000. 4004. 0000. 40>04 .0000 .0000 .04
0000. 0m40. 404m. 0004. 0000. 4000. 4000. 00000 .0000 .0000 .04
0000. 00mm. 0000. 44mo. 0000. mmom. 00mm. .4 I .0.0.< .04
0000. 4moo.I 4000.: m004.I 0000. 4400. 0004. x00 .04
0000.: m00o.I 0040. 0000.: 0440. 0004. 4000.: 004 .44
0040. 0004. 4000. 0000. 0000. 0400. 0000. .0 I .0.0.4 .04
0440. 0004. 0000. @004. 040m. 404m. 4000. 0040300 .000 .0000 .04
0400. 4040. 4000. 0404. mmmm. 4000. 0000. .4002 .0000 .44
0000. 0004. 0004. 0m4o. 0000. 0004. 0040. 00444000 .2.0 .0000 .04
4000. 0000. 0040. @040. 0000. 0004. 0444. 0000: .2.0 .0000 .0

0000. 0000, 0400. 4004. 0000. 0000. 0000.: .3000 .0002 .N.o .0

00mm. 4040. 4000. 0040. 44mm. 4404. 4000. 40440> .N.0 .0

00mm. 0000. @400. 4000. 04mm. 4404. 0044. 400 .00040 .m.m .0

4044. 0404. 4o4m. mmmm. 44mm. 0000. 0040. 400 .400 .002 .m.m .0

0044. 000m. 0000. 0000. 0004. 4000. 4000. «00 .400 .000 .0.0 .4

404m. 0000. o00m. 0004. 4400. 0000. 0004. 400 .0004 .400 .m.m .0

4004. 0004. 000m. 000m. 0000. 4004. 0004. <00 0002 .0.0 .0

0400. 0000. 0000. 0040. 4000. 0000. @040. 400 0044000 .m.m .4

>004O4m .0000 04¢ .n04< .ousus< 0040050004 ..>4coI444

 

 

MBHmmm>HZD MBdfim ZOBOZHmmdz Bfl mflmmfifiu Zmzmwmmm omelwmmH NEE mom
mmqdeM§> fimm¢ ZOHMMBHKU wm EBHZ mmqmflHm¢> mOBUHQmMQ mH m0 NHmfifiz UHmBQEEWmIZOZ

m NHQmemfl

153

 

0404. 0000. 0000. 0000. 0400. 0004. 0000. 40>04 .0000 .0000 .04
0000. 0400. 0000. 4000. 0400. 0040. 4040. 00000 .0000 .0000 .04
0000. 4400. 0000. 0000. 0000. 4000. 0000. .4 : .m.0.0 .04
0004. 0000. 0044. 0000. 4000.: 0040. 4000.: x00 .04
0000. 0000.: 0000.: 0000. 0000.: 4000.: 0000.: 000 .44
4400. 0000. 0040.: 4000. 0000. 0040. 4000. .0 : .m.0.< .04
0004. 0000. 0004. 0000. 4040. 0400. 0000. 0040000 .000 .0000 .04
0000. 0004. 4404. 0000. 0000. 0000. 0400. .0002 .0000 .44
0000. 4004. 0000. 0000. 0040. 4004. 0004. .44000 .2.0 .0000 .04
0000. 0000. 0004. 0400. 0000. 4004. 0000. 000mb .2.0 .0000 .0
0444.: 0000.: 0000.: 0400. 4000.: 0040. 0000. .3000 .0002 .N.0 .0
0000. 0000. 0000. 0000. 0040. 0000. 4000. 40QM0> .0.0 .0
0000. 0000. 0004. 0004. 4440. 0404. 0000. 000 .00040 .0.0 .0
0000. 0404. 0004. 0000. 0004. 0000. 0004. 000 .400 .002 .0.0 .0
0440. 0004. 0004. 0040. 0000. 0000. 4040. 000 .400 .000 .0.0 .4
0040. 0000. 0000. 4040. 0400. 0000. 0040. 000 .0004 .000 .0.0 .0
0000. 0000. 0404. 0000. 0004. 0000. 0004. 000 0002 .0.0 .0
0404. 4004. 0040. 0400. 04mm. 4400. 04mm. 000 0m44mcm .m.0 .4
.0800 .4000 004000500 0E0HQ .0000 ..8000 .04800 .050 000000

 

 

 

.002040000 .0

04000000

154

0040. 4040. 0000. 0000. 0040. 0000. 0000. 0400. 40>04 .0000 .0000 .04

 

0000. 4000. 0400. 0000. 0040. 0000. 0000. 0400. 00000 .0000 .0000 .04
0000. 4040. 0400. 0000. 0000. 4400. 0004. 0000. .4 : .0.U.< .04
4000. 0000. 0040.: 0400. 0004.: 4040.: 0000. 4040. X00 .04
0000.: 4440.: 0040. 0400.: 0000.: 0040.: 0004.: 4040.: 00< .44
0044. 0400. 0040. 4040. 0000. 0000. 0000. 0004. .0 : .0.U.< .04
4400. 0000. 0000. 0000. 0440. 0000. 0400. 0000. 0040000 .000 .0000 .04
4044. 0000. 0004. 0000. 4000. 0000. 0000. 0004. .0002 .0000 .44
0000. 0044. 0000. 0004. 0400. 4004. 0004. 0004. .44000 .2.0 .0000 .04
0400. 0400. 0000. 0000. 0004. 4004. 4004. 0000. 00000 .2.0 .0000 .0
0000.: 0000. 0400. 0000. 0004. 0000. 0004.: 0400.: .3000 .0002 .0.0 .0
0040. 0400. 0040. 0000. 0440. 0000. 4000. 0000. 40QH0> .0.0 .0
0000. 0000. 0000. 0004. 0404. 0004. 4000. 0000. <00 .00040 .0.0 .0
0000. 0404. 0040. 4440. 0400. 0000. 4000. 0400. <00 .400 .002 .0.0 .0
4000. 0000. 0000. 4400. 0000. 0000. 0004. 0400. <00 .400 .000 .0.0 .4
0000. 0000. 0404. 0400. 0004. 4000. 0000. 0000. <00 .0004 .000 .0.0 .0
0000. 0000. 4040. 4040. 4040. 0000. 0000. 0040. <00 .0002 .0.0 .0
0000. 0004. 0000. 0040. 0000. 4040. 0440. 4000. <00 0044000 .0.0 .4
.0004 .00 504 0000040 0004000 .0000 00000000 .044

0800 :4000500 .000

 

 

 

.000040000 .0 0400000<

155

 

0044. 0400. 0000. 0040. 0000. 0000. 0004. 40>04 .0000 .0000 .04
4000. 0000. 4040. 0000. 0000. 0000. 0004. 00000 .0000 .0000 .04
@000. 0000. 0000. 0040. 0004. 0000. 0004. .4 : .0.0.< .04
0000. 0000. 0040. 0044. 0000. 4000.: 0000. x00 .04
¢Hao.- 0mmo.u owmv.: nmmo. 0000.: osmo.- hmoo.: 00¢ .00
0004. 0004. 0000. 0000. 4004. 0044. 0004. .0 : .0.0.< .04
0004. 0000. 0004. 0000. 4004. 4000. 0004. 0040500 .000 .0000 .04
0000. 0000. 0044. 0000. 0000. 0004. 0000. .0002 .0000 .44
0004. 0040. 0000. 0004. 0040. 0000. 0000. .44000 .2.0 .0000 .04
0000. 0040. 0000. 0000. 0000. 0000. 0000. 00000 .2.0 .0000 .0

0000. 0000.: 0004.: 0000. 0000. 0000. 0004.: .3000 .0002 .0.0 .0

0004. 0000. 0004. 0000. 0004. 4004. 0004. 40900> .N.0 .0

0000. 0004. 0404. 0000. 0000. 0004. 0000. <00 .00040 .0.0 .0

0000. 0000. 0004. 0040. 0000. 0404. 0000. <00 .400 .002 .0.0 .0

0000. 0000. 0004. 0000. 0000. 0004. 0040. <00 .400 .000 .0.0 .0

0000. 0000. 0004. @000. 0000. 0004. 0040. <00 .0004.000 .0.0 .m

0000. 0000. 0400. 0000. 0400. 00040 0000. <00 .0002 .0.0 .0

0000. 0000. 0000. 4000. 0040. 0004. 0000. <00 0044000 .0.0 .4
0040000 0000004400 >00EH0£0 004u4uusz 0040052 040:: £00:

 

 

 

.005040000

.0 x40c000<

156

 

mmmm. omwm. mfimm. mmma. omvm. wflmm. aw>wa .Ummm .mooo .ma
vomm. Nmmm. mmm¢. moma. momm. mwmm. Ummmm .wmmm .mooo .ma
Nomm. wmom. mmafi. mmmo. mmfim. mmmm. .a I .m.o.< .ma
mmoo. aflma. mmmo. mamo. mamo.u mwaa. xmm .ma
maoo. momo.l awoo.l omma.l mmmo.l ¢¢v0.I mm< .Va
mmwm. mama. ommm. @mmo.u mwmm. owam. .0 I .m.o.¢ .ma
Nomm. mmvm. ¢mvv. Nmmm. mamm. wmmm. mmaUSpm .00m .9000 .ma
momm. @Nma. amam. mama. mmam. mfimm. .numz .Qoov .aa
comm. mmam. aamm. aoam. mmam. mwmm. .aammm .z.o .Qoou .oa
moom. omvm. mmmm. momo. m¢am. Nmmm. mmmmb .E.o .moou .o
omao. mmmo.l aaoo.| mmma. mvvo. mmmo.| .BOQx .nomz .N.o .w
vomm. Nmom. wmww. mmma. mmmm. mfimm. amnum> .N.0 .m
mamm. momm. Nmmm. Nmma. mmom. omom. «m0 .uomam .m.m .@
mamw. mmmm. ammm. vao. aoov. moav. <mw .aom .umz .m.: .m
wamm. moon. mmmv. ommo. 000%. awmw. «mo .aom .oom .m.m .¢
ammm. m¢am. o¢¢m. mmma. mmmm. vmvm. mmw .mcmq .uom .m.: .m
mmav. mOOM. momm. ¢m¢a.| mao¢. mmmm. «mo sum: .m.m .m
ommm. moov. mmaw. mmma. mmom. mmov. amo smaamcm .m.m .a
mmOaOON Sommmm amanauom >9 ¢ oacmm wmoaono>mm .aom .aom

 

 

 

.©®5:apcoo .m xawcmmm<

APPENDIX C

UNIVERSITY OF MINNESOTA
OFFICE OF EDUCATIONAL RESEARCH
211 BURTON HALL

INTERPRETATION OF N. FATTUus NOMOGRAPH - January 21, 1946

Purpose:

To test hypothesis that two observed proportions, p1 and p ,
obtained from samples of size N and N2 are consistent wit
sampling from a common population.

On this hypothesis, pl - p2 is p1 - p = 0, hence wish to
test if observed difference is Significantly greater than
zero. The standard deviation of the difference (with zero
correlation between proportions) is

\

 

(pl - p2) pl p2 N N '.N N

l 2 l 2
Since p = p = p is unknown, it must be estimated in order
to evaluate. Theoretical considerations indicate that a
good estimate is the weighted mean of the two p's, or

 

 

 

 

p1 = N1 p1 + N2 p2
N1 + N2
The t-ratio becomes,
t pl ‘ p2 p1 ‘ p2
Nl pl + N2 P2 .l .l 1/2 (p + p )1/2 g
N + N N + N l 2 n
1 ‘ 2 1 2

By suitable algebraic manipulation of the formula it is

possible to make a chart with a fixed N1 and N2.

157

158

However, it is much easier to make a computing chart if one
uses Fisher's transformation of proportions, t = 2 arc

sine p as Zubinl did.

. . l
The standard error of the transformed proportion, t, 15 ‘WE;
hence the standard error for the difference becomes

 

(1"
l
H
m

(1.

[—1.

0
II
I

I prepared the Committee nonograph in 1939 because the two
that Zubin had were too small and too cumbersome for practical
use. Reference from one to the other also introduced another
possible source of error.

The tridk was accomplished by calibrating the scales p and
p in terms of the arc sine transformation units. From p1
and p it would therefore be possible to read T = t1 - t2
direcgly. (This is accomplished by joining p and p2 by
means of a hair line. The required value of T is given by
the reading at the point where the hair line crosses the

vertical T-scale. Similarly, connecting N1 and N2 by means

of a hair line gives the reading, D =JLL +-£ , at the
N1 N2

point where the hair line crosses the vertical D—scale.)

 

To facilitate the computation of tests of significance,
the T values were divided by 1.96 for the .05 level and 2.58
for the .01 level. A test of significance then is made by

comparing the D with the T values. If the values of D = T 05.

the difference is significant at .05 level. If the value of

 

lJ. Zubin, ”NOnographs for Determining the Signifi-
cance of the Differences between the Frequencies of Events
in Two Contrasted Series or Groups," Journal of American
Statistical Association, 34:539-544, 1939.

 

 

159

||/\

D T 01, the difference is significant at the .01 level.
To find the actual t-ratio multiply the T

and divide by D, or multiply the T
divide by D. °

05 value by 1.96

01 value by 2.58 and

.wpam Xa so a

a ma uomamu

sHOo

MH

.mmam am so a - e

 

 

 

 

 

ma mammnuomwn aasc powwow .mo. n ma .9 mommouo pa mHmQB muoz .mmwo unmamuum m cuaz mm
cam am pomccou .Q mommouo pa me£3 muo .wmpm unmamuum m £ua3 NZ ocm z uomccoo “mommo NZ
2 so comma Ammv usmoumm mnu spas mommo z :0 comma Aamv usmoumm map mo mocmummmao mcu ummu 09
N
O 5 O 5 O 5 0 no 0 no 0‘0 0 Au OnVOmw
2 2 3 3 4. 4 5 ,b .I go 9.0 a. O nXUnuO
l l 2 3451
2.
P
O ,3 O 5 O O O O O 0 AU :3 O 5 O
.1. l 2 3 4 5 6 7 8 8 9 9 O
l
O
6
l
1 no 9 8 _/ ,b 5 4 3 9. .i O 03 A5 7 .0 5 4.3 a2.l O .3
3 3 2 2 2 2 2 2 2 2 2 2 l l l l 1 1111.1, 0
D o o o o o o o o o o o c u o o o oo o u o 0
mm 6 5 4 a. .2 1i AU 1 n2 .3 4 5 no
2 o o o o o o o o o o o e
T -
0/0 o o o o o o
5
m :3 0 :3 O :3 nu O O O nonu 0 AU 00.0Mw
2 3 3 4 4 5 6 7 8 90 5 O 0000
l l 2 34R...U
N1
1i0 :3 O .5 no 0 O O O O O ,3
O 5
pm 9 9 8 8 _/ r0 5 4 3 2 l l O
_