Citation
Factors affecting persistence in education and employment of occupational therapists

Material Information

Title:
Factors affecting persistence in education and employment of occupational therapists
Creator:
Stewart, Carolyn Dockstader, 1943-
Publication Date:
Language:
English
Physical Description:
x, 84 leaves : ; 28 cm.

Subjects

Subjects / Keywords:
Academic achievement ( jstor )
College students ( jstor )
Colleges ( jstor )
Internships ( jstor )
Mathematical dependent variables ( jstor )
Mathematical variables ( jstor )
Modeling ( jstor )
Nursing students ( jstor )
Occupational medicine ( jstor )
Registry ( jstor )
Curriculum and Instruction thesis Ph. D
Dissertations, Academic -- Curriculum and Instruction -- UF
Occupational therapists ( lcsh )
Persistence ( lcsh )
Prediction of occupational success ( lcsh )
Prediction of scholastic success ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 68-77.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Carolyn D. Stewart.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
023342390 ( ALEPH )
06571546 ( OCLC )

Downloads

This item has the following downloads:


Full Text


















FACTORS AFFECTING PERSISTENCE IN
EDUCATION AND EMPLOYMENT OF
OCCUPATIONAL THERAPISTS






By

CAROLYN D. STEWART


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY










UNIVERSITY OF FLORIDA


1979
























DEDICATED TO
ALICE C. JANTZEN



An educator who has brought the future to the field
of occupational therapy. An inspiration to those who are
fortunate enough to be touched by her.
















ACKNOWLEDGMENTS


The researcher wishes to express sincere appreciation to those

persons who had the patience and persistence in helping to make this

study possible. Since most of the preparation was done by correspon-

dence, a debt of gratitude goes to members of the faculty of the

University of Florida and the University of Montana.

Dr. Alice Jantzen provided initial encouragement and inspiration

for entering graduate school. Dr. Myron Cunningham contributed his

personal warmth and appreciation of the educational process and served

as an encouragement to proceed past the masters level.

Dr. James lensel, Chairman of the supervisory committee, provided

efficient guidance of the doctoral program and coordination of the

supervisory committee members. His inspiration, helpfulness, and

willingness to go beyond the input that is normally expected of an

advisor in terms of meeting requirements is held in deep appreciation.

Without Dr. Hensel my candidacy would not have been possible.

Dr. Margaret Morgan gave valuable assistance with the technical

aspects of the study and the manuscript. )r. Lela Llorens contributed

knowledge of her specialty to add to the depth of the study.

Dr. Larry Gianchetta of the University of Montana was extremely

generous in contributing his assistance to the statistical design of

this study. Dr. Gianchetta and Dan Ballas, at the University of

Montana computer center, were most helpful in the technical analysis of

the data.

















TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS ........................ iii

LIST OF TABLES .......... ......................... .. vi

LIST OF FIGURES .......... ........................ vii

ABSTRACT ............ ............................ viii


CHAPTER

I INTRODUCTION ........ ...................... 1

General Background and Statement of Problem .. ..... 1

General Purpose and Need for the Study .... ........ 5

Need for Study .......... .................... 6

Assumptions ......... .. ..................... 8

Hypotheses ........... ...................... 8

Limitations of the Study ....... ............... 9

Definition of Terms ........ ................. 10

Significance of the Study .... .............. 11

Organization of Subsequent Chapters .. ......... 12

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

Academic Performance, Persistence, or Withdrawal 14

Practical Application ..... ................ 19

Tests of Professional Knowledge ... ........... 20

Professional Performance and Persistence ........ .. 22

Concluding Remarks ...... .................. 24









Page

III DESIGN AND METHODOLOGY ...... ................. .. 26

Population ........ ...................... 26

Data Collection ...... ................... 29

Method of Analysis ...... .................. 32

Definitions of Statistical Terms ... ........... 36

Pitfalls of Regression and Correlation Techniques . 41

Organization of Data ...... ................. .. 42

Summary ........ ....................... 42

IV ANALYSIS OF DATA AND RESULTS .... .............. 43

Introduction ....... ..................... 43

The Computer Programs ..... ................ 44

Stepwise Multiple Regression Analysis ........ 46

Model I--National Registry Examination Score
(REGIST) ....... ..................... 46
Model Il--Psychiatric Internship Score (PYCHIN) 48
Model Ill--Physical Disabilities Internship
Score (PDINT) ...... .................. 50
Model IV--Work Status Score ... ............ 51

Stepwise Discriminant Analysis ... ............ 53

Model V--Complete and Not Complete the Program
for Registry in Occupational Therapy ........ .. 53
Models VII and VIII--Working or Not Working in
the Field of Occupational Therapy ........ .. 57

V CONCLUSIONS, OBSERVATIONS AND RECOMMENDATIONS ..... 60

Conclusions ....... .................... 60

Observations ....... ..................... 65

Recommendations ...... ................... 67

BIBLIOGRAPHY .......... .......................... 68

APPENDIX ........... ............................ 78

BIOGRAPHICAL SKETCH ........ ...................... 84
















LIST OF TABLES


Table Page

1 Work Status Scale ........ ...................... 35

2 Model I--Variables Entered by Step ..... .............. .47

3 Model I--Score Comparison for National Registry
Examination-Actual and Predicted .... ............. 48

4 Model II--Variables Entered by Step .... ............. 49

5 Model II--Score Comparison for Psychiatric
Internship-Actual and Predicted ..... .............. 50

6 Model III--Variables Entered by Step ... ............ 51

7 Model III--Score Comparison for Physical Disabilities
Internship-Actual and Predicted ..... .............. 52

8 Model IV--Variables Entered by Step .... ............. 52

9 Model IV--Score Comparison for Work Status
Actual and Predicted ....... ................... 53

10 Models V and VI--Classification Functions ... .......... 54

11 Models V and VI--Classification Matrix ... ........... 54

12 Models V and VI--Jacknifed Classification ... .......... 55

13 Models V and VI--Summary-Variables Entered by Step ..... 56

14 Model V--Score Comparison for Complete-Actual and Predicted 56

15 Model VI--Score Comparison for Did Not Complete-
Actual and Predicted ....... ................... 57

16 Models VII and VIII--Classification Functions .. ........ 58

17 Models VII and VIII--Classification Matrix .. ......... 58

18 Models VII and VIII--Jacknifed Classification .. ........ ..58

19 Frequency of Independent Variable Use .... ............ 64
















LIST OF FIGURES


Figure Page

1. Simple Linear Regression ..... ................ 38

2. Multiple Regression ....... ................... 39

3. Discriminant Function ...... .................. .. 40















Abstract of Dissertation Presented to the Graduate
Council of the University of Florida in Partial
Fulfillment of the Requirements for the
Degree of Doctor of Philosophy



FACTORS AFFECTING PERSISTENCE IN
EDUCATION AND EMPLOYMENT OF
OCCUPATIONAL THERAPISTS

By

CAROLYN D. STEWART

December 1979


Chairman: Dr. James W. Hensel
Major Department: Division of Curriculum and Instruction

The purpose of this study was to explore independent variables to

determine which, if any, could be used to predict success in an occupa-

tional therapy curriculum, internship, registry examination, and job

persistence prior to the time professional course work is begun for use

in student selection. Ultimately the development of a process model

which could be used with future data for selection of candidates was a

desirable outcome.

The 169 subjects in this study all had been accepted into the

occupational therapy program for graduating classes spanning the five

year period from 1969 to 1973. Of the 160 women and nine men, 19 with-

drew from the undergraduate curriculum. The 25 selected variables

occurring prior to professional course work were taken from University

of Florida transcripts which were available for all subjects. These


viii









variables were used as probable factors explaining the movement of the

dependent variables.

The study was executed in two statistical parts. The first part

investigated the best pre-professional course work indicators for scores

on the following dependent variables:

1. Psychiatric Internship

2. Physical Disabilities Internship

3. National Registry Examination

4. Work Status Follow-up

Part one of the study was designed to compute a stepwise multiple

regressional analysis using the Biomedical Computer packaged Programs P

Series, BMDP2R.

The second part used stepwise discriminant analysis to find the

best indicators of completion and non-completion of the occupational

therapy program and the best indicators for work persistence and non-

persistence in the field of occupational therapy using Biomedical

Computer packaged program, BMDP7M.

The multiple regression analysis of the 25 variables showed five

statistically significant at the .05 level: college entry test score,

college grade point average for the first two years, transfer student or

entry as a University of Florida freshman, number of credit hours in

English and literature and number of credit hours in psychology. The

discriminant analysis used all but one of the same variables (CEEB) and

added one more: number of credit hours in music. In each case the

eight corresponding null hypotheses were rejected.

AlthougLi statistically s significant variables were revealed, they

did not explain a high enough percentage of the movement in the










dependent variables to be useful for practical application. This was

demonstrated for each dependent variable by a random sample of subjects'

actual data used in each predictive equation. The accuracy of the

discriminant analysis was far less than might have occurred by chance.

The results of the study indicated that combining both academic and

practical application parameters in both the independent variables and

the dependent variables does not enhance the practical use of the

predictive selection equation versus using just one of these parameters

in both. It was found that the same five variables were statistically

significant for both the academic and the practical application dependent

variables. Implications of the study were that a few variables could

be considered in the selection of potential candidates: college grade-

point average for the first two years, transfer student or University of

Florida freshman, number of credit hours in English and literature,

College Entrance Examination Board Test score, number of credit hours

in psychology, and number of credit hours in music. However, human

factors and subjective data should also be included. In conclusion, this

researcher found that a practical predictive equation cannot be developed

with the objective data used in the study.
















CHAPTER I


INTRODUCTION


General Background and Statement of Problem


Accountability was a movement of the 70s. The ultimate account-

ability of an expensive technical education may be measured by

statistics showing a high percentage of completion of the program and

persistence within the field.

The concepts of predicting persistence and graduation from college

have been the substance of many studies, especially in the last ten

years. Since a 50 per cent attrition rate of college students prior to

graduation represents a great loss of time, money and energy to all

concerned (Morgan 1970), there has been a thrust to find ways to

decrease this high rate. Predicting professional knowledge, application,

and persistence have also been studied. However, combining all these

variables, both academic and applied measures, into one study on the

same population has rarely been done. This researcher did not find

information about such a study that had thus far been completed.

The objective of streamlining the complex process of selecting the

best qualified students who will complete oversubscribed programs has

been a part of accountability. Selection processes for persistence

in working in the field past the first year have not been adequately

addressed. To expedite the selection process, to cut valuable faculty











time spent in selection, and to keep prejudices to a minimum, it would

be desirable to be able to use information that has been collected and

is available on institution transcripts. Most past studies involved

administration of additional tests of questionnaires to prospective

candidates. If specific objective data from transcripts could be used

as valid predictors of completion and persistence in the work field,

unskilled office help could do most of the selecting of candidates.

Faculty could then be minimally involved making the final decisions

concerning the few borderline cases.

In a survey of 39 out of the 40 schools contacted for bachelor's

degree occupational therapists in 1974, Johnson, Arbes, and Thompson

found that 17 percent of the schools accepted fewer than half of their

respective applicants. They also found that:

1. All but three used interviews

2. All used grades in college for the first two years

3. All used aptitude and achievement test scores

4. About half used letters of recommendation

5. Ten used an application letter or essay

6. Seven used a personality or interest inventory

7. All used high school grades

8. Two used a biographical questionnaire

9. One used a lottery

The University of Florida Department of Occupational Therapy

annually has 35 openings for students. The Office of the Registrar

sends that department approximately 100 applications. The selection

procedure is similar to that used by most other occupational therapy










programs. An error in judgment in selecting a student who will later

drop out of a program or field of occupational therapy represents a

personal and professional loss of two people: the one who dropped out

and the one who missed the original opportunity to take that opening

(Morgan, 1970).

How can the best students for the health-related professions be

selected? What factors, such as abilities, past experiences, person-

ality attributes, styles of life, and interests, are related to

proficiency in a professional world? What information about an

individual is associated with or will permit prediction of future

competencies in his training and persistence in his health care field?

Are specialized programs training students who will remain in that work

force? With the large numbers of applicants, one knows that many are

not qualified academically or personally. However, a significant number

are. How then does one objectify in part or in toto this process in

order to select those most suitable?

Historically the public wants the best health service available

regardless of the expense. Encouraged by the mass media, the public

expects services from the health system that it did not know existed

prior to the 1970s. People of the 70s began demanding a drastic

increase in health service, including comprehensive health care for all.

Health care delivery system became a catch phrase with a powerful impact.

Mase, as reported by Brown (1972), stated that health care was the second

largest industry in the United States. This demand for care precipitated

a drain on the health system and created a need for additional allied

health manpower to provide support services for physicians and dentists








4

in the health care delivery system. DuVal reported some 34,000 graduates

of allied health educational programs in 1971, compared with about 6,000

in 1960. The development of health manpower has meaning, not only to

society, but to education and the individual (Brown, 1973).

College students have also shown an increased interest in the

health care fields. An interest survey by the American Council of

Education in the fall of 1972 indicated fulltime freshmen showed 1.2

percent more interest in health care fields than those of the previous

year (Lind, 1970:91). The American College Testing (ACT) Center in Iowa

shows that for the 1977-78 school year, 17 percent of the high school

juniors and seniors tested identified health professions as a chosen

field. The number of qualified potential candidates has grown to the

point where traditional selection criteria are inadequate for several

health fields. Since the health-related curriculum is relatively more

expensive than the "academic" curriculum, selecting students who will

complete the program and be successful within their chosen health field

becomes a major problem.

Private and public funding agencies want to know where, how, and to

what effect tax dollars or private dollars are being spent. Cost

effectiveness in the health fields is of increasing concern. In addition

to financial accountability, funding agencies demand service account-

ability such as how many trainees actually go to work in their field.

How many who enter their field continued to work in that field?

Therefore, predictive studies should include job entry and long term

persistence in the chosen health field as part of successful selection

processes.











This study explored possible combinations of variables, such as

undergraduate courses, college entrance examination scores and other

transcript data as indicators of students who completed and worked in

one health care field--occupational therapy. All of the students in

this study graduated from the University of Florida.


General Purpose and Need for the Study


A need exists to objectively identify prospective college students

who would be able to complete the requirements for registration in

occupational therapy, enter the job market and continue within the field

of occupational therapy. The purpose of the present study was to explore

independent variables such as college entry test scores, courses taken,

and grade point averages to determine which, if any, could be used prior

to the beginning of professional course work as indicators of those

who would be most likely to graduate, then be employed, and continue

working in the field of occupational therapy.

Previous studies sought predictors of (1) academic performance,

persistence, or withdrawal, (2) performance in practical application of

training such as internships, (3) professional knowledge as measured by

written tests, (4) short term professional performance and long term

persistence. None of these studies included all four variables on the

same population. This study sought to generate meaningful results by

using variables from each category as they related to a larger number

of subjects (n = 169). More specifically the purposes were to

1. Identify the best pre-professional indicators of persistence

in working in the field of occupational therapy.











2. Identify the best pre-professional indicators

of the program in occupational therapy.

3. Identify the best pre-professional indicators

of Physical Disabilities Internship.

4. Identify the best pre-professional indicators

of Psychiatric Internship.

5. Identify the best pre-professional indicators

passing the National Registry Examination.

6. Identify the best pre-professional indicators

in the field of occupational therapy.


of completion



of completion



of completion



of success on



of work status


Need for the Study


The training of an occupational therapist is expensive in time,

money and equipment, encompassing a four-year to five-year educational

program. Therefore, the selection and retention of successful students

should be studied to help identify criteria for the selection of future

applicants for training. Selection of potential occupational therapists

has been primarily the domain of educational personnel because career

entry has required completion of a formal, accredited educational

program. Many areas of student ability and background have been con-

sidered as elements of prognosis of success in college and for gradua-

tion from college. Methods employed have included extensive use of

written applications, recommendations, and personal interviews, consuming

great amounts of faculty time. Criteria for selection have emphasized

academic achievement, participation in extracurricular activities, and

personal qualities.











The "Reports on Curriculum Study" issued in 1963, which reported

on the 1958 to 1963 American Occupational Therapy Association study,

stated that student selection is considered one of the eight major areas

to include in future planning of curricula. Several studies have been

conducted on student selection for occupational therapy including those

by Bailey et al. (1969), Booth (1957), Crane (1962), Engelhart (1957),

Hendrickson (1962), and Lind (1970). However, results have not

been uniformly positive (Lind 1970). Negative correlations have

resulted within a few. Research on student selection procedures in

occupational therapy has been relatively meager and limited in scope.

Most of the studies on this topic have been based on small samples

ranging from 18 to 37 students (Johnson RW et al. 1974). Sample sizes

of at least 100 and preferably 200 students should be obtained from

successive classes at the same institution or by the use of comparative

data from several institutions (Johnson RW et al. 1974). Pedrini and

Pedrini (1974) concluded that there is a need for multivariate research

after an extensive review of the literature related to college success.

Prendergast (1978) concurred with this need in occupational therapy.

The ultimate impact for change in selection processes will depend

on the results of longitudinal research studies which will provide data

on the success of applicants in the field (Bistreich 1977). Johnson JA

(1974) pinpointed that the education of clinicians who can move with the

trends and continue in the field of occupational therapy should be a

major emphasis.

In 1977 Morgan stated the need to identify characteristics of

effective practitioners:











There is a need to determine if certain characteristics
discriminate between those who remain in a field and those
who drop out. (p. 72)

The present study was needed to determine if objective factors

obtained from records available at the time of selection could be used

as criteria to predict the completion of training, entry into, and

persistence in the field of occupational therapy.


Assumptions


1. It was assumed that the transcripts of the students involved in

this study at the University of Florida were accurate.

2. It was assumed that scores and grades which appeared in the

student's personnel file located in the occupational therapy

department were accurate.

3. It was assumed that the data received from the American

Occupational Therapy Association were accurate.

4. It was assumed that follow-up studies are a practical means of

determining strengths and weaknesses of selection processes.

5. It was assumed that academic success is related to occupational

success.


Hypotheses


Four generalized, multivariate null hypotheses were proposed for

multiple regression analysis:

It will not be possible to predict scores for

(I) the National Registry Examination

(II) Psychiatry internship










(III) Physical disabilities internship

(IV) Work status

on the basis of multiple combinations of variables which include college

entry test scores, courses taken, grade point averages, and other

objective data taken from transcripts.

Four generalized, multivariate null hypotheses were proposed for

discriminant analysis:

It will not be possible to discriminate among groups of students

(V) those who completed and (VI) those who did not complete the occupa-

tional therapy program on the basis of linear combinations of variables

which include college entry test scores, courses taken, grade point

averages, and other objective data taken from transcripts.

It will not be possible to discriminate among groups of students

(VII) those who were working and (VIII) those who were not working in

the field of occupational therapy four to eight years later on the basis

of linear combinations of variables which include college entry test

scores, courses taken, grade point averages, and other objective data

taken from transcripts.


Limitations of the Study


1. This study was limited to obtaining, organizing and presenting

data of students who graduated from the University of Florida in

the Occupational Therapy curriculum from 1969 to 1973. During

this period the course training and faculty from year to year were

stable. Subsequent to 1973, there were major changes in the

faculty and curriculum.











2. This study was concerned with 169 students' records, including

9 men and 160 women.

3. The students represented various environmental and educational

backgrounds. However, the only minority group represented was of

Spanish surnames (n = 3).

4. This study was not intended to determine the effect that the

different student backgrounds had upon individuals in the study

group. Nor was it intended that particular school systems should

be compared to one another as reflected by the achievement of their

graduates. Rather, the purpose was to review, as objectively as

possible, a group of students in an effort to determine which of

those would have difficulty in the pursuit of a specific college

program.

5. No attempt was made to evaluate the content of specific courses.

6. No attempt was made to appraise the teaching methods of individual

faculty members.

7. This study did not investigate why students left the curriculum.


Definition of Terms


The following is a list of terms to be defined as used within the

text of this study.

1. ACT--American College Test for scholastic aptitude.

2. AOTA--American Occupational Therapy Association.

3. CEEB--College Entrance Examination Board. The sponsoring body for

the Scholastic Aptitude Test (SAT).

3. Complete the occupational therapy program--to complete all under-











graduate training plus internships and the National Registry

examination to be eligible for Registry in AOTA.

5. GPA--Grade Point Average. The average is computed by dividing the

total number of grade points by the total number of course credit

hours attempted.

6. Non-persistence in working in the field of occupational therapy--

not working in the field of occupational therapy four to eight years

after graduation.

7. Persistence in working in the field of occupational therapy--working

in the field of occupational therapy four to eight years after

graduation.

8. Physical Disabilities Internship--a three-month supervised on-the-

job training and learning situation approved by AOTA. This intern-

ship is required as partial fulfillment for completion of the

occupational therapy program prior to the National Registry

Examination.

9. Psychiatric Internship--a three-month supervised on-the-job

training and learning situation approved by AOTA. This internship

is required as partial fulfillment for completion of the occupa-

tional therapy program prior to the National Registry Examination.


Significance of the Study


With the contemporary emphasis on accountability, all schools are

being asked to justify their continuation based upon their purpose,

methods of achieving that purpose, and allocation of money, personnel

and equipment, thereby scrutinizing priorities and responsibilities to









12

the individual and society. A faculty could benefit from being able to

select students who would continue to work in their field of training.

A persistent effort is needed in the departments of all health

related professions to eliminate any waste of time, money and resources

among students who plan to enter health field training or further their

present training. Through the 1970s, few longitudinal studies have

been done which followed subjects through college and into their work

experiences.

The writer recognizes the differences among various institutions

and curricula, but the individuals and conditions involved in the

Department of Occupational Therapy at the University of Florida may

represent the entire field. The University of Florida accommodates

some 30,000 students in all phases of university education. The De-

partment of Occupational Therapy serves approximately 100 occupational

therapy undergraduate majors per year who are juniors, seniors or

interns. The first two years of the occupational therapy program are

in liberal arts. The four-and-one-half year undergraduate program leads

to the Bachelor of Health Science in occupational therapy. Should the

conclusions of this study be desirable for use in another health field

or another institution, modification would undoubtedly be needed to suit

the unique institutional and curricular characteristics.


Organization of Subsequent Chapters


Chapter II contains a review of the literature concerning comple-

tion of training and job success in education, psychology, sociology

and allied health fields.








13

Chapter III describes the manner in which the data were obtained

and processed, specifically dealing with the design and mathematical

methodology. Chapter IV contains the presentation of these data while

Chapter V contains a summary of the results, conclusions, observations

and recommendations.
















CHAPTER II


REVIEW OF RELATED LITERATURE


The literature related to this study was reviewed within four

major categories. They were (1) academic survival, performance, and

persistence, (2) performance in practical application of training such

as internships, (3) professional knowledge as measured by written tests,

and (4) short-term professional performance, and long-term persistence.

The studies were all concerned with specific training curricula at the

junior college, college, and graduate levels of higher education,

psychology, sociology, and allied health.


Academic Performance, Persistence, or Withdrawal


The literature abounds with studies seeking predictors of college

performance, persistence, or withdrawal. Some notable ones by Aiken

(1968), A'Hearn (1974), Bowers and Loeb (1972), Green (1962), and Hinkle

and Houston (1977) all found that the best predictors of future academic

achievement were previous achievement and intellectual characteristics,

meaning: high school rank, American College Testing (ACT) and Scholastic

Aptitude Testing (SAT) test scores, reading test scores, and Adjective

Check List scores. Their focus was toward first semester or first

quarter college grades.

In studying persistence and withdrawal of students at the end of the

freshman year of college Ilumphreys et al. (1973), Knopf (1972), Knauer








15

(1968), and Reed et al. (1972) concurred that lower ACT scores and lower

high school rank were more closely related to withdrawal than higher rank

and scores. Less than 30 per cent of the lower scoring group stayed to

the second semester of the junior year (Humphreys et al. 1973).

Hall and Coates (1973), Pedrini and Pedrini (1974), and Pfeifer and

Sedlacek (1974) stated again that high school achievement and aptitude

test scores predicted first year college performance for racial minority

groups about the same as for whites. Bistreich (1977) found overall high

school grade point average (GPA) and high school GPA in science and

English were significant predictors of junior college GPA. Inkenberry

(1961) tried to predict first year college persistence using noncognitive

variables and intellective functions. He also found that the intellec-

tive function was the most significant variable with a part of the

cultural area: sex and social background.

Brazziel (1977) and Chase and Johnson (1977) both found high school

rank to be a statistically significant predictor of success in college

through the second year. They found two other significant variables:

hours of high school English and level of family income.

Within specific technical curricula, R. Brown and Stones (1972)

found no close association between intelligence factors and withdrawal

rates in the first year of nurses' training. Davies and Khosla (1974)

studied health visitor students using intelligence test scores as a

predictive tool for those who performed at high levels in academics.

Their multiple correlation was R = .52.

Reed et al. (1973) found some noncognitive variables to be signifi-

cant predictors of first year success of nursing students: age, year











of entry into the program, level of previous education, and schools

attended. Stress and satisfaction of first year nursing students were

assessed by a questionnaire developed by Katzell (1968). Statistically

significant negative correlations were shown between withdrawal and

experienced satisfaction and expectations. First semester grade point

average in a physical therapy program was best predicted by the Educa-

tional Interest Inventory Fine Arts Scale and the score of the Attitude

Toward Disabled Persons Scale (ATDPS) (Buca 1967). In a study by

Schimpfhauser and Broski (1976) of first year academic success in a

health related professions program the ACT score was found to be the

best predictor. Lowenthal et al. (1977) found that the Pharmacy College

Admission Test (PCAT) accounted for 25 per cent of the variation in

completion of a pharmacy curriculum. However, Liao and Adams (1977)

found that prepharmacy academic performance was the best predictor of

first year pharmacy CPA, not the PCAT. Shalik (1978) presented a con-

vincing study showing no statistical difference between the usual

selection process for a physical therapy program and a stratified

sampling process for student selection.

Specific engineering course grades and GPA were best predicted by

the Achievement and Affiliation scales of the Edwards Personal Prefer-

ence Schedule (EPPS) and the scores on American Council on Education

Psychological Examination (Muchinsky 1973).

In studying transfer students, Nickens (1972) stated that junior

college CPA was the best predictor of first term Florida State

University CPA.











Baker (1975), Bailey (1968), Bean (1975), Holmstrom (1975), and

Lee and Shepard (1973) conducted descriptive studies of students who

were persistors and withdrawals of programs in health related fields.

The persistors and graduates were generally better adjusted, more inde-

pendent, self-confident, responsible, and intelligent than those who

withdrew.

Prediger (1965), Reavley and Wilson (1972), and Rose (1965) found

that counseling served as a deterrent to freshman withdrawals. However,

Prediger found that the past achievement dimension made the most

difference.

Morgan (1970) used Omnibus Personality Inventory factors and ACT

scores to discriminate between groups for persistence in college seven

semesters after beginning. Her results showed that students differed by

sex in personality configurations, ability, and college persistence.

However, the best predictive variables for classifying the students into

groups were ACT scores, followed by an authoritarianism measure. Aldag

and Martin (1975), Cypres (1969), and Wittmeyer et al. (1971) concurred

that ACT scores differentiated the students who completed the college

curriculum and those who withdrew. Cypres suggested that high school

class rank should also be a part of a selection process. In addition,

Wittmeyer et al. found that students who completed the program tended to

be less independent and less venturesome as measured by the 16 Person-

ality Factor Indicator (PFI) and the Myers-Briggs Type Indicator.

Johnson (1970) used the 16 PFI, Strong Vocational Interest Blank (SVIB),

and a college qualification test to predict nursing theory and practice

grades. His results showed that abstract reasoning ability was the best

predictor. Completion of a teacher training program was studied by










Belcastro (1975) in which he used the Edwards Personal Preference

Schedule (EPPS) and the SVIB. The Autonomy Scale of the EPPS and the

Architect Scale of the SVIB ranked second and third, respectively. In

first place was the students' GPA. One- to four-year success in a

health education program was again determined to be best predicted by

GPA by Heit et al. (1978). Success in graduating from a three year

pharmacy program was also predicted by prepharmacy GPA (Palmieri 1976).

Burgess (1972), Lewis and Welch (1975), and Owen and Feldhusen (1970)

found that prenursing CPA was the best indicator of nursing course grade

point average.

Crane (1962) and Levitt et al. (1971) both showed that the Kuder

Preference Record and the EPPS as statistically significant in predict-

ing students who would complete an occupational therapy program and a

nursing program, respectively.

Graduate school success as measured by the graduate level GPA was

studied by Bean and Covert (1973), Feldman (1976), Federici and Schuerger

(1974), R. Stein and Green (1970) and Willingham (1974). The best vari-

able in all of these studies was the Graduate Record Examination scores.

Clapp (1975), Gough and Hall (1975), Richards et al. (1974), and Roemer

(1965) agreed that the Medical College Admission Test (MCAT) was a signif-

icant predictor of medical school success (GPA). They also found that

premedical GPA and science GPA were significant. Predental GPA was the

most significant predictor of junior year dental GPA (Chen et al. 1967).

Halfter's (1966) study showed that the Law School Admissions Test (LSAT)

score predicted first year law school GPA somewhat better than the

undergraduate GPA did.











Neely (1977) showed high school GPA and the ACT scores were most

consistently associated with post graduate work. However, Stein et al.

(1974) showed that overall performance of first year social work

students evaluated by faculty at the end of the year could be predicted

from students' attitudinal ratings done at the beginning of the year.

In summary, most of these studies found that academic achievement

at the college or graduate level could best be predicted by previous

performance either on achievement tests or with GPAs. The results,

though often statistically significant, were not sufficient for

practical application toward the selection of students into various

curricula, as supported by Hall and Coates' (1972) review of studies

completed between 1950 and 1968. His review stated that no reliable

variables have emerged to be of general value in the selection of nurses.


Practical Application


Lind (1970) predicted clinical practice grades of occupational

therapy students with college GPA. Tidd and Conine (1974) studied the

GPA of physical therapy students and found that academically better

students also tended to perform better in the clinic than did their

peers. Englehart (1957) showed that on-the-job performance during

clinical training of occupational therapy students may be predicted

from performance in some required college courses but other courses do

not predict clinical performance. Dunteman (1966) was not able to

predict clinical ratings from freshman and sophomore course performance.

Bailey et al. (1969) found that the Interest Test was the most effective

predictor of clinical performance of occupational therapy students.








20

Blume (1976) correlated success in a medical technology internship with

responses on a Biographical Data inventory. Booth (1957) showed

significant correlations between pediatric internship of occupational

therapy students and the T score of Guilford's STDCR Inventory and the

Literary Interest scale of the Kuder Preference Record, although the

latter was a negative correlation. The C score of STDCR correlated with

grades from a tuberculosis internship and the total internship grade was

linked with the persuasive interest scale of the Kuder Preference Record.

Osipow and Walsh (1973) conducted two studies that indicated tests

of social intelligence were useful in predicting a desired facet of

performance of counselor trainees in counseling psychology. McClain

(1968) found that the 16 PFI test was more useful for both men and women

in identifying the superior counselors than in differentiating the

average from the poor. Cough (1978) stated that high premedical science

competance and performance was unrelated to clinical performance and

faculty ratings of clinical competence. Munro (1964) decided that the

Minnesota Teacher Attitude Inventory was an inadequate predictor of

teaching performance in internship.


Tests of Professional Knowledge


Ayers and Rohr (1974) and Walberg (1967) attempted to predict scores

on the National Teacher Examination (NTE). Ayers found that grades in

social science were the best indicators of success on the NTE, while

Walberg found that high school GPA and scores on the School and College

Ability Test (SCAT) were the best indicators.








21

High school performance and standardized achievement tests were the

best predictors of scores on nursing State Board Tests (Haglund 1975,

Tatham 1976, and Weber et al. 1973). Larken (1976) found that scores on

the comparative Guidance and Placement Tests (CGP) taken at entry to a

nursing curriculum appeared to discriminate between State Board Test

passes and failures. Brandt et al. (1966) showed a lack of relationship

between grades in nursing theory courses and scores on State Board Tests.

A low correlation between the occupational therapy National

Registry Examination score and Guilford's STDCR Inventory--R Factor was

revealed by Booth (1957). Engelhart (1957) found that some grades in

specific courses predict occupational therapy Registry Examination

scores.

Fredricks et al. (1970) investigated the relationship between medical

students' performance on the National Board Examinations and their

performance on the MCAT and their academic achievement over the four

years of medical school. The most effective predictor for performance

on the National Board Examination was the GPA from the second year of

medical school.

Bergman et al. (1974) conducted a comprehensive study in Israel on

589 nursing students. They found psychological tests correlated signif-

icantly with achievement in theoretical learning. Personality factors

were studied by Krall (1970) in relation to nursing success and failure.

The Interpretation of a Draw-A-Person Test and scores from the National

League of Nursing (NLN) test were used to compare successful and failed

students. The NLN test was found to be significantly related to

nursing school withdrawal in the first year for those who did not leave








22

for "personal reasons." Unsuccessful students produced more imcomplete

figures and drew the opposite sex first significantly more often than

did successful students.


Professional Performance and Persistence


Brandt (1967), Burton (1972), and Dubs (1975) studied the records

of nurses within one year of graduation. Brandt found that 79% of them

were working but no predictor variables were statistically significant.

Burton revealed that the 16 PF indicator predicted the performance of

new nurses slightly better than chance. Dubs used curriculum objectives

to develop an employer rating scale and then compared the results with

grades to see if the curriculum was doing what it was designed to do.

May and Chitty (1971) used the progressive matrices and Mill Hill

Vocabulary Test in their study of new state enrolled nurses in their

first jobs. They concluded that the tests could be used only as a means

of obtaining more intelligent student nurses, not for competent state

enrolled nurses. Di Marco and Hilliard (1978) decided there was no con-

elusive evidence that associate, diploma, and baccalaureate degree nurses

were either the same or different in terms of the professional performance

variables measured.

Several studies have found negative results. A one year follow-up

study of clerical workers by Schmidt (1975) showed that the low ability

group was more satisfied rather than the middle or high groups.

Cornett (1969) studied the effectiveness of CPA and grades in two

classes in predicting performance of first year teachers on the job.

The best single predictor for secondary teaching performance was college








23

GPA; however, this correlation was very low. McGreevy (1967) found the

Miller Analogies Test did an "acceptable" job of identifying academic-

ally talented people.

In physical therapy, Walker (1959) found no factor to be a signif-

icant predictor of graduation and entrance into the profession from

grades and achievement measures. Anderson and Jantzen (1965) agreed

that grades and achievement measures were ineffective predictors of

clinical performance in another health field, occupational therapy.

Muthard et al. (1976) did a short term follow-up on occupational therapy

graduates. They failed to show that the registry examination and

internship scores had predictive validity for specific indices of first

year job success.

Long-term follow-up in the health fields has been limited. The

studies that have been done were more descriptive than predictive.

Hendrickson (1962) and Brown and Stones (1972) described occupational

therapy personality characteristics and job satisfaction. Brollier (1970)

expanded her description to include not only occupational therapists,

but physical therapists and social workers in a small sample of each.

The differences appeared to be related to the different job concerns of

the social psychiatric group versus the physical disabilities group.

Sussman and Haug (1968) and Jantzen (1972) described the typical

person in their respective fields. The implication is that such data

could possibly be used for recruitment and selection of future students

who would be most likely to succeed in a career of these health related

fields.

Drugo (1975), Maynard et al. (1972), Nelson (1975), Thom and Hickcox

(1975) and Tucker and Strong (1962) carried out predictive studies. Drugo











found no significant correlations between GPA, internship scores, and

employer ratings of physical therapists' one-year job success. A seven

year study by Maynard et al. (1972) revealed a small, but significant,

relationship between job performance and practical experience grades.

Nelson (1975) concluded from an extensive review of the literature of

education, health, psychology, social science, and business that GPAs

from college are of little use in predicting later occupational success.

Tucker (1962) did a ten-year follow-up of medical students and found

that interest scales did not measure distinctions between the medical

specialty areas. Ginsberg concluded that college performance had no

differential significance for later performance; however, outstanding

grades in graduate school did indicate later superior performance

(Nelson 1975).

Thom and Hickcox (1975) did a different approach to prediction of

career success. He had faculty, students and principals predict on the

basis of file information for six students whether the students would do

well in academics and careers. The actual achievements of the students

were known to the researchers. He found that there was no difference in

selection accuracy. They were approximately 60 percent correct for

academic success but only 50 percent for job success. Transcripts and

letters of recommendations were the most useful for selection of

achievers.


Concluding Remarks


Most of the studies reviewed suggested that previous intellectual

functions such as academic achievement and standardized test scores for











achievement were the best predictors of later academic achievement and

of higher scores on written tests of professional knowledge. Many

studies required the administration of additional tests and question-

naires. Occasionally scores on psychological and interest measures were

significant. None of the predictive variables were considered by the

respective researcher to be adequate as a single variable when selecting

students for a specific curriculum. Only one study used both academic

and practical application data in both independent and dependent

variables; however, that study did not follow up the same population for

entry and persistence in the specific field. Relatively few studies

dealt with long-term follow-up.

A study of the literature revealed no study combining all of the

following:

1. practical application data in the dependent and independent

variables.

2. academic data in the dependent and independent variables.

3. a sample size larger than 50.

4. data on subjects for more than three successive years.

5. followup on the same subjects more than one year after

graduation.

6. followup on the subjects who withdrew from the undergraduate

curriculum.

7. use of only objective data from transcripts.
















CHAPTER III


DESIGN AND METHODOLOGY


The purpose of this study was to explore independent variables to

determine which, if any, could be used to predict success in an occupa-

tional therapy curriculum, internship, registry examination, and job

persistence prior to the time professional course work is begun for use

in student selection. Ultimately the development of a process model

which could be used with future data for selection of candidates was a

desirable outcome.

This chapter contains an explanation of the procedures used in

this study. Initially, the population and selection of the sample are

described, followed by a description of the data collection procedures.

Subsequently, the method of analysis (including definitions of statis-

tical terms) and the pitfalls of interpreting regression analysis are

discussed. The concluding section contains a description of how the

data are organized.


Population


The undergraduate occupational therapy curriculum at the University

of Florida is within the College of Health Related Professions and

served approximately 75 juniors, seniors and interns each year.

The five-year period from 1969 to 1973 was chosen for study because

during these years the course of study and the faculty remained quite










stable. After 1973, there were major changes in both the faculty and

the curriculum. During this period of time, 9 men and 161 women

students were enrolled in the occupational therapy curriculum.

Of the 169 students, 20 dropped out of the program before finishing

the undergraduate degree. One student failed on internship and never

finished. Three students completed the undergraduate program and both

interships but did not make a satisfactory score on the registry

examination.

These students were selected for admission to the occupational

therapy curriculum based on the following criteria:

Student Requirements:

1. The student must have been an undergraduate student at an

accredited college or university.

2. Transfer students at the junior level must have been residents

of the State of Florida.

3. Nonresidents must have been enrolled at the University of

Florida since the freshman year.

4. Since the program begins only in the fall of each year, all

applicants must have completed a minimum of 96 transferable

quarter credits and all departmental prerequisites prior to

enrollment.

5. The student must have been admitted to the University of

Florida.

6. The student must have completed all departmental forms by

March 15 of the spring before admission for the following

fall. These included










a. A student information form for the College of Health

Related Professions. This was a traditional form similar

to a curriculum vitae, including

1) extracurricular activities

2) honors and distinctions

3) employment

4) essay as to why the student was interested in a health

career.

b. Copies of all transcripts must have been received in

Occupational Therapy Department via the Registrar's Office.

c. Student must have completed the form listing student

organizations, activities, and publications.

Faculty evaluation commenced after the above steps were completed by the

applicant. The selection process considered the following:

1. Grades were important; the faculty scanned credentials of all

candidates possessing a 2.0 CPA or better.

2. Pre-professional grades must be 2.0 or better.

3. The essay was read and the following factors were evaluated:

a. The student's motivation.

b. Why the student was interested in occupational therapy.

c. flow well the student wrote.

4. Work and volunteer experience were considered.

5. Two interviews were conducted;

a. One was as an advisement discussion.

b. The second was more formal and was randomly assigned

to a different member of the faculty.










6. Each faculty member wrote a summary of impressions of the

student.

7. Meeting as a selections committee, the entire faculty

discussed each student, finally ranking all in order of

preference for admission.


Data Collection


Data collection began with the files in the occupational therapy

curriculum office at the University of Florida's J. Hillis Miller Health

Center. The social security number was obtained for each student

involved in the occupational therapy curriculum that did or would have

graduated between 1967 and 1973. These files yielded the scores for

three of the dependent variables on each student: National Registry

Examination, Psychiatric Internship, and Physical Disabilities Intern-

ship. Since some of the Registry Examination scores were never recorded

in these records, it was then necessary to consult the computer printout

from the American Occupational Therapy Association (AOTA). Many of the

scores for the internships had not been coded and computed. The faculty

member who had been in charge of this task was consulted. Most of the

internship evaluations were found, coded, computed, and recorded. Some

scores were obtained by telephone and were not recorded; therefore, they

were not available even from the internship institution. This affected

five students for the Psychiatric Internship and two students for the

Physical Disabilities Internship (see Appendix). One student who is

currently registered could not be reached to give permission to AOTA to

release her Registry Examination score. An average was computed for all











those who had scores and inserted for the missing scores rather than

using a zero to avoid distorting the statistical results (Winer 1962).

The means used were

Psychiatric Internship .... ...... ..255

Physical Disabilities Internship . 247

National Registry Examination . . 172

Permission was obtained from the Registrar's office of the

University of Florida to receive copies of transcripts for all students

involved in the occupational therapy curriculum for the period from

1967 to 1973. The social security number of each student, including

those who dropped out of the curriculum, was used to request these

documents. A staff member photocopied each transcript and the data

were forwarded to this researcher.

Collection of the data from the individual transcripts was done by

the researcher. Fifteen transcripts were selected at random and were

reworked several times. Various general categories were set up to

accommodate the variations in student course selections. A standard-

ization of categories was finally achieved. For example,

Social sciences: history, anthropology, western civilization

Biological sciences: zoology, biology

Physical sciences: chemistry, physics, geology

Art: studio

Music: appreciation, lessons, choral

Literature: English and literature

College course grades were converted to points: A = 4, B = 3, C = 2

and D = 1. Semester grades were converted to quarter hour credits. The










categories tabulated in grade points were common to all transcripts.

Several categories were recorded in number of credits taken which might

indicate a proficiency and/or interest in an area.

When this study was initiated, 76 variables from the transcripts

were considered. However, many observations were deleted because they

were represented by only one or two students. Other variables were

deleted after an initial stepwise multiple regression analysis showed

that they entered the equation very late, were not statistically

significant, and accounted for a miniscule amount of movement of the

dependent variables. Only 24 variables which were represented by at

least eight per cent of the students were used in the final analysis.

These 24 variables were then keypunched according to the Biomedical

Computer Program P-Series and analyzed by the University of Montana

computer.

Ten transcripts did not have a score for the Florida Twelfth Grade

Placement Test nor the Scholastic Aptitude Test of the College Entrance

Examination Board (CEEB). An average score for all those who had scores

was computed and used rather than a zero (Winer 1962). The means used

were:

CEEB ..... .............. .1006

Florida 12th Grade Test . . 384

A table was obtained from the University of Florida Admissions

Office to convert the total score for the Florida 12th Grade Test to

the equivalent total for the Scholastic Aptitude Test scores of the

CEEB.

The American Occupational Therapy Association was contacted con-

cerninig follow-up on all subjects who had continued their national










registration. Specific information requested from the 1973 and 1977

member data surveys was forwarded to this researcher (see Appendix).

A telephone followup was conducted for all nonrespondents to these

surveys. There was no number available for six of these subjects and

17 were successfully reached by telephone.

Telephone followup was also attempted for all 20 subjects who

dropped out of the occupational therapy curriculum. No number was

available for seven of them. Contact was achieved with 13 of them.

The follow-up data yielded the fourth dependent variable score,

work status, for the multiple regression analyses. It also provided

the dependent variables, (1) complete/not complete the occupational

therapy program and (2) working/not working in the field of occupational

therapy, for the discriminant analyses.


Method of Analysis


This study was executed in two statistical parts. The first part

encompassed the objectives of finding the best pre-professional

indicators for completion of internships and the registry examination in

occupational therapy and for working in the field of occupational therapy

four to eight years after graduation. The first part was designed to

compute a multiple regression analysis using Biomedical Computer Program

P-Series (BMPP), stepwise multiple regression analysis (MR).

The second statistical part used stepwise discriminant (DIS)

analysis of the Bio-Med package to find the best indicators of work

persistence and non-persistence in the field of occupational therapy and

completion or non-completion of the occupational therapy program.








33

The 169 students studied were enrolled in the Occupational Therapy

curriculum from 1969 to 1973. For each student these follow-up records

were available:

1. University of Florida transcript

2. Internship scores, if the subjects completed the internships

3. National Registry Examination scores for the subjects who

took this examination

4. Work-entry statement from departmental follow-up newsletter

5. Follow-up information from the AOTA of those who continued

their registry

6. Telephone followup of those subjects who did not respond to

the AOTA surveys

7. Telephone followup of those subjects who did not continue

their registry in occupational therapy

8. Telephone followup of those subjects who dropped the occupa-

tional therapy program before passing the registry examination.

From the 169 University of Florida transcripts examined, 24 inde-

pendent variables occurred prior to entering professional course work.

These independent variables identified as possible factors explaining

the movement of the dependent variables are listed below:

1. Florida resident or out-of-state resident

2. University of Florida freshman or transfer

3. Grade in points in Introduction to the Health Related

Professions course

4. Gap in years from college freshman to upper division status

5. Number of hours in speech courses











6. Number of hours in

7. Number of hours in

8. Number of hours in

9. Grade in points in

10. Number of hours in

11. Number of hours in

12. Number of hours in

13. Number of hours in

14. number of hours in

15. Number of hours in

16. Number of hours in

17. Number of hours in

18. Number of hours in

19. Number of hours in

20. Number of hours in

21. Number of hours in


business courses

nursing courses

library science courses

growth and development courses

special education courses

drama and theater courses

health courses

social sciences courses

physical sciences courses

biological sciences courses

English and literature courses

mathematics courses

psychology courses

foreign language courses

arts and crafts courses

music


22. Grade point average for the first two years

23. Florida 12th Grade Placement Test total score

24. College Entrance Examination Board total test score

The dependent variables established for this study were:

1. American Occupational Therapy Association National Registry

Examination score

2. Psychiatric Internship score

3. Physical Disabilities Internship score

4. Work status at the time of this study

5. Completion of non-completion of the occupational therapy

program








6. Working or not working in the field of occupational therapy

at the time of this study.

A scale employing the numbers zero through ten was developed to

describe different levels of work status (see Table 1). The least

desirable level, indicated by a zero, represented no available followup

on those who did not complete the undergraduate curriculum in occupa-

tional therapy. The number ten represented the most favorable follow-up

information of working full time in the field of occupational therapy.

Table 1

Work Status Scale


Level
Number Description

0 No followup and did not complete the undergraduate
curriculum in occupational therapy

1 Did not complete the undergraduate curriculum in
occupational therapy and not working

2 Did not complete the undergraduate curriculum in
occupational therapy and is working in an unrelated field

3 Did not complete the undergraduate curriculum in
occupational therapy and is working in a health related
profession

4 No followup data; did complete the undergraduate curriculum
in occupational therapy, and recorded as not registered.

5 Did complete the undergraduate curriculum in occupational
therapy, not registered, and working in an unrelated field.

6 No followup data and recorded as registered

7 Registered and working in unrelated field

8 Registered, not presently working but plans to be back
to occupational therapy

9 Registered and working parttime in occupational therapy

10 Registered and working fulltime in occupational therapy











A stepwise multiple regression analysis was calculated with the BMDP

computer package using the 30 independent and dependent variables. The

dependent variable was identified as y, the independent variable as X.

The process of narrowing the predictive independent variables to just a

2
few is called variable select. This yields R which is a coefficient

of determination; that is, it yields the percentage of the movement of

Y that is being explained by the X(s), or, the per centage of variance

in Y due to the regression on X(s).

Discriminant analysis is an extension of multiple regression

analysis which allows use of a dependent variable in the form of a

positive or negative occurance. Stepwise discriminant analysis uses a

similar process of identifying the best predictor at each step for the

dependent variable. It is used to determine correct classification

within specific groups, rather than identifying variables which relate

to better performance of the dependent variable.

The independent variables were tested for significance at the .05

level of confidence. The hypotheses were then accepted or rejected.

A variable was deleted when its F ratio became too low. The "F to enter"

in the BMDP package program was 4.0 (given the size of the data set in

this problem, and "F to enter" corresponds to a 5 per cent level of

significance).


Definitions of Statistical Terms


1. Correlation analysis--the degree to which two variables are related

(Lapin 1973).










2. Correlation coefficient--shows how closely two variables move to-

gether by means of a single number calculated from the same data.

3. Coefficient of determination--expresses the relative reduction in

the variation of Y that can be attributed to knowledge of X and

its relationship to Y by way of the regression line or the pro-

portion of the total variation in Y explained by the regression

line. Expresses how strongly X is associated with Y. Represented
2 R2
by r or R (Lapin 1973).

4. Dependent variable--the variable whose value is being predicted,

represented by the y axis (Lapin 1973).

5. Independent variable--the variable whose value is known, and

represented by the X axis (Lapin 1973).

6. Estimated regression line--expresses the average relationship

between the X and Y variables. This provides an estimate of the

mean level of the dependent variable Y when the value of X is

specified. Because the estimated regression line provides

estimates only, the symbol is Y (Lapin 1973).
x

7. Simple linear regression analysis--is used to obtain predictions

of one variable using known values of another variable. The

regression equation looks like this:

Y = a + bX

Sample observations are taken and used to predict another

variable. To avoid a large sampling error, a sample size

considerably greater than ten should be used (Lapin 1973).


























a

a


2


S40 on To R Oc



Simple Linear Regression (p. 447)


Fr



Figure 1.









39

8. Multiple regression analysis--is an expansion of the same techniques

used in simple regression. Multiple regression uses several

independent variables as predictors. The movement or variation

of Y can then be explained by two or more variables. This

formula is:

Y = a + b, X, = b2 X2+. + bk Xk


This multi-dimensional analysis can be shown by the following

figure:


Figure 2.


Multiple Regression (p. 500)


An analogy would be, using the walls and floor of a room. The

corner represents the point where all three variables have a

value of zero. The observation points are represented by marbles

suspended in space at various distances from the floor and the

two walls. The value of Y is represented by the marble's height

above the floor. The observed value for X1 is represented by








40

its distance from the wall on the right, while the distance from

the left wall represents the value for X2 (Lapin 1973).

There are no theoretical limits to the number of variables that

may be used in multiple regression. Until the recent accessibil-

ity to high-speed computers, the higher-dimensional multiple

regression was severely limited in scope because of the vast

quantity of calculations involved (Lapin 1973). This study would

have been practically impossible before the advent of the

computer.

9. Discriminant function of two groups--is an expansion of multiple

regression. It is illustrated in the following figure:


A /




/0
/0
/ 0


Figure 3. I)iscriminant Function (p. 370)








41

With two groups, A and B, and two measurements X and Y, on each

member of the two groups, a scatter plot could be done for each

group. Then a circle is drawn around the groups, respectively.

A straight line is then drawn through the two points where the

ellipses intersect and then projected to a new axis, Z. The

overlap between the distributions A' and B' is smaller than would

be obtained by any other line drawn through the ellipses. The Z

axis is the discriminant function in this case. Group averages

and mean-square deviations of group values are computed. Then

that result is compared to the mean square deviations of the

group averages from the overall average. If the latter is large,

then the linear function used in these calculations separates the

groups well (Green and Tull 1970).

10. Standard error of the estimate--the square root of the mean squared

deviations about the regression line or plane. It may be used to

estimate the true variability in Y about the model. It directly

expresses the degree of scatter in the data (Lapin 1973).


Pitfalls of Regression and Correlation Techniques


When reading a correlation coefficient the reader must remember

that it reflects, in numerical value only, the strength of association

between variables. This is a statistical relationship; a large positive
orneaiv alefo 2 2
or negative value for R or r does not indicate a cause and effect

relationship. Examples of nonsense or spurious correlations abound.

For most practical problems there is usually some reasonable connection.

Often a relationship between two variables may be explained by their











interactions with a common factor. Care must be taken in using past

data to determine a future relationship. The conditions under which

the Y and X data were obtained may have changed over time. Caution

should be used when making predictions of Y whenever X falls outside

the range of values used in the original collections of data (Lapin

1973).


Organization of the Data


For the purposes of this study the students' data were organized

into their respective five classes and arranged alphabetically within

each class. The earliest metriculated class was listed first. Each

subject was then numbered consecutively, starting with A in the earliest

class, 1967. Those who dropped out of the undergraduate curriculum were

alphabetized and consecutively numbered at the end of the before

mentioned groups.


Summary


This chapter has presented the procedures used in this study. The

population and selection of the sample were described followed by data

collection procedures and the method of analysis. Definitions of

statistical terms used in this study were given. The pitfalls of

interpreting regression analysis were briefly discussed and, finally,

the organization of the data was presented.
















CHAPTER IV


ANALYSIS OF DATA AND RESULTS


Introduction


This chapter was organized around the analysis of data in order to

develop a process model. After a discussion of the specific computer

programs used in this study the results of the data analysis were

presented in six major sections based on six dependent variables. Each

dependent variable was shown with its respective prediction equation,

significance level of the independent variables, and examples of student

performance predictions with comparisons of actual data realized. All

data from 169 subjects were used throughout. Stepwise multiple regres-

sion analysis was used for predictors of the following dependent

variables:

1. National Registry Examination Score

2. Psychiatry Internship Score

3. Physical Disabilities Internship Score

4. Work Status Score

Stepwise discriminant analysis was used for predictors of correctly

classifying the subjects into dichotomous groups represented by the

following dependent variables:

1. Complete or not complete the program for registry in

occupational therapy.











2. Working or not working in the field of occupational therapy.

The last section discusses the process model.


The Computer Programs


The statistical package used for computing the data in this study

was the Biomedical Computer Programs known as "BID." The recent

revision of these programs was known as "BMDP" (Dixon and Brown 1979).

The BM)P program package, allowing the use of several dependent variables

and a large number of independent variables, offered the two programs

pertinent to this study: stepwise multiple regression "P2R" and stepwise

discriminant analysis "P7M." Both of these analyses used the forward

variable select procedure. This was described by Lapin (1973) as a

sequence of multiple linear regression equations in a stepwise manner.

At each step one variable is added to the regression equation. First,

the program selects the independent variable (X1) which accounts for the

most movement in the dependent variable Y. Then the next independent

variable is (X2) added to the model. The program continues in succes-

sively higher dimensions until there is insufficient F ratio to enter the

regression function. The F ratio for significance of the regression

coefficient was set at 4.0. If the statistical criteria were not met,

the variable was not entered into the analysis.

Separate regression analysis was done to isolate from a set of 25

predictor variables a subset which would yield an optimal predictive

equation with the fewest possible terms to predict each of the four

dependent variables (1) National Registry Examination Score, (2)

Psychiatry Internship Score, (3) Physical Disabilities Internship Score,

and (4) Work Status Score.








45

Output from stepwise regression analysis (BMDP2R) includes at each

step,

1. Multiple R

2. Standard error of estimate

3. Analysis-of-variance table

4. For variables in the equation:

a. Regression coefficient

b. Standard error

c. Tolerance

d. F to remove

5. For variables not in the equation:

a. Tolerance

b. Partial correlation coefficient

c. F to enter

Tiedeman (1951) explained that discriminant analysis determines

whether discrimination between groups is possible or not and then

reduces the size of the space in which it is necessary to think about

regions of classification (Morgan 1970, p. 37). Again, the forward

variable select method was employed. At each step one variable was

selected from the variables not in the analysis and entered into the set

of discriminating variables. The variable selected at each step was the

one which provided the greatest improvement in classification between

the two groups. A variable was considered for selection only if its

partial multivariate F ratio was larger than 3.999. Separate stepwise

discriminant analysis was done for predictor variables for each of the

two dependent variables: (1) complete/not complete, and (2) work/not











work. Output from stepwise discriminant analysis program (BMDP7M)

includes at each step:

1. Wilk's Lambda

2. Approximate F statistic

3. Degrees of freedom

4. Constant

5. For variables in the equation:

a. F to remove

b. Tolerance

6. For variables not in the equation:

a. F to enter

b. Tolerance


Stepwise Multiple Regression Analysis


Each model has been described through its development at each step.

An equation was shown for each step and some examples of using the final

equation was shown through the randomly chosen samples.


Model I--National Registry Examination Score (REGIST)

The first step in the model accounted for six percent of the

movement or a multiple R2 of .057. This equation was as follows:

Y = a + bX

Y = the National Registry Examination Score (REGIST)

a = the constant as computed at each step

b = the coefficient as computed for the independent variable

at each step











X = the independent variable--in this case number 29, total

score on CEEB

Therefore, REGIST = 67.188 + .084X

When the second variable was added, X2 = number 18, number of hours

in English and literature (ENGLT), the two yielded a multiple R2 of

.075. The equation for the second step was:

Y = a + blX1 + b2X2

REGIST = 88.733 + .085X1 2.063X2

The model was cut off at the second step because there was

insufficient F ratio to continue the stepping process. Table 2 shows

the variables entered by step, percent of movement at each step, the

overall F ratio at each step. The prediction equation developed by the

step wise multiple regression analysis for the score on the National

Registry Examination, as shown in step two description, gave statisti-

cally significant results but the reader should focus on the fact that


Table 2

Model I--Variables Entered by Step


Percent of
Dependent Independent Variable Movement at Overall
Variable Variable Number Each Step F

National CEEB 29 .057 10.04
Registry
Examination ENGLT 18 .075 7.84
(#26)

the model explains only eight percent of the movement in the National

Registry Examination scores. That means that 92 percent of the movement

in scores were not accounted for.








48

Null hypothesis I was rejected since it was statistically possible

to predict scores on the occupational therapy National Registry Examina-

tion on the basis of multiple combinations of the objective variables

used in this study. The weakness of the prediction accuracy of Model I

is shown in Table 2 using a random sample of subjects' actual scores

for examples.


Table 3

Model I--Score Comparison for National Registry
Examination-Actual and Predicted


Subject Actual Predicted
Number Registry CEEB ENGLT Score

007 186 860 08 129.7
051 158 1048 16.6 110.56
11 155 1006 08 157.74
149 174 1100 09 163.67
155 0 1020 28 117.67
165 0 820 32.1 92.21



Model II--Psychiatric Internship Score (PYCHIN)

The first step in the model accounted for five percent of the

movement or a multiple R2 of .053. The first significant variable was

number 4, University of Florida freshman or transfer student (TRANS).

The equation was:

Y =a + bX

PYCHIN = 241.634 39.606X1

The second most significant variable was grade point average of the

first two years of college, number 24 (GPA-I). Adding this to the equa-

tion brought the multiple R2 to .095. The equation at this point looked










like this:

Y = a + bY b X

PYCHIN = 154,004 49.124X1 + 34.994X2

At the third step the number of hours in English and literature

number 18 (ENGLT) brought the movement up to 12 percent. The equation

follows:

Y = a + b1X1 + b2X2 = b3X3

PYCHIN + 171.749 43.864X1 = 39.128X2 s.813X3


The number of credit hours in psychology, number 20 (PSYCH) was

significant at the fourth step giving a multiple R2 of .146. The

predictive equation developed for the score in psychiatric internship

was as follows:

Y = a + blX2 = b2X2 b3X3 + b4X4

PYCHIN = 151.52 46.716XI 3.119X2 + 38.49X3 + s.706X4


After this step the F ratio was not sufficient for further stepping

(see Table 4). Null hypothesis II was rejected since it was statis-

tically possible to predict scores in psychiatry internship on the basis


Table 4

Model IT--Variables Entered by Step

Percent of
Dependent Independent Variable Movement Overall
Variable Variable Number at Each Step F

Psychiatry TRANS 4 .054 9.50
Internship GPA-l 24 .095 8.74
#27
ENGLT 18 .120 7.51
PSYCH 20 .146 7.04










of multiple combinations of objective variables used in this study.

Care should be taken in use of this model because only 15 percent of

the movement in Psychiatry Internship scores was explained by the

statistically significant variables, leaving 85 percent unaccountable.

By examining Table 5 for a random sample of this equation in use, one

can see that it does not work well in practical application.


Table 5

Model II--Score Comparison for Psychiatric
Internship-Actual and Predicted



Psychiatry
Subject Internship Predicted
Number Actual Score TRANS GPA-l ENGLT PSYCH Score


007 262 1 2.7 08 16 477.62

051 254 0 1.97 16.6 10 659.85

i1 186 1 1.44 08 14 446.137

149 236 0 2.61 09 09 514.35

155 0 0 2.09 28 04 1082.025

165 0 0 3.0 32.1 09.5 1403.419


Model III--Physical Disabilities Internship Score (PDINT)

Variable number 4, University of Florida freshman or transfer

student, was the first significant variable with a multiple R2 of .031.

The equation appeared as:

Y = a + bX

PDINT = 231.082 30.478X1


The second step included in grade point average of the first two

years of college, number 24, raising the R2 to .077. The model

terminated at this step because of insufficient F ratio (see Table 6).










Table 6

Model Ill--Variables Entered by Step


Percent of
Dependent Independent Variable Movement Overall
Variable Variable Number at Each Step F


Physical
Disabilities TRANS 4 .032 5.44
Internship GPA-l 24 .077 6.95
#28


The prediction equation for scores in physical disabilities internship

follows:

Y = a b 1lX1 + b2X2


PDINT = 138.714 40.511X1 + 36.877X2


Again the amount of movement accounted for is very small, eight

percent. Therefore, the reader should exercise care in interpreting

the use of this model. Null hypothesis III was rejected since it was

statistically possible to predict scores in physical disabilities

internship on the basis of multiple combinations of objective variables

used in this study. Table 7 shows this model demonstrated by a random

sample of subjects' data.


Model IV--Work Status Score

This model computed only one step which accounted for only three

percent of the movement in work status scores (Table 8). The variable

of significance was CEEB, number 29. The first and final step showed

the following equation:

Y = a + bX

WKSTA = 5.352 .003X











Table 7

Model III--Score Comparison for Physical Disabilities
Internship-Actual and Predicted


Subject
Number

007
051
ill
149
155
165


Actual
Phys. Dis.
Internship

256
270.5
144
174
0
0


TRANS

1
0
1
0
0
0


GPA First
Two Years

2.7
1.97
1.44
2.61
2.09
3.0


Predicted
Score

294.97
211.36
151.306
235.963
215.787
388.059


Table 8

Model IV--Variables Entered by Step


Percent of
Dependent Independent Variable Movement Overall
Variable Variable Number at Each Step F

Work Status CEEB 29 .028 4.80
#31



Null hypothesis IV was rejected since it was statistically possible

to predict scores in work status on the basis of multiple combinations

of the objective variables used in this study. The accuracy in practical

application of this prediction equation is very low as shown in Table 9

using a random sample of subjects.












Table 9

Model IV--Score Comparison for Work Status
Actual and Predicted


Subject Actual Work Predicted Score
Number Status CEEB for Work Status


007 10 860 7.932

051 10 1048 8.496

ill 10 1006 8.370

149 10 1100 8.652

155 3 1020 8.412

165 2 820 7.812


Stepwise Discrimination Analysis


Model V--Complete and Not Complete the Program for
Registry in Occupational Therapy

The analysis of the data developed a prediction for the dichotomy

of completing or not completing the program for registration in

occupational therapy. As Shalik (1978) synthesized:

The method used for this analysis is the Wilk's lambda. The
criterion was the overall multivariate F ratio for the test
of differences among group centroids. The variable which
maximized the F ratio also minimized Wilk's lambda, a measure
of group discrimination. Group assignment was based on the
distance from the centroid with the actual group assignment
determined by the closest distance to one of the centroids.
(p. 48)

Table 10 shows the classification functions.

This discriminant function correctly classified 66.7 percent of the

subjects into the completion group and 33.3 percent were incorrectly

classified. Correct classification into the non-completion group was









54

68.7 percent with an error of 31.3 percent. A total average for correct

placement of the subjects was 68 percent. This is a very low percentage

since the correct prediction by chance would be 93.5 percent, as

illustrated in Table 11, the classification matrix. The "jacknifed

classification" is a procedure of the cross validation type (see Table

12). The "jacknifed classification" table indicates again that the

chance rate of correct classification was 92 percent while the actual

correct classification rate was only 66 percent.


Table 10

Model V and VI--Classification Functions


Variable Group = Complete Did Not Complete

Number 4 TRANS -0.946 -2.443
18 ENGLT 0.322 0.202
20 PSYCH 0.207 0.334
24 GPA-l 8.983 10.218



Table 11

Model V and VI--Classification Matrix



Percent Number of Cases Classified into Group
Group Correct Complete Did Not Complete

Complete 66.7 14 7
Not Complete 68.2 47 101
Total 68.0 61 108










Table 12

Models V and VI--Jacknifed Classification


Percent Number of Cases Classified into Group
Group Correct Complete Did Not Complete


Complete 57.1 12 9

Not Complete 67.6 48 100

Total 66.3 60 109



The first significant variable entered was number 4, University of

Florida freshman or transfer student (TRANS); second was the variable

number 20, number of hours in psychology (PSYCH); third, number 24, GPA

for the first two years of college (GPA-l); and fourth, number 18,

number of hours in English and literature (ENGLT). These four variables

are the same ones that entered into the multiple regression analysis for

predicting scores in psychiatry internship.

Null hypotheses V and VI were both rejected since it was statis-

tically possible to discriminate among groups of students, those who

completed (V), and those who did not complete (VI) the occupational

therapy program on the basis of linear combinations of objective

variables used in this study.

The same basic equation applied to discriminant analysis as to

multiple regression analysis:

Y = a + b1X1 + b2X2 + b3X3 + b4X4

V Complete = 14.387 .946X (TRANS) + .207X2 (PSYCH) + 8.983X3


(GPA-I) + .322X4 (ENGLT)











VI No Complete = -16.47 2.443X1 (TRANS) + .334X2 (PSYCH)

+ 10.218X3 (GPA-I) + .202X4 (ENGLT)

The reader can refer to Table 13 for a summary of variables entered

by step for completion of the occupational therapy program. Table 14

shows the actual and predicted scores using a random sample of data.


Table 13

Models V and VI--Summary-Variables Entered by Step


Variable Degrees of
Step Number Entered F ratio Overall F Freedom

1 TRANS 5.806 5.806 1 167
2 PSYCH 5.729 5.850 2 166
3 GPA-I 4.311 5.414 3 165
4 ENGLT 5.078 5.431 4 164




Table 14

Model V--Score Comparison for Completed-Actual and Predicted


Subject Number Actual TRANS PSYCH GPA-1 ENGLT Predicted

007 1 1 16 2.7 08 28.637
051 1 0 10 1.97 16.6 25.112
111 1 1 14 1.44 08 17.464
149 1 0 09 2.61 09 28.207
155 0 0 04 2.09 28 28.618
165 0 0 9.5 3.0 32.1 39.252


Model VI--No completion
other half of the dichotomy.

usefulness when applied to a


of occupational therapy program is the
Table 15 shows that this equation lacks


practical situation.











Table 15

Model VI--Score Comparison for Did Not Complete-
Actual and Predicted


Subject Number Actual TRANS PSYCH GPA-l ENGLT Predicted

007 1 1 16 2.7 08 15.636
051 1 0 10 1.97 16.6 10.352
11 1 1 14 1.44 08 2.093
149 1 0 09 2.61 09 15.023
155 0 0 04 2.09 28 11.878
165 0 0 9.5 3.0 32.1 23.841



Models VII and VIII--Working or Not Working in the

Field of Occupational Therapy

A prediction was developed for the dichotomy of working or not

working in the field of occupational therapy four to eight years after

graduation. Table 16 shows the classification functions. This

discriminant function correctly classified 42.9 percent of the subjects

into the working group and 57.1 percent were incorrectly classified.

Correct classification into the non-working group was 71.7 percent with

an error of 28.3 percent. The total average for correct classification

of the subjects was 62.1 percent. This is again low since the correct

prediction by chance would be 71.7 percent, as shown in Table 17, the

classification matrix. The "jacknifed procedure" indicates a concurrence

with these findings (see Table 18).













Models VII and


Table 16

VITI--Classification Functions


Degrees of
Variable Group = Work No Work Overall F Freedom


#23 Music .268 0.121 9.021 1 167

Constant -1.093 -0.775




Table 17

Models VII and VIII--Classification Matrix


Percent Number of Cases Classified into Groups
Group Correct Work No Work


Work 42.9 24 32

No Work 71.7 32 81

Total 62.1 56 113




Table 18

Models VII and VIII--Jacknifed Classification


Percent Number of Cases Classified into Group
Group Correct Work No Work


Work 42.9 24 32

No Work 71.7 32 81

Total 62.1 56 113


The first and only

of hours in music.


significant variable to enter was number 23, number









59

Null hypotheses VII and VIII were rejected since it was statistically

possible to discriminate among groups of students (VII) those who are

working and (VIII) those who are not working in the field of occupational

therapy four to eight years after graduation on the basis of linear

combinations of objective variables used in this study.

Y = a + bX

Work = 1.095 + .268 (MUSIC)

No Work = -0.775 + .121 (MUSIC)

Practical application shows the predictive weakness of this equation

also.

Model VIII not working in the field of occupational therapy was the

other half of the dichotomy. This again was not a practical equation,

especially since there was only one variable that was statistically

significant.

It was the intent of this study to develop a process model for use

in selecting students who would complete the program in occupational

therapy and remain working in the field. The process model could have

been the combintation of all the variables that proved to be significant

for the different dependent variables. Since five of the variables were

seen two or more times within the eight models, these could be considered

more important. However, the amount of movement that could be accounted

for was too small to be of practical use.
















CHAPTER V


CONCLUSIONS, OBSERVATIONS AND RECOMmENDATIONS


Conclusions


The purpose of this study was to identify objective variables which

might prove useful for selection of potential candidates for entry into

a health related curriculum who would most likely be able to complete

the academic requirements and meet the qualifications for registry in

occupational therapy, enter the job market, and continue within the

field of occupational therapy.

A study of the literature revealed that no study combined all

of the following:

1. practical application data in the dependent and independent

variables.

2. academic data in the dependent and independent variables.

3. a sample size larger than 50.

4. data on subjects for more than three successive years.

5. followup on the same subjects more than one year after

graduation.

6. followup on the subjects who withdrew from the undergraduate

curriculum.

7. use of only objective data from transcripts.

This study included all of the above on the same population.










The 169 subjects in this study all had been accepted into the

curriculum for graduating classes in occupational therapy spanning the

five year period from 1969 to 1973. Of the 160 women and nine men,

19 withdrew from the undergraduate curriculum. The 25 selected

variables occurring prior to professional course work were taken from

University of Florida transcripts. University of Florida transcripts

were available for all subjects. These variables were used as probable

factors explaining the movement of the dependent variables.

The study was executed in two statistical parts. The first part

encompassed the objectives of finding the best pre-professional course

work indicators for scores on the following:

1. Psychiatric Internship

2. Physical Disabilities Internship

3. National Registry Examination

4. Work Status Follow-up.

This part was designed to compute a stepwise multiple regression

analysis using the Bio-Med BMDP2R statistical packaged program.

The four multivariate null hypotheses proposed for the multiple

regression analysis were all statistically rejected.

It was not possible to predict scores for

(I) the National Registry Examination

(II) Psychiatry internship

(III) Physical disabilities internship

(IV) Work status









62

on the basis of multiple combinations of variables which include college

entry test scores, courses taken, grade point averages, and other

objective data taken from transcripts.

The second part used stepwise discriminant analysis to find the

best indicators of completion and non-completion of the occupational

therapy program and the best indicators for work persistence and non-

persistence in the field of occupational therapy using Biomedical

computer packaged program P-Series, BMDP7M.

The four multivariate null hypotheses proposed for the discriminant

analysis were all statistically rejected.

It was not possible to discriminate among groups of students (V)

those who completed and (VI) those who did not complete the occupational

therapy program on the basis of linear combinations of variables which

include college entry test scores, courses taken, grade point averages,

and other objective data taken from transcripts.

It was not possible to discriminate among groups of students (VII)

those who were working and (VIII) those who were not working in the

field of occupational therapy four to eight years later on the basis of

linear combinations of variables which include college entry test

scores, courses taken, grade point averages, and other objective data

taken from transcripts.











Results of the multiple regression analysis of the 25 variables

showed five to be statistically significant at the .05 level: college

entry test score, college grade point average for the first two years,

transfer student or entry as a University of Florida freshman, number

of credit hours in English and literature, number of credit hours in

psychology. The discriminant analysis used all but one of the same

variables (CEEB) and added only one more: number of credit hours in

music.

In each analysis the eight corresponding null hypotheses were

rejected. Although statistically significant variables were revealed,

they were not useful for practical application in selection because

1. they accounted for such a small amount of the movement of the

dependent variables.

2. there were not enough significant variables to make a good

model.

3. the sign of the regression coefficient did not always seem

practical. A negative sign meant that selection of students

would be based on lower scores or fewer hours in these

variables.

All of the subjects in this study were part of a pre-selected

group which obviously narrowed the range of variability and affected

the results. This was demonstrated for each dependent variable by a

random sample of subjects' actual data used in each predictive equation.

The accuracy of the discriminant analysis was far less than might have

occurred by chance.









64

An initial regression analysis on approximately 55 variables showed

that better predictors could be identified that accounted for up to 50

percent of the movement of some of the dependent variables. However,

the use of these variables was not rational in terms of when the data

were available and when they were needed to make the selection of

students. These more significant variables occurred after selection

had been accomplished and were factors in the professional curriculum.

It was the intent of this study to develop a process model for use

in selecting students who would complete the program in occupational

therapy and remain working in the field. The process model could have

been the combination of all the variables that proved to be significant

for the different dependent variables. Table 19 shows all the

independent variables used in the eight equations and the frequency of

use.


Table 19

Frequency of Independent Variable Use


Frequency of Use in
Variable Eight Equations

College Entrance Examination Board Test Score 2.... 2
Number of credit hours in English and literature 3
Transfer student or University of Florida freshman 3
Number of credit hours in psychology .... ......... 2
College grade point average for the first two years . 3
Number of credit hours in music ...... ............ 1



Since five of the variables were seen two or more times within the

eight models, these could be considered more important. However, the











amount of movement that could be accounted for was too small to be of

practical use.

The results of the study indicated that combining both academic

aiid p ractia I a I app I cat ion pa rameters in both the independent and the

dependent variables does not enhance the practical application of a

predictive selection equation versus using just one of these parameters

in both. It was found that the same five variables were statistically

significant for both the academic and the practical application

dependent variables. Implications of the study were that a few

variables could be considered in the selection of potential candidates:

college grade-point average for the first two years, transfer student

or University of Florida freshman, number of credit hours in English

and literature, College Entrance Examination Board Test score, number

of credit hours in psychology, and number of credit hours in music.

However, human factors and subjective data should also be included.

In conclusion, this researcher found that a practical predictive

equation for use in selecting potential health curriculum students

cannot be developed with the objective data used in this study.


Observations


The dropout rate at the University of Florida undergraduate level

in occupational therapy was small, only 11 percent, and persistence of

graduates working in the field of occupational therapy was 75 percent

four to eight years after graduation. Johnson (1978) said that 75

percent of occupational therapists quit working in their first five

years of employment. However, Jantzen (1972) pointed out that











80 percent of all the qualified occupational therapists in the United

States were working in the field of occupational therapy. This was

only five percent more than was shown by this study. With these facts

in mind, it seemed appropriate to say that the selection process for

potential occupational therapy students at the University of Florida

for the population of this study was doing its job well.

An employment rate of 75 to 80 percent was extremely good compared

to the 60 percent in nursing as cited by Knopf (1979). Expectations

for new graduates might be too high, considering that occupational therapy

is mostly a female field with the typical therapist being married, aged

27, and having one or no children (Jantzen 1972). The time span of five

to ten years after graduation was during the prime child bearing years.

A study from the physical therapy literature shows an employment pattern

similar to occupational therapy. Five years after graduation from a

physical therapy curriculum did not show a higher percent (32 percent)

of graduates employed than at one year after graduation (11 percent)

(Cypres 1969).

Johnson (1978) pointed out that 63 percent of those therapists

who drop out of employment within the first five years return after 20

years. A future study of work status 15 to 25 years after graduation

might reveal more accurate information on persistence in employment of

occupational therapists.

The literature showed that previous academic performance and

achievement were the best indicators of later success in similar areas.


This study concurred with these findings.











Other observations from this study include:

1. Transfer students from other colleges did not do as well as

students who began at the University of Florida as freshmen.

2. Classes in other health fields were more of a hindrance than

a help to success In tie occupational therapy curriculum at

the University of Florida.


Recommendations


Based on the findings of this study and the review of the

literature, this researcher recommends the following:

1. Further research is needed in followup of students 15 to 25

years after graduation.

2. Further research is needed using all applicants to the specific

curriculum and compare those who were accepted and those who

were rejected in the following areas:

a. Different combinations of course categories should be

tried in future research

b. Biographical data should be utilized more in future

research

c. Use of measures of maturity should be considered in

future research

3. Future attention should be given to the following variables in

the development of a student selection procedure:

a. College grade point average for the first two years

b. Transfer student or entry as a University of Florida

freshman.

c. Number of credit hours in English and literature.















BIBLIOGRAPHY


Acquaviva, FA: AOTA Human Resources Project III, The Member Data Survey.
An J OT 29(5):293-300, May-June L975.

Acquaviva, FA: The Member Data Survey IV--Demographic Characteristics
of Occupational Therapy Personnel. Am J OT 29(7):426-431, Aug.
1975.

A'Hearn, RT: A Model for Predicting Academic Achievement of Technical
College Students. Doctoral Dissert, Marquette Univ, 1974.

Aiken, LR Jr: "Three Alignment Charts for Use in Selective Admissions."
J Ed Res, 62:14-17, Sept 1968.

Aldag, FC and Martin, MF: Physical Therapist Assistant Selection and
Academic Success. Phys Ther 55(7):747-750, 1975.

AOTA: Occ Ther Newletter 29(4):l, Jan 1975.

AOTA: "Occupational Therapy: 2001 AD" Papers presented at the special
session of the Representative Assembly, Nov. 1978, AOTA, Rockville,
Maryland, 1979.

Anderson, HE and Jantzen, AC: A Prediction of Clinical Performance.
Am J OT 19(2):76-78, 1965.

Ayers, JB and Rohr, ME: Relationship of Selected Variables to Success
in a Teacher Preparation Program. Ed Psych Meas 34:933-937, 1974.

Bailey, JP: Some Comparisons of Occupational Therapy Biographical Facts
with Implications for Recruiting. Am J OT 22(4):259-263, 1968.

Bailey, JP, Jantzen, AC and Dunteman, GH: Relative Effectiveness of
Personality Achievement and Interest Measures in the Prediction of
a Performance Criterion. Am J OT 23(l):27-29, 1969.

Baker, E Jr: Associate Degree Nursing Students: Nonintellective Dif-
ferences between Dropouts and Graduates. Nurs Res 24(l):42-45,
Jan-Feb 1975.

Bean, AG: The Prediction of Performance in an Educational Psychology
Master's Degree Program. Ed & Psych Meas 35(4):957-962, 1975.










Bean, AG and Covert, RW: Prediction of College Persistence Withdrawal
and Academic Dismissal: A Discriminant Analysis. Ed & Psych Meas
33:407-411, 1973.

Belcastro, FP: Use of Selected Factors as Predictors of Success in
Completing a Secondary Teacher Preparation Program. Ed & Psych
Meas 35:957-962, 1975.

Bergman, R, Edelstein, A, Rotenberg, A and Melamed, Y: Psychological
Tests: Their Use and Validity in Selecting Candidates for Schools
of Nursing in Israel. Int J Nurs St 11:85-109, 1974.

Bernstein, L: Medicare and the Occupational Therapist. Am JOT 22(5):
390-395, 1968.

Bistreich, AM: Predicting Grade Point Average, Withdrawal and Graduation
from Four Allied Health Programs at Miami-Dade Community College
Medical Center Campus. Doctoral Dissert, Nova Univ, 1977.

Blaisdell, EA Jr and Gordon, D: Selection of Occupational Therapy
Students. Am J OT 33(4):223-229, Apr 1979.

Blume, CS: Use of a Biographical Data Inventory to Identify Behavioral
Correlates of Practicing Medical Technologists: An Exploratory
Study. J Alld 111th 5(l):54-61, W 1976.

Booth, MD: A Study of the Relationship between Certain Personality
Factors and Success in Clinical Training of Occupational Therapy
Students. Am J OT 11(2):93-96, 1957.

Bowers, J and Loeb, J: Changes in Predictor Weights in GPA Regression
Equations. Meas & Eval Guid 5(3):438-442, Oct 1972.

Brandt, EM: Comparison of On-the-Job Performance of Graduates with
School of Nursing Objectives. Nurs Res 16(l):50-60, 1967.

Brandt, EM, Hastie, B and Schuman, D: Predicting Success on State Board
Examinations. Nurs Res 15(l):62-69, 1966.

Brazziel, W: Non-intellective Predictors of Student Persistence/
Attrition for College and University Research and Planning. Paper
presented at the Tnt Forum of the Assoc for Insti Res, Montreal,
May 9, 1977.

Brollier, C: Personality Characteristics of Three Allied Health Profes-
sional Groups. Am J OT 24(7):500-505, ]970.

Brown, PU: Career Development and Satisfaction of Occupational Therapists
in Florida. Doctoral Dissert, Univ of Fla 1973.

Brown, RGS and Stones, RNII: Personality and Intelligence Characteristics
of Male Nurses. Int J Nurs St 9:167-177, 1972.










Buca, JT: An Investigation of the Attrition and Retention of Physical
Therapy Students at Univ of Maryland. Unpub Thesis, Univ of Mary-
land, Baltimore 1967.

Buhmeyer, KJ: Predicting Success in a Physician Extender (Medex) Program.
Doctoral Dissert, Walden Univ, July 1975.

Burgess, MM: Two Studies of Prediction of Success in a Collegiate
Program of Nursing. Nurs Res 21(4):357-366, July-Aug 1972.

Burton, DA: The Selection of Nurses by Discriminant Analysis. Int J
Nurs St 9:77-84, 1972.

Chase, DI and Johnson, JJ: Predicting College Success with Nontradi-
tional Data for Inner-city Students. J Col St Pers 210-214,
May 1977.

Chen, MK, Podshadley, DW and Shrock, JG: A Factorial Study of Some
Psychological, Vocational Interest and Mental Ability Variables as
Predictors of Success in Dental School. J Appl Psych 51(3):236-
241, 1967.

Clapp, TT: Comparison of Three Models for Selection of Applicants and
Prediction of Success in Medical School. Doctoral Dissert, Univ
of Missouri, Columbia, 1975.

Conine, TA: A Survey of the Graduates of a Professional Physical
Therapy Program. Phys Ther 52(8):855-861, 1972.

Cornett, JD: Effectiveness of Three Selective Admissions Criteria in
Predicting Performance of First-Year Teachers. J Ed Res 62:247-
250, Feb 1969.

Crane, WJ: Screening Devices for Occupational Therapy Majors. Am J OT
16(3):131-132, 1962.

Cramer, EM: Brief Report: The Use of Highly Correlated Predictors in
Regression Analysis. Mult Beh Res 241-243, Apr 1974.

Cypres, RN: Attrition and Related Variables in a Four Year Baccalaureate
Physical Therapy Curriculum. Unpub Thesis, Boston Univ, Sargeant
College 1969.

Davies, ME and Khosla, T: Predictive Value of Two Intelligence Tests as
Criteria of Success in a Health Visitor Examination. Nurs Times
28:109-112, 1974.

Di Marco, N and Hilliard, M: Comparisons of associate, diploma, and bac-
calaureate degree nurses' state board, quality of patient care,
competency rating supervisor rating, subordinates' satisfaction
with supervision and self-report job satisfaction scores. Int J
Nurs St 15:163-170, 1978.










Dixon, WJ, ed: Biomedical Computer Programs, 3rd ed. Berkeley:
University of California Press, 1976 pp 285-387.

Dixon, WJ and Brown, MB, Eds: BMDP Biomedical Computer Programs
P-Series University of California Press, 1979 pp. 399-436.

Drugo, JG: Regression models and noncognitive predictors. Doctoral
Dissert, Uiiiversity of Pittsburgh, 1975.

Dubs, R: Comparison of student achievement with performance ratings of
graduates and state board examination scores. Nurs Res 24(l):59-
62, 1975.

Dunning, ER: An accountability model for occupational therapy. Am J OT
29(l):35-38, 1975.

Dunteman, GH: Characteristics of students in the health related profes-
sions. Univ of Fla, Rehab Res Monograph Series, Number 2,
June 1966.

Engelhart, HV: An investigation of the relationship between college
grades and on-the-job performance during clinical training of
occupational therapy students. Am J OT 11(2):97-107, Part II
1957.

Federici, L and Schuerger, J: Prediction of success in an applied M A
psychology program. Ed & Psych Meas 34:945-952, 1974.

Feldman, EW: Characteristics associated with persistence in premedical
education at Duke Univ: A discriminant analysis. Doctoral Dissert,
Duke Univ. 1976.

Fredricks, MA, Mundy, P, Robertson, LS and Kosa, J: The national board
examination and academic achievement in a medical school. J Ed Res
63:275-278, Feb 1970.

Gough, HG: Some predictive implications of premedical scientific
competence and preferences. J Med Ed 53:291-300, Apr 1978.

Gough, HG and Hall, WB: The prediction of academic and clinical perform-
ance in medical school. Res Higher Ed 3:301-314, 1975.

Green, LH: Predicting academic survival--attrition in first semester of
Marquette University physical therapy curriculum. Unpub Thesis,
Milwaukee, Wisc. Jan 1962.

Green, PE and Tull, DS: Research for Marketing Decisions, 2nd ed.
Prentice-Hall, Inc. New Jersey, 1970.

Haglund, AH: Predicting success in collegiate nursing programs.
Doctoral Dissert, Univ of Wisc, Madison, 1975.










Halfter, IT: Validity studies: Law school. J Ed Res 59:307-309, Mar
1966.

Hall, JN and Andrews, WT: Apparent homogeneity in characteristics of
student nurse groups. Int J Nurs St 9:103-109, 1972.

Hall, R and Coates, CJ: Prediction of academic success for special
program and regularly admitted college freshmen. Col & Univ,
14-18, Fall 1973.

Heit, P, Johnson, G, Meeks, LB and Paxon, C: A selective admissions
process as a predictor of academic success in health education.
J Sch Hlth 48(3):146-500, 1978.

Hendrickson, D: Personality variables: Significant departures of
Occupational therapy from population norms. Am J OT 16(3):127-
130, 1962.

Hinkel, D and Houston, CA: A comparative study of the effectiveness of
two Bayesian models for predicting the academic successes of
selected allied health students enrolled in the comprehensive
community college. Paper presented at the Annual Mtg of Am Ed Res
Assoc, NY, NY Apr 4-8, 1977.

Holmstrom, EI: Promising prospects: Students choosing therapy as a
career. Am J OT 29(10):608-614, 1975.

Holt, LI: A study to identify predictors of success in academic programs
and in occupations for students in medical laboratory technician
programs in community colleges, Unpub Doctoral Dissert, Univ of
Connec, 1978.

Humphreys, LG, Levy, J and Taber, T: Predictability of academic grades
for student of high and low academic promise. Ed & Psych Meas 33:
385-392, 1973.

Inkenberry, SO: Factors in college persistence. J Coun Psych 8(4):322-
329, 1961.

Jantzen, AC: Some characteristics of female occupational therapists.
Am J OT 26(l):19-26, 1972.

Johnson, JA: Nationally speaking. Am J OT 28(9):521-526, 1974.

: Nationally speaking: Issues in education. Am J OT
32(6):355-358, 1978.

Johnson, RW, Arbes, BH and Thompson, CG: Selection of occupational
therapy students. Am J 0OT 28(10):597-601, 1974.

Johnson, RW and Leonard, LC: Psychological test characteristics and
performance of nursing students. Nurs Res 19(2):147-150, 1970.










Katzell, ME: Expectations and dropouts in schools of nursing. J Appl
Psych 52(2):154-157, 1968.

Kim, JO and Kohout, FJ: Multiple regression analysis: Subprogram
regression. Univ of Iowa in SPSS--Statistical Package for the
Social Sciences, Nie, Norman H, et al McGraw Hill--N 2nd ed,
1975, pp 320-367.


Knauer, FE: A study of the relationship of selected variables to
persistence of academically capable former students of the Univ
of SD, Unpub Doctoral Dissert, U of SD, 1968.

Knopf, L: From student to RN: A report of the nurse career--pattern
study. Pub Hlth Ser (DHEW) Wash DC, Div Nurs, 1972.

: Nurse Career--Pattern study: Baccalaureate degree nurses ten
years after graduation. Hosp Tpcs 57(2):5-6, 1979.

Krall, V: Personality factors in nursing school success and failure.
Nurs Res 19(3):265-268, 1970.

Lapin, Lawrence L: Statistics for Modern Business Decisions, Harcourt
Brace Javanovich, Inc. NY, 1973.

Larken, PG: Admission test results as criteria for entrance into the
nursing program and regression analysis of entrance tests as pre-
dictors of success in nursing state board examinations. Prince
George's Comm. Col, Largo, MD Office of Institutional Res, 1976.

Lee, JE and Shepard, KF: A comparison of biographical data of physical
therapy students in California. Phys Ther 52(3):272-281, 1973.

Lewis, J and Welch, M: Predicting achievement in an upper-division
bachelor's degree nursing major. Ed & Psych Meas 35:467-469, 1975.

Levitt, EE, Lubin, B, DeWitt, KN: An attempt to develop an objective
test battery for the selection of nursing students. Nurs Res 20(3):
255-261, 1971.

Liao, WC and Adams, JP: Methodology for the prediction of pharmacy
student academic success I: Preliminary aspects. J Col Stud Per
18(3):210-214, 1977.

Lind, Al: An exploratory study of predictive factors for success in the
clinical affiliation experience. Am J OT 24(3):222-226, 1970.

Lowenthal, W, Wergin, J and Smith, HL: Predictors of success in pharmacy
school: PCAT vs other admission criteria. Am J Pharm Ed 41(3):
267-269, 1977.










Lucci, JA: An approach to the selection of students. Am J OT 28(2):
91-93, 1974.

May, AE and Chitty, E: Selecting candidates for pupil nurse training.
Nurs Mirror pp 12-13, Dec 17, 1971.

Maynard, M, Bilkey, A, Hyre, K: Predicting Job performance. Am J OT
26(5):265-268, 1972.

McClain, EW: Sixteen Personality Factor questionnaire scores and success
in counseling. J Couns Psych 14(6):492-496, 1968.

McGreevy, CP: Factor analysis of measures used in the selection and
evaluation of counselor education candidates. J Couns Psych 14(1):
51-56, 1967.

Morgan, MK: The OPI, the ACT and university attrition: A discriminant
analysis. Doctoral Dissert, U of Kentucky, 1970.

Morgan, MK: Selecting candidates for over-subscribed programs.
Community College Review 5:65-73, Fall 1977.

Muchinsky, PM and Hoyt, DP: Predicting college grades of engineering
graduates from selected personality and aptitude variables. Ed &
Psych Meas 33:935-937, 1973.

Munro, BC: The Minnesota Teacher Attitude Inventory as a predictor of
teaching success. J Ed Res 57:138-139, Nov 1964.

Muthard, JE, Morris, JD, Crocker, LM and Slaymaker, JE: Field work
experience ratings and certification examination scores as
predictors of job performance and satisfaction in occupational
therapy. Am J OT 30(3):151-156, Mar 1976.

Neeley, R: Discriminant analysis for prediction of college graduation.
Ed & Psych Meas 37:965-970, 1977.

Nelson, AM: Undergraduate academic achievement in college as an
indication of occupational success. Professional Series 75-5.
Civil Service Comm, Wash DC, Pers Meas Res & Devel Ctr, Sept 1975.

Newsletter of MOTA. 3(2):4, 1979.

Nickens, JM: A comparison of accuracy of selected models for predicting
academic performance of junior college transfer students. J Ed Res
66(3):111-114, 1972.

Osipow, SH and Walsh, WB. Social intelligence and the selection of
counselors. J Couns Psych 20(4):366-369, 1973.

Owen, SV and Feldhusen, JF: Effectiveness of three models of multi-
variate prediction of academic success in nursing education.
Nurs Res 19(6):517-525, 1970.










Palmieri AA III: A model for multivariate prediction of academic
success of "transfer students" in pharmacy school. Am J Pharm Ed
41(3):264-600, 1976.

Patterson, TW and Davidshofer, CO: A counseling and training program
for occupational therapy students. Am J OT 26:419-423, 1972.

Pedrini, BC and Pedrini, DT: Predictors of college success. ERIC
microfische ED086102, 1974.

Pennell, MY and Hoover, DB: Health manpower source book: Allied health
manpower supply and requirements 1950-80, Wash DC: Govt Printing
Ofc. No 263, Sec 21, 1970.

Pfeifer, CM Jr and Sedlacek, WE: Predicting black student grades with
nonintellectual measures. J Neg Ed 43(l):67-76, 1974.

Poole, MA and Kassalow, S: Manpower survey report: Wisconsin Occupa-
tional therapy association. Am J OT 22(4):304-306, 1968.

Prediger, DJ: Prediction of persistence in college. J Couns Psych
12(l):62-67 1965.

Prendergast, N: Comments on issues in education. Am J OT 32(9):595-
596, 1978.

Progress Notes, Oct 1970 through Fall 1974, Univ of Fla, Dept of OT.

Reavley, W and Wilson, LJ: Selection or diagnosis? The use of
psychometric tests with candidates for student nurse training.
Nursing Times pp. 183-184, Nov 16, 1972.

Reed, CL: Prediction of grade point averages using cognitive and non-
cognitive predictor variables. Psych Rep 32:143-148, 1973.

Reed, CL, Feldhusen, JF and Van Mondfrans, AP: Prediction of grade
point averages in nursing schools using second-order multiple
regression models. J Ed Meas 9(3):181-187, 1972.

Redfering, DL and Biasco, F: Selection and elimination of candidates
in counselor education programs. Couns Ed & Supr 34:298-303, 1976.

Richards, JM Jr, Calkins, EV, McCanse, A and Burgess, MM: Predicting
performance in a combined undergraduate and medical education
program. Ed & Psych Meas 34:923-931, 1974.

Roemer, RE: Nine year validity study of predictors of medical school
success. J Ed Res 59:183-185, 1965.

Rose, HA: Prediction and prevention of freshman attrition. J Couns
Psych 12(l):399-403, 1965.










Schimpfhauser, FT and Broski, DC: Predicting academic success in allied
health curricula. J Alld Hlth 25:35-46, Winter 1976.

Schmidt, JB: Prediction of success in clerical occupations from ability
test scores. Voc Gdnc Qrtly 30:68-72, 1975.

Seitz, LA: Towards a selection and admissions model: Predicting
academicsuccess in veterinary school. Doctoral Dissert,
Ohio State Univ, 1973.

Shalik, H: A multivariate process model to evaluate undergraduate PT
students admitted by stratified sample and by ordinal rank pre-
selection. Doctoral Dissert, Univ of Fla, 1978.

Singh, A: The predictive value of cognitive tests for selection of
pupil nurses. Nurs Times 20:89-92, 1972.

Stein, RF and Green, EJ: The graduate record examination as a predictive
potential in the nursing major. Nurs Res 19(l):42-47, 1970.

Stein, S, Linn, MW and Furdon, J: Predicting social work student
performance. J Ed Soc Wrk 10(3):85-92, 1974.

Stone, RG and Acquaviva, FA: AOTA Human Resources Project II. The
Education Study. Jm J OT 29(4):233-239, 1975.

Sussman, MB and Haug, MR: Rehabilitation counselor recruits. J Couns
Psych 15(3):250-256, 1968.

Tatham, EL: Nursing and dental hygiene selection procedures. Part 2:
An examination of academic variables as predictors of success.
Johnson Co Community College, Overland Park, Kansas, Mar 1976.

Thom, DJ and Hickcox, ES: The selection of new graduate student for an
educational administration program. J Ed Adm 13(l):23-34, 1975.

Tidd, GS and Conine, TA: Do better students perform better in the
Clinic? Phys Ther 54(5):500-505, 1974.

Tiedeman, DV: The utility of the discriminant function in psychological
and guidance investigations. (The Multiple Discriminant Function--
A Symposium). Harvard Ed Rev 121:71-80, 1951.

Tucker, AC and Strong, EK Jr: Ten-year followup of vocational interest
scores of 150 medical college seniors. J Appl Psych 46(2):81-86,
1972.

U.S. Dept of Health Manpower. Health Resources Statistices: Health
Manpower Facilities, 1974. Wash DC, Govt Printing Office, Library
of Congress No. 66-62580, 1974.











U.S. Dept of Labor. Occupational Outlook Handbook 1976-77. Wash DC,
Govt Printing Office, Bulletin 1875, 1976.

Walberg, HJ: Scholastic Aptitude, the National Teacher Examinations,
and teaching success. J Ed Res 61:129-131, 1967.

Walker, SM: A study of the value of the admissions requirements in
predicting graduation from a physical therapy program and entrance
into the profession, Masters Thesis, Boston Univ, 1959.

Weber, L, King, MF and Pitts, G: Predicting success in schools of
practical nursing. Ed & Psych Meas 33:943-943, 1973.

Wittmeyer, AL, Camiscioni, JS, and Purdy, PA: A longitudinal study of
attrition and academic performance in a collegiate nursing program.
Nurs Res 20(4):339-347, 1971.

Willingham, WW: Predicting success in graduate education. Sci 183(25):
273-278, 1974.

Winer, BJ: Statistical Principles in Experimental Design. McGraw-Hill
Book Company, Inc. New York 1962.
































APPENDIX
















Key to Abbreviations


WK/INK = working or not working in the field of occupational therapy
four to eight years after graduation.

REGIST = American Occupational Therapy National Registry
Examination score.

PYCHIN = Psychiatric Internship score.

PD INT = Physical Disabilities Internship score

COMP P = Completion of no completion of the educational program in
occupational therapy

WK STA = Work status score four to eight years after expected
graduation.







80

Subject #2 #26 #27 #28 #30 #31
Number WK/NWK REGIST PYCHIN PD INT COMP P WK STA


1 0 172 223.0 228.0 1 08
2 1 200 229.5 267.0 1 10
3 1 175 226.5 288.0 1 10
4 0 195 264.0 233.0 1 08
5 0 188 224.0 216.5 1 08
6 1 163 277.5 244.5 1 10
7 1 186 262.0 256.0 1 10
8 1 175 297.5 298.0 1 10
9 1 170 265.5 253.0 1 10
10 1 213 269.5 262.0 1 10
11 1 207 275.5 281.5 1 10
12 1 190 275.5 221.5 1 10
13 1 181 243.5 239.0 1 10
14 0 201 213.0 192.5 1 09
15 1 174 244.5 221.5 1 10

16 1 187 273.5 279.5 1 10
17 1 177 244.5 296.0 1 10
18 1 201 236.5 241.0 1 10
19 1 206 270.0 229.0 1 10
20 1 205 269.0 249.5 1 10
21 1 174 275.5 286.0 1 10
22 1 207 291.5 283.5 1 09
23 1 173 280.5 251.5 1 10
24 0 188 276.0 258.0 1 08
25 1 187 272.5 199.0 1 10
26 1 194 242.0 222.0 1 09
27 1 174 228.0 234.5 1 10
28 1 159 238.0 230.5 1 10
29 0 189 275.0 209.0 1 08
30 0 194 294.5 256.5 1 08
31 1 157 225.5 275.5 1 09
32 1 177 219.0 291.0 1 10
33 0 175 253.0 244.0 1 08
34 0 179 281.5 264.0 0 08
35 1 188 259.0 267.0 1 10
36 1 187 261.5 230.0 1 10
37 0 183 249.0 206.0 1 08
38 1 171 242.0 229.0 1 09
39 0 191 273.0 277.0 1 09
40 1 164 297.0 265.5 1 10
41 1 182 281.5 266.0 1 10
42 0 160 Mean 255.0 198.0 1 07
43 1 162 245.0 141.0 1 10
44 1 191 251.0 230.0 1 10
45 1 158 232.5 295.0 1 10










#27
PYCHIN


164
200
164
180
Mean 172
158
149
172
137
193
163
183
177
165
151
172
150
165
146
159
169
144
161
181
166


274.0
206.5
288.5
148.5
Mean 255.0


Subject
Number


#2
WK/NWK


#26
REGIST


#28
PD INT


#30
COMP P


#31
WK STA


151.5
245.0
270.0
262.5
184.0
254.0
228.5
231.0
203.5
233.0
269.0
272.0
239.0
237.0
285.0
Mean 255.0
274.0
289.0
246.0
222.0
273.0
281.0
249.0
269.0
246.5
Mean 255.0 Mean
256.0
231.0
232.0
269.0
264.0
226.0
242.5
256.0
169.0
276.5
274.0
283.0
269.5
255.5


165
164
150
184
178
197
149
147
138
163
141
178
179
000
154


281.0
217.0
228.0
266.0
149.0
270.5
226.0
231.0
232.0
254.0
205.0
283.0
270.5
272.0
300.0
287.0
282.0
267.0
175.0
243.0
242.0
289.0
218.0
224.0
236.0
247.0
234.5
196.0
284.0
269.0
238.5
290.0
233.0
285.5
180.0
164.5
267.0
230.0
166.5
263.0
284.5
274.0
245.5
203.0
276.5










Subject #2 #26 #27 #28 #30 #31
Number WK/NWK REGIST PYCHIN PD INT COMP P WK STA


91 1 155 273.5 277.5 1 10
92 1 167 256.0 272.0 1 09
93 0 150 210.0 189.0 1 04
94 1 162 224.0 207.0 1 10
95 1 175 279.0 250.0 1 10
96 1 172 274.5 277.0 1 10
97 1 178 279.5 277.0 1 10
98 0 173 276.0 200.0 1 08
99 0 163 230.0 211.5 1 04
100 1 176 253.0 270.0 1 10
101 1 164 Mean 255.0 219.5 1 10
102 1 144 268.0 201.0 1 10
103 1 180 251.0 267.0 1 10
104 1 166 294.5 278.0 1 10
105 1 170 300.0 280.0 1 10
106 1 180 285.0 280.5 1 09
107 1 194 284.0 266.0 1 10
108 1 168 267.0 218.5 1 10
109 1 177 239.0 194.0 1 10
110 1 183 242.0 284.0 1 10
111 1 155 186.0 144.0 1 10
112 1 157 279.0 229.5 1 10
113 1 174 231.0 298.0 1 10
114 1 187 243.0 280.5 1 10
115 1 174 283.0 268.0 1 10
116 1 151 238.0 155.0 1 10
117 1 143 279.0 280.0 1 10
118 1 159 292.5 233.0 1 10
119 1 165 208.0 292.5 1 10
120 1 201 291.0 290.0 1 10
121 0 171 250.0 265.0 1 08
122 0 183 246.0 213.0 1 08
123 1 188 265.0 265.0 1 10
124 1 144 245.0 238.5 1 10
125 1 167 220.5 248.0 1 10
126 1 180 289.0 243.5 1 10
127 1 189 253.5 212.0 1 10
128 1 160 288.0 292.5 1 10
129 0 166 237.5 194.5 1 06
130 1 161 236.0 219.0 1 10
131 1 168 287.0 284.5 1 10
132 1 168 202.0 236.5 1 10
133 1 197 275.5 251.5 1 10
134 1 164 289.0 260.0 1 10
135 1 179 274.0 299.0 1 10











Subject #2 #26 #27 #28 #30 #31
Number WK/NWK REGIST PYCHIN PD INT COMP P WK STA


136 1 189 291.0 282.0 1 10
137 1 174 255.5 240.5 1 10
138 0 000 244.5 237.0 0 02
139 1 150 233.0 288.5 1 10
140 1 188 232.0 218.0 1 09
141 1 162 262.5 280.5 1 10
142 0 159 235.0 285.0 1 05
143 1 184 166.5 244.0 1 10
144 1 167 267.0 Mean 247.0 1 09
145 1 148 237.0 248.0 1 10
146 0 169 257.0 169.0 1 06
147 1 151 260.5 151.0 1 10
148 1 190 219.0 190.0 1 09
149 1 174 236.0 174.0 1 10
150 0 0 02
151 0 0 02
152 0 0 00
153 1 197 265.0 255.0 1 09
154 0 00
155 0 0 03
156 0 0 01
157 0 0 02
158 0 0 01
159 0 0 00
160 0 0 01
161 0 0 00
162 0 0 00
163 0 0 00
164 0 0 00
165 0 0 03
166 0 0 02
167 0 0 00
168 0 0 03
169 0 0 02
















BIOGRAPHICAL SKETCH


Carolyn Dockstader Stewart was born Feburary 4, 1943, in El Paso,

Texas. She attended numerous public schools all over the northwast and

central United States, and graduated from Clinton High School, Clinton,

Iowa, in 1961. She attended the University of Iowa and received the

Bachelor of Science degree in occupational therapy in June 1965.

Upon graduation she worked as a staff therapist and assistant

director of occupational therapy in several clinical settings from 1966

to 1972. In 1971 she entered graduate school on a parttime basis at

the University of Florida. She received the Master of Education degree

in March 1973 with a major in special education. In December of 1973

she received the Specialist in Education degree with a major in

curriculum and instruction. Since that date she has been studying part-

time toward the Doctor of Philosophy degree in curriculum and instruction

While studying for the doctorate she has worked parttime as Assis-

tant Professor of Education at the University of Montana, occupational

therapy and educational consultant to the Comprehensive Developmental

Clinic In Missoula, Montana, and special education consultant to the

Missoula County high schools.

From her marriage to Keith C. Stewart, M.D., of Cedar Falls, Iowa,

they have two sons, Chad and Dirk.

She is a member and a registered occupational therapist with the

American Occupational Therapy Association since 1966.











I certify that I have read this study and in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate,
in scope and quality, as a dissertation for the degree of Doctor of
Philosophy.





James W. Hensel, Chairman
Professor of Curriculum and Instruction







I certify that I have read this study and in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate,
in scope and quality, as a dissertation for the degree of Doctor of
Philosophy.





Lela A. Llorens
Professor of Occupational Therapy







I certify that I have read this study and in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate,
in scope and quality, as a dissertation for the degree of Doctor of
Philosophy.





Margaret Y. Morgan 61
Associate Professor of Curriculum
and Instruction










I certify that I have read this study and in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate,
in scope and quality, as a dissertation for the degree of Doctor of
Philosophy.






Pro essor of Special Education








This dissertation was submitted to the Graduate Faculty of the Division
of Curriculum and Instruction in the College of Education and to the
Graduate Council, and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy.

December 1979


Dean, Graduate School






















































UNIVERSITY OF FLORIDA

1IIIl lllll1111111 IlGIM 1111 i i 11111111 UliM I
3 1262 08555 2817




Full Text
6
2. Identify the best pre-professional indicators of completion
of the program in occupational therapy.
3. Identify the best pre-professional indicators of completion
of Physical Disabilities Internship.
4. Identify the best pre-professional indicators of completion
of Psychiatric Internship.
5. Identify the best pre-professional indicators of success on
passing the National Registry Examination.
6. Identify the best pre-professional indicators of work status
in the field of occupational therapy.
Need for the Study
The training of an occupational therapist is expensive in time,
money and equipment, encompassing a four-year to five-year educational
program. Therefore, the selection and retention of successful students
should be studied to help identify criteria for the selection of future
applicants for training. Selection of potential occupational therapists
has been primarily the domain of educational personnel because career
entry has required completion of a formal, accredited educational
program. Many areas of student ability and background have been con
sidered as elements of prognosis of success in college and for gradua
tion from college. Methods employed have included extensive use of
written applications, recommendations, and personal interviews, consuming
great amounts of faculty time. Criteria for selection have emphasized
academic achievement, participation in extracurricular activities, and
personal qualities.


APPENDIX


32
registration. Specific information requested from the 1973 and 1977
member data surveys was forwarded to this researcher (see Appendix).
A telephone followup was conducted for all nonrespondents to these
surveys. There was no number available for six of these subjects and
17 were successfully reached by telephone.
Telephone followup was also attempted for all 20 subjects who
dropped out of the occupational therapy curriculum. No number was
available for seven of them. Contact was achieved with 13 of them.
The follow-up data yielded the fourth dependent variable score,
work status, for the multiple regression analyses. It also provided
the dependent variables, (1) complete/not complete the occupational
therapy program and (2) working/not working in the field of occupational
therapy, for the discriminant analyses.
Method of Analysis
This study was executed in two statistical parts. The first part
encompassed the objectives of finding the best pre-professional
indicators for completion of internships and the registry examination in
occupational therapy and for working in the field of occupational therapy
four to eight years after graduation. The first part was designed to
compute a multiple regression analysis using Biomedical Computer Program
P-Series (BMPP), stepwise multiple regression analysis (MR).
The second statistical part used stepwise discriminant (DIS)
analysis of the Bio-Med package to find the best indicators of work
persistence and non-persistence in the field of occupational therapy and
completion or non-completion of the occupational therapy program.


66
80 percent of all the qualified occupational therapists in the United
States were working in the field of occupational therapy. This was
only five percent more than was shown by this study. With these facts
in mind, it seemed appropriate to say that the selection process for
potential occupational therapy students at the University of Florida
for the population of this study was doing its job well.
An employment rate of 75 to 80 percent was extremely good compared
to the 60 percent in nursing as cited by Knopf (1979). Expectations
for new graduates might be too high, considering that occupational therapy
is mostly a female field with the typical therapist being married, aged
27, and having one or no children (Jantzen 1972). The time span of five
to ten years after graduation was during the prime child bearing years.
A study from the physical therapy literature shows an employment pattern
similar to occupational therapy. Five years after graduation from a
physical therapy curriculum did not show a higher percent (32 percent)
of graduates employed than at one year after graduation (11 percent)
(Cypres 1969).
Johnson (1978) pointed out that 63 percent of those therapists
who drop out of employment within the first five years return after 20
years. A future study of work status 15 to 25 years after graduation
might reveal more accurate information on persistence in employment of
occupational therapists.
The literature showed that previous academic performance and
achievement were the best indicators of later success in similar areas.
This study concurred with these findings.


ACKNOWLEDGMENTS
The researcher wishes to express sincere appreciation to those
persons who had the patience and persistence in helping to make this
study possible. Since most of the preparation was done by correspon
dence, a debt of gratitude goes to members of the faculty of the
University of Florida and the University of Montana.
Dr. Alice Jantzen provided initial encouragement and inspiration
for entering graduate school. Dr. Myron Cunningham contributed his
personal warmth and appreciation of the educational process and served
as an encouragement to proceed past the masters level.
Dr. James Hensel, Chairman of the supervisory committee, provided
efficient guidance of the doctoral program and coordination of the
supervisory committee members. His inspiration, helpfulness, and
willingness to go beyond the input that is normally expected of an
advisor in terms of meeting requirements is held in deep appreciation.
Without Dr. Hensel my candidacy would not have been possible.
Dr. Margaret Morgan gave valuable assistance with the technical
aspects of the study and the manuscript. Dr. Lela Llorens contributed
knowledge of her specialty to add to the depth of the study.
Dr. Larry Gianchetta of the University of Montana was extremely
generous in contributing his assistance to the statistical design of
this study. Dr. Gianchetta and Dan Balias, at the University of
Montana computer center, were most helpful in the technical analysis of
the data.


58
Table 16
Models VII and VIIIClassification Functions
Variable
Group
= Work
No Work Overall F
Degrees
Freedom
of
//23 Music
.268
0.121 9.021
1 167
Constant
-1.093
-0.775
Table 17
Models VII and
VIIIClassification Matrix
Group
Percent
Correct
Number of Cases Classified into
Work No Work
Groups
Work
42.9
24
32
No Work
71.7
32
81
Total
62.1
56
113
Table 18
Models
VII and VIIIJacknifed Classification
Group
Percent
Correct
Number of Cases Classified into
Work No Work
Group
Work
42.9
24
32
No Work
71.7
32
81
Total
62.1
56
113
The first and only significant variable to enter was number 23, number
of hours in music.


22
for "personal reasons." Unsuccessful students produced more imcomplete
figures and drew the opposite sex first significantly more often than
did successful students.
Professional Performance and Persistence
Brandt (1967), Burton (1972), and Dubs (1975) studied the records
of nurses within one year of graduation. Brandt found that 79% of them
were working but no predictor variables were statistically significant.
Burton revealed that the 16 PF indicator predicted the performance of
new nurses slightly better than chance. Dubs used curriculum objectives
to develop an employer rating scale and then compared the results with
grades to see if the curriculum was doing what it was designed to do.
May and Chitty (1971) used the progressive matrices and Mill Hill
Vocabulary Test in their study of new state enrolled nurses in their
first jobs. They concluded that the tests could be used only as a means
of obtaining more intelligent student nurses, not for competent state
enrolled nurses. Di Marco and Hilliard (1978) decided there was no con
clusive evidence that associate, diploma, and baccalaureate degree nurses
were either the same or different in terms of the professional performance
variables measured.
Several studies have found negative results. A one year follow-up
study of clerical workers by Schmidt (1975) showed that the low ability
group was more satisfied rather than the middle or high groups.
Cornett (1969) studied the effectiveness of GPA and grades in two
classes in predicting performance of first year teachers on the job.
The best single predictor for secondary teaching performance was college


64
An initial regression analysis on approximately 55 variables showed
that better predictors could be identified that accounted for up to 50
percent of the movement of some of the dependent variables. However,
the use of these variables was not rational in terms of when the data
were available and when they were needed to make the selection of
students. These more significant variables occurred after selection
had been accomplished and were factors in the professional curriculum.
It was the intent of this study to develop a process model for use
in selecting students who would complete the program in occupational
therapy and remain working in the field. The process model could have
been the combination of all the variables that proved to be significant
for the different dependent variables. Table 19 shows all the
independent variables used in the eight equations and the frequency of
use.
Table 19
Frequency of Independent Variable Use
Frequency of Use in
Variable Eight Equations
College Entrance Examination Board Test Score 2
Number of credit hours in English and literature ... 3
Transfer student or University of Florida freshman . 3
Number of credit hours in psychology 2
College grade point average for the first two years . 3
Number of credit hours in music 1
Since five of the variables were seen two or more times within the
eight models, these could be considered more important. However, the


74
Lucci, JA: An approach to the selection of students. Am J OT 28(2):
91-93, 1974.
May, AE and Chitty, E: Selecting candidates for pupil nurse training.
Nurs Mirror pp 12-13, Dec 17, 1971.
Maynard, M, Bilkey, A, Hyre, K: Predicting Job performance. Am J OT
26(5):265-268, 1972.
McClain, EW: Sixteen Personality Factor questionnaire scores and success
in counseling. J Couns Psych 14(6):492-496, 1968.
McGreevy, CP: Factor analysis of measures used in the selection and
evaluation of counselor education candidates. J Couns Psych 14(1):
51-56, 1967.
Morgan, MK: The OPI, the ACT and university attrition: A discriminant
analysis. Doctoral Dissert, U of Kentucky, 1970.
Morgan, MK: Selecting candidates for over-subscribed programs.
Community College Review 5:65-73, Fall 1977.
Muchinsky, PM and Hoyt, DP: Predicting college grades of engineering
graduates from selected personality and aptitude variables. Ed &
Psych Meas 33:935-937, 1973.
Munro, BC: The Minnesota Teacher Attitude Inventory as a predictor of
teaching success. J Ed Res 57:138-139, Nov 1964.
Muthard, JE, Morris, JD, Crocker, LM and Slaymaker, JE: Field work
experience ratings and certification examination scores as
predictors of job performance and satisfaction in occupational
therapy. Am J OT 30(3):151-156, Mar 1976.
Neeley, R: Discriminant analysis for prediction of college graduation.
Ed & Psych Meas 37:965-970, 1977.
Nelson, AM: Undergraduate academic achievement in college as an
indication of occupational success. Professional Series 75-5.
Civil Service Comm, Wash DC, Pers Meas Res & Devel Ctr, Sept 1975.
Newsletter of MOTA. 3(2):4, 1979.
Nickens, JM: A comparison of accuracy of selected models for predicting
academic performance of junior college transfer students. J Ed Res
66(3):111-114, 1972.
Osipow, SH and Walsh, WB. Social intelligence and the selection of
counselors. J Couns Psych 20(4):366-369, 1973.
Owen, SV and Feldhusen, JF: Effectiveness of three models of multi
variate prediction of academic success in nursing education.
Nurs Res 19(6):517-525, 1970.


9
(III)Physical disabilities internship
(IV)Work status
on the basis of multiple combinations of variables which include college
entry test scores, courses taken, grade point averages, and other
objective data taken from transcripts.
Four generalized, multivariate null hypotheses were proposed for
discriminant analysis:
It will not be possible to discriminate among groups of students
(V)those who completed and (VI) those who did not complete the occupa
tional therapy program on the basis of linear combinations of variables
which include college entry test scores, courses taken, grade point
averages, and other objective data taken from transcripts.
It will not be possible to discriminate among groups of students
(VII) those who were working and (VIII) those who were not working in
the field of occupational therapy four to eight years later on the basis
of linear combinations of variables which include college entry test
scores, courses taken, grade point averages, and other objective data
taken from transcripts.
Limitations of the Study
1. This study was limited to obtaining, organizing and presenting
data of students who graduated from the University of Florida in
the Occupational Therapy curriculum from 1969 to 1973. During
this period the course training and faculty from year to year were
stable. Subsequent to 1973, there were major changes in the
faculty and curriculum.


36
A stepwise multiple regression analysis was calculated with the BMDP
computer package using the 30 independent and dependent variables. The
dependent variable was identified as y, the independent variable as X.
The process of narrowing the predictive independent variables to just a
2
few is called variable select. This yields R which is a coefficient
of determination; that is, it yields the percentage of the movement of
Y that is being explained by the X(s), or, the per centage of variance
in Y due to the regression on X(s).
Discriminant analysis is an extension of multiple regression
analysis which allows use of a dependent variable in the form of a
positive or negative occurance. Stepwise discriminant analysis uses a
similar process of identifying the best predictor at each step for the
dependent variable. It is used to determine correct classification
within specific groups, rather than identifying variables which relate
to better performance of the dependent variable.
The independent variables were tested for significance at the .05
level of confidence. The hypotheses were then accepted or rejected.
A variable was deleted when its .F ratio became too low. The "F to enter"
in the BMDP package program was 4.0 (given the size of the data set in
this problem, and "_F to enter" corresponds to a 5 per cent level of
significance).
Definitions of Statistical Terms
1. Correlation analysisthe degree to which two variables are related
(Lapin 1973).


11
graduate training plus internships and the National Registry
examination to be eligible for Registry in AOTA.
5. GPAGrade Point Average. The average is computed by dividing the
total number of grade points by the total number of course credit
hours attempted.
6. Non-persistence in working in the field of occupational therapy
not working in the field of occupational therapy four to eight years
after graduation.
7. Persistence in working in the field of occupational therapyworking
in the field of occupational therapy four to eight years after
graduation.
8. Physical Disabilities Internshipa three-month supervised on-the-
job training and learning situation approved by AOTA. This intern
ship is required as partial fulfillment for completion of the
occupational therapy program prior to the National Registry
Examination.
9. Psychiatric Internshipa three-month supervised on-the-job
training and learning situation approved by AOTA. This internship
is required as partial fulfillment for completion of the occupa
tional therapy program prior to the National Registry Examination.
Significance of the Study
With the contemporary emphasis on accountability, all schools are
being asked to justify their continuation based upon their purpose,
methods of achieving that purpose, and allocation of money, personnel
and equipment, thereby scrutinizing priorities and responsibilities to


19
Neely (1977) showed high school GPA and the ACT scores were most
consistently associated with post graduate work. However, Stein et al.
(1974) showed that overall performance of first year social work
students evaluated by faculty at the end of the year could be predicted
from students' attitudinal ratings done at the beginning of the year.
In summary, most of these studies found that academic achievement
at the college or graduate level could best be predicted by previous
performance either on achievement tests or with GPAs. The results,
though often statistically significant, were not sufficient for
practical application toward the selection of students into various
curricula, as supported by Hall and Coates' (1972) review of studies
completed between 1950 and 1968. His review stated that no reliable
variables have emerged to be of general value in the selection of nurses.
Practical Application
Lind (1970) predicted clinical practice grades of occupational
therapy students with college GPA. Tidd and Conine (1974) studied the
GPA of physical therapy students and found that academically better
students also tended to perform better in the clinic than did their
peers. Englehart (1957) showed that on-the-job performance during
clinical training of occupational therapy students may be predicted
from performance in some required college courses but other courses do
not predict clinical performance. Dunteman (1966) was not able to
predict clinical ratings from freshman and sophomore course performance.
Bailey et al. (1969) found that the Interest Test was the most effective
predictor of clinical performance of occupational therapy students.


4
in the health care delivery system. DuVal reported some 34,000 graduates
of allied health educational programs in 1971, compared with about 6,000
in 1960. The development of health manpower has meaning, not only to
society, but to education and the individual (Brown, 1973).
College students have also shown an increased interest in the
health care fields. An interest survey by the American Council of
Education in the fall of 1972 indicated fulltime freshmen showed 1.2
percent more interest in health care fields than those of the previous
year (Lind, 1970:91). The American College Testing (ACT) Center in Iowa
shows that for the 1977-78 school year, 17 percent of the high school
juniors and seniors tested identified health professions as a chosen
field. The number of qualified potential candidates has grown to the
point where traditional selection criteria are inadequate for several
health fields. Since the health-related curriculum is relatively more
expensive than the "academic" curriculum, selecting students who will
complete the program and be successful within their chosen health field
becomes a major problem.
Private and public funding agencies want to know where, how, and to
what effect tax dollars or private dollars are being spent. Cost
effectiveness in the health fields is of increasing concern. In addition
to financial accountability, funding agencies demand service account
ability such as how many trainees actually go to work in their field.
How many who enter their field continued to work in that field?
Therefore, predictive studies should include job entry and long term
persistence in the chosen health field as part of successful selection
processes.


50
of multiple combinations of objective variables used in this study.
Care should be taken in use of this model because only 15 percent of
the movement in Psychiatry Internship scores was explained by the
statistically significant variables, leaving 85 percent unaccountable.
By examining Table 5 for a random sample of this equation in use, one
can see that it does not work well in practical application.
Table 5
Model IIScore Comparison for Psychiatric
Internship-Actual and Predicted
Subj ect
Number
Psychiatry
Internship
Actual Score
TRANS
GPA-1
ENGLT
PSYCH
Predicted
Score
007
262
1
2.7
08
16
477.62
051
254
0
1.97
16.6
10
659.85
111
186
1
1.44
08
14
446.137
149
236
0
2.61
09
09
514.35
155
0
0
2.09
28
04
1082.025
165
0
0
3.0
32.1
09.5
1403.419
Model III
Physical Disabilities Internship
Score
(PDINT)
Variable number 4,
University
of Florida freshman or transfer
student,
was the first
significant
variable
with a
multiple
R2 of .031.
The equation appeared as:
Y = a + bX
PDINT = 231.082 lO^SX^^
The second step included in grade point average of the first two
2
years of college, number 24, raising the R to .077. The model
terminated at this step because of insufficient F ratio (see Table 6).


BIOGRAPHICAL SKETCH
Carolyn Dockstader Stewart was born Feburary 4, 1943, in El Paso,
Texas. She attended numerous public schools all over the northwast and
central United States, and graduated from Clinton High School, Clinton,
Iowa, in 1961. She attended the University of Iowa and received the
Bachelor of Science degree in occupational therapy in June 1965.
Upon graduation she worked as a staff therapist and assistant
director of occupational therapy in several clinical settings from 1966
to 1972. In 1971 she entered graduate school on a parttime basis at
the University of Florida. She received the Master of Education degree
in March 1973 with a major in special education. In December of 1973
she received the Specialist in Education degree with a major in
curriculum and instruction. Since that date she has been studying part-
time toward the Doctor of Philosophy degree in curriculum and instruction
While studying for the doctorate she has worked parttime as Assis
tant Professor of Education at the University of Montana, occupational
therapy and educational consultant to the Comprehensive Developmental
Clinic In Missoula, Montana, and special education consultant to the
Missoula County high schools.
From her marriage to Keith C. Stewart, M.D., of Cedar Falls, Iowa,
they have two sons, Chad and Dirk.
She is a member and a registered occupational therapist with the
American Occupational Therapy Association since 1966.
84


30
those who had scores and inserted for the missing scores rather than
using a zero to avoid distorting the statistical results (Winer 1962).
The means used were
Psychiatric Internship 255
Physical Disabilities Internship . 247
National Registry Examination .... 172
Permission was obtained from the Registrar's office of the
University of Florida to receive copies of transcripts for all students
involved in the occupational therapy curriculum for the period from
1967 to 1973. The social security number of each student, including
those who dropped out of the curriculum, was used to request these
documents. A staff member photocopied each transcript and the data
were forwarded to this researcher.
Collection of the data from the individual transcripts was done by
the researcher. Fifteen transcripts were selected at random and were
reworked several times. Various general categories were set up to
accommodate the variations in student course selections. A standard
ization of categories was finally achieved. For example,
Social sciences: history, anthropology, western civilization
Biological sciences: zoology, biology
Physical sciences: chemistry, physics, geology
Art: studio
Music: appreciation, lessons, choral
Literature: English and literature
College course grades were converted to points: A=4, B=3, C = 2
and D = 1. Semester grades were converted to quarter hour credits. The


63
Results of the multiple regression analysis of the 25 variables
showed five to be statistically significant at the .05 level: college
entry test score, college grade point average for the first two years,
transfer student or entry as a University of Florida freshman, number
of credit hours in English and literature, number of credit hours in
psychology. The discriminant analysis used all but one of the same
variables (CEEB) and added only one more: number of credit hours in
music.
In each analysis the eight corresponding null hypotheses were
rejected. Although statistically significant variables were revealed,
they were not useful for practical application in selection because
1. they accounted for such a small amount of the movement of the
dependent variables.
2. there were not enough significant variables to make a good
model.
3. the sign of the regression coefficient did not always seem
practical. A negative sign meant that selection of students
would be based on lower scores or fewer hours in these
variables.
All of the subjects in this study were part of a pre-selected
group which obviously narrowed the range of variability and affected
the results. This was demonstrated for each dependent variable by a
random sample of subjects' actual data used in each predictive equation.
The accuracy of the discriminant analysis was far less than might have
occurred by chance.


10
2. This study was concerned with 169 students' records, including
9 men and 160 women.
3. The students represented various environmental and educational
backgrounds. However, the only minority group represented was of
Spanish surnames (n = 3).
4. This study was not intended to determine the effect that the
different student backgrounds had upon individuals in the study
group. Nor was it intended that particular school systems should
be compared to one another as reflected by the achievement of their
graduates. Rather, the purpose was to review, as objectively as
possible, a group of students in an effort to determine which of
those would have difficulty in the pursuit of a specific college
program.
5. No attempt was made to evaluate the content of specific courses.
6. No attempt was made to appraise the teaching methods of individual
faculty members.
7. This study did not investigate why students left the curriculum.
Definition of Terms
The following is a list of terms to be defined as used within the
text of this study.
1. ACTAmerican College Test for scholastic aptitude.
2. AOTAAmerican Occupational Therapy Association.
3. CEEBCollege Entrance Examination Board. The sponsoring body for
the Scholastic Aptitude Test (SAT).
3. Complete the occupational therapy programto complete all under-


I certify that I have read this study and in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate,
in scope and quality, as a dissertation for the degree of Doctor of
Philosophy.
This dissertation was submitted to the Graduate Faculty of the Division
of Curriculum and Instruction in the College of Education and to the
Graduate Council, and was accepted as partial fulfillment of the
requirements for the degree of Doctor of Philosophy.
December 1979
Dean, Graduate School


Abstract of Dissertation Presented to the Graduate
Council of the University of Florida in Partial
Fulfillment of the Requirements for the
Degree of Doctor of Philosophy
FACTORS AFFECTING PERSISTENCE IN
EDUCATION AND EMPLOYMENT OF
OCCUPATIONAL THERAPISTS
By
CAROLYN D. STEWART
December 1979
Chairman: Dr. James W. Hensel
Major Department: Division of Curriculum and Instruction
The purpose of this study was to explore independent variables to
determine which, if any, could be used to predict success in an occupa
tional therapy curriculum, internship, registry examination, and job
persistence prior to the time professional course work is begun for use
in student selection. Ultimately the development of a process model
which could be used with future data for selection of candidates was a
desirable outcome.
The 169 subjects in this study all had been accepted into the
occupational therapy program for graduating classes spanning the five
year period from 1969 to 1973. Of the 160 women and nine men, 19 with
drew from the undergraduate curriculum. The 25 selected variables
occurring prior to professional course work were taken from University
of Florida transcripts which were available for all subjects. These
viii


CHAPTER III
DESIGN AND METHODOLOGY
The purpose of this study was to explore independent variables to
determine which, if any, could be used to predict success in an occupa
tional therapy curriculum, internship, registry examination, and job
persistence prior to the time professional course work is begun for use
in student selection. Ultimately the development of a process model
which could be used with future data for selection of candidates was a
desirable outcome.
This chapter contains an explanation of the procedures used in
this study. Initially, the population and selection of the sample are
described, followed by a description of the data collection procedures.
Subsequently, the method of analysis (including definitions of statis
tical terms) and the pitfalls of interpreting regression analysis are
discussed. The concluding section contains a description of how the
data are organized.
Population
The undergraduate occupational therapy curriculum at the University
of Florida is within the College of Health Related Professions and
served approximately 75 juniors, seniors and interns each year.
The five-year period from 1969 to 1973 was chosen for study because
during these years the course of study and the faculty remained quite
26


Subject
Number
it 2
WK/NWK
it 26
REGIST
#27
PYCHIN
it 28
PD INT
it 30
COMP
46
1
164
151.5
281.0
1
47
1
200
245.0
217.0
1
48
0
164
270.0
228.0
1
49
1
180
262.5
266.0
1
50
1
Mean 172
184.0
149.0
1
51
1
158
254.0
270.5
1
52
1
149
228.5
226.0
1
53
0
172
231.0
231.0
1
54
1
137
203.5
232.0
1
55
0
193
233.0
254.0
1
56
0
163
269.0
205.0
1
57
0
183
272.0
283.0
1
58
0
111
239.0
270.5
1
59
0
165
237.0
272.0
1
60
1
151
285.0
300.0
1
61
1
172
Mean 255.0
287.0
1
62
1
150
274.0
282.0
1
63
1
165
289.0
267.0
1
64
1
146
246.0
175.0
1
65
1
159
222.0
243.0
1
66
1
169
273.0
242.0
1
67
1
144
281.0
289.0
1
68
1
161
249.0
218.0
1
69
1
181
269.0
224.0
1
70
0
166
246.5
236.0
1
71
1
160
Mean 255.0
Mean 247.0
1
72
1
145
256.0
234.5
1
73
0
154
231.0
196.0
1
74
1
175
232.0
284.0
1
75
1
173
269.0
269.0
1
76
0
165
264.0
238.5
1
77
0
164
226.0
290.0
1
78
1
150
242.5
233.0
1
79
1
184
256.0
285.5
1
80
1
178
169.0
180.0
1
81
1
197
276.5
164.5
1
82
1
149
274.0
267.0
1
83
1
147
283.0
280.0
1
84
0
138
269.5
166.5
1
85
1
163
255.5
263.0
1
86
1
141
274.0
284.5
2
87
0
178
206.5
274.0
1
88
0
179
288.5
245.5
1
89
0
000
148.5
203.0
1
90
-
154
Mean 255.0
276.5
1
#31
WK STA
09
10
08
10
10
10
10
08
10
08
08
08
08
08
10
10
10
10
10
10
10
09
10
10
08
09
10
05
09
10
08
08
10
10
10
10
10
10
04
10
10
10
08
00
08


75
Palmieri AA III: A model for multivariate prediction of academic
success of "transfer students" in pharmacy school. Am J Pharm Ed
41(3):264-600, 1976.
Patterson, TW and Davidshofer, CO: A counseling and training program
for occupational therapy students. Am J OT 26:419-423, 1972.
Pedrini, BC and Pedrini, DT: Predictors of college success. ERIC
microfische ED086102, 1974.
Pennell, MY and Hoover, DB: Health manpower source book: Allied health
manpower supply and requirements 1950-80, Wash DC: Govt Printing
Ofc. No 263, Sec 21, 1970.
Pfeifer, CM Jr and Sedlacek, WE: Predicting black student grades with
nonintellectual measures. J Neg Ed 43(l):67-76, 1974.
Poole, MA and Kassalow, S: Manpower survey report: Wisconsin Occupa
tional therapy association. Am J OT 22(4):304-306, 1968.
Prediger, DJ: Prediction of persistence in college. J Couns Psych
12(1):62-67 1965.
Prendergast, N: Comments on issues in education. Am J OT 32(9):595
596, 1978.
Progress Notes, Oct 1970 through Fall 1974, Univ of Fla, Dept of OT.
Reavley, W and Wilson, LJ: Selection or diagnosis? The use of
psychometric tests with candidates for student nurse training.
Nursing Times pp. 183-184, Nov 16, 1972.
Reed, CL: Prediction of grade point averages using cognitive and non-
cognitive predictor variables. Psych Rep 32:143-148, 1973.
Reed, CL, Feldhusen, JF and Van Mondfrans, AP: Prediction of grade
point averages in nursing schools using second-order multiple
regression models. J Ed Meas 9(3):181-187, 1972.
Redfering, DL and Biasco, F: Selection and elimination of candidates
in counselor education programs. Couns Ed & Supr 34:298-303, 1976.
Richards, JM Jr, Calkins, EV, McCanse, A and Burgess, MM: Predicting
performance in a combined undergraduate and medical education
program. Ed & Psych Meas 34:923-931, 1974.
Roemer, RE: Nine year validity study of predictors of medical school
success. J Ed Res 59:183-185, 1965.
Rose, HA: Prediction and prevention of freshman attrition. J Couns
Psych 12(1):399-403, 1965.


Key to Abbreviations
WK/NWK
= working or not working in the field of
four to eight years after graduation.
occupational therapy
REGIST
= American Occupational Therapy National
Examination score.
Registry
PYCHIN
= Psychiatric Internship score.
PD INT
= Physical Disabilities Internship score
COMP P
= Completion of no completion of the educational program in
occupational therapy
WK STA
= Work status score four to eight years after expected
graduation.
79


27
stable. After 1973, there were major changes in both the faculty and
the curriculum. During this period of time, 9 men and 161 women
students were enrolled in the occupational therapy curriculum.
Of the 169 students, 20 dropped out of the program before finishing
the undergraduate degree. One student failed on internship and never
finished. Three students completed the undergraduate program and both
interships but did not make a satisfactory score on the registry
examination.
These students were selected for admission to the occupational
therapy curriculum based on the following criteria:
Student Requirements:
1. The student must have been an undergraduate student at an
accredited college or university.
2. Transfer students at the junior level must have been residents
of the State of Florida.
3. Nonresidents must have been enrolled at the University of
Florida since the freshman year.
4. Since the program begins only in the fall of each year, all
applicants must have completed a minimum of 96 transferable
quarter credits and all departmental prerequisites prior to
enrollment.
5. The student must have been admitted to the University of
Florida.
6. The student must have completed all departmental forms by
March 15 of the spring before admission for the following
fall. These included


dependent variables to be useful for practical application. This was
demonstrated for each dependent variable by a random sample of subjects'
actual data used in each predictive equation. The accuracy of the
discriminant analysis was far less than might have occurred by chance.
The results of the study indicated that combining both academic and
practical application parameters in both the independent variables and
the dependent variables does not enhance the practical use of the
predictive selection equation versus using just one of these parameters
in both. It was found that the same five variables were statistically
significant for both the academic and the practical application dependent
variables. Implications of the study were that a few variables could
be considered in the selection of potential candidates: college grade-
point average for the first two years, transfer student or University of
Florida freshman, number of credit hours in English and literature,
College Entrance Examination Board Test score, number of credit hours
in psychology, and number of credit hours in music. However, human
factors and subjective data should also be included. In conclusion, this
researcher found that a practical predictive equation cannot be developed
with the objective data used in the study.
x


31
categories tabulated in grade points were common to all transcripts.
Several categories were recorded in number of credits taken which might
indicate a proficiency and/or interest in an area.
When this study was initiated, 76 variables from the transcripts
were considered. However, many observations were deleted because they
were represented by only one or two students. Other variables were
deleted after an initial stepwise multiple regression analysis showed
that they entered the equation very late, were not statistically
significant, and accounted for a miniscule amount of movement of the
dependent variables. Only 24 variables which were represented by at
least eight per cent of the students were used in the final analysis.
These 24 variables were then keypunched according to the Biomedical
Computer Program P-Series and analyzed by the University of Montana
computer.
Ten transcripts did not have a score for the Florida Twelfth Grade
Placement Test nor the Scholastic Aptitude Test of the College Entrance
Examination Board (CEEB). An average score for all those who had scores
was computed and used rather than a zero (Winer 1962). The means used
were:
CEEB 1006
Florida 12th Grade Test .... 384
A table was obtained from the University of Florida Admissions
Office to convert the total score for the Florida 12th Grade Test to
the equivalent total for the Scholastic Aptitude Test scores of the
CEEB.
The American Occupational Therapy Association was contacted con
cerning follow-up on all subjects who had continued their national


41
With two groups, A and B, and two measurements X and Y, on each
member of the two groups, a scatter plot could be done for each
group. Then a circle is drawn around the groups, respectively.
A straight line is then drawn through the two points where the
ellipses intersect and then projected to a new axis, Z. The
overlap between the distributions A' and B' is smaller than would
be obtained by any other line drawn through the ellipses. The Z
axis is the discriminant function in this case. Group averages
and mean-square deviations of group values are computed. Then
that result is compared to the mean square deviations of the
group averages from the overall average. If the latter is large,
then the linear function used in these calculations separates the
groups well (Green and Tull 1970).
10. Standard error of the estimatethe square root of the mean squared
deviations about the regression line or plane. It may be used to
estimate the true variability in Y about the model. It directly
expresses the degree of scatter in the data (Lapin 1973).
Pitfalls of Regression and Correlation Techniques
When reading a correlation coefficient the reader must remember
that it reflects, in numerical value only, the strength of association
between variables. This is a statistical relationship; a large positive
2 2
or negative value for R or r does not indicate a cause and effect
relationship. Examples of nonsense or spurious correlations abound.
For most practical problems there is usually some reasonable connection.
Often a relationship between two variables may be explained by their


BIBLIOGRAPHY
Acquaviva, FA: AOTA Human Resources Project III, The Member Data Survey.
Am J OT 29(5):293-300, May-June L975.
Acquaviva, FA: The Member Data Survey IVDemographic Characteristics
of Occupational Therapy Personnel. Am J OT 29(7):426-431, Aug.
1975.
A'Hearn, RT: A Model for Predicting Academic Achievement of Technical
College Students. Doctoral Dissert, Marquette Univ, 1974.
Aiken, LR Jr: "Three Alignment Charts for Use in Selective Admissions."
J Ed Res, 62:14-17, Sept 1968.
Aldag, FC and Martin, MF: Physical Therapist Assistant Selection and
Academic Success. Phys Ther 55(7):747-750, 1975.
AOTA: Occ Ther Newletter 29(4):1, Jan 1975.
AOTA: "Occupational Therapy: 2001 AD" Papers presented at the special
session of the Representative Assembly, Nov. 1978, AOTA, Rockville,
Maryland, 1979.
Anderson, HE and Jantzen, AC: A Prediction of Clinical Performance.
Am J OT 19(2):76-78, 1965.
Ayers, JB and Rohr, ME: Relationship of Selected Variables to Success
in a Teacher Preparation Program. Ed Psych Meas 34:933-937, 1974.
Bailey, JP: Some Comparisons of Occupational Therapy Biographical Facts
with Implications for Recruiting. Am J OT 22(4):259-263, 1968.
Bailey, JP, Jantzen, AC and Dunteman, GH: Relative Effectiveness of
Personality Achievement and Interest Measures in the Prediction of
a Performance Criterion. Am J OT 23(1):27-29, 1969.
Baker, E Jr: Associate Degree Nursing Students: Nonintellective Dif
ferences between Dropouts and Graduates. Nurs Res 24(1):42-45,
Jan-Feb 1975.
Bean, AG: The Prediction of Performance in an Educational Psychology
Master's Degree Program. Ed & Psych Meas 35(4):957-962, 1975.
68


15
(1968), and Reed et al. (1972) concurred that lower ACT scores and lower
high school rank were more closely related to withdrawal than higher rank
and scores. Less than 30 per cent of the lower scoring group stayed to
the second semester of the junior year (Humphreys et al. 1973).
Hall and Coates (1973), Pedrini and Pedrini (1974), and Pfeifer and
Sedlacek (1974) stated again that high school achievement and aptitude
test scores predicted first year college performance for racial minority
groups about the same as for whites. Bistreich (1977) found overall high
school grade point average (GPA) and high school GPA in science and
English were significant predictors of junior college GPA. Inkenberry
(1961) tried to predict first year college persistence using noncognitive
variables and intellective functions. He also found that the intellec
tive function was the most significant variable with a part of the
cultural area: sex and social background.
Brazziel (1977) and Chase and Johnson (1977) both found high school
rank to be a statistically significant predictor of success in college
through the second year. They found two other significant variables:
hours of high school English and level of family income.
Within specific technical curricula, R. Brown and Stones (1972)
found no close association between intelligence factors and withdrawal
rates in the first year of nurses' training. Davies and Khosla (1974)
studied health visitor students using intelligence test scores as a
predictive tool for those who performed at high levels in academics.
Their multiple correlation was R = .52.
Reed et al. (1973) found some noncognitive variables to be signifi
cant predictors of first year success of nursing students: age, year


56
VI No Complete = -16.47 2.443X1 (TRANS) + .334X2 (PSYCH)
+ 10.218X3 (GPA-1) + .202X4 (ENGLT)
The reader can refer to Table 13 for a summary of variables entered
by step for completion of the occupational therapy program. Table 14
shows the actual and predicted scores using a random sample of data.
Table 13
Models V and VISummary-Variables Entered by Step
Step Number
Variable
Entered
F ratio
Overall F
Degrees of
Freedom
1
TRANS
5.806
5.806
1 167
2
PSYCH
5.729
5.850
2 166
3
GPA-1
4.311
5.414
3 165
4
ENGLT
5.078
5.431
4 164
Table 14
Model
VScore Comparison for
Comple t ed-Ac tua1
and Predicted
Subject Number
Actual
TRANS
PSYCH
GPA-1
ENGLT
Predicted
007
1
1
16
2.7
08
28.637
051
1
0
10
1.97
16.6
25.112
111
1
1
14
1.44
08
17.464
149
1
0
09
2.61
09
28.207
155
0
0
04
2.09
28
28.618
165
0
0
9.5
3.0
32.1
39.252
Model
VINo
completion
of occupational
therapy program
is the
other half
of the
dichotomy.
Table
15 shows
that this equation lacks
usefulness when applied to a practical situation.


CHAPTER V
CONCLUSIONS, OBSERVATIONS AND RECOMMENDATIONS
Conclusions
The purpose of this study was to identify objective variables which
might prove useful for selection of potential candidates for entry into
a health related curriculum who would most likely be able to complete
the academic requirements and meet the qualifications for registry in
occupational therapy, enter the job market, and continue within the
field of occupational therapy.
A study of the literature revealed that no study combined all
of the following:
1. practical application data in the dependent and independent
variables.
2. academic data in the dependent and independent variables.
3. a sample size larger than 50.
A. data on subjects for more than three successive years.
5. followup on the same subjects more than one year after
graduation.
6. followup on the subjects who withdrew from the undergraduate
curriculum.
7. use of only objective data from transcripts.
This study included all of the above on the same population.
60


3
programs. An error in judgment in selecting a student who will later
drop out of a program or field of occupational therapy represents a
personal and professional loss of two people: the one who dropped out
and the one who missed the original opportunity to take that opening
(Morgan, 1970).
How can the best students for the health-related professions be
selected? What factors, such as abilities, past experiences, person
ality attributes, styles of life, and interests, are related to
proficiency in a professional world? What information about an
individual is associated with or will permit prediction of future
competencies in his training and persistence in his health care field?
Are specialized programs training students who will remain in that work
force? With the large numbers of applicants, one knows that many are
not qualified academically or personally. However, a significant number
are. How then does one objectify in part or in toto this process in
order to select those most suitable?
Historically the public wants the best health service available
regardless of the expense. Encouraged by the mass media, the public
expects services from the health system that it did not know existed
prior to the 1970s. People of the 70s began demanding a drastic
increase in health service, including comprehensive health care for all.
Health care delivery system became a catch phrase with a powerful impact.
Mase, as reported by Brown (1972), stated that health care was the second
largest industry in the United States. This demand for care precipitated
a drain on the health system and created a need for additional allied
health manpower to provide support services for physicians and dentists


time spent in selection, and to keep prejudices to a minimum, it would
be desirable to be able to use information that has been collected and
is available on institution transcripts. Most past studies involved
administration of additional tests of questionnaires to prospective
candidates. If specific objective data from transcripts could be used
as valid predictors of completion and persistence in the work field,
unskilled office help could do most of the selecting of candidates.
Faculty could then be minimally involved making the final decisions
concerning the few borderline cases.
In a survey of 39 out of the 40 schools contacted for bachelor's
degree occupational therapists in 1974, Johnson, Arbes, and Thompson
found that 17 percent of the schools accepted fewer than half of their
respective applicants. They also found that:
1. All but three used interviews
2. All used grades in college for the first two years
3. All used aptitude and achievement test scores
4. About half used letters of recommendation
5. Ten used an application letter or essay
6. Seven used a personality or interest inventory
7. All used high school grades
8. Two used a biographical questionnaire
9. One used a lottery
The University of Florida Department of Occupational Therapy
annually has 35 openings for students. The Office of the Registrar
sends that department approximately 100 applications. The selection
procedure is similar to that used by most other occupational therapy


21
High school performance and standardized achievement tests were the
best predictors of scores on nursing State Board Tests (Haglund 1975,
Tatham 1976, and Weber et al. 1973). Larken (1976) found that scores on
the comparative Guidance and Placement Tests (CGP) taken at entry to a
nursing curriculum appeared to discriminate between State Board Test
passes and failures. Brandt et al. (1966) showed a lack of relationship
between grades in nursing theory courses and scores on State Board Tests.
A low correlation between the occupational therapy National
Registry Examination score and Guilford's STDCR InventoryR Factor was
revealed by Booth (1957). Engelhart (1957) found that some grades in
specific courses predict occupational therapy Registry Examination
scores.
Fredricks et al. (1970) investigated the relationship between medical
students' performance on the National Board Examinations and their
performance on the MCAT and their academic achievement over the four
years of medical school. The most effective predictor for performance
on the National Board Examination was the GPA from the second year of
medical school.
Bergman et al. (1974) conducted a comprehensive study in Israel on
589 nursing students. They found psychological tests correlated signif
icantly with achievement in theoretical learning. Personality factors
were studied by Krall (1970) in relation to nursing success and failure.
The Interpretation of a Draw-A-Person Test and scores from the National
League of Nursing (NLN) test were used to compare successful and failed
students. The NLN test was found to be significantly related to
nursing school withdrawal in the first year for those who did not leave


5
This study explored possible combinations of variables, such as
undergraduate courses, college entrance examination scores and other
transcript data as indicators of students who completed and worked in
one health care fieldoccupational therapy. All of the students in
this study graduated from the University of Florida.
General Purpose and Need for the Study
A need exists to objectively identify prospective college students
who would be able to complete the requirements for registration in
occupational therapy, enter the job market and continue within the field
of occupational therapy. The purpose of the present study was to explore
independent variables such as college entry test scores, courses taken,
and grade point averages to determine which, if any, could be used prior
to the beginning of professional course work as indicators of those
who would be most likely to graduate, then be employed, and continue
working in the field of occupational therapy.
Previous studies sought predictors of (1) academic performance,
persistence, or withdrawal, (2) performance in practical application of
training such as internships, (3) professional knowledge as measured by
written tests, (4) short term professional performance and long term
persistence. None of these studies included all four variables on the
same population. This study sought to generate meaningful results by
using variables from each category as they related to a larger number
of subjects (n = 169). More specifically the purposes were to
1. Identify the best pre-professional indicators of persistence
in working in the field of occupational therapy.


35
6. Working or not working in the field of occupational therapy
at the time of this study.
A scale employing the numbers zero through ten was developed to
describe different levels of work status (see Table 1). The least
desirable level, indicated by a zero, represented no available followup
on those who did not complete the undergraduate curriculum in occupa
tional therapy. The number ten represented the most favorable follow-up
information of working full time in the field of occupational therapy.
Table 1
Work Status Scale
Level
Number
Description
0
1
2
3
A
5
6
7
8
9
10
No followup and did not complete the undergraduate
curriculum in occupational therapy
Did not complete the undergraduate curriculum in
occupational therapy and not working
Did not complete the undergraduate curriculum in
occupational therapy and is working in an unrelated field
Did not complete the undergraduate curriculum in
occupational therapy and is working in a health related
profession
No followup data; did complete the undergraduate curriculum
in occupational therapy, and recorded as not registered.
Did complete the undergraduate curriculum in occupational
therapy, not registered, and working in an unrelated field.
No followup data and recorded as registered
Registered and working in unrelated field
Registered, not presently working but plans to be back
to occupational therapy
Registered and working parttime in occupational therapy
Registered and working fulltime in occupational therapy


59
Null hypotheses VII and VIII were rejected since it was statistically
possible to discriminate among groups of students (VII) those who are
working and (VIII) those who are not working in the field of occupational
therapy four to eight years after graduation on the basis of linear
combinations of objective variables used in this study.
Y = a + bX
Work = 1.095 + .268 (MUSIC)
No Work = -0.775 + .121 (MUSIC)
Practical application shows the predictive weakness of this equation
also.
Model VIII not working in the field of occupational therapy was the
other half of the dichotomy. This again was not a practical equation,
especially since there was only one variable that was statistically
significant.
It was the intent of this study to develop a process model for use
in selecting students who would complete the program in occupational
therapy and remain working in the field. The process model could have
been the combintation of all the variables that proved to be significant
for the different dependent variables. Since five of the variables were
seen two or more times within the eight models, these could be considered
more important. However, the amount of movement that could be accounted
for was too small to be of practical use.


CHAPTER IV
ANALYSIS OF DATA AND RESULTS
Introduction
This chapter was organized around the analysis of data in order to
develop a process model. After a discussion of the specific computer
programs used in this study the results of the data analysis were
presented in six major sections based on six dependent variables. Each
dependent variable was shown with its respective prediction equation,
significance level of the independent variables, and examples of student
performance predictions with comparisons of actual data realized. All
data from 169 subjects were used throughout. Stepwise multiple regres
sion analysis was used for predictors of the following dependent
variables:
1. National Registry Examination Score
2. Psychiatry Internship Score
3. Physical Disabilities Internship Score
4. Work Status Score
Stepwise discriminant analysis was used for predictors of correctly
classifying the subjects into dichotomous groups represented by the
following dependent variables:
1. Complete or not complete the program for registry in
occupational therapy.


70
Buca, JT: An Investigation of the Attrition and Retention of Physical
Therapy Students at Univ of Maryland. Unpub Thesis, Univ of Mary
land, Baltimore 1967.
Buhmeyer, KJ: Predicting Success in a Physician Extender (Medex) Program.
Doctoral Dissert, Walden Univ, July 1975.
Burgess, MM: Two Studies of Prediction of Success in a Collegiate
Program of Nursing. Nurs Res 21(4):357-366, July-Aug 1972.
Burton, DA: The Selection of Nurses by Discriminant Analysis. Int J
Nurs St 9:77-84, 1972.
Chase, DI and Johnson, JJ: Predicting College Success with Nontradi-
tional Data for Inner-city Students. J Col St Pers 210-214,
May 1977.
Chen, MK, Podshadley, DW and Shrock, JG: A Factorial Study of Some
Psychological, Vocational Interest and Mental Ability Variables as
Predictors of Success in Dental School. J Appl Psych 51(3):236-
241, 1967.
Clapp, TT: Comparison of Three Models for Selection of Applicants and
Prediction of Success in Medical School. Doctoral Dissert, Univ
of Missouri, Columbia, 1975.
Conine, TA: A Survey of the Graduates of a Professional Physical
Therapy Program. Phys Ther 52(8):855-861, 1972.
Cornett, JD: Effectiveness of Three Selective Admissions Criteria in
Predicting Performance of First-Year Teachers. J Ed Res 62:247-
250, Feb 1969.
Crane, WJ: Screening Devices for Occupational Therapy Majors. Am J 0T
16(3):131-132, 1962.
Cramer, EM: Brief Report: The Use of Highly Correlated Predictors in
Regression Analysis. Mult Beh Res 241-243, Apr 1974.
Cypres, RM: Attrition and Related Variables in a Four Year Baccalaureate
Physical Therapy Curriculum. Unpub Thesis, Boston Univ, Sargeant
College 1969.
Davies, ME and Khosla, T: Predictive Value of Two Intelligence Tests as
Criteria of Success in a Health Visitor Examination. Nurs Times
28:109-112, 1974.
Di Marco, N and Hilliard, M: Comparisons of associate, diploma, and bac
calaureate degree nurses' state board, quality of patient care,
competency rating supervisor rating, subordinates' satisfaction
with supervision and self-report job satisfaction scores. Int J
Nurs St 15:163-170, 1978.


16
of entry into the program, level of previous education, and schools
attended. Stress and satisfaction of first year nursing students were
assessed by a questionnaire developed by Katzell (1968). Statistically
significant negative correlations were shown between withdrawal and
experienced satisfaction and expectations. First semester grade point
average in a physical therapy program was best predicted by the Educa
tional Interest Inventory Fine Arts Scale and the score of the Attitude
Toward Disabled Persons Scale (ATDPS) (Buca 1967). In a study by
Schimpfhauser and Broski (1976) of first year academic success in a
health related professions program the ACT score was found to be the
best predictor. Lowenthal et al. (1977) found that the Pharmacy College
Admission Test (PCAT) accounted for 25 per cent of the variation in
completion of a pharmacy curriculum. However, Liao and Adams (1977)
found that prepharmacy academic performance was the best predictor of
first year pharmacy GPA, not the PCAT. Shalik (1978) presented a con
vincing study showing no statistical difference between the usual
selection process for a physical therapy program and a stratified
sampling process for student selection.
Specific engineering course grades and GPA were best predicted by
the Achievement and Affiliation scales of the Edwards Personal Prefer
ence Schedule (EPPS) and the scores on American Council on Education
Psychological Examination (Muchinsky 1973) .
In studying transfer students, Nickens (1972) stated that junior
college GPA was the best predictor of first term Florida State
University GPA.


72
Halfter, IT: Validity studies: Law school. J Ed Res 59:307-309, Mar
1966.
Hall, JN and Andrews, WT: Apparent homogeneity in characteristics of
student nurse groups. Int J Nurs St 9:103-109, 1972.
Hall, R and Coates, CJ: Prediction of academic success for special
program and regularly admitted college freshmen. Col & Univ,
14-18, Fall 1973.
Heit, P, Johnson, G, Meeks, LB and Paxon, C: A selective admissions
process as a predictor of academic success in health education.
J Sch Hlth 48(3):146-500, 1978.
Hendrickson, D: Personality variables: Significant departures of
Occupational therapy from population norms. Am J OT 16(3):127-
130, 1962.
Hinkel, D and Houston, CA: A comparative study of the effectiveness of
two Bayesian models for predicting the academic successes of
selected allied health students enrolled in the comprehensive
community college. Paper presented at the Annual Mtg of Am Ed Res
Assoc, NY, NY Apr 4-8, 1977.
Holmstrom, El: Promising prospects: Students choosing therapy as a
career. Am J OT 29(10):608-614, 1975.
Holt, LI: A study to identify predictors of success in academic programs
and in occupations for students in medical laboratory technician
programs in community colleges, Unpub Doctoral Dissert, Univ of
Connec, 1978.
Humphreys, LG, Levy, J and Taber, T: Predictability of academic grades
for student of high and low academic promise. Ed & Psych Meas 33:
385-392, 1973.
Inkenberry, SO: Factors in college persistence. J Coun Psych 8(4):322-
329, 1961.
Jantzen, AC: Some characteristics of female occupational therapists.
Am J OT 26(1):19-26, 1972.
Johnson, JA: Nationally speaking. Am J OT 28(9):521-526, 1974.
: Nationally speaking: Issues in education. Am J OT
32(6):355-358, 1978.
Johnson, RW, Arbes, BH and Thompson, CG: Selection of occupational
therapy students. Am J OT 28(10):597-601, 1974.
Johnson, RW and Leonard, LC: Psychological test characteristics and
performance of nursing students. Nurs Res 19(2):147-150, 1970.


52
Table 7
Model IIIScore Comparison for Physical Disabilities
Internship-Actual and Predicted
Subj ect
Number
Actual
Phys. Dis.
Internship
TRANS
GPA First
Two Years
Predicted
Score
007
256
1
2.7
294.97
051
270.5
0
1.97
211.36
111
144
1
1.44
151.306
149
174
0
2.61
235.963
155
0
0
2.09
215.787
165
0
0
3.0
388.059
Table
8
Model IV-
-Variables
Entered by Step
Dependent
Variable
Independent
Variable
Variable
Number
Percent of
Movement
at Each Step
Overall
F
Work Status
#31
CEEB
29
.028
4.80
Null hypothesis IV was rejected since it was statistically possible
to predict scores in work status on the basis of multiple combinations
of the objective variables used in this study. The accuracy in practical
application of this prediction equation is very low as shown in Table 9
using a random sample of subjects.


DEDICATED TO
ALICE C. JANTZEN
An educator who has brought the future to the field
of occupational therapy. An inspiration to those who are
fortunate enough to be touched by her.


38
Figure 1. Simple Linear Regression (p. 447)


42
interactions with a common factor. Care must be taken in using past
data to determine a future relationship. The conditions under which
the Y and X data were obtained may have changed over time. Caution
should be used when making predictions of Y whenever X falls outside
the range of values used in the original collections of data (Lapin
1973).
Organization of the Data
For the purposes of this study the students data were organized
into their respective five classes and arranged alphabetically within
each class. The earliest metriculated class was listed first. Each
subject was then numbered consecutively, starting with A in the earliest
class, 1967. Those who dropped out of the undergraduate curriculum were
alphabetized and consecutively numbered at the end of the before
mentioned groups.
Summary
This chapter has presented the procedures used in this study. The
population and selection of the sample were described followed by data
collection procedures and the method of analysis. Definitions of
statistical terms used in this study were given. The pitfalls of
interpreting regression analysis were briefly discussed and, finally,
the organization of the data was presented.


I certify that I have read this study and in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate,
in scope and quality, as a dissertation for the degree of Doctor of
Philosophy.
/ 'Lz.
UJ
"James W. Hensel, Chairman
Professor of Curriculum and Instruction
I certify that I have read this study and in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate,
in scope and quality, as a dissertation for the degree of Doctor of
Philosophy.
ti.
Lela A. Llorens
Professor of Occupational Therapy
I certify that I have read this study and in my opinion it conforms
to acceptable standards of scholarly presentation and is fully adequate,
in scope and quality, as a dissertation for the degree of Doctor of
Philosophy.
Margaret K. Morgan
Associate Professor of Curriculum
and Instruction


83
Subject #2 #26 #27 #28 #30 #31
Number WK/NWK REGIST PYCHIN PD INT COMP P WK STA
136
1
189
291.0
282.0
1
10
137
1
174
255.5
240.5
1
10
138
0
000
244.5
237.0
0
02
139
1
150
233.0
288.5
1
10
140
1
188
232.0
218.0
1
09
141
1
162
262.5
280.5
1
10
142
0
159
235.0
285.0
1
05
143
1
184
166.5
244.0
1
10
144
1
167
267.0 Mean
247.0
1
09
145
1
148
237.0
248.0
1
10
146
0
169
257.0
169.0
1
06
147
1
151
260.5
151.0
1
10
148
1
190
219.0
190.0
1
09
149
1
174
236.0
174.0
1
10
150
0
-
-
-
0
02
151
0

-
-
0
02
152
0
-
-
-
0
00
153
1
197
265.0
255.0
1
09
154
-
-
-
-
0
00
155
0
-
-
-
0
03
156
0
-
-
-
0
01
157
0
-
-
-
0
02
158
0
-
-
-
0
01
159
0
-
-
-
0
00
160
0
-
-
-
0
01
161
0
_
-
-
0
00
162
0
-
-
-
0
00
163
0
-
-
-
0
00
164
0
-
-
-
0
00
165
0
-
-
-
0
03
166
0

-
-
0
02
167
0
-
-
-
0
00
168
0
-
-
-
0
03
169
0
-
-
-
0
02


CHAPTER I
INTRODUCTION
General Background and Statement of Problem
Accountability was a movement of the 70s. The ultimate account
ability of an expensive technical education may be measured by
statistics showing a high percentage of completion of the program and
persistence within the field.
The concepts of predicting persistence and graduation from college
have been the substance of many studies, especially in the last ten
years. Since a 50 per cent attrition rate of college students prior to
graduation represents a great loss of time, money and energy to all
concerned (Morgan 1970), there has been a thrust to find ways to
decrease this high rate. Predicting professional knowledge, application,
and persistence have also been studied. However, combining all these
variables, both academic and applied measures, into one study on the
same population has rarely been done. This researcher did not find
information about such a study that had thus far been completed.
The objective of streamlining the complex process of selecting the
best qualified students who will complete oversubscribed programs has
been a part of accountability. Selection processes for persistence
in working in the field past the first year have not been adequately
addressed. To expedite the selection process, to cut valuable faculty
1


76
Schimpfhauser, FT and Broski, DC: Predicting academic success in allied
health curricula. J Alld Hlth 25:35-46, Winter 1976.
Schmidt, JB: Prediction of success in clerical occupations from ability
test scores. Voc Gdnc Qrtly 30:68-72, 1975.
Seitz, LA: Towards a selection and admissions model: Predicting
academicsuccess in veterinary school. Doctoral Dissert,
Ohio State Univ, 1973.
Shalik, H: A multivariate process model to evaluate undergraduate PT
students admitted by stratified sample and by ordinal rank pre
selection. Doctoral Dissert, Univ of Fla, 1978.
Singh, A: The predictive value of cognitive tests for selection of
pupil nurses. Nurs Times 20:89-92, 1972.
Stein, RF and Green, EJ: The graduate record examination as a predictive
potential in the nursing major. Nurs Res 19(l):42-47, 1970.
Stein, S, Linn, MW and Furdon, J: Predicting social work student
performance. J Ed Soc Wrk 10(3):85-92, 1974.
Stone, RG and Acquaviva, FA: AOTA Human Resources Project II. The
Education Study. Jm J OT 29(4):233-239, 1975.
Sussman, MB and Haug, MR: Rehabilitation counselor recruits. J Couns
Psych 15(3):250-256, 1968.
Tatham, EL: Nursing and dental hygiene selection procedures. Part 2:
An examination of academic variables as predictors of success.
Johnson Co Community College, Overland Park, Kansas, Mar 1976.
Thom, DJ and Hickcox, ES: The selection of new graduate student for an
educational administration program. J Ed Adm 13(l):23-34, 1975.
Tidd, GS and Conine, TA: Do better students perform better in the
Clinic? Phys Ther 54(5):500-505, 1974.
Tiedeman, DV: The utility of the discriminant function in psychological
and guidance investigations. (The Multiple Discriminant Function
A Symposium). Harvard Ed Rev 121:71-80, 1951.
Tucker, AC and Strong, EK Jr: Ten-year followup of vocational interest
scores of 150 medical college seniors. J Appl Psych 46(2):8186,
1972.
U.S. Dept of Health Manpower. Health Resources Statistices: Health
Manpower Facilities, 1974. Wash DC, Govt Printing Office, Library
of Congress No. 66-62580, 1974.


23
GPA; however, this correlation was very low. McGreevy (1967) found the
Miller Analogies Test did an "acceptable" job of identifying academic
ally talented people.
In physical therapy, Walker (1959) found no factor to be a signif
icant predictor of graduation and entrance into the profession from
grades and achievement measures. Anderson and Jantzen (1965) agreed
that grades and achievement measures were ineffective predictors of
clinical performance in another health field, occupational therapy.
Muthard et al. (1976) did a short term follow-up on occupational therapy
graduates. They failed to show that the registry examination and
internship scores had predictive validity for specific indices of first
year job success.
Long-term follow-up in the health fields has been limited. The
studies that have been done were more descriptive than predictive.
Hendrickson (1962) and Brown and Stones (1972) described occupational
therapy personality characteristics and job satisfaction. Brollier (1970)
expanded her description to include not only occupational therapists,
but physical therapists and social workers in a small sample of each.
The differences appeared to be related to the different job concerns of
the social psychiatric group versus the physical disabilities group.
Sussman and Haug (1968) and Jantzen (1972) described the typical
person in their respective fields. The implication is that such data
could possibly be used for recruitment and selection of future students
who would be most likely to succeed in a career of these health related
fields.
Drugo (1975), Maynard et al. (1972), Nelson (1975), Thom and Hickcox
(1975) and Tucker and Strong (1962) carried out predictive studies. Drugo


48
Null hypothesis I was rejected since it was statistically possible
to predict scores on the occupational therapy National Registry Examina
tion on the basis of multiple combinations of the objective variables
used in this study. The weakness of the prediction accuracy of Model I
is shown in Table 2 using a random sample of subjects' actual scores
for examples.
Table 3
Model IScore Comparison for National Registry
Examination-Actual and Predicted
Subject
Number
Actual
Registry
CEEB
ENGLT
Predicted
Score
007
186
860
08
129.7
051
158
1048
16.6
110.56
111
155
1006
08
157.74
149
174
1100
09
163.67
155
0
1020
28
117.67
165
0
820
32.1
92.21
Model IIPsychiatric Internship Score (PYCHIN)
The first step in the model accounted for five percent of the
2
movement or a multiple R of .053. The first significant variable was
number 4, University of Florida freshman or transfer student (TRANS).
The equation was:
Y = a + bX
PYCHIN = 241.634 39.606X1
The second most significant variable was grade point average of the
first two years of college, number 24 (GPA-1). Adding this to the equa-
2
tion brought the multiple R to .095. The equation at this point looked


77
U.S. Dept of Labor. Occupational Outlook Handbook 1976-77. Wash DC,
Govt Printing Office, Bulletin 1875, 1976.
Walberg, HJ: Scholastic Aptitude, the National Teacher Examinations,
and teaching success. J Ed Res 61:129-131, 1967.
Walker, SM: A study of the value of the admissions requirements in
predicting graduation from a physical therapy program and entrance
into the profession, Masters Thesis, Boston Univ, 1959.
Weber, L, King, ME and Pitts, G: Predicting success in schools of
practical nursing. Ed & Psych Meas 33:943-943, 1973.
Wittmeyer, AL, Camiscioni, JS, and Purdy, PA: A longitudinal study of
attrition and academic performance in a collegiate nursing program.
Nurs Res 20(4):339-347, 1971.
Willingham, WW: Predicting success in graduate education. Sci 183(25):
273-278, 1974.
Winer, BJ: Statistical Principles in Experimental Design. McGraw-Hill
Book Company, Inc. New York 1962.


46
work. Output from stepwise discriminant analysis program (BMDP7M)
includes at each step:
1. Wilk's Lambda
2. Approximate _F statistic
3. Degrees of freedom
4. Constant
5. For variables in the equation:
a. F to remove
b. Tolerance
6. For variables not in the equation:
a. £ to enter
b. Tolerance
Stepwise Multiple Regression Analysis
Each model has been described through its development at each step.
An equation was shown for each step and some examples of using the final
equation was shown through the randomly chosen samples.
Model INational Registry Examination Score (REGIST)
The first step in the model accounted for six percent of the
2
movement or a multiple R of .057. This equation was as follows:
Y = a + bX
Y = the National Registry Examination Score (REGIST)
a = the constant as computed at each step
b = the coefficient as computed for the independent variable
at each step


61
The 169 subjects in this study all had been accepted into the
curriculum for graduating classes in occupational therapy spanning the
five year period from 1969 to 1973. Of the 160 women and nine men,
19 withdrew from the undergraduate curriculum. The 25 selected
variables occurring prior to professional course work were taken from
University of Florida transcripts. University of Florida transcripts
were available for all subjects. These variables were used as probable
factors explaining the movement of the dependent variables.
The study was executed in two statistical parts. The first part
encompassed the objectives of finding the best pre-professional course
work indicators for scores on the following:
1. Psychiatric Internship
2. Physical Disabilities Internship
3. National Registry Examination
4. Work Status Follow-up.
This part was designed to compute a stepwise multiple regression
analysis using the Bio-Med BMDP2R statistical packaged program.
The four multivariate null hypotheses proposed for the multiple
regression analysis were all statistically rejected.
It was not possible to predict scores for
(I)the National Registry Examination
(II)Psychiatry internship
(III)Physical disabilities internship
(IV)Work status


LIST OF FIGURES
Figure Page
1. Simple Linear Regression 38
2. Multiple Regression 39
3. Discriminant Function 40
vii


17
Baker (1975), Bailey (1968), Bean (1975), Holmstrom (1975), and
Lee and Shepard (1973) conducted descriptive studies of students who
were persistors and withdrawals of programs in health related fields.
The persistors and graduates were generally better adjusted, more inde
pendent, self-confident, responsible, and intelligent than those who
withdrew.
Prediger (1965), Reavley and Wilson (1972), and Rose (1965) found
that counseling served as a deterrent to freshman withdrawals. However,
Prediger found that the past achievement dimension made the most
difference.
Morgan (1970) used Omnibus Personality Inventory factors and ACT
scores to discriminate between groups for persistence in college seven
semesters after beginning. Her results showed that students differed by
sex in personality configurations, ability, and college persistence.
However, the best predictive variables for classifying the students into
groups were ACT scores, followed by an authoritarianism measure. Aldag
and Martin (1975), Cypres (1969), and Wittmeyer et al. (1971) concurred
that ACT scores differentiated the students who completed the college
curriculum and those who withdrew. Cypres suggested that high school
class rank should also be a part of a selection process. In addition,
Wittmeyer et al. found that students who completed the program tended to
be less independent and less venturesome as measured by the 16 Person
ality Factor Indicator (PFI) and the Myers-Briggs Type Indicator.
Johnson (1970) used the 16 PFI, Strong Vocational Interest Blank (SVIB),
and a college qualification test to predict nursing theory and practice
grades. His results showed that abstract reasoning ability was the best
predictor. Completion of a teacher training program was studied by


13
Chapter III describes the manner in which the data were obtained
and processed, specifically dealing with the design and mathematical
methodology. Chapter IV contains the presentation of these data while
Chapter V contains a summary of the results, conclusions, observations
and recommendations.


47
X = the independent variablein this case number 29, total
score on CEEB
Therefore, REGIST = 67.188 + .084X
When the second variable was added, = number 18, number of hours
2
in English and literature (ENGLT), the two yielded a multiple R of
.075. The equation for the second step was:
Y = a + b-^X + ^>2^2
REGIST = 88.733 + .085X1 2.063X2
The model was cut off at the second step because there was
insufficient F ratio to continue the stepping process. Table 2 shows
the variables entered by step, percent of movement at each step, the
overall F ratio at each step. The prediction equation developed by the
step wise multiple regression analysis for the score on the National
Registry Examination, as shown in step two description, gave statisti
cally significant results but the reader should focus on the fact that
Table 2
Model IVariables Entered by Step
Percent of
Dependent
Independent
Variable
Movement at
Overall
Variable
Variable
Number
Each Step
F
National
Registry
CEEB
29
.057
10.04
Examination
ENGLT
18
.075
7.84
(#26)
the model explains only eight
percent of
the movement in the
National
Registry Examination scores.
That means
that 92 percent of
the movement
in scores were not accounted for.


49
like this:
Y = a +
biY1 +
b2X2
PYCHIN = 154,004 49.124X][ + 34.994X2
At the third step the number of hours in English and literature
number 18 (ENGLT) brought the movement up to 12 percent. The equation
follows:
Y = a + b1X1 + b2X2 = b^
PYCHIN + 171.749 43.864X1 = 39.128X0 s.813X3
The number of credit hours in psychology, number 20 (PSYCH) was
2
significant at the fourth step giving a multiple R of .146. The
predictive equation developed for the score in psychiatric internship
was as follows:
Y = a + b X = bX b.X- + b.X,
12 22 33 44
PYCHIN = 151.52 46.716X, 3.119X- + 38.49X0 + S.706X,
1 2 3 4
After this step the _F ratio was not sufficient for further stepping
(see Table 4). Null hypothesis II was rejected since it was statis
tically possible to predict scores in psychiatry internship on the basis
Table 4
Model IIVariables Entered by Step
Dependent
Variable
Independent
Variable
Variable
Number
Percent of
Movement
at Each Step
Overall
F
Psychiatry
TRANS
4
.054
9.50
Internship
#27
GPA-1
24
.095
8.74
ENCLT
18
.120
7.51
PSYCH
20
.146
7.04


TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS iii
LIST OF TABLES vi
LIST OF FIGURES vii
ABSTRACT viii
CHAPTER
I INTRODUCTION 1
General Background and Statement of Problem 1
General Purpose and Need for the Study 5
Need for Study 6
Assumptions 8
Hypotheses 8
Limitations of the Study 9
Definition of Terms 10
Significance of the Study 11
Organization of Subsequent Chapters 12
II REVIEW OF RELATED LITERATURE 14
Academic Performance, Persistence, or Withdrawal ... 14
Practical Application 19
Tests of Professional Knowledge 20
Professional Performance and Persistence 22
Concluding Remarks 24
Iv


51
Table 6
Model IIIVariables Entered by Step
4
Percent of
Dependent
Independent
Variable
Movement
Overall
Variable
Variable
Number at Each Step
F
Physical
Disabilities
TRANS
4
.032
5.44
Internship
#28
GPA-1
24
.077
6.95
The prediction
equation for
scores in physical
disabilities
internship
follows:
Y = a b X + b2X2
PDINT = 138.714 40.511X + 36.877X2
Again the amount of movement accounted for is very small, eight
percent. Therefore, the reader should exercise care in interpreting
the use of this model. Null hypothesis III was rejected since it was
statistically possible to predict scores in physical disabilities
internship on the basis of multiple combinations of objective variables
used in this study. Table 7 shows this model demonstrated by a random
sample of subjects' data.
Model IV--Work Status Score
This model computed only one step which accounted for only three
percent of the movement in work status scores (Table 8). The variable
of significance was CEEB, number 29. The first and final step showed
the following equation:
Y = a + bX
WKSTA = 5.352
. 003X


73
Katzell, ME: Expectations and dropouts in schools of nursing. J Appl
Psych 52(2):154-157, 1968.
Kim, JO and Kohout, FJ: Multiple regression analysis: Subprogram
regression. Univ of Iowa in SPSSStatistical Package for the
Social Sciences, Nie, Norman H, et al McGraw HillNY, 2nd ed,
1975, pp 320-367.
Knauer, FE: A study of the relationship of selected variables to
persistence of academically capable former students of the Univ
of SD, Unpub Doctoral Dissert, U of SD, 1968.
Knopf, L: From student to RN: A report of the nurse careerpattern
study. Pub Hlth Ser (DHEW) Wash DC, Div Nurs, 1972.
: Nurse CareerPattern study: Baccalaureate degree nurses ten
years after graduation. Hosp Tpcs 57(2):5-6, 1979.
Krall, V: Personality factors in nursing school success and failure.
Nurs Res 19(3):265-268, 1970.
Lapin, Lawrence L: Statistics for Modern Business Decisions, Harcourt
Brace Javanovich, Inc. NY, 1973.
Larken, PG: Admission test results as criteria for entrance into the
nursing program and regression analysis of entrance tests as pre
dictors of success in nursing state board examinations. Prince
George's Comm. Col, Largo, MD Office of Institutional Res, 1976.
Lee, JE and Shepard, KF: A comparison of biographical data of physical
therapy students in California. Phys Ther 52(3):272-281, 1973.
Lewis, J and Welch, M: Predicting achievement in an upper-division
bachelor's degree nursing major. Ed & Psych Meas 35:467-469, 1975.
Levitt, EE, Lubin, B, DeWitt, KN: An attempt to develop an objective
test battery for the selection of nursing students. Nurs Res 20(3):
255-261, 1971.
Liao, WC and Adams, JP: Methodology for the prediction of pharmacy
student academic success I: Preliminary aspects. J Col Stud Per
18(3):210-214, 1977.
Lind, AI: An exploratory study of predictive factors for success in the
clinical affiliation experience. Am J OT 24(3):222-226, 1970.
Lowenthal, W, Wergin, J and Smith, HL: Predictors of success in pharmacy
school: PCAT vs other admission criteria. Am J Pharm Ed 41(3):
267-269, 1977.


67
Other observations from this study include:
1. Transfer students from other colleges did not do as well as
students who began at the University of Florida as freshmen.
2. Classes in other health fields were more of a hindrance than
a help to success in the occupational therapy curriculum at
the University of Florida.
Recommendations
Based on the findings of this study and the review of the
literature, this researcher recommends the following:
1. Further research is needed in followup of students 15 to 25
years after graduation.
2. Further research is needed using all applicants to the specific
curriculum and compare those who were accepted and those who
were rejected in the following areas:
a. Different combinations of course categories should be
tried in future research
b. Biographical data should be utilized more in future
research
c. Use of measures of maturity should be considered in
future research
3. Future attention should be given to the following variables in
the development of a student selection procedure:
a. College grade point average for the first two years
b. Transfer student or entry as a University of Florida
freshman.
c.Number of credit hours in English and literature.


UNIVERSITY OF FLORIDA
3 1262 08555 2817


variables were used as probable factors explaining the movement of the
dependent variables.
The study was executed in two statistical parts. The first part
investigated the best pre-professional course work indicators for scores
on the following dependent variables:
1. Psychiatric Internship
2. Physical Disabilities Internship
3. National Registry Examination
4. Work Status Follow-up
Part one of the study was designed to compute a stepwise multiple
regressional analysis using the Biomedical Computer packaged Programs P
Series, BMDP2R.
The second part used stepwise discriminant analysis to find the
best indicators of completion and non-completion of the occupational
therapy program and the best indicators for work persistence and non
persistence in the field of occupational therapy using Biomedical
Computer packaged program, BMDP7M.
The multiple regression analysis of the 25 variables showed five
statistically significant at the .05 level: college entry test score,
college grade point average for the first two years, transfer student or
entry as a University of Florida freshman, number of credit hours in
English and literature, and number of credit hours in psychology. The
discriminant analysis used all but one of the same variables (CEEB) and
added one more: number of credit hours in music. In each case the
eight corresponding null hypotheses were rejected.
Although statistically significant variables were revealed, they
did not explain a high enough percentage of the movement in the


12
the individual and society. A faculty could benefit from being able to
select students who would continue to work in their field of training.
A persistent effort is needed in the departments of all health
related professions to eliminate any waste of time, money and resources
among students who plan to enter health field training or further their
present training. Through the 1970s, few longitudinal studies have
been done which followed subjects through college and into their work
experiences.
The writer recognizes the differences among various institutions
and curricula, but the individuals and conditions involved in the
Department of Occupational Therapy at the University of Florida may
represent the entire field. The University of Florida accommodates
some 30,000 students in all phases of university education. The De
partment of Occupational Therapy serves approximately 100 occupational
therapy undergraduate majors per year who are juniors, seniors or
interns. The first two years of the occupational therapy program are
in liberal arts. The four-and-one-half year undergraduate program leads
to the Bachelor of Health Science in occupational therapy. Should the
conclusions of this study be desirable for use in another health field
or another institution, modification would undoubtedly be needed to suit
the unique institutional and curricular characteristics.
Organization of Subsequent Chapters
Chapter II contains a review of the literature concerning comple
tion of training and job success in education, psychology, sociology
and allied health fields.


LIST OF TABLES
Table Page
1 Work Status Scale 35
2 Model IVariables Entered by Step 47
3 Model IScore Comparison for National Registry
Examination-Actual and Predicted 48
4 Model IIVariables Entered by Step 49
5 Model IIScore Comparison for Psychiatric
Internship-Actual and Predicted 50
6 Model IIIVariables Entered by Step 51
7 Model IIIScore Comparison for Physical Disabilities
Internship-Actual and Predicted 52
8 Model IVVariables Entered by Step 52
9 Model IVScore Comparison for Work Status
Actual and Predicted 53
10 Models V and VIClassification Functions 54
11 Models V and VIClassification Matrix 54
12 Models V and VIJacknifed Classification 55
13 Models V and VISummary-Variables Entered by Step 56
14 Model VScore Comparison for Complete-Actual and Predicted 56
15 Model VIScore Comparison for Did Not Complete-
Actual and Predicted 57
16 Models VII and VIIIClassification Functions 58
17 Models VII and VIIIClassification Matrix 58
18 Models VII and VIIIJacknifed Classification 58
19 Frequency of Independent Variable Use 64
vi


44
2. Working or not working in the field of occupational therapy.
The last section discusses the process model.
The Computer Programs
The statistical package used for computing the data in this study
was the Biomedical Computer Programs known as "BMD." The recent
revision of these programs was known as "BMDP" (Dixon and Brown 1979).
The BMDP program package, allowing the use of several dependent variables
and a large number of independent variables, offered the two programs
pertinent to this study: stepwise multiple regression "P2R" and stepwise
discriminant analysis "P7M." Both of these analyses used the forward
variable select procedure. This was described by Lapin (1973) as a
sequence of multiple linear regression equations in a stepwise manner.
At each step one variable is added to the regression equation. First,
the program selects the independent variable (X^) which accounts for the
most movement in the dependent variable Y. Then the next independent
variable is (X9) added to the model. The program continues in succes
sively higher dimensions until there is insufficient _F ratio to enter the
regression function. The _F ratio for significance of the regression
coefficient was set at 4.0. If the statistical criteria were not met,
the variable was not entered into the analysis.
Separate regression analysis was done to isolate from a set of 25
predictor variables a subset which would yield an optimal predictive
equation with the fewest possible terms to predict each of the four
dependent variables (1) National Registry Examination Score, (2)
Psychiatry Internship Score, (3) Physical Disabilities Internship Score,
and (4) Work Status Score.


69
Bean, AG and Covert, RW: Prediction of College Persistence Withdrawal
and Academic Dismissal: A Discriminant Analysis. Ed & Psych Meas
33:407-411, 1973.
Belcastro, FP: Use of Selected Factors as Predictors of Success in
Completing a Secondary Teacher Preparation Program. Ed & Psych
Meas 35:957-962, 1975.
Bergman, R, Edelstein, A, Rotenberg, A and Melamed, Y: Psychological
Tests: Their Use and Validity in Selecting Candidates for Schools
of Nursing in Israel. Int J Nurs St 11:85-109, 1974.
Bernstein, L: Medicare and the Occupational Therapist. Am J OT 22(5):
390-395, 1968.
Bistreich, AM: Predicting Grade Point Average, Withdrawal and Graduation
from Four Allied Health Programs at Miami-Dade Community College
Medical Center Campus. Doctoral Dissert, Nova Univ, 1977.
Blaisdell, EA Jr and Gordon, D: Selection of Occupational Therapy
Students. Am J OT 33(4):223-229, Apr 1979.
Blume, CS: Use of a Biographical Data Inventory to Identify Behavioral
Correlates of Practicing Medical Technologists: An Exploratory
Study. J Alld Hlth 5(1):54-61, W 1976.
Booth, MD: A Study of the Relationship between Certain Personality
Factors and Success in Clinical Training of Occupational Therapy
Students. Am J OT 11(2):93-96, 1957.
Bowers, J and Loeb, J: Changes in Predictor Weights in GPA Regression
Equations. Meas & Eval Guid 5(3):438-442, Oct 1972.
Brandt, EM: Comparison of On-the-Job Performance of Graduates with
School of Nursing Objectives. Nurs Res 16(l):50-60, 1967.
Brandt, EM, Hastie, B and Schuman, D: Predicting Success on State Board
Examinations. Nurs Res 15(1):62-69, 1966.
Brazziel, W: Non-intellective Predictors of Student Persistence/
Attrition for College and University Research and Planning. Paper
presented at the Int Forum of the Assoc for Insti Res, Montreal,
May 9, 1977.
Brollier, C: Personality Characteristics of Three Allied Health Profes
sional Groups. Am J OT 24(7):500-505, ]970.
Brown, PU: Career Development and Satisfaction of Occupational Therapists
in Florida. Doctoral Dissert, Univ of Fla 1973.
Brown, RGS and Stones, RWH: Personality and Intelligence Characteristics
of Male Nurses. Int J Nurs St 9:167-177, 1972.


65
amount of movement that could be accounted for was too small to be of
practical use.
The results of the study indicated that combining both academic
and practical application parameters in both the independent and the
dependent variibles does not enhance the practical application of a
predictive selection equation versus using just one of these parameters
in both. It was found that the same five variables were statistically
significant for both the academic and the practical application
dependent variables. Implications of the study were that a few
variables could be considered in the selection of potential candidates:
college grade-point average for the first two years, transfer student
or University of Florida freshman, number of credit hours in English
and literature, College Entrance Examination Board Test score, number
of credit hours in psychology, and number of credit hours in music.
However, human factors and subjective data should also be included.
In conclusion, this researcher found that a practical predictive
equation for use in selecting potential health curriculum students
cannot be developed with the objective data used in this study.
Observations
The dropout rate at the University of Florida undergraduate level
in occupational therapy was small, only 11 percent, and persistence of
graduates working in the field of occupational therapy was 75 percent
four to eight years after graduation. Johnson (1978) said that 75
percent of occupational therapists quit working in their first five
years of employment. However, Jantzen (1972) pointed out that


33
The 169 students studied were enrolled in the Occupational Therapy
curriculum from 1969 to 1973. For each student these follow-up records
were available:
1. University of Florida transcript
2. Internship scores, if the subjects completed the internships
3. National Registry Examination scores for the subjects who
took this examination
4. Work-entry statement from departmental follow-up newsletter
5. Follow-up information from the AOTA of those who continued
their registry
6. Telephone followup of those subjects who did not respond to
the AOTA surveys
7. Telephone followup of those subjects who did not continue
their registry in occupational therapy
8. Telephone followup of those subjects who dropped the occupa
tional therapy program before passing the registry examination.
From the 169 University of Florida transcripts examined, 24 inde
pendent variables occurred prior to entering professional course work.
These independent variables identified as possible factors explaining
the movement of the dependent variables are listed below:
1. Florida resident or out-of-state resident
2. University of Florida freshman or transfer
3. Grade in points in Introduction to the Health Related
Professions course
4. Gap in years from college freshman to upper division status
5. Number of hours in speech courses


Subject
Number
#2
WK/NWK
#26
REGIST
#27
PYCHIN
#28
PD INT
#30
COMP
1
0
172
223.0
228.0
1
2
1
200
229.5
267.0
1
3
1
175
226.5
288.0
1
A
0
195
26A.0
233.0
1
5
0
188
22A.0
216.5
1
6
1
163
277.5
2AA.5
1
7
1
186
262.0
256.0
1
8
1
175
297.5
298.0
1
9
1
170
265.5
253.0
1
10
1
213
269.5
262.0
1
11
1
207
275.5
281.5
1
12
1
190
275.5
221.5
1
13
1
181
2A3.5
239.0
1
1A
0
201
213.0
192.5
1
15
1
17A
2AA.5
221.5
1
16
1
187
273.5
279.5
1
17
1
177
2AA.5
296.0
1
18
1
201
236.5
2A1.0
1
19
1
206
270.0
229.0
1
20
1
205
269.0
2A9.5
1
21
1
17A
275.5
286.0
1
22
1
207
291.5
283.5
1
23
1
173
280.5
251.5
1
2A
0
188
276.0
258.0
1
25
1
187
272.5
199.0
1
26
1
19A
2A2.0
222.0
1
27
1
17A
228.0
23A.5
1
28
1
159
238.0
230.5
1
29
0
189
275.0
209.0
1
30
0
19A
29A.5
256.5
1
31
1
157
225.5
275.5
1
32
1
177
219.0
291.0
1
33
0
175
253.0
2AA.0
1
3A
0
179
281.5
26A.0
0
35
1
188
259.0
267.0
1
36
1
187
261.5
230.0
1
37
0
183
2A9.0
206.0
1
38
1
171
2A2.0
229.0
1
39
0
191
273.0
277.0
1
AO
1
16A
297.0
265.5
1
A1
1
182
281.5
266.0
1
A2
0
160
Mean 255.0
198.0
1
A3
1
162
2A5.0
1A1.0
1
AA
1
191
251.0
230.0
1
A5
1
158
232.5
295.0
1
#31
WK STA
08
10
10
08
08
10
10
10
10
10
10
10
10
09
10
10
10
10
10
10
10
09
10
08
10
09
10
10
08
08
09
10
08
08
10
10
08
09
09
10
10
07
10
10
10


CHAPTER II
REVIEW OF RELATED LITERATURE
The literature related to this study was reviewed within four
major categories. They were (1) academic survival, performance, and
persistence, (2) performance in practical application of training such
as internships, (3) professional knowledge as measured by written tests,
and (4) short-term professional performance, and long-term persistence.
The studies were all concerned with specific training curricula at the
junior college, college, and graduate levels of higher education,
psychology, sociology, and allied health.
Academic Performance, Persistence, or Withdrawal
The literature abounds with studies seeking predictors of college
performance, persistence, or withdrawal. Some notable ones by Aiken
(1968), A'Hearn (1974), Bowers and Loeb (1972), Green (1962), and Hinkle
and Houston (1977) all found that the best predictors of future academic
achievement were previous achievement and intellectual characteristics,
meaning: high school rank, American College Testing (ACT) and Scholastic
Aptitude Testing (SAT) test scores, reading test scores, and Adjective
Check List scores. Their focus was toward first semester or first
quarter college grades.
In studying persistence and withdrawal of students at the end of the
freshman year of college Humphreys et al. (1973), Knopf (1972), Knauer


The "Reports on Curriculum Study" issued in 1963, which reported
on the 1958 to 1963 American Occupational Therapy Association study,
stated that student selection is considered one of the eight major areas
to include in future planning of curricula. Several studies have been
conducted on student selection for occupational therapy including those
by Bailey et al. (1969), Booth (1957), Crane (1962), Engelhart (1957),
Hendrickson (1962), and Lind (1970). However, results have not
been uniformly positive (Lind 1970). Negative correlations have
resulted within a few. Research on student selection procedures in
occupational therapy has been relatively meager and limited in scope.
Most of the studies on this topic have been based on small samples
ranging from 18 to 37 students (Johnson RW et al. 1974). Sample sizes
of at least 100 and preferably 200 students should be obtained from
successive classes at the same institution or by the use of comparative
data from several institutions (Johnson RW et al. 1974). Pedrini and
Pedrini (1974) concluded that there is a need for multivariate research
after an extensive review of the literature related to college success.
Prendergast (1978) concurred with this need in occupational therapy.
The ultimate impact for change in selection processes will depend
on the results of longitudinal research studies which will provide data
on the success of applicants in the field (Bistreich 1977). Johnson JA
(1974) pinpointed that the education of clinicians who can move with the
trends and continue in the field of occupational therapy should be a
major emphasis.
In 1977 Morgan stated the need to identify characteristics of
effective practitioners:


25
achievement were the best predictors of later academic achievement and
of higher scores on written tests of professional knowledge. Many
studies required the administration of additional tests and question
naires. Occasionally scores on psychological and interest measures were
significant. None of the predictive variables were considered by the
respective researcher to be adequate as a single variable when selecting
students for a specific curriculum. Only one study used both academic
and practical application data in both independent and dependent
variables; however, that study did not follow up the same population for
entry and persistence in the specific field. Relatively few studies
dealt with long-term follow-up.
A study of the literature revealed no study combining all of the
following:
1. practical application data in the dependent and independent
variables.
2. academic data in the dependent and independent variables.
3. a sample size larger than 50.
4. data on subjects for more than three successive years.
5. followup on the same subjects more than one year after
graduation.
6. followup on the subjects who withdrew from the undergraduate
curriculum.
7. use of only objective data from transcripts.


34
6. Number of hours in business courses
7. Number of hours in nursing courses
8. Number of hours in library science courses
9. Grade in points in growth and development courses
10. Number of hours in special education courses
11. Number of hours in drama and theater courses
12. Number of hours in health courses
13. Number of hours in social sciences courses
14. number of hours in physical sciences courses
15. Number of hours in biological sciences courses
16. Number of hours in English and literature courses
17. Number of hours in mathematics courses
18. Number of hours in psychology courses
19. Number of hours in foreign language courses
20. Number of hours in arts and crafts courses
21. Number of hours in music
22. Grade point average for the first two years
23. Florida 12th Grade Placement Test total score
24. College Entrance Examination Board total test score
The dependent variables established for this study were:
1. American Occupational Therapy Association National Registry
Examination score
2. Psychiatric Internship score
3. Physical Disabilities Internship score
4. Work status at the time of this study
5. Completion of non-completion of the occupational therapy
program


57
Table 15
Model VIScore Comparison for Did Not Complete-
Actual and Predicted
Subject Number
Actual
TRANS
PSYCH
GPA-1
ENGLT
Predicted
007
1
1
16
2.7
08
15.636
051
1
0
10
1.97
16.6
10.352
111
1
1
14
1.44
08
2.093
149
1
0
09
2.61
09
15.023
155
0
0
04
2.09
28
11.878
165
0
0
9.5
3.0
32.1
23.841
Models VII and VIII-
-Working
or Not Working in the
Field of Occupational Therapy
A prediction was developed for the dichotomy of working or not
working in the field of occupational therapy four to eight years after
graduation. Table 16 shows the classification functions. This
discriminant function correctly classified 42.9 percent of the subjects
into the working group and 57.1 percent were incorrectly classified.
Correct classification into the non-working group was 71.7 percent with
an error of 28.3 percent. The total average for correct classification
of the subjects was 62.1 percent. This is again low since the correct
prediction by chance would be 71.7 percent, as shown in Table 17, the
classification matrix. The "jacknifed procedure" indicates a concurrence
with these findings (see Table 18).


39
8. Multiple regression analysisis an expansion of the same techniques
used in simple regression. Multiple regression uses several
independent variables as predictors. The movement or variation
of Y can then be explained by two or more variables. This
formula is:
Yc = a + b, X, = b2 X2 + . + bk Xk
This multi-dimensional analysis can be shown by the following
figure:
Figure 2. Multiple Regression (p. 500)
An analogy would be, using the walls and floor of a room. The
corner represents the point where all three variables have a
value of zero. The observation points are represented by marbles
suspended in space at various distances from the floor and the
two walls. The value of Y is represented by the marble's height
above the floor. The observed value for X^ is represented by


Page
III DESIGN AND METHODOLOGY 26
Population 26
Data Collection 29
Method of Analysis 32
Definitions of Statistical Terms 36
Pitfalls of Regression and Correlation Techniques . 41
Organization of Data 42
Summary 42
IV ANALYSIS OF DATA AND RESULTS 43
Introduction 43
The Computer Programs 44
Stepwise Multiple Regression Analysis 46
Model INational Registry Examination Score
(REGIST) 46
Model IIPsychiatric Internship Score (PYCHIN) . 48
Model IIIPhysical Disabilities Internship
Score (PDINT) 50
Model IVWork Status Score 51
Stepwise Discriminant Analysis 53
Model VComplete and Not Complete the Program
for Registry in Occupational Therapy 53
Models VII and VIIIWorking or Not Working in
the Field of Occupational Therapy 57
V CONCLUSIONS, OBSERVATIONS AND RECOMMENDATIONS 60
Conclusions 60
Observations 65
Recommendations 67
BIBLIOGRAPHY 68
APPENDIX 78
BIOGRAPHICAL SKETCH 84
v


53
Table 9
Model IVScore Comparison for Work Status
Actual and Predicted
Subj ect
Number
Actual Work
Status
CEEB
Predicted Score
for Work Status
007
10
860
7.932
051
10
1048
8.496
111
10
1006
8.370
149
10
1100
8.652
155
3
1020
8.412
165
2
820
7.812
Stepwise Discrimination Analysis
Model V
-Complete and Not Complete
the Program for
Registry
in Occupational Therapy
The analysis of the data developed a prediction for the dichotomy
of completing or not completing the program for registration in
occupational therapy. As Shalik (1978) synthesized:
The method used for this analysis is the Wilk's lambda. The
criterion was the overall multivariate _F ratio for the test
of differences among group centroids. The variable which
maximized the _F ratio also minimized Wilk's lambda, a measure
of group discrimination. Group assignment was based on the
distance from the centroid with the actual group assignment
determined by the closest distance to one of the centroids.
(p. 48)
Table 10 shows the classification functions.
This discriminant function correctly classified 66.7 percent of the
subjects into the completion group and 33.3 percent were incorrectly
classified. Correct classification into the non-completion group was


28
a. A student information form for the College of Health
Related Professions. This was a traditional form similar
to a curriculum vitae, including
1) extracurricular activities
2) honors and distinctions
3) employment
4) essay as to why the student was interested in a health
career.
b. Copies of all transcripts must have been received in
Occupational Therapy Department via the Registrar's Office.
c. Student must have completed the form listing student
organizations, activities, and publications.
Faculty evaluation commenced after the above steps were completed by the
applicant. The selection process considered the following:
1. Grades were important; the faculty scanned credentials of all
candidates possessing a 2.0 GPA or better.
2. Pre-professional grades must be 2.0 or better.
3. The essay was read and the following factors were evaluated:
a. The student's motivation.
b. Why the student was interested in occupational therapy.
c. How well the student wrote.
4. Work and volunteer experience were considered.
5. Two interviews were conducted;
a. One was as an advisement discussion.
b. The second was more formal and was randomly assigned
to a different member of the faculty.


37
2. Correlation coefficientshows how closely two variables move to
gether by means of a single number calculated from the same data.
3. Coefficient of determinationexpresses the relative reduction in
the variation of Y that can be attributed to knowledge of X and
its relationship to Y by way of the regression line or the pro
portion of the total variation in Y explained by the regression
line. Expresses how strongly X is associated with Y. Represented
by r^ or (Lapin 1973).
4. Dependent variablethe variable whose value is being predicted,
represented by the y axis (Lapin 1973) .
5. Independent variablethe variable whose value is known, and
represented by the X axis (Lapin 1973).
6. Estimated regression lineexpresses the average relationship
between the X and Y variables. This provides an estimate of the
mean level of the dependent variable Y when the value of X is
specified. Because the estimated regression line provides
estimates only, the symbol is Y^ (Lapin 1973).
7. Simple linear regression analysisis used to obtain predictions
of one variable using known values of another variable. The
regression equation looks like this:
Y = a + bX
Sample observations are taken and used to predict another
variable. To avoid a large sampling error, a sample size
considerably greater than ten should be used (Lapin 1973) .



PAGE 1

)$&7256 $))(&7,1* 3(56,67(1&( ,1 ('8&$7,21 $1' (03/2<0(17 2) 2&&83$7,21$/ 7+(5$3,676 %\ &$52/<1 67(:$57 $ ',66(57$7,21 35(6(17(' 72 7+( *5$'8$7( &281&,/ 2) 7+( 81,9(56,7< 2) )/25,'$ ,1 3$57,$/ )8/),//0(17 2) 7+( 5(48,5(0(176 )25 7+( '(*5(( 2) '2&725 2) 3+,/2623+< 81,9(56,7< 2) )/25,'$

PAGE 2

'(',&$7(' 72 $/,&( & -$17=(1 $Q HGXFDWRU ZKR KDV EURXJKW WKH IXWXUH WR WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ $Q LQVSLUDWLRQ WR WKRVH ZKR DUH IRUWXQDWH HQRXJK WR EH WRXFKHG E\ KHU

PAGE 3

$&.12:/('*0(176 7KH UHVHDUFKHU ZLVKHV WR H[SUHVV VLQFHUH DSSUHFLDWLRQ WR WKRVH SHUVRQV ZKR KDG WKH SDWLHQFH DQG SHUVLVWHQFH LQ KHOSLQJ WR PDNH WKLV VWXG\ SRVVLEOH 6LQFH PRVW RI WKH SUHSDUDWLRQ ZDV GRQH E\ FRUUHVSRQn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

PAGE 4

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

PAGE 5

3DJH ,,, '(6,*1 $1' 0(7+2'2/2*< 3RSXODWLRQ 'DWD &ROOHFWLRQ 0HWKRG RI $QDO\VLV 'HILQLWLRQV RI 6WDWLVWLFDO 7HUPV 3LWIDOOV RI 5HJUHVVLRQ DQG &RUUHODWLRQ 7HFKQLTXHV 2UJDQL]DWLRQ RI 'DWD 6XPPDU\ ,9 $1$/<6,6 2) '$7$ $1' 5(68/76 ,QWURGXFWLRQ 7KH &RPSXWHU 3URJUDPV 6WHSZLVH 0XOWLSOH 5HJUHVVLRQ $QDO\VLV 0RGHO ,f§1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ 6FRUH 5(*,67f 0RGHO ,,f§3V\FKLDWULF ,QWHUQVKLS 6FRUH 3<&+,1f 0RGHO ,,,f§3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLS 6FRUH 3',17f 0RGHO ,9f§:RUN 6WDWXV 6FRUH 6WHSZLVH 'LVFULPLQDQW $QDO\VLV 0RGHO 9f§&RPSOHWH DQG 1RW &RPSOHWH WKH 3URJUDP IRU 5HJLVWU\ LQ 2FFXSDWLRQDO 7KHUDS\ 0RGHOV 9,, DQG 9,,,f§:RUNLQJ RU 1RW :RUNLQJ LQ WKH )LHOG RI 2FFXSDWLRQDO 7KHUDS\ 9 &21&/86,216 2%6(59$7,216 $1' 5(&200(1'$7,216 &RQFOXVLRQV 2EVHUYDWLRQV 5HFRPPHQGDWLRQV %,%/,2*5$3+< $33(1',; %,2*5$3+,&$/ 6.(7&+ Y

PAGE 6

/,67 2) 7$%/(6 7DEOH 3DJH :RUN 6WDWXV 6FDOH 0RGHO ,f§9DULDEOHV (QWHUHG E\ 6WHS 0RGHO ,f§6FRUH &RPSDULVRQ IRU 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ$FWXDO DQG 3UHGLFWHG 0RGHO ,,f§9DULDEOHV (QWHUHG E\ 6WHS 0RGHO ,,f§6FRUH &RPSDULVRQ IRU 3V\FKLDWULF ,QWHUQVKLS$FWXDO DQG 3UHGLFWHG 0RGHO ,,,f§9DULDEOHV (QWHUHG E\ 6WHS 0RGHO ,,,f§6FRUH &RPSDULVRQ IRU 3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLS$FWXDO DQG 3UHGLFWHG 0RGHO ,9f§9DULDEOHV (QWHUHG E\ 6WHS 0RGHO ,9f§6FRUH &RPSDULVRQ IRU :RUN 6WDWXV $FWXDO DQG 3UHGLFWHG 0RGHOV 9 DQG 9,f§&ODVVLILFDWLRQ )XQFWLRQV 0RGHOV 9 DQG 9,f§&ODVVLILFDWLRQ 0DWUL[ 0RGHOV 9 DQG 9,f§-DFNQLIHG &ODVVLILFDWLRQ 0RGHOV 9 DQG 9,f§6XPPDU\9DULDEOHV (QWHUHG E\ 6WHS 0RGHO 9f§6FRUH &RPSDULVRQ IRU &RPSOHWH$FWXDO DQG 3UHGLFWHG 0RGHO 9,f§6FRUH &RPSDULVRQ IRU 'LG 1RW &RPSOHWH $FWXDO DQG 3UHGLFWHG 0RGHOV 9,, DQG 9,,,f§&ODVVLILFDWLRQ )XQFWLRQV 0RGHOV 9,, DQG 9,,,f§&ODVVLILFDWLRQ 0DWUL[ 0RGHOV 9,, DQG 9,,,f§-DFNQLIHG &ODVVLILFDWLRQ )UHTXHQF\ RI ,QGHSHQGHQW 9DULDEOH 8VH YL

PAGE 7

/,67 2) ),*85(6 )LJXUH 3DJH 6LPSOH /LQHDU 5HJUHVVLRQ 0XOWLSOH 5HJUHVVLRQ 'LVFULPLQDQW )XQFWLRQ YLL

PAGE 8

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n WLRQDO WKHUDS\ FXUULFXOXP LQWHUQVKLS UHJLVWU\ H[DPLQDWLRQ DQG MRE SHUVLVWHQFH SULRU WR WKH WLPH SURIHVVLRQDO FRXUVH ZRUN LV EHJXQ IRU XVH LQ VWXGHQW VHOHFWLRQ 8OWLPDWHO\ WKH GHYHORSPHQW RI D SURFHVV PRGHO ZKLFK FRXOG EH XVHG ZLWK IXWXUH GDWD IRU VHOHFWLRQ RI FDQGLGDWHV ZDV D GHVLUDEOH RXWFRPH 7KH VXEMHFWV LQ WKLV VWXG\ DOO KDG EHHQ DFFHSWHG LQWR WKH RFFXSDWLRQDO WKHUDS\ SURJUDP IRU JUDGXDWLQJ FODVVHV VSDQQLQJ WKH ILYH \HDU SHULRG IURP WR 2I WKH ZRPHQ DQG QLQH PHQ ZLWKn GUHZ IURP WKH XQGHUJUDGXDWH FXUULFXOXP 7KH VHOHFWHG YDULDEOHV RFFXUULQJ SULRU WR SURIHVVLRQDO FRXUVH ZRUN ZHUH WDNHQ IURP 8QLYHUVLW\ RI )ORULGD WUDQVFULSWV ZKLFK ZHUH DYDLODEOH IRU DOO VXEMHFWV 7KHVH YLLL

PAGE 9

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n SHUVLVWHQFH LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ XVLQJ %LRPHGLFDO &RPSXWHU SDFNDJHG SURJUDP %0'30 7KH PXOWLSOH UHJUHVVLRQ DQDO\VLV RI WKH YDULDEOHV VKRZHG ILYH VWDWLVWLFDOO\ VLJQLILFDQW DW WKH OHYHO FROOHJH HQWU\ WHVW VFRUH FROOHJH JUDGH SRLQW DYHUDJH IRU WKH ILUVW WZR \HDUV WUDQVIHU VWXGHQW RU HQWU\ DV D 8QLYHUVLW\ RI )ORULGD IUHVKPDQ QXPEHU RI FUHGLW KRXUV LQ (QJOLVK DQG OLWHUDWXUH DQG QXPEHU RI FUHGLW KRXUV LQ SV\FKRORJ\ 7KH GLVFULPLQDQW DQDO\VLV XVHG DOO EXW RQH RI WKH VDPH YDULDEOHV &((%f DQG DGGHG RQH PRUH QXPEHU RI FUHGLW KRXUV LQ PXVLF ,Q HDFK FDVH WKH HLJKW FRUUHVSRQGLQJ QXOO K\SRWKHVHV ZHUH UHMHFWHG $OWKRXJK VWDWLVWLFDOO\ VLJQLILFDQW YDULDEOHV ZHUH UHYHDOHG WKH\ GLG QRW H[SODLQ D KLJK HQRXJK SHUFHQWDJH RI WKH PRYHPHQW LQ WKH

PAGE 10

GHSHQGHQW YDULDEOHV WR EH XVHIXO IRU SUDFWLFDO DSSOLFDWLRQ 7KLV ZDV GHPRQVWUDWHG IRU HDFK GHSHQGHQW YDULDEOH E\ D UDQGRP VDPSOH RI VXEMHFWVn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

PAGE 11

&+$37(5 ,1752'8&7,21 *HQHUDO %DFNJURXQG DQG 6WDWHPHQW RI 3UREOHP $FFRXQWDELOLW\ ZDV D PRYHPHQW RI WKH V 7KH XOWLPDWH DFFRXQWn DELOLW\ RI DQ H[SHQVLYH WHFKQLFDO HGXFDWLRQ PD\ EH PHDVXUHG E\ VWDWLVWLFV VKRZLQJ D KLJK SHUFHQWDJH RI FRPSOHWLRQ RI WKH SURJUDP DQG SHUVLVWHQFH ZLWKLQ WKH ILHOG 7KH FRQFHSWV RI SUHGLFWLQJ SHUVLVWHQFH DQG JUDGXDWLRQ IURP FROOHJH KDYH EHHQ WKH VXEVWDQFH RI PDQ\ VWXGLHV HVSHFLDOO\ LQ WKH ODVW WHQ \HDUV 6LQFH D SHU FHQW DWWULWLRQ UDWH RI FROOHJH VWXGHQWV SULRU WR JUDGXDWLRQ UHSUHVHQWV D JUHDW ORVV RI WLPH PRQH\ DQG HQHUJ\ WR DOO FRQFHUQHG 0RUJDQ f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

PAGE 12

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n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

PAGE 13

SURJUDPV $Q HUURU LQ MXGJPHQW LQ VHOHFWLQJ D VWXGHQW ZKR ZLOO ODWHU GURS RXW RI D SURJUDP RU ILHOG RI RFFXSDWLRQDO WKHUDS\ UHSUHVHQWV D SHUVRQDO DQG SURIHVVLRQDO ORVV RI WZR SHRSOH WKH RQH ZKR GURSSHG RXW DQG WKH RQH ZKR PLVVHG WKH RULJLQDO RSSRUWXQLW\ WR WDNH WKDW RSHQLQJ 0RUJDQ f +RZ FDQ WKH EHVW VWXGHQWV IRU WKH KHDOWKUHODWHG SURIHVVLRQV EH VHOHFWHG" :KDW IDFWRUV VXFK DV DELOLWLHV SDVW H[SHULHQFHV SHUVRQn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f VWDWHG WKDW KHDOWK FDUH ZDV WKH VHFRQG ODUJHVW LQGXVWU\ LQ WKH 8QLWHG 6WDWHV 7KLV GHPDQG IRU FDUH SUHFLSLWDWHG D GUDLQ RQ WKH KHDOWK V\VWHP DQG FUHDWHG D QHHG IRU DGGLWLRQDO DOOLHG KHDOWK PDQSRZHU WR SURYLGH VXSSRUW VHUYLFHV IRU SK\VLFLDQV DQG GHQWLVWV

PAGE 14

LQ WKH KHDOWK FDUH GHOLYHU\ V\VWHP 'X9DO UHSRUWHG VRPH JUDGXDWHV RI DOOLHG KHDOWK HGXFDWLRQDO SURJUDPV LQ FRPSDUHG ZLWK DERXW LQ 7KH GHYHORSPHQW RI KHDOWK PDQSRZHU KDV PHDQLQJ QRW RQO\ WR VRFLHW\ EXW WR HGXFDWLRQ DQG WKH LQGLYLGXDO %URZQ f &ROOHJH VWXGHQWV KDYH DOVR VKRZQ DQ LQFUHDVHG LQWHUHVW LQ WKH KHDOWK FDUH ILHOGV $Q LQWHUHVW VXUYH\ E\ WKH $PHULFDQ &RXQFLO RI (GXFDWLRQ LQ WKH IDOO RI LQGLFDWHG IXOOWLPH IUHVKPHQ VKRZHG SHUFHQW PRUH LQWHUHVW LQ KHDOWK FDUH ILHOGV WKDQ WKRVH RI WKH SUHYLRXV \HDU /LQG f 7KH $PHULFDQ &ROOHJH 7HVWLQJ $&7f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n DELOLW\ VXFK DV KRZ PDQ\ WUDLQHHV DFWXDOO\ JR WR ZRUN LQ WKHLU ILHOG +RZ PDQ\ ZKR HQWHU WKHLU ILHOG FRQWLQXHG WR ZRUN LQ WKDW ILHOG" 7KHUHIRUH SUHGLFWLYH VWXGLHV VKRXOG LQFOXGH MRE HQWU\ DQG ORQJ WHUP SHUVLVWHQFH LQ WKH FKRVHQ KHDOWK ILHOG DV SDUW RI VXFFHVVIXO VHOHFWLRQ SURFHVVHV

PAGE 15

7KLV VWXG\ H[SORUHG SRVVLEOH FRPELQDWLRQV RI YDULDEOHV VXFK DV XQGHUJUDGXDWH FRXUVHV FROOHJH HQWUDQFH H[DPLQDWLRQ VFRUHV DQG RWKHU WUDQVFULSW GDWD DV LQGLFDWRUV RI VWXGHQWV ZKR FRPSOHWHG DQG ZRUNHG LQ RQH KHDOWK FDUH ILHOGf§RFFXSDWLRQDO WKHUDS\ $OO RI WKH VWXGHQWV LQ WKLV VWXG\ JUDGXDWHG IURP WKH 8QLYHUVLW\ RI )ORULGD *HQHUDO 3XUSRVH DQG 1HHG IRU WKH 6WXG\ $ QHHG H[LVWV WR REMHFWLYHO\ LGHQWLI\ SURVSHFWLYH FROOHJH VWXGHQWV ZKR ZRXOG EH DEOH WR FRPSOHWH WKH UHTXLUHPHQWV IRU UHJLVWUDWLRQ LQ RFFXSDWLRQDO WKHUDS\ HQWHU WKH MRE PDUNHW DQG FRQWLQXH ZLWKLQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ 7KH SXUSRVH RI WKH SUHVHQW VWXG\ ZDV WR H[SORUH LQGHSHQGHQW YDULDEOHV VXFK DV FROOHJH HQWU\ WHVW VFRUHV FRXUVHV WDNHQ DQG JUDGH SRLQW DYHUDJHV WR GHWHUPLQH ZKLFK LI DQ\ FRXOG EH XVHG SULRU WR WKH EHJLQQLQJ RI SURIHVVLRQDO FRXUVH ZRUN DV LQGLFDWRUV RI WKRVH ZKR ZRXOG EH PRVW OLNHO\ WR JUDGXDWH WKHQ EH HPSOR\HG DQG FRQWLQXH ZRUNLQJ LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ 3UHYLRXV VWXGLHV VRXJKW SUHGLFWRUV RI f DFDGHPLF SHUIRUPDQFH SHUVLVWHQFH RU ZLWKGUDZDO f SHUIRUPDQFH LQ SUDFWLFDO DSSOLFDWLRQ RI WUDLQLQJ VXFK DV LQWHUQVKLSV f SURIHVVLRQDO NQRZOHGJH DV PHDVXUHG E\ ZULWWHQ WHVWV f VKRUW WHUP SURIHVVLRQDO SHUIRUPDQFH DQG ORQJ WHUP SHUVLVWHQFH 1RQH RI WKHVH VWXGLHV LQFOXGHG DOO IRXU YDULDEOHV RQ WKH VDPH SRSXODWLRQ 7KLV VWXG\ VRXJKW WR JHQHUDWH PHDQLQJIXO UHVXOWV E\ XVLQJ YDULDEOHV IURP HDFK FDWHJRU\ DV WKH\ UHODWHG WR D ODUJHU QXPEHU RI VXEMHFWV Q f 0RUH VSHFLILFDOO\ WKH SXUSRVHV ZHUH WR ,GHQWLI\ WKH EHVW SUHSURIHVVLRQDO LQGLFDWRUV RI SHUVLVWHQFH LQ ZRUNLQJ LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\

PAGE 16

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n VLGHUHG DV HOHPHQWV RI SURJQRVLV RI VXFFHVV LQ FROOHJH DQG IRU JUDGXDn WLRQ IURP FROOHJH 0HWKRGV HPSOR\HG KDYH LQFOXGHG H[WHQVLYH XVH RI ZULWWHQ DSSOLFDWLRQV UHFRPPHQGDWLRQV DQG SHUVRQDO LQWHUYLHZV FRQVXPLQJ JUHDW DPRXQWV RI IDFXOW\ WLPH &ULWHULD IRU VHOHFWLRQ KDYH HPSKDVL]HG DFDGHPLF DFKLHYHPHQW SDUWLFLSDWLRQ LQ H[WUDFXUULFXODU DFWLYLWLHV DQG SHUVRQDO TXDOLWLHV

PAGE 17

7KH 5HSRUWV RQ &XUULFXOXP 6WXG\ LVVXHG LQ ZKLFK UHSRUWHG RQ WKH WR $PHULFDQ 2FFXSDWLRQDO 7KHUDS\ $VVRFLDWLRQ VWXG\ VWDWHG WKDW VWXGHQW VHOHFWLRQ LV FRQVLGHUHG RQH RI WKH HLJKW PDMRU DUHDV WR LQFOXGH LQ IXWXUH SODQQLQJ RI FXUULFXOD 6HYHUDO VWXGLHV KDYH EHHQ FRQGXFWHG RQ VWXGHQW VHOHFWLRQ IRU RFFXSDWLRQDO WKHUDS\ LQFOXGLQJ WKRVH E\ %DLOH\ HW DO f %RRWK f &UDQH f (QJHOKDUW f +HQGULFNVRQ f DQG /LQG f +RZHYHU UHVXOWV KDYH QRW EHHQ XQLIRUPO\ SRVLWLYH /LQG f 1HJDWLYH FRUUHODWLRQV KDYH UHVXOWHG ZLWKLQ D IHZ 5HVHDUFK RQ VWXGHQW VHOHFWLRQ SURFHGXUHV LQ RFFXSDWLRQDO WKHUDS\ KDV EHHQ UHODWLYHO\ PHDJHU DQG OLPLWHG LQ VFRSH 0RVW RI WKH VWXGLHV RQ WKLV WRSLF KDYH EHHQ EDVHG RQ VPDOO VDPSOHV UDQJLQJ IURP WR VWXGHQWV -RKQVRQ 5: HW DO f 6DPSOH VL]HV RI DW OHDVW DQG SUHIHUDEO\ VWXGHQWV VKRXOG EH REWDLQHG IURP VXFFHVVLYH FODVVHV DW WKH VDPH LQVWLWXWLRQ RU E\ WKH XVH RI FRPSDUDWLYH GDWD IURP VHYHUDO LQVWLWXWLRQV -RKQVRQ 5: HW DO f 3HGULQL DQG 3HGULQL f FRQFOXGHG WKDW WKHUH LV D QHHG IRU PXOWLYDULDWH UHVHDUFK DIWHU DQ H[WHQVLYH UHYLHZ RI WKH OLWHUDWXUH UHODWHG WR FROOHJH VXFFHVV 3UHQGHUJDVW f FRQFXUUHG ZLWK WKLV QHHG LQ RFFXSDWLRQDO WKHUDS\ 7KH XOWLPDWH LPSDFW IRU FKDQJH LQ VHOHFWLRQ SURFHVVHV ZLOO GHSHQG RQ WKH UHVXOWV RI ORQJLWXGLQDO UHVHDUFK VWXGLHV ZKLFK ZLOO SURYLGH GDWD RQ WKH VXFFHVV RI DSSOLFDQWV LQ WKH ILHOG %LVWUHLFK f -RKQVRQ -$ f SLQSRLQWHG WKDW WKH HGXFDWLRQ RI FOLQLFLDQV ZKR FDQ PRYH ZLWK WKH WUHQGV DQG FRQWLQXH LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ VKRXOG EH D PDMRU HPSKDVLV ,Q 0RUJDQ VWDWHG WKH QHHG WR LGHQWLI\ FKDUDFWHULVWLFV RI HIIHFWLYH SUDFWLWLRQHUV

PAGE 18

7KHUH LV D QHHG WR GHWHUPLQH LI FHUWDLQ FKDUDFWHULVWLFV GLVFULPLQDWH EHWZHHQ WKRVH ZKR UHPDLQ LQ D ILHOG DQG WKRVH ZKR GURS RXW S f 7KH SUHVHQW VWXG\ ZDV QHHGHG WR GHWHUPLQH LI REMHFWLYH IDFWRUV REWDLQHG IURP UHFRUGV DYDLODEOH DW WKH WLPH RI VHOHFWLRQ FRXOG EH XVHG DV FULWHULD WR SUHGLFW WKH FRPSOHWLRQ RI WUDLQLQJ HQWU\ LQWR DQG SHUVLVWHQFH LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ $VVXPSWLRQV ,W ZDV DVVXPHG WKDW WKH WUDQVFULSWV RI WKH VWXGHQWV LQYROYHG LQ WKLV VWXG\ DW WKH 8QLYHUVLW\ RI )ORULGD ZHUH DFFXUDWH ,W ZDV DVVXPHG WKDW VFRUHV DQG JUDGHV ZKLFK DSSHDUHG LQ WKH VWXGHQWn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f WKH 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ ,,f 3V\FKLDWU\ LQWHUQVKLS

PAGE 19

,,,f3K\VLFDO GLVDELOLWLHV LQWHUQVKLS ,9f:RUN VWDWXV RQ WKH EDVLV RI PXOWLSOH FRPELQDWLRQV RI YDULDEOHV ZKLFK LQFOXGH FROOHJH HQWU\ WHVW VFRUHV FRXUVHV WDNHQ JUDGH SRLQW DYHUDJHV DQG RWKHU REMHFWLYH GDWD WDNHQ IURP WUDQVFULSWV )RXU JHQHUDOL]HG PXOWLYDULDWH QXOO K\SRWKHVHV ZHUH SURSRVHG IRU GLVFULPLQDQW DQDO\VLV ,W ZLOO QRW EH SRVVLEOH WR GLVFULPLQDWH DPRQJ JURXSV RI VWXGHQWV 9fWKRVH ZKR FRPSOHWHG DQG 9,f WKRVH ZKR GLG QRW FRPSOHWH WKH RFFXSDn WLRQDO WKHUDS\ SURJUDP RQ WKH EDVLV RI OLQHDU FRPELQDWLRQV RI YDULDEOHV ZKLFK LQFOXGH FROOHJH HQWU\ WHVW VFRUHV FRXUVHV WDNHQ JUDGH SRLQW DYHUDJHV DQG RWKHU REMHFWLYH GDWD WDNHQ IURP WUDQVFULSWV ,W ZLOO QRW EH SRVVLEOH WR GLVFULPLQDWH DPRQJ JURXSV RI VWXGHQWV 9,,f WKRVH ZKR ZHUH ZRUNLQJ DQG 9,,,f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

PAGE 20

7KLV VWXG\ ZDV FRQFHUQHG ZLWK VWXGHQWVn UHFRUGV LQFOXGLQJ PHQ DQG ZRPHQ 7KH VWXGHQWV UHSUHVHQWHG YDULRXV HQYLURQPHQWDO DQG HGXFDWLRQDO EDFNJURXQGV +RZHYHU WKH RQO\ PLQRULW\ JURXS UHSUHVHQWHG ZDV RI 6SDQLVK VXUQDPHV Q f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f§$PHULFDQ &ROOHJH 7HVW IRU VFKRODVWLF DSWLWXGH $27$f§$PHULFDQ 2FFXSDWLRQDO 7KHUDS\ $VVRFLDWLRQ &((%f§&ROOHJH (QWUDQFH ([DPLQDWLRQ %RDUG 7KH VSRQVRULQJ ERG\ IRU WKH 6FKRODVWLF $SWLWXGH 7HVW 6$7f &RPSOHWH WKH RFFXSDWLRQDO WKHUDS\ SURJUDPf§WR FRPSOHWH DOO XQGHU

PAGE 21

JUDGXDWH WUDLQLQJ SOXV LQWHUQVKLSV DQG WKH 1DWLRQDO 5HJLVWU\ H[DPLQDWLRQ WR EH HOLJLEOH IRU 5HJLVWU\ LQ $27$ *3$f§*UDGH 3RLQW $YHUDJH 7KH DYHUDJH LV FRPSXWHG E\ GLYLGLQJ WKH WRWDO QXPEHU RI JUDGH SRLQWV E\ WKH WRWDO QXPEHU RI FRXUVH FUHGLW KRXUV DWWHPSWHG 1RQSHUVLVWHQFH LQ ZRUNLQJ LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\f§ QRW ZRUNLQJ LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ IRXU WR HLJKW \HDUV DIWHU JUDGXDWLRQ 3HUVLVWHQFH LQ ZRUNLQJ LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\f§ZRUNLQJ LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ IRXU WR HLJKW \HDUV DIWHU JUDGXDWLRQ 3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLSf§D WKUHHPRQWK VXSHUYLVHG RQWKH MRE WUDLQLQJ DQG OHDUQLQJ VLWXDWLRQ DSSURYHG E\ $27$ 7KLV LQWHUQn VKLS LV UHTXLUHG DV SDUWLDO IXOILOOPHQW IRU FRPSOHWLRQ RI WKH RFFXSDWLRQDO WKHUDS\ SURJUDP SULRU WR WKH 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ 3V\FKLDWULF ,QWHUQVKLSf§D WKUHHPRQWK VXSHUYLVHG RQWKHMRE WUDLQLQJ DQG OHDUQLQJ VLWXDWLRQ DSSURYHG E\ $27$ 7KLV LQWHUQVKLS LV UHTXLUHG DV SDUWLDO IXOILOOPHQW IRU FRPSOHWLRQ RI WKH RFFXSDn WLRQDO WKHUDS\ SURJUDP SULRU WR WKH 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ 6LJQLILFDQFH RI WKH 6WXG\ :LWK WKH FRQWHPSRUDU\ HPSKDVLV RQ DFFRXQWDELOLW\ DOO VFKRROV DUH EHLQJ DVNHG WR MXVWLI\ WKHLU FRQWLQXDWLRQ EDVHG XSRQ WKHLU SXUSRVH PHWKRGV RI DFKLHYLQJ WKDW SXUSRVH DQG DOORFDWLRQ RI PRQH\ SHUVRQQHO DQG HTXLSPHQW WKHUHE\ VFUXWLQL]LQJ SULRULWLHV DQG UHVSRQVLELOLWLHV WR

PAGE 22

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n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n WLRQ RI WUDLQLQJ DQG MRE VXFFHVV LQ HGXFDWLRQ SV\FKRORJ\ VRFLRORJ\ DQG DOOLHG KHDOWK ILHOGV

PAGE 23

&KDSWHU ,,, GHVFULEHV WKH PDQQHU LQ ZKLFK WKH GDWD ZHUH REWDLQHG DQG SURFHVVHG VSHFLILFDOO\ GHDOLQJ ZLWK WKH GHVLJQ DQG PDWKHPDWLFDO PHWKRGRORJ\ &KDSWHU ,9 FRQWDLQV WKH SUHVHQWDWLRQ RI WKHVH GDWD ZKLOH &KDSWHU 9 FRQWDLQV D VXPPDU\ RI WKH UHVXOWV FRQFOXVLRQV REVHUYDWLRQV DQG UHFRPPHQGDWLRQV

PAGE 24

&+$37(5 ,, 5(9,(: 2) 5(/$7(' /,7(5$785( 7KH OLWHUDWXUH UHODWHG WR WKLV VWXG\ ZDV UHYLHZHG ZLWKLQ IRXU PDMRU FDWHJRULHV 7KH\ ZHUH f DFDGHPLF VXUYLYDO SHUIRUPDQFH DQG SHUVLVWHQFH f SHUIRUPDQFH LQ SUDFWLFDO DSSOLFDWLRQ RI WUDLQLQJ VXFK DV LQWHUQVKLSV f SURIHVVLRQDO NQRZOHGJH DV PHDVXUHG E\ ZULWWHQ WHVWV DQG f VKRUWWHUP SURIHVVLRQDO SHUIRUPDQFH DQG ORQJWHUP SHUVLVWHQFH 7KH VWXGLHV ZHUH DOO FRQFHUQHG ZLWK VSHFLILF WUDLQLQJ FXUULFXOD DW WKH MXQLRU FROOHJH FROOHJH DQG JUDGXDWH OHYHOV RI KLJKHU HGXFDWLRQ SV\FKRORJ\ VRFLRORJ\ DQG DOOLHG KHDOWK $FDGHPLF 3HUIRUPDQFH 3HUVLVWHQFH RU :LWKGUDZDO 7KH OLWHUDWXUH DERXQGV ZLWK VWXGLHV VHHNLQJ SUHGLFWRUV RI FROOHJH SHUIRUPDQFH SHUVLVWHQFH RU ZLWKGUDZDO 6RPH QRWDEOH RQHV E\ $LNHQ f $n+HDUQ f %RZHUV DQG /RHE f *UHHQ f DQG +LQNOH DQG +RXVWRQ f DOO IRXQG WKDW WKH EHVW SUHGLFWRUV RI IXWXUH DFDGHPLF DFKLHYHPHQW ZHUH SUHYLRXV DFKLHYHPHQW DQG LQWHOOHFWXDO FKDUDFWHULVWLFV PHDQLQJ KLJK VFKRRO UDQN $PHULFDQ &ROOHJH 7HVWLQJ $&7f DQG 6FKRODVWLF $SWLWXGH 7HVWLQJ 6$7f WHVW VFRUHV UHDGLQJ WHVW VFRUHV DQG $GMHFWLYH &KHFN /LVW VFRUHV 7KHLU IRFXV ZDV WRZDUG ILUVW VHPHVWHU RU ILUVW TXDUWHU FROOHJH JUDGHV ,Q VWXG\LQJ SHUVLVWHQFH DQG ZLWKGUDZDO RI VWXGHQWV DW WKH HQG RI WKH IUHVKPDQ \HDU RI FROOHJH +XPSKUH\V HW DO f .QRSI f .QDXHU

PAGE 25

f DQG 5HHG HW DO f FRQFXUUHG WKDW ORZHU $&7 VFRUHV DQG ORZHU KLJK VFKRRO UDQN ZHUH PRUH FORVHO\ UHODWHG WR ZLWKGUDZDO WKDQ KLJKHU UDQN DQG VFRUHV /HVV WKDQ SHU FHQW RI WKH ORZHU VFRULQJ JURXS VWD\HG WR WKH VHFRQG VHPHVWHU RI WKH MXQLRU \HDU +XPSKUH\V HW DO f +DOO DQG &RDWHV f 3HGULQL DQG 3HGULQL f DQG 3IHLIHU DQG 6HGODFHN f VWDWHG DJDLQ WKDW KLJK VFKRRO DFKLHYHPHQW DQG DSWLWXGH WHVW VFRUHV SUHGLFWHG ILUVW \HDU FROOHJH SHUIRUPDQFH IRU UDFLDO PLQRULW\ JURXSV DERXW WKH VDPH DV IRU ZKLWHV %LVWUHLFK f IRXQG RYHUDOO KLJK VFKRRO JUDGH SRLQW DYHUDJH *3$f DQG KLJK VFKRRO *3$ LQ VFLHQFH DQG (QJOLVK ZHUH VLJQLILFDQW SUHGLFWRUV RI MXQLRU FROOHJH *3$ ,QNHQEHUU\ f WULHG WR SUHGLFW ILUVW \HDU FROOHJH SHUVLVWHQFH XVLQJ QRQFRJQLWLYH YDULDEOHV DQG LQWHOOHFWLYH IXQFWLRQV +H DOVR IRXQG WKDW WKH LQWHOOHFn WLYH IXQFWLRQ ZDV WKH PRVW VLJQLILFDQW YDULDEOH ZLWK D SDUW RI WKH FXOWXUDO DUHD VH[ DQG VRFLDO EDFNJURXQG %UD]]LHO f DQG &KDVH DQG -RKQVRQ f ERWK IRXQG KLJK VFKRRO UDQN WR EH D VWDWLVWLFDOO\ VLJQLILFDQW SUHGLFWRU RI VXFFHVV LQ FROOHJH WKURXJK WKH VHFRQG \HDU 7KH\ IRXQG WZR RWKHU VLJQLILFDQW YDULDEOHV KRXUV RI KLJK VFKRRO (QJOLVK DQG OHYHO RI IDPLO\ LQFRPH :LWKLQ VSHFLILF WHFKQLFDO FXUULFXOD 5 %URZQ DQG 6WRQHV f IRXQG QR FORVH DVVRFLDWLRQ EHWZHHQ LQWHOOLJHQFH IDFWRUV DQG ZLWKGUDZDO UDWHV LQ WKH ILUVW \HDU RI QXUVHVn WUDLQLQJ 'DYLHV DQG .KRVOD f VWXGLHG KHDOWK YLVLWRU VWXGHQWV XVLQJ LQWHOOLJHQFH WHVW VFRUHV DV D SUHGLFWLYH WRRO IRU WKRVH ZKR SHUIRUPHG DW KLJK OHYHOV LQ DFDGHPLFV 7KHLU PXOWLSOH FRUUHODWLRQ ZDV 5 5HHG HW DO f IRXQG VRPH QRQFRJQLWLYH YDULDEOHV WR EH VLJQLILn FDQW SUHGLFWRUV RI ILUVW \HDU VXFFHVV RI QXUVLQJ VWXGHQWV DJH \HDU

PAGE 26

RI HQWU\ LQWR WKH SURJUDP OHYHO RI SUHYLRXV HGXFDWLRQ DQG VFKRROV DWWHQGHG 6WUHVV DQG VDWLVIDFWLRQ RI ILUVW \HDU QXUVLQJ VWXGHQWV ZHUH DVVHVVHG E\ D TXHVWLRQQDLUH GHYHORSHG E\ .DW]HOO f 6WDWLVWLFDOO\ VLJQLILFDQW QHJDWLYH FRUUHODWLRQV ZHUH VKRZQ EHWZHHQ ZLWKGUDZDO DQG H[SHULHQFHG VDWLVIDFWLRQ DQG H[SHFWDWLRQV )LUVW VHPHVWHU JUDGH SRLQW DYHUDJH LQ D SK\VLFDO WKHUDS\ SURJUDP ZDV EHVW SUHGLFWHG E\ WKH (GXFDn WLRQDO ,QWHUHVW ,QYHQWRU\ )LQH $UWV 6FDOH DQG WKH VFRUH RI WKH $WWLWXGH 7RZDUG 'LVDEOHG 3HUVRQV 6FDOH $7'36f %XFD f ,Q D VWXG\ E\ 6FKLPSIKDXVHU DQG %URVNL f RI ILUVW \HDU DFDGHPLF VXFFHVV LQ D KHDOWK UHODWHG SURIHVVLRQV SURJUDP WKH $&7 VFRUH ZDV IRXQG WR EH WKH EHVW SUHGLFWRU /RZHQWKDO HW DO f IRXQG WKDW WKH 3KDUPDF\ &ROOHJH $GPLVVLRQ 7HVW 3&$7f DFFRXQWHG IRU SHU FHQW RI WKH YDULDWLRQ LQ FRPSOHWLRQ RI D SKDUPDF\ FXUULFXOXP +RZHYHU /LDR DQG $GDPV f IRXQG WKDW SUHSKDUPDF\ DFDGHPLF SHUIRUPDQFH ZDV WKH EHVW SUHGLFWRU RI ILUVW \HDU SKDUPDF\ *3$ QRW WKH 3&$7 6KDOLN f SUHVHQWHG D FRQn YLQFLQJ VWXG\ VKRZLQJ QR VWDWLVWLFDO GLIIHUHQFH EHWZHHQ WKH XVXDO VHOHFWLRQ SURFHVV IRU D SK\VLFDO WKHUDS\ SURJUDP DQG D VWUDWLILHG VDPSOLQJ SURFHVV IRU VWXGHQW VHOHFWLRQ 6SHFLILF HQJLQHHULQJ FRXUVH JUDGHV DQG *3$ ZHUH EHVW SUHGLFWHG E\ WKH $FKLHYHPHQW DQG $IILOLDWLRQ VFDOHV RI WKH (GZDUGV 3HUVRQDO 3UHIHUn HQFH 6FKHGXOH (336f DQG WKH VFRUHV RQ $PHULFDQ &RXQFLO RQ (GXFDWLRQ 3V\FKRORJLFDO ([DPLQDWLRQ 0XFKLQVN\ f ,Q VWXG\LQJ WUDQVIHU VWXGHQWV 1LFNHQV f VWDWHG WKDW MXQLRU FROOHJH *3$ ZDV WKH EHVW SUHGLFWRU RI ILUVW WHUP )ORULGD 6WDWH 8QLYHUVLW\ *3$

PAGE 27

%DNHU f %DLOH\ f %HDQ f +ROPVWURP f DQG /HH DQG 6KHSDUG f FRQGXFWHG GHVFULSWLYH VWXGLHV RI VWXGHQWV ZKR ZHUH SHUVLVWRUV DQG ZLWKGUDZDOV RI SURJUDPV LQ KHDOWK UHODWHG ILHOGV 7KH SHUVLVWRUV DQG JUDGXDWHV ZHUH JHQHUDOO\ EHWWHU DGMXVWHG PRUH LQGHn SHQGHQW VHOIFRQILGHQW UHVSRQVLEOH DQG LQWHOOLJHQW WKDQ WKRVH ZKR ZLWKGUHZ 3UHGLJHU f 5HDYOH\ DQG :LOVRQ f DQG 5RVH f IRXQG WKDW FRXQVHOLQJ VHUYHG DV D GHWHUUHQW WR IUHVKPDQ ZLWKGUDZDOV +RZHYHU 3UHGLJHU IRXQG WKDW WKH SDVW DFKLHYHPHQW GLPHQVLRQ PDGH WKH PRVW GLIIHUHQFH 0RUJDQ f XVHG 2PQLEXV 3HUVRQDOLW\ ,QYHQWRU\ IDFWRUV DQG $&7 VFRUHV WR GLVFULPLQDWH EHWZHHQ JURXSV IRU SHUVLVWHQFH LQ FROOHJH VHYHQ VHPHVWHUV DIWHU EHJLQQLQJ +HU UHVXOWV VKRZHG WKDW VWXGHQWV GLIIHUHG E\ VH[ LQ SHUVRQDOLW\ FRQILJXUDWLRQV DELOLW\ DQG FROOHJH SHUVLVWHQFH +RZHYHU WKH EHVW SUHGLFWLYH YDULDEOHV IRU FODVVLI\LQJ WKH VWXGHQWV LQWR JURXSV ZHUH $&7 VFRUHV IROORZHG E\ DQ DXWKRULWDULDQLVP PHDVXUH $OGDJ DQG 0DUWLQ f &\SUHV f DQG :LWWPH\HU HW DO f FRQFXUUHG WKDW $&7 VFRUHV GLIIHUHQWLDWHG WKH VWXGHQWV ZKR FRPSOHWHG WKH FROOHJH FXUULFXOXP DQG WKRVH ZKR ZLWKGUHZ &\SUHV VXJJHVWHG WKDW KLJK VFKRRO FODVV UDQN VKRXOG DOVR EH D SDUW RI D VHOHFWLRQ SURFHVV ,Q DGGLWLRQ :LWWPH\HU HW DO IRXQG WKDW VWXGHQWV ZKR FRPSOHWHG WKH SURJUDP WHQGHG WR EH OHVV LQGHSHQGHQW DQG OHVV YHQWXUHVRPH DV PHDVXUHG E\ WKH 3HUVRQn DOLW\ )DFWRU ,QGLFDWRU 3),f DQG WKH 0\HUV%ULJJV 7\SH ,QGLFDWRU -RKQVRQ f XVHG WKH 3), 6WURQJ 9RFDWLRQDO ,QWHUHVW %ODQN 69,%f DQG D FROOHJH TXDOLILFDWLRQ WHVW WR SUHGLFW QXUVLQJ WKHRU\ DQG SUDFWLFH JUDGHV +LV UHVXOWV VKRZHG WKDW DEVWUDFW UHDVRQLQJ DELOLW\ ZDV WKH EHVW SUHGLFWRU &RPSOHWLRQ RI D WHDFKHU WUDLQLQJ SURJUDP ZDV VWXGLHG E\

PAGE 28

%HOFDVWUR f LQ ZKLFK KH XVHG WKH (GZDUGV 3HUVRQDO 3UHIHUHQFH 6FKHGXOH (336f DQG WKH 69,% 7KH $XWRQRP\ 6FDOH RI WKH (336 DQG WKH $UFKLWHFW 6FDOH RI WKH 69,% UDQNHG VHFRQG DQG WKLUG UHVSHFWLYHO\ ,Q ILUVW SODFH ZDV WKH VWXGHQWVn *3$ 2QH WR IRXU\HDU VXFFHVV LQ D KHDOWK HGXFDWLRQ SURJUDP ZDV DJDLQ GHWHUPLQHG WR EH EHVW SUHGLFWHG E\ *3$ E\ +HLW HW DO f 6XFFHVV LQ JUDGXDWLQJ IURP D WKUHH \HDU SKDUPDF\ SURJUDP ZDV DOVR SUHGLFWHG E\ SUHSKDUPDF\ *3$ 3DOPLHUL f %XUJHVV f /HZLV DQG :HOFK f DQG 2ZHQ DQG )HOGKXVHQ f IRXQG WKDW SUHQXUVLQJ *3$ ZDV WKH EHVW LQGLFDWRU RI QXUVLQJ FRXUVH JUDGH SRLQW DYHUDJH &UDQH f DQG /HYLWW HW DO f ERWK VKRZHG WKDW WKH .XGHU 3UHIHUHQFH 5HFRUG DQG WKH (336 DV VWDWLVWLFDOO\ VLJQLILFDQW LQ SUHGLFWn LQJ VWXGHQWV ZKR ZRXOG FRPSOHWH DQ RFFXSDWLRQDO WKHUDS\ SURJUDP DQG D QXUVLQJ SURJUDP UHVSHFWLYHO\ *UDGXDWH VFKRRO VXFFHVV DV PHDVXUHG E\ WKH JUDGXDWH OHYHO *3$ ZDV VWXGLHG E\ %HDQ DQG &RYHUW f )HOGPDQ f )HGHULFL DQG 6FKXHUJHU f 5 6WHLQ DQG *UHHQ f DQG :LOOLQJKDP f 7KH EHVW YDULn DEOH LQ DOO RI WKHVH VWXGLHV ZDV WKH *UDGXDWH 5HFRUG ([DPLQDWLRQ VFRUHV &ODSS f *RXJK DQG +DOO f 5LFKDUGV HW DO f DQG 5RHPHU f DJUHHG WKDW WKH 0HGLFDO &ROOHJH $GPLVVLRQ 7HVW 0&$7f ZDV D VLJQLIn LFDQW SUHGLFWRU RI PHGLFDO VFKRRO VXFFHVV *3$f 7KH\ DOVR IRXQG WKDW SUHPHGLFDO *3$ DQG VFLHQFH *3$ ZHUH VLJQLILFDQW 3UHGHQWDO *3$ ZDV WKH PRVW VLJQLILFDQW SUHGLFWRU RI MXQLRU \HDU GHQWDO *3$ &KHQ HW DO f +DOIWHUnV f VWXG\ VKRZHG WKDW WKH /DZ 6FKRRO $GPLVVLRQV 7HVW /6$7f VFRUH SUHGLFWHG ILUVW \HDU ODZ VFKRRO *3$ VRPHZKDW EHWWHU WKDQ WKH XQGHUJUDGXDWH *3$ GLG

PAGE 29

1HHO\ f VKRZHG KLJK VFKRRO *3$ DQG WKH $&7 VFRUHV ZHUH PRVW FRQVLVWHQWO\ DVVRFLDWHG ZLWK SRVW JUDGXDWH ZRUN +RZHYHU 6WHLQ HW DO f VKRZHG WKDW RYHUDOO SHUIRUPDQFH RI ILUVW \HDU VRFLDO ZRUN VWXGHQWV HYDOXDWHG E\ IDFXOW\ DW WKH HQG RI WKH \HDU FRXOG EH SUHGLFWHG IURP VWXGHQWVn DWWLWXGLQDO UDWLQJV GRQH DW WKH EHJLQQLQJ RI WKH \HDU ,Q VXPPDU\ PRVW RI WKHVH VWXGLHV IRXQG WKDW DFDGHPLF DFKLHYHPHQW DW WKH FROOHJH RU JUDGXDWH OHYHO FRXOG EHVW EH SUHGLFWHG E\ SUHYLRXV SHUIRUPDQFH HLWKHU RQ DFKLHYHPHQW WHVWV RU ZLWK *3$V 7KH UHVXOWV WKRXJK RIWHQ VWDWLVWLFDOO\ VLJQLILFDQW ZHUH QRW VXIILFLHQW IRU SUDFWLFDO DSSOLFDWLRQ WRZDUG WKH VHOHFWLRQ RI VWXGHQWV LQWR YDULRXV FXUULFXOD DV VXSSRUWHG E\ +DOO DQG &RDWHVn f UHYLHZ RI VWXGLHV FRPSOHWHG EHWZHHQ DQG +LV UHYLHZ VWDWHG WKDW QR UHOLDEOH YDULDEOHV KDYH HPHUJHG WR EH RI JHQHUDO YDOXH LQ WKH VHOHFWLRQ RI QXUVHV 3UDFWLFDO $SSOLFDWLRQ /LQG f SUHGLFWHG FOLQLFDO SUDFWLFH JUDGHV RI RFFXSDWLRQDO WKHUDS\ VWXGHQWV ZLWK FROOHJH *3$ 7LGG DQG &RQLQH f VWXGLHG WKH *3$ RI SK\VLFDO WKHUDS\ VWXGHQWV DQG IRXQG WKDW DFDGHPLFDOO\ EHWWHU VWXGHQWV DOVR WHQGHG WR SHUIRUP EHWWHU LQ WKH FOLQLF WKDQ GLG WKHLU SHHUV (QJOHKDUW f VKRZHG WKDW RQWKHMRE SHUIRUPDQFH GXULQJ FOLQLFDO WUDLQLQJ RI RFFXSDWLRQDO WKHUDS\ VWXGHQWV PD\ EH SUHGLFWHG IURP SHUIRUPDQFH LQ VRPH UHTXLUHG FROOHJH FRXUVHV EXW RWKHU FRXUVHV GR QRW SUHGLFW FOLQLFDO SHUIRUPDQFH 'XQWHPDQ f ZDV QRW DEOH WR SUHGLFW FOLQLFDO UDWLQJV IURP IUHVKPDQ DQG VRSKRPRUH FRXUVH SHUIRUPDQFH %DLOH\ HW DO f IRXQG WKDW WKH ,QWHUHVW 7HVW ZDV WKH PRVW HIIHFWLYH SUHGLFWRU RI FOLQLFDO SHUIRUPDQFH RI RFFXSDWLRQDO WKHUDS\ VWXGHQWV

PAGE 30

%OXPH f FRUUHODWHG VXFFHVV LQ D PHGLFDO WHFKQRORJ\ LQWHUQVKLS ZLWK UHVSRQVHV RQ D %LRJUDSKLFDO 'DWD LQYHQWRU\ %RRWK f VKRZHG VLJQLILFDQW FRUUHODWLRQV EHWZHHQ SHGLDWULF LQWHUQVKLS RI RFFXSDWLRQDO WKHUDS\ VWXGHQWV DQG WKH 7 VFRUH RI *XLOIRUGnV 67'&5 ,QYHQWRU\ DQG WKH /LWHUDU\ ,QWHUHVW VFDOH RI WKH .XGHU 3UHIHUHQFH 5HFRUG DOWKRXJK WKH ODWWHU ZDV D QHJDWLYH FRUUHODWLRQ 7KH & VFRUH RI 67'&5 FRUUHODWHG ZLWK JUDGHV IURP D WXEHUFXORVLV LQWHUQVKLS DQG WKH WRWDO LQWHUQVKLS JUDGH ZDV OLQNHG ZLWK WKH SHUVXDVLYH LQWHUHVW VFDOH RI WKH .XGHU 3UHIHUHQFH 5HFRUG 2VLSRZ DQG :DOVK f FRQGXFWHG WZR VWXGLHV WKDW LQGLFDWHG WHVWV RI VRFLDO LQWHOOLJHQFH ZHUH XVHIXO LQ SUHGLFWLQJ D GHVLUHG IDFHW RI SHUIRUPDQFH RI FRXQVHORU WUDLQHHV LQ FRXQVHOLQJ SV\FKRORJ\ 0F&ODLQ f IRXQG WKDW WKH 3), WHVW ZDV PRUH XVHIXO IRU ERWK PHQ DQG ZRPHQ LQ LGHQWLI\LQJ WKH VXSHULRU FRXQVHORUV WKDQ LQ GLIIHUHQWLDWLQJ WKH DYHUDJH IURP WKH SRRU *RXJK f VWDWHG WKDW KLJK SUHPHGLFDO VFLHQFH FRPSHWDQFH DQG SHUIRUPDQFH ZDV XQUHODWHG WR FOLQLFDO SHUIRUPDQFH DQG IDFXOW\ UDWLQJV RI FOLQLFDO FRPSHWHQFH 0XQUR f GHFLGHG WKDW WKH 0LQQHVRWD 7HDFKHU $WWLWXGH ,QYHQWRU\ ZDV DQ LQDGHTXDWH SUHGLFWRU RI WHDFKLQJ SHUIRUPDQFH LQ LQWHUQVKLS 7HVWV RI 3URIHVVLRQDO .QRZOHGJH $\HUV DQG 5RKU f DQG :DOEHUJ f DWWHPSWHG WR SUHGLFW VFRUHV RQ WKH 1DWLRQDO 7HDFKHU ([DPLQDWLRQ 17(f $\HUV IRXQG WKDW JUDGHV LQ VRFLDO VFLHQFH ZHUH WKH EHVW LQGLFDWRUV RI VXFFHVV RQ WKH 17( ZKLOH :DOEHUJ IRXQG WKDW KLJK VFKRRO *3$ DQG VFRUHV RQ WKH 6FKRRO DQG &ROOHJH $ELOLW\ 7HVW 6&$7f ZHUH WKH EHVW LQGLFDWRUV

PAGE 31

+LJK VFKRRO SHUIRUPDQFH DQG VWDQGDUGL]HG DFKLHYHPHQW WHVWV ZHUH WKH EHVW SUHGLFWRUV RI VFRUHV RQ QXUVLQJ 6WDWH %RDUG 7HVWV +DJOXQG 7DWKDP DQG :HEHU HW DO f /DUNHQ f IRXQG WKDW VFRUHV RQ WKH FRPSDUDWLYH *XLGDQFH DQG 3ODFHPHQW 7HVWV &*3f WDNHQ DW HQWU\ WR D QXUVLQJ FXUULFXOXP DSSHDUHG WR GLVFULPLQDWH EHWZHHQ 6WDWH %RDUG 7HVW SDVVHV DQG IDLOXUHV %UDQGW HW DO f VKRZHG D ODFN RI UHODWLRQVKLS EHWZHHQ JUDGHV LQ QXUVLQJ WKHRU\ FRXUVHV DQG VFRUHV RQ 6WDWH %RDUG 7HVWV $ ORZ FRUUHODWLRQ EHWZHHQ WKH RFFXSDWLRQDO WKHUDS\ 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ VFRUH DQG *XLOIRUGnV 67'&5 ,QYHQWRU\f§5 )DFWRU ZDV UHYHDOHG E\ %RRWK f (QJHOKDUW f IRXQG WKDW VRPH JUDGHV LQ VSHFLILF FRXUVHV SUHGLFW RFFXSDWLRQDO WKHUDS\ 5HJLVWU\ ([DPLQDWLRQ VFRUHV )UHGULFNV HW DO f LQYHVWLJDWHG WKH UHODWLRQVKLS EHWZHHQ PHGLFDO VWXGHQWVn SHUIRUPDQFH RQ WKH 1DWLRQDO %RDUG ([DPLQDWLRQV DQG WKHLU SHUIRUPDQFH RQ WKH 0&$7 DQG WKHLU DFDGHPLF DFKLHYHPHQW RYHU WKH IRXU \HDUV RI PHGLFDO VFKRRO 7KH PRVW HIIHFWLYH SUHGLFWRU IRU SHUIRUPDQFH RQ WKH 1DWLRQDO %RDUG ([DPLQDWLRQ ZDV WKH *3$ IURP WKH VHFRQG \HDU RI PHGLFDO VFKRRO %HUJPDQ HW DO f FRQGXFWHG D FRPSUHKHQVLYH VWXG\ LQ ,VUDHO RQ QXUVLQJ VWXGHQWV 7KH\ IRXQG SV\FKRORJLFDO WHVWV FRUUHODWHG VLJQLIn LFDQWO\ ZLWK DFKLHYHPHQW LQ WKHRUHWLFDO OHDUQLQJ 3HUVRQDOLW\ IDFWRUV ZHUH VWXGLHG E\ .UDOO f LQ UHODWLRQ WR QXUVLQJ VXFFHVV DQG IDLOXUH 7KH ,QWHUSUHWDWLRQ RI D 'UDZ$3HUVRQ 7HVW DQG VFRUHV IURP WKH 1DWLRQDO /HDJXH RI 1XUVLQJ 1/1f WHVW ZHUH XVHG WR FRPSDUH VXFFHVVIXO DQG IDLOHG VWXGHQWV 7KH 1/1 WHVW ZDV IRXQG WR EH VLJQLILFDQWO\ UHODWHG WR QXUVLQJ VFKRRO ZLWKGUDZDO LQ WKH ILUVW \HDU IRU WKRVH ZKR GLG QRW OHDYH

PAGE 32

IRU SHUVRQDO UHDVRQV 8QVXFFHVVIXO VWXGHQWV SURGXFHG PRUH LPFRPSOHWH ILJXUHV DQG GUHZ WKH RSSRVLWH VH[ ILUVW VLJQLILFDQWO\ PRUH RIWHQ WKDQ GLG VXFFHVVIXO VWXGHQWV 3URIHVVLRQDO 3HUIRUPDQFH DQG 3HUVLVWHQFH %UDQGW f %XUWRQ f DQG 'XEV f VWXGLHG WKH UHFRUGV RI QXUVHV ZLWKLQ RQH \HDU RI JUDGXDWLRQ %UDQGW IRXQG WKDW b RI WKHP ZHUH ZRUNLQJ EXW QR SUHGLFWRU YDULDEOHV ZHUH VWDWLVWLFDOO\ VLJQLILFDQW %XUWRQ UHYHDOHG WKDW WKH 3) LQGLFDWRU SUHGLFWHG WKH SHUIRUPDQFH RI QHZ QXUVHV VOLJKWO\ EHWWHU WKDQ FKDQFH 'XEV XVHG FXUULFXOXP REMHFWLYHV WR GHYHORS DQ HPSOR\HU UDWLQJ VFDOH DQG WKHQ FRPSDUHG WKH UHVXOWV ZLWK JUDGHV WR VHH LI WKH FXUULFXOXP ZDV GRLQJ ZKDW LW ZDV GHVLJQHG WR GR 0D\ DQG &KLWW\ f XVHG WKH SURJUHVVLYH PDWULFHV DQG 0LOO +LOO 9RFDEXODU\ 7HVW LQ WKHLU VWXG\ RI QHZ VWDWH HQUROOHG QXUVHV LQ WKHLU ILUVW MREV 7KH\ FRQFOXGHG WKDW WKH WHVWV FRXOG EH XVHG RQO\ DV D PHDQV RI REWDLQLQJ PRUH LQWHOOLJHQW VWXGHQW QXUVHV QRW IRU FRPSHWHQW VWDWH HQUROOHG QXUVHV 'L 0DUFR DQG +LOOLDUG f GHFLGHG WKHUH ZDV QR FRQn FOXVLYH HYLGHQFH WKDW DVVRFLDWH GLSORPD DQG EDFFDODXUHDWH GHJUHH QXUVHV ZHUH HLWKHU WKH VDPH RU GLIIHUHQW LQ WHUPV RI WKH SURIHVVLRQDO SHUIRUPDQFH YDULDEOHV PHDVXUHG 6HYHUDO VWXGLHV KDYH IRXQG QHJDWLYH UHVXOWV $ RQH \HDU IROORZXS VWXG\ RI FOHULFDO ZRUNHUV E\ 6FKPLGW f VKRZHG WKDW WKH ORZ DELOLW\ JURXS ZDV PRUH VDWLVILHG UDWKHU WKDQ WKH PLGGOH RU KLJK JURXSV &RUQHWW f VWXGLHG WKH HIIHFWLYHQHVV RI *3$ DQG JUDGHV LQ WZR FODVVHV LQ SUHGLFWLQJ SHUIRUPDQFH RI ILUVW \HDU WHDFKHUV RQ WKH MRE 7KH EHVW VLQJOH SUHGLFWRU IRU VHFRQGDU\ WHDFKLQJ SHUIRUPDQFH ZDV FROOHJH

PAGE 33

*3$ KRZHYHU WKLV FRUUHODWLRQ ZDV YHU\ ORZ 0F*UHHY\ f IRXQG WKH 0LOOHU $QDORJLHV 7HVW GLG DQ DFFHSWDEOH MRE RI LGHQWLI\LQJ DFDGHPLFn DOO\ WDOHQWHG SHRSOH ,Q SK\VLFDO WKHUDS\ :DONHU f IRXQG QR IDFWRU WR EH D VLJQLIn LFDQW SUHGLFWRU RI JUDGXDWLRQ DQG HQWUDQFH LQWR WKH SURIHVVLRQ IURP JUDGHV DQG DFKLHYHPHQW PHDVXUHV $QGHUVRQ DQG -DQW]HQ f DJUHHG WKDW JUDGHV DQG DFKLHYHPHQW PHDVXUHV ZHUH LQHIIHFWLYH SUHGLFWRUV RI FOLQLFDO SHUIRUPDQFH LQ DQRWKHU KHDOWK ILHOG RFFXSDWLRQDO WKHUDS\ 0XWKDUG HW DO f GLG D VKRUW WHUP IROORZXS RQ RFFXSDWLRQDO WKHUDS\ JUDGXDWHV 7KH\ IDLOHG WR VKRZ WKDW WKH UHJLVWU\ H[DPLQDWLRQ DQG LQWHUQVKLS VFRUHV KDG SUHGLFWLYH YDOLGLW\ IRU VSHFLILF LQGLFHV RI ILUVW \HDU MRE VXFFHVV /RQJWHUP IROORZXS LQ WKH KHDOWK ILHOGV KDV EHHQ OLPLWHG 7KH VWXGLHV WKDW KDYH EHHQ GRQH ZHUH PRUH GHVFULSWLYH WKDQ SUHGLFWLYH +HQGULFNVRQ f DQG %URZQ DQG 6WRQHV f GHVFULEHG RFFXSDWLRQDO WKHUDS\ SHUVRQDOLW\ FKDUDFWHULVWLFV DQG MRE VDWLVIDFWLRQ %UROOLHU f H[SDQGHG KHU GHVFULSWLRQ WR LQFOXGH QRW RQO\ RFFXSDWLRQDO WKHUDSLVWV EXW SK\VLFDO WKHUDSLVWV DQG VRFLDO ZRUNHUV LQ D VPDOO VDPSOH RI HDFK 7KH GLIIHUHQFHV DSSHDUHG WR EH UHODWHG WR WKH GLIIHUHQW MRE FRQFHUQV RI WKH VRFLDO SV\FKLDWULF JURXS YHUVXV WKH SK\VLFDO GLVDELOLWLHV JURXS 6XVVPDQ DQG +DXJ f DQG -DQW]HQ f GHVFULEHG WKH W\SLFDO SHUVRQ LQ WKHLU UHVSHFWLYH ILHOGV 7KH LPSOLFDWLRQ LV WKDW VXFK GDWD FRXOG SRVVLEO\ EH XVHG IRU UHFUXLWPHQW DQG VHOHFWLRQ RI IXWXUH VWXGHQWV ZKR ZRXOG EH PRVW OLNHO\ WR VXFFHHG LQ D FDUHHU RI WKHVH KHDOWK UHODWHG ILHOGV 'UXJR f 0D\QDUG HW DO f 1HOVRQ f 7KRP DQG +LFNFR[ f DQG 7XFNHU DQG 6WURQJ f FDUULHG RXW SUHGLFWLYH VWXGLHV 'UXJR

PAGE 34

IRXQG QR VLJQLILFDQW FRUUHODWLRQV EHWZHHQ *3$ LQWHUQVKLS VFRUHV DQG HPSOR\HU UDWLQJV RI SK\VLFDO WKHUDSLVWVn RQH\HDU MRE VXFFHVV $ VHYHQ \HDU VWXG\ E\ 0D\QDUG HW DO f UHYHDOHG D VPDOO EXW VLJQLILFDQW UHODWLRQVKLS EHWZHHQ MRE SHUIRUPDQFH DQG SUDFWLFDO H[SHULHQFH JUDGHV 1HOVRQ f FRQFOXGHG IURP DQ H[WHQVLYH UHYLHZ RI WKH OLWHUDWXUH RI HGXFDWLRQ KHDOWK SV\FKRORJ\ VRFLDO VFLHQFH DQG EXVLQHVV WKDW *3$V IURP FROOHJH DUH RI OLWWOH XVH LQ SUHGLFWLQJ ODWHU RFFXSDWLRQDO VXFFHVV 7XFNHU f GLG D WHQ\HDU IROORZXS RI PHGLFDO VWXGHQWV DQG IRXQG WKDW LQWHUHVW VFDOHV GLG QRW PHDVXUH GLVWLQFWLRQV EHWZHHQ WKH PHGLFDO VSHFLDOW\ DUHDV *LQVEHUJ FRQFOXGHG WKDW FROOHJH SHUIRUPDQFH KDG QR GLIIHUHQWLDO VLJQLILFDQFH IRU ODWHU SHUIRUPDQFH KRZHYHU RXWVWDQGLQJ JUDGHV LQ JUDGXDWH VFKRRO GLG LQGLFDWH ODWHU VXSHULRU SHUIRUPDQFH 1HOVRQ f 7KRP DQG +LFNFR[ f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

PAGE 35

DFKLHYHPHQW ZHUH WKH EHVW SUHGLFWRUV RI ODWHU DFDGHPLF DFKLHYHPHQW DQG RI KLJKHU VFRUHV RQ ZULWWHQ WHVWV RI SURIHVVLRQDO NQRZOHGJH 0DQ\ VWXGLHV UHTXLUHG WKH DGPLQLVWUDWLRQ RI DGGLWLRQDO WHVWV DQG TXHVWLRQn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

PAGE 36

&+$37(5 ,,, '(6,*1 $1' 0(7+2'2/2*< 7KH SXUSRVH RI WKLV VWXG\ ZDV WR H[SORUH LQGHSHQGHQW YDULDEOHV WR GHWHUPLQH ZKLFK LI DQ\ FRXOG EH XVHG WR SUHGLFW VXFFHVV LQ DQ RFFXSDn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n WLFDO WHUPVf DQG WKH SLWIDOOV RI LQWHUSUHWLQJ UHJUHVVLRQ DQDO\VLV DUH GLVFXVVHG 7KH FRQFOXGLQJ VHFWLRQ FRQWDLQV D GHVFULSWLRQ RI KRZ WKH GDWD DUH RUJDQL]HG 3RSXODWLRQ 7KH XQGHUJUDGXDWH RFFXSDWLRQDO WKHUDS\ FXUULFXOXP DW WKH 8QLYHUVLW\ RI )ORULGD LV ZLWKLQ WKH &ROOHJH RI +HDOWK 5HODWHG 3URIHVVLRQV DQG VHUYHG DSSUR[LPDWHO\ MXQLRUV VHQLRUV DQG LQWHUQV HDFK \HDU 7KH ILYH\HDU SHULRG IURP WR ZDV FKRVHQ IRU VWXG\ EHFDXVH GXULQJ WKHVH \HDUV WKH FRXUVH RI VWXG\ DQG WKH IDFXOW\ UHPDLQHG TXLWH

PAGE 37

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

PAGE 38

D $ VWXGHQW LQIRUPDWLRQ IRUP IRU WKH &ROOHJH RI +HDOWK 5HODWHG 3URIHVVLRQV 7KLV ZDV D WUDGLWLRQDO IRUP VLPLODU WR D FXUULFXOXP YLWDH LQFOXGLQJ f H[WUDFXUULFXODU DFWLYLWLHV f KRQRUV DQG GLVWLQFWLRQV f HPSOR\PHQW f HVVD\ DV WR ZK\ WKH VWXGHQW ZDV LQWHUHVWHG LQ D KHDOWK FDUHHU E &RSLHV RI DOO WUDQVFULSWV PXVW KDYH EHHQ UHFHLYHG LQ 2FFXSDWLRQDO 7KHUDS\ 'HSDUWPHQW YLD WKH 5HJLVWUDUnV 2IILFH F 6WXGHQW PXVW KDYH FRPSOHWHG WKH IRUP OLVWLQJ VWXGHQW RUJDQL]DWLRQV DFWLYLWLHV DQG SXEOLFDWLRQV )DFXOW\ HYDOXDWLRQ FRPPHQFHG DIWHU WKH DERYH VWHSV ZHUH FRPSOHWHG E\ WKH DSSOLFDQW 7KH VHOHFWLRQ SURFHVV FRQVLGHUHG WKH IROORZLQJ *UDGHV ZHUH LPSRUWDQW WKH IDFXOW\ VFDQQHG FUHGHQWLDOV RI DOO FDQGLGDWHV SRVVHVVLQJ D *3$ RU EHWWHU 3UHSURIHVVLRQDO JUDGHV PXVW EH RU EHWWHU 7KH HVVD\ ZDV UHDG DQG WKH IROORZLQJ IDFWRUV ZHUH HYDOXDWHG D 7KH VWXGHQWnV PRWLYDWLRQ E :K\ WKH VWXGHQW ZDV LQWHUHVWHG LQ RFFXSDWLRQDO WKHUDS\ F +RZ ZHOO WKH VWXGHQW ZURWH :RUN DQG YROXQWHHU H[SHULHQFH ZHUH FRQVLGHUHG 7ZR LQWHUYLHZV ZHUH FRQGXFWHG D 2QH ZDV DV DQ DGYLVHPHQW GLVFXVVLRQ E 7KH VHFRQG ZDV PRUH IRUPDO DQG ZDV UDQGRPO\ DVVLJQHG WR D GLIIHUHQW PHPEHU RI WKH IDFXOW\

PAGE 39

(DFK IDFXOW\ PHPEHU ZURWH D VXPPDU\ RI LPSUHVVLRQV RI WKH VWXGHQW 0HHWLQJ DV D VHOHFWLRQV FRPPLWWHH WKH HQWLUH IDFXOW\ GLVFXVVHG HDFK VWXGHQW ILQDOO\ UDQNLQJ DOO LQ RUGHU RI SUHIHUHQFH IRU DGPLVVLRQ 'DWD &ROOHFWLRQ 'DWD FROOHFWLRQ EHJDQ ZLWK WKH ILOHV LQ WKH RFFXSDWLRQDO WKHUDS\ FXUULFXOXP RIILFH DW WKH 8QLYHUVLW\ RI )ORULGDnV +LOOLV 0LOOHU +HDOWK &HQWHU 7KH VRFLDO VHFXULW\ QXPEHU ZDV REWDLQHG IRU HDFK VWXGHQW LQYROYHG LQ WKH RFFXSDWLRQDO WKHUDS\ FXUULFXOXP WKDW GLG RU ZRXOG KDYH JUDGXDWHG EHWZHHQ DQG 7KHVH ILOHV \LHOGHG WKH VFRUHV IRU WKUHH RI WKH GHSHQGHQW YDULDEOHV RQ HDFK VWXGHQW 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ 3V\FKLDWULF ,QWHUQVKLS DQG 3K\VLFDO 'LVDELOLWLHV ,QWHUQn VKLS 6LQFH VRPH RI WKH 5HJLVWU\ ([DPLQDWLRQ VFRUHV ZHUH QHYHU UHFRUGHG LQ WKHVH UHFRUGV LW ZDV WKHQ QHFHVVDU\ WR FRQVXOW WKH FRPSXWHU SULQWRXW IURP WKH $PHULFDQ 2FFXSDWLRQDO 7KHUDS\ $VVRFLDWLRQ $27$f 0DQ\ RI WKH VFRUHV IRU WKH LQWHUQVKLSV KDG QRW EHHQ FRGHG DQG FRPSXWHG 7KH IDFXOW\ PHPEHU ZKR KDG EHHQ LQ FKDUJH RI WKLV WDVN ZDV FRQVXOWHG 0RVW RI WKH LQWHUQVKLS HYDOXDWLRQV ZHUH IRXQG FRGHG FRPSXWHG DQG UHFRUGHG 6RPH VFRUHV ZHUH REWDLQHG E\ WHOHSKRQH DQG ZHUH QRW UHFRUGHG WKHUHIRUH WKH\ ZHUH QRW DYDLODEOH HYHQ IURP WKH LQWHUQVKLS LQVWLWXWLRQ 7KLV DIIHFWHG ILYH VWXGHQWV IRU WKH 3V\FKLDWULF ,QWHUQVKLS DQG WZR VWXGHQWV IRU WKH 3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLS VHH $SSHQGL[f 2QH VWXGHQW ZKR LV FXUUHQWO\ UHJLVWHUHG FRXOG QRW EH UHDFKHG WR JLYH SHUPLVVLRQ WR $27$ WR UHOHDVH KHU 5HJLVWU\ ([DPLQDWLRQ VFRUH $Q DYHUDJH ZDV FRPSXWHG IRU DOO

PAGE 40

WKRVH ZKR KDG VFRUHV DQG LQVHUWHG IRU WKH PLVVLQJ VFRUHV UDWKHU WKDQ XVLQJ D ]HUR WR DYRLG GLVWRUWLQJ WKH VWDWLVWLFDO UHVXOWV :LQHU f 7KH PHDQV XVHG ZHUH 3V\FKLDWULF ,QWHUQVKLS 3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLS 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ 3HUPLVVLRQ ZDV REWDLQHG IURP WKH 5HJLVWUDUn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n L]DWLRQ RI FDWHJRULHV ZDV ILQDOO\ DFKLHYHG )RU H[DPSOH 6RFLDO VFLHQFHV KLVWRU\ DQWKURSRORJ\ ZHVWHUQ FLYLOL]DWLRQ %LRORJLFDO VFLHQFHV ]RRORJ\ ELRORJ\ 3K\VLFDO VFLHQFHV FKHPLVWU\ SK\VLFV JHRORJ\ $UW VWXGLR 0XVLF DSSUHFLDWLRQ OHVVRQV FKRUDO /LWHUDWXUH (QJOLVK DQG OLWHUDWXUH &ROOHJH FRXUVH JUDGHV ZHUH FRQYHUWHG WR SRLQWV $ % & DQG 6HPHVWHU JUDGHV ZHUH FRQYHUWHG WR TXDUWHU KRXU FUHGLWV 7KH

PAGE 41

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f $Q DYHUDJH VFRUH IRU DOO WKRVH ZKR KDG VFRUHV ZDV FRPSXWHG DQG XVHG UDWKHU WKDQ D ]HUR :LQHU f 7KH PHDQV XVHG ZHUH &((% )ORULGD WK *UDGH 7HVW $ WDEOH ZDV REWDLQHG IURP WKH 8QLYHUVLW\ RI )ORULGD $GPLVVLRQV 2IILFH WR FRQYHUW WKH WRWDO VFRUH IRU WKH )ORULGD WK *UDGH 7HVW WR WKH HTXLYDOHQW WRWDO IRU WKH 6FKRODVWLF $SWLWXGH 7HVW VFRUHV RI WKH &((% 7KH $PHULFDQ 2FFXSDWLRQDO 7KHUDS\ $VVRFLDWLRQ ZDV FRQWDFWHG FRQn FHUQLQJ IROORZXS RQ DOO VXEMHFWV ZKR KDG FRQWLQXHG WKHLU QDWLRQDO

PAGE 42

UHJLVWUDWLRQ 6SHFLILF LQIRUPDWLRQ UHTXHVWHG IURP WKH DQG PHPEHU GDWD VXUYH\V ZDV IRUZDUGHG WR WKLV UHVHDUFKHU VHH $SSHQGL[f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f FRPSOHWHQRW FRPSOHWH WKH RFFXSDWLRQDO WKHUDS\ SURJUDP DQG f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f VWHSZLVH PXOWLSOH UHJUHVVLRQ DQDO\VLV 05f 7KH VHFRQG VWDWLVWLFDO SDUW XVHG VWHSZLVH GLVFULPLQDQW ',6f DQDO\VLV RI WKH %LR0HG SDFNDJH WR ILQG WKH EHVW LQGLFDWRUV RI ZRUN SHUVLVWHQFH DQG QRQSHUVLVWHQFH LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ DQG FRPSOHWLRQ RU QRQFRPSOHWLRQ RI WKH RFFXSDWLRQDO WKHUDS\ SURJUDP

PAGE 43

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n WLRQDO WKHUDS\ SURJUDP EHIRUH SDVVLQJ WKH UHJLVWU\ H[DPLQDWLRQ )URP WKH 8QLYHUVLW\ RI )ORULGD WUDQVFULSWV H[DPLQHG LQGHn SHQGHQW YDULDEOHV RFFXUUHG SULRU WR HQWHULQJ SURIHVVLRQDO FRXUVH ZRUN 7KHVH LQGHSHQGHQW YDULDEOHV LGHQWLILHG DV SRVVLEOH IDFWRUV H[SODLQLQJ WKH PRYHPHQW RI WKH GHSHQGHQW YDULDEOHV DUH OLVWHG EHORZ )ORULGD UHVLGHQW RU RXWRIVWDWH UHVLGHQW 8QLYHUVLW\ RI )ORULGD IUHVKPDQ RU WUDQVIHU *UDGH LQ SRLQWV LQ ,QWURGXFWLRQ WR WKH +HDOWK 5HODWHG 3URIHVVLRQV FRXUVH *DS LQ \HDUV IURP FROOHJH IUHVKPDQ WR XSSHU GLYLVLRQ VWDWXV 1XPEHU RI KRXUV LQ VSHHFK FRXUVHV

PAGE 44

1XPEHU RI KRXUV LQ EXVLQHVV FRXUVHV 1XPEHU RI KRXUV LQ QXUVLQJ FRXUVHV 1XPEHU RI KRXUV LQ OLEUDU\ VFLHQFH FRXUVHV *UDGH LQ SRLQWV LQ JURZWK DQG GHYHORSPHQW FRXUVHV 1XPEHU RI KRXUV LQ VSHFLDO HGXFDWLRQ FRXUVHV 1XPEHU RI KRXUV LQ GUDPD DQG WKHDWHU FRXUVHV 1XPEHU RI KRXUV LQ KHDOWK FRXUVHV 1XPEHU RI KRXUV LQ VRFLDO VFLHQFHV FRXUVHV QXPEHU RI KRXUV LQ SK\VLFDO VFLHQFHV FRXUVHV 1XPEHU RI KRXUV LQ ELRORJLFDO VFLHQFHV FRXUVHV 1XPEHU RI KRXUV LQ (QJOLVK DQG OLWHUDWXUH FRXUVHV 1XPEHU RI KRXUV LQ PDWKHPDWLFV FRXUVHV 1XPEHU RI KRXUV LQ SV\FKRORJ\ FRXUVHV 1XPEHU RI KRXUV LQ IRUHLJQ ODQJXDJH FRXUVHV 1XPEHU RI KRXUV LQ DUWV DQG FUDIWV FRXUVHV 1XPEHU RI KRXUV LQ PXVLF *UDGH SRLQW DYHUDJH IRU WKH ILUVW WZR \HDUV )ORULGD WK *UDGH 3ODFHPHQW 7HVW WRWDO VFRUH &ROOHJH (QWUDQFH ([DPLQDWLRQ %RDUG WRWDO WHVW VFRUH 7KH GHSHQGHQW YDULDEOHV HVWDEOLVKHG IRU WKLV VWXG\ ZHUH $PHULFDQ 2FFXSDWLRQDO 7KHUDS\ $VVRFLDWLRQ 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ VFRUH 3V\FKLDWULF ,QWHUQVKLS VFRUH 3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLS VFRUH :RUN VWDWXV DW WKH WLPH RI WKLV VWXG\ &RPSOHWLRQ RI QRQFRPSOHWLRQ RI WKH RFFXSDWLRQDO WKHUDS\ SURJUDP

PAGE 45

:RUNLQJ RU QRW ZRUNLQJ LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ DW WKH WLPH RI WKLV VWXG\ $ VFDOH HPSOR\LQJ WKH QXPEHUV ]HUR WKURXJK WHQ ZDV GHYHORSHG WR GHVFULEH GLIIHUHQW OHYHOV RI ZRUN VWDWXV VHH 7DEOH f 7KH OHDVW GHVLUDEOH OHYHO LQGLFDWHG E\ D ]HUR UHSUHVHQWHG QR DYDLODEOH IROORZXS RQ WKRVH ZKR GLG QRW FRPSOHWH WKH XQGHUJUDGXDWH FXUULFXOXP LQ RFFXSDn WLRQDO WKHUDS\ 7KH QXPEHU WHQ UHSUHVHQWHG WKH PRVW IDYRUDEOH IROORZXS LQIRUPDWLRQ RI ZRUNLQJ IXOO WLPH LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ 7DEOH :RUN 6WDWXV 6FDOH /HYHO 1XPEHU 'HVFULSWLRQ $ 1R IROORZXS DQG GLG QRW FRPSOHWH WKH XQGHUJUDGXDWH FXUULFXOXP LQ RFFXSDWLRQDO WKHUDS\ 'LG QRW FRPSOHWH WKH XQGHUJUDGXDWH FXUULFXOXP LQ RFFXSDWLRQDO WKHUDS\ DQG QRW ZRUNLQJ 'LG QRW FRPSOHWH WKH XQGHUJUDGXDWH FXUULFXOXP LQ RFFXSDWLRQDO WKHUDS\ DQG LV ZRUNLQJ LQ DQ XQUHODWHG ILHOG 'LG QRW FRPSOHWH WKH XQGHUJUDGXDWH FXUULFXOXP LQ RFFXSDWLRQDO WKHUDS\ DQG LV ZRUNLQJ LQ D KHDOWK UHODWHG SURIHVVLRQ 1R IROORZXS GDWD GLG FRPSOHWH WKH XQGHUJUDGXDWH FXUULFXOXP LQ RFFXSDWLRQDO WKHUDS\ DQG UHFRUGHG DV QRW UHJLVWHUHG 'LG FRPSOHWH WKH XQGHUJUDGXDWH FXUULFXOXP LQ RFFXSDWLRQDO WKHUDS\ QRW UHJLVWHUHG DQG ZRUNLQJ LQ DQ XQUHODWHG ILHOG 1R IROORZXS GDWD DQG UHFRUGHG DV UHJLVWHUHG 5HJLVWHUHG DQG ZRUNLQJ LQ XQUHODWHG ILHOG 5HJLVWHUHG QRW SUHVHQWO\ ZRUNLQJ EXW SODQV WR EH EDFN WR RFFXSDWLRQDO WKHUDS\ 5HJLVWHUHG DQG ZRUNLQJ SDUWWLPH LQ RFFXSDWLRQDO WKHUDS\ 5HJLVWHUHG DQG ZRUNLQJ IXOOWLPH LQ RFFXSDWLRQDO WKHUDS\

PAGE 46

$ VWHSZLVH PXOWLSOH UHJUHVVLRQ DQDO\VLV ZDV FDOFXODWHG ZLWK WKH %0'3 FRPSXWHU SDFNDJH XVLQJ WKH LQGHSHQGHQW DQG GHSHQGHQW YDULDEOHV 7KH GHSHQGHQW YDULDEOH ZDV LGHQWLILHG DV \ WKH LQGHSHQGHQW YDULDEOH DV ; 7KH SURFHVV RI QDUURZLQJ WKH SUHGLFWLYH LQGHSHQGHQW YDULDEOHV WR MXVW D IHZ LV FDOOHG YDULDEOH VHOHFW 7KLV \LHOGV 5 ZKLFK LV D FRHIILFLHQW RI GHWHUPLQDWLRQ WKDW LV LW \LHOGV WKH SHUFHQWDJH RI WKH PRYHPHQW RI < WKDW LV EHLQJ H[SODLQHG E\ WKH ;Vf RU WKH SHU FHQWDJH RI YDULDQFH LQ < GXH WR WKH UHJUHVVLRQ RQ ;Vf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f 'HILQLWLRQV RI 6WDWLVWLFDO 7HUPV &RUUHODWLRQ DQDO\VLVf§WKH GHJUHH WR ZKLFK WZR YDULDEOHV DUH UHODWHG /DSLQ f

PAGE 47

&RUUHODWLRQ FRHIILFLHQWf§VKRZV KRZ FORVHO\ WZR YDULDEOHV PRYH WRn JHWKHU E\ PHDQV RI D VLQJOH QXPEHU FDOFXODWHG IURP WKH VDPH GDWD &RHIILFLHQW RI GHWHUPLQDWLRQf§H[SUHVVHV WKH UHODWLYH UHGXFWLRQ LQ WKH YDULDWLRQ RI < WKDW FDQ EH DWWULEXWHG WR NQRZOHGJH RI ; DQG LWV UHODWLRQVKLS WR < E\ ZD\ RI WKH UHJUHVVLRQ OLQH RU WKH SURn SRUWLRQ RI WKH WRWDO YDULDWLRQ LQ < H[SODLQHG E\ WKH UHJUHVVLRQ OLQH ([SUHVVHV KRZ VWURQJO\ ; LV DVVRFLDWHG ZLWK < 5HSUHVHQWHG E\ UA RU /DSLQ f 'HSHQGHQW YDULDEOHf§WKH YDULDEOH ZKRVH YDOXH LV EHLQJ SUHGLFWHG UHSUHVHQWHG E\ WKH \ D[LV /DSLQ f ,QGHSHQGHQW YDULDEOHf§WKH YDULDEOH ZKRVH YDOXH LV NQRZQ DQG UHSUHVHQWHG E\ WKH ; D[LV /DSLQ f (VWLPDWHG UHJUHVVLRQ OLQHf§H[SUHVVHV WKH DYHUDJH UHODWLRQVKLS EHWZHHQ WKH ; DQG < YDULDEOHV 7KLV SURYLGHV DQ HVWLPDWH RI WKH PHDQ OHYHO RI WKH GHSHQGHQW YDULDEOH < ZKHQ WKH YDOXH RI ; LV VSHFLILHG %HFDXVH WKH HVWLPDWHG UHJUHVVLRQ OLQH SURYLGHV HVWLPDWHV RQO\ WKH V\PERO LV
PAGE 48

)LJXUH 6LPSOH /LQHDU 5HJUHVVLRQ S f

PAGE 49

0XOWLSOH UHJUHVVLRQ DQDO\VLVf§LV DQ H[SDQVLRQ RI WKH VDPH WHFKQLTXHV XVHG LQ VLPSOH UHJUHVVLRQ 0XOWLSOH UHJUHVVLRQ XVHV VHYHUDO LQGHSHQGHQW YDULDEOHV DV SUHGLFWRUV 7KH PRYHPHQW RU YDULDWLRQ RI < FDQ WKHQ EH H[SODLQHG E\ WZR RU PRUH YDULDEOHV 7KLV IRUPXOD LV
PAGE 50

LWV GLVWDQFH IURP WKH ZDOO RQ WKH ULJKW ZKLOH WKH GLVWDQFH IURP WKH OHIW ZDOO UHSUHVHQWV WKH YDOXH IRU /DSLQ f 7KHUH DUH QR WKHRUHWLFDO OLPLWV WR WKH QXPEHU RI YDULDEOHV WKDW PD\ EH XVHG LQ PXOWLSOH UHJUHVVLRQ 8QWLO WKH UHFHQW DFFHVVLELOn LW\ WR KLJKVSHHG FRPSXWHUV WKH KLJKHUGLPHQVLRQDO PXOWLSOH UHJUHVVLRQ ZDV VHYHUHO\ OLPLWHG LQ VFRSH EHFDXVH RI WKH YDVW TXDQWLW\ RI FDOFXODWLRQV LQYROYHG /DSLQ f 7KLV VWXG\ ZRXOG KDYH EHHQ SUDFWLFDOO\ LPSRVVLEOH EHIRUH WKH DGYHQW RI WKH FRPSXWHU 'LVFULPLQDQW IXQFWLRQ RI WZR JURXSVf§LV DQ H[SDQVLRQ RI PXOWLSOH UHJUHVVLRQ ,W LV LOOXVWUDWHG LQ WKH IROORZLQJ ILJXUH )LJXUH 'LVFULPLQDQW )XQFWLRQ S f

PAGE 51

:LWK WZR JURXSV $ DQG % DQG WZR PHDVXUHPHQWV ; DQG < RQ HDFK PHPEHU RI WKH WZR JURXSV D VFDWWHU SORW FRXOG EH GRQH IRU HDFK JURXS 7KHQ D FLUFOH LV GUDZQ DURXQG WKH JURXSV UHVSHFWLYHO\ $ VWUDLJKW OLQH LV WKHQ GUDZQ WKURXJK WKH WZR SRLQWV ZKHUH WKH HOOLSVHV LQWHUVHFW DQG WKHQ SURMHFWHG WR D QHZ D[LV = 7KH RYHUODS EHWZHHQ WKH GLVWULEXWLRQV $n DQG %n LV VPDOOHU WKDQ ZRXOG EH REWDLQHG E\ DQ\ RWKHU OLQH GUDZQ WKURXJK WKH HOOLSVHV 7KH = D[LV LV WKH GLVFULPLQDQW IXQFWLRQ LQ WKLV FDVH *URXS DYHUDJHV DQG PHDQVTXDUH GHYLDWLRQV RI JURXS YDOXHV DUH FRPSXWHG 7KHQ WKDW UHVXOW LV FRPSDUHG WR WKH PHDQ VTXDUH GHYLDWLRQV RI WKH JURXS DYHUDJHV IURP WKH RYHUDOO DYHUDJH ,I WKH ODWWHU LV ODUJH WKHQ WKH OLQHDU IXQFWLRQ XVHG LQ WKHVH FDOFXODWLRQV VHSDUDWHV WKH JURXSV ZHOO *UHHQ DQG 7XOO f 6WDQGDUG HUURU RI WKH HVWLPDWHf§WKH VTXDUH URRW RI WKH PHDQ VTXDUHG GHYLDWLRQV DERXW WKH UHJUHVVLRQ OLQH RU SODQH ,W PD\ EH XVHG WR HVWLPDWH WKH WUXH YDULDELOLW\ LQ < DERXW WKH PRGHO ,W GLUHFWO\ H[SUHVVHV WKH GHJUHH RI VFDWWHU LQ WKH GDWD /DSLQ f 3LWIDOOV RI 5HJUHVVLRQ DQG &RUUHODWLRQ 7HFKQLTXHV :KHQ UHDGLQJ D FRUUHODWLRQ FRHIILFLHQW WKH UHDGHU PXVW UHPHPEHU WKDW LW UHIOHFWV LQ QXPHULFDO YDOXH RQO\ WKH VWUHQJWK RI DVVRFLDWLRQ EHWZHHQ YDULDEOHV 7KLV LV D VWDWLVWLFDO UHODWLRQVKLS D ODUJH SRVLWLYH RU QHJDWLYH YDOXH IRU 5 RU U GRHV QRW LQGLFDWH D FDXVH DQG HIIHFW UHODWLRQVKLS ([DPSOHV RI QRQVHQVH RU VSXULRXV FRUUHODWLRQV DERXQG )RU PRVW SUDFWLFDO SUREOHPV WKHUH LV XVXDOO\ VRPH UHDVRQDEOH FRQQHFWLRQ 2IWHQ D UHODWLRQVKLS EHWZHHQ WZR YDULDEOHV PD\ EH H[SODLQHG E\ WKHLU

PAGE 52

LQWHUDFWLRQV ZLWK D FRPPRQ IDFWRU &DUH PXVW EH WDNHQ LQ XVLQJ SDVW GDWD WR GHWHUPLQH D IXWXUH UHODWLRQVKLS 7KH FRQGLWLRQV XQGHU ZKLFK WKH < DQG ; GDWD ZHUH REWDLQHG PD\ KDYH FKDQJHG RYHU WLPH &DXWLRQ VKRXOG EH XVHG ZKHQ PDNLQJ SUHGLFWLRQV RI < ZKHQHYHU ; IDOOV RXWVLGH WKH UDQJH RI YDOXHV XVHG LQ WKH RULJLQDO FROOHFWLRQV RI GDWD /DSLQ f 2UJDQL]DWLRQ RI WKH 'DWD )RU WKH SXUSRVHV RI WKLV VWXG\ WKH VWXGHQWVf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

PAGE 53

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n VLRQ DQDO\VLV ZDV XVHG IRU SUHGLFWRUV RI WKH IROORZLQJ GHSHQGHQW YDULDEOHV 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ 6FRUH 3V\FKLDWU\ ,QWHUQVKLS 6FRUH 3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLS 6FRUH :RUN 6WDWXV 6FRUH 6WHSZLVH GLVFULPLQDQW DQDO\VLV ZDV XVHG IRU SUHGLFWRUV RI FRUUHFWO\ FODVVLI\LQJ WKH VXEMHFWV LQWR GLFKRWRPRXV JURXSV UHSUHVHQWHG E\ WKH IROORZLQJ GHSHQGHQW YDULDEOHV &RPSOHWH RU QRW FRPSOHWH WKH SURJUDP IRU UHJLVWU\ LQ RFFXSDWLRQDO WKHUDS\

PAGE 54

:RUNLQJ RU QRW ZRUNLQJ LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ 7KH ODVW VHFWLRQ GLVFXVVHV WKH SURFHVV PRGHO 7KH &RPSXWHU 3URJUDPV 7KH VWDWLVWLFDO SDFNDJH XVHG IRU FRPSXWLQJ WKH GDWD LQ WKLV VWXG\ ZDV WKH %LRPHGLFDO &RPSXWHU 3URJUDPV NQRZQ DV %0' 7KH UHFHQW UHYLVLRQ RI WKHVH SURJUDPV ZDV NQRZQ DV %0'3 'L[RQ DQG %URZQ f 7KH %0'3 SURJUDP SDFNDJH DOORZLQJ WKH XVH RI VHYHUDO GHSHQGHQW YDULDEOHV DQG D ODUJH QXPEHU RI LQGHSHQGHQW YDULDEOHV RIIHUHG WKH WZR SURJUDPV SHUWLQHQW WR WKLV VWXG\ VWHSZLVH PXOWLSOH UHJUHVVLRQ 35 DQG VWHSZLVH GLVFULPLQDQW DQDO\VLV 30 %RWK RI WKHVH DQDO\VHV XVHG WKH IRUZDUG YDULDEOH VHOHFW SURFHGXUH 7KLV ZDV GHVFULEHG E\ /DSLQ f DV D VHTXHQFH RI PXOWLSOH OLQHDU UHJUHVVLRQ HTXDWLRQV LQ D VWHSZLVH PDQQHU $W HDFK VWHS RQH YDULDEOH LV DGGHG WR WKH UHJUHVVLRQ HTXDWLRQ )LUVW WKH SURJUDP VHOHFWV WKH LQGHSHQGHQW YDULDEOH ;Af ZKLFK DFFRXQWV IRU WKH PRVW PRYHPHQW LQ WKH GHSHQGHQW YDULDEOH < 7KHQ WKH QH[W LQGHSHQGHQW YDULDEOH LV ;f DGGHG WR WKH PRGHO 7KH SURJUDP FRQWLQXHV LQ VXFFHVn VLYHO\ KLJKHU GLPHQVLRQV XQWLO WKHUH LV LQVXIILFLHQW B) UDWLR WR HQWHU WKH UHJUHVVLRQ IXQFWLRQ 7KH B) UDWLR IRU VLJQLILFDQFH RI WKH UHJUHVVLRQ FRHIILFLHQW ZDV VHW DW ,I WKH VWDWLVWLFDO FULWHULD ZHUH QRW PHW WKH YDULDEOH ZDV QRW HQWHUHG LQWR WKH DQDO\VLV 6HSDUDWH UHJUHVVLRQ DQDO\VLV ZDV GRQH WR LVRODWH IURP D VHW RI SUHGLFWRU YDULDEOHV D VXEVHW ZKLFK ZRXOG \LHOG DQ RSWLPDO SUHGLFWLYH HTXDWLRQ ZLWK WKH IHZHVW SRVVLEOH WHUPV WR SUHGLFW HDFK RI WKH IRXU GHSHQGHQW YDULDEOHV f 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ 6FRUH f 3V\FKLDWU\ ,QWHUQVKLS 6FRUH f 3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLS 6FRUH DQG f :RUN 6WDWXV 6FRUH

PAGE 55

2XWSXW IURP VWHSZLVH UHJUHVVLRQ DQDO\VLV %0'35f LQFOXGHV DW HDFK VWHS 0XOWLSOH 5 6WDQGDUG HUURU RI HVWLPDWH $QDO\VLVRIYDULDQFH WDEOH )RU YDULDEOHV LQ WKH HTXDWLRQ D 5HJUHVVLRQ FRHIILFLHQW E 6WDQGDUG HUURU F 7ROHUDQFH G B) WR UHPRYH )RU YDULDEOHV QRW LQ WKH HTXDWLRQ D 7ROHUDQFH E 3DUWLDO FRUUHODWLRQ FRHIILFLHQW F ) WR HQWHU 7LHGHPDQ f H[SODLQHG WKDW GLVFULPLQDQW DQDO\VLV GHWHUPLQHV ZKHWKHU GLVFULPLQDWLRQ EHWZHHQ JURXSV LV SRVVLEOH RU QRW DQG WKHQ UHGXFHV WKH VL]H RI WKH VSDFH LQ ZKLFK LW LV QHFHVVDU\ WR WKLQN DERXW UHJLRQV RI FODVVLILFDWLRQ 0RUJDQ S f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f FRPSOHWHQRW FRPSOHWH DQG f ZRUNQRW

PAGE 56

ZRUN 2XWSXW IURP VWHSZLVH GLVFULPLQDQW DQDO\VLV SURJUDP %0'30f LQFOXGHV DW HDFK VWHS :LONnV /DPEGD $SSUR[LPDWH B) VWDWLVWLF 'HJUHHV RI IUHHGRP &RQVWDQW )RU YDULDEOHV LQ WKH HTXDWLRQ D ) WR UHPRYH E 7ROHUDQFH )RU YDULDEOHV QRW LQ WKH HTXDWLRQ D e WR HQWHU E 7ROHUDQFH 6WHSZLVH 0XOWLSOH 5HJUHVVLRQ $QDO\VLV (DFK PRGHO KDV EHHQ GHVFULEHG WKURXJK LWV GHYHORSPHQW DW HDFK VWHS $Q HTXDWLRQ ZDV VKRZQ IRU HDFK VWHS DQG VRPH H[DPSOHV RI XVLQJ WKH ILQDO HTXDWLRQ ZDV VKRZQ WKURXJK WKH UDQGRPO\ FKRVHQ VDPSOHV 0RGHO ,f§1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ 6FRUH 5(*,67f 7KH ILUVW VWHS LQ WKH PRGHO DFFRXQWHG IRU VL[ SHUFHQW RI WKH PRYHPHQW RU D PXOWLSOH 5 RI 7KLV HTXDWLRQ ZDV DV IROORZV < D E; < WKH 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ 6FRUH 5(*,67f D WKH FRQVWDQW DV FRPSXWHG DW HDFK VWHS E WKH FRHIILFLHQW DV FRPSXWHG IRU WKH LQGHSHQGHQW YDULDEOH DW HDFK VWHS

PAGE 57

; WKH LQGHSHQGHQW YDULDEOHf§LQ WKLV FDVH QXPEHU WRWDO VFRUH RQ &((% 7KHUHIRUH 5(*,67 ; :KHQ WKH VHFRQG YDULDEOH ZDV DGGHG QXPEHU QXPEHU RI KRXUV LQ (QJOLVK DQG OLWHUDWXUH (1*/7f WKH WZR \LHOGHG D PXOWLSOH 5 RI 7KH HTXDWLRQ IRU WKH VHFRQG VWHS ZDV < D EA; A!A 5(*,67 ; ; 7KH PRGHO ZDV FXW RII DW WKH VHFRQG VWHS EHFDXVH WKHUH ZDV LQVXIILFLHQW ) UDWLR WR FRQWLQXH WKH VWHSSLQJ SURFHVV 7DEOH VKRZV WKH YDULDEOHV HQWHUHG E\ VWHS SHUFHQW RI PRYHPHQW DW HDFK VWHS WKH RYHUDOO ) UDWLR DW HDFK VWHS 7KH SUHGLFWLRQ HTXDWLRQ GHYHORSHG E\ WKH VWHS ZLVH PXOWLSOH UHJUHVVLRQ DQDO\VLV IRU WKH VFRUH RQ WKH 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ DV VKRZQ LQ VWHS WZR GHVFULSWLRQ JDYH VWDWLVWLn FDOO\ VLJQLILFDQW UHVXOWV EXW WKH UHDGHU VKRXOG IRFXV RQ WKH IDFW WKDW 7DEOH 0RGHO ,f§9DULDEOHV (QWHUHG E\ 6WHS 3HUFHQW RI 'HSHQGHQW ,QGHSHQGHQW 9DULDEOH 0RYHPHQW DW 2YHUDOO 9DULDEOH 9DULDEOH 1XPEHU (DFK 6WHS ) 1DWLRQDO 5HJLVWU\ &((% ([DPLQDWLRQ (1*/7 f WKH PRGHO H[SODLQV RQO\ HLJKW SHUFHQW RI WKH PRYHPHQW LQ WKH 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ VFRUHV 7KDW PHDQV WKDW SHUFHQW RI WKH PRYHPHQW LQ VFRUHV ZHUH QRW DFFRXQWHG IRU

PAGE 58

1XOO K\SRWKHVLV ZDV UHMHFWHG VLQFH LW ZDV VWDWLVWLFDOO\ SRVVLEOH WR SUHGLFW VFRUHV RQ WKH RFFXSDWLRQDO WKHUDS\ 1DWLRQDO 5HJLVWU\ ([DPLQDn WLRQ RQ WKH EDVLV RI PXOWLSOH FRPELQDWLRQV RI WKH REMHFWLYH YDULDEOHV XVHG LQ WKLV VWXG\ 7KH ZHDNQHVV RI WKH SUHGLFWLRQ DFFXUDF\ RI 0RGHO LV VKRZQ LQ 7DEOH XVLQJ D UDQGRP VDPSOH RI VXEMHFWVn DFWXDO VFRUHV IRU H[DPSOHV 7DEOH 0RGHO ,f§6FRUH &RPSDULVRQ IRU 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ$FWXDO DQG 3UHGLFWHG 6XEMHFW 1XPEHU $FWXDO 5HJLVWU\ &((% (1*/7 3UHGLFWHG 6FRUH 0RGHO ,,f§3V\FKLDWULF ,QWHUQVKLS 6FRUH 3<&+,1f 7KH ILUVW VWHS LQ WKH PRGHO DFFRXQWHG IRU ILYH SHUFHQW RI WKH PRYHPHQW RU D PXOWLSOH 5 RI 7KH ILUVW VLJQLILFDQW YDULDEOH ZDV QXPEHU 8QLYHUVLW\ RI )ORULGD IUHVKPDQ RU WUDQVIHU VWXGHQW 75$16f 7KH HTXDWLRQ ZDV < D E; 3<&+,1 ; 7KH VHFRQG PRVW VLJQLILFDQW YDULDEOH ZDV JUDGH SRLQW DYHUDJH RI WKH ILUVW WZR \HDUV RI FROOHJH QXPEHU *3$f $GGLQJ WKLV WR WKH HTXD WLRQ EURXJKW WKH PXOWLSOH 5 WR 7KH HTXDWLRQ DW WKLV SRLQW ORRNHG

PAGE 59

OLNH WKLV < D EL< E; 3<&+,1 ;@> ; $W WKH WKLUG VWHS WKH QXPEHU RI KRXUV LQ (QJOLVK DQG OLWHUDWXUH QXPEHU (1*/7f EURXJKW WKH PRYHPHQW XS WR SHUFHQW 7KH HTXDWLRQ IROORZV < D E; E; EA 3<&+,1 ; ; V; 7KH QXPEHU RI FUHGLW KRXUV LQ SV\FKRORJ\ QXPEHU 36<&+f ZDV VLJQLILFDQW DW WKH IRXUWK VWHS JLYLQJ D PXOWLSOH 5 RI 7KH SUHGLFWLYH HTXDWLRQ GHYHORSHG IRU WKH VFRUH LQ SV\FKLDWULF LQWHUQVKLS ZDV DV IROORZV < D E ;f Ef;f E; E; 3<&+,1 ; ; ; 6; $IWHU WKLV VWHS WKH B) UDWLR ZDV QRW VXIILFLHQW IRU IXUWKHU VWHSSLQJ VHH 7DEOH f 1XOO K\SRWKHVLV ,, ZDV UHMHFWHG VLQFH LW ZDV VWDWLVn WLFDOO\ SRVVLEOH WR SUHGLFW VFRUHV LQ SV\FKLDWU\ LQWHUQVKLS RQ WKH EDVLV 7DEOH 0RGHO ,,f§9DULDEOHV (QWHUHG E\ 6WHS 'HSHQGHQW 9DULDEOH ,QGHSHQGHQW 9DULDEOH 9DULDEOH 1XPEHU 3HUFHQW RI 0RYHPHQW DW (DFK 6WHS 2YHUDOO ) 3V\FKLDWU\ 75$16 ,QWHUQVKLS *3$ (1&/7 36<&+

PAGE 60

RI PXOWLSOH FRPELQDWLRQV RI REMHFWLYH YDULDEOHV XVHG LQ WKLV VWXG\ &DUH VKRXOG EH WDNHQ LQ XVH RI WKLV PRGHO EHFDXVH RQO\ SHUFHQW RI WKH PRYHPHQW LQ 3V\FKLDWU\ ,QWHUQVKLS VFRUHV ZDV H[SODLQHG E\ WKH VWDWLVWLFDOO\ VLJQLILFDQW YDULDEOHV OHDYLQJ SHUFHQW XQDFFRXQWDEOH %\ H[DPLQLQJ 7DEOH IRU D UDQGRP VDPSOH RI WKLV HTXDWLRQ LQ XVH RQH FDQ VHH WKDW LW GRHV QRW ZRUN ZHOO LQ SUDFWLFDO DSSOLFDWLRQ 7DEOH 0RGHO ,,f§6FRUH &RPSDULVRQ IRU 3V\FKLDWULF ,QWHUQVKLS$FWXDO DQG 3UHGLFWHG 6XEM HFW 1XPEHU 3V\FKLDWU\ ,QWHUQVKLS $FWXDO 6FRUH 75$16 *3$ (1*/7 36<&+ 3UHGLFWHG 6FRUH 0RGHO ,,, f§3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLS 6FRUH 3',17f 9DULDEOH QXPEHU 8QLYHUVLW\ RI )ORULGD IUHVKPDQ RU WUDQVIHU VWXGHQW ZDV WKH ILUVW VLJQLILFDQW YDULDEOH ZLWK D PXOWLSOH 5 RI 7KH HTXDWLRQ DSSHDUHG DV < D E; 3',17 O2A6;AA 7KH VHFRQG VWHS LQFOXGHG LQ JUDGH SRLQW DYHUDJH RI WKH ILUVW WZR \HDUV RI FROOHJH QXPEHU UDLVLQJ WKH 5 WR 7KH PRGHO WHUPLQDWHG DW WKLV VWHS EHFDXVH RI LQVXIILFLHQW ) UDWLR VHH 7DEOH f

PAGE 61

7DEOH 0RGHO ,,,f§9DULDEOHV (QWHUHG E\ 6WHS 3HUFHQW RI 'HSHQGHQW ,QGHSHQGHQW 9DULDEOH 0RYHPHQW 2YHUDOO 9DULDEOH 9DULDEOH 1XPEHU DW (DFK 6WHS ) 3K\VLFDO 'LVDELOLWLHV 75$16 ,QWHUQVKLS *3$ 7KH SUHGLFWLRQ HTXDWLRQ IRU VFRUHV LQ SK\VLFDO GLVDELOLWLHV LQWHUQVKLS IROORZV < D E ; E; 3',17 ; ; $JDLQ WKH DPRXQW RI PRYHPHQW DFFRXQWHG IRU LV YHU\ VPDOO HLJKW SHUFHQW 7KHUHIRUH WKH UHDGHU VKRXOG H[HUFLVH FDUH LQ LQWHUSUHWLQJ WKH XVH RI WKLV PRGHO 1XOO K\SRWKHVLV ,,, ZDV UHMHFWHG VLQFH LW ZDV VWDWLVWLFDOO\ SRVVLEOH WR SUHGLFW VFRUHV LQ SK\VLFDO GLVDELOLWLHV LQWHUQVKLS RQ WKH EDVLV RI PXOWLSOH FRPELQDWLRQV RI REMHFWLYH YDULDEOHV XVHG LQ WKLV VWXG\ 7DEOH VKRZV WKLV PRGHO GHPRQVWUDWHG E\ D UDQGRP VDPSOH RI VXEMHFWVn GDWD 0RGHO ,9:RUN 6WDWXV 6FRUH 7KLV PRGHO FRPSXWHG RQO\ RQH VWHS ZKLFK DFFRXQWHG IRU RQO\ WKUHH SHUFHQW RI WKH PRYHPHQW LQ ZRUN VWDWXV VFRUHV 7DEOH f 7KH YDULDEOH RI VLJQLILFDQFH ZDV &((% QXPEHU 7KH ILUVW DQG ILQDO VWHS VKRZHG WKH IROORZLQJ HTXDWLRQ < D E; :.67$ ;

PAGE 62

7DEOH 0RGHO ,,,f§6FRUH &RPSDULVRQ IRU 3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLS$FWXDO DQG 3UHGLFWHG 6XEM HFW 1XPEHU $FWXDO 3K\V 'LV ,QWHUQVKLS 75$16 *3$ )LUVW 7ZR
PAGE 63

7DEOH 0RGHO ,9f§6FRUH &RPSDULVRQ IRU :RUN 6WDWXV $FWXDO DQG 3UHGLFWHG 6XEM HFW 1XPEHU $FWXDO :RUN 6WDWXV &((% 3UHGLFWHG 6FRUH IRU :RUN 6WDWXV 6WHSZLVH 'LVFULPLQDWLRQ $QDO\VLV 0RGHO 9f§ &RPSOHWH DQG 1RW &RPSOHWH WKH 3URJUDP IRU 5HJLVWU\ LQ 2FFXSDWLRQDO 7KHUDS\ 7KH DQDO\VLV RI WKH GDWD GHYHORSHG D SUHGLFWLRQ IRU WKH GLFKRWRP\ RI FRPSOHWLQJ RU QRW FRPSOHWLQJ WKH SURJUDP IRU UHJLVWUDWLRQ LQ RFFXSDWLRQDO WKHUDS\ $V 6KDOLN f V\QWKHVL]HG 7KH PHWKRG XVHG IRU WKLV DQDO\VLV LV WKH :LONnV ODPEGD 7KH FULWHULRQ ZDV WKH RYHUDOO PXOWLYDULDWH B) UDWLR IRU WKH WHVW RI GLIIHUHQFHV DPRQJ JURXS FHQWURLGV 7KH YDULDEOH ZKLFK PD[LPL]HG WKH B) UDWLR DOVR PLQLPL]HG :LONnV ODPEGD D PHDVXUH RI JURXS GLVFULPLQDWLRQ *URXS DVVLJQPHQW ZDV EDVHG RQ WKH GLVWDQFH IURP WKH FHQWURLG ZLWK WKH DFWXDO JURXS DVVLJQPHQW GHWHUPLQHG E\ WKH FORVHVW GLVWDQFH WR RQH RI WKH FHQWURLGV S f 7DEOH VKRZV WKH FODVVLILFDWLRQ IXQFWLRQV 7KLV GLVFULPLQDQW IXQFWLRQ FRUUHFWO\ FODVVLILHG SHUFHQW RI WKH VXEMHFWV LQWR WKH FRPSOHWLRQ JURXS DQG SHUFHQW ZHUH LQFRUUHFWO\ FODVVLILHG &RUUHFW FODVVLILFDWLRQ LQWR WKH QRQFRPSOHWLRQ JURXS ZDV

PAGE 64

SHUFHQW ZLWK DQ HUURU RI SHUFHQW $ WRWDO DYHUDJH IRU FRUUHFW SODFHPHQW RI WKH VXEMHFWV ZDV SHUFHQW 7KLV LV D YHU\ ORZ SHUFHQWDJH VLQFH WKH FRUUHFW SUHGLFWLRQ E\ FKDQFH ZRXOG EH SHUFHQW DV LOOXVWUDWHG LQ 7DEOH WKH FODVVLILFDWLRQ PDWUL[ 7KH MDFNQLIHG FODVVLILFDWLRQ LV D SURFHGXUH RI WKH FURVV YDOLGDWLRQ W\SH VHH 7DEOH f 7KH MDFNQLIHG FODVVLILFDWLRQ WDEOH LQGLFDWHV DJDLQ WKDW WKH FKDQFH UDWH RI FRUUHFW FODVVLILFDWLRQ ZDV SHUFHQW ZKLOH WKH DFWXDO FRUUHFW FODVVLILFDWLRQ UDWH ZDV RQO\ SHUFHQW 7DEOH 0RGHO 9 DQG 9,f§&ODVVLILFDWLRQ )XQFWLRQV 9DULDEOH *URXS &RPSOHWH 'LG 1RW &RPSOHWH 1XPEHU 75$16 (1*/7 36<&+ *3$ 0RGHO 9 DQG 7DEOH 9,f§&ODVVLILFDWLRQ 0DWUL[ *URXS 3HUFHQW &RUUHFW 1XPEHU RI &DVHV &RPSOHWH &ODVVLILHG LQWR *URXS 'LG 1RW &RPSOHWH &RPSOHWH 1RW &RPSOHWH 7RWDO

PAGE 65

7DEOH 0RGHOV 9 DQG 9,f§-DFNQLIHG &ODVVLILFDWLRQ *URXS 3HUFHQW &RUUHFW 1XPEHU RI &DVHV &RPSOHWH &ODVVLILHG LQWR *URXS 'LG 1RW &RPSOHWH &RPSOHWH 1RW &RPSOHWH 7RWDO 7KH ILUVW VLJQLILFDQW YDULDEOH HQWHUHG ZDV QXPEHU 8QLYHUVLW\ RI )ORULGD IUHVKPDQ RU WUDQVIHU VWXGHQW 75$16f VHFRQG ZDV WKH YDULDEOH QXPEHU QXPEHU RI KRXUV LQ SV\FKRORJ\ 36<&+f WKLUG QXPEHU *3$ IRU WKH ILUVW WZR \HDUV RI FROOHJH *3$f DQG IRXUWK QXPEHU QXPEHU RI KRXUV LQ (QJOLVK DQG OLWHUDWXUH (1*/7f 7KHVH IRXU YDULDEOHV DUH WKH VDPH RQHV WKDW HQWHUHG LQWR WKH PXOWLSOH UHJUHVVLRQ DQDO\VLV IRU SUHGLFWLQJ VFRUHV LQ SV\FKLDWU\ LQWHUQVKLS 1XOO K\SRWKHVHV 9 DQG 9, ZHUH ERWK UHMHFWHG VLQFH LW ZDV VWDWLVn WLFDOO\ SRVVLEOH WR GLVFULPLQDWH DPRQJ JURXSV RI VWXGHQWV WKRVH ZKR FRPSOHWHG 9f DQG WKRVH ZKR GLG QRW FRPSOHWH 9,f WKH RFFXSDWLRQDO WKHUDS\ SURJUDP RQ WKH EDVLV RI OLQHDU FRPELQDWLRQV RI REMHFWLYH YDULDEOHV XVHG LQ WKLV VWXG\ 7KH VDPH EDVLF HTXDWLRQ DSSOLHG WR GLVFULPLQDQW DQDO\VLV DV WR PXOWLSOH UHJUHVVLRQ DQDO\VLV < D E ; E ; EB;B E; 9 &RPSOHWH ; 75$16f ; 36<&+f ; *3$f ; (1*/7f

PAGE 66

9, 1R &RPSOHWH ; 75$16f ; 36<&+f ; *3$f ; (1*/7f 7KH UHDGHU FDQ UHIHU WR 7DEOH IRU D VXPPDU\ RI YDULDEOHV HQWHUHG E\ VWHS IRU FRPSOHWLRQ RI WKH RFFXSDWLRQDO WKHUDS\ SURJUDP 7DEOH VKRZV WKH DFWXDO DQG SUHGLFWHG VFRUHV XVLQJ D UDQGRP VDPSOH RI GDWD 7DEOH 0RGHOV 9 DQG 9,f§6XPPDU\9DULDEOHV (QWHUHG E\ 6WHS 6WHS 1XPEHU 9DULDEOH (QWHUHG ) UDWLR 2YHUDOO ) 'HJUHHV RI )UHHGRP 75$16 36<&+ *3$ (1*/7 7DEOH 0RGHO 9f§6FRUH &RPSDULVRQ IRU &RPSOH W HG$F WXD DQG 3UHGLFWHG 6XEMHFW 1XPEHU $FWXDO 75$16 36<&+ *3$ (1*/7 3UHGLFWHG 0RGHO 9,f§1R FRPSOHWLRQ RI RFFXSDWLRQDO WKHUDS\ SURJUDP LV WKH RWKHU KDOI RI WKH GLFKRWRP\ 7DEOH VKRZV WKDW WKLV HTXDWLRQ ODFNV XVHIXOQHVV ZKHQ DSSOLHG WR D SUDFWLFDO VLWXDWLRQ

PAGE 67

7DEOH 0RGHO 9,f§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f

PAGE 68

7DEOH 0RGHOV 9,, DQG 9,,,f§&ODVVLILFDWLRQ )XQFWLRQV 9DULDEOH *URXS :RUN 1R :RUN 2YHUDOO ) 'HJUHHV )UHHGRP RI 0XVLF &RQVWDQW 7DEOH 0RGHOV 9,, DQG 9,,,f§&ODVVLILFDWLRQ 0DWUL[ *URXS 3HUFHQW &RUUHFW 1XPEHU RI &DVHV &ODVVLILHG LQWR :RUN 1R :RUN *URXSV :RUN 1R :RUN 7RWDO 7DEOH 0RGHOV 9,, DQG 9,,,f§-DFNQLIHG &ODVVLILFDWLRQ *URXS 3HUFHQW &RUUHFW 1XPEHU RI &DVHV &ODVVLILHG LQWR :RUN 1R :RUN *URXS :RUN 1R :RUN 7RWDO 7KH ILUVW DQG RQO\ VLJQLILFDQW YDULDEOH WR HQWHU ZDV QXPEHU QXPEHU RI KRXUV LQ PXVLF

PAGE 69

1XOO K\SRWKHVHV 9,, DQG 9,,, ZHUH UHMHFWHG VLQFH LW ZDV VWDWLVWLFDOO\ SRVVLEOH WR GLVFULPLQDWH DPRQJ JURXSV RI VWXGHQWV 9,,f WKRVH ZKR DUH ZRUNLQJ DQG 9,,,f WKRVH ZKR DUH QRW ZRUNLQJ LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ IRXU WR HLJKW \HDUV DIWHU JUDGXDWLRQ RQ WKH EDVLV RI OLQHDU FRPELQDWLRQV RI REMHFWLYH YDULDEOHV XVHG LQ WKLV VWXG\ < D E; :RUN 086,&f 1R :RUN 086,&f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

PAGE 70

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

PAGE 71

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fWKH 1DWLRQDO 5HJLVWU\ ([DPLQDWLRQ ,,f3V\FKLDWU\ LQWHUQVKLS ,,,f3K\VLFDO GLVDELOLWLHV LQWHUQVKLS ,9f:RUN VWDWXV

PAGE 72

RQ WKH EDVLV RI PXOWLSOH FRPELQDWLRQV RI YDULDEOHV ZKLFK LQFOXGH FROOHJH HQWU\ WHVW VFRUHV FRXUVHV WDNHQ JUDGH SRLQW DYHUDJHV DQG RWKHU REMHFWLYH GDWD WDNHQ IURP WUDQVFULSWV 7KH VHFRQG SDUW XVHG VWHSZLVH GLVFULPLQDQW DQDO\VLV WR ILQG WKH EHVW LQGLFDWRUV RI FRPSOHWLRQ DQG QRQFRPSOHWLRQ RI WKH RFFXSDWLRQDO WKHUDS\ SURJUDP DQG WKH EHVW LQGLFDWRUV IRU ZRUN SHUVLVWHQFH DQG QRQn SHUVLVWHQFH LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ XVLQJ %LRPHGLFDO FRPSXWHU SDFNDJHG SURJUDP 36HULHV %0'30 7KH IRXU PXOWLYDULDWH QXOO K\SRWKHVHV SURSRVHG IRU WKH GLVFULPLQDQW DQDO\VLV ZHUH DOO VWDWLVWLFDOO\ UHMHFWHG ,W ZDV QRW SRVVLEOH WR GLVFULPLQDWH DPRQJ JURXSV RI VWXGHQWV 9f WKRVH ZKR FRPSOHWHG DQG 9,f WKRVH ZKR GLG QRW FRPSOHWH WKH RFFXSDWLRQDO WKHUDS\ SURJUDP RQ WKH EDVLV RI OLQHDU FRPELQDWLRQV RI YDULDEOHV ZKLFK LQFOXGH FROOHJH HQWU\ WHVW VFRUHV FRXUVHV WDNHQ JUDGH SRLQW DYHUDJHV DQG RWKHU REMHFWLYH GDWD WDNHQ IURP WUDQVFULSWV ,W ZDV QRW SRVVLEOH WR GLVFULPLQDWH DPRQJ JURXSV RI VWXGHQWV 9,,f WKRVH ZKR ZHUH ZRUNLQJ DQG 9,,,f WKRVH ZKR ZHUH QRW ZRUNLQJ LQ WKH ILHOG RI RFFXSDWLRQDO WKHUDS\ IRXU WR HLJKW \HDUV ODWHU RQ WKH EDVLV RI OLQHDU FRPELQDWLRQV RI YDULDEOHV ZKLFK LQFOXGH FROOHJH HQWU\ WHVW VFRUHV FRXUVHV WDNHQ JUDGH SRLQW DYHUDJHV DQG RWKHU REMHFWLYH GDWD WDNHQ IURP WUDQVFULSWV

PAGE 73

5HVXOWV RI WKH PXOWLSOH UHJUHVVLRQ DQDO\VLV RI WKH YDULDEOHV VKRZHG ILYH WR EH VWDWLVWLFDOO\ VLJQLILFDQW DW WKH OHYHO FROOHJH HQWU\ WHVW VFRUH FROOHJH JUDGH SRLQW DYHUDJH IRU WKH ILUVW WZR \HDUV WUDQVIHU VWXGHQW RU HQWU\ DV D 8QLYHUVLW\ RI )ORULGD IUHVKPDQ QXPEHU RI FUHGLW KRXUV LQ (QJOLVK DQG OLWHUDWXUH QXPEHU RI FUHGLW KRXUV LQ SV\FKRORJ\ 7KH GLVFULPLQDQW DQDO\VLV XVHG DOO EXW RQH RI WKH VDPH YDULDEOHV &((%f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n DFWXDO GDWD XVHG LQ HDFK SUHGLFWLYH HTXDWLRQ 7KH DFFXUDF\ RI WKH GLVFULPLQDQW DQDO\VLV ZDV IDU OHVV WKDQ PLJKW KDYH RFFXUUHG E\ FKDQFH

PAGE 74

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

PAGE 75

DPRXQW RI PRYHPHQW WKDW FRXOG EH DFFRXQWHG IRU ZDV WRR VPDOO WR EH RI SUDFWLFDO XVH 7KH UHVXOWV RI WKH VWXG\ LQGLFDWHG WKDW FRPELQLQJ ERWK DFDGHPLF DQG SUDFWLFDO DSSOLFDWLRQ SDUDPHWHUV LQ ERWK WKH LQGHSHQGHQW DQG WKH GHSHQGHQW YDUL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f VDLG WKDW SHUFHQW RI RFFXSDWLRQDO WKHUDSLVWV TXLW ZRUNLQJ LQ WKHLU ILUVW ILYH \HDUV RI HPSOR\PHQW +RZHYHU -DQW]HQ f SRLQWHG RXW WKDW

PAGE 76

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f ([SHFWDWLRQV IRU QHZ JUDGXDWHV PLJKW EH WRR KLJK FRQVLGHULQJ WKDW RFFXSDWLRQDO WKHUDS\ LV PRVWO\ D IHPDOH ILHOG ZLWK WKH W\SLFDO WKHUDSLVW EHLQJ PDUULHG DJHG DQG KDYLQJ RQH RU QR FKLOGUHQ -DQW]HQ f 7KH WLPH VSDQ RI ILYH WR WHQ \HDUV DIWHU JUDGXDWLRQ ZDV GXULQJ WKH SULPH FKLOG EHDULQJ \HDUV $ VWXG\ IURP WKH SK\VLFDO WKHUDS\ OLWHUDWXUH VKRZV DQ HPSOR\PHQW SDWWHUQ VLPLODU WR RFFXSDWLRQDO WKHUDS\ )LYH \HDUV DIWHU JUDGXDWLRQ IURP D SK\VLFDO WKHUDS\ FXUULFXOXP GLG QRW VKRZ D KLJKHU SHUFHQW SHUFHQWf RI JUDGXDWHV HPSOR\HG WKDQ DW RQH \HDU DIWHU JUDGXDWLRQ SHUFHQWf &\SUHV f -RKQVRQ f SRLQWHG RXW WKDW SHUFHQW RI WKRVH WKHUDSLVWV ZKR GURS RXW RI HPSOR\PHQW ZLWKLQ WKH ILUVW ILYH \HDUV UHWXUQ DIWHU \HDUV $ IXWXUH VWXG\ RI ZRUN VWDWXV WR \HDUV DIWHU JUDGXDWLRQ PLJKW UHYHDO PRUH DFFXUDWH LQIRUPDWLRQ RQ SHUVLVWHQFH LQ HPSOR\PHQW RI RFFXSDWLRQDO WKHUDSLVWV 7KH OLWHUDWXUH VKRZHG WKDW SUHYLRXV DFDGHPLF SHUIRUPDQFH DQG DFKLHYHPHQW ZHUH WKH EHVW LQGLFDWRUV RI ODWHU VXFFHVV LQ VLPLODU DUHDV 7KLV VWXG\ FRQFXUUHG ZLWK WKHVH ILQGLQJV

PAGE 77

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

PAGE 78

%,%/,2*5$3+< $FTXDYLYD )$ $27$ +XPDQ 5HVRXUFHV 3URMHFW ,,, 7KH 0HPEHU 'DWD 6XUYH\ $P 27 f 0D\-XQH / $FTXDYLYD )$ 7KH 0HPEHU 'DWD 6XUYH\ ,9f§'HPRJUDSKLF &KDUDFWHULVWLFV RI 2FFXSDWLRQDO 7KHUDS\ 3HUVRQQHO $P 27 f $XJ $n+HDUQ 57 $ 0RGHO IRU 3UHGLFWLQJ $FDGHPLF $FKLHYHPHQW RI 7HFKQLFDO &ROOHJH 6WXGHQWV 'RFWRUDO 'LVVHUW 0DUTXHWWH 8QLY $LNHQ /5 -U 7KUHH $OLJQPHQW &KDUWV IRU 8VH LQ 6HOHFWLYH $GPLVVLRQV (G 5HV 6HSW $OGDJ )& DQG 0DUWLQ 0) 3K\VLFDO 7KHUDSLVW $VVLVWDQW 6HOHFWLRQ DQG $FDGHPLF 6XFFHVV 3K\V 7KHU f $27$ 2FF 7KHU 1HZOHWWHU f -DQ $27$ 2FFXSDWLRQDO 7KHUDS\ $' 3DSHUV SUHVHQWHG DW WKH VSHFLDO VHVVLRQ RI WKH 5HSUHVHQWDWLYH $VVHPEO\ 1RY $27$ 5RFNYLOOH 0DU\ODQG $QGHUVRQ +( DQG -DQW]HQ $& $ 3UHGLFWLRQ RI &OLQLFDO 3HUIRUPDQFH $P 27 f $\HUV -% DQG 5RKU 0( 5HODWLRQVKLS RI 6HOHFWHG 9DULDEOHV WR 6XFFHVV LQ D 7HDFKHU 3UHSDUDWLRQ 3URJUDP (G 3V\FK 0HDV %DLOH\ -3 6RPH &RPSDULVRQV RI 2FFXSDWLRQDO 7KHUDS\ %LRJUDSKLFDO )DFWV ZLWK ,PSOLFDWLRQV IRU 5HFUXLWLQJ $P 27 f %DLOH\ -3 -DQW]HQ $& DQG 'XQWHPDQ *+ 5HODWLYH (IIHFWLYHQHVV RI 3HUVRQDOLW\ $FKLHYHPHQW DQG ,QWHUHVW 0HDVXUHV LQ WKH 3UHGLFWLRQ RI D 3HUIRUPDQFH &ULWHULRQ $P 27 f %DNHU ( -U $VVRFLDWH 'HJUHH 1XUVLQJ 6WXGHQWV 1RQLQWHOOHFWLYH 'LIn IHUHQFHV EHWZHHQ 'URSRXWV DQG *UDGXDWHV 1XUV 5HV f -DQ)HE %HDQ $* 7KH 3UHGLFWLRQ RI 3HUIRUPDQFH LQ DQ (GXFDWLRQDO 3V\FKRORJ\ 0DVWHUnV 'HJUHH 3URJUDP (G t 3V\FK 0HDV f

PAGE 79

%HDQ $* DQG &RYHUW 5: 3UHGLFWLRQ RI &ROOHJH 3HUVLVWHQFH :LWKGUDZDO DQG $FDGHPLF 'LVPLVVDO $ 'LVFULPLQDQW $QDO\VLV (G t 3V\FK 0HDV %HOFDVWUR )3 8VH RI 6HOHFWHG )DFWRUV DV 3UHGLFWRUV RI 6XFFHVV LQ &RPSOHWLQJ D 6HFRQGDU\ 7HDFKHU 3UHSDUDWLRQ 3URJUDP (G t 3V\FK 0HDV %HUJPDQ 5 (GHOVWHLQ $ 5RWHQEHUJ $ DQG 0HODPHG < 3V\FKRORJLFDO 7HVWV 7KHLU 8VH DQG 9DOLGLW\ LQ 6HOHFWLQJ &DQGLGDWHV IRU 6FKRROV RI 1XUVLQJ LQ ,VUDHO ,QW 1XUV 6W %HUQVWHLQ / 0HGLFDUH DQG WKH 2FFXSDWLRQDO 7KHUDSLVW $P 27 f %LVWUHLFK $0 3UHGLFWLQJ *UDGH 3RLQW $YHUDJH :LWKGUDZDO DQG *UDGXDWLRQ IURP )RXU $OOLHG +HDOWK 3URJUDPV DW 0LDPL'DGH &RPPXQLW\ &ROOHJH 0HGLFDO &HQWHU &DPSXV 'RFWRUDO 'LVVHUW 1RYD 8QLY %ODLVGHOO ($ -U DQG *RUGRQ 6HOHFWLRQ RI 2FFXSDWLRQDO 7KHUDS\ 6WXGHQWV $P 27 f $SU %OXPH &6 8VH RI D %LRJUDSKLFDO 'DWD ,QYHQWRU\ WR ,GHQWLI\ %HKDYLRUDO &RUUHODWHV RI 3UDFWLFLQJ 0HGLFDO 7HFKQRORJLVWV $Q ([SORUDWRU\ 6WXG\ $OOG +OWK f : %RRWK 0' $ 6WXG\ RI WKH 5HODWLRQVKLS EHWZHHQ &HUWDLQ 3HUVRQDOLW\ )DFWRUV DQG 6XFFHVV LQ &OLQLFDO 7UDLQLQJ RI 2FFXSDWLRQDO 7KHUDS\ 6WXGHQWV $P 27 f %RZHUV DQG /RHE &KDQJHV LQ 3UHGLFWRU :HLJKWV LQ *3$ 5HJUHVVLRQ (TXDWLRQV 0HDV t (YDO *XLG f 2FW %UDQGW (0 &RPSDULVRQ RI 2QWKH-RE 3HUIRUPDQFH RI *UDGXDWHV ZLWK 6FKRRO RI 1XUVLQJ 2EMHFWLYHV 1XUV 5HV Of %UDQGW (0 +DVWLH % DQG 6FKXPDQ 3UHGLFWLQJ 6XFFHVV RQ 6WDWH %RDUG ([DPLQDWLRQV 1XUV 5HV f %UD]]LHO : 1RQLQWHOOHFWLYH 3UHGLFWRUV RI 6WXGHQW 3HUVLVWHQFH $WWULWLRQ IRU &ROOHJH DQG 8QLYHUVLW\ 5HVHDUFK DQG 3ODQQLQJ 3DSHU SUHVHQWHG DW WKH ,QW )RUXP RI WKH $VVRF IRU ,QVWL 5HV 0RQWUHDO 0D\ %UROOLHU & 3HUVRQDOLW\ &KDUDFWHULVWLFV RI 7KUHH $OOLHG +HDOWK 3URIHVn VLRQDO *URXSV $P 27 f @ %URZQ 38 &DUHHU 'HYHORSPHQW DQG 6DWLVIDFWLRQ RI 2FFXSDWLRQDO 7KHUDSLVWV LQ )ORULGD 'RFWRUDO 'LVVHUW 8QLY RI )OD %URZQ 5*6 DQG 6WRQHV 5:+ 3HUVRQDOLW\ DQG ,QWHOOLJHQFH &KDUDFWHULVWLFV RI 0DOH 1XUVHV ,QW 1XUV 6W

PAGE 80

%XFD -7 $Q ,QYHVWLJDWLRQ RI WKH $WWULWLRQ DQG 5HWHQWLRQ RI 3K\VLFDO 7KHUDS\ 6WXGHQWV DW 8QLY RI 0DU\ODQG 8QSXE 7KHVLV 8QLY RI 0DU\n ODQG %DOWLPRUH %XKPH\HU .3UHGLFWLQJ 6XFFHVV LQ D 3K\VLFLDQ ([WHQGHU 0HGH[f 3URJUDP 'RFWRUDO 'LVVHUW :DOGHQ 8QLY -XO\ %XUJHVV 00 7ZR 6WXGLHV RI 3UHGLFWLRQ RI 6XFFHVV LQ D &ROOHJLDWH 3URJUDP RI 1XUVLQJ 1XUV 5HV f -XO\$XJ %XUWRQ '$ 7KH 6HOHFWLRQ RI 1XUVHV E\ 'LVFULPLQDQW $QDO\VLV ,QW 1XUV 6W &KDVH ', DQG -RKQVRQ -3UHGLFWLQJ &ROOHJH 6XFFHVV ZLWK 1RQWUDGL WLRQDO 'DWD IRU ,QQHUFLW\ 6WXGHQWV &RO 6W 3HUV 0D\ &KHQ 0. 3RGVKDGOH\ ': DQG 6KURFN -* $ )DFWRULDO 6WXG\ RI 6RPH 3V\FKRORJLFDO 9RFDWLRQDO ,QWHUHVW DQG 0HQWDO $ELOLW\ 9DULDEOHV DV 3UHGLFWRUV RI 6XFFHVV LQ 'HQWDO 6FKRRO $SSO 3V\FK f &ODSS 77 &RPSDULVRQ RI 7KUHH 0RGHOV IRU 6HOHFWLRQ RI $SSOLFDQWV DQG 3UHGLFWLRQ RI 6XFFHVV LQ 0HGLFDO 6FKRRO 'RFWRUDO 'LVVHUW 8QLY RI 0LVVRXUL &ROXPELD &RQLQH 7$ $ 6XUYH\ RI WKH *UDGXDWHV RI D 3URIHVVLRQDO 3K\VLFDO 7KHUDS\ 3URJUDP 3K\V 7KHU f &RUQHWW -' (IIHFWLYHQHVV RI 7KUHH 6HOHFWLYH $GPLVVLRQV &ULWHULD LQ 3UHGLFWLQJ 3HUIRUPDQFH RI )LUVW
PAGE 81

'L[RQ :HG %LRPHGLFDO &RPSXWHU 3URJUDPV UG HG %HUNHOH\ 8QLYHUVLW\ RI &DOLIRUQLD 3UHVV SS 'L[RQ :DQG %URZQ 0% (GV %0'3 %LRPHGLFDO &RPSXWHU 3URJUDPV 36HULHV 8QLYHUVLW\ RI &DOLIRUQLD 3UHVV SS 'UXJR -* 5HJUHVVLRQ PRGHOV DQG QRQFRJQLWLYH SUHGLFWRUV 'RFWRUDO 'LVVHUW 8QLYHUVLW\ RI 3LWWVEXUJK 'XEV 5 &RPSDULVRQ RI VWXGHQW DFKLHYHPHQW ZLWK SHUIRUPDQFH UDWLQJV RI JUDGXDWHV DQG VWDWH ERDUG H[DPLQDWLRQ VFRUHV 1XUV 5HV f 'XQQLQJ (5 $Q DFFRXQWDELOLW\ PRGHO IRU RFFXSDWLRQDO WKHUDS\ $P 27 f 'XQWHPDQ *+ &KDUDFWHULVWLFV RI VWXGHQWV LQ WKH KHDOWK UHODWHG SURIHVn VLRQV 8QLY RI )OD 5HKDE 5HV 0RQRJUDSK 6HULHV 1XPEHU -XQH (QJHOKDUW +9 $Q LQYHVWLJDWLRQ RI WKH UHODWLRQVKLS EHWZHHQ FROOHJH JUDGHV DQG RQWKHMRE SHUIRUPDQFH GXULQJ FOLQLFDO WUDLQLQJ RI RFFXSDWLRQDO WKHUDS\ VWXGHQWV $P 27 f 3DUW ,, )HGHULFL / DQG 6FKXHUJHU 3UHGLFWLRQ RI VXFFHVV LQ DQ DSSOLHG 0 $ SV\FKRORJ\ SURJUDP (G t 3V\FK 0HDV )HOGPDQ (: &KDUDFWHULVWLFV DVVRFLDWHG ZLWK SHUVLVWHQFH LQ SUHPHGLFDO HGXFDWLRQ DW 'XNH 8QLY $ GLVFULPLQDQW DQDO\VLV 'RFWRUDO 'LVVHUW 'XNH 8QLY )UHGULFNV 0$ 0XQG\ 3 5REHUWVRQ /6 DQG .RVD 7KH QDWLRQDO ERDUG H[DPLQDWLRQ DQG DFDGHPLF DFKLHYHPHQW LQ D PHGLFDO VFKRRO (G 5HV )HE *RXJK +* 6RPH SUHGLFWLYH LPSOLFDWLRQV RI SUHPHGLFDO VFLHQWLILF FRPSHWHQFH DQG SUHIHUHQFHV 0HG (G $SU *RXJK +* DQG +DOO :% 7KH SUHGLFWLRQ RI DFDGHPLF DQG FOLQLFDO SHUIRUPn DQFH LQ PHGLFDO VFKRRO 5HV +LJKHU (G *UHHQ /+ 3UHGLFWLQJ DFDGHPLF VXUYLYDOf§DWWULWLRQ LQ ILUVW VHPHVWHU RI 0DUTXHWWH 8QLYHUVLW\ SK\VLFDO WKHUDS\ FXUULFXOXP 8QSXE 7KHVLV 0LOZDXNHH :LVH -DQ *UHHQ 3( DQG 7XOO '6 5HVHDUFK IRU 0DUNHWLQJ 'HFLVLRQV QG HG 3UHQWLFH+DOO ,QF 1HZ -HUVH\ +DJOXQG $+ 3UHGLFWLQJ VXFFHVV LQ FROOHJLDWH QXUVLQJ SURJUDPV 'RFWRUDO 'LVVHUW 8QLY RI :LVH 0DGLVRQ

PAGE 82

+DOIWHU ,7 9DOLGLW\ VWXGLHV /DZ VFKRRO (G 5HV 0DU +DOO -1 DQG $QGUHZV :7 $SSDUHQW KRPRJHQHLW\ LQ FKDUDFWHULVWLFV RI VWXGHQW QXUVH JURXSV ,QW 1XUV 6W +DOO 5 DQG &RDWHV &3UHGLFWLRQ RI DFDGHPLF VXFFHVV IRU VSHFLDO SURJUDP DQG UHJXODUO\ DGPLWWHG FROOHJH IUHVKPHQ &RO t 8QLY )DOO +HLW 3 -RKQVRQ 0HHNV /% DQG 3D[RQ & $ VHOHFWLYH DGPLVVLRQV SURFHVV DV D SUHGLFWRU RI DFDGHPLF VXFFHVV LQ KHDOWK HGXFDWLRQ 6FK +OWK f +HQGULFNVRQ 3HUVRQDOLW\ YDULDEOHV 6LJQLILFDQW GHSDUWXUHV RI 2FFXSDWLRQDO WKHUDS\ IURP SRSXODWLRQ QRUPV $P 27 f +LQNHO DQG +RXVWRQ &$ $ FRPSDUDWLYH VWXG\ RI WKH HIIHFWLYHQHVV RI WZR %D\HVLDQ PRGHOV IRU SUHGLFWLQJ WKH DFDGHPLF VXFFHVVHV RI VHOHFWHG DOOLHG KHDOWK VWXGHQWV HQUROOHG LQ WKH FRPSUHKHQVLYH FRPPXQLW\ FROOHJH 3DSHU SUHVHQWHG DW WKH $QQXDO 0WJ RI $P (G 5HV $VVRF 1< 1< $SU +ROPVWURP (O 3URPLVLQJ SURVSHFWV 6WXGHQWV FKRRVLQJ WKHUDS\ DV D FDUHHU $P 27 f +ROW /, $ VWXG\ WR LGHQWLI\ SUHGLFWRUV RI VXFFHVV LQ DFDGHPLF SURJUDPV DQG LQ RFFXSDWLRQV IRU VWXGHQWV LQ PHGLFDO ODERUDWRU\ WHFKQLFLDQ SURJUDPV LQ FRPPXQLW\ FROOHJHV 8QSXE 'RFWRUDO 'LVVHUW 8QLY RI &RQQHF +XPSKUH\V /* /HY\ DQG 7DEHU 7 3UHGLFWDELOLW\ RI DFDGHPLF JUDGHV IRU VWXGHQW RI KLJK DQG ORZ DFDGHPLF SURPLVH (G t 3V\FK 0HDV ,QNHQEHUU\ 62 )DFWRUV LQ FROOHJH SHUVLVWHQFH &RXQ 3V\FK f -DQW]HQ $& 6RPH FKDUDFWHULVWLFV RI IHPDOH RFFXSDWLRQDO WKHUDSLVWV $P 27 f -RKQVRQ -$ 1DWLRQDOO\ VSHDNLQJ $P 27 f 1DWLRQDOO\ VSHDNLQJ ,VVXHV LQ HGXFDWLRQ $P 27 f -RKQVRQ 5: $UEHV %+ DQG 7KRPSVRQ &* 6HOHFWLRQ RI RFFXSDWLRQDO WKHUDS\ VWXGHQWV $P 27 f -RKQVRQ 5: DQG /HRQDUG /& 3V\FKRORJLFDO WHVW FKDUDFWHULVWLFV DQG SHUIRUPDQFH RI QXUVLQJ VWXGHQWV 1XUV 5HV f

PAGE 83

.DW]HOO 0( ([SHFWDWLRQV DQG GURSRXWV LQ VFKRROV RI QXUVLQJ $SSO 3V\FK f .LP -2 DQG .RKRXW )0XOWLSOH UHJUHVVLRQ DQDO\VLV 6XESURJUDP UHJUHVVLRQ 8QLY RI ,RZD LQ 6366f§6WDWLVWLFDO 3DFNDJH IRU WKH 6RFLDO 6FLHQFHV 1LH 1RUPDQ + HW DO 0F*UDZ +LOOf§1< QG HG SS .QDXHU )( $ VWXG\ RI WKH UHODWLRQVKLS RI VHOHFWHG YDULDEOHV WR SHUVLVWHQFH RI DFDGHPLFDOO\ FDSDEOH IRUPHU VWXGHQWV RI WKH 8QLY RI 6' 8QSXE 'RFWRUDO 'LVVHUW 8 RI 6' .QRSI / )URP VWXGHQW WR 51 $ UHSRUW RI WKH QXUVH FDUHHUf§SDWWHUQ VWXG\ 3XE +OWK 6HU '+(:f :DVK '& 'LY 1XUV 1XUVH &DUHHUf§3DWWHUQ VWXG\ %DFFDODXUHDWH GHJUHH QXUVHV WHQ \HDUV DIWHU JUDGXDWLRQ +RVS 7SFV f .UDOO 9 3HUVRQDOLW\ IDFWRUV LQ QXUVLQJ VFKRRO VXFFHVV DQG IDLOXUH 1XUV 5HV f /DSLQ /DZUHQFH / 6WDWLVWLFV IRU 0RGHUQ %XVLQHVV 'HFLVLRQV +DUFRXUW %UDFH -DYDQRYLFK ,QF 1< /DUNHQ 3* $GPLVVLRQ WHVW UHVXOWV DV FULWHULD IRU HQWUDQFH LQWR WKH QXUVLQJ SURJUDP DQG UHJUHVVLRQ DQDO\VLV RI HQWUDQFH WHVWV DV SUHn GLFWRUV RI VXFFHVV LQ QXUVLQJ VWDWH ERDUG H[DPLQDWLRQV 3ULQFH *HRUJHnV &RPP &RO /DUJR 0' 2IILFH RI ,QVWLWXWLRQDO 5HV /HH -( DQG 6KHSDUG .) $ FRPSDULVRQ RI ELRJUDSKLFDO GDWD RI SK\VLFDO WKHUDS\ VWXGHQWV LQ &DOLIRUQLD 3K\V 7KHU f /HZLV DQG :HOFK 0 3UHGLFWLQJ DFKLHYHPHQW LQ DQ XSSHUGLYLVLRQ EDFKHORUnV GHJUHH QXUVLQJ PDMRU (G t 3V\FK 0HDV /HYLWW (( /XELQ % 'H:LWW .1 $Q DWWHPSW WR GHYHORS DQ REMHFWLYH WHVW EDWWHU\ IRU WKH VHOHFWLRQ RI QXUVLQJ VWXGHQWV 1XUV 5HV f /LDR :& DQG $GDPV -3 0HWKRGRORJ\ IRU WKH SUHGLFWLRQ RI SKDUPDF\ VWXGHQW DFDGHPLF VXFFHVV 3UHOLPLQDU\ DVSHFWV &RO 6WXG 3HU f /LQG $, $Q H[SORUDWRU\ VWXG\ RI SUHGLFWLYH IDFWRUV IRU VXFFHVV LQ WKH FOLQLFDO DIILOLDWLRQ H[SHULHQFH $P 27 f /RZHQWKDO : :HUJLQ DQG 6PLWK +/ 3UHGLFWRUV RI VXFFHVV LQ SKDUPDF\ VFKRRO 3&$7 YV RWKHU DGPLVVLRQ FULWHULD $P 3KDUP (G f

PAGE 84

/XFFL -$ $Q DSSURDFK WR WKH VHOHFWLRQ RI VWXGHQWV $P 27 f 0D\ $( DQG &KLWW\ ( 6HOHFWLQJ FDQGLGDWHV IRU SXSLO QXUVH WUDLQLQJ 1XUV 0LUURU SS 'HF 0D\QDUG 0 %LONH\ $ +\UH 3UHGLFWLQJ -RE SHUIRUPDQFH $P 27 f 0F&ODLQ (: 6L[WHHQ 3HUVRQDOLW\ )DFWRU TXHVWLRQQDLUH VFRUHV DQG VXFFHVV LQ FRXQVHOLQJ &RXQV 3V\FK f 0F*UHHY\ &3 )DFWRU DQDO\VLV RI PHDVXUHV XVHG LQ WKH VHOHFWLRQ DQG HYDOXDWLRQ RI FRXQVHORU HGXFDWLRQ FDQGLGDWHV &RXQV 3V\FK f 0RUJDQ 0. 7KH 23, WKH $&7 DQG XQLYHUVLW\ DWWULWLRQ $ GLVFULPLQDQW DQDO\VLV 'RFWRUDO 'LVVHUW 8 RI .HQWXFN\ 0RUJDQ 0. 6HOHFWLQJ FDQGLGDWHV IRU RYHUVXEVFULEHG SURJUDPV &RPPXQLW\ &ROOHJH 5HYLHZ )DOO 0XFKLQVN\ 30 DQG +R\W '3 3UHGLFWLQJ FROOHJH JUDGHV RI HQJLQHHULQJ JUDGXDWHV IURP VHOHFWHG SHUVRQDOLW\ DQG DSWLWXGH YDULDEOHV (G t 3V\FK 0HDV 0XQUR %& 7KH 0LQQHVRWD 7HDFKHU $WWLWXGH ,QYHQWRU\ DV D SUHGLFWRU RI WHDFKLQJ VXFFHVV (G 5HV 1RY 0XWKDUG -( 0RUULV -' &URFNHU /0 DQG 6OD\PDNHU -( )LHOG ZRUN H[SHULHQFH UDWLQJV DQG FHUWLILFDWLRQ H[DPLQDWLRQ VFRUHV DV SUHGLFWRUV RI MRE SHUIRUPDQFH DQG VDWLVIDFWLRQ LQ RFFXSDWLRQDO WKHUDS\ $P 27 f 0DU 1HHOH\ 5 'LVFULPLQDQW DQDO\VLV IRU SUHGLFWLRQ RI FROOHJH JUDGXDWLRQ (G t 3V\FK 0HDV 1HOVRQ $0 8QGHUJUDGXDWH DFDGHPLF DFKLHYHPHQW LQ FROOHJH DV DQ LQGLFDWLRQ RI RFFXSDWLRQDO VXFFHVV 3URIHVVLRQDO 6HULHV &LYLO 6HUYLFH &RPP :DVK '& 3HUV 0HDV 5HV t 'HYHO &WU 6HSW 1HZVOHWWHU RI 027$ f 1LFNHQV -0 $ FRPSDULVRQ RI DFFXUDF\ RI VHOHFWHG PRGHOV IRU SUHGLFWLQJ DFDGHPLF SHUIRUPDQFH RI MXQLRU FROOHJH WUDQVIHU VWXGHQWV (G 5HV f 2VLSRZ 6+ DQG :DOVK :% 6RFLDO LQWHOOLJHQFH DQG WKH VHOHFWLRQ RI FRXQVHORUV &RXQV 3V\FK f 2ZHQ 69 DQG )HOGKXVHQ -) (IIHFWLYHQHVV RI WKUHH PRGHOV RI PXOWLn YDULDWH SUHGLFWLRQ RI DFDGHPLF VXFFHVV LQ QXUVLQJ HGXFDWLRQ 1XUV 5HV f

PAGE 85

3DOPLHUL $$ ,,, $ PRGHO IRU PXOWLYDULDWH SUHGLFWLRQ RI DFDGHPLF VXFFHVV RI WUDQVIHU VWXGHQWV LQ SKDUPDF\ VFKRRO $P 3KDUP (G f 3DWWHUVRQ 7: DQG 'DYLGVKRIHU &2 $ FRXQVHOLQJ DQG WUDLQLQJ SURJUDP IRU RFFXSDWLRQDO WKHUDS\ VWXGHQWV $P 27 3HGULQL %& DQG 3HGULQL '7 3UHGLFWRUV RI FROOHJH VXFFHVV (5,& PLFURILVFKH (' 3HQQHOO 0< DQG +RRYHU '% +HDOWK PDQSRZHU VRXUFH ERRN $OOLHG KHDOWK PDQSRZHU VXSSO\ DQG UHTXLUHPHQWV :DVK '& *RYW 3ULQWLQJ 2IF 1R 6HF 3IHLIHU &0 -U DQG 6HGODFHN :( 3UHGLFWLQJ EODFN VWXGHQW JUDGHV ZLWK QRQLQWHOOHFWXDO PHDVXUHV 1HJ (G Of 3RROH 0$ DQG .DVVDORZ 6 0DQSRZHU VXUYH\ UHSRUW :LVFRQVLQ 2FFXSDn WLRQDO WKHUDS\ DVVRFLDWLRQ $P 27 f 3UHGLJHU '3UHGLFWLRQ RI SHUVLVWHQFH LQ FROOHJH &RXQV 3V\FK f 3UHQGHUJDVW 1 &RPPHQWV RQ LVVXHV LQ HGXFDWLRQ $P 27 ff§ 3URJUHVV 1RWHV 2FW WKURXJK )DOO 8QLY RI )OD 'HSW RI 27 5HDYOH\ : DQG :LOVRQ /6HOHFWLRQ RU GLDJQRVLV" 7KH XVH RI SV\FKRPHWULF WHVWV ZLWK FDQGLGDWHV IRU VWXGHQW QXUVH WUDLQLQJ 1XUVLQJ 7LPHV SS 1RY 5HHG &/ 3UHGLFWLRQ RI JUDGH SRLQW DYHUDJHV XVLQJ FRJQLWLYH DQG QRQ FRJQLWLYH SUHGLFWRU YDULDEOHV 3V\FK 5HS 5HHG &/ )HOGKXVHQ -) DQG 9DQ 0RQGIUDQV $3 3UHGLFWLRQ RI JUDGH SRLQW DYHUDJHV LQ QXUVLQJ VFKRROV XVLQJ VHFRQGRUGHU PXOWLSOH UHJUHVVLRQ PRGHOV (G 0HDV f 5HGIHULQJ '/ DQG %LDVFR ) 6HOHFWLRQ DQG HOLPLQDWLRQ RI FDQGLGDWHV LQ FRXQVHORU HGXFDWLRQ SURJUDPV &RXQV (G t 6XSU 5LFKDUGV -0 -U &DONLQV (9 0F&DQVH $ DQG %XUJHVV 00 3UHGLFWLQJ SHUIRUPDQFH LQ D FRPELQHG XQGHUJUDGXDWH DQG PHGLFDO HGXFDWLRQ SURJUDP (G t 3V\FK 0HDV 5RHPHU 5( 1LQH \HDU YDOLGLW\ VWXG\ RI SUHGLFWRUV RI PHGLFDO VFKRRO VXFFHVV (G 5HV 5RVH +$ 3UHGLFWLRQ DQG SUHYHQWLRQ RI IUHVKPDQ DWWULWLRQ &RXQV 3V\FK f

PAGE 86

6FKLPSIKDXVHU )7 DQG %URVNL '& 3UHGLFWLQJ DFDGHPLF VXFFHVV LQ DOOLHG KHDOWK FXUULFXOD $OOG +OWK :LQWHU 6FKPLGW -% 3UHGLFWLRQ RI VXFFHVV LQ FOHULFDO RFFXSDWLRQV IURP DELOLW\ WHVW VFRUHV 9RF *GQF 4UWO\ 6HLW] /$ 7RZDUGV D VHOHFWLRQ DQG DGPLVVLRQV PRGHO 3UHGLFWLQJ DFDGHPLFVXFFHVV LQ YHWHULQDU\ VFKRRO 'RFWRUDO 'LVVHUW 2KLR 6WDWH 8QLY 6KDOLN + $ PXOWLYDULDWH SURFHVV PRGHO WR HYDOXDWH XQGHUJUDGXDWH 37 VWXGHQWV DGPLWWHG E\ VWUDWLILHG VDPSOH DQG E\ RUGLQDO UDQN SUHn VHOHFWLRQ 'RFWRUDO 'LVVHUW 8QLY RI )OD 6LQJK $ 7KH SUHGLFWLYH YDOXH RI FRJQLWLYH WHVWV IRU VHOHFWLRQ RI SXSLO QXUVHV 1XUV 7LPHV 6WHLQ 5) DQG *UHHQ (7KH JUDGXDWH UHFRUG H[DPLQDWLRQ DV D SUHGLFWLYH SRWHQWLDO LQ WKH QXUVLQJ PDMRU 1XUV 5HV Of 6WHLQ 6 /LQQ 0: DQG )XUGRQ 3UHGLFWLQJ VRFLDO ZRUN VWXGHQW SHUIRUPDQFH (G 6RF :UN f 6WRQH 5* DQG $FTXDYLYD )$ $27$ +XPDQ 5HVRXUFHV 3URMHFW ,, 7KH (GXFDWLRQ 6WXG\ -P 27 f 6XVVPDQ 0% DQG +DXJ 05 5HKDELOLWDWLRQ FRXQVHORU UHFUXLWV &RXQV 3V\FK f 7DWKDP (/ 1XUVLQJ DQG GHQWDO K\JLHQH VHOHFWLRQ SURFHGXUHV 3DUW $Q H[DPLQDWLRQ RI DFDGHPLF YDULDEOHV DV SUHGLFWRUV RI VXFFHVV -RKQVRQ &R &RPPXQLW\ &ROOHJH 2YHUODQG 3DUN .DQVDV 0DU 7KRP 'DQG +LFNFR[ (6 7KH VHOHFWLRQ RI QHZ JUDGXDWH VWXGHQW IRU DQ HGXFDWLRQDO DGPLQLVWUDWLRQ SURJUDP (G $GP Of 7LGG *6 DQG &RQLQH 7$ 'R EHWWHU VWXGHQWV SHUIRUP EHWWHU LQ WKH &OLQLF" 3K\V 7KHU f 7LHGHPDQ '9 7KH XWLOLW\ RI WKH GLVFULPLQDQW IXQFWLRQ LQ SV\FKRORJLFDO DQG JXLGDQFH LQYHVWLJDWLRQV 7KH 0XOWLSOH 'LVFULPLQDQW )XQFWLRQf§ $ 6\PSRVLXPf +DUYDUG (G 5HY 7XFNHU $& DQG 6WURQJ (. -U 7HQ\HDU IROORZXS RI YRFDWLRQDO LQWHUHVW VFRUHV RI PHGLFDO FROOHJH VHQLRUV $SSO 3V\FK ff§ 86 'HSW RI +HDOWK 0DQSRZHU +HDOWK 5HVRXUFHV 6WDWLVWLFHV +HDOWK 0DQSRZHU )DFLOLWLHV :DVK '& *RYW 3ULQWLQJ 2IILFH /LEUDU\ RI &RQJUHVV 1R

PAGE 87

86 'HSW RI /DERU 2FFXSDWLRQDO 2XWORRN +DQGERRN :DVK '& *RYW 3ULQWLQJ 2IILFH %XOOHWLQ :DOEHUJ +6FKRODVWLF $SWLWXGH WKH 1DWLRQDO 7HDFKHU ([DPLQDWLRQV DQG WHDFKLQJ VXFFHVV (G 5HV :DONHU 60 $ VWXG\ RI WKH YDOXH RI WKH DGPLVVLRQV UHTXLUHPHQWV LQ SUHGLFWLQJ JUDGXDWLRQ IURP D SK\VLFDO WKHUDS\ SURJUDP DQG HQWUDQFH LQWR WKH SURIHVVLRQ 0DVWHUV 7KHVLV %RVWRQ 8QLY :HEHU / .LQJ 0( DQG 3LWWV 3UHGLFWLQJ VXFFHVV LQ VFKRROV RI SUDFWLFDO QXUVLQJ (G t 3V\FK 0HDV :LWWPH\HU $/ &DPLVFLRQL -6 DQG 3XUG\ 3$ $ ORQJLWXGLQDO VWXG\ RI DWWULWLRQ DQG DFDGHPLF SHUIRUPDQFH LQ D FROOHJLDWH QXUVLQJ SURJUDP 1XUV 5HV f :LOOLQJKDP :: 3UHGLFWLQJ VXFFHVV LQ JUDGXDWH HGXFDWLRQ 6FL f :LQHU %6WDWLVWLFDO 3ULQFLSOHV LQ ([SHULPHQWDO 'HVLJQ 0F*UDZ+LOO %RRN &RPSDQ\ ,QF 1HZ
PAGE 88

$33(1',;

PAGE 89

.H\ WR $EEUHYLDWLRQV :.1:. ZRUNLQJ RU QRW ZRUNLQJ LQ WKH ILHOG RI IRXU WR HLJKW \HDUV DIWHU JUDGXDWLRQ RFFXSDWLRQDO WKHUDS\ 5(*,67 $PHULFDQ 2FFXSDWLRQDO 7KHUDS\ 1DWLRQDO ([DPLQDWLRQ VFRUH 5HJLVWU\ 3<&+,1 3V\FKLDWULF ,QWHUQVKLS VFRUH 3' ,17 3K\VLFDO 'LVDELOLWLHV ,QWHUQVKLS VFRUH &203 3 &RPSOHWLRQ RI QR FRPSOHWLRQ RI WKH HGXFDWLRQDO SURJUDP LQ RFFXSDWLRQDO WKHUDS\ :. 67$ :RUN VWDWXV VFRUH IRXU WR HLJKW \HDUV DIWHU H[SHFWHG JUDGXDWLRQ

PAGE 90

6XEMHFW 1XPEHU :.1:. 5(*,67 3<&+,1 3' ,17 &203 $ $ $ $$ $ $ $ $$ $$ $ $ $ $ $ $ $ $ $ $ $$ $ $ $ $ $2 $ $ $ 0HDQ $ $ $ $$ $ :. 67$

PAGE 91

6XEMHFW 1XPEHU LW :.1:. LW 5(*,67 3<&+,1 LW 3' ,17 LW &203 0HDQ 0HDQ 0HDQ 0HDQ 0HDQ :. 67$

PAGE 92

6XEMHFW 1XPEHU :.1:. 5(*,67 3<&+,1 3' ,17 &203 0HDQ :. 67$

PAGE 93

6XEMHFW 1XPEHU :.1:. 5(*,67 3<&+,1 3' ,17 &203 3 :. 67$ 0HDQ f§ B f§

PAGE 94

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n WDQW 3URIHVVRU RI (GXFDWLRQ DW WKH 8QLYHUVLW\ RI 0RQWDQD RFFXSDWLRQDO WKHUDS\ DQG HGXFDWLRQDO FRQVXOWDQW WR WKH &RPSUHKHQVLYH 'HYHORSPHQWDO &OLQLF ,Q 0LVVRXOD 0RQWDQD DQG VSHFLDO HGXFDWLRQ FRQVXOWDQW WR WKH 0LVVRXOD &RXQW\ KLJK VFKRROV )URP KHU PDUULDJH WR .HLWK & 6WHZDUW 0' RI &HGDU )DOOV ,RZD WKH\ KDYH WZR VRQV &KDG DQG 'LUN 6KH LV D PHPEHU DQG D UHJLVWHUHG RFFXSDWLRQDO WKHUDSLVW ZLWK WKH $PHULFDQ 2FFXSDWLRQDO 7KHUDS\ $VVRFLDWLRQ VLQFH

PAGE 95

, FHUWLI\ WKDW KDYH UHDG WKLV VWXG\ DQG LQ P\ RSLQLRQ LW FRQIRUPV WR DFFHSWDEOH VWDQGDUGV RI VFKRODUO\ SUHVHQWDWLRQ DQG LV IXOO\ DGHTXDWH LQ VFRSH DQG TXDOLW\ DV D GLVVHUWDWLRQ IRU WKH GHJUHH RI 'RFWRU RI 3KLORVRSK\  n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

PAGE 96

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

PAGE 97

81,9(56,7< 2) )/25,'$


24
found no significant correlations between GPA, internship scores, and
employer ratings of physical therapists' one-year job success. A seven
year study by Maynard et al. (1972) revealed a small, but significant,
relationship between job performance and practical experience grades.
Nelson (1975) concluded from an extensive review of the literature of
education, health, psychology, social science, and business that GPAs
from college are of little use in predicting later occupational success.
Tucker (1962) did a ten-year follow-up of medical students and found
that interest scales did not measure distinctions between the medical
specialty areas. Ginsberg concluded that college performance had no
differential significance for later performance; however, outstanding
grades in graduate school did indicate later superior performance
(Nelson 1975).
Thom and Hickcox (1975) did a different approach to prediction of
career success. He had faculty, students and principals predict on the
basis of file information for six students whether the students would do
well in academics and careers. The actual achievements of the students
were known to the researchers. He found that there was no difference in
selection accuracy. They were approximately 60 percent correct for
academic success but only 50 percent for job success. Transcripts and
letters of recommendations were the most useful for selection of
achievers.
Concluding Remarks
Most of the studies reviewed suggested that previous intellectual
functions such as academic achievement and standardized test scores for


71
Dixon, WJ, ed: Biomedical Computer Programs, 3rd ed. Berkeley:
University of California Press, 1976 pp 285-387.
Dixon, WJ and Brown, MB, Eds: BMDP Biomedical Computer Programs
P-Series University of California Press, 1979 pp. 399-436.
Drugo, JG: Regression models and noncognitive predictors. Doctoral
Dissert, University of Pittsburgh, 1975.
Dubs, R: Comparison of student achievement with performance ratings of
graduates and state board examination scores. Nurs Res 24(1):59-
62, 1975.
Dunning, ER: An accountability model for occupational therapy. Am J OT
29(1):35-38, 1975.
Dunteman, GH: Characteristics of students in the health related profes
sions. Univ of Fla, Rehab Res Monograph Series, Number 2,
June 1966.
Engelhart, HV: An investigation of the relationship between college
grades and on-the-job performance during clinical training of
occupational therapy students. Am J OT 11(2):97-107, Part II
1957.
Federici, L and Schuerger, J: Prediction of success in an applied M A
psychology program. Ed & Psych Meas 34:945-952, 1974.
Feldman, EW: Characteristics associated with persistence in premedical
education at Duke Univ: A discriminant analysis. Doctoral Dissert,
Duke Univ. 1976.
Fredricks, MA, Mundy, P, Robertson, LS and Kosa, J: The national board
examination and academic achievement in a medical school. J Ed Res
63:275-278, Feb 1970.
Gough, HG: Some predictive implications of premedical scientific
competence and preferences. J Med Ed 53:291-300, Apr 1978.
Gough, HG and Hall, WB: The prediction of academic and clinical perform
ance in medical school. Res Higher Ed 3:301-314, 1975.
Green, LH: Predicting academic survivalattrition in first semester of
Marquette University physical therapy curriculum. Unpub Thesis,
Milwaukee, Wise. Jan 1962.
Green, PE and Tull, DS: Research for Marketing Decisions, 2nd ed.
Prentice-Hall, Inc. New Jersey, 1970.
Haglund, AH: Predicting success in collegiate nursing programs.
Doctoral Dissert, Univ of Wise, Madison, 1975.


40
its distance from the wall on the right, while the distance from
the left wall represents the value for (Lapin 1973).
There are no theoretical limits to the number of variables that
may be used in multiple regression. Until the recent accessibil
ity to high-speed computers, the higher-dimensional multiple
regression was severely limited in scope because of the vast
quantity of calculations involved (Lapin 1973). This study would
have been practically impossible before the advent of the
computer.
9. Discriminant function of two groupsis an expansion of multiple
regression. It is illustrated in the following figure:
1
Figure 3. Discriminant Function (p. 370)


8
There is a need to determine if certain characteristics
discriminate between those who remain in a field and those
who drop out. (p. 72)
The present study was needed to determine if objective factors
obtained from records available at the time of selection could be used
as criteria to predict the completion of training, entry into, and
persistence in the field of occupational therapy.
Assumptions
1. It was assumed that the transcripts of the students involved in
this study at the University of Florida were accurate.
2. It was assumed that scores and grades which appeared in the
student's personnel file located in the occupational therapy
department were accurate.
3. It was assumed that the data received from the American
Occupational Therapy Association were accurate.
4. It was assumed that follow-up studies are a practical means of
determining strengths and weaknesses of selection processes.
5. It was assumed that academic success is related to occupational
success.
Hypotheses
Four generalized, multivariate null hypotheses were proposed for
multiple regression analysis:
It will not be possible to predict scores for
(I) the National Registry Examination
(II) Psychiatry internship


Subject
Number
112
WK/NWK
#26
REGIST
#27
PYCHIN
#28
PD INT
#30
COMP
91
1
155
273.5
277.5
1
92
1
167
256.0
272.0
1
93
0
150
210.0
189.0
1
94
1
162
224.0
207.0
1
95
1
175
279.0
250.0
1
96
1
172
274.5
277.0
1
97
1
178
279.5
277.0
1
98
0
173
276.0
200.0
1
99
0
163
230.0
211.5
1
100
1
176
253.0
270.0
1
101
1
164
Mean 255.0
219.5
1
102
1
144
268.0
201.0
1
103
1
180
251.0
267.0
1
104
1
166
294.5
278.0
1
105
1
170
300.0
280.0
1
106
1
180
285.0
280.5
1
107
1
194
284.0
266.0
1
108
1
168
267.0
218.5
1
109
1
177
239.0
194.0
1
110
1
183
242.0
284.0
1
111
1
155
186.0
144.0
1
112
1
157
279.0
229.5
1
113
1
174
231.0
298.0
1
114
1
187
243.0
280.5
1
115
1
174
283.0
268.0
1
116
1
151
238.0
155.0
1
117
1
143
279.0
280.0
1
118
1
159
292.5
233.0
1
119
1
165
208.0
292.5
1
120
1
201
291.0
290.0
1
121
0
171
250.0
265.0
1
122
0
183
246.0
213.0
1
123
1
188
265.0
265.0
1
124
1
144
245.0
238.5
1
125
1
167
220.5
248.0
1
126
1
180
289.0
243.5
1
127
1
189
253.5
212.0
1
128
1
160
288.0
292.5
1
129
0
166
237.5
194.5
1
130
1
161
236.0
219.0
1
131
1
168
287.0
284.5
1
132
1
168
202.0
236.5
1
133
1
197
275.5
251.5
1
134
1
164
289.0
260.0
1
135
1
179
274.0
299.0
1
#31
WK STA
10
09
04
10
10
10
10
08
04
10
10
10
10
10
10
09
10
10
10
10
10
10
10
10
10
10
10
10
10
10
08
08
10
10
10
10
10
10
06
10
10
10
10
10
10


54
68.7 percent with an error of 31.3 percent. A total average for correct
placement of the subjects was 68 percent. This is a very low percentage
since the correct prediction by chance would be 93.5 percent, as
illustrated in Table 11, the classification matrix. The "jacknifed
classification" is a procedure of the cross validation type (see Table
12). The "jacknifed classification" table indicates again that the
chance rate of correct classification was 92 percent while the actual
correct classification rate was only 66 percent.
Table 10
Model V and VIClassification Functions
Variable Group =
Complete
Did Not Complete
Number 4 TRANS
-0.946
-2.443
18 ENGLT
0.322
0.202
20 PSYCH
0.207
0.334
24 GPA-1
8.983
10.218
Model V and
Table 11
VIClassification Matrix
Group
Percent
Correct
Number of Cases
Complete
Classified into Group
Did Not Complete
Complete
66.7
14
7
Not Complete
68.2
47
101
Total
68.0
61
108


62
on the basis of multiple combinations of variables which include college
entry test scores, courses taken, grade point averages, and other
objective data taken from transcripts.
The second part used stepwise discriminant analysis to find the
best indicators of completion and non-completion of the occupational
therapy program and the best indicators for work persistence and non
persistence in the field of occupational therapy using Biomedical
computer packaged program P-Series, BMDP7M.
The four multivariate null hypotheses proposed for the discriminant
analysis were all statistically rejected.
It was not possible to discriminate among groups of students (V)
those who completed and (VI) those who did not complete the occupational
therapy program on the basis of linear combinations of variables which
include college entry test scores, courses taken, grade point averages,
and other objective data taken from transcripts.
It was not possible to discriminate among groups of students (VII)
those who were working and (VIII) those who were not working in the
field of occupational therapy four to eight years later on the basis of
linear combinations of variables which include college entry test
scores, courses taken, grade point averages, and other objective data
taken from transcripts.


FACTORS AFFECTING PERSISTENCE IN
EDUCATION AND EMPLOYMENT OF
OCCUPATIONAL THERAPISTS
By
CAROLYN D. STEWART
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1979


45
Output from stepwise regression analysis (BMDP2R) includes at each
step,
1. Multiple R
2. Standard error of estimate
3. Analysis-of-variance table
4. For variables in the equation:
a. Regression coefficient
b. Standard error
c. Tolerance
d. _F to remove
5. For variables not in the equation:
a. Tolerance
b. Partial correlation coefficient
c. F to enter
Tiedeman (1951) explained that discriminant analysis determines
whether discrimination between groups is possible or not and then
reduces the size of the space in which it is necessary to think about
regions of classification (Morgan 1970, p. 37). Again, the forward
variable select method was employed. At each step one variable was
selected from the variables not in the analysis and entered into the set
of discriminating variables. The variable selected at each step was the
one which provided the greatest improvement in classification between
the two groups. A variable was considered for selection only if its
partial multivariate _F ratio was larger than 3.999. Separate stepwise
discriminant analysis was done for predictor variables for each of the
two dependent variables: (1) complete/not complete, and (2) work/not


29
6. Each faculty member wrote a summary of impressions of the
student.
7. Meeting as a selections committee, the entire faculty
discussed each student, finally ranking all in order of
preference for admission.
Data Collection
Data collection began with the files in the occupational therapy
curriculum office at the University of Florida's J. Hillis Miller Health
Center. The social security number was obtained for each student
involved in the occupational therapy curriculum that did or would have
graduated between 1967 and 1973. These files yielded the scores for
three of the dependent variables on each student: National Registry
Examination, Psychiatric Internship, and Physical Disabilities Intern
ship. Since some of the Registry Examination scores were never recorded
in these records, it was then necessary to consult the computer printout
from the American Occupational Therapy Association (AOTA). Many of the
scores for the internships had not been coded and computed. The faculty
member who had been in charge of this task was consulted. Most of the
internship evaluations were found, coded, computed, and recorded. Some
scores were obtained by telephone and were not recorded; therefore, they
were not available even from the internship institution. This affected
five students for the Psychiatric Internship and two students for the
Physical Disabilities Internship (see Appendix). One student who is
currently registered could not be reached to give permission to AOTA to
release her Registry Examination score. An average was computed for all


55
Table 12
Models V and VIJacknifed Classification
Group
Percent
Correct
Number of Cases
Complete
Classified into Group
Did Not Complete
Complete
57.1
12
9
Not Complete
67.6
48
100
Total
66.3
60
109
The first significant variable entered was number 4, University of
Florida freshman or transfer student (TRANS); second was the variable
number 20, number of hours in psychology (PSYCH); third, number 24, GPA
for the first two years of college (GPA-1); and fourth, number 18,
number of hours in English and literature (ENGLT). These four variables
are the same ones that entered into the multiple regression analysis for
predicting scores in psychiatry internship.
Null hypotheses V and VI were both rejected since it was statis
tically possible to discriminate among groups of students, those who
completed (V), and those who did not complete (VI) the occupational
therapy program on the basis of linear combinations of objective
variables used in this study.
The same basic equation applied to discriminant analysis as to
multiple regression analysis:
Y = a + b X1 + b X + b_X_ + b.X.
V Complete = 14.387 .946X1 (TRANS) + .207X2 (PSYCH) + 8.983X3
(GPA-1) + .322X (ENGLT)
4


18
Belcastro (1975) in which he used the Edwards Personal Preference
Schedule (EPPS) and the SVIB. The Autonomy Scale of the EPPS and the
Architect Scale of the SVIB ranked second and third, respectively. In
first place was the students' GPA. One- to four-year success in a
health education program was again determined to be best predicted by
GPA by Heit et al. (1978). Success in graduating from a three year
pharmacy program was also predicted by prepharmacy GPA (Palmieri 1976).
Burgess (1972), Lewis and Welch (1975), and Owen and Feldhusen (1970)
found that prenursing GPA was the best indicator of nursing course grade
point average.
Crane (1962) and Levitt et al. (1971) both showed that the Kuder
Preference Record and the EPPS as statistically significant in predict
ing students who would complete an occupational therapy program and a
nursing program, respectively.
Graduate school success as measured by the graduate level GPA was
studied by Bean and Covert (1973), Feldman (1976), Federici and Schuerger
(1974), R. Stein and Green (1970) and Willingham (1974). The best vari
able in all of these studies was the Graduate Record Examination scores.
Clapp (1975), Gough and Hall (1975), Richards et al. (1974), and Roemer
(1965) agreed that the Medical College Admission Test (MCAT) was a signif
icant predictor of medical school success (GPA). They also found that
premedical GPA and science GPA were significant. Predental GPA was the
most significant predictor of junior year dental GPA (Chen et al. 1967).
Halfter's (1966) study showed that the Law School Admissions Test (LSAT)
score predicted first year law school GPA somewhat better than the
undergraduate GPA did.


20
Blume (1976) correlated success in a medical technology internship with
responses on a Biographical Data inventory. Booth (1957) showed
significant correlations between pediatric internship of occupational
therapy students and the T score of Guilford's STDCR Inventory and the
Literary Interest scale of the Kuder Preference Record, although the
latter was a negative correlation. The C score of STDCR correlated with
grades from a tuberculosis internship and the total internship grade was
linked with the persuasive interest scale of the Kuder Preference Record.
Osipow and Walsh (1973) conducted two studies that indicated tests
of social intelligence were useful in predicting a desired facet of
performance of counselor trainees in counseling psychology. McClain
(1968) found that the 16 PFI test was more useful for both men and women
in identifying the superior counselors than in differentiating the
average from the poor. Gough (1978) stated that high premedical science
competance and performance was unrelated to clinical performance and
faculty ratings of clinical competence. Munro (1964) decided that the
Minnesota Teacher Attitude Inventory was an inadequate predictor of
teaching performance in internship.
Tests of Professional Knowledge
Ayers and Rohr (1974) and Walberg (1967) attempted to predict scores
on the National Teacher Examination (NTE). Ayers found that grades in
social science were the best indicators of success on the NTE, while
Walberg found that high school GPA and scores on the School and College
Ability Test (SCAT) were the best indicators.