University of Southern California Dissertations and Theses

Prediction Of Success In Training Among Electronics Technicians

(USC Thesis Other)

Prediction Of Success In Training Among Electronics Technicians

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content This dissertation has been 6 2—6040
microfilmed exactly as received
BEOE, John Richard, 1926-
PREDICTION OP SUCCESS IN TRAINING AMONG
ELECTRONICS TECHNICIANS.
University of Southern California, Ph.D., 1962
Education, psychology
U n iversity M icrofilm s, Inc., A n n A rbor, M ich igan
PREDICTION OF SUCCESS IN TRAINING
AMONG ELECTRONICS TECHNICIANS
by
John Richard Broe
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(Education)
June 1962
UNIVERSITY O F S O U T H E R N CA LIFO R N IA
G R A D U A TE SCHOO L.
U N IV E R S IT Y P A R K
L O S A N G E L E S 7, C A L IF O R N IA
This dissertation, written by
JOHN RICHARD BROE
under the direction of h.^S....Dissertation Com
mittee, and approved by all its members, has
been presented to and accepted by the Dean of
the Graduate School, in partial fulfillment of
requirements for the degree of
D O C T O R O F P H I L O S O P H Y
t D ate— —V
Dean
-DISS SSERTATION C
^Udeli
.OM M ITT EE
i..
ACKNOWLEDGMENTS
Acknowledgment is hereby gratefully made to
Dr. Floyd L. Ruch, co-author of the Employee Apt
tude Survey Tests, who very generously arranged
for Psychological Services, Inc., of Los Angeles
to furnish research test materials to this inves
tigator, as well as giving of his time in advise
ment of this project; to Western Data Processing
Center, Graduate School of Business Administra
tion, University of California at Los Angeles,
for the use of machine facilities and program
information in the analysis of the data; and to
Dr. D. Welty Lefever, for his extensive efforts
in the initial organization and direction of the
dissertation.
TABLE OF CONTENTS
Page
ACKNOWLEDGMENTS.................................... ii
LIST OF TABLES . , ............................... v
LIST OF FIGURES . . '................................ vii
Chapter
I. THE PROBLEM................................ 1
Introduction
Statement of the Problem
The Hypotheses
Organization of the Remainder of the
Dissertation
II. REVIEW OF THE LITERATURE ........... 13
Introduction
Need for More Investigation
Studies Related to Training Electronics
Technicians
Predictive Tools
Summary
III. METHOD OF PROCEDURE............. 53
The Sample
The Criteria
Predictor Variables
Method of Data Analysis
Summary
Chapter Page
IV ANALYSIS OF DATA AND FINDINGS 96
Correlation Analysis of Single Variables
Multiple Correlation Analysis
Analysis of Sample Differences
Summary
General Summary
Summary of Biographical Data
Summary of the Influence of Motivational
Characteristics
Summary of the Influence of SORT Variables
Summary of the Influence of EAS Test
Variables
Summary of the Data Related to the
Hypotheses
Description of the Sample
Predictive Capacity of Single Variables
Predictive Capacity of Characteristics
of Motivation
Predictive Capacity of Multiple Variables
Implications of the Investigation
Need for Further Study
V. SUMMARY 155
VI CONCLUSIONS 175
APPENDIXES 193
BIBLIOGRAPHY 214
LIST OF TABLES
Table Page
1. Sample Breakdown by Colleges and Training
Levels....................................... 57
2. Factorial Composition of the EAS Tests .... 73
3. Correlations between Predictors and Grade-
Received Criterion.......................... 98
4. Correlation between Predictors and the Sum-of-
Ratings Criterion . ........................ 101
5. Correlation between Predictors and Composite
Criterion..................................... 104
6. Correlation Matrix of Motivation Ratings to
Three Criteria via Total and Sub Samples . . 107
7. Intercorrelation Matrix of Criteria, Total
Sample N=176 . Ill
8. Intercorrelation Matrix of Ability Variables,
Total Sample N = 1 7 6 .......................... 113
9. Intercorrelation Matrix between Certain
Predictors.................................. 116
10. Intercorrelation Matrix between Certain Person
ality Predictors from the SORT, Total
Sample N=176 119
11. Intercorrelation Matrix of Personality Pre
dictors from the SORT, Total Sample N=176 . , 120
v
Table Page
12. Multiple Correlation Analysis Approach One,
First Semester N=56.......................... 123
13. Multiple Correlation Analysis Approach One,
First Year N = 1 1 4 ............................ 124
14. Multiple Correlation Analysis Approach One,
Second Year N = 6 2 ............................ 125
15. Multiple Correlation Analysis Approach Two,
First Semester N-56.............. 129
16. Multiple Correlation Analysis Approach Two,
First Year N = 1 1 4 ........................... 130
17. Multiple Correlation Analysis Approach Two,
Second Year N = 6 2 ............................ 131
18. Multiple Correlation Analysis Approach Three,
Total Sample N = 1 7 6 ....................... . 134
19. Multiple Correlation Analysis Approach Three,
First Year N = 1 1 4 ............................ 136
20. Analysis of Criteria Differences between High
and Low Achievers............................ 140
21. Analysis of Predictor Differences between High
and Low Achievers . ....................... 141
22. Analysis of Sample Differences between Schools 145
23. Raw Score Norm Comparisons between Samples . . 149
V? tr
24. EAS Test Score Comparisons between 176 Elec
tronics Trainees and 86 First Year Engineer
ing Students from University of Southern
California. Suggested Cutting Scores Recom
mended and Percentage of Predictive Effi
ciency as Suggested by F o y ................. 153
25. Projected Occupations of Total Sample ........ 164
LIST OF FIGURES
Figure ' Page
I. Expected Grade for Total and Sub-samples . . 163
CHAPTER I
THE PROBLEM
Introduction
The need today for qualified electronics techni
cians is accepted as a definite reality. The rapid growth
of the entire field of electronics in the decade of the
fifties has created acute shortages in personnel.
The tremendous industry growth is pointed up in an
article in Financial World for October 11, 1961, which
stated:
Total [electronics] industry volume rose at an
average annual rate of about 16% between 1950 and
1958, but since then has dropped in half, amounting
to only about 8% during the past: two and a half
years. (32)
And, as The Magazine of Wall Street reported on December
16, 1961:
In a talk before the New York Security Anal
ysts on November 1, Mr. Richard E. Krafve, Presi
dent of Raytheon, stated: ". . . The industry will
continue to grow, certainly, but at a rate sub
stantially less than in the past ten years. We
1
2
estimate future growth at a rate of 7 or 8% each
year, which will give us an industry volume of
between $18 and $20 billion by 1970 . . (35)
Conversely, the acute shortages in trained indus
trial electronics technicians has been pointed out in a
recent survey by Ryan and Allen. Commenting on the re
sults of their survey, they state:
The companies emphasized that they are critic
ally short of employees who have an overall skill,
ability and related knowledge applicable to the
electronics technician occupation.
Companies are not interested in electronics
technicians who have ability in one operation.
Only in the case of a dire national emergency,
when unit cost is not a factor, can jobs be broken
down into many small operations which permit the
use of operators. The outlook for the future is
for the well-trained person. (34)
Pointing up the shortage of technicians, as re
cently as March, 1960, a study of trade and industrial stu
dents in California reported that 16 per cent of entering
college electronics trainees failed to complete the re
quirements necessary for graduation from the two-year
course (10). Another study estimated that fully 60 per
cent of the entering two-year college electronics trainees
failed to complete their fourth semester of training (9).
The situation in the electronics industry and the
shortage of electronics technicians as stated apply to the
United States as a whole. In Southern California, however
the foregoing facts are multiplied many-fold, since the
concentration of companies in the electronics field is
greatest in this area. After indicating that the electron
ics industry is spiraling upward, Ryan and Allen pointed
out that in 1949 there were approximately 43 manufacturing
firms in the Los Angeles area concerned with electronics,
with sales of approximately 97 million dollars, and in
1959, ten years later, there were 461 manufacturing plants
in the same area with sales of 1,157 million dollars (34).
Throughout the western states, electronics has become a
major industry, with 17.8 per cent of all electronics
firms in the United States and 19 per cent of all elec
tronics personnel located here. Relative size of these
western companies may be seen in the fact that they ac
count for 23 per cent of all United States electronics
sales (55).
Yet, recalling studies on the drop out rate of
electronics trainees in California junior colleges, in
structors and administrators alike are concerned because
human training resources and facilities for training are
not being fully utilized. They feel that some percentage
of the drop-out rate is the natural result of trainee
frustration and maladjustment when failure and reorienta
tion of vocational goals occur. These college personnel
recognize, to be sure, that not all drop outs are due en
tirely to failure in training but that many of them may
also occur as a result of a host of personal reasons, in
cluding the attraction of immediate employment. These ad
ministrators and instructors who are responsible for the
education of electronics trainees are vitally interested
in the development and testing of selective devices and
guidance procedures in order that they may more efficiently
discuss the specialized training in electronics with stu
dents considering enrollment in such courses.
Statement of the Problem
The problem under investigation, then, is a con
sideration of and an attempt at a solution of the diffi-
■ \ .
culties electronics trainees have in successfully complet
ing college training program. Guidance, based on proven
predictors of success-in-training, seemed to be a key to
solve the problem, but in a survey of this area the inves
tigator was unable to find a single electronics instructor
or director of technical studies who felt that the guidance
and selection of electronics trainees in public colleges
were even close to achieving the results needed.
Therefore, in view of the tremendous current need,
and the potentially even greater need in the future, for
trained electronics technicians, this investigation was
instigated in the hope of being able to offer assistance
toward answering the following questions.
1. What characteristics have predictive value in
the guidance of students into, or away from,
electronics training programs?
2. How may comprehensive information for guidance
personnel be provided regarding the beneficial
or detrimental characteristics of full-time
day electronics trainees enrolled in public
junior colleges in Southern California?
3. What information can be provided which will be
helpful in planning the instruction of elec
tronics technicians in public colleges?
4. What criteria can be developed which may be
helpful in identifying trainees who are more
likely to be satisfactorily and successfully
employed as technicians?
More specifically, it was considered desirable to
learn the average personal characteristics of students cur
rently enrolled in electronics training and the deviation
of these characteristics. By computing the relationships
between these characteristics and their currently-achieved
success in their electronics courses, a set of predictors
might be established which would be of value to guidance
counsellors.
Elements against which predictor variables would
be measured fall, in general, into four broad categories:
(1) individual aptitudes in certain specific areas, (2)
background or biographical data, (3) motivational charac
teristics, and (4) personality characteristics or charac
teristics of thinking.
Two well-known and highly-regarded batteries of
tests for measuring aptitude were used: (1) the Employee
Aptitude Survey and (2) the School and College Ability
Tests.
Some of the areas of background or biographical
level which would be measured by predictor variables were:
1. Age
2. Units in progress
3. Completed college units
4. High school grade point average
5. Extent of employment experience
6. Grade expected by trainees in their current
electronics courses
7. Occupation anticipated ten years hence
8. Extent of prior training in electronics
9. Occupational level of fathers of trainees.
Personality characteristics were compiled through
the use of a standard personality test, the Structured-
Objective Rorschach Test, in order to compile a broad
range of interests, mental functioning characteristics,
responsiveness, and temperament.
In addition to discovering the extent and signifi
cance of relationships between these personal character
istics and success in training, this investigation will
suggest that, given sufficient levels of capactiy and op
portunity to learn, motivation and personality character
istics are of great importance in the degree of success
experienced by students in general. It was hoped, there
fore, that data collected from the electronics trainee
sample from interviews could be rated and classified as to
Source of Motivation, Intensity of Motivation, Extent of
Planning, and Manifest Interest.
Another positive approach was an attempt to dis
cover the degree of relationship which exists between such
ratings of motivation and training success. By separating
the students in electronics courses into two groups--those
with the better grades being called "high achievers" and
those with the lower grades being called "low achievers"--
it was hypothesized that a number of predictor variables
could be found which would be related significantly to the
success or failure of students in electronics training.
The manner in which personal characteristics differentiate
between these high and low achievers was also scheduled to
be explored.
If some of the predictors under investigation in
this study were to prove valuable in portending the suc
cess of a candidate for electronics course work, it is
conceivable that college admissions counsellors could use
these predictors in guiding potential trainees into elec
tronics courses where they would be successfully trained
to take their places in an industry which admittedly needs
many more employees. Conversely, if the predictors to be
explored and evaluated here proved valuable in screening
f
out candidates for electronics courses whose ability,
personal background, and temperament indicated that they
were unlikely to be successful in the field of electronics,
guidance counsellors would have a tool upon which they
could rely as they advised potential enrollees.
The Hypotheses
In order to explicate the specific areas under
consideration in this investigation, the following hypoth
eses are set forth:
1. Significantly higher scholastic aptitude scores
will be manifested by the high achievers as
compared to the scores of the low achievers.
2. Significantly higher scores will be manifested
by the high achievers as compared to the scores
of the low achievers in specific Employee Ap
titude Survey Tests, namely:
a. EAS #1 - Verbal Comprehension
b. EAS #2 - Numerical Ability
c. EAS #5 - Space Visualization
■ d. EAS #6 - Numerical Reasoning
e. EAS #10 - Symbolic Reasoning
3. Significantly higher high school grade point
averages will be manifested by high achievers
as compared to low achievers.
4. Significant differences will not exist between
high and low achievers on the basis of other
specific tests in the Employee Aptitude Survey
battery of tests, namely:
a. EAS #3 - Visual Pursuit
b. EAS #4 - Visual Speed and Accuracy
c. EAS #7 - Verbal Reasoning
d. EAS #8 - Word Fluency
e. EAS #9 - Manual Speed and Accuracy
Significant differences will not exist between
high and low achievers on the basis of 25 var
iables taked from the results of the Structured-
Objective Rorschach Test.
Significant differences will not exist in the
accuracy of the relationship between the grade
expected and the grade actually received in
electronics courses between high achievers and
low achievers.
Significant differences will not exist in the
ratings of the Father's Occupational Level
between high achievers and low achievers.
11
Organization of the Remainder
of the Dissertation
The remainder of this dissertation is organized
as follows:
Chapter II will review the literature which is re
lated to predicting success in electronics trainees in
various settings, prediction studies of success for related
curricular groups, appropriate studies which relate to the
use of aptitude tests, and finally personality tests and
biographical characteristics which have been used in the
prediction of success in schools.
Chapter III will describe the research procedures,
the sample used in this investigation, the criteria, and
the predictors which were studied. In addition, the
statistical analysis which was conducted is described.
Chapter IV will contain the findings and results
of the correlation analysis. This material will be pre
sented in the form of correlations for single variables,
of multiple correlations and analyses of the results of
the combinations, tests of sample differences, and compar
isons of appropriate mean data.
Chapter V will summarize the findings as they
relate to the questions raised and the hypotheses stated
by this investigation. Conclusions are drawn and recom
mendations set down for further study in Chapter VI.
CHAPTER II
REVIEW OF THE LITERATURE
Introduction
A review of the literature relating to predictions
of success-in-training electronics technicians reveals a
paucity of studies either in the research covering the
growing need for students in this field or in the research
covering the guidance of students into or away from this
field. There are few studies which are closely related
and which were conducted on the extensive training of
electronics technicians for the armed services. These are
reviewed in this chapter. Two studies were discovered
which related to technicians in general, and these will be
reviewed insofar as they relate to the problems investi
gated in this study.
In addition, there is an abundance of research
available on the prediction of success in the training of
the closely related profession of engineering. These
13
. 14
studies, particularly those sections of the studies which
are of greatest pertinence, are covered here.
General studies of prediction of success in train
ing and/or in college are in greater abundance. Some of
these have touched upon areas which are readily adaptable
to predictions of success in training electronics techni
cians; other research has gone deeply into certain specific
areas. In either case, research of this nature is reviewed
here briefly in an effort to delineate the applicable work
which has already been done and to point out the wide
areas in which additional research and study would be of
great value.
Heed for More Investigation
The barrenness of specific research relating to
the electronics technician, either on-the-job or in-train
ing, is no doubt the result of the technological changes
which have brought about, in a relatively short time, the
entirely new occupation of electronics technicians.
In a January 24, 1962, newspaper article covering
the monetary aspects of the field, financial columnist
R.L.W. wrote:
The whole electronics industry has expanded in
10 years to almost unbelievable proporations in the
15
consumer, industrial and military markets. Factory
sales of $10.8 billion compared to an even $10 bil
lion in 1961 are predicted for the year 1962. (71)
He stated further that the Electronic Industries Associa
tion says that large and rising expenditures on research,
development, and new products are creating additional com
petition, and the implication in terms of this investiga
tion is that manpower requirements, especially for trained
electronics technicians, already tremendously expanded,
give no indication of abating.
The shortage of trained technicians in electronics
has been evident for most of the past decade. Prior to
World War II, there were workers in a similar field who
were called radio repairmen or radio technicians, but the
complexity of their work was limited by the embryonic state
of the materials with which they worked. With the onset
of more advanced technology, the growing demand for com
puters, the stepped-up race with the Soviet Union for
space superiority, and the concomitant related fields,
American industry created an expanding need for trained
personnel. In 1957, the National Association of Manufac
turers, in a bulletin entitled "Your Opportunities in
Industry as a Technician," commented on the growth of
technology and research and pointed out that the nation
was spending at the rate of four billion dollars annually
for research and development (61).
By 1959, the shortage of trained personnel was al
ready so acute that a few studies were being conducted to
point out the need for additional trained technicians. For
example, in a book entitled The Manpower Future: Its Chal
lenges to Vocational Guidance, Helen Wood projected a 60
per cent increase in the required number of trained tech
nicians by 1970. She went on to prophesy that fewer jobs
would be available at that time to the unskilled or semi
skilled (98) . That same year, the State of California
issued several studies covering the same general subject,
one reason for California's concern being the increasing
concentration of electronics manufacturers in that state.
One of the California State Department of Education pub
lications titled A Study of Technical Education in Cali
fornia: Guidelines for the Development and Operation of
Technical Education Programs in the Junior Colleges re
viewed the serious and growing shortage of adequately-
trained technicians. This study stated that in 1959, be
cause of the phenomenal expansion of industry and the many
< / ■
new scientific developments, the growing recognition of
the need for adequately qualified electronics technicians
17
is great. Unfortunately, as the Department further com
mented, there is a serious lack of adequate facilities for
training electronics technicians (8).
Concurrent with the dearth of a sufficient number
of electronics trainees was a dearth of predictive studies
on vocational guidance in this field. As far back as
1956, C. H. Patterson pointed out this fact in an article
for the Journal of Applied Psychology entitled "The Pre
diction of Attrition in Trade School Courses." At that
time, he indicated that four out of five youths between
the ages of 17 and 21 were in formal training of a voca
tional nature, as compared to the one out of five enrolled
in academic programs. In spite of the pressing need for
guidance of these students, Patterson reported that almost
no guidance literature was available in the electronics
field, and the vocational counselors who were advising the
potential trainees were not in a position to do research
(63).
Consequently, as Patterson also mentioned, tests
of general intelligence were being used as late as 1956,
tests which bore but slight relationship to training
achievement, tests which were certainly not effective to
the extent that the more specific aptitude tests in other
18
fields related to potential training success (60).
Studies Related to Training
Electronics Technicians
General studies
One of the earlier studies, and even today one of
the most comprehensive, was a correlational analysis of
achievement in a generalized electronics troubleshooting
course in Los Angeles authorized by the Air Research and
Development Command at Lockland Air Force Base, San
Antonio, Texas, released in 1957. The authors, Warren,
Dossett, and Ford, reported this experimental study of 90
high school electronics trainees which used the Employee
Aptitude Survey Tests (EAS Tests), along with a number of
other variables in predicting success in electronics
training.
Because Warren, Dossett, and Ford's experimental
study provided a partial background upon which the pro
cedures and hypotheses of the present investigation rest,
their study will be summarized here briefly. They first
established training levels, which were defined in terms
of the number of practice problems presented to the train
ees. The criteria of training success were measured in
19
terms of scores from four tests: (1) a paper and pencil
achievement survey, (2) a performance efficiency score
based on number of actions requared to solve the perform
ance problems, (3).a performance time score, and (4) a
composite measure of performance, which was a function of
the performance time score and an inverse function of the
performance efficiency score. The results of this multi-
ple-correlation analysis indicated that when verbal com
prehension (EAS #1), space visualization (EAS #5), and
symbolic reasoning (EAS #10), had a multiple-R of .80 with
the written achievement criterion, and then visual pursuit
(EAS #3) was added to the multiple-R, the coefficient was
raised to .82. However, as the experimenters commented,
this addition was not worth the increased efficiency inas-
much as R , the coefficient of determination, was increased
only from ,645 to .669. The coefficients of correlation
between individual Employee Aptitude Survey tests and
written achievement criterion scores were as follows:
Verbal Comprehension to criterion .57
Numerical Ability to criterion .61
Visual Pursuit to criterion .52
Visual Speed and Accuracy to criterion .35
Space Visualization to criterion .56
20
Numerical Reasoning to criterion , .63
Verbal Comprehension to criterion . .65
Word Fluency to criterion ,28
Symbolic Reasoning to criterion .69
The authors concluded that regardless of the criterion
employed in predicting training success, Verbal Reasoning
and/or Symbolic Reasoning tests from the Employee Aptitude
Survey tests were somewhat predictive of training suc
cess (94) .
In spite of the background basis of the Warren,
Dossett, and Ford experimental study for this investiga
tion, there are fundamental differences: (1) The nature of
the trainees was apparently different in many respects:
age, previous experience, and characteristics of motiva
tion, to name a few. (2) Measures of success in training
were certainly different; Warren, Dossett, and FordTs
study employed a much more "academic type" of criteria
measures than are proposed in this investigation. (3) The
nature of the training experience itself was, of course,
another fundamental difference, inasmuch as Warren, Dos
sett, and Ford essentially used artifacts of training,
whereas the present investigation proposes a broader range
21
of training objectives, with electronics trainees studied
*
in typical college learning situations.
Perhaps of much closer approximation to the char
acteristics of the present investigation was an unpublished
study of the selection of technical students at Glendale
College, conducted by the Glendale Unified School District
in 1960 and known as the "Glendale Study." It proposed to
provide data upon which counselors could base their advice
in helping students select technical programs, among which
were technical illustration, electronics, drafting, machine
shop, and aviation airframe and engine programs. In all,
215 students from the Technical Arts Department of Glendale
College were given the Employee Aptitude Survey tests, and
grades in eight major unit courses were then used as the
criterion of training success. For the sample of 69 elec
tronics trainees in the Glendale study, moderately high
significant coefficients were obtained between Verbal Com
prehension (EAS #1), Numerical Ability (EAS #2), Numerical
Reasoning (EAS #6), and Symbolic Reasoning (EAS #10) and
the criterion of the grade received in major electronics
courses. These coefficients were .32, .58, .45, and .48
respectively--all significant at or beyond the .01 level
of confidence. Space Visualization (EAS #5) correlated
22
with the criterion at .24, which was at the .05 level of
confidence (33).
The criterion in the Glendale Study of the grades
received in the eight courses was identical to the Grade
Received Criterion, which was used as one of the criteria
in the present investigation. Additionally, the findings
in the Glendale Study were similar to those found by the
present investigation in studying a sample of 176 trainees
from three similar public junior colleges.
Prior to the present investigation, these two
studies were the only ones discovered by this researcher
which specifically predicted the potential success of
civilian electronics trainees enrolled in college programs.
Studies of Armed Services schools
As mentioned earlier, there have been a number of
studies of electronics trainees enrolled in Armed Services
training courses. Unfortunately, the results of these
studies are not directly applicable to the present inves
tigation because of the prior selection of trainees for
training and because of significant differences in the
nature of the training programs. However, the results of
these studies are significant in themselves as they have
23
been carried out under standard conditions with relatively
large samples of trainees.
Birnbaum, et al., reported in 1955 a predictive
study of 1,019 enlisted men enrolled in three radio repair
courses taught by the United States Army Signal School.
Composites of the results of three tests from the Army
Classification Battery--(l) Electrical Test, (2) Radio
Information Test, and (3) Arithmetic Reasoning Test--
yielded the most promising prediction of course success
(2). The authors reported that cross-validity coefficients
ranged from .66 to ,85 (2).
In 1958, in a similar study at the United States
Army Signal School, Helme reported on the prediction of
success in course training for electronics and electrical
maintenance jobs. Again using the Army Classification
Battery of Tests, Helme found that the Mechanical Aptitude
Test plus the Electrical and Radio Information Test
yielded the best multiple-R*s with course grades--between
.46 and .83 (42). Using the same predictors from the Army
Classification Battery--the Electrical Test and the Radio
Information Test, plus the Mechanical Aptitude Test--Helme
had reported a year earlier that the prediction of success
of trainees in carrier repair, teletype maintenance, and
24
power equipment courses yielded a multiple-R of .81 against
course grades (43).
The amount of prior education as it related to
success in training was tested on a large sample of men in
Air Force Training Schools by Zachert and Levine in 1952.
Their findings showed that the amount of prior education
was highly related to success in training (99). This find
ing, however, may be somewhat misleading because these
investigators found that when results of the Army General
Classification Tests were combined with high school grade-
point averages, the additional computations using this
prior education predictor contributed little to the mul
tiple prediction of training success (99).
These studies of Armed Services personnel were of
interest to this investigation despite the differences in
the conditions, especially in the area of the population
which made up the fields of testing. The principle find
ings which were of value to this investigator were the
high correlations of the Birnbaum, et al., and the two
Helme studies of high school grade point average with
course grades.
25
Related fields of study
Although not directly related to this investiga
tion, studies which have been conducted in technical arts
or vocational areas, as well as some of the many studies
concerned with engineering students in college, were re
viewed and are reported here.
Technical arts or vocational training.--Four stud
ies which used different test batteries were checked, since
their findings were of considerable interest to this proj
ect .
An unpublished study by Hugh Halsey, entitled
"The Predictive Value of Certain Measures Used in Selecting
Freshmen for the Technical Curricula in a Community Col
lege," was made in 1957. The differential Aptitude Test
battery, high school grade-point average, and achievement
scores on an instructor-designed mathematics test were
related to freshman-year grades. Using the Wherry-Doo-
little selection method to arrive at his multiple correla
tion, Halsey found that high school grade-point average,
DAT Numerical Ability, DAT Sentences, and DAT Mechanical
Reasoning provided the highest multiple-R with the cri
terion--.601 (40). In a further analysis of his findings,
26
Halsey reported that trainees in a technical curriculum
who had taken advanced algebra and plane geometry in high -
school were not differentiated from trainees who had taken
only elementary algebra in high school with respect to
potential success in college (40).
In an extensive study of the attrition in trade
school courses, C. H. Patterson in 1956 reported his find
ings from several different batteries of tests. With a
sample of 350 students from a large midwestem private
training school, Patterson compared the 224 trainees who
remained with the 126 trainees who had dropped out within
a six months period. Employing first a battery of bio
graphical predictor variables, Patterson reported that
very few of the eighteen variables studied differentiated
between the groups. His results indicated a tendency for
the remaining trainees to be older (especially between
ages 20 and 30 years), for them to have had more prior
work experience, and for them to have had several shop,
science, and mathematics courses in high school, but he
found much overlapping in test scores between the two
groups. Using secondly the combined results of the Army
General Classification Test (AGCT), the Bennett Mechanical
Comprehension Test, and the Revised Minnesota Paper Form
27
Board Test, Patterson stated that it was necessary to
eliminate 21 to 29 per cent of the passing or successful
group if a sufficiently high per cent (46 to 50 per cent)
of the potential failure, or drop out, group were to be
eliminated. He stated, however, that from these combined
test scores, continuance in training could be predicted
-.significantly better than by chance. Although Patterson's
results were interesting, they fail to offer vocational
counselors much reassurance as they make recommendations
to individual trainees (63).
Emphasizing chiefly the Spatial Relations Ability
test, Form A, Case and Ruch studied a small sample of
elementary electric shop trainees and advanced shop train
ees in 1944. They found significant, low-positive coef
ficients of correlation between instructors' ratings of
quality and quantity of shop achievement and the results
from the Spacial Relations Ability test (14).
Correlating instructors' ratings to mechanical
trainees' aptitude for learning, quality and precision,
and interest and enthusiasm for mechanical course work,
McDaniel and Reynolds utilized the McQuarrie and O'Rourke
Mechanical Aptitude Tests, and discovered a multiple-R of
.47 (58).
28
Unfortunately all four of the preceding studies
in this section are several years old and both the tests
which are available and the purposes of the tests in guid
ing potential electronics trainees have advanced consider
ably since the investigations were conducted. Later, more
up-to-date information is sadly lacking and at the same
time vitally necessary.
Engineering training in college.--The early re
searchers on predictions of success of engineering students
in college generally used the American Council Psycholog
ical Examination (ACE) tests. As early as 1940, Remmer
and Grieger reported on the prediction of success of en
gineering students at Purdue University. Their findings
showed coefficients of correlation as high as .62 between
grades of first and second semester students at Purdue
with scores on the ACE (69). In 1951, Berdie combined
results of the ACE battery with the Differential Aptitude
Test battery in studying 472 engineering students at the
University of Minnesota. He discovered that the combina
tion of the ACE battery of tests with the DAT battery con
tributed little, if anything, to the multiple prediction
of grade achievement (1). Berdie’s findings were sup
29
ported by Coleman, who reported in 1953 that the scores
on the ACE did not add significantly to the results of the
Cooperative Algebra, Cooperative English, and Bennett
Mechanical Comprehension tests as correlated against course
grades (18).
The effectiveness of four measures--high school
grade average, General Mathematical Test, Scientific Verbal
Ability Test, and Comprehension of Scientific Materials
Tests--in predicting a weighted four-year grade point aver
age was studied by Long and Perry in 1953. When they
studied a large sample of 433 engineering school graduates,
they found correlations between the criterion and the pre
dictors which varied from .30 to .50 (54). They also re
ported a correlation for the composite score from measures
with the criterion of .53 (54). According to Long and
Perry, measures of interests and data from personal ques
tionnaires were not sufficiently significant to justify
their inclusion in the final selection battery (54).
In the review of the literature in this area, it
was apparent that the researchers mentioned felt that high
school grade-point average and then achievement tests
(especially in mathematics) and specific aptitude tests
were, in that order, related to success in training of
engineers. In an American Council on Education study in
1949, Stuit, Dickson, and Jordan reviewed the literature
extensively to that time on the relationship of high school
grade-point average to engineering achievement and obtained
an average correlation of .55 (83). Five years later, in
1954, W., L. Layton again conducted an extensive review of
the literature. He reported that mechanical and spatial
aptitude tests correlate about .35 with grades in engin
eering, science aptitude tests correlate about .45 with
grades, and mathematics aptitude tests correlate about .50
with grades (52).
Perhaps the study of engineering students most
pertinent to the present investigation was an unpublished
doctoral thesis by Foy in 1959 at the University of South
ern California. When the grades of engineering students
from the entire first year of course work were used as the
criterion, Foy determined the validity and cross-validity
coefficients of correlation between scores on Employee
Aptitude Survey tests and grade achievement for two separ
ate samples, one containing 86 and the second containing
139 engineering students (27). Foy reported that the va
lidity coefficients ranged between a low coefficient of
.18 and a high correlation of .63. The Verbal Reasoning
test produced the weakest relationship with grades, while
Numerical Ability had the highest. According to Foy, the
best combination of Employee Aptitude Survey test results
included Verbals Comprehension, Numerical Ability, Space
Visualization, Numerican Reasoning, and Symbolic Reason
ing. The multiple-R for the first sample of 86 engineering
students was .87; when cross-validated against the second
sample of 139 engineering students, the multiple-R was
reduced to .77 against first-year grades. However, this
cross-validation multiple-R accounted for 75 per cent of
the variance of the criterion (27). Foy also determined
the most effective cutoff scores and pointed out the re
sulting efficiency therefrom, as these scores applied to
first year engineering students at the University of
Southern California.
Each of the above studies dealing with prediction
of success of engineering students utilized batteries of
tests or multiple predictors in correlation with grades to
indicate success. These results, therefore, confirm the
findings indicated by Patterson that there has been a
definite trend toward the use of a factored battery of
tests rather than a single test (64).
32
Predictive Tools
It would seem appropriate at this point, in the
review of the literature, to look at some of the studies
which support the feasibility and validity of using the
predictor variables which are proposed by the present
investigation.
Aptitude tests
Many different batteries of tests have been de
veloped through the years to test a wide variety of
abilities. Some of these might have been applicable to
the present investigation and a large number would not
have been applicable. Out of the wide choice of test bat
teries available to this investigator, the Employee Apti
tude Survey tests and the School and College Ability Tests
are the two series which appeared to have the greatest
meaning for this investigation.
Employee Aptitude Survey tests.--Foy used the
Employee Aptitude Survey (EAS) tests in 1959 in his unpub
lished doctoral thesis entitled "A Study of the Relation
ship between Certain Factor-Analyzed Ability Measures and
Success in College Engineering.” (27). This test battery
was also used by Warren, Dossett, and Ford in 1957 (94),
33
and by the Glendale Unified School District in 1960 (33).
Psychological Services, Incorporated, publishers
of the Employee Aptitude Survey tests, suggests that the
nature and content of the test items provides a face
validity for many of the tests. For example, an electronic
trainee should recognize immediately the reading of blue
prints of electronics schematics as related to test items
in the Visual Pursuit test. Similarly, the figures used
in the Symbolic Reasoning test are the symbols often used
in technical training programs. The titles of the tests
give a clue to their contents, and are as follows:
Verbal Comprehension
Numerical Ability
Visual Pursuit
Visual Speed and Accuracy
Space Visualization
Numerical Reasoning
Verbal Reasoning
Word Fluency
Manual Speed and Accuracy
Symbolic Reasoning
A. sufficient number of studies have been com
pleted, both in industry and in educational institutions,
to suggest that these tests offer a real opportunity in
combination with measures of personality and personal back
ground data to predict success in electronics training.
By way of describing the advantages and the validity of
the EAS tests, the publishers themselves have reported
some of the studies of current validity. For example,
Psychological Services, Incorporated states that the
Numerical Ability test was correlated with instructors'
ratings of achievement in an intensive 16-week basic
electronics course at Lockheed Aircraft Corporation, pro
ducing a coefficient of .65. In fact, according to the
publisher, the Numerical Ability test has proved of con
siderable value to Lockheed in predicting training success
in a number of different technical training courses, such
as engineering drafting, and the Lockheed junior engineer
ing training program. The Visual Pursuit test has also
demonstrated a significant relationship to training course
achievement— specifically training success in electronics--
at Lockheed, with validity coefficients ranging from .46
to .58. Also at Lockheed, the Numerical Reasoning test,
when correlated with instructors' ratings of electronics
trainee success, evidenced a validity coefficient of .69
(68).
35
In summarizing validity studies and the feasibility
of using the Employee Aptitude Survey tests, it seemed to
this investigator that these tests have, in varying degrees,
moderately high relationships with various criteria of
training success in a number of different technical-type
courses.
School and College Ability Tests.--The School and
College Ability Tests (SCAT) are now being used extensively
by many colleges for various functions in student person
nel work. It seemed desirable, therefore, to include
results of the School and College Ability Tests as part of
this investigation.
Three separate test results comprise the School
and College Ability Test--Verbal (V) score, Quantitative
(Q) score, and Total (T) score. In a 1958 Test Supplement,
the Educational Testing Service, publishers of SCAT, re
ported some of the validity studies of this test. For
example, the Educational Testing Service reported validity
coefficients between college achievement, average grades,
and the School and College Ability Test results of between
.33 and .57. In general, Total and Verbal scores were
36
higher in relationship to college achievement than the
Quantitative scores (24).
The publishers1 stated results were corroborated
in 1958 by Kennedy in an article in the California Journal
of Educational Research (49). She reported validity coef
ficients for SCAT-V and SCAT-T scores with course grades
in various curricula of between .52 and .79. Kennedy also
found SCAT-Q scores to be much lower in correlation with
academic success, with correlations in the .20's (49).
Another confirmation came in 1959 from Klugh and
Bierley in an article in Educational and Psychological
Measurement entitled "The School and College Ability Test
and High School Grades as Predictors of College Achieve
ment" (51). Klugh and Bierley studied groups of 106 male
students and 125 male students at Alma College by corre
lating SCAT results with average freshmen grades and found
validity coefficients of .68 and .81 respectively for
SCAT-V and SCAT-Q (51).
Although the School and College Ability Test is a
test of relatively recent origin, the success it has
achieved as a predictor of success in college courses has
been noted in other studies which are not specifically
related to the present investigation, and are therefore
37
not reviewed here. It was felt, however, that the high
correlations which have been reported for SCAT with college
grades, and because of its wide usage, made it a suitable
test for use in this current study.
Biographical data
Many studies reflected the importance of personal
background data as predictors of school attainment. Those
which seemed to be most nearly pertinent to this investi
gation are reported upon here.
High school grade-point average.--In the studies
which were reviewed, high school grade-point averages have
consistently shown themselves to be the best single pre
dictors of college achievement. For example, in a follow-
up study of all graduates from Staunton Military Academy,
Brice discovered that grade-point averages, together with
high school teachers' evaluations of students, were super
ior predictors of college success as compared with the com
bined results from the American College of Education (ACE)
Quantitative, Language, and Total test results and the
Otis test results (4). In an unpublished dissertation
entitled "The Prediction of Academic Success of Freshmen
in a Community College," Kern also discovered that high
school grade-point average was an evaluator which was much
superior toa number of test instruments in predicting the
success of over 500 students in a large community college
(50)♦ In an extensive study of the relationship of high
school grade averages to college achievement in sixteen
public colleges in Georgia, Franz, et al., reported in
Educational and Psychological Measurement that they had
found consistently low positive correlations (28). They
stated, in addition, that the lowest correlations between
grade achievement in high school and college were from
students enrolled in technical training type colleges
(28), a fact which was of specific interest to the present
study.
Work experience.--Another avenue explored by some
investigators was the relationship of academic achievement
to previous and/or current employment. In an interesting
research study by Dennis Lee Trueblood entitled "Effects
of Employment on Academic achievement," the relationship
of work experience was related to student academic achieve
ment at Indiana University for a sample of 568 full-time
students- (88). This study, which was related to the in
fluence of part-time, concurrent work experience upon
academic success, reported three findings of interest to
this investigation: (X) there were no significant differ
ences in ACE scores between workers and non-workers; (2)
the grade-point average of the workers was essentially
the same as it was for the total of the student body; and
(3) as the number of hours of part-time work increased,
achievement for the 'workers remained the same as achieve
ment for non-workers (88). Three years previously, in an
unpublished dissertation entitled "Selected Characteris
tics including Academic Achievement of Employed and Non
employed Students in the Indiana University School of Bus
iness," Trueblood discovered, after investigation of the
employment records of the entire senior class at Indiana
University, that the extent of employment experience--past
and concurrent--did not relate positively to differences
in achievement levels (89) .
Somewhat related to Trueblood’s findings were
those reported by Motto, in 1960, in an article in the
Journal of Personnel Guidance entitled "Stability of Work
Experience as a Predictor of Success." Comparing the
stability of past employment experience with vocational
training success of disabled veterans for a total sample
of 64 veterans with a median age of 26 years and Army
General Classification Test median scores of 103, Motto
concluded that the trainees with stable employment back
grounds were significantly more likely to complete train
ing (60). However, Motto also discovered that veterans
with stable employment records were not significantly dif
ferent in their Army General Classification Test scores
from the veterans with unstable employment records (60).
In 1954, Gordon reported on the degree of past
work experience as an influence, interrelated with previ
ous culture conditions as an influence, on aptitude and
grade course results in mechanics courses offered by the
United States Air Force. In general she found that cul
tural background factors were not related, but that the
amount of past work experience was significantly related
both to the Army General Classification Test results and
to course grade results (36).
The consensus of all three investigators, there
fore, was that previous work experience correlated posi
tively with success in training. The findings of these
researchers are, therefore, of significance in this cur
rent study because of the attempt to establish predictors
of success in the training of electronics technicians.
41
Father’s occupation.--Another predictor which has
been utilized by some workers is the occupation of the
father of students. In 1954 Porter published an article
entitled "Predicting Vocational Plans of High School
Senior Boys," which was a follow-up study of the influ
ence of students' own plans for their future occupation
coupled with their father's occupational levels as related
to subsequent success in those occupations. Porter con
cluded that high school seniors are fairly stable in their
plans for vocations; in fact, vocational plans were more
stable than vocational preferences (65). Porter reported
that the extent of vocational plans was significantly re
lated to father's occupational level with a coefficient
of correlation of .44. A multiple correlation among extent
of vocational plans, father's occupational level, and Emo
tional Stability score from the Bernreuter test as related
to the degree of vocational success was .53 (65).
While Porter's study was not directly pertinent
to this investigation, his findings were of interest as
background material in early choice of occupation.
Motivation.--The manner and degree in which motiva
tion contributed to success in training and in success in
42
college is an area which has been investigated. Two of
these studies which are of special interest to this inves
tigation are reported.
In 1954 Briggs published a study, "Development and
Appraisal of a Measure of Student Motivation" (5). In
structors from the Fire Control Training School were asked
to rate fifty items of behavior related to characteristics
of trainee motivation. These ratings were correlated to
Army General Classification Tests and course grades, and
Briggs reported that he found moderately high, significant
relationships between characteristics of motivation and
both ability scores and course grades. Two forms of a
rating scale of motivation were developed and Briggs
stated that reliability coefficients of .91 and .81 from
alternate forms were obtained for two groups of instruc
tors who used both forms (5).
In another study on motivation, Teahan reported
in 1957 that high achievers in college tend to look more
to the future in regard to satisfaction. Low achievers
were typically not as optimistic nor did they tend to look
forward to future satisfactions to the same significant
extent that the high achievers did (87).
43
Personal background.--In a more extensive and com
prehensive study, Malloy attempted to predict college
grades by relating personal background influences to suc
cess-in-training measures. In the item analysis study of
what he termed the Life Experience Inventory, 434 students
were used initially and then 441 additional students were
employed in the cross-validation of his instrument. Items
in this inventory included school experiences, attitudes
toward education, self-appraisal, family relationships,
and choice and type of friends. He found that the results
from his inventory yielded a higher multiple-R coeffi
cient (.49) than did the combined results of the American
Council of Education Language test and the Nebraska Eng
lish Placement Test when correlated with the average grades
received (56).
Personality factors
Personality variables have long been thought to be
related to training achievement, and a number of research
workers have either developed their own standards or tests,
or have used those already available to them, to predict
training achievement. A few of the relevant studies are
reveiwed here.
44
One of the new aptitude examinations specifically-
designed for engineering students was tested by Jones and
Case in 1955. Jones and Case used the four sections of an
engineering aptitude test constructed at the University of
California at Los Angeles and correlated the results of
this test with freshman course grades. They reported
multiple-R validity coefficients of .50. However, one of
the conclusions reached was that measures of interest and
personality are needed to account for more of the variance
of the criterion measure (47).
In another study relating college engineering
grades with personality, Crooks, in an unpublished disser
tation entitled "The Value of Personality Adjustment Ques
tionnaire Items in the Predicting of Out-comes in College
Engineering Training," in 1953, reported that, although
results of a personality adjustment questionnaire were not
in themselves highly predictive of success in college
engineering, the items which entered the results contrib
uted substantially to the prediction of college engineering
achievement when combined with ability variables (19).
National Merit Scholarship Competition.--In a study
based on 641 males and 311 females drawn from 7,500 final
ists in the National Merit Scholarship Competition, Holland
measured grade-course results for the freshmen year. In an
article in the Journal of Educational Psychology entitled
"The Prediction of College Grades from Personality and
Aptitude Variables," Holland reported correlations between
the results of an experimental personality inventory and
freshman grade-point averages from 277 colleges and uni
versities (44). Low but significant coefficients were de
termined between the subtest scores on strength of Superego,
Persistence, and Play and the average grade criterion. The
author concluded that the results suggest that non-intel
lectual variables, such as Superego, Persistence, and
f t
Deferred Gratification (the opposite end of the Play
Scale) were useful in predicting success and in understand
ing the nature of top academic achievers (44).
In a study titled "Use of Projective Tests in
Predicting College Achievement," Chahbazi supported the
findings of Holland (15). When he administered the Picture
and Sound Stimuli tests to 813 freshmen at Cornell Univer
sity to predict first year college grade averages, Chahbazi
revealed an increase in the multiple coefficient of deter
mination from .26 when ability predictors were used alone
to .40 when projective test results were included (15).
46
Minnesota Multiphasic Inventory.— Many studies
have been completed involving the relationship of the
Minnesota Multiphasic Personality Inventory (MMPI) to col
lege achievement, but in spite of the number of articles
which have been written on these correlations, the results
have generally proved to be disappointing.
In 1954, Hancock and Carter studied approximately
300 students at the University of Illinois enrolled in var
ious curricula, and failed to show any significant rela
tionships between MMPI data and grades. They did, however,
discover significant differences between students enrolled
in engineering as compared with students enrolled in other
*
curricula and reported that engineering students were sig
nificantly less hysterical and paranoiac in responsiveness
as measured by the MMPI, concluding that engineering stu
dents seemed to be more free from symptoms which could
lead to poor vocational adjustment (41).
In 1955, Frink studied 267 freshmen women at the
University of California at Santa Barbara. His paper was
titled "Improving the Prediction of Academic Achievement
by Use of the MMPI," but he reported his findings as show
ing significant low negative coefficients of correlation
between some of the MMPI scales and grade point average (29) .
47
This negative result was confirmed by Stone and
Ganung in 1956, reporting in the Journal of Educational
Research. These research workers reported that college
students who received high scores of 70 or above on one or
more of the Minnesota Multiphasic Inventory scales had
significantly lower grade-point averages than students who
scored in the normal range of the MMPI (81).
The following year, 1957, saw an article by Drake
and Getting in the Journal of Counselling Psychology in
which the authors indicated that by a pattern analysis of
the MMPI results, college achievement was predictable (21).
In 1959, however, another investigator, Gallese, failed
to find any significant differences between the MMPI
scales and academic achievement in an engineering school.
From the above results, it does not appear that
the Minnesota Multiphasic Inventory tests and scales are
overly valuable in predicting scholastic success, which is
not to say that these tests do not have value in other
areas of evaluation and prediction.
Rorschach Tests.--As reported here, studies which
employed the Structure-Objective Rorschach Tests (SORT)
are reviewed separately from those studies which used other
adaptations of the Rorschach Test.
In the Manual of the Structure-Objective Rorschach
Test, the statement is made that when supervisors' ratings
are used as a criterion for the construct validity of the
SORT, a reasonably strong relationship can be demonstrated
between the supervisors' ratings and the SORT T-scores
(13).
In an article entitled "Differential Prediction of
Academic Success at Brigham Young University," Stone re
ported the relationship of scores from the SORT to fresh
men grades for a sample of 966 freshmen. Several of the
single SORT scores, when correlated with the criterion,
produced low but significant results; responses of other
SORT tests showed low and significant positive correla
tions with criterion measures. Still other responses from
the SORT produced low and significant negative correla
tions with the criterion (82). A multiple-R of .64 was
obtained between an optimum combination of SORT scores
with the criterion, but when the optimum combination of
high school grade-point average and optimum SORT scores
were related to the criterion, the multiple-R rose to .68
(82).
When Rorschach adaptations were tried, other
researchers arrived at somewhat different conclusions.
For example, Blechner and Carter, working with standard
score equivalents of specified Rorschach tests, demon
strated low positive coefficients, from .05 to .15, with
course grades in educational psychology courses for 362
students (3). Two years later, in 1958, Clark, using the
multiple-choice group-Rorschach technique, did not find
that measures derived from the Rorschach significantly im
proved the multiple prediction of college course grades
when combined with ability predictors (17). Furthermore,
Sopchak, in his study of single Rorschach variables, con
cluded that they were less valuable in predicting college
course grades than other tests of personality commonly
used (79). Sopchak went on the point out that only the
number of M (human movement) responses resulted in a low
tetrachoric correlation of .24 with grade-point average
(79).
Summary
The review of the literature has led to the fol
lowing summary and analysis of research findings:
1. There have been relatively few studies cover
ing the prediction of success in the training
of electronics technicians.
Studies which have been carried out relating
specifically to trainees in electronics used
the Army General Classification Test or,
recently with civilian samples, the Employee
Aptitude Survey Tests. The validity coeffi
cients from correlating these ability tests
with the usual criteria have been encouraging
in general.
Tests of general intelligence have produced
low correlations with training success in a
variety of technical arts training programs.
Investigators generally feel that the multiple
test battery of relatively unique abilities
shows more promise in predicting training suc
cess than single aptitude tests.
There is some evidence to suggest that bio
graphical and personality characteristics may
be related to college achievement. Those re
searchers who correlated data of this type
agreed, in general, that these characteristics,
if adequately measured, could account for more
of the variance between predictors and criteria
than ability predictor variables by themselves.
One of the serious limitations discovered in
many of the studies related to prediction in
the technical arts courses was that the size
of the sample was often woefully inadequate.
In general, such studies have not been in
cluded in this review of the literature.
Another limitation encountered in published
accounts of these studies has been the lack of
information regarding characteristics of sam
ples.
Patterson found paper and pencil tests gener
ally have been more predictive of training
success than performance-type tests, and he
felt that this result could be attributed to
the fact that the elements measured primarily
by performance tests, such as dexterity, are
not, in themselves, related to training activ
ities (64) .
Although research in the area of learning
generally recognized the importance of motiva
tion in the success of training, little has
been done to measure, or relate, the charac
teristics of motivation to training success
CHAPTER III
METHOD OF PROCEDURE
The objective of this investigation was to deter
mine the extent of the relationships existing between
specific predictor variables and various criteria measures
of success in electronics training in large public col
leges. In order to present this information, it is neces
sary to describe the three elements of the study which
are inherent parts of the preceding statement. These
elements are (1) the specific predictor variables, (2) the
various criterion-measures of success, and (3) the sample
of students involved in the electronics training courses.
The general plan was to select a sample which was
readily available and which would be willing and able to
contribute to this investigation. The predetermined pre
dictor variables would then be applied to this sample and
the results correlated with the pre-established criteria
of success. Analysis of the significance of the differ-
54
ences would then be made in terms of the way these differ
ences related to high achievers and low achievers. Pre
dictor variables which showed promise would then be com
bined in multiple correlations to make possible meaningful
predictions of training success from the test data.
It is the purpose of this chapter to present the
details of the sample selected, the criteria employed, and
the various methods of collecting and analyzing the data.
The Sample
The sample employed in this investigation was
recruited from the students attending three public Southern
California junior colleges, namely, El Camino College,
Santa Monica City College, and Orange Coast College. The
college settings were selected because they represent
large public junior colleges with well-developed technical
education programs in electronics.
The courses from which the trainees were origi
nally selected and from which letter grades were taken as
a criterion of training success were the four major elec
tronics courses at El Camino College and Santa Monica City
College. For identification purposes, these are listed as
follows:
55
El Camino College
1st semester: Basic Electronics 1-A
2nd semester: Electronics 1-B
3rd semester: Electronics 1-C
4th semester: Electronics 1-D
Santa Monica City College
1st semester: Basic Electricity and Lab 1-^
Electronics Shop 1-B
2nd semester: AC Circuits, Transistor,
etc. 2-A
Introduction to Network
Analysis, 2-B
3rd semester: Vacuum Tube and Transistor
Circuits, 3-A
Industrial Electronics
Controls, 3-B
4th semester: Advanced Electronic Circuits,
4-A
Electronic Controls, 4-B
The complete recommended programs for study of
training at El Camino College and Santa Monica
lege will be found in Appendix B.
8 units
8 units
8 units
8 units
5 units
2 units
5 units
3 units
5 units
5 units
5 units
5 units
electronics
City Col-
56
It was initially decided that it would be desirable
for the total sample to fall into three separate classi
fications of approximately equal size: one-third of the
sample to be then enrolled in advanced or second-year
courses, and one-third from each the first semester and
the second semester of training. After determining the
size of the classes at El Camino College and Santa Monica
City College, it was discovered that an insufficient number
of advanced trainees were available to be included in the
sample. Accordingly, the necessary number of fourth
semester students to complete the number required was
enlisted from Orange Coast College--a total of 16 students.
The addition of these 16 advanced trainees from Orange
Coast College brought the number (N) in the sample up to
a total of 176. Table 1 illustrates the breakdown of the
sample by colleges and semesters of training.
The selection of the electronics trainees who were
included in the sample was based upon the following con
ditions:
1. All selected trainees were full-time, day-
school students enrolled, in one of the basic, required
electronics courses, which totaled approximately 15 hours
57
TABLE 1
SAMPLE BREAKDOWN BY COLLEGES
AND TRAINING LEVELS
Training Level
Colleges
Total.
ECC SMCC OCC
First semester 21 35
■ -
56
Second semester 42 16
-
58
Third semester 9 8
-
17
Fourth semester 16 13 16 45
Totals 88 72 16 176
58
1
of class and laboratory time per week. The courses were
offered in sequence only; that is, the beginning course
was a prerequisite for entry into the next level course,
and so on, until the four-course sequence was completed.
Naturally, under these conditions, duplication'of subjects
was automatically eliminated, since it was not possible
for one trainee to be in more than one of these major
electronics courses during the investigation.
2. Because the purposes of trainees forty years
of age and older were quite atypical of their younger
counterparts, it was decided to eliminate from the sample
all such electronics trainees. This decision caused the
elimination of six of the oldest of the training students.
3. Only electronics trainees who were enrolled
and consistently attending their major electronics courses
after three weeks of experience in those courses of the
spring semester of 1961 were included. The results of
this investigation, therefore, cannot be generalized to
include all entering trainees. In other words, entering
students who wanted to "try electronics out,” but who
1
LeBold points out the hazards in combining full-
and part-time students, in making prediction studies
(53).
59
became discouraged and dropped out because of lack of
genuine interest, low aptitude, or greater attraction to
some other occupation have not been included in this study.
Electronics trainees who remained in their major courses
for three weeks, according to reports of the electronics
instructors, usually completed these courses. These re
ports were borne out by statistics, inasmuch as only 2 out
of the 176 trainees included in the sample received drop
out grades.
In summary, the sample used in this investigation
was composed of 176 full-time, day students from electron
ics courses in three Southern California junior colleges,
equally divided among 1st semester, 2nd semester, and 2nd
year enrollees who had already completed the first three
weeks of the semester.
The Criteria
The criterion-measures of success were divided
into two major categories: (1) the Sum-of-Nine-Ratings-
Criterion and (2) the Grade-Received Criterion. These two
separate and distinct elements were subsequently combined
to form a Composite Criterion, which became the basis for
all multiple correlation analyses.
60
Sum-of-Nine-Ratings Criterion
In order to establish a set of criteria on which
tentative predictions of success in electronics training
could be made, an extensive survey of the literature per
taining to descriptions of work experience and training
background as requirements for electronics technicians
was undertaken. At the same time, to supplement the liter
ature with the current opinions of key workers in the field
of electronics training, a number of personal interviews
■ ^
were conducted. The men interviewed were instructors of
electronics courses, personnel men in companies working
in the electronics field, division chairmen of technical
arts training programs, and technical training directors.
As a result of the findings in the literature and
of the interviews with these specialists, a list of nine
areas of ability, interest, and personality characteristics
that would indicate success in the electronics courses
and/or success in on-the-job electronics work was compiled.
The characteristics initially selected were then put into
the form of a rating scale to be used in measuring success
in electronics training. The following list presents the
nine areas of success:
1. Training Achievement--Theory
2. Training Achievement--Laboratory Applications
3. Ability for Precision Work
4. Persistence for Details of Work
5. Ability to Get Along with Other Trainees
6. Ability to Get Along with Instructor
7. Clarity of Written Work
8. Comprehension of Mathematics
9. Ability to Learn from Technical Materials
In scoring these characteristics, instructors were
asked to rate students from 1 to 5, 1 being the lowest
and 5 being the highest score possible. This criterion,
therefore, had a range of from 9 to 45 when scores from
each of the above nine characteristics were added, and the
total figure was termed "Sum-of-Nine-Ratings Criterion.”
The list of nine characteristics was then submitted
to instructors of electronics courses and to technical
training directors for suggestions and review. It was the
consensus among the men consulted that the Sum-of-Nine-
Ratings Criterion as developed was probably a better
instrument for estimating potential employment success in
electronics fields than course grades in college electronics
courses. These opinions conform to the findings of Siegel
62
and Schultz, who support the use of rating scales as a
criterion. Siegel and Schultz discovered that job per
formance in electronics could be as adequately scaled and
rated as attitudes and sensory phenomena have been scaled
psychophysically (76).
With the Sum-of-Nine-Ratings Criterion established
as one of the instruments to be utilized in this investi
gation, a rating form and uniform rating procedures were
drawn up and given to instructors of the students in El
Camino College, Santa Monica City College, and Orange
Coast College who constituted the sample. A copy of this
form is found in Appendix A. The electronics instructors
were asked to rate their students, using this form, during
the fourth week of the spring semester of 1961, rating
each trainee in a major electronics course.
Grade-Received Criterion
The criterion of the grades received by the sample
in the appropriate electronics courses is self-explana
tory. These grades were assembled as usual by all instruc
tors at the completion of the spring semester of 1961 and
permission was given to analyze these grades in this in
vestigation. The letter grades received by the students
63
who constituted the sample were then assigned quantitative
scores and these values were designated the Grade-Received
Criterion.
Composite Criterion
To obtain a single criterion for comparison with
the predictor variables, the Sum-of-Nine-Ratings Criterion
was combined with the Grade-Received Criterion for each
student in the sample, and the resulting figure was named
the Composite Criterion. The authority for this combina
tion of criteria is Gordon's bulletin published by the
United States Air Force Personnel Training Research Center,
entitled Stability of Final School Grade. Gordon deter
mined, after extensive investigation, that grade-in-course
was not a sufficient criterion measure to assess training
achievement in technical courses. She stated that
instructors' rating would add much to the stability of
the criterion measure (36). This Composite Criterion, or
overall criterion of trainee success, was used as the basis
for all multiple correlation analyses and the analysis of
the significance of the differences between high and low
achievers.
As analyzed in Chapter IV, the correlation coef-
64
ficient for the Sum-of-Nine-Ratings compared with the
Grade-Received criteria was .61. This degree of relation
ship suggests that instructors1 ratings of the character
istics described are significantly related to the letter
grades assigned, and forms the basis for the validity of
using these ratings as a measure of training success.
Further support of this validity may be found in the fact
that the elapsed time between making the ratings and as
signing the grades was approximately three months.
Predictor Variables
The predictor variables assembled for use in this
research were of four different varieties. Each variety
was selected as a result of findings reported by research
workers in this and related fields. A total of 50 pre
dictor variables was the result of assembling the following
components:
Aptitude variables from the results of the
Employee Aptitude Survey tests . . 10
School-related abilities from the results of
the School and College Aptitude Tests .... 3
65
Biographical variables taken from the results of
the investigation questionnaire and search
of school records 8 .
Personality characteristics variables from the
Structured-Objective Rorschach Tests .... 25
Motivational characteristics variables from the
results of ratings of characteristics of
motivation obtained through personal inter
views with high and low achievers ...... _4
Total predictor variables ........................ 50
Employee Aptitude Survey
Tests (EAS)
The Employee Aptitude Survey tests have been de
veloped and published within the past ten years. They
were designed to test a broad variety of aptitudes which
have low intercorrelations and are, additionally, rela
tively unique. The Employee Aptitude Survey tests were
developed on the principle that maximum validity per min
ute of testing time is perhaps best achieved through a
battery of short, mutually-independent tests, each of which
makes unique contributions to validity.
It is the total battery of tests included in the
Employee Aptitude Survey which makes this series valid and
66
helpful. As Guilford has indicated:
If a composite score from a battery is to be
used . . . it is likely that there is not much to
be gained by achieving reliabilities for single
tests higher than .60. . . . In general, if there
is a choice between lengthening of tests in a
battery to make them more reliable and adding
more tests of different kinds that contribute
unique valid variances, the decision should be
to the second alternative. (38).
As discussed in Chapter II, the test publisher,
Psychological Services, Inc., of Los Angeles, has reported
on the validity of on-the-job success correlated with
Employee Aptitude Survey tests in extensive research at
Lockheed Aircraft Corporation (68). In addition, research
search using the EAS tests has proved the significance of
these tests in correlation with grades received for elec
tronics courses. The results of these tests administered
at Glendale College (33), and the previously discussed
research of Warren and Dossett (94), and of Foy (27),
made it seem reasonable that these same tests might prove
valid for predicting the potential success of the trainees
in this investigation.
It seemed likely that much could be gained by the
use of such a battery as the Employee Aptitude Survey
tests, and this battery was therefore incorporated as one
part of the total of 50 predictor variables. The descrip-
tion of these tests which follows is abstracted from the
Manual of Employee Aptitude Survey: A Battery of Practical
Employment Tests:
Test 1--Verbal Comprehension.--Test 1 measures the
ability to use words meaningfully in communication, think
ing and planning. Performance on this test is highly in
dicative of reading speed and ability to understand written
or spoken instructions. Verbal ability is the most impor
tant single aspect of "general intelligence." This test
requires the subject to find a synonym among four choices
for a given word. There are 30 items and the time limit
is five minutes. The test has an alternate-form reliability
of .83.
Test 2— Numerical Ability.--Numerical ability meas
ures the ability to work easily with numbers and to do
simple arithmetic fast and accurately. The test requires
the subject to perform simple addition, subtraction, mul
tiplication and divisional problems and to indicate which
of five answers is correct. The score is the number right
minus one-fourth the number wrong. The test is set up in
three parts which are separately timed and has altemate-
form reliability that is estimated at .92. Scores may be
68
obtained for each part separately; the total score, how
ever, is recommended to obtain maximum reliability and
validity and was used in this study.
Test 3— Visual Pursuit.--This test measures the
speed and accuracy in visually tracing lines through en
tangled networks of lines. The task is to follow each
line from right to left through the network and select one
of five choices on the right-hand side of the page. The
score is the number right minus one-fourth the number
wrong. Altemate-form reliability is estimated at .82.
Test 4--Visual Speed and Accuracy.--Test 4 meas
ures the ability to see small details quickly and accu
rately, as in visual inspection and clerical work. The
task is to compare visually like and unlike pairs of num
bers and letters and to indicate whether each pair is the
same or different. The score is the total number of rights
minus the total number of wrong responses. Alternate-form
reliability for the test has been determined at .93. Total
testing time is five minutes.
Test 5-~Space Visualization.--Space Visualization
measures ability to visualize forms in space and manipulate
69
objects mentally. The task is to look at a pile of blocks
and determine how many other blocks are in the pile which
touch a particularly lettered block. There are 50 items
and total testing time is 5 minutes. The score is the
number right minus one-fifth the number wrong. Altemate-
form estimated reliability is .93.
Test 6--Numerical Reasoning.--This test measures
ability to analyze into logical relationship and discover
the underlying principles in such relationships. This is
essentially the process of making conclusions based on
inductive reasoning. There are 7 numbers in series of 20
problems. The task is to choose one of five possible
choices that would replace the question mark and which
continues the series correctly. The time limit on the
test is five minutes. Alternate-form reliability is .60,
while the test-retest reliability is estimated at .76.
Test 7--Verbal Reasoning.--Test 7 measures the
ability to analyze verbally-stated facts and to make valid
judgments on the basis of logical implications of such
facts. A significant' feature of this test is that it meas
ures the ability to decide whether or not the available
facts provide sufficient information to support a definite
70
conclusion. Altemate-form reliability is estimated at
.70. The task is to draw conclusions which may be true,
false, or uncertain on the basis of four or five facts.
There are 30 possible conclusions based upon 6 sets of
facts. The scoring is number of rights minus one-half
number of wrongs.
Test 8--Word Fluency.--Word Fluency measures flex
ibility and ease in verbal communications. In contrast
to verbal comprehension, word fluency involves speed and
freedom in using words rather than understanding of verbal
meanings. It is a test to measure how rapidly the subject
thinks of words. Altemate-form reliability is estimated
at .75. The task is to write as many words as possible
beginning with either the letter "S" or MCM. The score is
the number of words written in five minutes.
Test 9--Manual Speed and Accuracy.--This test
measures the ability to make fine finger movements rapidly
and accurately. An important factor affecting the score,
on this test is the willingness to perform a type of
monotonous and repetitive task which is involved. The task
is to place a pencil dot in as many "O's" that are arranged
in lines as possible in five minutes. The score is the
71
number marked correctly minus five times the number marked
incorrectly. The test-retest reliability estimate at .79
has been determined.
Test 10--Symbolic Reasoning.--Test 10 measures the
ability to manipulate abstract symbols mentally and to make
judgments and decisions which are logically valid. An
important aspect of this test is to evaluate whether ade
quate information is available and to make definite deci
sions. A set of six abstract symbols are used. Each prob
lem contains a statement and conclusion both given in
symbolic form. The task is to determine whether each con
clusion is true, false, or impossible to determine on the
basis of the statement. There are thirty problems and
five minutes of working time. The score is the number of
right responses minus one-half the number of wrong. Reli
ability estimate at .69 was obtained on alternate-forms of
the test (68).
It is interesting to reflect on Warren, et al.1s
factorial analysis of the composition of the Employee
Aptitude Survey tests. Manual Speed and Accuracy (EA.S
Test #9) and Visual Speed and Accuracy (EAS Test #4) were
omitted in Warren and Dossett's investigation since they
72
were clearly unrelated to any of the other EAS tests (94).
The eight test characteristics were factored out and are
described:
1. Verbal Comprehension:--Ability to understand the
English language.
2* Number:--Facility in handling numbers or working
with numerical material.
3. Visual pursuit:— Ability to make rapid accurate
scanning movements with the eyes.
4. Visualization:--Ability to manipulate mentally
visual objects in order to visualize them in
changed appearance or location.
5. Spatial relationships:--Ability to perceive accu
rately the arrangement and interrelationships of
elements in a visual pattern.
6. Word fluency:--Ability to produce words rapidly,
without regard to meaning or quality.
7. Induction:--Reasoning from the specific to the
general in order to derive general principles.
8. Syntactic evaluation:--A form of evaluative reason
ing, concerned with the ability to determine
whether available information is sufficient for
making valid decisions (68).
Warren and Dossett then computed a load factor for
each of the characteristics of eight of the Employee Apti
tude Survey tests and analyzed the factorial composition
of these tests. Their results, as taken from the publish
er's test manual, are presented in Table 2.
After deciding that the Employee Aptitude Survey
73
TABLE 2
FACTORIAL COMPOSITION OF THE EAS TESTS2
EAS Test Factor
Per
centage
Loading
#1
Verbal Comprehension Verbal Comprehension .71
Induction .28
#2 Numerical Ability Number .43
Syntactic evaluation .40
, Verbal comprehension .36
Visual pursuit .34
#3 Visual Pursuit Visual pursuit .67
Visualization .31
#5 Space Visualization Visualization .62
Spatial relations .35
#6 Numerical Reasoning Induction .51
Number .28
#7 Verbal Reasoning Induction .51
Syntactic evaluation .40
Verbal comprehension .32
#8 Word Fluency Word fluency
.57
#10 Symbolic Reasoning Syntactic evaluation .64
Spatial relations .29
Visual pursuit .28
2
From Psychological Services, Inc., Test Manual.
74
tests were sufficiently valid to include in the 50 pre
dictor variables in this investigation, these tests were
administered to the total sample of trainees in their
natural class settings during the fourth and fifth weeks
of the 1961 spring semester. The answer sheets were then
machine-scored insofar as possible through the use of
facilities at El Camino College.
School and College Ability
Tests (SCAT)
The School and College Ability Tests, which were
developed about 1957 to predict general success in college
before students had enrolled, were included among the pre
dictor variables of this investigation because of the
growing general acceptance of these tests by colleges and
universities throughout the United States. In many schools
the School and College Ability Tests have replaced or are
replacing the older predictive instrument which had been
relatively widely used: The American Council on Education
Psychological Examination (usually referred to as ACE).
Both sets of tests have been used to predict success or
failure of students applying for admission to colleges and
both measure verbal, quantitative, and total scores to
arrive at the predictive rating.
75
The School and College Ability tests, being newer
and more recently standardized, were deemed important to
this investigation. In addition, these tests are currently
being used by admissions departments in both El Camino
College and Santa Monica City College. The manual pub
lished by Educational Testing Service, distributors of the
School and College Ability Tests, defined SCAT as:
. , . an aid in estimating the capacity of
students to undertake the academic work of the
next higher level of schooling. They measure
the two kinds of school-related abilities which
are most important in the greatest number of
school and college endeavors: verbal and quanti
tative. (22)
As reported in Chapter II, Kennedy (49) and Klugh
and Bierley (51) investigated the validity of the results
of SCAT as a predictor for college grades and reported a
relatively high coefficient of correlation. In addition,
the SCAT Supplement reports that validity coefficients for
students in Virginia Polytechnic Institute ranged between
.36 to .56 for final course grades.
In a technical report, the Educational Testing
Service stated that the reliabilities for SCAT are the
results of internal analyses based on single administra
tions of the tests. They are, therefore, estimates of
internal consistency. Correlations between alternate forms
76
and test-retest correlations apparently have not been
obtained. With very large numbers of subjects in all
studies, there are reliability coefficients in the high
,80's for the Quantitative Scores and low .90’s for the
Verbal Scores. Reliability coefficients for the Total
Score are reported also in the low to mid .90's.
The School and College Ability Tests were not given
at the same time to all trainees who were concerned with
this investigation. The vast majority of students were
tested between September, 1959, and March, 1961, For those
subjects who had not taken these tests, pick-up testing
was arranged so that 100 per cent of the sample had SCAT
data available.
Biographical characteristics
As a prelude to the testing procedure, all trainees
were asked to complete the Trainee Questionnaire (see Ap
pendix A). Data for seven of the eight biographical pre
dictor variables were thus obtained from the information
supplied by the subjects themselves as noted in answers
to the questions. These variables are as follows:
1. Age in months
2. Units in progress
' 77
3. Completed college units
4. (see below)
5. Extent of employment experience
6. Grade expected by trainee in course
7. Extent of prior training in electronics
8. Father’s occupational level of trainees
Item 4: High school grade-point average was com
puted from the transcript records of the trainees. When
transcripts were not available, telephone conversations or
written communications supplied the needed data until the
information was complete for 100 per cent of the sample.
Age in months.--The questionnaire requested stu
dents to fill in blank spaces calling for the birthday and
also the age, and the age was then computed to the closest
chronological age month.
Units in progress.--Specifically the questionnaire
asked: "Number of units this semester," and a blank was
left for the figure to be filled in by the trainee. The
figure recorded represented the number of college units of
work being carried at the time the questionnaire was com
pleted, which was the fourth week of the spring semester
of 1961.
78
Completed college units.--A blank to be filled in
by the student called very simply for the number of col
lege units completed by the trainees prior to the spring,
1961, college semester.
High school grade-point average.--As previously
noted, the high school grade record was obtained from the
student's transcript when available. If the transcript
was not available, the information was obtained either
through telephone conversations or by correspondence. The
average was computed on the basis of the following: A. = 4,
B = 3, C = 2, and D = 1 grade points for each unit of
credit in a course. F and Inc. earned 0 grade points.
The total of all units of credit assigned by the high
schools were divided into the total number of correspond
ing grade points. Thus, the high school grade-point aver
age was determined.
Extent of employment experience.— At the bottom of
the questionnaire, space was allowed for the student to
fill in information concerning his total paid or unpaid
previous employment. The information requested included
the name of the employer, the period of employment, the
title or duties of the job, and the earnings. A partially
79
ordered scale was devised by assigning arbitrarily equal
units of rating scores. Length of experience, and whether
the experience was related or unrelated, were the bases
upon which the rating scores were assigned. Appendix C
contains a copy of the rating scale upon which the pre
dictor rating was determined.
Grade expected by trainees in course.--This vari
able related to the major electronics course in which the
trainee was currently enrolled and was simply the letter
grade, including pluses and minuses, which the students
anticipated receiving after having had three weeks' exper
ience in the course. The question to be filled in on the
questionnaire asked, "Grade expected this course," with a
blank for the letter to be inserted.
Extent of prior training in electronics.--The
questionnaire contained the following line which related
to prior electronics training: "Have you had electronics
training before: __________ Where: _______________________
Yes or No
How much _________________ ." The replies which were accept
able as contributing to this predictor were those which
expressed closely-related prior semesters of training.
80
For example, having had radio shop in high school was
credited, but no credit was allowed if trainees had had
general shop or metal shop training, or some similar tech
nical arts courses. Many of the trainees had had elec
tronics training in the armed services. Because of the
intensity of training in these programs, it was arbitrarily
decided to credit six months of such training with two
semesters of "prior training" credit.
Father’s occupational level of trainees.--The ques
tionnaire asked: "Father’s principal occupation: (If
deceased or retired, kindly indicate the last principal
occupation held) Title: ____________________. Describe his
duties . . ." and four lines were allowed for the answer.
The United States Bureau of the Census Classification of
Occupational Levels was the guide in assessing the informa
tion (see Appendix D). In order to keep the rating scales
parallel in direction, the scores for "father's occupation
al level" were reversed: for example, a father in a pro
fessional occupation was assigned a score of 6 instead of
1 as identified in the Bureau of Census Classification.
Occasionally the response on the questionnaire was ambigu
ous, and in cases of this kind the trainees were contacted
81
individually so that accurate ratings could be determined
for this predictor variable. An example was the response,
"sales"; this was always clarified.
The eight biographical characteristics were sim
plified to offer information which supplement the balance
of the predictor variables comprising the 50 used in this
investigation.
The Structured-Objective
Rorschach Test (SORT)
The Structured-Objective Rorschach Tests were used
for four principal reasons:
1. Ease of administration
2. Ease of scoring
3. Variety of personality characteristics
4. Test of personality not easily faked.
SORT is a broad test of 25 personality character
istics or, more specifically, of 25 characteristics of
thought processes and temperament factors. This forced-
choice test has been employed considerably in industry
since its initial development and use at Brigham Young
University by Stone (82). Among companies in Southern
California which have made use of SORT are General Tele
phone and Kaiser Steel Corporations. It lends itself quite
82
readily to testing in industry since it may be either
machine- or hand-scored.
In addition to the studies reported in Chapter II,
two test-retest reliability studies were reported by the
test's publisher, the California Test Bureau, for SORT. A
sample of 79 college students and 94 industrial supervis
ors were tested and then retested one week later. It was
reported that "This period was considered long enough to
obviate the effects of memory and short enough to minimize
the effects of personality change" (13). The correlation
coefficients for test-retest reliability ranged from a
low of .62 for white space responses (S) to a high of .90
for color-form responses (FC).
The validity of the claims by the publisher was
accepted and, after reflecting on the validity study for
first-year college grades at Brigham Young University
(13) and employment success at General Telephone Company
(13), it was decided that certain of the variables from
this test might have value in predicting success of train
ing in electronics.
The following descriptions of the areas covered by
the tests are taken from the SORT Manual, although some of
the designations have been altered slightly to conform to
the designations of predictors used in this study:
MENTAL FUNCTIONING CHARACTERISTICS
Intellectual level does not necessarily reflect
intellectual performance. It is desirable to know
such features as the type of.approach to intellec
tual situations used, adaptability to the reason
ing process, flexibility of ideas, and ability to
organize (structure) mental processes.
Theoretical Approach: Facility for thinking
in broad general, or abstract terms; facility for
getting perspective, visualizing the overall pic
ture, and seeing relationships between the parts.
Practical Approach: Tendency for thinking or
attacking problems on the basis of practical,
concrete, or very definite details.
Pedantic -Approach: Preference for thinking and
attacking problems from the standpoint of fine,
minute details; tendency to be perfectionistic
and to focus on precise, sometimes trivial de
tails.
Inductive Reasoning: Facility for logical
thinking based upon inferences from elements;
utilization of their accumulative synthesis to
lead to conclusions, principles, or generaliza
tions; ability to organize details into a mean
ingful whole.
Deductive Reasoning: Readiness to employ the
logical approach in which established or specula
tive theories, principles, or generalizations are
applied to data or details for the-purpose of
analyzing their relationships to one another (and
to the principle probably involved). A balance
between facilities for inductive and deductive
thinking, especially when both are high, would
point toward a mental adaptiveness of "efficiency"
wherein such intellectual potential as the indi-
vidual has is the more effective because of versa
tility in logical processes.
Rigidity in Approach: Tendency toward the
dogmatic or toward fixed ideas. Higher scores
suggest an unwillingness to change a point of view
in spite of evidence to the contrary; low scores
suggest an uncritical acceptance of others' view
points .
Structure in Approach: Facility for mental
alertness and precision and exactitude in percep
tion of reality. Occasionally this relates to a
somewhat rigid and formalistic way of solving
problems, but usually indicates an awareness of
and conformity to the environment and its demands.
Concentration: Capacity for attending to the
task at hand or for avoiding distractions from
one's environment or from one's own extraneous
thoughts.
INTERESTS
These facets of behavior refer to the range
of reactions to perceptual experience. Sensitiv
ity to a variety of kinds of percepts implies a
broader range of interests than does a paucity of
percept-types.
Range of Interests: Tendency of interests to
be either expansive or to be narrow and confined.
Human Relations Interest: Disposition toward
the perception of and attention to elements having
human connotations.
RESPONSIVENESS
Two frames of reference are involved here.
The first derives from the modality of responses,
the second from the frequency of responses. It
Is assumed that responses to items most frequently
seen by the majority of the normative group are
indicative of conformity. Conversely, consistent
selection of rarely observed items implies a dis
position toward uniqueness.
Popular Type Interest: Tendency to perceive
the same features in the same way as others; to
see things as other persons do; empathic tenden
cies .
Original Type Interest: Disposition to per
ceive the unique, the different, and the non-con
forming, perhaps even the eccentric; emphasis on
individualism of actions.
TEMPERAMENT
The attributes listed under this heading re
late largely to deep inner feeling, for which
there often are compensations in outwardly ob
served behavior. Many of the compensations can
become occupational advantages.
Persistence: The determination not to deviate
from a set course. It may appear as doggedness
or stick-to-itiveness. It can range from inabil
ity to stick to or complete a task along to the
further extreme of stubbornness, defiance, or con
tentiousness.
Aggressiveness: The aspiration toward goals
by means of well-accepted and morally developed
procedures; willingness and desire to work; sense
of a mature self-control with social conformity.
Social Responsibility: Willingness to sub
serve oneself, even though no personal gains are
evident; energetic acceptance of one's obligations
to himself, to his family, and to society.
Cooperation: Willingness to use a team-work
approach; sensitivity toward others in combination
with appreciation and responsiveness in human
relationships. Willingness to submerge one's
immediate needs to the long-range interests of
other persons is implied.
Social Tact: Control of impulses and biases;
maturity expressed in the ability to maintain a
stable relationship with supervisors, peers, and
inferiors. There is balance between inner impulses
conscious self-control, and demands of the social
environment.
Confidence Level; Ego-strength, self-confi
dence, morale; inner feelings of prestige or
personal worth, ranging from feelings of inferi
ority to strong feelings of self-assurance. It
implies ability to withstand stresses and strains
to maintain feelings of self-worth (prestige) in
the face of adversity.
Consistency of Behavior; Predictability of
actions; tendency for characteristic behavior pat
terns to be stable and well established.
Anxiety: Generalized apprehensiveness, un
easiness, or internal disquietude; self-concern
and preoccupation with personal well-being, feel
ings, emotions, and sensations, resulting from a
feeling of insecurity. A low anxiety score indi
cates composure; however, excessive composure, or
almost complete absence of anxiety, may indicate
a tendency to smother feelings to the point of
seeming cold and insensitive. Anxiety may reflect
itself in feelings of insecurity, expressions of
inadequacy, or construction of behavior; it may
also reflect itself in erratic behavior.
Moodiness in Behavior: Sharp fluctuations in
mood, ranging from elation to depression. The
intensity and duration of either phase may vary
greatly.
Activity Potential: Control of emotional
energy; energy endowment; capacity to follow
87
through on a planned course of action; concentra
tion of energies in a given direction, as opposed
to dissipation of strength in non-productive chan
nels .
Impulsiveness: Tendency to act upon impulse
rather than on the basis of a considered plan;
reflected in spur-of-the-moment decisions.
Flexibility: Adaptability; faculty for ac
cepting and handling most life situations in a
mature manner; capacity to adjust readily from
one type of situation to another.
Tendency to Conform: Tendency to accept and
be directed by the socially accepted codes, cus
toms, and mores. (13)
The Structured-Objective Rorschach Test was admin
istered during the fourth week of the 1961 spring semester.
The results were machine-scored and translated into T
scores for the purposes of this investigation. The above
25 predictor variables were used and the results of these
characteristics were analyzed in terms of success in elec
tronics training. These results will be presented in the
following chapter.
Measures of motivation
In addition to the foregoing predictor variables,
it was decided to investigate certain characteristics of
motivation behind the selection of electronics training and
to add these characteristics to complete the set of 50
88
variables to be used as the basis of predicting success in
electronics courses. Four characteristics of motivation
in regard to the trainee’s curriculum selection and em
ployment projections were determined from the results of a
semi-guided or structured interview:
1. Source of motivation. Within this variable,
such elements as a discussion of why the trainee wanted to
be an electronics technician were discussed; his experience,
if any, in this line of work; did the trainee elect the
field because his father was in it, or some other person
who was important to the trainee; the benefits the trainee
expected from this field, such as remuneration, compensa
tions including fringe benefits, time off, etc.; or some
other motivational factor.
2. Intensity of motivation. The purpose of this
section of the interview was to attempt to learn how
strongly the trainee felt about his desire to become an
electronics technician,
3. Extent of planning. Two elements were involved
in this item. The first covered the plans which had been
made prior to entering the college course and the trainee
89
was asked to be specific in discussing these plans. The
second element covered the plans the trainee had already
made for entering into his chosen occupation upon comple
tion of his present courses.
4. Extent of manifest interest. The interviewee
was asked to discuss his favorite hobbies or activities,
going back to the time he was about eight years of age and
bringing these hobbies and activities up to the present.
In the discussion, it was hoped that a pattern of interest
would be discovered which might relate to the current choice
of training.
Approximately 25 per cent of the highest achievers
and 25 per cent of the lowest achievers from the total
sample were selected on the basis of the Sum-of-Ratings
Criterion to be interviewed. A set of four questions to
be used as a guide were developed (see Appendix E) in
order to approach each interview with a generally consis
tent plan.
All interviews were held during the months of
March, April, and May, 1961, and were taped for the pur
pose of making ratings of the interviewee after the comple
tion of the interview. Ratings were on the basis of
90
"little," "some," or "much," with relative scores of 1, 2,
or 3 assigned to the factors of intensity of motivation,
extent of planning, and extent of manifest interest. The
primary source of motivation was scored on the basis of 1
for associations, 2 for benefits, and 3 for the activity
itself. These ratings are set forth on the semi-guided
interview form which is Appendix E.
It was felt that the open-end questions covering
the points of motivation were valuable in arriving at a
similarity of stimulus for building a rating scale. The
ratings themselves were made covering all four character
istics of motivation without knowledge as to whether the
student being interviewed was selected because of high
criterion scores or low criterion scores.
Method of Data Analysis
The predominant objective of this investigation
was to determine whether individual predictor variables,
or a combination of various predictor variables, could be
found to predict success in electronics training with any
degree of reliability.
The selected list of 50 predictor variables had
been established. The criteria consisting of instructors'
91
ratings of electronics trainees' behavior, final course
grades in electronics courses, and a weighted composite of
ratings and grades had been determined. Guilford's state
ment that it must be remembered that "The coefficient of
validity in the restricted group is almost invariably
smaller than what it would be in an unrestricted group"
(38) had been taken into consideration. The data were,
therefore, ready for analysis.
This report would be needlessly encumbered if it
were to contain the results of all of the hundreds of cor
relations which were computed in order to analyze properly
the interrelationships and intrarelationships of predictors
and criteria measures. The data were analyzed on the basis
of total group, first semester group, second semester
group, and second year group. Furthermore, many additional
correlations were computed among the interrelationships of
predictors and criteria. The data were also analyzed on
the basis of samples of high and low achievers.
Because of the quantity of data to be analyzed,
approval was sought and obtained forJuse of computer as
sistance from the Western Data Processing Center through
the sponsorship of the University of Southern California.
The WDCORR Program producing the Pearson product-
moment correlation analysis was applied to both predictors
and criteria, resulting in many intercorrelations of pre
dictors and criteria. (37) . Multiple correlation analysis
was obtained by use of the BIMD Program 29 (97) in the first
two multiple correlation approaches. The multiple corre
lation as described by Guilford (39) was used in a third
approach of multiple correlation analysis.
Accordingly, the treatment of the data resolves
itself into calculating the partial regression weights to
be associated with the independent variables in predicting
the Composite Criterion.
It is generally recognized that a coefficient of
multiple correlation represents the maximum amount of re
lationship between predictor variables optimally weighted
and a single criterion. It is also recognized that this
correlation is subject to some shrinkage when the weights
are applied to predicting the criterion in a new sample
. t
from which the weights were not derived. Consequently,
shrinkage formulas, as described by Guilford (39), were
applied to the multiple correlation coefficients.
Frequency distribution of the Composite Criterion
and certain highly related predictors were plotted in the
93
form of a scatter diagram for the entire sample. By in
spection it was noted that the scatter diagrams of the
same predictor-criterion were essentially linear and homo-
scedastic. In addition, the distribution of these demon
strated essential symmetry. The normality was, therefore,
accepted for the sample population's data. Thus, require
ments for the product-moment correlations were satisfied.
In addition, the Kolmogorov-Smirnov Two-Sample
Test, as described by Siegel (77) was used to assess the
significance of the differences between high and low
achievers for criteria and predictor variables. This non-
parametric statistic was also used to determine the degree
of similarity between trainees from El Camino College and
Santa Monica City College on criteria and predictor vari
ables.
Finally, for general comparative purposes, certain
percentiles were determined for the total sample on Employ
ee Aptitude Survey test results, and consideration was
given to the fact, as The Psychological Corporation has
indicated (66), that only relevant comparisons should be
considered.
94
Summary
This chapter has set forth in detail an analysis
o£ the sample which was used in this investigation. As
described, the sample was composed of students at El Camino
College, Santa Monica City College, and Orange Coast Col
lege who were enrolled in major electronics courses.
The criteria which were established were also
described. These criteria consisted of (1) the Sum-of-
Nine-Ratings Criterion, which was essentially an evalua
tion of the members who composed the sample by the in
structors of the electronics courses, based on nine arbi
trarily-selected elements judged to be valuable in elec
tronics training success and (2) the Grade-Received Cri
terion, which was the course grade for the students in the
sample.
A description of the 50 predictor variables se
lected for this investigation has been presented. Essen
tially these variables may be grouped as follows:
Employee Aptitude Survey tests 10
School and College Aptitude Tests 3
Biographical variables 8
Structured-Objective Rorschach Tests 25
Motivational characteristics 4
The total of the predictor variables was 50'.
This chapter has stated the reasons for the selec
tion of these specific variables and has described them in
detail in order that the reader might understand why each
of the 50 predictor variables was chosen.
The method of data analysis has been described,
although in view of the tremendous amount of detailed com
putation, not attempt has been made to describe all of the
steps entailed in analyzing the various data and the many
intercorrelations which were computed. Those which are
relevant to the findings will be further described in de
tail in Chapter IV.
CHAPTER IV
ANALYSIS OF DATA AND FINDINGS
This chapter will set forth an analysis of the
findings and an analysis of the data, which are arranged
into three general sections: (1) correlation analysis of
single variables, (2) multiple correlation analysis, and
(3) analysis of sample differences.
Correlation Analysis of Single Variables
Each of the fifty predictor variables established
for this investigation were correlated with the three pre
viously established criteria: (1) the final course grades
received in major electronics courses; (2) the Sum-of-
Ratings Criterion, consisting of nine trainee characteris
tics as reported by instructors; and (3) the weighted com
bination of the first two criteria, known as the Composite
Criterion.
96
Grade-received criterion
The tabulation of the coefficients of correlation
between the predictor variables and the final grade re
ceived in the major electronics course is presented as
Table 3. For additional facility in comparison, the cri
terion has been set forth in individual coefficients of
correlation for (1) the total group, (2) the first semes
ter of the first year, (3) the second Isemester of the
first year, and (4) the second year as a whole.
It may be noted from Table 3 that wide range
exists in the correlations as they relate to various pre
dictor variables. Predictor variable #18, Grade Expected
in Course, showed the highest positive consistent coeffi
cient of correlation for all three of the training levels.
In fact, even at individual training levels, the grade
expected correlated with the actual grade received at the
highest correlation factor of any of the predictors with
the exception of predictor variable #23, Intensity of
Motivation, for the first semester, and #24, Extent of
Planning, for the second semester students. A. total of
22 of the 50 predictor variables correlated at or above
the .05 level of significance with the Grade-Received
Criterion.
98
TABLE 3
CORRELATIONS BETWEEN PREDICTORS AND
GRADE-RECEIVED CRITERION
— ■■ 1— — i^— **-1 ■ I ' ■ ■ ’■ J ' — — — —
Training Levels
Predictor Variables
Total
N=176
1st
Sem.
N=56
2nd
Sem.
N-58
2nd
Year
N=62
1. EAS #1 Verbal Comprehension
34**
44** 36** 11
2. EAS #2 Numerical Ability 40** 40** 43** 29*
3. EAS #3 Visual Pursuit 16* 24* 04 28*
4. EAS #4 Visual Speed & Accuracy 15* 19 17 09
5. EAS #5 Space Visualization 06 03 10 -02
6 . EAS #6 Numerical Reasoning 18** 12 28* -02
7. EAS #7 Verbal Reasoning 09 12 22* -17
8 . EAS #8 Word Fluency 06 -02 21* -17
9. EAS #9 Manual Speed & Accuracy 11
39**
-08 16
10. EAS #10 Symbolic Reasoning 21** 29* 13 30**
11. SCAT (Q) 26** 25* 13 30**
12. SCAT (L)
33**
44** 33** 12
13. SCAT (T)
30**
26* 33** 25*
14. Age in Months 18** 28* 22* -02
15. Father's Occupational Level 05 04 03 07
16. Prior Training in Electronics 00 12 -16 22*
17. Employment Experience 18** 23* 18 11
18. Grade Expected in Course 48**
46** 54**
53**
19. Units in Progress 20** 20 17 22*
20. Units Completed 14* 24*
34**
03
21. High School Grade Point Average 21** 32** 08 25*
N-79 N=27 N-30 N=22
22. Source of Motivation 02 20 09 00
23. Intensity of Motivation 30** 57** 52** 04
24. Extent of Planning 32**
47**
60** 07
25. Manifest Interest 23* 26 53** 07
99
TABLE 3— Continued
Predictor Variables
Training Levels
_ , 1st 2nd 2nd
A ax Sem. Sem. Year
N=176 N=56 N=58 N=62
26. Theoretical Approach 17* 11 14 27*
27. Practical Approach 00 10 03 -09
28. Pedantic Approach -20**
-30**
-14 -27*
29. Rigidity in Approach -07 -24* -08 -15
30. Structure in Approach -08 -05 -04 -15
31. Human Relations Interest 04 -06 17 -08
32. Popular Type Interest 17* 28* 11 20
33. Original Type Interest -17*
-47**
-04 -11
34. Persistence -07 -24* -08 -15
35. Anxiety 07 17 02 00
36. Activity Potential -01 -15 07 00
37. Inductive Reasoning 12 -01 13 16
38. Deductive Reasoning 02 06 06 07
39. Concentration -04 -02 -01 -06
40. Range of Interests 13 19 19 13
41. Aggressiveness -06 -12 -02 -07
42. Social Responsibility -04 -04 08 -13
43. Cooperation 13* 08 09 19
44. Social Tact 13* 06 -07 26*
45. Confidence Level 00 -03 -03 02
46. Consistency of Behavior -13* -04 -06 -17
47. Moodiness in Behavior 04 01 -01 08
48. Impulsiveness 12 07 -01 18
49. Flexibility 10 -10 09 -22
50. Tendency to Conform 11
49**
07 05
* Significant at .05 level.
** Significant at .01 level or better.
Sum-of-Ratings Criterion
Table 4 reports the coefficients of correlation
between the 50 predictor variables and the nine trainee
characteristics which were grouped to form the Sum-of-
Ratings Criterion. Again, the results are segregated into
the three levels of training and then combined into the
total sample. A wide range of correlations is noted in
this table, also, but it may be important that 27 of the
50 predictor variables show positive correlations at or
above the .05 level of significance, while 5 of the 50
predictor variables produced negative correlations at the
same level. Predictor variables #23, #24, and #25, which
are Intensity of Motivation, Extent of Planning, and
Manifest Interest, respectively, showed the highest posi
tive correlations for the total sample, and it is inter
esting that this result is largely due to the correlations
achieved by the first year students in both the first and
second semesters. By the second year, these three pre
dictors appear to have lost their value.
In general, the Employee Aptitude Survey test pre
dictors ranked comparatively high as predictors for the
total sample, although again, except for predictor variable
#10, Symbolic Reasoning, which was significantly correlated
101
TABLE 4
CORRELATION1 BETWEEN PREDICTORS AND THE
SUM-OF-RATINGS CRITERION
Training Levels
Predictor Variables
Total
N=176
1st
Sem.
N=56
2nd
Sem.
N=58
2nd
Year
N=62
1. EAS #1 Verbal Comprehension 35**
45**
56** 20
2. EAS #2 Numerical Ability 36** 31** 23* 12
3. EAS #3 Visual Pursuit 12* 09
34**
06
4. EAS #4 Visual Speed & Accuracy 16* 20 27* -08
5. EAS #5 Space Visualization 11 05
37**
-06
6 . EAS #6 Numerical Reasoning
19**
23*
41**
-09
7. EAS #7 Verbal Reasoning 14* 19 24* -19
8. EAS #8 Word Fluency 12* 19 07 -10
9. EAS #9 Manual Speed 6c Accuracy 13* 20 38** 12
10. EAS #10 Symbolic Reasoning 35**
37**
44** 30**
11. SCAT (Q) 26** 19 42** 12
12. SCAT (L) 36** 42** 46** 24*
13. SCAT (T)
30** 20
44**
22*
14. Age in Months 31**
23* 44**
19
15. Father's Occupational Level -01 00 06 -09
16. Prior Training in Electronics 03 -01 03 10
17. Employment Experience 24** 27* 34** 11
18. Grade Expected in Course 48** 30** 57** 57**
19. Units in Progress 16* 20 15 14
20. Units Completed 07 23* 21* -14
21. High School Grade Point Average 25** 20
32**
25*
N-79 N=27 N=30 N~22
22. Source of Motivation 14 22 07 14
23. Intensity of Motivation
54** 57**
75** 20
24. Extent of Planning 58** 57**
75**
20
25. Manifest Interest 58** 11
57**
12
^Pearson product-moment correlation coefficients,
decimal points omitted.
102
TABLE 4--Continued
Training Levels
Predictor Variables
Total
N=176
1st
Sem.
N=56
2nd
Sem.
N=58
2nd
Year
N=62
26. Theoretical Approach 14* -02 15 27*
27. Practical Approach 02 16 05 -14
28. Pedantic Approach 25** -19 -29* -28*
29. Rigidity in Approach -27**
-34**
-12 -25**
30. Structure in Approach -18** -23* -06 -24*
31. Human Relations Interest 01 -09 08 01
32. Popular Type Interest 21** 22* 19 23*
33. Original Type Interest -16* -28* -16 -08
34. Persistence -27**
-34**
-12 -35**
35. Anxiety 02 09 00 02
36. Activity Potential 03 09 04 15
37.
Inductive Reasoning 09 -06 10 21
38. Deductive Reasoning 03 01 06 00
39. Concentration -09 -18 06 -13
40. Range of Interests 04 04 15 -19
41. Aggres s ivenes s -12* -26* -01 -01
42. Social Responsibility 17* 06 12 33*
43. Cooperation 13* 16 14 08
44. Social Tact 06 -04 09 10
45. Confidence Level 01 -18 06 12
46. Consistency of Behavior -08 -18 -03 -05
47. Moodiness in Behavior -01 24* -16 -07
48. Impu1s ivene s s 06 19 -11 06
49. Flexibility 15* 00 08 36**
50. Tendency to Conform 19** 33** 15 16
* Significant at -05 level.
** Significant at .01 level or better.
103
for second year students, it was the first year students
who influenced the total positive correlation.
It is interesting that Grade-Expected-in-Course
had higher correlations for all levels of training for the
instructor-rated criteria than was consistently found for
any other predictor variable.
Composite Criterion
When correlations were computed for the combined
Sum-of-Ratings Criterion and the Grade-Received Criterion,
the resulting Composite Criterion showed an even larger
number of significant areas than did either of the criteria
alone, as seen in Table 5. On the other hand, certain
predictor variables are significantly high in correlation
with the Composite Criterion for trainees in their first
two semesters but lack potency at the more advanced second
year training level. This latter condition suggests that
as trainees advance they tend, as a group, to become rather
homogeneous in certain critical characteristics. It is
also possible that the lower and less significant correla
tions between certain of the predictor variables and the
Composite Criterion at the second year level are due to the
weeding out of the less apt and less serious trainees.
104
TABLE 5
CORRELATION2 BETWEEN PREDICTORS AND
COMPOSITE CRITERION
Training Levels
Predictor Variables
Total
N=176
1st
Sem.
N=56
2nd
Sem.
N=58
2nd
Year
N=62
1. EAS #1 Verbal Comprehension 38** 52** 42** 15
2. EAS #2 Numerical Ability 42** 42** 54** 19
3. EAS #3 Visual Pursuit 15* 20 14 14
4. EAS #4 Visual Speed & Accuracy 15* 22* 23* -01
5. EAS #5 Space Visualization 10 04 21* -05
6 . EAS #6 Numerical Reasoning 21** 20 37** -08
7. EAS #7 Verbal Reasoning 12* 18 32** -20
8 . EAS #8 Word Fluency 10 09 24* -14
9. EAS #9 Manual Speed 6c Accuracy 13*
35**
-08 16
10. EAS #10 Symbolic Reasoning 30** 38**
27* 29*
11. SCAT (Q) 30** 26*
37**
22*
12. SCAT (L)
39**
50** 41**
19
13. SCAT (T)
34**
27* 44** 24*
14. Age in Months 26** 30** 32** 10
15. Father’s Occupational Level 03 02 07 -03
16. Prior Training in Electronics 02 07 -07 17
17. Grade Expected in Course
54** 45**
61** 60**
18. Employment Experience 23**
29**
28** 13
19. Units in Progress 20** 23* 19 17
20. Units Completed 12* 27* 02 -14
21. High School Grade Point Average 26**
31**
18 28*
N*=79 N=27 N=30 N=22
22. Source of Motivation 10 24 09 07
23. Intensity of Motivation 58**
69** 62** 18
24. Extent of Planning 62** 60**
73**
24
25. Manifest Interest
39**
23 60** 20
2
Pearson product-moment correlation coefficients,
decimal points omitted.
105
TABLE 5--Continued
Training Levels
Predictor Variables
Total
N=176
1st
Sem.
N=56
2nd
Sem.
N=58
2nd
Year
N=62
26. Theoretical Approach
18**
06 16 30**
27. Practical Approach 01 15 04 -14
28. Pedantic Approach -25** -29* -21* -28*
29. Rigidity in Approach -18** -33** -01 -27*
30. Structure in Approach -14* -15 -06 -21*
31. Human Relations Interest 04 -09 17 -04
32. Popular Type Interest 21** 30** 16 23*
33. Original Type Interest -18**
-45**
-07 -10
34. Persistence -18** -35** -01 -27*
35. Anxiety 06 16 00 02
36. Activity Potential 01 -15 08 06
37. Inductive Reasoning 12* -03 14 23*
38. Deductive Reasoning 01 03 09 06
39. Concentration -07 -11 04 -14
40. Range of Interests 10 11 21* 21*
41. Aggressiveness -10 -21* -02 -02
42.- Social Responsibility 12* 00 12 24*
43. Cooperation 13* 14 11 16
44. Social Tact 05 02 -01 17
45. Confidence Level 01 -11 01 09
46. Consistency of Behavior -10 -11 -06 -12
47. Moodiness in Behavior -01 13 -09 -02
48. Impulsiveness 07 14 -04 11
49. Flexibility 08 -06 08 23
50. Tendency to Conform 21**
49**
10 18
* Significant at .05 level.
** Significant at .01 level or better.
106
Again, it is interesting to note the correlation coeffi
cients of Grade-Expected-in-Course, predictor #18, at all
levels, and to see that the level of significance of the
coefficients does not decrease with an advance in the .
training. In fact, after a relatively significant corre
lation in the first semester, the correlation grew in sig
nificance during the second semester and the second year.
A few of the predictor variables demonstrate mod
erately high significant relationships to criteria. From
the Employee Aptitude Survey, tests on verbal comprehen
sion and numerical ability correlate with significant rela
tionships to criteria, especially at the first and second
semester levels. The School and College Aptitude Test in
Language (SCAT[L]) demonstrates the same moderately high
significant relationship to Composite Criterion.
Motivational characteristics
A separate computation was made to correlate the
criteria with the measures of motivation. Table 6 recapit
ulates the data for intensity of motivation and extent of
planning to show that these two predictors are signifi
cantly high in correlation with criteria. There is an
apparent tendency for manifest interest to correlate
TABLE 6
CORRELATION MATRIX OF MOTIVATION RATINGS TO THREE CRITERIA
VIA TOTAL AND SUB SAMPLES
Motiva
tional
Character
Correlations and Significance Levels
Composite Criterion Sum-of-Ratings Criterion Grade-■Received Criterion
istic
Rated in
Interview
Total
1st
Sem.
2nd
Sem.
2nd
Year
Total
1st
Sem.
2nd
Sem.
2nd
Year
Total
1st
Sem.
2nd
Sem.
2nd
Year
N=79 N=27 N=30 N=22 N=79 N=27 N=30 N=22 N=79 N=27 N=30’ N=22
Trainee’s
Experience
Source of
. 10 .24 .09 .07 .14 .22 .07 .14 .02 .20 .09 .00
Motivation
Trainee1s
Experience
Strength of
Interest
.58
.01
.69
.01
.62
.01
.18 .54
.01
.61
.01
.64
.01
.20 .30
.01
•57
.01
.52
.01
.04
Trainee's
Experience
Plans for
Future in
. 62
.01
.60
.01
.73
.01
.24 .58
.01
.57
.01
.75
.01
.20 .32
.01
.47
.0!
.60
.0!
.07'
Electronics
TABLE 6--Continued
Motiva
Correlations and Significance Levels
■
tional
Character
Composite Criterion Sum-of-Ratings Criterion Grade-Received Criterion
istic
Rated in
Interview
Total
1st
Sem.
2nd
Sem.
2nd
Year
Total
1st
Sem.
2nd 2nd
Sem. Year
Total
1st
Sem.
2nd
Sem.
2nd
Year
N=79 N=27 N-30 N=22 N=29 N*27 N=30 N=22 N=79 N=27 N=30 N=22
Manifest
Interest in .39 .23 .60 .20 .58 .11 .57 .12 .23 .26 .53 .07 ■
Electronics
from .01 .01 .01 .01 .05 .01
childhood
108
109
significantly with the criteria, even though this relation
ship did not hold up at the first semester level of train
ing. Source of motivation demonstrates essentially only a
chance relationship to Grade-Received Criterion, Sum-of-
Ratings Criterion, and to the Composite Criterion.
Summary
It is interesting to note that approximately 60
per cent of the 50 predictor variables are significantly
correlated at, or beyond, the .05 level of confidence with
the Composite Criterion. Generally speaking, these corre
lation coefficients are low and positive. There is a
tendency for some of the predictor variables from the
Structured-Objective Rorschach Tests to show low negative
correlation with the criteria. For example, Pedantic Ap
proach, Rigidity in Approach, and Persistence are illus
trative of this point.
In general, ability variables as covered by the
Employee Aptitude Survey tests and the School and College
Ability Tests, Items 1 through 13 of the 50 predictor var
iables, are more consistently related to criteria than are
the biographical or Structured-Objective Rorschach Tests,
as noted on Tables 3, 4, and 5.
1X0
Multiple Correlation Analysis
Multiple correlation analysis is provided in order
that prediction of success in training from a combination
of individual variables could have practical value in the
selection and guidance of trainees. Actually, there is
little in the magnitude of the single coefficients of cor
relation thus far discussed to excite the interest of the
guidance counsellor.
Intercorrelation analysis
The results of intercorrelations among variables,
both predictor variables and criteria, and between various
combinations of these predictor variables and criteria,
will be discussed first, followed by three approaches made
to the analysis of the multiple correlations.
Table 7 is an intercorrelation matrix of coeffi
cients among the 12 criterion variables. It is noted that
the interrelationship between criterion ratings and the
Grade-Received Criterion are moderately high and signifi
cant. Martoccia and Nelson showed significant high corre
lation between instructors' opinions of naval student pre
check flights and later grades received in training (57).
All intercorrelation coefficients in Table 7 are
i TABLE 7
INTERCORRELATION MATRIX OF CRITERIA
TOTAL SAMPLE N=176......
Criteria
Descriptions of Criteria 1 2 3 4 5 6 7 8 9 1 0 1 1 1 2
1. Training Achievement
(Theory)
f
2. Training Achievement
(Lab Application)
.58
w
3. Ability for Precision Work .53 .70 -
4. Persistence for Details
.54 .64 .71
(Stick-to-itiveness)
5. Ability to Get along with
Other Trainees .39 .42 . 56 .57 -
6. Ability to Get along with
Instructor
.49 .46 .59 .66 .59 -
7. Clarity of Written Work
(Technical)
.56 .60 .58 .58 .53 .57
-
8. Comprehension of Math
.70 .58 .54 .47 .45 .47 . 56
(Capacity & Ability)
•
9. Ability to Learn from
.81 .58 .54 .45 .56 .74
Technical Materials
.59 .57
—
10. Sum-of-Nine-Ratings .79 .79 .82 .81 .70 .76 .77 .77 .81
11. Grade in Course .64 .48 .46 .45 .24. .44 .45 .54 .60
12. Composite Criterion
(Comb. Ratings & Grade)
.78 .69 .69 .69 .49 .65 .66 .71 .77 .92
All coefficients significant at or beyond the .01 level.
Ill
112
significant at or beyond the .005 level of confidence.
One exception to this finding is the comparatively low
positive correlation coefficient of .24 between the instruc
tors' ratings of Ability to Get Along with Other Trainees
and Grade-Received Criterion. Sum-of-Ratings Criterion
is understandably high in relationship to the Grade-Re
ceived Criterion. This correlation of .61 suggests the
degree of validity of instructors' use of a rating scale
after a short period of experience in electronics training
courses as an estimate of how successful trainees will be
as measured by the usual criterion of success, the course
grade.
Intercorrelations between the results of the Em
ployee Aptitude Survey tests and the School and College
Ability Tests are presented in Table 8, and in this table,
too, are the intercorrelations among these ability pre
dictor variables. In general, the EAS tests have low
positive intercorrelations, a finding which corresponds
with the intercorrelations reported by the EAS Test pub
lishers (68). Also of interest is the report of factorial
loadings-of the EAS tests as reflected in Table 2.
As was to be expected, there are high and signifi
cant correlations among SCAT (L), (Q), and (T) scores.
TABLE 8
INTERCORREIATION MATRIX OF ABILITY VARIABLES
TOTAL SAMPLE N=176
Ability Variables
Ability Variables
5 6 7 8 9 10 11 12 13
1. EAS #1, Verbal
Comprehension
2. EAS #2, Numerical
Ability
3. EAS #3, Visual
Pursuit
4. EAS #4, Visual Speed
& Accuracy
5. EAS #5, Space
Visualization
6. EAS #6, Numerical
Reasoning
7. EAS #7, Verbal
Reasoning
. 24 -
. 12s . 20b -
.09 .38b .32b -
.16a .27b .34b .31b -
.28 .56 .12
a
.26b .38b
.45b .28b .I6a .20b .30b .36
8. EAS #8, Word Fluency .17a .17a .06 .17a .07 .23d .09
9. EAS #9, Manual Speed Q2 ^ igb 25b >3]b 02 >05
a Accuracy
10. EAS #10, Symbolic
Reasoning
.25b .40 ,20b .15a .29b .41b .281
.06 . -
.22b .10
TABLE 8— Continued
Ability Variables
Ability Variables
5 6 7 8 9 10 11 12 13
11. SCAT (Q) .32? .56? .12? .30 .31? .41? .33? .08 .11 .37? -
12. SCAT (L) .68? .33? .14? .15? .26? .27? .44? .16a .07 .33? .58? -
13. SCAT (T) .47 .51 .19 .26° .37° .40° .43° .15a .10 .42° .82° .82°
a = Significant at .05 level,
b = Significant at .01 level.
115
SCAT (L) apparently contributes more to the Total score
prediction than does SCAT (Q), a finding which corresponds
to the reports by the tests’ publishers (22).
Biographical characteristics were next intercorre
lated with both the Employee Aptitude Survey tests and the
School and College Ability Tests, and these results are
presented in Table 9. Item #6, Number of Units in Prog
ress, has a low correlation by comparison with all of the
other variables and, as might be expected, a low but sig
nificantly negative coefficient of correlation exists
between this item and Employment Experience. The charac
teristic of Age is also negatively related to Number of
Units in Progress with a coefficient of -.26, but Age in
Months produced a moderately high correlation to Employment
Experience, .49. These findings would tend to suggest
that older electronics trainees tend to take fewer units
of work at one time but have more past and present employ
ment experience.
Table 9 also shows that item #5, Grade Expected in
Course, is low and significant in correlation with most of
the ability variables. High School Grade Point Average,
#8, is also in low positive correlation with several pre
dictor ability variables, which in turn are significantly
TABLE 9
INTERCORRELATION MATRIX BETWEEN CERTAIN PREDICTORS
Predictor Variables
1
First Eight Predictor Variables
2 3 4 5 6 7 8
1. Age in months
-
2. Father's Occupational Level -.09
-
3. Prior Training in Electronics -.00 -.01
-
4. Employment Experience .49a -.02 . 12a
-
.
5. Grade Expected in Course . 13a .07 . 15a .11
-
6. Units in Progress -.26 .11 -.13 - -.22 .16
7- Units Completed . 14a .01 .10 .07 .08 .11
-
8. High School Grade Point Av. .03 .06 -.09 ,06
'lll
.03 . 13a
i
9. EAS #1, Verbal Comprehension - . 14a .04 .08 .23 . 26 .12 .03 . 19b
10. EAS #2, Numerical Ability . 15a .04 -.05 .11 . 23 . 16a .04 . 15a
11. EAS #3, Visual Pursuit .03 -.07 .12 . 14a .00
*°\
.01 .11
12. EA.S #4, Visual Speed & Ac. .10 .02 .00 . 18b . 14a . 22° .00 .09'
13. EAS #5, Space Visualization .08 -.04 -.01 . I4a .02
*
.00 .02 .10
14. EAS #6, Numerical Reasoning .05 .02 -.07 .09 . I8b .12 .08 -.01
15. EAS #7, Verbal Reasoning .02 .00 .07 .04 .10 .11 .01 .10
16. EAS #8, Word Fluency .01 . 12a -. 13a .10 .11 .11 .02 .07
17. EAS #9, Manual Speed & Ac. .12 -.03 .01 - -.03 .09
v
.11 .04 .00
18. EAS #10, Symbolic Reasoning . 12a .10 -.01 .02 . 33 . 15a .11 . 17a
i - *
h-1
o\
TABLE 9--Continued
Predictor Variables
First Eight Predictor Variables
1 2 3 4 5 6 7 8
19. SCAT (Q)
20. SCAT (L)
21. SCAT (T)
.04
. 16a
. 16a
.03
.03
.08
.08
.10
.04
.07
. 16a
.16*
.23b
•23h
. 18
.18b
.11
.10
.10
.01
.04
.15:
.20
.17
a = Significant at .05 level,
b = Significant at .01 level.
i , i i t t?- r i — u s.: j— i e ^ . r ■■■ t'T1 ' - = 7 , i ,i. - i - u - 1 .. i T.f -i iii:-rn s = s c a a m a a ; r ■ iiijf.,, ii"1ji i s g g
118
related to the criteria. This findings, perhaps, explains
why High School Grade Point Average does not add much in
the multiple correlation analysis. It should also be
recalled from previous discussion of research in Chapter II
that the high school grade-point average often is not
highly related to success in technical arts curricula (28).
In this investigation, the coefficient of correlation is
only .26 between High School Grade Point Average and the
Composite Criterion (see Table 5).
It is apparent from the inspection of Table 9 that
the biographical characteristics, except Grade Expected
and High School Grade Point Average, are related to abil
ity variables essentially only by chance. This condition,
of course, provides an excellent opportunity for the bio
graphical characteristics to be combined with the ability
variables in the multiple prediction of the criterion, as
will be discussed later.
In order to complete the intercorrelations among
the predictor variables, data for the 25 personality pre
dictors as gleaned from the SORT tests are presented in
Tables 10 and 11. The nature of the tests, that is the
forcing of the selection of one response over another
possible response of a different scoring category, spuri-
TABLE 10
INTERCORRELATION MATRIX BETWEEN CERTAIN PERSONALITY PREDICTORS FROM THE SORT
TOTAL SAMPLE N=176
Personality Predictor Personality Predictor Variables
Variables
1 2 3 4 5 6 7 8 9 10 11
1. Theoretical Thinking
Approach
2. Practical Thinking
-.64b
Approach
3. Pedantic Thinking
-.47b -.03
Approach
•
4. Rigidity in Thinking
Approach
-.if .01 ,36b
-
■
5. Structuring in
Approach
- .57b .41b ,52b . 39b
-
6. Human Relations
Interest
.23b -.03 18b -.08 -,27b
« •
7. Popular Type Interest .64b
- .49
-.19b -.42^
.4?
-.13* -.39b .24b
8. Original Type Interest .If
•l7l
.3?
-.11 -.61
m .
9. Persistence -.17? • 36b .98 -.08 . 17s
10. Anxiety •45b -,20b -.09 -.38b . 24b -. 16a -.09 -
11. Activity Potential .25 -.08 -.26 -. 12s -.30 .34 -.17® -.12® -.24 -
a = Significant at .05 level,
b = Significant at .01 level.
TABLE 11
INTERCORRELATION MATRIX OF PERSONALITY PREDICTORS FROM THE SORT
TOTAL SAMPLE N=176
Pers. Personality Variables
Var. 1 2 3 4 5 6 7 8 9 10 11 12 13 14
1. I.R.
2. D.R. .29a
3. Con. -.34® .05 -
4. R.I. .57 .27“ -.24“ -
5. Agg. .12® .61“ .62“ .08 -
6. S.R. .72 .59 -.24° .42° .26®
7. Coop. .16® -.06 -.22“ .09 -.34° .27°
8. S.T. .31° .21° -.19° .47° .10 .46b .22b -
9. Conf. .51° .47 -.17® .60° .37b .56b -.04 .84b
10. C.B. -.43 .16® .75“ -.35“ .68b -,22b -.32b -.18b -.13®
11. M.B. -.16® -.38 -.60b -.27“ -.73® -.26® .15® -.52® -.63b -.46b -
12. Imp. .30“ -.06 -.93® ,20b -.59b .20b .20b .17® .14® -.70b .61b -
13. Flex. .30° .62 -.17® .40b .41b .89b -.13® .37b .59b -.09 -.33b .12® -
14. T.C. .67 .04 -. 19b .31b -.13® .47b .26b .06 .10 -.32b .03 .14® .38b -
a = Significant at .05 level,
b = Significant at .01 level.
i —1
to
o
121
ously increases the relationships among these SORT vari
ables. For example, it is understandable that there is a
spuriously high negative correlation between Theoretical
Approach, W responses, and Practical Approach, D responses,
Pedantic Approach, Dd responses, and Rigidity in Approach,
S responses, as the selection of one of these alternatives
in the triad automatically bars the selection of any one
of the others. Consequently, the intercorrelations among
the SORT variables must be regarded, in general, as arti
facts.
Multiple correlation analysis--
Approach number one
In the first approach to an analysis of the mul
tiple correlations, it was decided to select variables
which accounted for at least .01 of the variance of the
Composite Criterion. Based upon BIMD Program (97), step
wise multiple correlation analyses for 46 of the 50 pre
dictor variables were scanned, selected, and combined in
the first approach. The four predictor variables related
to the characteristics of motivation were not included
inasmuch as data on these variables were not available for
the total sample. The Composite Criterion was used as the
122
dependent variable in all of the multiple analysis ap
proaches .
As described in the BIMD Program of multiple cor
relation, an arbitrary basis for selection of variables
must be determined. After all variables were scanned and
selected on the basis of .01 of the variance of the Com
posite Criterion, the most significant were reported in
descending order of contribution. The computed F values
were evolved from the ratio of the sum of squares due to
regression and the deviation about the regression. In the
BIMD Program 29 the multiple correlation analysis F value
indicates the confidence in the extent of the variance
between the first, most significant, predictor variable in
relation to the criterion and the subsequent F value for
the additional contribution of each of the lessor predictor
variables. Naturally, as the lessor predictors are accu
mulated in multiple correlation equation, the F value
diminishes (97) . As additional independent variables were
added to the more significant variables, there occurred a
build-up of multiple-R and the corresponding coefficient of
determination, R . As can be seen in Tables 12, 13, and
14, however, the percentage of contribution which each
variable makes in the multiple correlation decreases as the
123
TABLE 12
MULTIPLE CORRELATION ANALYSIS APPROACH ONE
FIRST SEMESTER N=56
Variable
Coefficient
of Det. (R^)
Computed
F Value
Step-Wise
R Values
EAS #1, Verbal Comp. .270 19.954 .519
EAS #2, Numerical Ability .380 16.256 .617
Age in Months .456 14.551 .656
Rigidity in Approach .508 13.173 .713
Units in Progress .558 12.650 .747
Tendency to Conform .601 12.314 .775
Grade Expected in Course .628 11.552 .793
Theoretical Approach .639 10.420 .800
EAS #6, Numerical Reason. .650 9.484 .806
HS Grade-Point Average .662 8.823 .814
Multiple Correlation Summary
First Six Best Predictors (R)
Coefficient of Determination
Shrunken R9
Shrunken R
(R2)
.775
.601
.748
.561
All Ten Predictors (%)
Coefficient of Determination
Shrunken R2
Shrunken R
(R2)
.814
.662
.774
.600
Computed F Value Significances Level .01
124
TABLE 13
MULTIPLE CORRELATION ANALYSIS APPROACH ONE
FIRST YEAR N=114
Coefficient Computed
Variable of Det p £alue
Step-Wise'*'
R Values
Grade Expected in Course .262 39.727 .512
EAS #2, Numerical Ability .391 35.606 .625
EAS #1, Verbal Comp. .456 30.761 .675
Age in Months .482 25.387 .694
Units in Progress .511 22.579 .715
Tendency to Conform .531 20.211 .729
Multiple Correlation Summary
Multiple R of Six Variables .729
Coefficient of Determination (R^) .531
Computed F Value Significances Level .01
Shrunken R .713
Shrunken R^ .509
■^The most significant predictor variable is pre
sented first, followed by the cumulative addition of
contribution by each preceding variable.
125
TABLE 14
MULTIPLE CORRELATION ANALYSIS APPROACH ONE
SECOND YEAR N=62
Variable
Coefficient
of Det. (R^)
Computed
F Value
Step-Wise
R Values
Grade Expected in Course .351 32.408 .592
SCAT (T) .422 21.494 .649
Theoretical Approach .468 17.043 .684
EAS #3, Visual Pursuit .491 13.752 .701
EAS #4, Visual Speed & Ac. .514 11.830 .717
Rigidity .530 10.318 .728
Popular Type Interest .540 9.066 .735
Employment Experience .553 8.192 .744
EAS #2, Numerical Ability .567 7.557 .753
EAS #6, Numerical Reason. .580 7.050 .762
Units in Progress .592 6.601 .770
Multiple Correlation Summary
First Five Best Predictors ' R .717
Coefficient of Determination (R^) .514
Shrunken R„ .693
Shrunken R .580
Eleven Predictors 1 R .770
9
Coefficient of Determination R .592
Shrunken R .717
Shrunken R^ .513
Computed F Value Significance Level .01
lessor significant variables are added.
By comparing Tables 12 and 13 with Table 14, it
may be seen that there are very few predictor variables in
common between the first and second years of training.
Expected Grade does not show the prominence as a predictor
«
at the first semester level that it does when applied to
more advanced levels of training.
By combining the results of the six most signifi
cant variables for the first semester sample, Table 12, a
rather high and significant multiple-R is obtained--.775.
The percentage of variance of the Composite Criterion--
explained best by the combination of these variables--is
60 per cent, or = .601.
The regression equation which best predicts the
Composite Criterion from the six most significant independ
ent variables indicated on Table 12 is:
Y - . 6A + . 4B 4- . 5C - . 3D + .9E + . 5F
where Y is the expected Criterion score; A represents the
EAS #1 score, B the EAS #2 score, C the Age in Months, D
the T-score for S responses, E the Units in Progress, and
F the T-score for Tendency to Conform or ratio of 0 to P
responses.
127
Reference to Table 13 reveals that the multiple-R
is slightly reduced for the first year sample, .729. The
variance explained by this multiple-R is 53 per cent, or
?
R = » .531. The regression equation which best predicts
the Criterion Score for this sample is:
Y = . 9A + . 3B + . 5C + .3D + . 5E + . 2F
where Y is the Criterion Score expected; A represents the
Grade Expected in Course, B the EAS #2 score, C the EAS #1
score, D the Age in Months, E the Units in Progress, and
F the T-score for Tendency to Conform.
The second year sample multiple-R results are
found on Table 14. Of the first five most important pre
dictor variables, only Grade Expected in Course is in com
mon with the first year sample results, as may be seen by
comparing Table 14 with Table 12. When the first five most
significant variables are combined in multiple prediction
of the Composite Criterion, a substantial multiple-R of
.717 is obtained. The variance explained by this degree
2
of relationship is 51 per cent, or R = .514.
The regression equation which best predicts the
Composite Criterion score from the combination of these
five variables is:
Y = .9A - .4B - .3C - .5D - .IE
128
where Y equals the expected Criterion Score; A represents
the Grade Expected in Course equivalent, B the SCAT (T)
score, C the Theoretical Approach or T-Score of W responses,
D the EAS #3 score, and E the EAS #4 score.
Multiple correlation analysis--
Approach number two
There were two important differences between the
first and second approaches to the multiple correlation
analyses. In the second approach, the Grade Expected in
Course was excluded from consideration and, in addition,
it was decided to tighten the basis upon which the pre
dictor variables were selected for the multiple correla
tion analysis. Thus, for the first semester sample it was
decided to establish a minimum standard of .05 contribution
v
to the variance of the Composite Criterion, as shown in
Table 15; a .01 contribution at the first year level, as
shown in Table 16; and a .03 contribution at the second
year level, as shown in Table 17.
Reference to Table 15 for the first semester sam
ple will show that the multiple-R is .747. The correlation
accounts for 56 per cent of the variance of the Composite
Criterion, or = .558.
129
TABLE 15
MULTIPLE CORRELATION ANALYSIS APPROACH TWO
FIRST SEMESTER N=56
Variable
Coefficient
of Det. (R^)
Computed
F Value
Step-Wise
R Values
EAS #1, Verbal Comp. .270 19.954 .519
EAS #2, Numerical Ability .380 16.256 .617
Age in Months .456 14.551 .676
Rigidity in Approach .508 13.703 .713
Units in Progress .558 12.650 . 747
Multiple Correlation Summary
Multiple Correlation Coefficient (R)
2
Coefficient of Determination (R )
Computed F Value Significance Level
Shrunken R
o
Shrunken R
.747
.558
.01
.724
.524
130
TABLE 16
MULTIPLE CORRELATION ANALYSIS APPROACH TWO
FIRST YEAR N=114
Variable
Coefficient
of Det. (R.2)
Computed
F Value
Step-Wise
R Values
EAS #2, Numerical Ability .241 35.503 .491
EAS #1, Verbal Comp. .367 32.223 . 606
Age in Months .399 24.308 .631
Units in Progress .441 21.511 .664
Tendency to Conform .460 18.375 .678
Employment Experience .474 16.066 .688
Multiple Correlation Summary
Multiple Correlation Coefficient (R) .688
O
Coefficient of Determination (R ) .474
Computed F Value Significances Level .01
Shrunken R .649
Shrunken R^ .421
131
TABLE 17
MULTIPLE CORRELATION ANALYSIS APPROACH TWO
SECOND YEAR N=62
Variable
Coefficient
of Det. (R2)
Computed
F Value
Step-Wise
R Values
Social Responsibility .115 7.765 .338
EAS #10, Symbolic Reason .189 6.874 .435
HS Grade-Point Average .244 6.238 .494
Rigidity in Approach .282 5.612 .532
Multiple Correlation Summary
-
Multiple Correlation Coefficient (R) .532
Coefficient of Determination (R2) .282
Computed F Value Significances Level .01
Shrunken R .495
Shrunken R2 .246
132
The regression equation which best predicts the
Composite Criterion score is:
Y “ .9A + .2B + .5C - .3D + .9E
Where Y equals the expected Criterion Score; A represents
the score from EAS #1, B the score from EAS #2, C the Age
in Months, D the T-score for S responses or Rigidity in
Approach, and E the number of Units in Progress.
The multiple-R for the entire first-year sample is
reported in Table 16. It can be seen that by including the
second semester sample with the first semester sample, two
variables are added to the multiple correlation analysis,
namely Tendency to Conform and Employment Experience.
Rigidity in Approach, however, is eliminated. The multiple-
R coefficient is .688, which accounts for 47 per cent of
2
the variance of the Composite Criterion, or R = .474.
The regression equation which best predicts the
Composite Criterion Score is:
Y = .3A - .7B - .2C - .3D - .2E - .9F
where Y equals the expected Composite Criterion Score; A
represents the EAS #2 score, B the EAS #1 score, C the
number of Units in Progress, D the T-score of the ratio of
0 to P responses or Tendency to Conform, and E the extent
of Employment Experience, past and present.
. 133
Table 17 indicates the multiple correlation anal
ysis of data for the second year sample. The multiple-®,
of .532 accounts for 28 per cent of the variance of the
Composite Criterion, or R^ = * .282.
The regression equation which best predicts the
Composite Criterion Score at the second year level of
training is:
Y - .4A + . 5B + .3C - .3D
where Y equals the expected Criterion Score; A represents
the T-score for Social Responsibility, B the EAS #10 score,
C the High School Grade Point Average, and D the T-score
for Rigidity in Approach or number of S responses.
Multiple correlation analysis--
Approach number three
. Ten of the predictor variables which consistently
showed significant relationships to criteria at the first
year level of training but failed to show this relation
ship at the second year level of training were combined
and used in the third approach to multiple correlation
analysis. In making this third approach, the multiple
correlation analysis outlined by Guilford (39) was applied.
Table 18 presents the multiple correlation analysis
of the data for the total sample of 176 trainees. Verbal
134
TABLE 18
MULTIPLE CORRELATION ANALYSIS APPROACH THREE
TOTAL SAMPLE N=176
Variable
Regres^
sion
Coeffi
cient
Standard
Error of Computed
Regular t Value
Coef.
Prop
Var.
Account
for
EAS #1, Verbal Comp. .387 .224 1.729 .148
EAS #2, Numerical Ab. .398 .094 4.230 .113
EAS #4, Vis. Sp. & Ac . .021 .057 .359 .000
EAS #6, Num. Reason. .249 .292 .853 .008
EAS #7, Verb. Reason. -.364 .190 -1.913 .012
SCAT (L) .196 .125 1.158 .012
Age in Months .189 .104 1.815 .028
Employment Experience .288 .425 .676 .002
Units Completed -.026 .047 - .544 .002
Tendency to Conform .114 .080 1.422 .008
Multiple Correlation Summary
Multiple Correlation Coefficient (R) .578
Coefficient of Determination (R2) .334
Computed F Value (significant at .01 level) 8.275
Shrunken Multiple Correlation Coefficient (R) .546
Shrunken Coefficient
o
of Determination (R ) .298
135
Comprehension, EAS #1, and Numerical Ability, EAS #2,
account for at least 25 per cent of the variance of the
Composite Criterion. By adding Verbal Reasoning, EAS #7,
SCAT (L), and Age in Months, approximately 4 per cent more
of the variance can be accounted for. The multiple-R is
seen to be .578 and when the ten predictor variables are
included a maximum of 33 per cent of the variance of the
Criterion is known.
Table 19 reflects the same ten variables in multi
ple correlation for the first year sample with an N of 114.
Approximately 37 per cent of the variance of the Criterion
is determined by Verbal Comprehension and Numerical Ability
scores from the EAS tests. Age in Months and Numerical
Reasoning, EAS #6, results add approximately 4 more per
cent to the variance predicted. The multiple-R is .656,
which accounts for 43 per cent of the variance when all
ten predictor variables are thus combined.
At this point, it might be mentioned that there
are three predictor variables which failed to correlated
significantly at the first year of training, but which did
relate significantly to criteria at the more advanced level
in the second year. These reversals from expected findings
are limited to SORT predictor variables: (1) Structure in
136
TABLE 19
MULTIPLE CORRELATION ANALYSIS APPROACH THREE
FIRST YEAR N=114
Regres Standard Prop
Variable
sion Error of Computed Var.
Coeffi Regular t Value Account
cient Coef. for
EAS #1, Verb. Comp. .678 .271 2.503 .216
EAS #2, Num. Ability .427 .125 3.424 .152
EAS #4, Vis. Sp. & Ac. -.025 .074 - .334 .000
EAS #6, Num. Reason. -.356 .379 - .939 .012
EAS #7, Verb. Reason. -.076 .246 - .308 .001
SCAT (L) .147 .152 .970 .004
Age in Months .184 .118 1.555 .029
Employment Experience .500 .549 .910 .006
Units Completed .087 .091 .954 .005
Tendency to Conform .121 .113 1.065 .006
Multiple Correlation Summary
Multiple Correlation Coefficient (R) .656
Coefficient of Determination (R^) .431
Computed F Value (significant at .01 level) 7.800
Shrunken Multiple Correlation Coefficient (R) .613
o
Shrunken Coefficient of Determination (R ) .376
137
Approach (F responses), (2) Inductive Reasoning (W to M
ratio of responses), and (3) Flexibility (ratio of W to a
combination of FC and CF responses).
Summary
In review of all approaches to the multiple corre
lation analysis, it was apparent that all of the multiple
coefficients of correlation were significant at or beyond
the .01 level of confidence. This finding was indicated
by the analysis of variance, in accordance with Guilford
(39), and by the computed F values for all coefficients.
The standard errors of the multiple coefficients of corre
lations were negligible, as indicated on Tables 18 and 19.
The expected shrinkage for all of the multiple coeffi
cients of correlations, as is also indicated in the tables,
was also negligible. However, it must be remembered, as
Guilford suggests (39), that shrunken multiple-R coeffi
cients tend to assume the optimum values that may be
expected if the results are cross-validated.
Analysis of Sample Differences
Data were analyzed via subsamples in order that
many of the questions and hypotheses raised by this inves
tigation might be answered. Furthermore, it was believed
138
desirable to know to what extent, if any, the results of
this investigation might be applied to other electronics
trainees in similar settings in public junior colleges.
In order to reflect this consideration, therefore, an
analysis of the data by college subsamples was provided.
Finally, it was felt desirable to have raw score centiles
for this sample for the general purpose of comparing EAS
test results with other samples previously normed (68).
Analysis of differences &
between high and low
achievers
As previously mentioned, 39 high achievers and 40
low achievers were identified and selected on the basis of
the Sum-of-Ratings Criterion. These students were then
interviewed individually without knowledge on the part of
the interviewer as to which of the two groups each of the
trainees represented. Each interview was based on the pre
viously-prepared schedule (see Appendix E). The four
characteristics of expressed motivation were rated by
careful examination of the results of these interviews.
Subsequently, information was added to the inter
view data to indicate whether the interviewees represented
the high- or the low-achievement groups so that the
139
Kolmogorov-Smirnov Two-Sample Test could be used to deter
mine the significance of the differences between samples.
Table 20 indicates the mean scores on all criterion
measures, the percentage of non-overlap between samples,
the Chi-Square.approximation of differences, and the level
of confidence to indicate whether there is a difference
between the samples. Unequivocally, this table shows that
significant differences were revealed between the criterion
measures for high and low achievers, with the high achiev
ers naturally having a significantly higher mean score.
Table 21 reflects the same analyses of differences
for the 50 predictor variables between high and low achiev
ers . As one might suspect after looking at the correlation
analyses, the sigyiificance of the differences between high
and low achievers on certain ability variables are pro
nounced. Verbal Comprehension, Numerical Ability, Numeri
cal Reasoning, Symbolic Reasoning, and the three scores
from the SCAT tests are all confidently higher in mean
scores for high achievers than for low achievers. High
achievers are also significantly older than low achievers.
Furthermore, the high achievers, after three weeks’ exper
ience in electronics courses, are significantly above low
TABLE 20
ANALYSIS OF CRITERIA DIFFERENCES BETWEEN HIGH AND LOW ACHIEVERS
Low High Percentage Chi-Square Signifi-
Criterion Achievement Achievement Non- Approxi cance
Mean Mean Overlap mation Level
Understanding Theory 2.11 4.46 87.96 39.324 .0005
Application to Laboratory 2.26 4.38 84.26 36.082 .0005
Precision in Work 2.22 4.38 79.63 32.226 .0005
Ability to Stick to Tasks 1.93 4.29 85.19 36.880 .0005
Ability to Get Along with
People
2.14 4.13 68.06 23.539 .0005
Ability to Get Along with
Instructor
2.11 4.42 84.26 36.082 .0005
Written Work 2.00 4.13 83.33 35.294 .0005
Understanding of Mathematics 2.26 4.46 84.26 36.082 .0005
Ability to Learn Technical
Materials 2.22 4.33 88.43 39.739 .0005
Sum-of-Nine-Ratings 19.85 38.96 100.00 50.823 .0005
Grade Received in Course 5.63 13.71 100.00 50.823 .0005
Composite Criterion (Sum-of-
36.74 79.58 100,00 50.823 .0005
Ratings and Grade)
o
TABLE 21
ANALYSIS OF PREDICTOR DIFFERENCES BETWEEN LOW AND HIGH ACHIEVERS
Low High Percentage Chi-Square Sigriifi
Predictors Achievement Achievement Non- Approxima- cance
Mean Mean Overlap tion Level
1. EAS #1 Verbal Comp. 12.19
2. EAS #2 Num. Ability 33.26
3. EAS #3 Visual Pursuit 17.56
4. EAS #4 Vis. Speed & Ac. 83.67
5. EAS #5 Space Visual. 26.63
6. EAS #6 Num. Reasoning 8.52
7. EAS #7 Verbal Reasoning 12.93
8. EAS #8 Word Fluency 39.83
9. EAS #9 Man. Speed & Ac. 458.68
10. EAS #10 Symbolic Reasoning 8.81
11. SCAT (Q) 27.15
12. SCAT (L) 23.00
13. SCAT (T) 51.86
14. Age in Months 234.00
15. Father's Occupational Level 3.37
16. Prior Training in Elec. 1.93
17. Employment Experience 3.07
18. Grade Expected in Course 9.44
19. Units in Progress 12.30
20. Units Completed 11.58
18.79 45.37 10.461 .01
50.54 69.44 24.509 .0005
19.50 29.17 4.323
93.48 28.66 5.081
27.50 29.63 4.461
11.21 37.50 7.147 .05
16.25 37.04 6.971
44.50 24.07 2.945
471.52 24.07 2.945
13.04 35.19 6,291 .05
35.38 45.13 9.020 .025
35.33 56.94 16.480 .0005
69.08 47.59 11.556 .005
264.00 47.22 11.333 .0005
3.54 9.26 .435
1.38 20.37 2.108
4.50 25.93 3.416
11.83 45.37 10.416 .01
14.33 31.94 5.186 .10
18.14 6.57 5.686 .10
■ ! >
TABLE 21--Continued
Low High Percentage Chi-Square Signifi-
Predictors Achievement Achievement Non- Approxi- cance
Mean Mean Overlap mation Level
21. High School Grade Pt. Av. 2.10 2.40 7.57 5.626 .10
22. Source of Motivation 2.39 2.42 6.86 .196 .90
23. Intensity of Motivation 1.61 2.37 38.67 6.224 .05
24. Extent of Planning 1.52 2.26 42.11 7.378 .05
25. Manifest Interest 1.39 1.95 29.29 3.370 .20
26. Theoretical Approach 43.73 44.86 17.40 1.406
27. Practical Approach 44.58 48.95 28.94 3.891
28. Pedantic Approach 59.77 51.62 38.83 7.0005 .05
29. Rigidity in Approach 52.00 47.76 27.29 3.460
30. Structure in Approach 56.04 53.95 22.53 2.358
31. Human Relations Interest 52.96 52.29 23.44 2.553
32. Popular Type Interest 28.73 43.29 25.27 2.968
33. Original Type Interest 47.92 42.00 37.00 6.360 .05
34. Persistence 52.00 47.76 27.29 3.460
35. Anxiety 48.15 48.90 15.93 1.179
36. Activity Potential 54.62 51.81 27.39 3.744
37. Inductive Reasoning 49.33 48.57 20.37 2.011
38. Deductive Reasoning 49.63 50.33 16.99 1.529
39. Concentration 51.96 51.00 21.69 2.223
40. Range of Interests 45.89 48.24 25.93 3.175
41. Aggressiveness 55.07 52.67 28.04 3.715
42. Social Responsibility 48.37 49.38 17.99 1.529
TABLE 21--Continued
Low High Percentage Chi-Square Signifi
Predictors Achievement Achievement Non- Approxi cance
Mean Mean Overlap mation Level
43. Cooperation 42.89 46.33 36.15 6.297 .05
44. Social Tact 48.63 48.76 19.59 1.810
45. Confidence Level 51.52 49.62 26.46 3.306
46. Consistency of Behavior 56.30 55.81 19.58 1.810
47. Moodiness 49.11 49.38 12.17 .699
48. Impulsiveness 48,59 49.33 16.93 1.354
49. Flexibility 51.78 51.00 21.16 2.116
50. Conformity 46.00 51.10 44.79 9.556 .01
- H
■ P *
W
144
achievers in Grade Expected in Course score. There is a
tendency, at the .10 level of confidence, for high achiev
ers to have better High School Grade-Point Averages, to
carry more Units in Progress, and to have completed more
college units. High achievers are significantly higher in
mean scores rating Intensity of Motivation and Extent of
Planning.
Low achievers are significantly higher in mean
T-scores, in the number of rare detail responses, that is,
in the Dd or Pedantic Approach. Low achievers are also
significantly higher in mean T-scores for the number of
Original responses, Os. In both Cooperation and Tendency
to Conform, high achievers are very close to the mean of
college students, whereas low achievers are below this
norm.
Analysis of college sample
differences
The significance of the differences between train
ees enrolled in Santa Monica City College and El Camino
College are resported in Table 22. These differences,
like the analysis discussed above, have been studied by
using the Kolmogorov-Smirnov Test (77).
TABLE 22
ANALYSIS OF SAMPLE DIFFERENCES BETWEEN SCHOOLS
Santa Monica El Camino Percentage Chi-Square Signifi-
Predictors City College College Non- Approxi- cance
Mean Mean Overlap mat ion Level
1. EAS #1 Verbal Comp. 15.69 15.42 9.97 1.575 .50
2. EAS #2 'Num. Ability 41.90 38.52 17.55 4.878 .10
3. EAS #3 Visual Pursuit 18.51 18.25 7.81 1.202 .60
4. EAS #4 Vis. Speed & Ac. 87.11 ■ 87.26 6.56 .678 .80
5. EAS #5 Space Visual. 30.07 27.03 18.31 5.309 .10
6. EAS #6 Num. Reasoning 11,13 9.05 27.53 12,000 .005
7. EAS #7 Verbal Reasoning 14.82 12.62 26.14 10.820 .005
8. EAS #8 Word Fluency 37.24 38.40 4.09 .409 .90
9. EAS #9 Man. Speed & Ac. 448.96 453.60 19.32 5.911 .10
10. EAS #10 Symbolic Reasoning 10.33 9.07 15.91 4.009 .20
11. SCAT (Q) 27.46 30.06 13.61 2.914 .30
12. SCAT (L) 26.28 28.05 15.53 3.820 .20
13. SCAT (T) 54.02 56.10 19.19 5.834 .10
14. Age in Months 250.84 256.12 11.49 2.091 .40
15. Father's Occupational Level 3.39 3.20 10.73 1.824 .50
16. Prior Training in Elec. .81 1.49 13.64 2.945 .30
17. Employment Experience 4.31 3.90 13.01 2.679 .30
18. Grade Expected in Course 10.19 10.52 15.78 3.945 ,20
19. Units in Progress 13.53 13.40 10.98 1.911 .40
20. Units Completed 19.68 21.34 22.84 8.224 .0250
H
in
TABLE 22--Continued
Santa Monica EX Camino Percentage Chi-Square Signifi-
Predictors City College College Non- Approxi- cance
Mean Mean Overlap mation Level
21. High School Grade Pt. Av. 2.20 2.27 12.64 2.519 .30
22. Source of Motivation 2.42 2.43 6.46 .320 .90
23. Intensity of Motivation 2.15 1.96 11.59 1.033 .60
24. Extent of Planning 1.97 1.91
3.82 .112 .90
25. Manifest Interest 1.73 1.63 10.34 .822 .70
26. Theoretical Approach 45.94 44.74 8.62 1.171 .60
27. Practical Approach 45.00 46.95 12.50 2.462 .30
28. Pedantic Approach 54.18 55.59 15.37 3.724 .20
29. Rigidity in Approach 51.63 52.11 9.15 1.318 .60
30. Structure in Approach 52.61 55.31 14.03 3,102 .30
31. Human Relations Interest 54.04 51.77 13.46 2.853 .30
32. Popular Type Interest 41.06 41.28 13.36 2.813 .30
33. Original Type Interest 44.35 45.10 8.67 1.184 .60
34. Persistence 51.63 52.21 9.15 1.318 .60
35. Anxiety 48.14 50.33 13.36 2.19 .30
36. Activity Potential 54.01 53.14 6.14 .962 .90
37. Inductive Reasoning 51.21 47.99 16.79 4.466 .20
38, Deductive Reasoning 50.56 49.00 16.04 4.072 .20
39. Concentration 50.40 51.49 11.24 2.000 .40
40. Range of Interests 50.60 47.57 24.75 9.700 .010
41. Aggressiveness 54.25 53.36 15.15 3.636 .20
42. Social Responsibility 50.17 48.02 18.81 5.606 .10
146
TABLE 22--Continued
Santa Monica El Camino Percentage Chi-Square Signifi
Predictors City College College Non- Approxi cance
Mean Mean Overlap mation Level
43. Cooperation 44.57 45.24 10.48 1.739 .50
44. Social Tact 49.99 48.56 12.88 2.627 .30
45. Confidence Level 52.02 50.92 7.27 .781 .70
46. Consistency of Behavior 54.86 55.31 7.58 .909 .70
47. Moodiness in Behavior 48.39 49.32 12.12 2.327 .40
48. Impulsiveness 49.51 48.53 12.37 2.425: .30
49. Flexibility 53.24 50.32 21.72 7.470 .025
50. Tendency to Conform 48.53 48.37 10.68 1.797 .50
t —*
4>
148
Of the 62 variables analyzed--12 criteria variables
and 50 predictor variables--only 5 predictor variables
disclosed any significant differences between trainees
drawn from the two colleges. It is conceivable that 5 to
7 additional significant differences might be arrived at
by chance. Consequently, the significant differences which
have been discovered cannot be considered to be a serious
handicap in the generalizability of the results to other
like samples.
Other sample comparisons
Table 23 reflects Employee Aptitude Survey test
scores with centile equivalents. This investigation’s
sample of 160 trainees from El Camino College and Santa
Monica City College is compared to the Test Publisher's
norm data (68) for several samples. The 10th, 25th, 50th,
75th, and 90th centile raw score equivalents are reported
for each of the 10 EAS tests.
Any conclusions must be drawn with considerable
caution. However, it certainly appears from inspection of
the results for a number of the EAS tests that this inves
tigation's sample is less capable generally than college
students but more capable than high school students.
149
TABLE 23
RAW SCORE NORM COMPARISONS BETWEEN SAMPLES
EAS Tests and Samples^-
Centiles
10th 25th 50th 75th 90th
Verbal Comprehension
General Population N=910 , 7 10 14 18 22
College Students N=212 12 14 19 22 24
Study Sample N=160 ' 9 12 15 19 22
High School Students N=s167 7 10
-
17 19
Numerical Ability
General Population N=795 12 17 25 34 43
College Students N=348 26 33 42 50 56
Study Sample N=160 26 31 40 51 55
High School Students N=169 14 21 29 38 48
Visual Pursuit
General Population N=279 7 11 14 17 20
Freshmen Eng. Students N=213 14 17 19 21 23
Study Sample N=160 14 16 18 21 23
High School Seniors N=90 9 12
— ■
17 20
Visual Speed and Accuracy
General Population N=891 59 69 83 92 104
College Students N=333 83 89 101 117 129
Study Sample N=160 69 76 87 97 108
High School Seniors N=160 65 75 85 95 109
Study sample includes 160 trainees from El Camino
and Santa Monica City Colleges from all four semesters.
All other sample data supplied by Test publisher.
150
TABLE 23--Continued
Centiles
EAS Tests and Samples
10th 25th 50th 75th 90tt
Space Visualization
General Population N=59Q 9 16 23 30 35
Freshmen Eng. Students N=168 18 24 31 36 41
College Students N=305 18 24 31 37 42
Study Sample N=160 16 22 29 34 39
High School Seniors N=90 11 18 24 29 33
Numerical Reasoning
General Population N=845 1 4 6 9 11
College Students N=4l2 7 9
-
14 15
Study Sample N=160 5 7 10 13 14
High School Students N=90 4 6 9 11 12
Verbal Reasoning
General Population N=765 1 5 10 15 18
College Students N=277 8 13
-
20 21
Study Sample N-160 6 11 14 17 20
High School Students N=90 4 8
—
17 20
Word Fluency (S)
General Population N=837 25 32 39 47 55
High School Students N=90 32 40 47 56 64
Study Sample N=160 28 33 42 50_ 57
Manual Speed and Accuracy
General Population N= 182 233 307 388 483
Study Sample N=160 328 364 452 524 572
151
TABLE 23--Continued
EAS Tests and Samples
Centiles
10th 25 th 50 th 75th 90th
Symbolic Reasoning
General Population N=601 0 1
—
8 12
College Students N=274 4 6 9 13 17
Freshmen Eng. Students N=204 4 7 10 13 18
Study Sample N=160 4 7 9 14 18
High School Seniors N=90 1 3 6 10 14
152
Specifically, Employee Aptitude Survey tests which have
demonstrated significant relationship to success in train
ing, such as EAS Tests #2, #6, and #10, indicate that the
trainee sample bears a closer relationship to college stu
dents in general and to freshmen engineering students than
it bears to high school students.
However, in comparing 86 first year engineering
students with the 176 electronics trainees of this inves
tigation, certain EAS test score differences are more
apparent, as reported in Table 24. This table presents
I
data from Foy s study (27), and compares them with data
from this investigation.
Summary
This chapter has presented an analysis of the data
which were accumulated through testing 176 students en
rolled in major electronics courses in three junior col
leges in California, and who became the sample in this
study. The criteria established for this investigation
have been analyzed and correlated with each of the 50
predictor variables which were set forth in Chapter III.
Correlations were first computed for single vari
ables and the results were presented in tabular form.
TABLE 24
EAS TEST SCORE COMPARISONS BETWEEN 176 ELECTRONICS TRAINEES AND 86 FIRST YEAR
ENGINEERING STUDENTS FROM UNIVERSITY OF SOUTHERN CALIFORNIA.
SUGGESTED CUTTING SCORES RECOMMENDED AND PERCENTAGE
OF PREDICTIVE EFFICIENCY AS SUGGESTED BY FOY.1
EAS Tests
Electronics
Trainees
Engineering Students
Mean SD Mean SD
Reccom-
mended
Cutting
Scores
% of
Effi
ciency
Verbal Comprehension 15.6 5.2 19.2 4.6 11 78
Numerical Ability 40.8 11.4 54.6 11.4 46 80
Visual Pursuit 18.3 3.9 16.7 4.6 11 81 '
Space Visualization 28.8 9.6 33.5 6.9 21 81
Numerical Reasoning 10.0 3.5 12.4 2.7 10 83
Verbal Reasoning 13.8 5.2 17.2 4.8 7 83
Word Fluency 42.9 12.0 48.0 9.7 28 83
Manual Speed & Accuracy 455.32 92.8 424.4 91.6 250 83
Symbolic Reasoning 9.6 5.1 14.1 4.9 7 84
Glenn Arthur Foy, "A Study of the Relationship between Certain Factor-Analyzed
Ability Measures and Success in College Engineering (unpublished Ph.D. thesis, University
of Southern California, 1959).
154
Multiple correlations were then presented and analyzed as
they were approached in three different ways. Finally,
an analysis was presented of the differences which were
discovered between high and low achievers, and among the
three colleges whose students constituted the sample.
CHAPTER V
SUMMARY
General Summary
As discussed briefly in Chapter II, the growth in
the need for adequately-trained and qualified electronics
technicians has been considerable in the last ten years
and this need is apparently destined to continue to in
crease in the forseeable future. However, the failure of
many technicians-in-training to complete their two-year
public college courses has been both noticeable and alarm
ing to all educators concerned. In order to investigate
the reasons for the drop-out rate, it was a premise of this
research that inadequacies of aptitude, motivation, per
sonal biographical characteristics (such as lack of previ
ous work experience), and individual temperament charac
teristics contributed toward the high rate of failure
among electronics trainees.
The primary purpose of this investigation was to
155
156
discover the characteristics -which relate significantly
with success in electronics training. Various measures of
aptitude, personal data information, personality indices,
and characteristics of motivation were correlated against
the results of instructor's ratings of trainees' behavior
and the final grade received in major electronics courses
in three Southern California public junior colleges, which
comprised the criteria of training success.
To discover the relationship between the measures
and the criteria, Pearson product-moment coefficients of
correlation and multiple correlation analyses were made
and analyzed. Following the correlation study, the sig
nificant differences between high- and low-achievers were
analyzed through the use of the non-parametric Kolmogorov-
Smirnoff Two-Sample Test. Comparisons between results for
trainees from El Camino College and from Santa Monica City
College were made by the same test of significance of the
differences. Employment Aptitude Survey test results for
the sample were expressed in percentiles, and comparisons
were made with other norm groups.
The results of the investigation indicated that
29 of the 50 predictor variables were significantly related
to the Composite Criterion (the Composite Criterion being
157
the weighted combination of the Sum-of-Nine-Ratings Cri
terion and Grade Received Criterion). Aptitude results,
as discovered through the use of the ten tests of the
Employee Aptitude Survey, were generally more significantly
related to training success than was any other group of
predictor variables. However, the best single predictor
of training success for the total sample was the Grade
Expected in Course predictor, which was the grade which the
trainee expected after having had three weeks' experience
in the course.
By combining the results from the Employee Aptitude
Survey tests, especially scores from the Verbal Compre
hension and Numerical Ability, with Age in Months of the
trainees and number of Units in Progress, the multiple
prediction of success in training at the first semester
level of training was substantial, with a multiple-R of
.747.
For the 114 students comprising the sample for the
entire first year, a multiple prediction of success in
training was best achieved by a combination of Grade Ex
pected in Course, Age in Months of the trainees, results
from the Employee Aptitude Survey tests on Verbal Compre
hension and Numerical Ability, the number of college units
158
the trainees were carrying, and the test of Tendency to
Conform. A multiple-!! of .729 was thereby obtained.
The predictor variables significantly capable of
predicting success in training during the first year were
not uniformly reliable for the second year of training.
Symbolic Reasoning from the Employee Aptitude Survey bat
tery, High School Grade Point Average, and the predictors
of Social Responsibility and Rigidity in Approach proved
more capable in the multiple prediction of success at this
second year level of training. The best combination of
predictor variables yielded a multiple-R of .532.
The capacity to combine and predict success by
multiple regression analysis showed that it was a more
productive task at the lower training levels. As the
trainees were drawn from higher levels of training, espec
ially the second year level, there apparently occurred
more restriction in the range of talent. Subsequently,
the capacity to make predictions of success in training
was diminished.
Summary of Biographical Data
Administrators and guidance personnel concerned
with the program of training in electronics have expressed
159
a desire to know more details about the type of person who
becomes a full-time electronics trainee. To this end,
data were collected from a questionnaire, from interviews,
and from school records. One of the intents of this in
vestigation was to provide this general information for
those vitally concerned with guidance and, in addition, it
was planned to investigate some of the predictors which
might be correlated--individually and/or in combination--
with criteria of training success. The following summary
relates to the biographical elements against which pre
dictor variables were measured as proposed in Chapter I.
Age
The range in age of the sample used in this
investigation was from 211 months to 481 months, or from
17 years and 7 months to 40 years and 1 month. The train
ees had a mean age of 254 months, and standard deviation
of 45 months. Consequently, two-thirds of the sample were
found to be between the age of the youngest student in the
sample (17 years and 7 months) to an age of 24 years and
11 months. Although not stated as a specific hypothesis,
age was positive and significant in relationship to success
in training, with a coefficient of correlation of .26.
Furthermore, age was a particularly significant contribu
tor to the multiple regression prediction of training suc
cess for the first year of training.
Units in Progress
A mean of 13.7 college units in progress was es
tablished for the total sample and the standard deviation
was 3 units. Those trainees who took more units tended
to be more successful in training. A significant coeffi
cient of correlation (.20) was found between Units in
Progress and Composite Criterion of training success.
Units Completed
At the time of the collection of the data for this
investigation, the fourth week of the 1961 spring semester,
a mean of 23.25 college units had been completed by the
total sample. This reflects the number of units completed
by trainees from all four semesters of training. A very
low but significant coefficient of correlation (.12) was
found between Units Completed and Composite Criterion.
High School Grade-Point Average
The mean High School Grade-Point Average for the
total sample was 2.24, with a standard deviation of .99.
161
In other words, approximately two-thirds of the total
sample had earned between a D-plus and a B-plus average
in high school. High School Grade-Point Average was sig
nificantly related to Composite Criterion with a .26 coef
ficient of correlation. The difference between high and
low achievers in electronics courses, however, ‘ was not
significant; there was a 35 per cent of non-overlap differ
ence between samples.
Employment Experience
The ratings of past and present employment exper
ience were made on the basis of the scale in Appendix C.
On the basis of this partially-ordered scale, a mean of
3.25 was obtained. Translated, this would indicate that
the average trainee from this sample had work experience
between "some general experience" to "little related ex
perience." A standard deviation of 1.5 scale points was
obtained, which would indicate that the majority of the
total sample had less than one full year of general exper
ience, and less than six months of related experience.
The relationship between Employment Experience and Com
posite Criterion was significant with a coefficient of
.23. However, the test for the difference between high
162
and low achievers was not significant.
Grade Expected in Course
The mean grade expected by trainees in their cur
rent electronics course by the total sample was 10.34,
with a standard deviation of 1.95. As illustrated in
Figure 1, this represents an expected grade of B-minus.
Figure 1 also reflects the mean grade expectancies for
the sub-training and achievement samples. Grade Expected
in Course significantly related to success in training at
all levels of training and clearly differentiated between
high and low achievers.
Occupation anticipated ten
years hence
As may be seen from Table 25, the projected occu
pation of two-thirds of the total sample was that of the
electronics technician. Approximately 25 per cent more
of the trainees saw themselves performing closely related
occupations, such as engineering, electronics buying,
electronics administration, and TV repair servicing. In
none of the sub-training and sub-college samples did less
than a majority of the trainees select the electronics
technician as their future occupation. Approximately 5 per
Mean 1st Semester
Mean 2nd Semester
Mean 2nd Year
1 2 3 4 5 6 7
Fig. 1.--Expected Grade
srurtui o
9.98
10.40
10.60
CO
3
r-f
f a .
<
8 9 10 11 12 13 - 14 15
11.83 - Mean High Achievers
10.34 - Mean' Total Sample
9.44 - Mean Low Achievers
for Total and Sub-samples.
t —*
W
TABLE 25
PROJECTED OCCUPATIONS OF TOTAL SAMPLE
School Sem.
Total
No.
Electronics
Technicians
Professional
Electrical
Engineers
Not Sure or
Unrelated
Occupations
Other
Related
Occupations
Number
Per
Cent
Number
Per
Cent
Number
Per
Cent
Number
Per
Cent
ECC 1st 21 14 67 1 5 2 10 4 19
SMCC 1st 35 19 54 11 31 1 3 4 11
Total 56 33 59 12 21 3 5 8 14
ECC 2nd 42 30 71 4 10 1 2
7 18
SMCC 2nd 16 13 81 0 0 1 6 2 12
Total 58 43 74 4 9 2 3 9 16
ECC 3rd 9 8 89 0 0 1 11 0 0
SMCC 3rd 8 6 75 2 25 0 0 0 0
Total 17 14 82 2 12 1 6 0 0
! - >
o\
4>
TABLE 25--Continued
School Sem.
Total Electronics
Technicians
Professional
Electrical
Not Sure or
Unrelated
Other
Related
. No,
Engineers Occupations Occupations
Number ^er
Cent
Number
Per
Cent
Number ]ler
Cent
Number
Per
Cent
ECC 4th 16 11 69 2 12 2 12 1 6
SMCC 4th 13 8 62 4 31 0 0 1 8
OCC 4th 16 8 50 3 18 2 12 3 19
Total 45 27 60 9 20 4 . 9 5 11
Grand Total 176 117 66 27 15 10a 6 22b 12
a
Unrelated Occupations
1 *
Related Occupations
Not sure 6 Production Engineer 1 Electronics Management 3
Rancher 1 Buyer Elec, 2 Electronics Sales 3
Pilots 3 Electrician 4 Industrial Arts Teachers 2
Draftsmen 2 TV Repair 5
O ' *
Ul
166
cent of the total sample indicated either that they were
not sure or that they were directing themselves toward
entirely unrelated occupations.
Prior Training in Electronics
A mean of 1.2 previous semesters of electronics
training, or its equivalent, was determined for the total
sample. Two-thirds of the sample had between .3 semester
and 2.1 semesters of prior training. Table 21, page 141,
reflects that essentially no relationship existed between
Prior Training in Electronics and success for either the
total sample or for any of the sub-training groups. Table
21, in addition, demonstrates that there was no significant
difference between high and low achievers; in fact, low
achievers had a slightly higher mean score for Prior Train
ing in Electronics. However, there was only a 2 per cent
non-overlap difference between samples.
Father’s Occupational Level
Using the United States Census classification
system (see Appendix D) in making the judgments of the
occupational level of fathers of trainees, the mean score
for the total sample was 3.24, with a standard deviation
of 1.46. This would indicate that the average occupational
167
level of the fathers of these trainees was at the level of
"Craftsmen and Foremen." No significant relationship was
discovered between these ratings and Composite Criterion.
Only 9 per cent of non-overlap difference was determined
between high and low achievers in electronics.
Summary of the Influence of
Motivational Characteristics
Motivation, as judged by the ratings made from the
results of a semi-structured interview, proved to be a
promising predictor of trainee success. Table 5 reflects
the moderately high correlations determined between the
predictors, the Intensity of Motivation and Extent of
Planning, to Composite Criterion. Extent of Manifest In
terest and Source of Motivation were not as significantly
related to success.
This investigator found it extremely difficult to
classify trainee responses as reflecting any predominant
Source of Motivation. This inability to single out the
most significant Source of Motivation may explain the
reason for finding essentially no relationship between the
rating characteristic and the Composite Criterion of suc
cess.
168
High achievers were rated significantly higher for
Intensity of Motivation and Extent of Planning. A near,
yet non-significant, difference existed between high and
low achievers on the basis of their ratings of Manifest
Interest; high achievers had a slightly higher mean score.
It would seem that a trainee's verbal expressions
as to why he was interested in, or motivated toward, elec
tronics was related to achievement only by chance. How
ever, the ratings of how much a trainee was interested in,
or motivated toward, electronics clearly was related to
success in training; and, furthermore, these ratings clear
ly demonstrated that high achievers had a higher degree of
interest.
Summary of the Influence of
SORT Variables
A few of the 25 predictor variables which comprise
the Structured-Objective Rorschach Test, as reported in
Table 5, were significantly related to Composite Criterion.
The correlations were generally low in relationship with
Composite Criterion. Also, in general, these variables
were more significantly related to the total sample train
ing success than for any of the sub-samples of trainees.
169
These data suggest a tendency for the more success
ful trainees to be more nearly typical of college students
in the characteristic of conventionality, while the low
achievers were significantly lower in their Tendency to
Conform. There was an indication that the high achievers
build up information and reason inductively to a degree
greater than do the low achievers. High achievers also
had a tendency to be more cooperative and more socially
responsible. Table 21 showed that in only 4 out of the 25
SORT predictor variables did the results clearly differen
tiate between high and low achievers. Tendency to Conform
was significantly higher in mean score for the high achiev
ers at the .01 level of confidence, and Cooperation was
significantly higher in mean score for the high achievers
at the .05 level of confidence. On the other hand,
Pedantic Approach and Original Type Interest were signifi
cantly higher for the low achievers at the .05 level of
confidence.
Summary of the Influence of
EAS Test Variables
Items 1 through 10 on Table 5 indicate the coeffi
cients of correlation between Employee Aptitude Survey
test results and Composite Criterion. Verbal Comprehen
170
sion, Numerical Ability, Numerical Reasoning, and Symbolic
Reasoning were moderately high in relationship to Composite
Criterion. When combined, results from the Verbal Compre
hension and Numerical Ability tests were particularly
valuable in predicting success at the first year level of
training (see Table 13).
In an analysis of the differences between the high
and low achievers (see Table 21), It was clear that high
achievers were significantly higher in mean scores on
Verbal Comprehension, Numerical Ability, Numerical Reason
ing, and Symbolic Reasoning, with differences between the
two groups especially significant for Numerical Ability.
Approximately a 70 per cent non-overlap difference existed
between these samples.
Summary of the Data Related
to the Hypotheses
Hypothesis One: Significantly higher scholastic
aptitude scores will be manifested by the high achievers
as compared to the scores of the low achievers.--Results
indicate the Hypothesis One can be accepted; the null
hypothesis is thereby rejected. As measured by the School
and College Ability Tests, high achievers made signifi-
171
cantly higher mean scores on the SCAT (Q) , SCAT (L) , and
SCAT (T). Table 21 reflects these mean scores, the per
centages of non-overlap, the Chi Square approximations,
and the confidence levels for the significance of the
differences.
Hypothesis Two: Significantly higher scores will
be manifested by the high achievers as compared to the
scores of the lew achievers in specific Employee Aptitude
Survey tests, namely: Verbal Comprehension, Numerical
Ability, Space Visualization, Numerical Reasoning, and
Symbolic Reasoning.--This hypothesis can be accepted in
part and must be rejected in part. Scores for the Em
ployee Aptitude Survey tests on Verbal Comprehension,
Numerical Ability, Numerical Reasoning, and Symbolic
Reasoning were significantly higher for high achievers as
compared to low achievers. Therefore, the null hypothesis
was rejected. However, results of the Space Visualization
test failed to show that high achievers were significantly
higher in mean score than the mean scores of low achievers;
a slightly higher mean score was obtained for high achiev
ers, but this difference was only significant at the 20 per
cent level of confidence.
172
Hypothesis Three: Significantly higher high school
grade point averages will be manifested by high achievers
as compared to low achievers.--The results did not support
the hypothesis. Therefore, the null hypothesis was ac
cepted. High achievers had a mean High School Grade-Point
Average of 2.AO, while low achievers had a mean High School
Grade-Point Average of 2.10, This difference was signifi
cant only at the 10 per cent level of confidence.
Hypothesis Four; Significant differences will not
exist between high and low achievers on the basis of other
specific tests in the Employee Aptitude Survey battery of
tests, namely: Visual Pursuit, Visual Speed and Accuracy,
Verbal Reasoning, Word Fluency, and Manual Speed and
Accuracy.--The results indicate that the hypothesis can be
accepted in full. That is to say, there was no signifi
cance in the differences between high and low achievers;
the null hypothesis was thereby rejected.
Hypothesis Five; Significant differences will not
exist between high and low achievers on the basis of 25
variables taken from the results of the Structured-Objec
tive Rorschach Test.--The results indicate that for 21 out
of the 25 Structured-Objective Rorschach Test variables,
173
the hypothesis can be accepted. However, four out of the
25 SORT variables clearly differentiated between high and
low achievers. High achievers had significantly lower mean
scores than did low achievers on the tests reflecting
Pedantic Thinking (Dd responses) and Original Type Thinking
(0 responses). High achievers had significantly higher
mean scores than did low achievers on the basis of scores
reflecting Tendency to Conform (0:P responses) and Cooper
ation (CF:FC responses). The significance of the differ
ence of Tendency to Conform was at the 1 per cent level,
and the significance of the differences for the other 3
SORT variables were at the 5 per cent level of confidence.
Hypothesis Six: Significant differences will not
exist in the accuracy of the relationship between the
grade expected and the grade actually received in elec
tronics courses between high achievers and low achievers.--
The results strongly reject the hypothesis. High achiev
ers were significantly higher in the Grade Expected in
Course than were low achievers. There was a non-overlap of
45 per cent difference between the two sub-samples. The
significance of this difference was supported at a 1 per
cent level of confidence.
174
Hypothesis Seven: Significant differences will not
exist in the ratings of the Father’s Occupational Level
between high achievers and low achievers.--For this sample,
it would appear that the hypothesis was accepted. High
achievers had a slightly higher mean Father’s Occupational
Level score, but the difference was not significant. The
null hypothesis was thereby accepted.
CHAPTER VI
CONCLUSIONS
The problem which this investigation attempted to
delineate and, in some measure, solve was based on the
current and constantly-expanding shortage of adequately
trained electronics technicians, and the drop-out or fail
ure rate of electronics trainees which was and is alarming
guidance personnel and educators.
If a method of predicting potential success for
students entering electronics courses or, on the other
hand, potential lack of success for such students, could
be established, administrative personnel, guidance coun
selors, and educators in colleges and more especially in
junior colleges offering training in electronics could
use such a method in guiding candidates for admission.
The conclusions of this investigation are set
forth in this chapter under two categories: (1) the analy
sis of the trainees themselves and (2) the predictive
175
capacity of single or multiple variables.
Description of the Sample
Electronics trainees enrolled in electronics day
classes in El Camino College, Santa Monica City College,
together with a small segment of trainees from the second
year classes at Orange Coast College, constituted the sam
ple in this investigation. This study has concluded that,
based on the sample used, these trainees tended to be
older than junior college students in general and that
they appeared to be more homogenious in terms of their
mental potentials. That is, this sample was similar to
the average range of college students' aptitudes but it
did not contain as many low aptitude students nor as many
high aptitude students. This condidition was especially
noted in the results of tests covering Verbal Comprehen
sion, Numerical Ability, Numerical Reasoning, and Symbolic
Reasoning from the Employee Aptitude Survey battery of
tests.
Approximately two-thirds of the trainees in the
sample saw their future occupations as being those of
electronics technicians. Approximately another 27 per
cent of the sample felt that their futures lay in fields
177
closely related to the occupation of electronics techni
cians, such as electronics engineering, electronics sales,
or electronics management.
Full-time day-school electronics trainees were
taking approximately the same number of college units as
were other junior college students. Many in the sample
had previous training in electronics in high school or in
armed services schools. The grade average from high
school experience suggested that the vast majority had high
school averages between a D plus and a B minus level. Few
of the trainees were higher or lower than this range of
high school grade points.
Many of the electronics trainees were either
employed part-time or had had previous experience, or, as
in many cases, had both past and present employment exper
ience .
The range of the father's occupational level of
the trainees was rather narrow. Few were from the lowest
levels of occupations, i.e., unskilled; and few were from
the highest levels of occupations, i.e., managerial and
professional. The typical occupation of the father was
that of a skilled worker, or perhaps a foreman, in a hard
goods manufacturing company.
178
Predictive Capacity of Single Variables
This investigation discovered a total of 29 pre
dictor variables which related significantly to success-
in-training for electronics students.
A total of 9 single predictor variables, from
among the 50 utilized, related at a very significant level
with the Composite Criterion of training success. All of
these significant variables were correlated at, or beyond,
the .005 level of confidence. Six of these variables were
aptitude predictions, namely: (1) EAS Verbal Comprehension,
(2) EAS Numerical Ability, (3) EAS Symbolic Reasoning,
(4) SCAT (Q)s (5) SCAT (L), and (6) SCAT (T). Another
one of the 9 significant variables was (7.) the Grade Ex
pected in Course. The last two remaining variables were
the results of the ratings of the questionnaire concerning
motivation: (8) Intensity of Motivation and (9) Extent of
Planning. These correlated at the significance level with
the Composite Criterion. For these latter two variables,
it must be remembered that only the extremely high and
extremely low achievers were included in the sample. The
coefficients of correlation would probably have been con
siderably weaker if the entire sample had been included
in this correlation analysis.
179
In addition, 11 variables related to the Composite
Criterion of success at .01, a significant level of con
fidence. These variables were (1) EAS Numerical Reasoning,
(2) Units in Progress, (3) Age in Months, (4) Experience
in Employment, (5) High School Grade-point Average, and
six variables from the SORT. Three of the SORT variables
were significantly positive in relationship, namely (6)
Theoretical Approach to thinking, (7) Popular Type Inter
est, and (8) Tendency to Conform, and three of the SORT
variables were significantly negative in relationship,
namely (9) Rigidity in Approach to thinking, (10) Pedantic
Approach, and (11) Original Type Interest.
At the .05 level of confidence, 9 additional pre
dictor variables were related to the Composite Criterion:
(1) EAS Visual Pursuit, (2) EAS Visual Speed and Accuracy,
(3) EAS Manual Speed and Accuracy, (4) EAS Verbal Reason
ing, (5) Units Completed, and the SORT results in (6)
Structure in Approach to thinking, (7) Inductive Reason
ing, (8) Social Responsibility, and (9) Tendency to Con
form.
Despite a relationship between High School Grade
Point Average and Composite Criterion that was lower than
might have been expected as a result of previous studies
180
of college populations, the determined relationship for
this sample was considered to be in line with the results
in previous studies where the sample was as delimited as
was the sample in this investigation.
The most significant single predictors correlated
with training success at coefficients of correlation of
between .30 and .62. These predictor variables were sig
nificant at or beyond the .005 level of confidence. The
second group of significant predictor variables ranged in
coefficients between .18 and .26. They were significant
at or beyond the .01 level of confidence. Of lesser sig
nificance were the last group of 9 predictor variables
which ranged in coefficients of correlation from .12 to
.15 and were significant at the .05 level of confidence.
It cannot be concluded that any one of these vari
ables was of sufficient magnitude to warrant trust in its
use as a selective device in itself.
Predictive Capacity of Characteristics
of Motivation
Based upon the ratings scores made from a single
semi-structured interview, 4 characteristics of motivation
were assessed. Only the top 25 per cent of the high
achievers and the bottom 25 per cent of the low achievers
from the total sample were interviewed.
Conclusions based upon this phase of the investi
gation must be tempered because of the fact that the
reliability of these ratings was not determined. Other
interviewers might have made significantly different rat
ings, or the investigator might have made significantly
different ratings of the characteristics of motivation
under a different set of given circumstances. Taking into
account these limitations, the data would suggest that
some of the characteristics of trainee motivation are pos
sible to describe and classify from information obtained
in interviews.
The ability to determine a predominant Source of
Motivation, as such, however, which could predict training
success, was not discovered. It is possible that the
Source of Motivation is too complex a characteristic to be
determined from a short interview. Indeed, it appears
unlikely that electronics trainees are generally aware of
the source of their desire to enter the field. Even if a
predominant Source of Motivation could be determined and
quantified, it is unlikely that it, in itself, would re
late significantly with training success. The degree of
motivation, giving weight to practical considerations,
182
and the opportunity or encouragements to make plans for
a future career or employment in the field, probably would
be far more important in predicting training success.
The investigator felt--from the results of previ
ous studies and from his personal observations--that the
Source of Motivation might well be described broadly and
then classified into three categories: (1) motivation
through experience which emphasized the activity itself;
(2) motivation because of the benefits to be expected from
the electronics field of employment, such as wage compen
sations, promotional opportunities, and working conditions
in general; and (3) motivation resting on the desire to
"be like someone" from his present or past associations,
such as his father, a teacher, or other significant figure
in his life experience. However, as already expressed, it
was difficult to rate which, if any, of these sources were
predominant.
The extent of Manifest Interest was judged pri
marily from the trainees1 responses indicating the types
of hobbies and activities in which they had engaged since
early childhood. There was a tendency for the trainees
who expressed manifest interests more closely related to
electronics to be stronger in their achievement in
183
training. The significance of this difference was at only
the 20 per cent level of confidence, however, and, there
fore, extent of Manifest Interest must be rejected as
being a truly reliable difference between high and low
achievers.
The results of the ratings of trainees both for
Extent of Planning and for Intensity of Motivation were
highly significant in their relationships with training
success at the first and second semester levels. These
characteristics were not significant, however, for the
sample of second year trainees. Perhaps here, again, the
possibility that advanced training has restricted the
range of trainees’ motivational characteristics and
thereby limited the data's value in predicting success
for electronics students.
The type of verbal expressions and other behavior
used by the Interviewee undoubtedly influenced the inves
tigator in his ratings of these two characteristics of
motivation. This is probably also true of the concrete
plans that the trainees expressed about what they expected
education to do for them and what they intended to do voca
tionally upon completion of their training. The degree
that the trainees seemed to be "excited" about the future
184
of electronics and about their role in that future deter
mined the rating assigned for the Intensity of Motivation
variable.
There was a coefficient of intercorrelation be
tween these two predictor variables of .68. This might
indicate a strong degree of overlap of judgment, although,
of course, it might also be reasonable to expect that those
trainees who were energetic and expressive of intense
interest in the field of electronics would also have cor
responding manifestations in the form of concrete plans
for future training and employment.
Predictive Capacity of Multiple Variables
The capacity to make predictions of success in
training varied inversely with the levels of training.
That is, the most potent combinations of predictor vari
ables of success were found at the first semester level,
somewhat less potent combinations at the first year level,
and the least potent combinations at the second year
level. The step-wise multiple correlation analysis con
sistently demonstrated that the combined results of the
relatively unique Employment Aptitude Survey test abilities
of Verbal Comprehension and Numerical Ability accounted
185
for a large part of the variance from the Composite
Criterion of training success. The Age in Months and the
SORT Tendency to Conform variables were also helpful, in
combination with the two Employee Aptitude Survey aptitude
variables, in increasing the multiple prediction capacity.
The best combinations of predictor variables at
the first semester and the first year levels were in the
multiple-R range of the .60's. The standard error of
measurement was small, and the multiple-R1 s were highly
significant (at or beyond the 1 per cent level of confi
dence) . The degree of shrinkage to be expected if the
multiple regression equation was to be applied to other
like samples was found to be negligible.
The multiple prediction of success at the second
year level of training was of a somewhat lower magnitude.
SORT Social Responsibility and Rigidity in Approach to
thinking, along with EAS Symbolic Reasoning and High School
Grade-Point Average were the most significant contributors
to the multiple prediction of success at this level. The
multiple-R was .532 and significant at a 1 per cent level
of confidence. The standard error was small and the
shrinkage to be expected if the multiple regression was
applied to another like sample was found to be negligible.
186
The importance of certain predictor variables
found significant at the earlier training levels, but not
significant at the second year level, was no doubt due to
the restriction of range of talent within the more ad
vanced sample. That is, as the critical characteristics
important for trainee success operated to assist some train
ees during their early training experience, the same
critical characteristics deterred others of the less-
endowed trainees. Consequently, the range of talent in
such variables as EAS Verbal Comprehension and EAS Numer
ical Ability was sufficiently restricted so that the rela
tionships between the aptitudes and success in training
were no longer as strong. As an illustration of this, the
predictor variables, those which were individually related
to the Composite Criterion in the .40's and ,50's at the
first year level, fell into ,10's as related to success
at the second year level.
On the other hand, such characteristics as EAS
Symbolic Reasoning, being more complex and challenging,
correlated with a relatively low and significant magnitude 1
with training success at all levels.
The results of this investigation generally sup
port the conclusion that there are a large number of
187
electronics trainee characteristics which, used in combin
ations, relate significantly with earned success in train
ing at all levels. Conversely, there is sufficient evi
dence available to conclude that no single characteristic,
in itself, predicts training success well.' The utility
of prediction appears to be greatly enhanced by employing
a multiple regression equation prediction against a com
posite criterion, such as that used in this investigation.
The best combination of predictor variables from
the 46 studied (the 4 characteristics of motivation were
not included in the multiple correlation analysis) and
applied to samples at various levels of training proved
to be highly significant in predicting the Composite Cri
terion of training success. The best combination of pre
dictors applied to the three levels of training all proved
to be significant at the 1 per cent level of confidence
with relatively small standard error of measurement. The
measurements of expected shrinkage, if applied to other
like samples, were small.
Implications of the Investigation
The meaning and value of this investigation may
best be realized by further intensive study. There should
188
be a willingness on the part of public colleges to allow
probable failures or drop-outs to continue their training
programs while the predictors identified in this investi
gation are cross-validated. Critical scores which could
clearly differentiate between those trainees likely to
have poor success--D's and F's--or those who are destined
to drop out of training because of low aptitude or for
other personal characteristics, should be determined and
established as guides for counselors.
The identifiable characteristics from the sample
used in this investigation are probably fairly typical of
the characteristics of other electronics trainees enrolled
in like situations. The characteristics herein identified,
however, may not be representative at all of the typical
characteristics of evening school electronics trainees.
Furthermore, the identified characteristics probably do
not adequately describe the typical characteristics of
electronics trainees enrolled in industrial-type training
programs inasmuch as industrial-type trainees are usually
pre-selected on the basis of aptitudes, employment experi
ence, and a host of other personal information.
It would seem that the most optimum time for the
selection of electronics trainees would be just prior to
189
the entry into the first semester of training. It is
reasonable to suspect that the predictor variables which
were significantly related to success at the first semester
of training in this sample would be even more significantly
related to, or predictive of, success or continuance in
training if applied to all students who indicated a desire
to "try out" the electronics training program. The range
of talents and personal characteristics would probably be
greater for the unselected population and should, therefore,
yield even higher degrees of relationship with criteria of
success. Naturally, this suggestion should be supported
by further research.
The results of this investigation appear to imply
that the optimum prediction of training success will re
sult from a variety of characteristics, i.e., aptitude,
biographical, personality, and perhaps many others which
have thus far been neither identified nor measured. How
ever, it would seem that unique aptitude predictor vari
ables offer more promise by far of significant relation
ships in their predictions for success in training than do
variables from personal history and personality character
istics, even though these non-intellectual variables
should not be overlooked.
190
Over 50 per cent of the variance from the Composite
Criterion of this investigation was accounted for by the
various best combinations of predictor variables taken at
different levels of training. This finding would suggest
that there is good promise in using the investigation's
results, especially the indicators from the Employee
Aptitude Survey tests, in increasing training efficiency
and in the guidance of students in electronics at all
levels. It also appears that the greatest significance
may be related to identification and guidance of students
at the earliest levels of training, including all stu
dents who are at the level of wanting to enroll in the
program.
Instructors' ratings of trainees after a three-
months' time span correlated significantly with final
grades assigned in training courses. In this investiga
tion, the instructors' Sum-of-Ratings score was used as
a criterion. However, would it not be desirable to con
sider using this rating score as a predictor of training
success? In any event, it would seem desirable for
instructors to use this rating device, or its equivalent,
early in the course, possibly after three weeks, and to
191
employ the results as one of the sources of information
that trainees could use in assessing their potentials in
the field. Ordinarily, after four or five weeks, college
students are not allowed to drop courses without receiving
an automatic "F” grade. If the rating results showed
extremely poor potential, perhaps below a sum or ratings
score of 20 as suggested from this sample’s data, trainees
could be assisted to drop the course without receiving an
"F”, and be allowed to enroll in other courses of study
without too much time lost. On the other hand, high
achievers would receive an official acknowledgment of
their strength and, for many of these trainees, this might
serve to reinforce their interest and ultimate achieve
ment .
Need for Further Study
As already suggested, the results of this inves
tigation should be cross-validated on other like samples
of electronics trainees. Expansion of the investigation
might well include a study of night school trainees and
trainees enrolled in industrial-type settings, or this
might be a separate investigation. Other predictor vari
ables should also be studied and correlated against train
192
ing success in electronics. Finally, reliability checks
should be made which would support the use of the ratings
of characteristics of motivation as predictors of training
success. It would seem that multiple criteria of training
success should be used with comprehensive predictor vari
ables in correlations of training success. Refinement of
the multiple regression equation could then be made and
applied with greater confidence.
APPENDIXES
APPENDIX A
INSTRUCTORS1 RATINGS AND
TRAINEE QUESTIONNAIRE
APPENDIX A
INFORMATION SHEET FOR INSTRUCTORS
General Information
The enclosed materials, the rating scales and ques
tionnaires, are part of a doctorate dissertation
study. The study is designed to determine the rela
tionship between certain student characteristics and
their success in electronics training.
We have decided to use your ratings of these students
as the criterion for determining "success" in the
training program. The nine characteristics which we
are asking you to rate for each student may overlap
'"to some extent, and the ratings will be reflected to a
certain extent in the grade you assign the student in
the course. However, we feel that we can obtain addi
tional information that only you are in a position to
judge from this, the rating scale.
The questionnaires are to be printed and returned to
you completed by the students who are in your first
semester electronics training class, or classes.
The Rating Scale
You are being asked to rate each individual student
in your first semester electronics training course.
(One class or several.) Only studenrs who are attend
ing school full time, during the day program, are be
ing included in this study.
195
196
In looking at the rating sheet you will note that
there are nine characteristics identified and described.
There are also five rating point categories. Number
one is the lowest and number five is the highest.
The meaning of these rating categories are described
as follows:
„ . . Approximate % of Your
Rating Category Description Qmdemt-Q
1 Very Weak Lowest 107.
2 Below Average Next Highest 207.
3 Average Middle 407o
4 Above Average Next Highest 207,
5 Very Strong Highest 1 0 7 o
There are three common errors which rating procedures
frequently encounter. These common errors are listed
below with a suggested approach to make in avoiding
them.
1. Leniency Error -- That is, rating too many indi
viduals average or above average. This error can
be minimized by trying to group approximate num
bers of trainees into the percentage categories
suggested.
2. Central Tendency Error -- That is, rating too
many individuals into the average category. Mini
mize this error by doing as suggested for the
leniency error.
3. Halo Effect Errors -- That is, halo effect takes
two forms. The first is that of rating an indi
vidual too high or too low on all characteristics
because he is outstandingly high or low on some
specific characteristic. The second type is that
of rating all individuals in the group too high or
197
too low because of the influence of one or two
outstandingly effective or poor individuals in
the group. Perhaps the best solution to this
common effect is to rate each characteristic for
each student in your entire class, or classes,
and then the next characteristic for each individ
ual in the entire class or classes.
Before you start rating you might figure out approxi
mately how many students might go into each of the
five categories for the nine characteristics. As an
example, if there were twenty-eight students in my
group, the actual number of students that I would at
tempt to get into each of the categories would be
something like this:
(For a total of 28 students)
Rating Category Approximate Number of Students
1 3 students
2 5 students
3 11 students
4 6 students
5 3 students
However, feel free to disregard this approximate per
centage grouping if your group isn’t a typical first
semester class.
The Questionnaire
1. Ask all students to print and complete all items
on the questionnaire.
2. Pick up the questionnaires on the same day that
they are distributed. However, extra copies are
supplied in case students fail to return the
198
first copy given to them. We need a 100% completed
return to have a valid study.
199
INSTRUCTOR’S RATING OF TRAINEES
Trainee________________Group_________ Date
Low to High
Characteristics--------------------------------- : -----
1 2 3 A 5
Trainee Achievement: (Understanding of
theory, vocabulary, L.C.& R)
Rating should be based upon results
of tests, class work, etc.
Trainee Achievement: Ability to use
hand tools, laboratory and testing
equipment. Achievement in the appli
cation of theory and skills
Skill in Precision Work: Care taken in
the accuracy of work
Ability to Stick to Detailed Tasks:
Willingness to stay with tasks until
solutions are discovered
Ability to Get Along with Other Train
ees ; Willingness to cooperate with
fellow trainees
Ability to Get Along with Instructor:
Attendance and completion of assign
ments. Willingness to cooperate.
Ability to Write Clearly: Ability to
express technical information
Understanding of Mathematics: Under
standing of mathematical principles
and skill in their application
Understanding of Technical Materials:
Ability to profit from technical
reading materials
SUM OF NINE RATINGS
201
TRAINEE QUESTIONNAIRE
INSTRUCTIONS: Please complete all questions and print.
If more space is needed please label item and put
on the back of the questionnaire.
Name ____________ Birthdate_____Age__
Last First Initial
Address:_________________________ Telephone No.:____________
High School Last Attended:_________________________________
Name Location
Education Major:__________ Number of Units Completed:_____
Number of Units This Semester:
Grade Expected This Course:____ ______
Have you had electronics training before:_________ Where:
Yes or No
How Much:
Occupation that you are likely to be employed in ten
years from the present:_____________________________________
Describe the duties that you are likely to be carrying out
in this occupation:__________________________________________
Father's Principal Occupation: (If deceased or retired,
kindly indicate the last principal occupation held)
Title:
201
TRAINEE QUESTIONNAIRE
INSTRUCTIONS: Please complete all questions and print.
If more space is needed please label item and put
on the back of the questionnaire.
Name :_______________ Birthdate_____Age_
Last First Initial
Address:_________________________ Telephone No. : _______
High School Last Attended:________________________________
Name Location
Education Major:__________ Number of Units Completed:____
Number of Units This Semester: ______
Grade Expected This Course:__________
Have you had electronics training before:_________ Where:
Yes or No
How Much:
Occupation*that you are likely to be employed in ten
years from the present:_____________________________________
Describe the duties that you are likely to be carrying.out
in this occupation:_________________________________________
Father's Principal Occupation: (If deceased or retired,
kindly indicate the last principal occupation held)
Title:
202
Describe his duties:
EMPLOYMENT EXPERIENCE (paid or unpaid)
Employer_____________From To Title or Duties Earnings
2.^
3
4.
5.
APPENDIX B
CURRICULUM REQUIREMENTS
204
APPENDIX B
Curriculum Requirements Industrial Electronics
El Camino College
First Semester
1. Basic Electronics, 1-A 8 units
2. English 1-A 3 units
3. Psychology 2 or 5 3 units
4. Physical Education 1/2 unit
Second Semester
1. Electronics 1-B 8 units
2 . English 1-B or Speech Arts 1 3 units
3. Health Education 1 2 units
4. Physical Education 1/2 unit
5. Elective* 3 units
Third i Semester
1. Electronics 1-C 8 units
2. History 1 3 units
3. Science Requirement 3 units
4 * Physical Education 1/2 unit
5. Elective* 2 units
Fourth Semester
1. Electronics 1-D 8 units
2.
Humanities Requirement 2 units
3. Political Science 1 3 units
4. Physical Education 1/2 unit
5. Elective* 3 units
Elective*
Mathematics recommended.
205
Curriculum Requirements Industrial Electronics
Santa Monica City College
First Semester
1. Basic Electricity and Lab 1-A 5 units
2 . Electronics Shop 1-B 2 units
3. Mathematics 17-A 3 units
4. Electives 5 units
Second Semester
1. AC Circuits, Transistor, etc., 2-A 5 units
2. Introduction to Network Analysis 2-B 3 units
3. Mathematics 17-B 3 units
4. Electives 4 units
Third Semester
1. Vac. Tube & Transistor Circ. 3-A 5 units
2 . Industrial Elect. Controls 3-B 5 units
3. Mathematics 17-C 3 units
4. Electives 2 units
Fourth Semester
1. Advanced Electronics Circuits 4-A 5 units
2 . Electronic Controls 4-B 5.un its
3. Electives 5 units
Electives: Elective courses discussed approved by
counselors.
APPENDIX C .
EATING SCALE FOR EMPLOYMENT
EXPERIENCE
APPENDIX C
Scale for Use in Rating Employment Experience
and Criteria of Quantity of Experience
SCALE OF EMPLOYMENT EXPERIENCE
_ . Rating
Descrxptive Category Score
1. No general or related experience. 1
2. Little general experience. 2
3. Some general experience. 3
4. Little related experience. 4
5. Much to very much general experience. 5
6 . Some related experience. 6
7. Some related and general experience
rated individually, but together
equalling much experience. 7
8. Much related experience. 8
9. Much general and related experience
rated individually, but together
equalling very much experience. 9
10. Very much related experience. 10
207
Criteria of Quantity of Experience
No experience
Little experience
Some experience
Much experience
Very much experience
Self explanatory
Less than 6 months
6 months to 1 year
1 year to 2 years
2 years or more
APPENDIX D
CLASSIFICATION OF OCCUPATIONS
APPENDIX D
U. S. CENSUS CLASSIFICATION OF OCCUPATIONS1
Rating Occupational Groups Included
6 Professional and Technical--
Accountants, Pilots, Architects, Artists,
Athletes, Authors, Clergymen, Dentists,
Designers, Dieticians, Draftsmen, Newspapermen,
Forestors, Pharmacists, etc.
5 Managers, Officials and Proprietors--
Buyers, Conductors, Credit Men, Pursers,
Administrators, etc.
4 Clerical and Sales--
Attendants, Librarians, Bookkeepers, Auction
eers, Peddlers, Sales Agents, etc.
3 Craftsmen and Foremen--
Bakers, Bookbinders, Masons, Foremen,
Inspectors, Plumbers, Projectionists,'etc.
2. Operatives--
Blasters, Boatmen, Brakemen, Cab Drivers,
Conductors, Grinders, Shoemakers, etc.
1 Service and Laborers--
Barbers, Bartenders, Cleaners and Choremen,
Firemen, Stewards, Laborers, Teamsters, etc.
^.S., Bureau of the Census (90).
210
APPENDIX E
SEMI-GUIDED INTERVIEW
APPENDIX E
SEMI-GUIDED INTERVIEW
INTERVIEWEE: DATE:
The purpose of this interview is to collect in
formation which may be used as a basis for rating
characteristics of motivation as described on the second
page.
Please describe why you wish to be a technician.
Discuss this occupation. What is its future? What does
it mean to you? How long have you been planning to enter
it, etc.?
Describe the plans that you have made in preparing and
your plans for entering into your chosen occupation.
212
213
Describe your favorite hobbies or activities which you
have had since you were about eight years of age.
Ratings of Motivation
1. Primary source of motivation:
Associations Benefits Activity itself
1 2 3
Little Some Much
1 2 3
2. Intensity of m o t i v a t i o n _________ ____ ____
3. Extent of planning ___________________
4. Extent of manifest interest
BIB LIOGRAPHY
BIBLIOGRAPHY
Berdie, R. F. "Differential Aptitude Tests as
Predictors in Engineering Training," Journal of
Educational Psychology, XLII (February, 1951),
114-123.
Birnbaum, Abraham H., Ardmore, Sidney J., and
Bolanovich, Daniel J. Prediction of Success in
Radio Repair Courses. U.S., The Adjutant General
Office (Army) , Personnel Research Branch,
Technical Reserve, No. 52., Government Printing
Office, 1955.
Blechner, Janet E., and Carter, Harold D. "Rorschach
Personality Factors and College Achievement,"
California Journal of Educational Research, VII
(March, 1956), 72-75.
Brice, Marshall M. "A Comparison of Subjective Pre
dictors with Objective Predictors of College
Achievement," Dissertation Abstracts, XVI (Sep
tember, 1956), 1622.
Briggs, Leslie J. Development and Appraisal of a
Measure of Student Motivation. U.S. Air Force
Personnel Research Center, Research Bulletin
No. 54-31. Washington: U. S. Printing Office,
1954.
Brokaw, L. D. "Comparative Validities of 'Short'
versus 'Long* Tests," Journal of Applied
Psychology, XXXV (October, 1951), 325-30.
Burgess, Elva. "Personality Factors in Over- and
Under-Achievers in Engineering," Dissertation
Abstracts, XVI (April, 1954), 536-43.
216
8 . California, State of, Department of Education, Super
intendent of Public Instruction. A Study of
Technical Education in California: Guidelines for
the Development and Operation of Technical Educa
tion Programs in the Junior Colleges, XXVIII, No.
7. Sacramento: California State Printing Office,
September, 1959.
9. . California Follow-up Study of Trade and
Industrial Education Students. California Bureau
of Industrial Education, Sacramento, California,
1959 (mimeographed).
10. . "Follow-up Study of Trade and Industrial
Education Students: A Cooperative Study Partici
pated in by Seventy-Two California Schools and
Bureau of Industrial Education, 1956 and 1957."
California Bureau of Industrial Education, Sacra
mento, California, 1960 (mimeographed).
11. . Mathematics and Science Competencies for
Technicians: A Study of the Training of Electronics
and Chemical Technicians with Special Emphasis
on Critical Mathematics and Science Requirements,
XXIX, No. 12. Sacramento: California State Print
ing Office, December, 1960.
12. . "Report of the California Steering Com
mittee on Electronics Education." California
State Department of Education, May, 1960 (mimeo
graphed) .
13. California Test Bureau. Manual S-Q Rorschach Test.
Preliminary edition. Los Angeles: California
Test Bureau, 1957-
14. Case, H. W., and Ruch, F. L. Survey of Spatial Rela
tions Ability., Form A . Revised Manual. Los
Angeles: California Test Bureau, 1944.
15. Chahbazi, Parviz. "Use of Projective Tests in Pre
dicting College Achievement," Educational and
Psychological Measurement, XVI (Annual, 1956).,
538-42.
217
16.
17.
18.
19.
20.
21.
22.
23.
24.
25 .
Chicago Institute for Research. Careers in Elec
tronics . Research No. 149. Chicago: Chicago
Institute for Research, 1956.
Clark, Selby G. "The Rorschach and Academic Achieve
ment," Personnel Guidance Journal, XXXVI (February,
1958), 339-41.
Coleman, William. "An Economical Test Battery for
Predicting Freshmen Engineering Course Grades,"
Journal of Applied Psychology, XXXVII (December,
1953), 465-67.
Crooks, W. R. "The Value of Personality Adjustment
Questionnaire Items in the Predicting of Outcomes
in College Engineering Training," Dissertation
Abstracts, XIII (February, 1953), 255-56.
Dailey, John T., and Brokaw, L. W. "The Comparative
Composite Validities of 'Short' versus 'Long'
Tests," American Psychologist, V (June, 1950),
288-89.
Drake, L. E., and Oetting, Eugene R. "An MMPI
Pattern and a Suppressor Variable Prediction of
Academic Achievement," Journal of Gounseling
Psychology, IV (Winter, 1957), 245-47.
Educational Testing Service. Manual for Interpreting
SCAT Scores. Los Angeles: Educational Testing
Service, Cooperative Test Division, 1957.
- Technical Report of SCAT Tests. Los Ange
les: Educational Testing Service, Cooperative Test
Division, 1957.
_________. 1958 SCAT-Step Supplement. Los Angeles:
Educational Testing Service, Cooperative Test
Division, 1958.
Fairweather, Paul. "Appropriateness of Field and
Level of Vocational Choice as Related to Self
Concepts, Intelligence, School Achievement, and
Socioeconomical Status." Unpublished Ph.D. the
sis, University of Southern California, 1960.
218
26.
27.
28.
29.
30.
31.
32.
33.
34.
Field, Lewis William. "Personality Correlations of
College Achievement and Major Areas of Study,"
Dissertation Abstracts, XIV (September, 1954),
1344-45.
Foy, Glenn Arthur. "A Study of the Relationship
Between Certain Factor-Analyzed Ability Measures
and Success in College Engineering," Unpublished
Ph.D. thesis, University of Southern California,
1959.
Franz, Gretchen, Davis, Junius, and Barcia, Dolores.
"Prediction of Grades from Pre-admissions Indices
in Georgia Tax-supported Colleges," Educational
and Psychological Measurements, XIII (Winter,
1958), 841-44.
Frink, J. W. "Improving the Prediction of Academic
Achievement by Use of the MMPI," Journal of
Applied Psychology, XXXIX (February, 1955), 49-52.
Gallese, Arthur James. "Personality Characteristics
and Academic Achievement in School of Engineering,"
Dissertation Abstracts, XIX (May, 1959), 3022.
General Motors Corporation. Can I Be a Technician?
Detroit: General Motors Corporation, 1959.
"Glamour Fades for Electronics Issues," Financial
World, October 11, 1961, p. 10.
Glendale Unified School District. A Study of the
Selection of Technical Students at Glendale Col
lege Under the National Defense Education Act.
Title VIII. Glendale, California: Glendale Uni
fied School District, 1960.
________ . "A Survey Conducted by the Glendale School
District of Skilled and Technical Occupation in
the Aircraft, Electronic, and Missile Industries."
Glendale, California: Glendale Unified School
District, February, 1960 (mimeographed).
219
35.
36.
37.
38.
39 .
40.
41.
42.
43.
Goldstein, Leslie. "Electronics Industry: In Transi
tion between Infancy and Maturity--What Next?"
The Magazine of Wall Street, CX (December 16, ~
1961), 356-57, 369-71.
Gordon, Mary Agnes. Stability of Final School Grade.
U. S. Airforce Personnel Training Research Cen
ter, Research Bulletin No. 54-11. San Antonio,
Texas: U.S. Airforce, Lockland Airforce Base,
1954.
________ . Interaction of Experience and Aptitude in
Predicting Success in Training Courses for Air
plane and Engine Mechanics. U. S. Airforce Per
sonnel Training Research Center, Research Bulle
tin 57-133. San Antonio, Texas: U.S. Airforce,
Lockland Airforce Base, 1957.
Guilford, J. P. Psychometric Methods. 2d ed. New
York: McGraw-Hill, 1954.
________ . Fundamental Statistics in Psychology and
Education. 3rd ed. New York: McGraw-Hill, 1956.
Halsey, Hugh. "The Predictive Value of Certain
Measures Used in Selecting Freshmen for the
Technical Curricula in a Community College,"
Dissertation Abstracts, XVII (March, 1957), 542.
Hancock, John, and Carter, Gerald. "Student Person
ality Traits and Curriculae of Enrollment,"
Journal of Educational Research, XLVIII (Novem
ber, 1954), 225-27.
Helme, W. H. Prediction of Success in Course Train
ing for Electronics and Electrical Maintenance
Jobs. U.S., The Adjutant General Office (Army),
Personnel Research Branch, Technical Reserve,
No. 91. Washington: Government Printing Office,
1958.
________ . Prediction of Success in Carrier Repair.
Teletype Maintenance and Power Equipment Mainte
nance Courses. U.S., The Adjutant General
220
44.
45.
46.
47.
48.
49.
50.
51.
Office (Army), Personnel Research Branch, Tech
nical Reserve, No. 83. Washington: Government
Printing Office, 1957.
Holland, J. L. "The Prediction of College Grades
from Personality and Aptitude Variables," Jour
nal of Educational Psychology, LI (October,
1960), 245-53.
Hountras, Peter T. "Relationship between Student
Load and Achievement," Journal of Educational
Research, LI (January, 1958), 355-60.
Isaacson, Lee E. "Predicting Success in the Work-
Experience Program," Personnel Guidance Journal,
XXXIII (January, 1955), 270-73.
Jones, Margaret Hubbard, and Case, Harry W. "The
Validation of a New Aptitude Examination for
Engineering Students," Educational and Psycholog
ical Measurement, XV (Annual, 1955), 502-508.
Karcher, E. Kenneth, Zeidner, Joseph, Brueckel,
Joyce E., and Bolanovich, Daniel J. Effect of
Weighting the Number of Tests on the Validity
of Aptitude Area Composites. U. S. Army, Person
nel Research Branch, No. 39. Washington: Govern
ment Printing Office, 1954.
Kennedy, Phyllis E. "The Validity of the School and
College Ability Test (SCAT) for Prediction of
College Achievement," California Journal of Edu
cational Research, IX (March, 1958), 67-71.
Kern, Donald Warren. "The Prediction of Academic
Success of Freshmen in a Community College,"
Dissertation Abstracts, XV (January, 1955), 85.
Klugh, Henry E., and Bierley, Robert. "The School
and College Ability Test and High School Grades
as Predictors of College Achievement," Educa
tional and Psychological Measurement, XIX (Winter,
1959), 625-26.
221
52.
53.
54.
55.
56.
57.
58.
59.
60.
Layton, W. L. Predicting Engineering Grades: Select
ing and Counseling of Students in Engineering.
Minneapolis: University of Minnesota Press, 1954.
Lebold, W. K. "Intellectual and Non-Intellectual
Factors in Predicting Academic Success," Journal
of Engineering Education, XLVIII (October, 1958),
514-19.
Long, Louis, and Perry, James D. "Academic Achieve
ment in Engineering Related to Selection Proced
ures and Interest," Journal of Applied Psychology,
XXXVII (December, 1953), 468-71.
Los Angeles Times, Business and Finance, October 21,
1960.
Malloy, John. "The Prediction of College Achievement
with the Life Experience Inventory," Educational
and Psychological Measurement, XV (Annual, 1955),
170-80.
Martoccia, Charles T., and Nelson, William H. Com
parison of Instructor Grade and Instructor
Expressed Opinion as Predictors of Student Suc
cess in Naval Air Flight Training. U. S. Navy
School of Aviation, Medical Research Notes, No.
NM 001 108 107, Rep. No. 3. Washington: Govern
ment Printing Office, 1956.
McDaniel, J. W., and Reynolds, W. A. "A Study of
the Use of the Mechanical Aptitude Tests in the
Selection of Trainees for Mechanical Occupations,"
Educational and Psychological Measurement. IV
(Annual, 1944), 191-97.
McGehee, W., and Moffie, D. J. "Psychological Tests
in the Selection of Enrollees in Engineering,
Science, Management Defense Courses," Journal of
Applied Psychology, XXVI (December, 1942), 584-86.
Motto, Joseph. "Stability of Work Experience as a
Predictor of Success," Personnel Guidance Jour-
nal> III (May, I960), 214-16.
222
61.
62.
63.
64.
65.
66.
67.
68.
69 .
70.
National Association of Manufacturers. Your Oppor
tunity in Industry as a Technician. New York:
National Association of Manufacturers, Educational
Department, April, 1957.
Orange Coast College. "Electronic Computer Survey."
Costa Mesa, California: Orange Coast College,
1960 (mimeographed).
Patterson, C. H. "The Prediction of Attrition in
Trade School Courses," Journal of Applied Psy
chology, XL (June, 1956), 154-58.
________ . "Predicting Success in Vocational School
Courses: A Review of the Literature," Educational
and Psychological Measurement, XVI (Annual,
1956), 352-400.
Porter, Richard J. "Predicting Vocational Plans of
High School Senior Boys," Personnel Guidance
Journal, XXXIII (December, 1954), 365-68.
Psychological Corporation, The. Norms Must Be
Relevant. Test Service Bulletin, 1948-1950,
Nos. 36-40. New York: The Psychological Cor
poration, 1950.
________ . Watch Your Weights. Test Service Bulle
tin, 1957-1958, Nos. 52-53. New York: The
Psychological Corporation, 1958.
Psychological Services, Inc. Manual of Employee
Aptitude Survey: A Battery of Practical Employ
ment Tests. Los Angeles: Psychological Services,
Inc., n .d.
Remmers, H. H., and Grieger, H. E. "Predicting Suc
cess and Failure of Engineering Students in the
School of Engineering in Purdue University,"
Purdue University Studies in Higher Education,
XXXVIII (May, 1940), 10-19.
Ritchie, Ralph W. "Electronics Training Survey."
Ontario, California: Chaffey College, 1958 (ditto
copy) .
223
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
R.L.W. ’’ Inside Industry,” Citizen News (Hollywood,
California), January 24, 1962.
Ruch, Floyd L., and Ruch, William W. "Predicting
Success in Draftsman Training with Short Time
Limit Aptitude Tests,” Educational and Psycho
logical Measurement, XX (Winter, 1960), 827-34.
Samuelson, Cecil 0., and Pearson, David T., Sr.
"Interest Scores in Identifying the Potential
Trade School Dropout," Journal of Applied Psy
chology, XL (October, 1956), 386-88).
Scollary, Robert W. "Personal History Data as a
Predictor of Success," Personnel Psychology, X
(Winter, 1957), 23-26.
Scudder, C. P., and Raubenheimer, A. S. "Are Stand
ardized Mechanical Aptitude Tests Valid?" Journal
of Juvenile Research, XIV (March, 1930), 120-23.
Siegel, A. I., and Schultz, D. G. Post-training
Performance Criterion Development and Applica
tion: Generalized Guttman and Thurstone Scales
for Electronic Job Performance Evaluation.
Wayne,'Pennsylvania: Wayne University, Applied
Psychological Services, 1960.
Siegel, Sidney. Non-Parametric Statistics. New York:
McGraw-Hill, 1958.
Soar, Robert S. "Personal History Data as a Pre
dictor of Success in Service Station Management,"
Journal of Applied Psychology, XL (October, 1956),
383-85.
Sopchak, Andrew L. "Prediction of College Perform
ance by Commonly Used Tests,” Journal of Clinical
Psychology, XIV (January, 1958), 194-97.
Spencer, George Minard. "An Investigation of Some
Non-intellectual Factors Presumably Affecting
the Academic Adjustment of College Students at
Florida State University," Dissertation Abstracts,
XV (August, 1955), 1436-37.
224
81.
82.
83.
84.
85 .
86 .
87.
88.
89.
Stone, David R., and Ganung, George R. "A Study of
Scholastic Achievement Related to Personality
as Measured by the Minnesota Multiphasic Inven
tory," Journal of Educational Research, L (Octo
ber, 1956), 155-56.
Stone, Joyce B. "Differential Prediction of Academic
Success at Brigham Young University," Journal of
Applied Psychology. XXXVIII (April, 1954), 109.
Stuit, D. B., Dickson, G, S., and Jordan, L. F.
Predicting Success in Professional Schools.
Washington, D. C.: American Council on Education,
1949 .
Super, Donald. "Vocational Development: The Process
of Compromise or Synthesis," Journal of Counsel
ing Psychology, III (Winter, 1956), 214-19.
_________. "The Use of Multifactor Test Batteries in
Guidance," Journal of Personnel Guidance, XXXIII
(September, 1956), 214-19.
Teahan, John E. "Future Time Perspective, Optimism
and Academic Achievement," Journal of Abnormal
and Social Psychology, LVII (November, 1958),
379-80.
Traxler, Arthur E. "Should SCAT Scat ACE?" Educa
tional Records Bulletin No. 67 ("1955 Fall Test
ing Program--Independent Schools and Supplementary
Studies”). New York: Educational Records Bureau,
1956, pp. 51-63.
Trueblood, Dennis Lee. "Effects of Employment on
Academic Achievement," Personnel Guidance Journal,
XXXVI (October, 1957), 481t86.
_________. "Selected Characteristics, Including Aca
demic Achievement, of Employed and Non-employed
Students in the Indiana University School of
Buisiness," Dissertation Abstracts, XIV (April,
1954), 643-44.
225
90. U.S., Bureau of the Census. United States Census
of Population: 1950. California: General Char
acteristics , V, B. Washington: Government Print
ing Office, 1952.
91. U.S., Department of Health, Education, and Welfare;
Office of Education. Electronic Technology.
No. 0E-80009. Washington: Government Printing
Office, 1960.
92 . Warner, W.. L, , and Lunt, P. S . The Social Life of
a Modern Community. Yankee City Series, Vol. I.
New Haven: Yale University Press, 1941.
93. Meeker, M. , and Eells, K. Social Class in
America. Chicago: Science Research Associates,
1949.
94. Warren, N. D., Dossett, W. F., and Ford, J. L. A
Correlational Analysis of Achievement in a Gener
alized Electronics Troubleshooting Course.
Research Report AFPTRC-TN-57-148, ASTIA Document
No. AD 146424. San Antonio, Texas: Air Research
and Development Command, Lockland Air Force Base,
December, 1957.
95. Weeks, James. "The Predictive Value of ACE and
SCAT," Personnel Guidance Journal, XXXVIII
(September, 1959), 314-317.
96. Wesman, Alexander G., and Bennett, George. "Multiple
Regression versus Simple Addition of Scores in
Prediction of College Grades," Educational and
Psychological Measurement, XIX (Spring, 1959),
243-46.
97. Western Data Processing Center. BIMD Multiple Cor
relational Analysis, No. 2,9. Los Angeles: Gradu
ate School of Business Administration, University
of California, 1958.
98. Wood, Helen. "The Manpower Future: Its Challenges
to Vocational Guidance," Personnel Guidance
Journal, XXXVIII (December, 1959), 300-04.
226
99. Zachert, Virginia, and Levine, A. '’ Education and
Prediction of Military School Success," Journal
of Applied Psychology, XXXVI (August, 1952),
266-68.
100. Zeaman, Jean. "Some of the Personality Attributes
Related to Achievement in College," Dissertation
Abstracts, XVIII (January, 1958), 290-91.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

An Analysis Of The Role Of The School Psychologist In The State Of California

PDF

Utilization Of The Earliest Childhood Recollection In Detecting Maladjustment Among Junior College Students

PDF

A Study Of The School Achievement And Adjustment Of Children From One-Parent Homes

PDF

The Stability Of The Self-Concept In Junior College Students

PDF

A Study To Compare The Effectiveness Of Individual And Group Counseling Approaches With Able Underachievers When Counselor Time Is Held Constant

PDF

An Investigation Of Selected Correlates Of Academic Achievement Among Seventh-Grade Pupils

PDF

'Cost' And 'Utility' In The Prediction Of The Successful Junior College Student

PDF

To Attend Or Not To Attend College: Some Factors In The Decision Of Qualified High School Graduates

PDF

A Comparison Of Students' Evaluation Of Guidance Services Under Varying Plans Of Organization

PDF

The Validity Of The Graduate Record Examinations As Used With English-Speaking Foreign Students

PDF

An Empirical Comparison Of The School And College Ability Tests And The American Council On Education Psychological Examination

PDF

The Rehabilitation Of School Dropouts Through An Intensive Summer School Program

PDF

Upperclassmen As Academic Advisers To Freshmen In An Undergraduate College: An Experiment

PDF

Factor Analysis Of Certain Aptitude And Personality Variables

PDF

An Experimental Study Of Self-Confrontation In Counseling

PDF

Non-Cognitive Factors Related To Achievement Which Can Be Ascertained From Freshman Autobiographies

PDF

A Comparative Guidance Study: Group Counseling Methods With Selected Underachieving Ninth Grade Students

PDF

Positional Response Set In The Multiple Choice Examination

PDF

A Study Of The Effectiveness Of Teaching Methods Of Study To Selected High School Freshmen

PDF

Nonverbal Communication In Counseling: An Exploratory Study

Asset Metadata

Creator Broe, John Richard (author)

Core Title Prediction Of Success In Training Among Electronics Technicians

Degree Doctor of Philosophy

Degree Program Education

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag education, educational psychology,OAI-PMH Harvest

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Lefever, David Welty (committee chair), Carnes, Earl F. (committee member), Ruch, Floyd L. (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c18-261251

Unique identifier UC11358643

Identifier 6206040.pdf (filename),usctheses-c18-261251 (legacy record id)

Legacy Identifier 6206040.pdf

Dmrecord 261251

Document Type Dissertation

Rights Broe, John Richard

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA