Citation
Factors related to student choice of medical specialty

Material Information

Title:
Factors related to student choice of medical specialty
Creator:
Duerson, Margaret C
Publication Date:
Language:
English
Physical Description:
ix, 155 leaves : ill. ; 28 cm.

Subjects

Subjects / Keywords:
College students ( jstor )
Medical education ( jstor )
Medical schools ( jstor )
Medical specialties ( jstor )
Medical students ( jstor )
Nursing students ( jstor )
Physicians ( jstor )
Primary health care ( jstor )
Questionnaires ( jstor )
Students ( jstor )
Career development ( lcsh )
Career education ( lcsh )
Dissertations, Academic -- Educational Leadership -- UF
Educational Leadership thesis Ph. D
Medicine -- Specialties and specialists ( lcsh )
City of Gainesville ( local )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1986.
Bibliography:
Bibliography: leaves 148-153.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Margaret C. Duerson.

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:
030378488 ( ALEPH )
16904862 ( OCLC )

Downloads

This item has the following downloads:


Full Text













FACTORS RELATED TO STUDENT CHOICE
OF MEDICAL SPECIALTY



By


MARGARET C. DUERSON


















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













UNIVERSITY OF FLORIDA 1986














ACKNOWLEDGEMENTS



My deepest gratitude is extended to each of my

committee members. Special thanks go to my chairperson, Dr.

James W. Hensel, who provided me with unending support and guidance. His was a warm and motivating force throughout

the project. As a role model he exempt if ies the teacher I hope to emulate as I interact with students and peers.

Each committee member played a unique and indispensable

role in assisting me in this endeavor. I would like to thank Dr. Parker A. Small for his extensive, constructive criticism and his encouragement to constantly strive for excellence. I am grateful to Dr. Robert E. Jester and Dr. Ronald G. Marks for their valuable comments and suggestions, especially in the areas of research design and analysis. It was my good fortune to have the advice and help of Dr. Forrest W. Parkway.

The Department of Community Health and Family Medicine made available to me information and resources during the conduct of this study. For this, I am deeply indebted and appreciative. I would also like to thank Robert Epting for his help in executing the data analysis procedures. In addition, support was provided by Health and Human Services Grant number 2D15PE-84000-07.




ii









Finally, and most especially, my sincerest appreciation goes to my husband, Kearney, my two sons, and my parents whose love and encouragement have enabled me to develop and grow. Kearney's faith in me ultimately made this possible.














TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS .................................... ii

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

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

CHAPTER

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

Statement of the Problem ..................... 3
Purpose of the Study ......................... 4
Hypotheses ................................... 5
Background and Justification ................. 7
Delimitations, Limitations, and
Assumptions ................................ 11
Definition of Terms .......................... 14
Organization of the Study .................... 16

Ii REVIEW OF THE LITERATURE ..................... 18

Introduction ................................. 18
Background Factors ........................... 19
Personality and Attitude Factors ............. 31
Factors Related to the Medical Training
System ..................................... 37
Career Factors ............................... 46
Summary of the Review of Literature .......... 49

III METHODOLOGY AND INSTRUMENTATION .............. 53

Setting for the Study ........................ 53
Revision and Development of the
Questionnaire .............................. 56
Administration of the Questionnaire .......... 62 Residency Match .............................. 66
Research Design .............................. 67
Data Analysis ................................ 70

IV PRESENTATION OF RESULTS ...................... 76

Sociodemographic Characteristics of the
Study Group ................................ 76
Results of the Data Analysis for Testing
the Hypotheses ............................. 82


iv










?aqe

v SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS .... 116

Summary ...................................... 116
Conclusions .................................. 121
implications of the Study .................... 127
Recommendations .............................. 128

APPENDICES

A MEDICAL STUDENT ENTERING QUESTIONNAIRE ....... 136 B MEDICAL STUDENT FOLLOW-UP QUESTIONNAIRE ...... 140

C LETTER OF EXPLANATION ATTACHED TO ENTERING
QUESTIONNAIRE ................................ 143

D LETTER OF EXPLANATION ATTACHED TO FOLLOW-UP
QUESTIONNAIRE ................................ 145

E LETTER OF REMINDER SENT WITH SECOND FOLLOW-UP
QUESTIONNAIRE ................................ 147

REFERENCES .......................................... 148

BIOGRAPHICAL SKETCH ................................. 154






























v












TLIST OF TABLES


Table Page

1 Summary of Sociodemographic Characteristics
of Subjects ..................................... 77

2 Medical Specialty Preferences of Third-Year
Medical Students ................................ 80

3 Residency Choice of Fourth-Year Medical
Students ........................................ 81

4 Association between Third-Year Specialty
Preference and Sociodemographic Characteristics ........................................ 83

5 Association between Medical Specialty
Preferences and Father's Occupation and
Parents' Annual Income ........................ 85

6 Association between Medical Specialty Choice
of Fourth-Year Students and Father's
Occupation ...................................... 87

7 Association between Fourth-Year Specialty
Choice and Sociodemographic Characteristics... 89

8 Association between Medical Specialty Choice
of Fourth-Year Students and Parents' Annual
Income and College Major ....................... 91

9 Sociodemographic Groupings of Medical
Students and Preclerkship Attitudes ........... 94

10 Sociodemographic Groupings of Medical
Students and Postclerkship Attitudes .......... 95

11 Results of Preclerkship Career Goal Rankings
and Three Preference Groups ................... 100

12 Results of Preclerkship and Postclerkship
Career Goal Rankings and Two Specialty
Choice Groups .................................. 102

13 Correlation Coefficients for Career Goal
Rankings and Attitude Scores .................. 105






vi









Paqe

14 Comparison of Specialty Choice Groups Based
on Differences in Preliminary and Follow-up
Career Goal Rankings .......................... 110

15 Association between Third-Year Specialty
Preferences and Fourth-Year Specialty
Choices ....................................... 112

















































vii










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


FACTORS RELATED TO STUDENT CHOICE OF MEDICAL SPECIALTY

by

Margaret C. Duerson

December 1986

Chairperson: James W. Hensel Major Department: Educational Leadership

The purpose of this study was to identify factors

related to medical students' third- and fourth-year specialty choices and to determine the extent to which a clerkship in family medicine affected attitudes toward primary care. A multidimensional questionnaire was administered to third-year medical students immediately before and after a family medicine clerkship. The questionnaire addressed sociodemographic characteristics of the medical students, their career goals, and attitudes toward primary care. On the initial administration, the students indicated whether they preferred a primary care specialty, a subspecialty area, or were undecided on their future plans for residency training. The results of the fourth-year residency match served as evidence of the students' choice.

Responses by 109 students revealed that (a) third-year specialty preference was significantly associated with

fourth-year specialty choice of residency; (b) socioeconomic



viii








measures, parents annual income greater than $50,000, (c) students with experience as teachers were distinguished by

more positive attitudes toward primary care; (d) students supported by parents scored lower on attitudes toward primary care than those whose major means of support came

from other sources; and (e) the importance students assigned to 12 career goals did not differentiate between third-year specialty preference or fourth-year residency choice groups. Based on their mean attitude scores, students choosing primary care residencies held more favorable attitudes toward primary care before and after a clerkship in family medicine than subspecialty-oriented students but the subspecialty scores demonstrated greater improvement in attitudes.

These findings support several educational

interventions. Since the majority of students are capable of making stable career decisions by the third year, educational strategies likely to impact on these decisions should be implemented early in the educational process. The link between socioeconomic status and specialty choice suggests that as tuition costs increase, adequate financial assistance becomes critical for the less affluent students

who aremost likely to fill the ranks of primary care specialties. Early training activities in ambulatory care settings should be fostered to maintain the interest of

students preferring primary care and to improve attitudes and attract other students.




ix















CHAPTER I
INTRODUCTION



An important goal of this nation is to provide an

effective, efficient health care system for its citizens. To accomplish and maintain such a system requires that shortages and surpluses of health care providers-particularly physicians--be avoided. When considering supply and requirements for physician manpower needs, the focus must not be limited solely to aggregate numbers but must take into consideration the distribution among specialties of physicians, also.

The Summary Report of the Graduate Medical Education National Advisory Committee (GMENAC) (U.S. Department of Health and Human Services, 1980) cited imbalances in physician specialty areas as a major health care problem facing the nation. There is general consensus that there are too many physicians in some specialty areas and not enough in others. Using projected manpower needs for 1990, surpluses are expected to occur in nonprimary care areas such as surgery, while many of the primary care specialties will fall short of the projected requirements (Jacoby, 1981).

Neither is the maldistribution across specialties

likely to abate in the near future unless some action is



-1-






-2


taken to correct.the situation. Using the supply assumptions of the GMENAC Report, Kindig and Dunham (1985) projected physician specialist growth into the 21st century. By the year 2020, the absolute number of physicians in primary care will increase by 55% but the absolute numbers in nonprimary care will increase by 111%. As these authors observed, it is conceivable that with the aging of the population and the competitive approaches to cost containment the additional primary care physicians will be needed but it is difficult to assume a need for an additional 260,000 nonprimary care physicians.

Intervening in the situation while the numbers of physicians yet to be trained is amenable to change is essential. Two approaches toward correcting the problem include (a) stringent government regulations to limit the types and numbers of residencies which train specialists, to bring reimbursement of specialists in line with primary care physicians, and to restrict the number of graduates of foreign medical schools and foreign physicians coming into the country; and (b) working within the educational environment to influence specialty choice decisions. Of the two approaches, the latter seems the most viable solution in a democratic society to meet health care needs and at the same time allow individuals freedom in the specialty decision process.

Until recently there has been little effort directed toward assuring that the medical education system produced






-3


the proper mix of primary care and nonprimary care physicians for the needs of society. When the medical education system has been targeted, modification has occurred at the residency level rather than undergraduate medical education. Establishing policies and finding solutions has been hampered because factors influencing specialty choices are complex and not well understood. Although a large body of research has been generated

the specialty choice research literature
suffers from major inadequacies in providing
guidance for policy-makers and only rarely can provide unequivocal evidence regarding various
influences on physician specialty choice. (U.S.
DHHS, 1980, Vol. V, p. 2)



Statement of the Problem


The problem addressed in this study is the need for

research-substantiated information on the factors related to student choice of medical specialties as stated broadly in the following questions:

Are sociodemographic characteristics of medical

students associated with third-year specialty preferences or choice of specialty as evidenced by selection of a primary care or nonprimary care residency in the fourth year of medical school?

Is there a difference between medical students grouped by sociodemographic characteristics with regard to their attitudes toward primary care and family medicine?






-4


Is there a difference in career goals between groups of medical students identified by third-year specialty preferences or fourth-year specialty choice?

I What is the relationship between the importance medical students attach to career goals and their attitudes toward primary care?

Is there reason to believe that career goal rankings before and after a family medicine clerkship are different for medical students choosing primary care from those choosing subspecialties in the fourth year of medical school?

Is specialty preference expressed in the third year of medical school associated with medical students' specialty choice decisions in the fourth year as evidenced by residency selection?

-4 Are medical students' attitudes toward primary care changed after a clerkship in family medicine?
Is there a difference in attitudes toward primary care between groups of medical students identified by third-year specialty preferences of fourth-year specialty choice as evidenced by residency selection?



Purpose of the Study


The primary purpose of the study was to identify

factors related to medical students' third- and fourth-year specialty choices and to determine the extent to which a clerkship in family medicine affected the students'






-5


attitudes toward-primary care. observations obtained from administration of a multidimensional questionnaire were used to explore (a) the association between sociodemographic characteristics of medical students, their career goal rankings, and their attitudes toward primary care and the students' specialty preferences or specialty choices; (b) whether significant differences in sociodemographic characteristics existed between medical students with favorable and unfavorable attitudes toward primary care; (c) the relationship between the relative importance students assigned to career goals and their favorable or unfavorable attitudes toward primary care; (d) the extent to which attitudes toward primary care changed after a clerkship in family medicine; and (e) the association between medical students' career preferences in the third year and their actual career choice in the fourth year as measured by residency specialty.



Hypotheses


This study involved the testing of the following hypotheses:

1. There is no association between sociodemographic characteristics of medical students and their medical specialty preferences expressed in the third year of medical school.

2. There is no association between sociodemographic characteristics of medical students and their medical






-6


specialty choices as evidenced by fourth-year residency selection.

3. There is no difference between medical students grouped by sociodemographic characteristics according to their attitudes toward primary care measured before or after a third-year family medicine clerkship.

4. There is no difference between groups of medical

students identified by medical specialty preference in the third year or medical specialty choice in the fourth year with respect to their rankings of career goals.

5. There is no relationship between medical students'

rankings of career goals and their attitudes toward primary care before or after a third-year family medicine clerkship.

6. There is no difference between the groups of

medical students choosing primary care and those choosing subspecialties in the fourth year with respect to their rankings of career goals before and after a family medicine clerkship.

7. There is no association between medical students' specialty preferences expressed in the third year and their career specialty choices as evidenced by fourth-year residency selection.

8. There is no difference in medical students'

attitudes toward primary care before and after a family medicine clerkship.






-7


9. There is no difference between groups of medical

students identified by specialty preferences with regard to their attitudes toward primary care.

10. There is no difference between groups of medical students identified by their fourth year specialty choice with respect to their attitudes toward primary care before and after a family medicine clerkship.

The .01 level of significance was used for rejection of the null hypotheses.



Background and Justification


Achieving a better balance between physicians in

primary care specialties and those in other specialties has become one of the major concerns of medical educators, health care policy makers, and others vitally interested in

the proper functioning of the health care industry. Approximately 80% of the physicians in this country are specialists. When the ratio of generalist to specialist (primary care to nonprimary care) is imbalanced several consequences are likely to result: (a) cost of medical care increases due to the high dependence on specialists for

primary care, (b) quality of health care may actually be compromised from iatrogenic complications resulting from unnecessary procedures and physicians practicing outside of their areas of expertise in order to maintain their income,

and (c) doctors' job satisfaction could suffer as a result






-8


of improper use or under utilization of their skills and knowledge (Schroeder, 1984).

From the post-World War II period through the 1960s and 1970s, physician manpower policies were primarily concerned

with the overall number of physicians and their geographic distribution. In the 1940s and 1950s task forces on health manpower needs predicted a shortage of 50,000 doctors by 1980 (Petersdorf, 1975). Health manpower legislation of the 1960s was directed toward remedying this perceived shortage

by increasing medical school enrollments, recruiting faculty, and expanding the existing facilities. The preoccupation with increasing the aggregate number of physicians was so intense that the trend toward increased

specialization went largely unnoticed. From 1931 to 1974, physicians identifying themselves as general practitioners decreased from 83% to 18% (Harris et al., 1982).

The health manpower legislation of the 1960s was so

successful, it is estimated that by the year 2000 there will be 145,000 more physicians than needed to provide physician services (U.S. DHHS, 1980, Vol. I). The problem of oversupply of physicians is not limited to the United

States, however. Virtually every Western country is facing a current or projected excess of physicians. In European countries the surpluses have produced varying results from a dramatic decrease in physician income to actual unemployment for a sizeable number of medical doctors. While the United States shares the problem of surpluses with other countries,

compared to its Western allies, this country stands alone in






-9


the higher ratio of specialists to generalists (Schroeder, 1984) .

Factors responsible for the trend toward increasing specialization are difficult to determine. Rapidly

expanding technology and scientific knowledge may cause some to conclude that the only way to achieve depth of knowledge and skill in an area or to remain current in an area of medicine is to narrow the field of interest. Petersdorf (1975) commented on the possible causes:

Simplistically speaking, every man has an innate
desire for self -advancement, intellectually and
f financial ly. In addition, subspecialists seem to acquire prestige, particularly in medical schools.
More than anything else, however, the structure of
training programs in medical schools has led to
specialization. (p. 697)

To f i 11 the void 1 ef t with the demise of the o ld generalists' role, a new specialty, family practice, developed specifically designed to deliver primary care. The final approval of family practice as the twentieth primary specialty was granted by the American Medical Association in 1969 (Rakel & Pisacano, 1984). In an effort to respond to the need for more primary care physicians and

aided by funding provided by the Comprehensive Health Manpower Training Act of 1971, medical schools initiated residency training programs in this new specialty. In a further attempt to influence specialty distribution, legislation was passed in 1976 which "set minimum requirements for percentages of first year residency positions in the primary care specialties" for medical






-10


schools to meet in order to continue to receive capitation grant funds (U.S. DHHS, 1980, Vol. V, p. 1).

All of these early efforts toward correcting the

shortage of primary care physicians were directed at the residency training period. Almost no attention was given to understanding and influencing the career preferences and choices made earlier in the medical education process. Until recent years, research on career decisions at the

undergraduate medical education level had been principally conducted by sociologists and psychologists with academic interest in furthering the knowledge base in their own discipline rather than an interest in obtaining information for program evaluation and policy formulation. Influencing career choice at the undergraduate medical education level was sanctioned by Califano, Secretary of Health, Education,

and Welfare, in an address to the Annual Meeting of the Association of American Medical Colleges in 1978:

I urge medical educators to exert leadership in
dealing with the dangerous and wasteful decline in
primary care physicians by encouraging course
offerings or other innovations that increase the exposure of students to primary care settings and
heighten the appeal of primary care. (Califano,
1979, p. 22)

Medical schools accepted the challenge and expanded the curricula to include programs in family medicine at the

undergraduate level of medical education as well as the residency level.

Since the medical education environment is recognized as one of the factors contributing to the specialty





_11


imbalances, it is appropriate to study the relationship between the undergraduate educational process and specialty and subspecialty choices. A knowledge of the variables within the educational environment which influence and predict specialty and subspecialty career choices would provide guidance for medical educators, policy makers, and

others vitally interested in correcting the present medical manpower imbalances and planning for future needs.



Delimitations, Limitations, andAssumptions Delimitations of This Study


This research study was delimited to

1. Medical students who were in their third year of

training at one state-supported medical school in the southeast in the years 1983-84 and who participated in the residency match in 1984-85.

2. Factors affecting medical specialty choice at one university medical school as represented by the group of students sampled.



Limitations of This Study


There were several limitations that must be recognized in conducting this study:

1. The results of this research project were

characteristic of this study only. Because of the unique characteristics of samples and the variations inherent in





-12


different settings, generalizations beyond the sample studied in the setting of interest are to be avoided. The generalizability of this study was further limited since subjects were not randomly selected but came from an intact class. The students were randomly selected by a lottery system into the six rotation groups, however.

2. A limitation of the study was a lack of control of intervening variables. Specific events other than those being studied may have occurred between the first and second measurements to produce a change in the students' attitudes toward primary care and family medicine.

3. In test-retest situations, exposure to the first test may bias or sensitize the subject to the second test. Since the questionnaire used to measure entering attitudes was the same as the questionnaire used after the family medicine clerkship, this source of error could have biased the results of this study.

4. The design of the research instrument provided

limitations. The questionnaire items were forced choice, closed-end, and very short answer format. The Attitudes about Primary Care and Family Medicine section, Part F, used a Likert scale with the following choices: strongly agree, tend to agree, not sure, tend to disagree, and strongly disagree. Variable interpretations of this and other sections of the questionnaire could potentially distort the study's results.






-13


5. The follow-up questionnaire was mailed to medical

students at the completion of the family medicine clerkship. Mailed surveys often result in low return rates and even lower percent usable data as some items are not completed.

6. Working in the real world of educational research has the advantages of practical applicability and the disadvantages of having to deal with many variables simultaneously. If there is to be any application of the method to this or other sites, the natural setting has obvious advantages.



Assumptions of This Study


The following assumptions are implicit in the limitations and delimitations of the study:

1. Participants responded voluntarily to questionnaire items.

2. Responses were made in a nonthreatening environment so they were truthful and those intended by participants.

3. Administration and collection of the questionnaire did not influence the responses.

4. The participants were representative of previous medical students at the same level of training and those that follow at the University of Florida College of Medicine.

5. Experiences while on the family practice rotation were similar for all students.






-14


6. Items on the questionnaire resembling attitude statements concerning primary care and family medicine were representative of generalizations of the domain of interest.

7. Attitudes and beliefs are not open to inspection but can be inferred from the answers to statements pertaining to behaviors and positions toward certain actions.



Definition of Terms


The terms used in this research study are defined as follows:

A primary care physician is one who (a) is the

physician of first contact for the patient; (b) makes the initial assessment and attempts to solve as many of the patient's problems as possible; (c) coordinates the remainder of the health care team, including ancillary health personnel as well as consultants, that are necessary in dealing with the patient's problems; (d) provides continued contact with the patient and often his or her

family; and (e) assumes continued responsibility for his or her care (Petersdorf, 1975). The primary care physician administers a highly personalized type of care which coordinates all of the health care needs of the individual

in sickness and in health. Few specialties consider this comprehensive and continuing care to be their responsibility





_15


and within their range of competence (Rakel & Pisacano, 1984).

Clerkships are student-selected elective and required clinical experiences occurring during the last two years of medical school. Working as a member of the health care team in the actual care of patients provides the student with the opportunity to apply and practice newly acquired knowledge and skills.

Preceptor refers to a licensed, practicing physician who teaches medical students. This physician may be in private practice or in an academic setting.

The terms occupation preference, choice, and attainment are frequently used interchangeably. Vroom's distinction between occupation choice, preference, and attainment as explicated by Matteson and Smith (1977) are

Occupation preference is that occupation which at a

given time an individual would most like or prefer to enter.

Occupation choice is that occupation which a person

chooses to enter and then engages in behaviors to implement that choice.

Occupation attainment is the occupation of which the person is a member.

This study concerns itself with two points of occupational decision, preference and choice.

Preference of specialty is a positive written or verbal expression of a desire to pursue the particular area of medical practice. The participants' preferences for a






-16


primary care or nonprimary care specialty was solicited in the initial administration of the questionnaire in the third year of medical school prior to the family medicine clerkship.

Choice of medical specialty requires that some action be taken in the direction of or participation in a particular area of specialty practice, for example, taking steps to secure a residency in a particular medical specialty area. For purposes of this study, specialty choice was considered to be evidenced by the selection of residency in a given specialty area in the fourth year of medical school.



Organization of the Study


Chapter I contained an introduction, a statement of the problem, and the purpose of the study. The research hypotheses underlying the study were stated. The limitations and delimitations of the project were explained along with the definitions of words and phrases specific to this study.

Chapter II is a review of the literature related to medical specialty choice. Chapter III addresses the research methodology and instrumentation. The data collection procedures, the setting for the study, research design, and data analysis procedures are explained. Statistical results are presented in Chatper IV.






-17


The final chapter, Chapter V, summarizes the findings of the study, the conclusions, and the implications. Recommendations for future research in the area of medical specialty career choice are suggested.














CHAPTER II
REVIEW OF THE LITERATURE



Introduction


During the last 10 years researchers have examined a

wide range of factors thought to influence medical specialty choice. Because of the importance of the problem and the consequences of these intraprofessional career decisions, medical educators, sociologists, psychologists, health policy-makers, and others concerned with medical care in general have attempted to understand why and how choice of medical specialization occurs. The Graduate Medical Education National Advisory Committee Report (U.S. DHHS, 1980), confirming the suspicion that the nation faced an oversupply in some specialty areas and actual shortages in other specialties, gave further impetus to the study of variables associated with specialty choice.

This great interest in the career choice process of young doctors has resulted in an extensive body of literature. Literature reviews on career choice within medicine were published in the mid-1970s. Due to the comprehensive nature of these publications, this review will examine the literature published since 1975.

Approaches to the study of the problem have been so

diverse, this review will consider related research together



-18-






-19


which shares a common approach or similar variable, rather than summarizing the studies chronologically. The factors will be categorized into four groups: (a) background factors, (b) variables related to personality and attitudes,

(c) factors pertaining to the medical environment or training system, and (d) career factors.



Background Factors


A large part of the literature on specialty choice

focuses on students' background characteristics including age, sex, marital status, hometown, socioeconomic status of family, and educational history. Two reasons for researchers' seeming preoccupation with these variables is assumed to be their accessibility from students' records and their potential use in the selection process for admission to medical school. What must not be forgotten, however, is that medical students are a highly selected, homogeneous group so that differences in these background characteristics are small while specialties and subspecialties are quite diverse making the background variables poor predictors of specialty choice.

Both Anderson (1975) and Zuckerman (1977) in extensive reviews of the literature reported evidence that the larger the community of origin the greater the likelihood of the student choosing specialty practice over general practice. Hutt (1976) reported similar findings in Britain where those






-20


from rural areas.tended more often to prefer general practice.

Closely related to choice of specialty is choice of

practice location. The relationship between practice location and specialization becomes more apparent when one considers that highly trained subspecialists would not likely find sufficient patients or the sophisticated equipment necessary for practice of their specialty in small towns. In a study of students entering the University of

Washington School of Medicine in 1975, 1976, and 1977, Carline and associates (1980) found that contrary to other studies there were few differences in preferences for specialties based on size of hometown, while students' attitudes toward location of practice indicated a preference for practice in communities similar in size to those in which they were raised. Support for this finding was also reported by Hadac (1984). In a review of the literature associated with specialty and location choices, he found that a preference for rural or small town practice was associated with high school attendance in a town of similar size.

A second variable of interest is socioeconomic status

or social class which has been commonly measured in terms of the father's occupation and level of education. Studies of the relationship between social class and choice of specialty are mixed. Results of some studies demonstrated

an absence of a relationship or ambiguous results, while






-21


other researchers found a positive relationship (Anderson, 1975; Zuckerman, 1977). Where positive associations between social class and specialty choice were found, the higher the socioeconomic status of the family, the more likely the student was to choose specialization over general practice. As Zuckerman (1977) pointed out "the economic security and availability of financial assistance associated with higher social class may encourage pursuit of the additional training required for specialty practice" (p. 1082). Generally, the primary care specialties require fewer years of residency training than the subspecialties and thus decreased cost of education and more rapid entry into wage earning status.

Medical students whose fathers and mothers are

physicians constitute an interesting group with regard to social status, especially since they seem to be overrepresented in medical schools based on the population census. Gough and Hall (1977) observed that approximately 16% of students in American medical schools come from medical families compared to a census estimate of 0.4% physicians among employed adult males. In their study of 1195 students at the University of California, San Francisco, School of Medicine, there were 162 from medical families. These investigators noted that there was a statistically insignificant shift away from family medicine and general practice by the subgroup of physicians'






-22


children. However, there was a pattern toward overrepresentation in some specialty areas for the physicians' children, namely ophthalmology, otolaryngology, dermatology, and surgery.

Mawardi (1979), in a longitudinal study of Case

Western Reserve University graduates, found that 39 of the 135 physicians studied gave credit for their interest in medicine to family members in the profession. The most frequently mentioned influential relative was a father. These 39 subjects cited 59 relatives who were physicians.

Given the rising cost of medical education and the concommitent decrease in financial aid available to students, it could be hypothesized that students from families with few financial resources will not apply to medical schools. This could lead to a profession of the socially elite. Based on the preliminary studies of socioeconomic status and specialty choice, a preponderance of students from affluent backgrounds will effect the mix of specialist toward nonprimary care specialties.

Factors such as age and marital status have provided some consistent clues to specialty choice decision. Using the data from a longitudinal study of Jefferson Medical College students from 1971 through 1975, Herman and Veloski (1977) demonstrated not only a rising trend in the proportion of the Jefferson students interested in family medicine but those senior students expressing an interest in






-23


family medicine tended to be slightly older than those choosing other specialties.

Age and marital status may be interrelated inthat in generalmarried students tend to be older. Assuming that students who choose family medicine share common social characteristics, Cole, Fox, and Lieberman (1983) examined questionnaire data obtained from 429 students at a public, northeastern medical school. The profile of the subgroup of students who chose family medicine was disproportionately older, married, and more often Protestant. It may be that as with socioeconomic status, it is not the age or marital status per se that effects specialty choice for the older, married student but rather the economic restraints that make long residency training difficult.

Hutt (1976) found similar outcomes in Britain.

Marriage had a precipitating influence on specialty choice. Twice as many married men opted for careers in general practice. Married women were more apt to take up careers outside of hospitals or, if inside, to take those not involving clinical duties.

In a unique approach to gain a better understanding of physicians' career choices, Skipper and Gliebe (1977) surveyed not only senior medical students on their plans for a medical specialty but also asked the students their perceptions of the influence of their spouses on those plans. Going one step further, the researchers collected data from 17 of the 21 wives of the married students. Due






-24


to their own career goals, the women preferred that their husbands settle in a medium-sized city or suburban fringe area. They also preferred group practice to solo practice so that the husbands would have more time for family activities. While the specialty choices expressed by the wives was varied, some of the wives viewed family medicine as less desirable because it would provide less free time than other specialties. Researchers conclude that if primary care specialties such as family medicine are to appear attractive to wives of medical students as well as the students themselves, there will need to be some attempt to also convince the spouse of the good points of the specialty.

The small numbers of subjects involved in the Skipper and Gliebe (1977) study preclude definitive statements. it does, however, suggest that as more marriages become two career families, there will be pressures on career choices to accommodate both husbands and wives.

Medical schools have the mixed blessing of having a large applicant pool of highly qualified candidates from which to choose. While medical schools do not publish the admission criteria, it is generally known that students must have demonstrated academic success by virtue of their undergraduate grades and their scores on the Medical College Admission Test (MCAT). Strong interest and a proven aptitude for science are given more weight than other factors in the selection process. This emphasis on






-25


undergraduate grades in science and the MCAT scores as criteria for admission is questionable. The Anderson (1975) review of the literature reported mixed results in studies that used MCAT scores to differentiate specialty choice. Subsequent studies have not provided more consistent data.

Gough (1978) used four measures, the MCAT science subtext scores, premed grade point average (GPA), a preference for scientific subjects, and a composite of the three, in this study of 1,135 graduates of the University of California, San Francisco, Medical School. When these factors were used to predict performance in medical school, all four factors were significantly correlated the first year of medical school but the following years became less and less predictive of performance until by the fourth year they were completely unrelated. With regard to specialty, the scientifically oriented students entered specialties such as surgery, anesthesiology, and pathology; those ranking lower went into internal medicine, pediatrics, and psychiatry.

In his review of the literature to 1977, Zuckerman found that students planning to enter general practice scored low on MCATs and had low academic standing in medical school. Conversely, those who planned academic or research careers had high MCAT scores and high academic standing. Those planning to specialize and go into private practice scored in the mid range on MCATs. Other studies published the same year as the Zuckerman review reported (Association






-26


of American Medical Colleges, 1977; Hadley, 1977) that students with high undergraduate grade point averages and high MCAT scores appear to be predisposed toward subspecialty areas rather than primary care.

More recent studies of academic variables indicate that unlike the students in years past who went into general practice, students choosing primary care or family medicine perform at least as well if not better than those tending toward subspecialty areas. In a study of trends in senior students' interest in primary care specialties at Jefferson Medical College, Herman and Velosk (1977) found that students interested in family medicine performed as well or better on measures of basic science knowledge and subjects emphasizing clinical information as those students interested in all other specialties except internal medicine. Similarly, Collins and Roessler (1975) found no significant differences in intellectual characteristics as judged by MCAT scores and undergraduate GPA between family medicine oriented students and any other group in their study of Baylor Medical College students. Results of a more recent study of a large national sample of medical students (Burkett & Gelula, 1982) added additional weight to the previously cited findings. In a comparison made between primary care versus nonprimary care fields, students in the two groups were similar on the basis of measures of premedical school academic performance and MCAT scores.






-27


Academic measures made later in the medical school

process have not proved much better in producing consistent results. Two studies conducted in the same state but at two separate medical schools used National Board of Medical Examiners (NBME) examination scores to study the relationship to career choice. The NBME exam results are available late in the third year or early fourth year of medical school. One group of investigators who examined the effects of NBME scores for 628 students at the University of Medicine and Dentistry of New Jersey-New Jersey Medical School, found that high scores on the NBME Part II subtests in psychiatry, medicine, and surgery were associated with selection of these specialties (Fadem et al., 1984). Contrary to these findings, Rosevear, Tickman, and Gary (1985) did not find a positive relationship between medical students' career choice and their performance on the NBME Part II subtest scores for 347 graduates (1979-1983) of the University of Medicine and Dentistry of New Jersey-Rutgers Medical School.

Factors predictive of specialty choice which occur early in the educational process have the advantage of providing more opportunity for impact and intervention. Unfortunately, the early parameters such as undergraduate grade point average, MCAT scores, and other cognitive measures do not consistently discriminate between those choosing primary care specialties and those choosing nonprimary care specialties. It may be that because all






-28


medical practice.requires a high intellectual functioning, there are no discernible differences based on scholastic factors.

With the growing numbers of women entering medical school, research on the effects of gender composition on specialty choice has been heightened. Historically, women have been underrepresented in medicine. Women constituted only 7.7% of medical students in 1964-1965 (Dube, 1973). Concurrent with the women's movement of the late 1960s and the efforts to include women in occupations where they had been formerly excluded, their representation in medicine increased. Presently females constitute over 30% of the medical student population. The percentage of women in the entering class of 1991 is expected to be close to 45.3% (Lanska et al., 1984). This change in gender distribution has possible implications for health care delivery in general and specialty choice in particular.

Several studies of sex differences related to specialty choice found significant differences between male and female medical students' propensity toward specialties. Male/female differences in specialty choice, beliefs about specialties, and personality characteristics were explored by McGrath and Zimet (1977). Students at two state university medical schools rank ordered their preference for five major medical specialties. Both men and women listed family medicine as their first choice but significant differences occurred at the levels of second and third






-29


choices as women.ranked pediatrics second and men chose internal medicine. Surgery ranked last for women and third for men.

Similarly, Cuca (1979), based on plans for boardcertification of 1978 U.S. medical school graduates, found that over half of both males and females preferred primary care while twice as many women preferred pediatrics as did the men. The category of surgery again evidenced dissimilar trends for women and, men.

The work of Zimny and Shelton (1982) supported these earlier studies. Responses of 380 third year medical students on the Medical Specialty Preference Inventory showed no significant differences between male and female students' preference for obstetrics, pediatrics, and psychiatry. men, on the other hand, had a slight preference for internal medicine and surgery.

Bergquist and his colleagues (1985) asked the 1983 entering class of one medical school to select their intended specialty choice from a list of 50 medical fields. When the specialties were categorized into primary care, surgery, medicine, and other, the results showed that women selected specialties in primary care more often than men. Men selected surgical categories significantly more frequently than women.

On the theory that as the proportion of women in

organizations reach parity with men, women will be less likely to be entrapped in limited or stereotypically female






-30


roles, Weisman (1984) found that, indeed, in the graduates of 1970 through 1976 higher proportions of women in medical schools were associated with smaller proportions of women entering traditionally female specialties. There was only a weak effect of gender on first-year residency choice.

There was disagreement in the studies as to whether women were becoming less traditional in their specialty choice. Both males and females expressed a strong interest in the primary care specialties indicating some convergence. However, there are also major differences in the factors perceived to be important in choosing a specialty. Women expressed a preference for patient contact specialties where patients are involved in solving their own health problems and were willing to forego career time for more family life (Bergquist et al., 1985; Zimny & Shelton, 1982). Males showed a high preference for complex medical problems and technical procedures which would necessitate functioning in highly specialized areas (Zimny & Shelton, 1982). Males also expected to have higher incomes than women expected but it is not clear whether this expectation dictated the career choice or was a consequence of it (Bergquist et al., 1985).

The eventual impact on medicine of greater

participation by women is unclear. The fact that primary care specialties are popular with women may help to correct the problem of overspecialization but it is not the entire solution to the national health problems. This solution would be an oversimplification of the problem and women's






-31


unique contribution to health care. As the number of women physicians increases, medicine may take on some of the characteristics common to female dominated occupations. The changing gender mix of medical school classes merits continued study.



Personality and Attitude Factors


It is not surprising given the homogeneity of medical students' intellectual ability and some background characteristics that the ability to associate specialty choice with these variables has been inconsistent. Because the sociodemographic data have been so capricious, some investigators have sought to determine whether measurable differences in personality variables are predictive of the type of specialty chosen.

Zuckerman (1977) in a comprehensive review of the literature found that psychological characteristics distinguished psychiatrists and surgeons from other specialty types but these factors have been less fruitful in distinguishing other types of specialties. He concluded, "The utility of personality characteristics to predict variation in career choice appears to be limited" (p. 1083). Hutt's (1976) review of previous research supports the consistent findings of personality differences between psychiatrists and surgeons. She observed that "there is evidence that these two choices are made early and are more consistently adhered to than others" (p. 466). It may be






-32


that those specialties where the choice is made later, based on experience in the fields, are much less successfully identified from personality characteristics. Hutt commented, "The field of personality differences in relation to specialty, important as they are, is not a promising one for policy-makers who wish to put right particular imbalances in the supply of doctors" (p. 466).

Though there have been problems in predicting specialty choice from comparison of personality variables, it is important to review the findings. For example, Collins and Roessler (1975) compared four different career groups, internal medicine, surgery, obstetrics-gynecology, and pediatrics, with family medicine residents. One hundred eighteen third-and fourth-year students at Baylor College of Medicine participated in a battery of personality, vocation, and attitude tests to investigate the relationship between the test data and subsequent specialty choice. The Baylor students who chose family medicine residencies were in most instances significantly different from the other four groups. They scored higher on the need for affiliation, were less aggressive and less materialistic. The family medicine students' scores were far below the population average of 50 on materialism on the Birkman Vocational Interest and Attitude Survey.

Plovnick (1980) also observed differences in attitudes and values between students oriented toward primary care and those choosing nonprimary care specialties. The small






-33


sample size of this study precluded meaningful statistical analysis but several important trends were identified using three factors, orientation to Patient Care, Orientation to

Work Conditions, and Orientation to the Profession. Students choosing primary care careers scored higher on the

factor "Orientation to Patient Care," indicating more concern for people and less orientation towards the profession. more unsettling was his finding that over the

four years of medical school, primary care respondents exhibited a shift in their attitudes away from concern for

patient care toward a somewhat greater self-concern. Plovnick explained this trend as the general socialization influence of medical school which might be changed by a further differentiation of programs designed for primary and tertiary care providers.

A widely used instrument for investigating personality differences and career choices is the Myers-Briggs Type Indicator (MBTI). Based on the personality theory of Jung, it compares four categories on a continuum of opposites:

extroversion- introversion, think ing-f eel ing, sensingintuition, and judgment-perception. These type differences are related to differences in preferences for activities, interests, personal outlooks, and choices of occupation. In

the early 1950s, Myers tested over 5000 medical students. Those still in practice were followed up in 1973 (McCaulley, 1981) and confirmed that their specialty choices were similar to those chosen at graduation. Harris, Kelly, and






-34


Coleman (1984) demonstrated the stability of the MBTI from entrance to graduation for the classes of 1980 and 1981 at the University of Utah School of Medicine. In a discussion of implications relevant to the findings of MBTI studies, Hadac (1983/1984) observed that sensing types tend to be denied admission to medical schools because they generally score lower on entrance exams. When they are admitted, they prefer primary care specialties, thus admission policies may be eliminating those students likely to correct the

specialty maldistribution. Hadac cautioned that psychological type alone cannot predict specialty choice for

all students. He stated, "Also, all personality types are found in all specialties and there is notatpresent, and may never be, strong evidence that any type should be excluded from any specialty" (p. 46).

When considering the literature on personality and

attitudes, it was considered worthwhile to address interests, values, and early orientations to medicine. The criterion group in the Leserman (1978) study were first year medical students at three North Carolina medical schools. Data from this group of students

suggest that incoming medical students are
concerned with helping people but not necessarily
through political means, committed to some
geographic and specialty areas of patient need,
choosing medicine for reasons other than economic
rewards but not opposed to physicians' large
income and status and somewhat unaware of
discrimination toward women physicians and
patients. (p. 330)






-35


Seventy-three percent of this group expressed interest in primary care and 54% were interested in rural health. This represents a higher percent than might be expected compared to national statistics. Steinwachs and his colleagues (1982) in their analysis of recent trends in graduate medical education reported that approximately 55% entering residency in 1980 chose primary care.

In a national survey of third-year students, Burkett and Gelula (1982) asked students to assign relative values to four motivating factors in their decision to enter medical school. Those students indicating a preference for primary care attributed greater importance to the desire to help people than to the other three choices--desire for

financial reward, desire for social status, and desire to apply scientific ability. They also demonstrated a tendency to consider the sociopsychological context of patient

problems and felt a need for change to improve health care. According to Burkett and Gelula (1982), "the overall pattern appears to confirm the notion of primary care as a 'personcentered' health care field which attracts students who have a more 'holistic' orientation" (p. 512).

Robbins et al. (1983) studied gender differences in interest and motivation for a career in medicine. Not surprising, both men and women scored highest on the medical

science category of the Strong-Campbell Interest Inventory. Both sexes were similar on a projective technique where 90% were judged to fear success. On the attitudinal






-36


questionnaire, males and females ranked the same reasons as important for going to medical school although they ranked them in different order. Males' order of choices were interest in science, helping people, and having a career. Women ranked helping people as first in importance, followed by desire for a career, and interest in science. All respondents listed a happy family life and job satisfaction as important life goals. Men tended to consider a high income as more important than the females. Although attitudes toward chores were egalitarian, responses indicated that in reality women performed most routine household activities. The researchers pointed out that moments value considerations may affect initial career choices as to training programs and work opportunities.

Assessment of personality characteristics has been done rather unsystematically using a variety of standardized and nonstandardized instruments administered at varying times throughout medical school. The characteristics observed have reflected the special interests of the investigators and have often measured vague concepts such as idealism and cynicism. There has been no attempt to replicate previous research in most cases.

A further difficulty with systematizing the assessment of personality factors is that psychological tests can be falsified if the respondent is motivated to do so. The person can answer the items in the way that is believed to be preferred by the investigator or is perceived to be






-37


socially correct.. Most students consider "helping people" as the appropriate answer rather than "economic gain" when asked the reason to pursue a career in medicine.

Disregarding the difficulties inherent in personality testing, there is a major problem inherent in manipulating specialty distribution based on personality characteristics, namely, the legality of and societal distaste for doing so (U.S. DHHS, 1980). The subjective interpretation of the results of personality tests will certainly not go uncontested by those denied admission were these tests given specific weighting in the admission process. Society would surely reject the routine use of personality tests as the basis of admission due to the perceived subjectivity, unreliability, and possibility of social regimentation implied by such action. The information gained from personality inventories and questionnaires is most appropriately reserved for counseling students.



Factors Related to the Medical Training System


As Zuckerman (1977) pointed out, sociodemographic and

personality characteristics present at admission become less useful over time in predicting specialty choice, while those factors associated with the institutional environment take on more importance. Students are presumed to come into the medical training setting with certain motives, values, and knowledge. These factors are subject to change or reinforcement by the pressures of agents or conditions






-38


present in the environment. There is a strong feeling that these structural influences and the socialization process of the medical education system are key determinants of career choice. Based on this premise a number of studies have focused on the institutional environment and manipulation of variables therein.

Early reports concerning the development of medical

students were generally qualitative but, nonetheless, paved the way for later quantitative studies on institutional influences. Becker et al. (1961), in a hallmark study of the socialization process, addressed the observed increase in cynicism of medical students during their training. He argued that students enter with idealistic concerns about the sick and serving mankind. From his observations, students did not lose this idealism for broader social

issues. What some interpreted as cynicism was really the response to pressures of day-to-day details with long hours and intense studying which caused a narrowing of concerns just to get through the rigors of medical school. This short term effect disappeared but the long term institutional effects did not. Becker thought there was enough congruence between students chosen for admission

and the institutional values to effect long range perspectives.

In a review of the literature published almost 20 years after the Becker book, cynicism continued to be reported as a by-product of the medical school environment.






-39


Rezler (1974) concluded from his review that attitude changes resulting from a special program will fade unless they are reinforced by the total environment.

One suggested strategy for examining the influences of the medical education environment on specialty choice is based on the immediacy of their impact on the students (U.S. DHHS, 1980). First-order effects have a direct influence on the knowledge, skills, values, attitudes, and interests of the students. The primary example of first-order effects are the role models, teachers, and attending physicians. Also included in this category are the opportunities for role playing and practicing the acquired knowledge, skills, values, and attitudes.

Second-order influences are filtered through the firstorder effects and are composed of the organizational and institutional components which determine the type of agents and opportunities available for the socialization process. Third-order and all subsequent influences are those which are further removed and are filtered through the first- and second-order effects. The effects of these influences are so indirect and confounded by other variables that it precludes definitive conclusions (U.S. DHHS, 1980).

Following a logical sequence, first-order effects will be considered first.

An exemplary example of how difficult it is to

introduce a major innovative experience into a traditionally structured medical school curriculum was demonstrated in the






-40


Harvard experiment where the teaching of family medicine was introduced in 1953. Rosenblatt and Alpert (1979) followed up the first three cohorts of this early program to determine the impact of the experience on subsequent career choice, amount of family-oriented practice, and other achievements. No statistically significant differences could be found among the various groups, either within cohorts or across time. Of critical importance to the interpretation of the results of this very early attempt to integrate family medicine into the curriculum is the fact that at the time the program was introduced, Harvard had no faculty in family medicine to serve as role models. In fact in the latter years (cohort II and III) the directors wer.e two pediatricians, not family practitioners.

It is reasonable to believe that the availability of

role models and opportunities to practice the role will have a beneficial effect on the numbers of students choosing a specialty. In an effort to determine the beneficial and detrimental parameters influencing the choice of family practice as a career, Brearley and his colleagues (1982) surveyed 134 first year residents of southeastern family practice residencies on 18 curriculum components and elements that influenced specialty choice. The positive value of preceptorship experiences in family medicine during the third and fourth years of medical school and association with family physicians was striking. Detrimental influences were peer group attitudes and the traditional curriculum






-41


which the authors interpreted as being in opposition to students' interest in family medicine.

Closely associated with this study is the work of Harris et al. (1982) who inquired into the question "Is participation in a family practice track program associated with career choice decisions in family practice as evidenced by residency selection?" (p. 610). Each year for two years, 20 students who applied for an elective in family medicine were randomly selected for the experience. The students who participated in the elective selected family medicine residency at a significantly higher rate than those who expressed an interest in the track but were not selected for the elective. Although the subjects may be considered a biased group, the positive effect of experiencing such a program merits consideration.

Data pertaining to influences of persons and

experiences on career decisions published the same year as the Harris study (Paiva et al., 1982) extended the information in this area. The 1980 and 1981 classes of Southern Illinois University School of Medicine were sampled at three points in the curriculum, at the end of the basic science period, the end of the clinical science period, and the end of the clinical clerkships. Analysis of the data showed that approximately three-fourths of the students reported that a faculty member had some or very much influence upon their specialty decisions. The major influence came from full-time clinical faculty. The






-42


influence of particular events was more pronounced in the clinical phase than in the basic science period and the responses were more dichotomous on the evaluation scale. Those students choosing specialties selected by small numbers of students such as psychiatry and anesthesiology perceived the faculty influence to be a more important factor than did students choosing the popular specialties of internal medicine, surgery, and family medicine. Another finding which has implications for curriculum planners is the influence of the sequence of the rotations. The order of the clerkships was particularly relevant for the students who indicated prior to the clinical rotations that they were undecided on their specialty as most of them ended up choosing a specialty they had encountered during the first few clinical rotations.

Changes in understanding and attitudes resulting from an experience with a particular specialty were studied by Samra et al. (1983). Students participating in an anesthesiology rotation completed an attitudinal questionnaire containing statements related to anesthesiology. They also indicated their career choices. The clerkship did not have any immediate impact on the students' choices of specialties but there was evidence of change in perception of the image of an anesthesiologist to that of a physician rather than a technician. The students also perceived the role of the anesthesiologist to be broader after the clerkship. Although limited by the research






-43


design, the results are interesting enough to prompt further research.

A concl us ion to be drawn from the studies of f irstorder effects is that clinical teaching faculty exert a very strong role modeling effect on decisions of specialty choice. Experiences during the clinical rotations are an important component in the decision process, also, but it is probably the interplay of the two which is critical. The findings have implications for recruitment of students into primary care specialties.

Second-order influences are those that indirectly

effect first-order factors. These might include financing of medical education, the structure and organization of the medical school, and research endeavors. Given the fact that second-order must exert force through first-order, it is

hard to see how it would impact career choice among medical students, but a number of researchers reported interesting findings.

Zuckerman (1978) investigated the structural factors within the educational milieu. He hypothesized these factors linked together to form particular patterns which

resulted in different career outcomes. Based on structural characteristics, 28 medical schools were dichotomized into academic and clinical categories. Students were classified according to academic standing. The type of residency

chosen was also categorized. Finally, the type of medical practice entered was divided into four general areas. The






-44


findings of the tabulation procedure revealed 1,146 different career patterns for 2,514 students sampled. Obviously, the rigid tracking system Zuckerman hypothesized was not supported by the findings of the study.

Friedman, Stritter, and Talbert (1978) examined closely three community hospitals and an academic teaching hospital where their students had clerkships. Based on the amount and types of clinical experiences provided, they concluded that it was fallacious to try to dichotomize facilities.

Some community hospitals paralleled teaching hospitals on many characteristics.

There are probably multiple opportunities and combinations of experiences at sites which might be

considered primarily clinical or strictly academic. These two studies highlight the imprecision of the criteria used

to evaluate medical schools, not only by researchers but also by accrediting agencies.

Medical school administrators and government policy

makers advocate founding new departments, increasing fulltime faculty, and increasing research resources in an effort to interest students in a given specialty. How valid is the assumption that such investments produce dividends? The Canadian study by Roos and Roos (1980) offers some insight

into this question which may be applicable to the U.S. situation. They collected aggregate data on four characteristics of Canadian medical schools: number of full-time faculty, number of part-time faculty, research grants,






-45


and residency programs. This information was compared with first-year medical student preferences for specialties, the

students' fourth-year choices, and their careers four years after graduation. The results suggest that hiring more full-time faculty in a specialty area and encouraging more research in one area than another does cause physicians to enter one specialty over another. Full-time faculty and research activities have more influence on career choice

than clinical inputs such as part-time faculty and residency programs. In future planning, then, it would seem that increasing efforts in these directions would encourage students in the direction where the faculty and research are located.

Concerned about the decreasing interest in family

medicine since the early 1970s, Goldsmith (1982) surveyed 135 medical schools with 71 returns on their administrative structure and number of graduates choosing family medicine. The data revealed that schools with departmental status had a higher percentage of graduates in family medicine. The

explanation for the observed association may be that those with departmental status can effect admission policies, command more time in the curriculum, attract more faculty, and participate more heavily in career counseling.

A survey of departments of anesthesiology supported

these findings (Chandra & Hughes, 1984). Staffing patterns

in the departments were significantly related to the number of students choosing to specialize in anesthesiology.






-46


Career Factors


The factors examined to this point have been intrinsic characteristics of the students and the medical school environment but there are potential extrinsic influences worth consideration. Hutt (1976) wrote, "Curiously enough, these have received considerably less attention than the earlier groups although it is, of course, the career factors which policy-makers could most easily use to alter the distribution of doctors between specialties" (p. 468). Extrinsic factors include such areas as working conditions, economic incentives, and societal demands or needs.

There is considerable diversity between specialties

with regard to working conditions. Some specialties such as general surgery, pediatrics, and obstetrics require long, erratic hours with responsibilities at night and on weekends. Other specialties are more amenable to regular 9to-5 schedules. Style of practice, solo, group, or hospital, is related to specialty. For example, primary care specialties are principally ambulatory care practices with little emphasis on hospital practice. Other specialties such as surgery are principally hospital-based practice.

As was noted earlier, women are willing to forego

career time for more family time (Bergquist et al., 1985; Zimny & Shelton, 1982). They want more flexible hours and will usually opt for specialties that allow for that time. In Cuca's study (1979), women's career plans differed from






-47


their male counterparts in (a) length of residency, (b) employment settings, and (c) intended areas of specialties. The women planned to spend fewer years in residency training and wanted salaried practice options rather than private practice styles. These first two factors could explain the third--the tendency for women to participate at a higher rate in primary care specialties as opposed to some of the nonprimary care specialties which require long training periods and provide less flexible hours.

This is not to say that males do not have personal

goals that include time for family, leisure activities, and other pursuits and may consider these interests in making their choices. Up to this point, however, these goals have not been evidenced. It may be that as more families become two career families and more couples share family responsibilities, the males will express some of these same needs and demands.

The financial compensation for specialists in

nonprimary care specialties is generally higher than primary care specialties. Testing for economic incentives as factors in career choice is difficult. As stated earlier, people will not usually admit to what might be considered materialistic motives.

Eunkenstein had some success with economic factors in relation to career choice. Funkenstein's (1978) longitudinal, prospective study of Harvard medical students covered a period of five eras from 1958-1976. From his






-48


investigations he concluded that economic incentives and

ideology are more compelling than basic characteristics or original plans in predicting career outcomes. Further,

societal forces prevailing at the time of the decision exert significant influence. As Funkenstein (1978) observed, the "Sputnik" era saw an increase in emphasis on mastery in the

sciences with increased resources available in scientific subjects. Students admitted to medical school during this period had heavy science background. The 1960s and 1970s were periods of social conscience-raising and a time when individualism was prominent. This was followed by decreased funding in some areas of medicine so that careers in academic medicine were not as attractive (Funkenstein, 1978, p. 30).

Society continues to be enamoured with technology and

is willing to reward those who provide it. High-tech in the health field has brought us heart transplants, microsurgery, and life-sustaining equipment (Naisbitt, 1982). It could be hypothesized that high tech translates into highly specialized physicians to administer this type of medical care and a renewed interest in family doctors to balance the high-tech with highly personal care (Naisbitt, 1982, p. 41).

On the other hand the organization of the health care system is changing (Ginzberg, 1984). As large corporations buy and operate hospitals, medical care has become a business--a business for profit (Weiss, 1982). High

technology is expensive; subspecialists are expensive; tertiary care hospitals are expensive. The most economic






-49


medical care is administered by primary care facilities (Neuhauser, 1983). It follows then that there will be a demand by the hospital corporations for more primary care physicians to staff their hospitals. Subspecialists will be used sparingly and only on consultation. This change in the organization of the health care system is likely to have an effect on specialization.

Some of the career factors discussed here are amenable to manipulation. Working conditions can be changed to be more responsive to the needs of those seeking primary care

specialties in order to increase their numbers. Societal factors are not so easily defined and interventions are difficult to implement. Economic incentives can be

addressed in several ways. Decreasing the discrepancy between primary care and nonprimary care specialties would help offset this factor. The area needs more study upon which to base policy changes.



Summary of the Review of Literature


Although the specialty choice research literature is

voluminous, only rarely does unequivocal evidence of specific influences on physician specialty choice surface.

The conceptual differences and methodological problems make comparisons across studies difficult.

Since admission criteria have not been well defined,

the weight given to academic and personality variables by admissions committees is unclear. The study of personality






-50


factors has yielded little help with regard to prediction of

specialties. Personality tests have distinguished between

psychiatrist and surgeons but have been less helpful with

regard to other specialties. Academic measures available on

admission become less useful over time being more predictive

of early medical school performance but not later clinical

abilities. Little difference has been found in recent

studies of academic performance between those choosing

primary care and others.

Recognizing the limitations of the current admission

data, a viable alternative to the present admission process

was offered in the Summary Report of the Graduate Medical

Education National Advisory Committee (U.S. DHHS, 1980):

An idea which has been put forth as a solution to
another problem, that of the inevitable
disappointment of qualified applicants not being
accepted to medical schools because of heavy
competition from many other equally qualified
applicants, might also be a solution to the issue of personality screening for admission to medical
schools. The idea is that of a two stage
screening process, the first stage involving
screening on the basis of intellectual and
academic qualifications, the second involving a
random lottery. Such a process would assure that
acceptees would (a) be intellectually qualified,
(b) have an equal chance for selection at the
second stage (with "fate" making the final
decision and thus shouldering the blame for
nonacceptance) and (c) display personality traits in proportions representative of those occurring
in the population of qualified applications.
(Vol. V, p. 17)

Sociodemographic variables have provided some limited

insight into specialty selection. Size of home community,

age, and marital status have been most consistent of the

background variables in distinguishing primary versus






-51


nonprimary care specialists. The older, married students

from small towns are most likely to choose primary care specialties. With all other background characteristics, the results are mixed.

It appears that the variables brought to the medical setting diminish in importance as factors related to the medical training system increase in influence on specialty choice. Preliminary data suggest that role models and the opportunity to practice a role reinforce or refute the beliefs and attitudes that students bring with them into the clinical setting and consequently effect their perception of the various medical specialties. However, the studies in this area have largely used self-selected groups of subjects who take an elective course in a clinical clerkship so that the results are not generalizable.

Career factors have not received adequate attention by researchers. Information on the career and personal goals of young doctors could provide clues as to the needed changes in the work setting to make shortage areas more attractive career options.

In sum, what the literature seems to indicate is that

(a) results with sociodemographic variables are mixed and therefore require further study; (b) even though there are methodological problems in preliminary studies, opportunities for interaction with role models and role

playing in the clinical setting holds promise for influencing specialty choice but further study is needed;






-52


and (c) insufficient data are available on career and personal goals of medical students from which to make important educational and policy decisions.















CHAPTER III
METHODOLOGY AND INSTRUMENTATION



Setting for the Study


The study was conducted at a large southeastern

university college of medicine, the University of Florida College of Medicine. Unless otherwise specified, the descriptive information pertaining to the setting for the study was summarized from the College of Medicine catalog (University of Florida, 1985). The University of Florida is a public, state-funded institution established in 1906. The College of Medicine, a component college of the J. Hillis Miller Health Center of the University of Florida, admitted its first class in 1956. College of Medicine educational offerings include undergraduate medical courses leading to an M.D. degree, graduate medical education experiences, residency programs in various specialties and subspecialties, graduate courses in the basic medical sciences leading to a Ph.D., and postgraduate fellowships in clinical and scientific areas.

Medical school faculty and students are housed in the J. Hillis Miller Health Center and the adjoining Shands Hospital located on the main campus of the University of Florida and in the Gainesville Veterans Administration




-53-






-54


Medical Center situated across the street from the Health Center. In addition to these sites, the College of Medicine has developed educational programs in community settings where students can gain insight into the day-to-day problems of minor and major illnesses as they occur in urban and rural settings. While university medical centers such as Stands Hospital play a unique and indispensable role in the student's educational experiences, fewer than 5% of physician/patient contacts occur in hospitals (White, Williams, & Greenberg, 1961). The Association of American Medical College's Panel on General Professional Education of the Physician (GPEP) (1984) recommended that medical students' general professional education be a balance between acute illnesses and working with patients and communities to prevent or ameliorate disease (p. 16). This requires ambulatory settings for required clerkships. The community health experiences available to University of Florida medical students include preceptorships with practicing physicians throughout the state of Florida, a rural health clinic in Dixie County, Florida, and family practice centers. The preceptors are primary care physicians who have volunteered to accept students for a specified period of time, usually two or four weeks of training. Generally the students work in the physician's office seeing patients with the physician. The student accompanies the doctor on hospital rounds and becomes involved in all medically related activities. Often the






-55


student is lodged in the preceptor's home so that she or he is provided an opportunity to view and understand nonmedical aspects of the physician's roles and obligations to the community as well as the usual medical responsibilities.

There are elective and required assignments in two

University of Florida affiliated family practice centers, the Family Practice Center in Gainesville and St. Vincent's Family Practice Center in Jacksonville. Here students work with residents in family practice training and faculty members to become acquainted with the health care delivery in urban areas.

The four years of undergraduate medical education

leading to the M.D. degree at the University of Florida are divided into three parts, Phases A, B, and C. During Phase A, the first 12 months of study, students are presented the core of basic medical science courses necessary for clinical training. The course schedule includes biochemistry, human genetics, anatomy, bacteriology, immunology, virology, physiology, neuroscience, human organ system development, and human behavior.

Phase B is divided into preclinical and clinical

experiences. The first 27 weeks of the second year students take pathology, pharmacology, physical diagnosis, and interviewing. A course in social and ethical issues is offered in the last part of the 27-week block preceding the clinical portion of Phase B.

The major portion of Phase B is the 12-month long

clinical clerkship period in which students rotate in groups






-56


of 20 among the 6 clinical specialty services, pediatrics, internal medicine, obstetrics/gynecology (ob/gyn), surgery, psychiatry, and family medicine. Students work as members of the medical patient care teams experiencing direct patient care responsibilities.

Phase C occupies the last 15 months of medical school. Students are required to complete four additional weeks of surgery and four weeks of medicine clerkships. They review clinical pharmacology, pathophysiology, and study infectious diseases for an eleven week period of time. The last 10 months are reserved primarily as elective time which allows students to select clinical activities of special interest and interview at the various residency programs where they are interested in applying for graduate medical education.



Revision and Development of the Questionnaire


Historical Background of the Questionnaire


The instruments used in this research project were adapted from two survey questionnaires used to study attrition from Florida's family medicine residency programs. In the mid-to late 1970s these residency programs were experiencing substantial numbers of residents who dropped out after one year of training (Dallman, Crandall & Haas, 1980). In order to study the issue of attrition among its residency programs, a cooperative project was begun in 1977






-57


by the Department of Community Health and Family Medicine at the University of Florida, the Florida Council of Graduate Education for Family Practice, and the Florida Academy of Family Physicians (Dallman et al., 1980, p. 833). One of the purposes of the residency study on attrition was to delineate characteristics of two groups of residents, the first group being those who after one year of training planned to continue on to completion of the residency program. The second group included residents who indicated at the end of the first year of residency that they planned to leave the residency program.

Dallman and his group (1980) mailed a questionnaire to all incoming residents in Florida Family Medicine Residency programs in August of 1977 and those entering in August, 1978. The questionnaire was designed to obtain data on each resident's social, familial, and educational background, as well as the resident's beliefs and attitudes concerning family medicine and primary care, and the importance the resident gave to financial security, autonomy, lifestyle, and social position.

The second questionnaire was mailed to each resident at the end of the first year of residency training. The same attitude and opinion items were included on the second questionnaire as appeared on the entering form except that the background items were omitted on the second survey (Dallman et al., 1980, p. 833).






-58


Analysis of the residency survey data demonstrated that a number of the questionnaire items discriminated between dropouts and nondropouts at the end of one year of family practice residency training. Daliman and his group (1980) found that while sociodemographic characteristics such as sex, marital status, age, race, citizenship, and hometown location and size did not discriminate between the dropouts and nondropouts, source of financial support did differentiate the two groups. Those residents who were their sole support were more likely to leave training than those who drew financial support from spouses or others (Dallman et al., 1980, p. 834). The researchers also found that the rank order of career goals differed between the two groups with dropouts more interested in economic security and time for family and leisure activities. The attitudes section showed a significant difference between the dropouts and nondropouts also. Those who left the programs felt curative medicine was more interesting than preventive or health maintenance (Dallman et al., 1980, p. 834). Based on the significant findings from the residency study and the imperative need for understanding of the factors influencing choice of family medicine and primary care specialties, it was decided to introduce the questionnaire earlier in the educational process to attempt to identify characteristics of those likely to choose these specialty areas over others. Thus, the present study concerning identification of factors related to student choice of a medical specialty was begun.






-59


Development of the Medical Student Questionnaire


The medical student questionnaire used in this study (Appendices A and B), and first administered in 1983, was essentially identical in content to the resident attrition survey conducted in 1977 and 1978. The changes to the items in the medical student questionnaire were principally to correct for differences in level of training between residency and undergraduate medical education. For example, the section in the residency questionnaire entitled "Family Practice Residency Content" allowed residents to address components of the ideal family medicine training program. In the medical student questionnaire this section was changed to "Medical Education Content" and retained some of the same choices of answers where appropriate.

The medical student questionnaire was designed to gather information from a number of broad categories. Section A, of the entering questionnaire, deals with background data. This section includes aspects of personal data, family, finances, education, and work experiences. Section B addresses obligations to federal agencies for loans and scholarships the student may have incurred and plans for future style of practice, e.g., solo, partnership, group. These two sections required the student to choose between alternative answers by placing a check or an X in the space beside the alternative of their choice. Some of the questions required the respondent to write in a short






-60


answer as when asked age and major subject of undergraduate education.

The importance the student placed on financial

security, autonomy, lifestyle, social position in the community, and position in the peer group was examined in Section C, Career Goals. Students were asked to rank in order of importance items that relate to career goals. The most important goal was assigned the rank of 1, the second most important goal received a rank of 2, and so forth with no choice being used more than once.

Section D, Medical Education Content, lists eight items pertaining to medical education. Students ranked the items from 1 to 8 according to how important each was considered to be in producing the best physicians.

Improving Health Status, Section E, lists five

different approaches to improving the health status of people not receiving adequate health care. The student ranked them in order of his/her opinion as to their effectiveness with the rank of 1 equaling the most effective and the rank of 5 being least effective in improving health.

The Attitudes Toward Primary Care and Family Medicine

section, Section F, contains 29 statements which reflect the goals and philosophy of primary care and delineate family medicine and primary care from specialty-oriented medicine. The attitude section concentrates on the following areas:

(a) preventive medicine, (b) continuity of care, (c) managing psycho-social problems, (d) ambulatory versus






-61


hospital-based care, (e) broad spectrum of medical care,

(f) attitudes toward consultants and paramedical personnel, and (g) perspectives toward research in primary care settings.

Student respondents were asked to indicate whether

they strongly agree, tend to agree, are not sure, tend to disagree, or strongly disagree with each of the 29 statements on attitudes and beliefs concerning primary care and family medicine. A Likert scale of numerical values was assigned to each response depending on the degree of agreement or disagreement. Strongly agree received a value of I while strongly disagree was valued at 5.

The last section of the entering survey of medical students asked the respondent to indicate to the best of his/her knowledge future plans for residency training. The alternatives from which the student could choose were (a) family medicine, (b) pediatrics, (c) general internal medicine, (d) other primary care, (e) other specialty/subspecialty, or (f) uncertain. Students answering under the category "other primary care and other specialty/subspecialty" were asked to specify the specialty area by writing in a line beside the choice indicated. The last section was included on all of the entering questionnaires except the questionnaires for Rotation 2. It was inadvertently omitted from this group's questionnaires. Thus, 20 students did not have the opportunity to express their specialty preferences. A comparison of the sociodemographic characteristics of the students who did not






-62


respond to this section revealed no significant differences at p .01 between those who responded and those who did not.

The follow-up survey questionnaire items were identical to the entering survey with the exception of the deletion of the first two sections containing social, personal, and demographic data. The last section on future plans for residency training was also not included in the follow-up survey. The abbreviated follow-up survey included four sections common to the entering questionnaire: Career Goals, Medical Education Content, Improving Health Status, and Attitudes Toward Primary Care and Family Medicine.



Administration of the Questionnaire Sample


The sampling frame for the study included all medical students in the third year of required clinical clerkships in 1983. Third-year medical students were chosen because they were further along in the professional and socialization process and because fourth-year students are geographically scattered due to the nature of the curriculum which allows for externship at other sites and time out of school for interviews for residency positions.

During the clinical portion of Phase B students rotate in small groups through the individual clinical courses or clerkships. These clerkships were scheduled in blocks of eight weeks each. The Community Health and Family Medicine clerkship was allotted six weeks of one of the eight-week






-63


blocks sharing the other two weeks with a neurology clerkship. The Community Health and Family Medicine clerkship ran concurrently with the students' clinical experience in medicine, pediatrics, psychiatry, surgery, and ob/gyn.

In the Community Health and Family Medicine rotation

all clinical experiences were in settings outside of the J. Hillis Miller Health Center and Shands Hospital. The six weeks were scheduled as follows: four weeks in a family practice center (either the Family Practice Center in Gainesville or St. Vincent's Family Practice Center in Jacksonville) and two weeks in a rural health clinic either in Cross City or in Mayo, Florida. Participation in the various community settings provided the students with firsthand experience in the medical and health problems as they exist in the different communities.

The rotation year begins in March each year and

continues for 12 months. The schedule for the 1983-84 rotation year was as follows:

Rotation 1 March 21-May 14, 1983
Rotation 2 May 15-July 9, 1983
Rotation 3 July lO-Sept. 3, 1983
Rotation 4 Sept. 11-Nov. 5, 1983
Rotation 5 Nov. 6, 1983-Jan. 14, 1984
Rotation 6 Jan. 15-March 9, 1984

Approximately three weeks prior to the beginning day of the rotation, an orientation meeting was held for the students. The goals and objectives were explained, the course syllabus was distributed, and the evaluation system was discussed. The sequencing of the various clinical






-64


experiences in Community Health and Family Medicine was explained. A description of the sites was provided in the syllabus and this was reiterated during the orientation. Students were given the opportunity to ask questions about lodging accommodations, travel arrangements, and any other concerns. Assignment to the clinical sequence was by lottery.



Administration of the Entering Survey


At the beginning of the orientation session, after all students had arrived and taken a seat, the entering survey questionnaire was passed out to each student. Attached to the front of the questionnaire was a cover letter (Appendix C) signed by the chairman of the Department of Community Health and Family Medicine stating the purpose of the survey and assuring confidentiality of responses. Students were encouraged to participate but were told that the participation was voluntary. The author thanked the respondents for their help with the study.

The written information was verbally reinforced as the questionnaires were distributed. Students were told (a) a follow-up questionnaire would be mailed to them at their campus mailboxes or at the clinical site at the end of the rotation; (b) to put their name and social security number on the questionnaire but that a code number would be assigned when the data were entered in a computer--this would maintain confidentiality while allowing the following






-65


data to be collated with the entering data; and (c) participation or nonparticipation would not affect their rotation grade as those clinical faculty members responsible for grading would not have access to individual's responses. It required approximately 30 minutes for students to complete the questionnaire.



Follow-Up Survey Questionnaire


The follow-up administration of the survey

questionnaire was accomplished at the end of the rotation time period. The students were not told that the follow-up questionnaire was identical to the entering survey with the exception of the background data questions. The survey questionnaire was included in the envelope with the student evaluation of rotation forms. Two self addressed envelopes were included for the respondent to return the evaluation of rotation and the survey questionnaire. A cover letter (Appendix D) stated that the College of Medicine required student evaluation of rotation. It reminded students that they had participated in the entering survey and asked their cooperation in the follow-up survey.

A follow-up code sheet was used to record the name,

social security number, and code number for each respondent. When the follow-up form was returned an X was placed in the appropriate column. Those students who did not respond within two weeks were sent a note reminding them to send the questionnaire back. If another two weeks passed with no






-66


response, another note (Appendix D) was sent with a copy of the questionnaire. The envelopes enclosed in the student correspondence were addressed to a campus mailbox and therefore required no postage if mailed at a university post office. Students returned the questionnaires in the envelopes provided. Since most students were physically located at the Health Center where access to campus mail drop boxes were readily available, this was felt to be an acceptable method of securing their return.



Residency Match


All students who enter the profession of medicine must at some point make a secondary career choice--the type of medical specialty they wish to practice. Training in the medical specialty of choice is acquired through residency programs. Students begin applying to residency programs and interviewing at the various sites by the beginning of the fourth year of medical school. The final selection is made through a nationwide, computerized matching system, the National Resident Matching Program (NRMP). A majority of students participating in the match process are placed in a residency program of their choice although it may not be their first ranked preference. In January of their senior year in medical school, students ranked the residency programs in which they wanted to take their first year of graduate training. The students' selections are matched against the residency programs' choices of candidates.






-67


Results of the match were received by the individual students, the residency programs, and the medical schools the following March.



Research Design


To investigate the research questions for this study, a pre- and post-research design was used with multiple group replication measures. The research procedure involved observations taken immediately before and after a clerkship and again after a passage of time. The design used intact groups of third-year medical students previously formed into rotation groups through a lottery system.

The design for this study was considered a variation of the Recurrent Institutional Cycle Design, a strategy for field research described by Campbell and Stanley (1963) as appropriate to those situations in which some aspect of the institutional process is on a cyclical schedule. In field research in educational settings where restrictions prevent an experimental design with control of who would be exposed to the experimental variable and the effects of a global and complex construct are sought, this design offers a measure of strength over the one-group pretest-posttest design. The fact that the treatment variable, the clerkship, is continually being presented to a group previously exposed and a group about to be exposed makes for some degree of experimental control as elements of the cross-sectional and longitudinal approaches are present.






-68


The Recurrent Institutional Cycle Design (Campbell &

Stanley, 1963) as adapted for use in this study is shown in Figure 1.

The design for this study required the collection of

preclerkship data from each medical student at the beginning of the family medicine rotation. Following preclerkship data collection, each rotation group received a learning exposure to family medicine. A second administration of the questionnaire occurred immediately following the family medicine rotation. The final component of the design involved residency match choices for each study participant.

The cyclical schedule for this design was such that at one and the same time a group which had been exposed to the clerkship (X) and a group which was about to be exposed were measured. As Figure 1 demonstrates, subjects in Rotation 1, which began in mid-March, were completing the follow-up questionnaire in mid-May at approximately the same time as the subjects in Rotation 2 received the entering questionnaire. Each subsequent rotation began a new cycle as the previous rotation ended their clerkship. Students ranked the residency programs for the National Resident Matching Program in January of their senior year of medical school. Residency match choice information was received at the sponsoring institution in March, 1985.







-69
Ln


< < < < "0

IT





0 0 0 0

-Z

--o


U)



ul
4J

0
A4
r_:
0
EA -4
C) 4J
z
0
0
CD u
V-1 PW fo
4-)
0 ..t-D
04
00
Un 0
ri cf, Q)
4 P >1 z 4
w a) r-j 4-)
u z Q) 1 0
w ro C) PP
1: Ell C) TU)
(1) 04 Q) () 41 4 4J 4
-rq :3
0 (15 LH En < (13
-H Z 0 0 0 4
4-) 1:1 r-4 U) 41 Ln
fa 0 -4 m Ej r : C>
4-) >1 -,q 0 (0 0) (1)
r r-A 4J rZ4 -ri 4 rO ()
(1) -rq U) 9 tyl -H -4
-r-A r4 Q) ro r. 0 ul 0
P (0 ::S : fU 4 Q) 4
0 44 CR cc Aq u m El I
CD ..z


x 0



00
m
Ln M CN



P 0 E-4 4PHOZ U)w CY D 41 z u w






-70


Data Analysis


One important criterion in the choice of which

statistical procedure to use to test a research hypothesis is that the data meet the assumption of a particular level of measurement (Marks, 1982) The elements of observation in this study were considered nominal- and ordinal-level data. When words or symbols are used to classify persons or characteristics into groups, they constitute a nominal scale. In this research, sociodemographic characteristics, specialty preference groups, and specialty choice groups were identified as being nominal-level variables. With ordinal-level data, the classes or categories stand in relation to each other so that rank ordering is permitted. Students' rankings of career goals and their attitudes toward primary care measured on a Likert scale resulted in information contained in an ordinal scale.

Since nonparametric statistical tests are appropriate for data inherently in ranks or numerical scores having the strength of ranks, and some apply to nominal-level data as well, techniques of inference drawn from this body of methods were chosen for this study. The following nonparametric statistical procedures were used to test the 10 null hypotheses:

1. Hypothesis 1, there is no association between sociodemographic traits of medical students and their specialty preferences in the third year of medical school, was tested using the chi-square test (X2). The chi-square






-71


test is a statistical technique that tests the difference between what is observed and what might be expected under some theoretical model. It can be applied to one-sample cases, two-sample cases, or larger cases such as the three subsamples of specialty preference contained in this hypothesis. The three specialty preference response variables were primary care, subspecialty, and undecided. Chi-square is an appropriate statistical tool when frequencies in classified categories constitute the data of research such as the sociodemographic characteristics of this study.

2. The chi-square test for two independent samples was applied to Hypothesis 2 to determine if significant associations existed between sociodemographic characteristics of the medical students and their medical specialty choices as evidenced by residency selection of either primary care or nonprimary care specialties.

The chi-square test is valid only if at least 80% of the cells in a contingency table have frequencies of at least five. Where small cell frequencies are involved, categories may be collapsed as long as it is done logically and not merely to achieve significance or create bias. Some of the sociodemographic characteristic categories were collapsed for purposes of statistical analysis. When small cell frequencies persisted, Fisher's exact test was used as an alternative test of significance for 2 x 2 tables (Champion, 1970). The Agresti-Wackerly method of computing






-72


exact tests of independence was used for R x C crossclassification tables when R or C was greater than two (Agresti & Wackerly, 1977).

3. Hypothesis 3 tested the null hypothesis that there is no difference between medical students grouped by sociodemographic characteristics according to their attitudes toward primary care before or after a family medicine clerkship. Individual attitude scores were formed by summing each student's responses to the Attitudes about Primary Care and Family Medicine section of the questionnaire. Attitudes were measured before and again after a family medicine clerkship. The Kruskal-Wallis oneway analysis of variance was used to determine whether or not there was a difference in distribution of attitude measurements between the groups of students identified by their sociodemographic characteristics. The primary assumption for this test is that the data be ordinal-level data. The data to be ranked should be continuous rather than discrete. The Kruskal-Wallis makes no assumptions concerning distributions and it is appropriate for unequal sample sizes, both large and small. Based on these assumptions it was considered a suitable test for this hypothesis. When there are more than five subjects in each element or group, as occurred in this study, the KruskalWallis H value is treated as a X2for interpretive purposes (Champion, 1970, p. 188).

4. The Kruskal-Wallis one-way analysis of variance

test was again deemed the appropriate choice of statistical






-73


analysis for Hypothesis 4 which sought to determine the difference between the three specialty preference groups in the third year of medical school and the two specialty choice groups from the fourth year with respect to career goal rankings. The career goal data met the primary assumption of this procedure, namely that it be ordinallevel measurement as well as the assumption that the samples were independent.

5. Spearman's rho measures the degree of association between two sets of ordinal-level data. Therefore, for Hypothesis 5 it was considered an appropriate statistical procedure to investigate whether or not medical students' rankings of career goals were associated with their attitudes toward primary care before a family medicine clerkship and again, after the family medicine clerkship.

6. Hypothesis 6 was tested with the Kruskal-Wallis

technique. In the computation, the difference in students' rankings of career goals before and after a family medicine clerkship were obtained. These difference scores were then entered into the Kruskal-Wallis test to decide whether the medical specialty groups were from different populations. Since the scores could be considered at least ordinal data, the Kruskal-Wallis was appropriate to the data.

7. Hypothesis 7 tested whether the association between medical students' third-year specialty preferences and their fourth-year specialty choices was different from what might be expected by chance. As the data were in discrete






-74

categories at the nominal level, the chi-square test was a suitable one.

8. Hypothesis 8, there is no difference in medical

students' attitudes toward primary care before and after a family medicine clerkship, was tested with the Wilcoxon matched-pairs-signed-ranks test. In this case, the medical students acted as their own controls in a before-and-after design. Another assumption is that the level of measurement be interval-level. As the attitude scores before-and-after the clerkship were ordered in a numerical scale, the information was judged to be at least equal-appearing. A definite advantage of the Wilcoxon test is that it considers not only the direction of the difference in scores but also the magnitude of the change. As Champion (1970) suggested for sample sizes exceeding 25, an alternative formula, a modification of the Z test, was used to complete the Wilcoxon matched-pairs-signed-ranks test (p. 169).

9. The Kruskal-Wallis one-way analysis of variance

test was judged to be the proper procedure to test whether attitude scores of one specialty preference group were significantly different from the two other specialty preference groups of third-year medical students. Since the three groups of specialty preference could be considered independent and the attitude scores represented at least ordinal-level data, the Kruskal-Wallis test was considered the correct method for Hypothesis 9.

10. Hypothesis 10 was also tested with the KruskalWallis test. The hypothesis was that there is no difference






-75


between groups of medical students identified by their fourth-year specialty choice with respect to their attitudes toward primary care before and after a family medicine clerkship. The two groups of specialty choice constitute independent groups and the measure of attitudes before and after the clerkship constitute at least ordinal-level data; therefore, the Kruskal-Wallis was concluded to be the right procedure.

When multiple statistical tests are conducted on data as they were in this study, the chance of making erroneous conclusions is increased and, therefore, the overall significance level (OSL) must be set conservatively. The significance level of < .01 was used for hypotheses tested in this study so that the OSL would not exceed .10.

Marks (1982) discusses two philosophies concerning the significance level used for testing the null hypothesis (p. 22). One philosophy is to perform the statistical test using a predetermined level of significance and to use the test results only to see if the predetermined level of significance is achieved. Another approach is to present the attained significance level, decide whether or not the evidence is sufficient to reject the null hypothesis, and let others draw their own conclusions from the results. The second rationale was used in presenting the results of this study.















CHAPTER IV
PRESENTATION OF RESULTS



Sociodemographic Characteristics of the Study Group


Data reported here were obtained from a survey of 109

medical students of the 1984-85 graduating class of a statesupported medical school in the south/southeast. The data are presented in Table 1. Some categories from the original survey have been combined for purposes of statistical analysis and ease in reporting. Students ranged in age from 22 to 35 with a mean age of 25 years. Males made up 73% of the respondents; females, 27%. Whites were heavily represented with 88% of the group while all other races combined accounted for the additional 12%. The religious preferences of the group were (a) 44% Protestant, (b) 22% Catholic, (c) 11% Jewish, and (d) the remainder indicated no religious preference. There was almost no variation between the percentages of students indicating that religion played a great deal of importance in their lives (37%), some importance (35%), and very little to none at all (28%).

Eighty-two percent of the students were from the south/southeast region of the country with about equal representation from small towns, mid-sized cities, and metropolitan areas. Forty-seven percent of the subjects



-76-








-77












7v (D 0 (0 to o 00 m r- C> 0 V) 1r)
Lo t- co r- 000 Go -r 00 -4 r4 a)
1-4 r-q N -7v Cq -I t- t- N DO 14


co N 00 D t- 0 001-4 0) to co cl v 1-4 'r
r-4 r-4 lzr co m to cn :)o cl) t- m r-4




1-4
1-4
0 CD
u
a
C, r- r.
P Q) 0 (D CD
_r_ r. u Z .,4 0 41
CL- r- 0 0 u a 4.1
n 0 C. a -H Q) <
E r-4 (D 41
CD
Q)
Q)
0 41 J9 4J
-4 Q) 04 0 4,1 r-4 a) -4
'i A 0
j A o bHr Q) 0
J 4 0 41
Cd H 0 f .14 > Cd 0 c a) w 0
-4 1-4
4-)
U
W
4.1


ll C O C 1 C! C 1 1 C 0! C
N t- t- cq C13 t- 08 1-4 T 00 r-4 r-4 00 00 H
co m m LO -V U) Ce) cl) t- 00 1-1 C)

C'l to C') CD r-I r-I t- tD 0 t- 3
o t- 0 U') to ce) m t- 00 C11 rU) 4.1
U

41
U)
.14 Cc 0
.1 41 P I
4 Q) 4
(A Cd a cl (3) 1-4 (1)
ri 0) CL d 41 a '0 Cd U)
4J CL c 0 :3 -1 z
0 0 0)
U) Cd a
4 41
4 -H T ED (1) 8 0 0 c -0
Q) r a) U' -" g: 9 bl)
u cs z Q) 10 1-- 1 :3 41 ID C 41 W 41
0 m +, N p A
:: 11 0) 2 A 0 0. -, a C: .0 6,
41 4--) +1
u



04 4.1


c
m to t- 04 cli t- t- 0 0 r-I tlo I-- N 00 1-4 00 r-i V. r-i M v
5
(n m cn m rq Iq t- to C11 a)
a) r. I L- Cl) H oo 1-4 0 li CO
ro
0

or.
0 Q)

4.1 r.
r 0
4-4 0 0 >1 0
0 -1 9 a,
cd w C r. U
cd 0
8 8 fi -N
4 U) tk In .01 E
U) 0 m -4
z 0
-H 0) r-q Cd bjD
g r ') cd
E 0 4j :j a) Cd
-H Al 4 o r-4 S a) .0
0 >
(D






-78


judged their hometown economic base as a mixed economy. Thirty-six percent came from places dominated by a business economy. Farming communities contributed 11% to the sample and only 6% were from industrial areas.

When asked to indicate their parents' occupations while they were growing up, 52% of the students checked physician, health care provider, or other profession. Forty-eight percent cited business, clerical/sales, skilled, unskilled, farm/farm worker, and other as their father's occupation. Mothers' occupations were equally divided between housewife and working outside the home. Considering their parents' annual income, 69% had parents with incomes that fell in the $50,000 or less categories compared with 31% in the above $50,000 stratum.

Seventy-one percent of the students indicated that their principal source of support came either from themselves, spouses, or a combination; parents supported 29%. The financial status is further clarified by the information that 81% of the respondents reported some educational debt. For those reporting indebtedness, 92% had debts greater than $5000.

For the majority of the students, their education took place in public high schools (78%) and public-supported colleges (71%). As would be expected, 86% majored in science with 91% obtaining a bachelor's degree and 9% achieving a graduate degree.






-79


Specialty Preferences


Prior to a third-year Family Medicine clerkship, the medical students were asked to indicate their future plans for residency training. General internal medicine, pediatrics, family medicine, and other primary care specialties were the most frequently chosen areas with 47% of the group. About 24% expected to go into subspecialty residencies and 29% were still undecided on a specialty at that point. Table 2 shows the students' career plans by their preferred specialties.



Residency Plans


The students' first choice of career plans by residency choice are presented in Table 3. Primary care specialties were the first choice of career activity of 66% of the students. Thirty-six percent planned residencies in internal medicine with 14% choosing family medicine, 7% pediatrics, and 9% transitional residencies. Approximately 33% of the 1985 graduates elected subspecialty residencies. The most frequently named subspecialty was obstetrics and gynecology with 10% of the subjects. Slightly less than 4% opted for residencies in surgery, radiology, and pathology.





-80






Table 2

Medical Specialty Preferences of Third-Year Medical
St-udents



Specialty Preference Groups n Percent


Primary care 34 47.2

Specialty-subspecialty 17 23.6

Undecided 21 29.2


Total 72 100.0






-81






Table 3

Residency Choice of Fourth-Year Medical Students



Specialty Choice Groups n Percent


Primary Care
family medicine 15 13.8
pediatrics 8 7.3
internal medicine 39 35.8
transitional 10 9.2

Total 72 66.1


Subspecialty
ob/gyn 11 10.1
ER medicine 1 .9
anesthesiology 2 1.8
neurology 1 .9
dermatology 1 .9
pathology 4 3.7
radiology 4 3.7
radiation therapy 3 2.7
physical med. & rehab. 1 .9
surgery 4 3.7
plastic surgery 1 .9
neuro surgery 1 .9
orthopedic surgery 1 .9
psychiatry 1 .9

Total 36 32.9






-82


Results of the Data Analysis for Testing the__Hypotheses


Hypothesis 1


There is no association between sociodemographic characteristics of medical students and their medical specialty preferences expressed in the third year of medical school. The chi-square test was used to determine

separately for each sociodemographic variable whether medical specialty preferences of third-year medical students were dependent on sociodemographic characteristics. Medical specialty preferences were categorized as primary care,

subspecialty, and undecided based on the students' responses to the questionnaire item, Please indicate to the best of your knowledge your future plans for residency training.

Since multiple statistical tests were performed, a conservative overall significance level of p = .01 was used to decrease the probability of falsely rejecting the null hypothesis. The results of the chi-square procedures used to test the first set of variables related to Hypothesis 1 are reported in Table 4.

The overall significance level of < .01 was not

achieved in the comparison of sociodemographic variables and third-year specialty preferences of medical students. Therefore, the null hypothesis was not rejected.

However, one sociodemographic category, parents' annual income, was significantly related to specialty preference at

the < .01 level. This test produced a chi-square value





-83



Table 4

Association between Third-Year Specialty Preference and Sociodemographic Characteristics



Sociodemographic X2 Significance
Characteristics df Vaue


Sex 2 5.513 .06
Marital Status 2 .946 .62
Religious Orientation 8 3.097 .94+
Religious Importance 4 6.027 .20
Race 2 4.010 .16+
Citizenship 2 1.133 1.00+
Hometown Size 6 3.134 .79+
Hometown Location 2 2.218 .39+
Hometown Economic Base 8 6.867 .61+
Father's Occupation 2 7.727 .02
Mother's Occupation 2 1.312 .52
Parents' Religion 8 7.291 .51+
Parents' Income 2 15.864 .0004***
Source of Present Support 2 4.412 .11
Amount of Debt 2 1.780 .56+
Indebtedness 2 3.279 .21+
Public or Private High School 2 3.432 .20+
Public or Private College 2 5.080 .08
General Health 2 3.281 .24+
Family's Medical Care Provider
Family Practitioner/General 2 2.257 .32
Internist 2 4.331 .11
Pediatrician 2 1.678 .43
Other 2 .825 .70+
Work Experiences
Medical 2 3.531 .19+
Technical (not medical) 2 .756 .69
Helping Activities 2 .705 .70+
Public (sales, etc.) 2 3.309 .22
Teaching 2 1.957 .37+
Unskilled 2 .873 .65
Other 2 2.068 .37+
Level of Education 2 1.788 .52+
College Major 2 4.013 .16+


< .05
** < .01
***P < .001

+Agresti-Wackerly Significance Level






-84


significant at the .0004 level. It should be noted that the association between father's occupation and specialty preference attained a probability of .02.

Father's occupation. Student responses to the item, indicate your parents' occupations whileyou were growing

up, were used to form two occupation groups for father's occupation: (a) profess ional/technical and (b) business/ other. The professional/technical occupational group

included the students' responses to the categories of physician, other health care occupation, and other professional work. The group labeled as business/other was

composed of responses to occupational categories of business, clerical/sales, skilled worker, farmer/farm worker, and other. The two occupational groups were compared with the three medical specialty preference levels:

(a) primary care, (b) subspecialty, and (c) undecided. The chi-square statistic, X2 (df 2, N = 72) = 7.727, p = .02, indicated that the probability of the observed association between medical specialty preference groups and fathers'

occupations being a function of chance was 2 in 100 or less.

Table 5 shows that for students whose fathers were in professional/technical occupations when they were growing up, 33% preferred primary care specialties, 26% indicated subspecialty preferences, while 41% were undecided as to their eventual specialty choice. Of the third-year students whose fathers engaged in business and other types of work, 64% preferred primary care, 21% indicated subspecialty





-85






Table 5

Association between Medical Specialty Preferences and
Father's Occupation and Parents' Annual Income



Specialty Preferences


Sociodemographic Primary Care Subspecialty Undecided
Characteristics n Row % n Row % n Row %


Father's Occupation
professional/
technical 13 33.33 10 25.64 1.6 41.03
business/other 21 63.64 7 21.21 5 15.15

Parents' Income
> $50,000 9 20.00 25 55.00 11 25.00
< $50,000 12 62.22 2 11.11 6 26.67






-86


practice, and only 15% indicated that they had not made a decision. Thus, the greatest differences were in the primary care and undecided medical preference levels for the two groups. Almost twice as many students whose fathers were in business/other indicated primary care preferences as

did those from the prof essional/technical group. A much larger percentage of those from the professional/technical group indicated a delay in decisions on specialty choice.

The importance of this large percentage of undecided

students from the prof essional/technical background can be seen in the results from testing Hypothesis 2 when the

actual specialty choice was made in the fourth year (Table 6). While the percent choosing residencies in subspecialty areas increased, the most dramatic gain was made in the professional/technical group where the percentage of

students selecting primary care went from 33.3% to 62.5%.

Parents' annual income. The findings related to

parents' annual income and third-year specialty preferences were consistent with the data on fathers' occupations. The parents' annual income question asked the students to choose from among four income categories: (a) less than $10,000,

(b) $10,000 to $25,000, (c) $25,000 to $50,000, or (d) above $50,000. Two income groups were formed from these four categories for comparison with the three specialty preference levels. The two income groups were divided between responses above and below $50,000. The chi-square





-87






Table 6

Association between Medical Specialty Choice of
Fourth-Year Students and Father's Occupation



Specialty Choices


Primary Care Subspecialty
Father's Occupation n Row % n Row %


Professional/technical 35 62.50 21 37.50

Business/other 36 70.59 15 29.41






_88


statistic, X2 (df 2, N = 65) = 15.864, p = .0004, led to the conclusion that there was a difference in preference for medical specialties of these third-year medical students

based on parents' income.

Of the students from the more affluent families (i.e., annual incomes greater than $50,000), 20% preferred primary care, 55% indicated that they wanted to pursue subspecialty practice, and 25% were undecided. Those students whose

parents' annual income was equal to or less than $50,000 chose primary care more often than any other of the preference levels. Sixty-two percent chose primary care, 11% preferred subspecialties, and 27% were undecided. The percentage preferring primary care was three times greater among the lower income students. A wider difference between the income groups occurred in their preference for subspecialties. The percent of undecided students was fairly consistent for both groups. These data are presented in Table 5.



Hypothesis 2


There is no association between sociodemographic characteristics of medical students and their medical specialty choices as evidenced by fourth-year residency

selection. The procedure for analyzing this hypothesis was the chi-square test of significance. The results of the analysis for Hypothesis 2 are listed in Table 7.





-89



Table 7

Association between Fourth-Year Specialty Choice and
Sociodemographic Characteristics



Sociodemographic X2 Significance
Characteristics df Value


Sex 1 .377 .54
Marital Status 1 1.506 .22
Religious Orientation 4 1.640 .80
Religious Importance 2 1.219 .54+
Race 1 -- .54
Citizenship 1 --.26+
Hometown Location 1 .128 .72
Hometown Size 3 2.228 .52
Hometown Economic Base 4 6.326 .18#
Father's Occupation 1 .782 .38
Mother's Occupation 1 2.667 .10#
Parents' Religion 8 3.046 :57#
Parents' Income 1 4.602 *Q3*
Source of Present Support 1 .090 .76
Indebtedness 1 .711 .40
Amount of Debt 1 -- .67k
General health 1 -- 1.00k
Public or Private High School 1 1.453 .23
Public or Private College 1 .153 .70
Family's Medical Care Provider
Family Practitioner/General 1 1.212 .27
Internist 1 .266 .61
Pediatrician 1 3.142 .08
Other 1 .000 1.00
Work Experiences
Medical 1 .000 1.00
Helping Activities 1 .000 1.00
Technical 1 .019 .89
Teaching 1 .026 .87
Public 1 .928 .34
Unskilled 1 3.157 .08
Other 1 .559 .45
College Major 1 7.338 0*
Level of Education 1 -- :30+


*p< .05 (* .01
+ Fihr' Exact Test

#Agresti..wackerly Significance Level






-90


In the comparisons of fourth-year medical students' specialty choices with their sociodemograpbic

characteristics, the overall significance level of < .01 was not achieved. It was concluded that there was insufficient evidence to reject the null hypothesis.

One sociodemographic variable, college major, was significantly related to the medical students' specialty choice at the p < .01 level of significance. A second comparison using the sociodemographic variable, parents' annual income reached a significance level of .03. For these two variables, a discussion of the results is as follows.

Parents' annual income. The parents' annual income category asked the students to indicate their parents'

combined income by checking one of four income categories. As in Hypothesis 1, two income groups were formed from the four categories for comparison with the two medical

specialty choice levels, primary care and subspecialty practice. The two income groups were (a) greater than

$50,000 annual income and (b) annual income equal to or less than $50,000. The resulting chi-square statistic was X2(df 1, N = 99) = 4.602, p = .03. The percentage of higher level income students choosing primary care and subspecialty residencies was nearly equal with approximately 52% choosing primary care and 48% going with subspecialty residencies. A

considerably higher percentage (74%) of students from the $50,000 or less income group chose primary care than chose residencies in subspecialty areas (26%). The results are presented in Table 8.






-91






Table 8

Association between Medical Specialty Choice of Fourth-Year
Students and Parents' Annual Income and College Major



Specialty Choices


Sociodernographic Primary Care Subspecialty
Characteristics n Row % n Row %


Parents' Annual Income
> $50,000 35 51.61 33 48.39
< $50,000 23 73.53 8 26.47


College Major
Science 65 72.22 25 27.78
Other 5 35.71 9 64.29




Full Text
-59-
Development of the Medical Student Questionnaire
The medical student questionnaire used in this study
(Appendices A and B) and first administered in 1983, was
essentially identical in content to the resident attrition
survey conducted in 1977 and 1978. The changes to the items
in the medical student questionnaire were principally to
correct for differences in level of training between
residency and undergraduate medical education. For example,
the section in the residency questionnaire entitled "Family
Practice Residency Content" allowed residents to address
components of the ideal family medicine training program.
In the medical student questionnaire this section was
changed to "Medical Education Content" and retained some of
the same choices of answers where appropriate.
The medical student questionnaire was designed to
gather information from a number of broad categories.
Section A, of the entering questionnaire, deals with
background data. This section includes aspects of personal
data, family, finances, education, and work experiences.
Section B addresses obligations to federal agencies for
loans and scholarships the student may have incurred and
plans for future style of practice, e.g., solo, partnership,
group. These two sections required the student to choose
between alternative answers by placing a check or an X in
the space beside the alternative of their choice. Some of
the questions required the respondent to write in a short


-87-
Table 6
Association between Medical Specialty Choice of
Fourth-Year Students and Father's Occupation
Specialty
Choices
Father's Occupation
Primary Care
n Row %
Subspecialty
n Row %
Professional/technical
35
62.50
21
37.50
Business/other
36
70.59
15
29.41


-92-
College major. The questionnaire provided space for
the students to write their major subject while in
undergraduate college. As shown in Table 8, the majority
of the students indicated science as a major subject. All
nonscience majors were included in the category labeled
other. Of the science majors, 72% chose primary care
residencies. The highest percentage, 64%, of the nonscience
majors, went into subspecialty residencies. The chi-square,
(d_f 1, N = 104) = 7.338, £ = .0067, led to the conclusion
that when college major was tested for independence with
specialty choice, there was a significant difference between
the science majors and nonscience majors with respect to
specialty choice in this group of students.
Hypothesis 3
There is no difference between medical students grouped
by sociodemographic characteristics according to their
attitudes toward primary care measured before or after a
third-year family medicine clerkship. Responses to 29
attitudinal statements were collected both before and after
a clerkship in family medicine. For purposes of analysis
the Likert scale rankings for each attitude statement were
ordered so that higher ranks indicated a favorable attitude
toward primary care and family medicine. To assess the
overall attitudes for individual students, the Likert scale
ranks were summed. The Kruskal-Wal 1 is one-way analysis of
variance was used to test this hypothesis at a significance


-66-
response, another note (Appendix D) was sent with a copy of
the questionnaire. The envelopes enclosed in the student
correspondence were addressed to a campus mailbox and
therefore required no postage if mailed at a university post
office. Students returned the questionnaires in the
envelopes provided. Since most students were physically
located at the Health Center where access to campus mail
drop boxes were readily available, this was felt to be an
acceptable method of securing their return.
Residency Match
All students who enter the profession of medicine must
at some point make a secondary career choicethe type of
medical specialty they wish to practice. Training in the
medical specialty of choice is acquired through residency
programs. Students begin applying to residency programs and
interviewing at the various sites by the beginning of the
fourth year of medical school. The final selection is made
through a nationwide, computerized matching system, the
National Resident Matching Program (NRMP). A majority of
students participating in the match process are placed in a
residency program of their choice although it may not be
their first ranked preference. In January of their senior
year in medical school, students ranked the residency
programs in which they wanted to take their first year of
graduate training. The students' selections are matched
against the residency programs' choices of candidates.


-Ill-
administered in the third year asked the medical students to
indicate to the best of their knowledge their future plans
for residency training. Their responses were categorized
into three specialty preference groups: (a) primary care
which included family medicine, pediatrics, and general
internal medicine and other primary care; (b) other
specialty/subspecialty; and (c) undecided. The fourth-year
specialty choices were derived from the residency match
information. These two specialty choice groups were
identified as primary care and subspecialty.
Eighty-five percent of the students who preferred
primary care in the third year selected residencies in
primary care while the remaining 15% chose subspecialties
instead. Of those initially indicating an interest in
subspecialty areas of medicine, 76% chose a subspecialty
residency. Fifty-seven percent of those undecided in the
third year ultimately selected primary care residencies.
Using Goodman and Kruskal's lambda statistic as a measure of
association, there was a 33% improvement in ability to
predict residency choices given that specialty preferences
in the third year was known. The results of the data
analysis are demonstrated in Table 15.
The chi-square test statistic produced by the analysis
of the data was X2 = 18.810 (df 2, N = 72) = £ .001. This
led to the rejection of the null hypothesis and the
conclusion that residency choices for this sample were
related to third-year preferences.


-64-
experiences in Community Health and Family Medicine was
explained. A description of the sites was provided in the
syllabus and this was reiterated during the orientation.
Students were given the opportunity to ask questions about
lodging accommodations, travel arrangements, and any other
concerns. Assignment to the clinical sequence was by
lottery.
Administration of the Entering Survey
At the beginning of the orientation session, after all
students had arrived and taken a seat, the entering survey
questionnaire was passed out to each student. Attached to
the front of the questionnaire was a cover letter (Appendix
C) signed by the chairman of the Department of Community
Health and Family Medicine stating the purpose of the survey
and assuring confidentiality of responses. Students were
encouraged to participate but were told that the
participation was voluntary. The author thanked the
respondents for their help with the study.
The written information was verbally reinforced as the
questionnaires were distributed. Students were told (a) a
follow-up questionnaire would be mailed to them at their
campus mailboxes or at the clinical site at the end of the
rotation; (b) to put their name and social security number
on the questionnaire but that a code number would be
assigned when the data were entered in a computerthis
would maintain confidentiality while allowing the following


FACTORS RELATED TO STUDENT CHOICE
OF MEDICAL SPECIALTY
By
MARGARET C. DUERSON
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1986

ACKNOWLEDGEMENTS
My deepest gratitude is extended to each of my
committee members. Special thanks go to my chairperson, Dr.
James W. Hensel, who provided me with unending support and
guidance. His was a warm and motivating force throughout
the project. As a role model he exemplifies the teacher I
hope to emulate as I interact with students and peers.
Each committee member played a unique and indispensable
role in assisting me in this endeavor. I would like to
thank Dr. Parker A. Small for his extensive, constructive
criticism and his encouragement to constantly strive for
excellence. I am grateful to Dr. Robert E. Jester and Dr.
Ronald G. Marks for their valuable comments and suggestions,
especially in the areas of research design and analysis. It
was my good fortune to have the advice and help of Dr.
Forrest W. Parkay.
The Department of Community Health and Family Medicine
made available to me information and resources during the
conduct of this study. For this, I am deeply indebted and
appreciative. I would also like to thank Robert Epting for
his help in executing the data analysis procedures. In
addition, support was provided by Health and Human Services
Grant number 2D15PE-84000-07.

Final
goes to my
whose love
and grow,
possible.
y, and most especially, my sincerest appreciation
husband, Kearney, my two sons, and my parents
and encouragement have enabled me to develop
Kearney's faith in me ultimately made this
ii i

TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS
LIST OF TABLES vi
ABSTRACT viii
CHAPTER
IINTRODUCTION 1
Statement of the Problem 3
Purpose of the Study 4
Hypotheses 5
Background and Justification 7
Delimitations, Limitations, and
Assumptions 11
Definition of Terms 14
Organization of the Study 16
IIREVIEW OF THE LITERATURE 18
Introduction 18
Background Factors 19
Personality and Attitude Factors 31
Factors Related to the Medical Training
System 37
Career Factors 46
Summary of the Review of Literature 49
IIIMETHODOLOGY AND INSTRUMENTATION 53
Setting for the Study 53
Revision and Development of the
Questionnaire 56
Administration of the Questionnaire 62
Residency Match 66
Research Design 67
Data Analysis 70
IVPRESENTATION OF RESULTS 76
Sociodemographic Characteristics of the
Study Group 76
Results of the Data Analysis for Testing
the Hypotheses 82
IV

Page
V SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS.... 116
Summary 116
Conclusions 121
Implications of the Study 127
Recommendations 128
APPENDICES
A MEDICAL STUDENT ENTERING QUESTIONNAIRE 136
B MEDICAL STUDENT FOLLOW-UP QUESTIONNAIRE 140
C LETTER OF EXPLANATION ATTACHED TO ENTERING
QUESTIONNAIRE 143
D LETTER OF EXPLANATION ATTACHED TO FOLLOW-UP
QUESTIONNAIRE 145
E LETTER OF REMINDER SENT WITH SECOND FOLLOW-UP
QUESTIONNAIRE 147
REFERENCES 148
BIOGRAPHICAL SKETCH 154
v

LIST OF TABLES
Table Page
1 Summary of Sociodemographic Characteristics
of Subjects 77
2 Medical Specialty Preferences of Third-Year
Medical Students 80
3 Residency Choice of Fourth-Year Medical
Students 81
4 Association between Third-Year Specialty
Preference and Sociodemographic Character
istics 83
5 Association between Medical Specialty
Preferences and Father's Occupation and
Parents' Annual Income 85
6 Association between Medical Specialty Choice
of Fourth-Year Students and Father's
Occupation 87
7 Association between Fourth-Year Specialty
Choice and Sociodemographic Characteristics... 89
8 Association between Medical Specialty Choice
of Fourth-Year Students and Parents' Annual
Income and College Major 91
9 Sociodemographic Groupings of Medical
Students and Preclerkship Attitudes 94
10 Sociodemographic Groupings of Medical
Students and Postclerkship Attitudes 95
11 Results of Preclerkship Career Goal Rankings
and Three Preference Groups 100
12 Results of Preclerkship and Postclerkship
Career Goal Rankings and Two Specialty
Choice Groups 102
13 Correlation Coefficients for Career Goal
Rankings and Attitude Scores 105
vi

Page
14 Comparison of Specialty Choice Groups Based
on Differences in Preliminary and Follow-up
Career Goal Rankings 110
15 Association between Third-Year Specialty
Preferences and Fourth-Year Specialty
Choices 112
v 11

Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
FACTORS RELATED TO STUDENT CHOICE
OF MEDICAL SPECIALTY
by
Margaret C. Duerson
December 1986
Chairperson: James W. Hensel
Major Department: Educational Leadership
The purpose of this study was to identify factors
related to medical students' third- and fourth-year
specialty choices and to determine the extent to which a
clerkship in family medicine affected attitudes toward
primary care. A multidimensional questionnaire was
administered to third-year medical students immediately
before and after a family medicine clerkship. The
questionnaire addressed sociodemographic characteristics of
the medical students, their career goals, and attitudes
toward primary care. On the initial administration, the
students indicated whether they preferred a primary care
specialty, a subspecialty area, or were undecided on their
future plans for residency training. The results of the
fourth-year residency match served as evidence of the
students' choice.
Responses by 109 students revealed that (a) third-year
specialty preference was significantly associated with
fourth-year specialty choice of residency; (b) socioeconomic
v i i i

measures, parents' annual income greater than $50,000, (c)
students with experience as teachers were distinguished by
more positive attitudes toward primary care; (d) students
supported by parents scored lower on attitudes toward
primary care than those whose major means of support came
from other sources; and (e) the importance students assigned
to 12 career goals did not differentiate between third-year
specialty preference or fourth-year residency choice groups.
Based on their mean attitude scores, students choosing
primary care residencies held more favorable attitudes
toward primary care before and after a clerkship in family
medicine than subspecialty-oriented students but the
subspecialty scores demonstrated greater improvement in
attitudes.
These findings support several educational
interventions. Since the majority of students are capable
of making stable career decisions by the third year,
educational strategies likely to impact on these decisions
should be implemented early in the educational process. The
link between socioeconomic status and specialty choice
suggests that as tuition costs increase, adequate financial
assistance becomes critical for the less affluent students
who are most likely to fill the ranks of primary care
specialties. Early training activities in ambulatory care
settings should be fostered to maintain the interest of
students preferring primary care and to improve attitudes
and attract other students.
IX

CHAPTER I
INTRODUCTION
An important goal of this nation is to provide an
effective, efficient health care system for its citizens.
To accomplish and maintain such a system requires that
shortages and surpluses of health care providers
particularly physiciansbe avoided. When considering
supply and requirements for physician manpower needs, the
focus must not be limited solely to aggregate numbers but
must take into consideration the distribution among
specialties of physicians, also.
The Summary Report of the Graduate Medical Education
National Advisory Committee (GMENAC) (U.S. Department of
Health and Human Services, 1980) cited imbalances in
physician specialty areas as a major health care problem
facing the nation. There is general consensus that there
are too many physicians in some specialty areas and not
enough in others. Using projected manpower needs for 1990,
surpluses are expected to occur in nonprimary care areas
such as surgery, while many of the primary care specialties
will fall short of the projected requirements (Jacoby,
1981) .
Neither is the maldistribution across specialties
likely to abate in the near future unless some action is
-1-

-2-
taken to correct the situation. Using the supply
assumptions of the GMENAC Report, Kindig and Dunham (1985)
projected physician specialist growth into the 21st century.
By the year 2020, the absolute number of physicians in
primary care will increase by 55% but the absolute numbers
in nonprimary care will increase by 111%. As these authors
observed, it is conceivable that with the aging of the
population and the competitive approaches to cost
containment the additional primary care physicians will be
needed but it is difficult to assume a need for an
additional 260,000 nonprimary care physicians.
Intervening in the situation while the numbers of
physicians yet to be trained is amenable to change is
essential. Two approaches toward correcting the problem
include (a) stringent government regulations to limit the
types and numbers of residencies which train specialists, to
bring reimbursement of specialists in line with primary care
physicians, and to restrict the number of graduates of
foreign medical schools and foreign physicians coming into
the country; and (b) working within the educational
environment to influence specialty choice decisions. Of the
two approaches, the latter seems the most viable solution in
a democratic society to meet health care needs and at the
same time allow individuals freedom in the specialty
decision process.
Until recently there has been little effort directed
toward assuring that the medical education system produced

-3-
the proper mix of primary care and nonprimary care
physicians for the needs of society. When the medical
education system has been targeted, modification has
occurred at the residency level rather than undergraduate
medical education. Establishing policies and finding
solutions has been hampered because factors influencing
specialty choices are complex and not well understood.
Although a large body of research has been generated
the specialty choice research literature
suffers from major inadequacies in providing
guidance for policy-makers and only rarely can
provide unequivocal evidence regarding various
influences on physician specialty choice. (U.S.
DHHS, 1980, Vol. V, p. 2)
Statement of the Problem
The problem addressed in this study is the need for
research-substantiated information on the factors related to
student choice of medical specialties as stated broadly in
the following questions:
Are sociodemographic characteristics of medical
students associated with third-year specialty preferences or
choice of specialty as evidenced by selection of a primary
care or nonprimary care residency in the fourth year of
medical school?
Is there a difference between medical students grouped
by sociodemographic characteristics with regard to their
attitudes toward primary care and family medicine?

-4-
Is there a difference in career goals between groups
of medical students identified by third-year specialty
preferences or fourth-year specialty choice?
What is the relationship between the importance medical
students attach to career goals and their attitudes toward
primary care?
Is there reason to believe that career goal rankings
before and after a family medicine clerkship are different
for medical students choosing primary care from those
choosing subspecialties in the fourth year of medical
school?
Is specialty preference expressed in the third year of
medical school associated with medical students' specialty
choice decisions in the fourth year as evidenced by
residency selection?
Y Are medical students' attitudes toward primary care
changed after a clerkship in family medicine?
Is there a difference in attitudes toward primary care
between groups of medical students identified by third-year
specialty preferences of fourth-year specialty choice as
evidenced by residency selection?
Purpose of the Study
The primary purpose of the study was to identify
factors related to medical students' third- and fourth-year
specialty choices and to determine the extent to which a
clerkship in family medicine affected the students'

-5-
attitudes toward primary care. Observations obtained from
administration of a multidimensional questionnaire were used
to explore (a) the association between sociodemographic
characteristics of medical students, their career goal
rankings, and their attitudes toward primary care and the
students' specialty preferences or specialty choices; (b)
whether significant differences in sociodemographic
characteristics existed between medical students with
favorable and unfavorable attitudes toward primary care; (c)
the relationship between the relative importance students
assigned to career goals and their favorable or unfavorable
attitudes toward primary care; (d) the extent to which
attitudes toward primary care changed after a clerkship in
family medicine; and (e) the association between medical
students' career preferences in the third year and their
actual career choice in the fourth year as measured by
residency specialty.
Hypotheses
This study involved the testing of the following
hypotheses:
1. There is no association between sociodemographic
characteristics of medical students and their medical
specialty preferences expressed in the third year of medical
school.
2. There is no association between sociodemographic
characteristics of medical students and their medical

-o -
specialty choices as evidenced by fourth-year residency
selection.
3. There is no difference between medical students
grouped by sociodemographic characteristics according to
their attitudes toward primary care measured before or after
a third-year family medicine clerkship.
4. There is no difference between groups of medical
students identified by medical specialty preference in the
third year or medical specialty choice in the fourth year
with respect to their rankings of career goals.
5. There is no relationship between medical students'
rankings of career goals and their attitudes toward
primary care before or after a third-year family medicine
clerkship.
6. There is no difference between the groups of
medical students choosing primary care and those choosing
subspecialties in the fourth year with respect to their
rankings of career goals before and after a family medicine
clerkship.
7. There is no association between medical students'
specialty preferences expressed in the third year and their
career specialty choices as evidenced by fourth-year
residency selection.
8. There is no difference in medical students'
attitudes toward primary care before and after a family
medicine clerkship.

-7-
9. There is no difference between groups of medical
students identified by specialty preferences with regard to
their attitudes toward primary care.
10. There is no difference between groups of medical
students identified by their fourth year specialty choice
with respect to their attitudes toward primary care before
and after a family medicine clerkship.
The .01 level of significance was used for rejection of
the null hypotheses.
Background and Justification
Achieving a better balance between physicians in
primary care specialties and those in other specialties has
become one of the major concerns of medical educators,
health care policy makers, and others vitally interested in
the proper functioning of the health care industry.
Approximately 80% of the physicians in this country are
specialists. When the ratio of generalist to specialist
(primary care to nonprimary care) is imbalanced several
consequences are likely to result: (a) cost of medical care
increases due to the high dependence on specialists for
primary care, (b) quality of health care may actually be
compromised from iatrogenic complications resulting from
unnecessary procedures and physicians practicing outside of
their areas of expertise in order to maintain their income,
and (c) doctors' job satisfaction could suffer as a result

-8-
of improper use or under utilization of their skills and
knowledge (Schroeder, 1984 ).
From the post-World War II period through the 1960s and
1970s, physician manpower policies were primarily concerned
with the overall number of physicians and their geographic
distribution. In the 1940s and 1950s task forces on health
manpower needs predicted a shortage of 50,000 doctors by
1980 (Petersdorf, 1975). Health manpower legislation of the
1960s was directed toward remedying this perceived shortage
by increasing medical school enrollments, recruiting
faculty, and expanding the existing facilities. The
preoccupation with increasing the aggregate number of
physicians was so intense that the trend toward increased
specialization went largely unnoticed. From 1931 to 1974,
physicians identifying themselves as general practitioners
decreased from 83% to 18% (Harris et al., 1982).
The health manpower legislation of the 1960s was so
successful, it is estimated that by the year 2000 there will
be 145,000 more physicians than needed to provide physician
services (U.S. DHHS, 1980, Vol. I). The problem of
oversupply of physicians is not limited to the United
States, however. Virtually every Western country is facing
a current or projected excess of physicians. In European
countries the surpluses have produced varying results from a
dramatic decrease in physician income to actual unemployment
for a sizeable number of medical doctors. While the United
States shares the problem of surpluses with other countries,
compared to its Western allies, this country stands alone in

-9-
the higher ratio of specialists to generalists (Schroeder,
1984) .
Factors responsible for the trend toward increasing
specialization are difficult to determine. Rapidly
expanding technology and scientific knowledge may cause some
to conclude that the only way to achieve depth of knowledge
and skill in an area or to remain current in an area of
medicine is to narrow the field of interest. Petersdorf
(1975) commented on the possible causes:
Simplistical ly speaking, every man has an innate
desire for self-advancement, intellectually and
financially. In addition, subspecialists seem to
acquire prestige, particularly in medical schools.
More than anything else, however, the structure of
training programs in medical schools has led to
specialization. (p. 697)
To fill the void left with the demi se of the old
generalists' role, a new specialty, family practice,
developed specifically designed to deliver primary care.
The final approval of family practice as the twentieth
primary specialty was granted by the American Medical
Association in 1969 (Rakel & Pisacano, 1984). In an effort
to respond to the need for more primary care physicians and
aided by funding provided by the Comprehensive Health
Manpower Training Act of 1971, medical schools initiated
residency training programs in this new specialty. In a
further attempt to influence specialty distribution,
legislation was passed in 1976 which "set minimum
requirements for percentages of first year residency
positions in the primary care specialties" for medical

-10-
schools to meet In order to continue to receive capitation
grant funds (U.S. DHHS, 1980, Vol. V, p. 1).
All of these early efforts toward correcting the
shortage of primary care physicians were directed at the
residency training period. Almost no attention was given to
understanding and influencing the career preferences and
choices made earlier in the medical education process.
Until recent years, research on career decisions at the
undergraduate medical education level had been principally
conducted by sociologists and psychologists with academic
interest in furthering the knowledge base in their own
discipline rather than an interest in obtaining information
for program evaluation and policy formulation. Influencing
career choice at the undergraduate medical education level
was sanctioned by Califano, Secretary of Health, Education,
and Welfare, in an address to the Annual Meeting of the
Association of American Medical Colleges in 1978:
I urge medical educators to exert leadership in
dealing with the dangerous and wasteful decline in
primary care physicians by encouraging course
offerings or other innovations that increase the
exposure of students to primary care settings and
heighten the appeal of primary care. (Califano,
1979, p. 22)
Medical schools accepted the challenge and expanded the
curricula to include programs in family medicine at the
undergraduate level of medical education as well as the
residency level.
Since the medical education environment is recognized
as one of the factors contributing to the specialty

-11-
imbalances, it is appropriate to study the relationship
between the undergraduate educational process and specialty
and subspecialty choices. A knowledge of the variables
within the educational environment which influence and
predict specialty and subspecialty career choices would
provide guidance for medical educators, policy makers, and
others vitally interested in correcting the present medical
manpower imbalances and planning for future needs.
Delimitations, Limitations, and Assumptions
Delimitations of This Study
This research study was delimited to
1. Medical students who were in their third year of
training at one state-supported medical school in the
southeast in the years 1983-84 and who participated in the
residency match in 1984-85.
2. Factors affecting medical specialty choice at one
university medical school as represented by the group of
students sampled.
Limitations of This Study
There were several limitations that must be recognized
in conducting this study:
1. The results of this research project were
characteristic of this study only. Because of the unique
characteristics of samples and the variations inherent in

-12-
different settings, generalizations beyond the sample
studied in the setting of interest are to be avoided. The
generalizability of this study was further limited since
subjects were not randomly selected but came from an intact
class. The students were randomly selected by a lottery
system into the six rotation groups, however.
2. A limitation of the study was a lack of control of
intervening variables. Specific events other than those
being studied may have occurred between the first and second
measurements to produce a change in the students' attitudes
toward primary care and family medicine.
3. In test-retest situations, exposure to the first
test may bias or sensitize the subject to the second test.
Since the questionnaire used to measure entering attitudes
was the same as the questionnaire used after the family
medicine clerkship, this source of error could have biased
the results of this study.
4. The design of the research instrument provided
limitations. The questionnaire items were forced choice,
closed-end, and very short answer format. The Attitudes
about Primary Care and Family Medicine section, Part F, used
a Likert scale with the following choices: strongly agree,
tend to agree, not sure, tend to disagree, and strongly
disagree. Variable interpretations of this and other
sections of the questionnaire could potentially distort the
study's results.

-13-
5. The follow-up questionnaire was mailed to medical
students at the completion of the family medicine clerkship.
Mailed surveys often result in low return rates and even
lower percent usable data as some items are not completed.
6. Working in the real world of educational research
has the advantages of practical applicability and the
disadvantages of having to deal with many variables
simultaneously. If there is to be any application of the
method to this or other sites, the natural setting has
obvious advantages.
Assumptions of This Study
The following assumptions are implicit in the
limitations and delimitations of the study:
1. Participants responded voluntarily to questionnaire
iterns.
2. Responses were made in a nonthreatening environment
so they were truthful and those intended by participants.
3. Administration and collection of the questionnaire
did not influence the responses.
4. The participants were representative of previous
medical students at the same level of training and those
that follow at the University of Florida College of
Medicine.
5. Experiences while on the family practice rotation
were similar for all students.

-14-
6. Items on the questionnaire resembling attitude
statements concerning primary care and family medicine
were representative of generalizations of the domain of
interest.
7. Attitudes and beliefs are not open to inspection
but can be inferred from the answers to statements
pertaining to behaviors and positions toward certain
actions.
Definition of Terms
The terms used in this research study are defined as
follows:
A primary care physician is one who (a) is the
physician of first contact for the patient; (b) makes the
initial assessment and attempts to solve as many of the
patient's problems as possible; (c) coordinates the
remainder of the health care team, including ancillary
health personnel as well as consultants, that are necessary
in dealing with the patient's problems; (d) provides
continued contact with the patient and often his or her
family; and (e) assumes continued responsibility for his or
her care (Petersdorf, 1975). The primary care physician
administers a highly personalized type of care which
coordinates all of the health care needs of the individual
in sickness and in health. Few specialties consider this
comprehensive and continuing care to be their responsibility

-15-
and within their range of competence (Rakel & Pisacano,
1984 ) .
Clerkships are student-selected elective and required
clinical experiences occurring during the last two years of
medical school. Working as a member of the health care team
in the actual care of patients provides the student with the
opportunity to apply and practice newly acquired knowledge
and skills.
Preceptor refers to a licensed, practicing physician
who teaches medical students. This physician may be in
private practice or in an academic setting.
The terms occupation preference, choice, and attainment
are frequently used interchangeably. Vroom's distinction
between occupation choice, preference, and attainment as
explicated by Matteson and Smith (1977) are
Occupation preference is that occupation which at a
given time an individual would most like or prefer to enter.
Occupation choice is that occupation which a person
chooses to enter and then engages in behaviors to implement
that choice.
Occupation attainment is the occupation of which the
person is a member.
This study concerns itself with two points of
occupational decision, preference and choice.
Preference of specialty is a positive written or verbal
expression of a desire to pursue the particular area of
medical practice. The participants' preferences for a

-16-
primary care or nonprimary care specialty was solicited in
the initial administration of the questionnaire in the third
year of medical school prior to the family medicine
clerkship.
Choice of medical specialty requires that some action
be taken in the direction of or participation in a
particular area of specialty practice, for example, taking
steps to secure a residency in a particular medical
specialty area. For purposes of this study, specialty
choice was considered to be evidenced by the selection of
residency in a given specialty area in the fourth year of
medical school.
Organization of the Study
Chapter I contained an introduction, a statement of the
problem, and the purpose of the study. The research
hypotheses underlying the study were stated. The
limitations and delimitations of the project were explained
along with the definitions of words and phrases specific to
this study.
Chapter II is a review of the literature related to
medical specialty choice. Chapter III addresses the
research methodology and instrumentation. The data
collection procedures, the setting for the study, research
design, and data analysis procedures are explained.
Statistical results are presented in Chatper IV.

-17-
The final chapter, Chapter V, summarizes the findings
of the study, the conclusions, and the implications.
Recommendations for future research in the area of medical
specialty career choice are suggested.

CHAPTER II
REVIEW OF THE LITERATURE
Introduction
During the last 10 years researchers have examined a
wide range of factors thought to influence medical specialty
choice. Because of the importance of the problem and the
consequences of these intraprofessional career decisions,
medical educators, sociologists, psychologists, health
policy-makers, and others concerned with medical care in
general have attempted to understand why and how choice of
medical specialization occurs. The Graduate Medical
Education National Advisory Committee Report (U.S. DHHS,
1980), confirming the suspicion that the nation faced an
oversupply in some specialty areas and actual shortages in
other specialties, gave further impetus to the study of
variables associated with specialty choice.
This great interest in the career choice process of
young doctors has resulted in an extensive body of
literature. Literature reviews on career choice within
medicine were published in the mid-1970s. Due to the
comprehensive nature of these publications, this review will
examine the literature published since 1975.
Approaches to the study of the problem have been so
diverse, this review will consider related research together
-18-

-19-
which shares a common approach or similar variable, rather
than summarizing the studies chronologically. The factors
will be categorized into four groups: (a) background
factors, (b) variables related to personality and attitudes,
(c) factors pertaining to the medical environment or
training system, and (d) career factors.
Background Factors
A large part of the literature on specialty choice
focuses on students' background characteristics including
age, sex, marital status, hometown, socioeconomic status of
family, and educational history. Two reasons for
researchers' seeming preoccupation with these variables is
assumed to be their accessibility from students' records and
their potential use in the selection process for admission
to medical school. What must not be forgotten, however, is
that medical students are a highly selected, homogeneous
group so that differences in these background
characteristics are small while specialties and
subspecialties are quite diverse making the background
variables poor predictors of specialty choice.
Both Anderson (1975) and Zuckerman (1977) in extensive
reviews of the literature reported evidence that the larger
the community of origin the greater the likelihood of the
student choosing specialty practice over general practice.
Hutt (1976) reported similar findings in Britain where those

-20-
from rural areas tended more often to prefer general
practice.
Closely related to choice of specialty is choice of
practice location. The relationship between practice
location and specialization becomes more apparent when one
considers that highly trained subspecialists would not
likely find sufficient patients or the sophisticated
equipment necessary for practice of their specialty in small
towns. In a study of students entering the University of
Washington School of Medicine in 1975, 1976, and 1977,
Carline and associates (1980) found that contrary to other
studies there were few differences in preferences for
specialties based on size of hometown, while students'
attitudes toward location of practice indicated a preference
for practice in communities similar in size to those in
which they were raised. Support for this finding was also
reported by Hadac (1984). In a review of the literature
associated with specialty and location choices, he found
that a preference for rural or small town practice was
associated with high school attendance in a town of similar
size.
A second variable of interest is socioeconomic status
or social class which has been commonly measured in terms of
the father's occupation and level of education. Studies of
the relationship between social class and choice of
specialty are mixed. Results of some studies demonstrated
an absence of a relationship or ambiguous results, while

-21-
other researchers found a positive relationship (Anderson,
1975; Zuckerman, 1977). Where positive associations between
social class and specialty choice were found, the higher the
socioeconomic status of the family, the more likely the
student was to choose specialization over general practice.
As Zuckerman (1977) pointed out "the economic security and
availability of financial assistance associated with higher
social class may encourage pursuit of the additional
training required for specialty practice" (p. 1082) .
Generally, the primary care specialties require fewer years
of residency training than the subspecialties and thus
decreased cost of education and more rapid entry into wage
earning status.
Medical students whose fathers and mothers are
physicians constitute an interesting group with regard to
social status, especially since they seem to be over
represented in medical schools based on the population
census. Gough and Hall (1977) observed that approximately
16% of students in American medical schools come from
medical families compared to a census estimate of 0.4%
physicians among employed adult males. In their study of
1195 students at the University of California, San
Francisco, School of Medicine, there were 162 from medical
families. These investigators noted that there was a
statistically insignificant shift away from family medicine
and general practice by the subgroup of physicians'

-22-
children. However, there was a pattern toward over
representation in some specialty areas for the physicians'
children, namely ophthalmology, otolaryngology, dermatology,
and surgery.
Mawardi (1979), in a longitudinal study of Case
Western Reserve University graduates, found that 39 of
the 135 physicians studied gave credit for their interest
in medicine to family members in the profession. The
most frequently mentioned influential relative was a
father. These 39 subjects cited 59 relatives who were
physicians.
Given the rising cost of medical education and the
concommitent decrease in financial aid available to
students, it could be hypothesized that students from
families with few financial resources will not apply to
medical schools. This could lead to a profession of the
socially elite. Based on the preliminary studies of
socioeconomic status and specialty choice, a preponderance
of students from affluent backgrounds will effect the mix of
specialist toward nonprimary care specialties.
Factors such as age and marital status have provided
some consistent clues to specialty choice decision. Using
the data from a longitudinal study of Jefferson Medical
College students from 1971 through 1975, Herman and Veloski
(1977) demonstrated not only a rising trend in the
proportion of the Jefferson students interested in family
medicine but those senior students expressing an interest in

-23-
family medicine tended to be slightly older than those
choosing other specialties.
Age and marital status may be interrelated in,that in
general, married students tend to be older. Assuming that
students who choose family medicine share common social
characteristics, Cole, Fox, and Lieberman (1983) examined
questionnaire data obtained from 429 students at a public,
northeastern medical school. The profile of the subgroup of
students who chose family medicine was disproportionately
older, married, and more often Protestant. It may be that
as with socioeconomic status, it is not the age or marital
status per se that effects specialty choice for the older,
married student but rather the economic restraints that make
long residency training difficult.
Hutt (1976) found similar outcomes in Britain.
Marriage had a precipitating influence on specialty choice.
Twice as many married men opted for careers in general
practice. Married women were more apt to take up careers
outside of hospitals or, if inside, to take those not
involving clinical duties.
In a unique approach to gain a better understanding of
physicians' career choices, Skipper and Gliebe (1977)
surveyed not only senior medical students on their plans for
a medical specialty but also asked the students their
perceptions of the influence of their spouses on those
plans. Going one step further, the researchers collected
data from 17 of the 21 wives of the married students. Due

-24-
to their own career goals, the women preferred that their
husbands settle in a medium-sized city or suburban fringe
area. They also preferred group practice to solo practice
so that the husbands would have more time for family
activities. While the specialty choices expressed by the
wives was varied, some of the wives viewed family medicine
as less desirable because it would provide less free time
than other specialties. Researchers conclude that if
primary care specialties such as family medicine are to
appear attractive to wives of medical students as well as
the students themselves, there will need to be some attempt
to also convince the spouse of the good points of the
specialty.
The small numbers of subjects involved in the Skipper
and Gliebe (1977) study preclude definitive statements. It
does, however, suggest that as more marriages become two
career families, there will be pressures on career choices
to accommodate both husbands and wives.
Medical schools have the mixed blessing of having a
large applicant pool of highly qualified candidates from
which to choose. While medical schools do not publish the
admission criteria, it is generally known that students must
have demonstrated academic success by virtue of their
undergraduate grades and their scores on the Medical College
Admission Test (MCAT). Strong interest and a proven
aptitude for science are given more weight than other
factors in the selection process. This emphasis on

-25-
undergraduate grades in science and the MCAT scores as
criteria for admission is questionable. The Anderson (1975)
review of the literature reported mixed results in studies
that used MCAT scores to differentiate specialty choice.
Subsequent studies have not provided more consistent data.
Gough (1978) used four measures, the MCAT science
subtest scores, premed grade point average (GPA), a
preference for scientific subjects, and a composite of the
three, in this study of 1,135 graduates of the University of
California, San Francisco, Medical School. When these
factors were used to predict performance in medical school,
all four factors were significantly correlated the first
year of medical school but the following years became less
and less predictive of performance until by the fourth year
they were completely unrelated. With regard to specialty,
the scientifically oriented students entered specialties
such as surgery, anesthesiology, and pathology; those
ranking lower went into internal medicine, pediatrics, and
psychiatry.
In his review of the literature to 1977, Zuckerman
found that students planning to enter general practice
scored low on MCATs and had low academic standing in medical
school. Conversely, those who planned academic or research
careers had high MCAT scores and high academic standing.
Those planning to specialize and go into private practice
scored in the mid range on MCATs. Other studies published
the same year as the Zuckerman review reported (Association

-26-
of American Medical Colleges, 1977; Hadley, 1977) that
students with high undergraduate grade point averages and
high MCAT scores appear to be predisposed toward
subspecialty areas rather than primary care.
More recent studies of academic variables indicate that
unlike the students in years past who went into general
practice, students choosing primary care or family medicine
perform at least as well if not better than those tending
toward subspecialty areas. In a study of trends in senior
students' interest in primary care specialties at Jefferson
Medical College, Herman and Velosk (1977) found that
students interested in family medicine performed as well or
better on measures of basic science knowledge and subjects
emphasizing clinical information as those students
interested in all other specialties except internal
medicine. Similarly, Collins and Roessler (1975) found no
significant differences in intellectual characteristics as
judged by MCAT scores and undergraduate GPA between family
medicine oriented students and any other group in their
study of Baylor Medical College students. Results of a more
recent study of a large national sample of medical students
(Burkett & Gelula, 1982) added additional weight to the
previously cited findings. In a comparison made between
primary care versus nonprimary care fields, students in the
two groups were similar on the basis of measures of
premedical school academic performance and MCAT scores.

-27-
Academic measures made later in the medical school
process have not proved much better in producing consistent
results. Two studies conducted in the same state but at two
separate medical schools used National Board of Medical
Examiners (N3ME) examination scores to study the
relationship to career choice. The NBME exam results are
available late in the third year or early fourth year of
medical school. One group of investigators who examined the
effects of NBME scores for 628 students at the University of
Medicine and Dentistry of New Jersey-New Jersey Medical
School, found that high scores on the NBME Part II subtests
in psychiatry, medicine, and surgery were associated with
selection of these specialties (Fadem et al., 1984).
Contrary to these findings, Rosevear, Tickman, and Gary
(1985) did not find a positive relationship between medical
students' career choice and their performance on the NBME
Part II subtest scores for 347 graduates (1979-1983) of the
University of Medicine and Dentistry of New Jersey-Rutgers
Medical School.
Factors predictive of specialty choice which occur
early in the educational process have the advantage of
providing more opportunity for impact and intervention.
Unfortunately, the early parameters such as undergraduate
grade point average, MCAT scores, and other cognitive
measures do not consistently discriminate between those
choosing primary care specialties and those choosing
nonprimary care specialties. It may be that because all

-28-
medical practice requires a high intellectual functioning,
there are no discernible differences based on scholastic
factors.
With the growing numbers of women entering medical
school, research on the effects of gender composition on
specialty choice has been heightened. Historically, women
have been underrepresented in medicine. Women constituted
r
only 7.7% of medical students in 1964-1965 (Dube, 1973).
Concurrent with the women's movement of the late 1960s and
the efforts to include women in occupations where they had
been formerly excluded, their representation in medicine
increased. Presently females constitute over 30% of the
medical student population. The percentage of women in the
entering class of 1991 is expected to be close to 45.3%
(Lanska et al., 1984). This change in gender distribution
has possible implications for health care delivery in
general and specialty choice in particular.
Several studies of sex differences related to specialty
choice found significant differences between male and female
medical students' propensity toward specialties.
Male/female differences in specialty choice, beliefs about
specialties, and personality characteristics were explored
by McGrath and Zimet (1977). Students at two state
university medical schools rank ordered their preference for
five major medical specialties. Both men and women listed
family medicine as their first choice but significant
differences occurred at the levels of second and third

-29-
choices as women ranked pediatrics second and men chose
internal medicine. Surgery ranked last for women and third
for men.
Similarly, Cuca (1979), based on plans for board-
certification of 1978 U.S. medical school graduates, found
that over half of both males and females preferred primary
care while twice as many women preferred pediatrics as did
the men. The category of surgery again evidenced dissimilar
trends for women and men.
The work of Zimny and Shelton (1982) supported these
earlier studies. Responses of 380 third year medical
students on the Medical Specialty Preference Inventory
showed no significant differences between male and female
students' preference for obstetrics, pediatrics, and
psychiatry. Men, on the other hand, had a slight preference
for internal medicine and surgery.
Bergquist and his colleagues (1985) asked the 1983
entering class of one medical school to select their
intended specialty choice from a list of 50 medical fields.
When the specialties were categorized into primary care,
surgery, medicine, and other, the results showed that women
selected specialties in primary care more often than men.
Men selected surgical categories significantly more
frequently than women.
On the theory that as the proportion of women in
organizations reach parity with men, women will be less
likely to be entrapped in limited or stereotypically female

-30-
roles, Weisman (1984) found that, indeed, in the graduates
of 1970 through 1976 higher proportions of women in medical
schools were associated with smaller proportions of women
entering traditionally female specialties. There was only a
weak effect of gender on first-year residency choice.
There was disagreement in the studies as to whether
women were becoming less traditional in their specialty
choice. Both males and females expressed a strong interest
in the primary care specialties indicating some convergence.
However, there are also major differences in the factors
perceived to be important in choosing a specialty. Women
expressed a preference for patient contact specialties where
patients are involved in solving their own health problems
and were willing to forego career time for more family life
(Bergquist et al., 1985; Zimny & Shelton, 1982). Males
showed a high preference for complex medical problems and
technical procedures which would necessitate functioning in
highly specialized areas (Zimny & Shelton, 1982). Males
also expected to have higher incomes than women expected but
it is not clear whether this expectation dictated the career
choice or was a consequence of it (Bergquist et al., 1985).
The eventual impact on medicine of greater
participation by women is unclear. The fact that primary
care specialties are popular with women may help to correct
the problem of overspecialization but it is not the entire
solution to the national health problems. This solution
oversimplification of the problem and women's
would be an

-31-
unique contribution to health care. As the number of women
physicians increases, medicine may take on some of the
characteristics common to female dominated occupations. The
changing gender mix of medical school classes merits
continued study.
Personality and Attitude Factors
It is not surprising given the homogeneity of medical
students' intellectual ability and some background
characteristics that the ability to associate specialty
choice with these variables has been inconsistent. Because
the sociodemographic data have been so capricious, some
investigators have sought to determine whether measureable
differences in personality variables are predictive of the
type of specialty chosen.
Zuckerman (1977) in a comprehensive review of the
literature found that psychological characteristics
distinguished psychiatrists and surgeons from other
specialty types but these factors have been less fruitful in
distinguishing other types of specialties. He concluded,
"The utility of personality characteristics to predict
variation in career choice appears to be limited" (p. 1083).
Hutt's (1976) review of previous research supports the
consistent findings of personality differences between
psychiatrists and surgeons. She observed that "there is
evidence that these two choices are made early and are more
consistently adhered to than others" (p. 466). It may be

-32-
that those specialties where the choice is made later, based
on experience in the fields, are much less successfully
identified from personality characteristics. Hutt
commented, "The field of personality differences in relation
to specialty, important as they are, is not a promising one
for policy-makers who wish to put right particular
imbalances in the supply of doctors" (p. 466).
Though there have been problems in predicting specialty
choice from comparison of personality variables, it is
important to review the findings. For example, Collins and
Roessler (1975) compared four different career groups,
internal medicine, surgery, obstetrics-gynecology, and
pediatrics, with family medicine residents. One hundred
eighteen third-and fourth-year students at Baylor College of
Medicine participated in a battery of personality, vocation,
and attitude tests to investigate the relationship between
the test data and subsequent specialty choice. The Baylor
students who chose family medicine residencies were in most
instances significantly different from the other four
groups. They scored higher on the need for affiliation,
were less aggressive and less materialistic. The family
medicine students' scores were far below the population
average of 50 on materialism on the Birkman Vocational
Interest and Attitude Survey.
Plovnick (1980) also observed differences in attitudes
and values between students oriented toward primary care and
those choosing nonprimary care specialties. The small

-33-
sample size of this study precluded meaningful statistical
analysis but several important trends were identified using
three factors, Orientation to Patient Care, Orientation to
Work Conditions, and Orientation to the Profession.
Students choosing primary care careers scored higher on the
factor "Orientation to Patient Care," indicating more
concern for people and less orientation towards the
profession. More unsettling was his finding that over the
four years of medical school, primary care respondents
exhibited a shift in their attitudes away from concern for
patient care toward a somewhat greater self-concern.
Plovnick explained this trend as the general socialization
influence of medical school which might be changed by a
further differentiation of programs designed for primary and
tertiary care providers.
A widely used instrument for investigating personality
differences and career choices is the Myers-Briggs Type
Indicator (MBTI). Based on the personality theory of Jung,
it compares four categories on a continuum of opposites:
extroversion-introversion, thinking-feeling, sensing-
intuition, and judgment-perception. These type differences
are related to differences in preferences for activities,
interests, personal outlooks, and choices of occupation. In
the early 1950s, Myers tested over 5000 medical students.
Those still in practice were followed up in 1973 (McCaulley,
1981) and confirmed that their specialty choices were
similar to those chosen at graduation. Harris, Kelly, and

-34-
Coleman (1984) demonstrated the stability of the MBTI from
entrance to graduation for the classes of 1980 and 1981 at
the University of Utah School of Medicine. In a discussion
of implications relevant to the findings of MBTI studies,
Hadac (1983/1984) observed that sensing types tend to be
denied admission to medical schools because they generally
score lower on entrance exams. When they are admitted, they
prefer primary care specialties, thus admission policies may
be eliminating those students likely to correct the
specialty maldistribution. Hadac cautioned that
psychological type alone cannot predict specialty choice for
all students. He stated, "Also, all personality types are
found in all specialties and there is not at present, and
may never be, strong evidence that any type should be
excluded from any specialty" (p. 46).
When considering the literature on personality and
attitudes, it was considered worthwhile to address
interests, values, and early orientations to medicine. The
criterion group in the Leserman (1978) study were first year
medical students at three North Carolina medical schools.
Data from this group of students
suggest that incoming medical students are
concerned with helping people but not necessarily
through political means, committed to some
geographic and specialty areas of patient need,
choosing medicine for reasons other than economic
rewards but not opposed to physicians' large
income and status and somewhat unaware of
discrimination toward women physicians and
patients. (p. 330)

-35-
Seventy-three percent of this group expressed interest in
primary care and 54% were interested in rural health. This
represents a higher percent than might be expected compared
to national statistics. Steinwachs and his colleagues
(1982) in their analysis of recent trends in graduate
medical education reported that approximately 55% entering
residency in 1980 chose primary care.
In a national survey of third-year students, Burkett
and Gelula (1982) asked students to assign relative values
to four motivating factors in their decision to enter
medical school. Those students indicating a preference for
primary care attributed greater importance to the desire to
help people than to the other three choicesdesire for
financial reward, desire for social status, and desire to
apply scientific ability. They also demonstrated a tendency
to consider the sociopsychological context of patient
problems and felt a need for change to improve health care.
According to Burkett and Gelula (1982), "the overall pattern
appears to confirm the notion of primary care as a 'person-
centered' health care field which attracts students who have
a more 'holistic' orientation" (p. 512).
Robbins et al. (1983) studied gender differences in
interest and motivation for a career in medicine. Not
surprising, both men and women scored highest on the medical
science category of the Strong-Campbel 1 Interest Inventory.
Both sexes were similar on a projective technique where
90% were judged to fear success. On the attitudinal

-36-
questionnaire, males and females ranked the same reasons as
important for going to medical school although they ranked
them in different order. Males' order of choices were
interest in science, helping people, and having a career.
Women ranked helping people as first in importance, followed
by desire for a career, and interest in science. All
respondents listed a happy family life and job satisfaction
as important life goals. Men tended to consider a high
income as more important than the females. Although
attitudes toward chores were egalitarian, responses
indicated that in reality women performed most routine
household activities. The researchers pointed out that
women's value considerations may affect initial career
choices as to training programs and work opportunities.
Assessment of personality characteristics has been done
rather unsystematically using a variety of standardized and
nonstandardized instruments administered at varying times
throughout medical school. The characteristics observed
have reflected the special interests of the investigators
and have often measured vague concepts such as idealism and
cynicism. There has been no attempt to replicate previous
research in most cases.
A further difficulty with systematizing the assessment
of personality factors is that psychological tests can be
falsified if the respondent is motivated to do so. The
person can answer the items in the way that is believed to
be preferred by the investigator or is perceived to be

-37-
socially correct-. Most students consider "helping people"
as the appropriate answer rather than "economic gain" when
asked the reason to pursue a career in medicine.
Disregarding the difficulties inherent in personality
testing, there is a major problem inherent in manipulating
specialty distribution based on personality characteristics,
namely, the legality of and societal distaste for doing so
(U.S. DHHS, 1980). The subjective interpretation of the
results of personality tests will certainly not go
uncontested by those denied admission were these tests given
specific weighting in the admission process. Society would
surely reject the routine use of personality tests as the
basis of admission due to the perceived subjectivity,
unreliability, and possibility of social regimentation
implied by such action. The information gained from
personality inventories and questionnaires is most
appropriately reserved for counseling students.
Factors Related to the Medical Training System
As Zuckerman (1977) pointed out, sociodemographic and
personality characteristics present at admission become less
useful over time in predicting specialty choice, while those
factors associated with the institutional environment take
on more importance. Students are presumed to come into the
medical training setting with certain motives, values, and
knowledge. These factors are subject to change or
reinforcement by the pressures of agents or conditions

-38-
present in the environment. There is a strong feeling that
these structural influences and the socialization process of
the medical education system are key determinants of career
choice. Based on this premise a number of studies have
focused on the institutional environment and manipulation of
variables therein.
Early reports concerning the development of medical
students were generally qualitative but, nonetheless, paved
the way for later quantitative studies on institutional
influences. Becker et al. (1961), in a hallmark study of
the socialization process, addressed the observed increase
in cynicism of medical students during their training. He
argued that students enter with idealistic concerns about
the sick and serving mankind. From his observations,
students did not lose this idealism for broader social
issues. What some interpreted as cynicism was really the
response to pressures of day-to-day details with long hours
and intense studying which caused a narrowing of concerns
just to get through the rigors of medical school. This
short term effect disappeared but the long term
institutional effects did not. Becker thought there was
enough congruence between students chosen for admission
and the institutional values to effect long range
perspectives.
In a review of the literature published almost 20
years after the Becker book, cynicism continued to be
reported as a by-product of the medical school environment.

-39-
Rezler (1974) concluded from his review that attitude
changes resulting from a special program will fade unless
they are reinforced by the total environment.
One suggested strategy for examining the influences of
the medical education environment on specialty choice is
based on the immediacy of their impact on the students (U.S.
DHHS, 1980). First-order effects have a direct influence on
the knowledge, skills, values, attitudes, and interests of
the students. The primary example of first-order effects
are the role models, teachers, and attending physicians.
Also included in this category are the opportunities for
role playing and practicing the acquired knowledge, skills,
values, and attitudes.
Second-order influences are filtered through the first-
order effects and are composed of the organizational and
institutional components which determine the type of agents
and opportunities available for the socialization process.
Third-order and all subsequent influences are those which
are further removed and are filtered through the first- and
second-order effects. The effects of these influences are
so indirect and confounded by other variables that it
precludes definitive conclusions (U.S. DHHS, 1980).
Following a logical sequence, first-order effects will
be considered first.
An exemplary example of how difficult it is to
introduce a major innovative experience into a traditionally
structured medical school curriculum was demonstrated in the

-40-
Harvard experiment where the teaching of family medicine was
introduced in 1953. Rosenblatt and Alpert (1979) followed
up the first three cohorts of this early program to
determine the impact of the experience on subsequent career
choice, amount of family-oriented practice, and other
achievements. No statistically significant differences
could be found among the various groups, either within
cohorts or across time. Of critical importance to the
interpretation of the results of this very early attempt to
integrate family medicine into the curriculum is the fact
that at the time the program was introduced, Harvard had no
faculty in family medicine to serve as role models. In fact
in the latter years (cohort II and III) the directors were
two pediatricians, not family practitioners.
It is reasonable to believe that the availability of
role models and opportunities to practice the role will have
a beneficial effect on the numbers of students choosing a
specialty. In an effort to determine the beneficial and
detrimental parameters influencing the choice of family
practice as a career, Brearley and his colleagues (1982)
surveyed 134 first year residents of southeastern family
practice residencies on 18 curriculum components and
elements that influenced specialty choice. The positive
value of preceptorship experiences in family medicine during
the third and fourth years of medical school and association
with family physicians was striking. Detrimental influences
were peer group attitudes and the traditional curriculum

-41-
which the authors interpreted as being in opposition to
students' interest in family medicine.
Closely associated with this study is the work of
Harris et al. (1982) who inquired into the question "Is
participation in a family practice track program associated
with career choice decisions in family practice as evidenced
by residency selection?" (p. 610). Each year for two years,
20 students who applied for an elective in family medicine
were randomly selected for the experience. The students who
participated in the elective selected family medicine
residency at a significantly higher rate than those who
expressed an interest in the track but were not selected for
the elective. Although the subjects may be considered a
biased group, the positive effect of experiencing such a
program merits consideration.
Data pertaining to influences of persons and
experiences on career decisions published the same year as
the Harris study (Paiva et al., 1982) extended the
information in this area. The 1980 and 1981 classes of
Southern Illinois University School of Medicine were sampled
at three points in the curriculum, at the end of the basic
science period, the end of the clinical science period, and
the end of the clinical clerkships. Analysis of the data
showed that approximately three-fourths of the students
reported that a faculty member had some or very much
influence upon their specialty decisions. The major
influence came from full-time clinical faculty. The

-42-
influence of particular events was more pronounced in the
clinical phase than in the basic science period and the
responses were more dichotomous on the evaluation scale.
Those students choosing specialties selected by small
numbers of students such as psychiatry and anesthesiology
perceived the faculty influence to be a more important
factor than did students choosing the popular specialties of
internal medicine, surgery, and family medicine. Another
finding which has implications for curriculum planners is
the influence of the sequence of the rotations. The order
of the clerkships was particularly relevant for the students
who indicated prior to the clinical rotations that they were
undecided on their specialty as most of them ended up
choosing a specialty they had encountered during the first
few clinical rotations.
Changes in understanding and attitudes resulting from
an experience with a particular specialty were studied by
Samra et al. (1983). Students participating in an
anesthesiology rotation completed an attitudinal
questionnaire containing statements related to
anesthesiology. They also indicated their career choices.
The clerkship did not have any immediate impact on the
students' choices of specialties but there was evidence of
change in perception of the image of an anesthesiologist to
that of a physician rather than a technician. The students
also perceived the role of the anestheiologist to be broader
after the clerkship. Although limited by the research

-43-
design, the results are interesting enough to prompt further
research.
A conclusion to be drawn from the studies of first-
order effects is that clinical teaching faculty exert a very
strong role modeling effect on decisions of specialty
choice. Experiences during the clinical rotations are an
important component in the decision process, also, but it is
probably the interplay of the two which is critical. The
findings have implications for recruitment of students into
primary care specialties.
Second-order influences are those that indirectly
effect first-order factors. These might include financing
of medical education, the structure and organization of the
medical school, and research endeavors. Given the fact that
second-order must exert force through first-order, it is
hard to see how it would impact career choice among medical
students, but a number of researchers reported interesting
findings.
Zuckerman (1978) investigated the structural factors
within the educational milieu. He hypothesized these
factors linked together to form particular patterns which
resulted in different career outcomes. Based on structural
characteristics, 28 medical schools were dichotomized into
academic and clinical categories. Students were classified
according to academic standing. The type of residency
chosen was also categorized. Finally, the type of medical
practice entered was divided into four general areas. The

-44-
findings of the tabulation procedure revealed 1,145
different career patterns for 2,514 students sampled.
Obviously, the rigid tracking system Zuckerman hypothesized
was not supported by the findings of the study.
Friedman, Stritter, and Talbert (1978) examined closely
three community hospitals and an academic teaching hospital
where their students had clerkships. Based on the amount
and types of clinical experiences provided, they concluded
that it was fallacious to try to dichotomize facilities.
Some community hospitals paralleled teaching hospitals on
many characteristics.
There are probably multiple opportunities and
combinations of experiences at sites which might be
considered primarily clinical or strictly academic. These
two studies highlight the imprecision of the criteria used
to evaluate medical schools, not only by researchers but
also by accrediting agencies.
Medical school administrators and government policy
makers advocate founding new departments, increasing full
time faculty, and increasing research resources in an effort
to interest students in a given specialty. How valid is the
assumption that such investments produce dividends? The
Canadian study by Roos and Roos (1980) offers some insight
into this question which may be applicable to the U.S.
situation. They collected aggregate data on four charac
teristics of Canadian medical schools: number of full-time
faculty, number of part-time faculty, research grants,

-45-
and residency programs. This information was compared with
first-year medical student preferences for specialties, the
students' fourth-year choices, and their careers four years
after graduation. The results suggest that hiring more
full-time faculty in a specialty area and encouraging more
research in one area than another does cause physicians to
enter one specialty over another. Full-time faculty and
research activities have more influence on career choice
than clinical inputs such as part-time faculty and residency
programs. In future planning, then, it would seem that
increasing efforts in these directions would encourage
students in the direction where the faculty and research are
located.
Concerned about the decreasing interest in family
medicine since the early 1970s, Goldsmith (1982) surveyed
135 medical schools with 71 returns on their administrative
structure and number of graduates choosing family medicine.
The data revealed that schools with departmental status had
a higher percentage of graduates in family medicine. The
explanation for the observed association may be that those
with departmental status can effect admission policies,
command more time in the curriculum, attract more faculty,
and participate more heavily in career counseling.
A survey of departments of anesthesiology supported
these findings (Chandra & Hughes, 1984). Staffing patterns
in the departments were significantly related to the number
of students choosing to specialize in anesthesiology.

-46-
Career Factors
The factors examined to this point have been intrinsic
characteristics of the students and the medical school
environment but there are potential extrinsic influences
worth consideration. Hutt (1976) wrote, "Curiously enough,
these have received considerably less attention than the
earlier groups although it is, of course, the career factors
which policy-makers could most easily use to alter the
distribution of doctors between specialties" (p. 468).
Extrinsic factors include such areas as working conditions,
economic incentives, and societal demands or needs.
There is considerable diversity between specialties
with regard to working conditions. Some specialties such as
general surgery, pediatrics, and obstetrics require long,
erratic hours with responsibilities at night and on
weekends. Other specialties are more amenable to regular 9-
to-5 schedules. Style of practice, solo, group, or
hospital, is related to specialty. For example, primary
care spcialties are principally ambulatory care practices
with little emphasis on hospital practice. Other
specialties such as surgery are principally hospital-based
practice.
As was noted earlier, women are willing to forego
career time for more family time (Bergquist et al., 1985;
Zimny & Shelton, 1982). They want more flexible hours and
will usually opt for specialties that allow for that time.
In Cuca's study (1979), women's career plans differed from

-47-
their male counterparts in (a) length of residency, (b)
employment settings, and (c) intended areas of specialties.
The women planned to spend fewer years in residency training
and wanted salaried practice options rather than private
practice styles. These first two factors could explain the
third--the tendency for women to participate at a higher
rate in primary care specialties as opposed to some of the
nonprimary care specialties which require long training
periods and provide less flexible hours.
This is not to say that males do not have personal
goals that include time for family, leisure activities, and
other pursuits and may consider these interests in making
their choices. Up to this point, however, these goals have
not been evidenced. It may be that as more families become
two career families and more couples share family
responsibilities, the males will express some of these same
needs and demands.
The financial compensation for specialists in
nonprimary care specialties is generally higher than primary
care specialties. Testing for economic incentives as
factors in career choice is difficult. As stated earlier,
people will not usually admit to what might be considered
materialistic motives.
Funkenstein had some success with economic factors in
relation to career choice. Funkenstein's (1978)
longitudinal, prospective study of Harvard medical students
covered a period of five eras from 1958-1976. From his

-48-
investigations he concluded that economic incentives and
ideology are more compelling than basic characteristics or
original plans in predicting career outcomes. Further,
societal forces prevailing at the time of the decision exert
significant influence. As Funkenstein (1978) observed, the
"Sputnik" era saw an increase in emphasis on mastery in the
sciences with increased resources available in scientific
subjects. Students admitted to medical school during this
period had heavy science background. The 1960s and 1970s
were periods of social conscience-raising and a time when
individualism was prominent. This was followed by decreased
funding in some areas of medicine so that careers in
academic medicine were not as attractive (Funkenstein, 1978,
p. 30) .
Society continues to be enamoured with technology and
is willing to reward those who provide it. High-tech in the
health field has brought us heart transplants, microsurgery,
and life-sustaining equipment (Naisbitt, 1982). It could be
hypothesized that high tech translates into highly
specialized physicians to administer this type of medical
care and a renewed interest in family doctors to balance the
high-tech with highly personal care (Naisbitt, 1982, p. 41).
On the other hand the organization of the health care
system is changing (Ginzberg, 1984). As large corporations
buy and operate hospitals, medical care has become a
businessa business for profit (Weiss, 1982). High
technology is expensive; subspecialists are expensive;
tertiary care hospitals are expensive. The most economic

-49-
medical care is administered by primary care facilities
(Neuhauser, 1983). It follows then that there will be a
demand by the hospital corporations for more primary care
physicians to staff their hospitals. Subspecialists will be
used sparingly and only on consultation. This change in the
organization of the health care system is likely to have an
effect on specialization.
Some of the career factors discussed here are amenable
to manipulation. Working conditions can be changed to be
more responsive to the needs of those seeking primary care
specialties in order to increase their numbers. Societal
factors are not so easily defined and interventions are
difficult to implement. Economic incentives can be
addressed in several ways. Decreasing the discrepancy
between primary care and nonprimary care specialties would
help offset this factor. The area needs more study upon
which to base policy changes.
Summary of the Review of Literature
Although the specialty choice research literature is
voluminous, only rarely does unequivocal evidence of
specific influences on physician specialty choice surface.
The conceptual differences and methodological problems make
comparisons across studies difficult.
Since admission criteria have not been well defined,
the weight given to academic and personality variables by
admissions committees is unclear. The study of personality

-50-
factors has yielded little help with regard to prediction of
specialties. Personality tests have distinguished between
psychiatrist and surgeons but have been less helpful with
regard to other specialties. Academic measures available on
admission become less useful over time being more predictive
of early medical school performance but not later clinical
abilities. Little difference has been found in recent
studies of academic performance between those choosing
primary care and others.
Recognizing the limitations of the current admission
data, a viable alternative to the present admission process
was offered in the Summary Report of the Graduate Medical
Education National Advisory Committee (U.S. DHHS, 1980):
An idea which has been put forth as a solution to
another problem, that of the inevitable
disappointment of qualified applicants not being
accepted to medical schools because of heavy
competition from many other equally qualified
applicants, might also be a solution to the issue
of personality screening for admission to medical
schools. The idea is that of a two stage
screening process, the first stage involving
screening on the basis of intellectual and
academic qualifications, the second involving a
random lottery. Such a process would assure that
acceptees would (a) be intellectually qualified,
(b) have an equal chance for selection at the
second stage (with "fate" making the final
decision and thus shouldering the blame for
nonacceptance) and (c) display personality traits
in proportions representative of those occurring
in the population of qualified applications.
(Vol. V, p. 17)
Sociodemographic variables have provided some limited
insight into specialty selection. Size of home community,
age, and marital status have been most consistent of the
background variables in distinguishing primary versus

-51-
nonprimary care specialists. The older, married students
from small towns are most likely to choose primary care
specialties. With all other background characteristics, the
results are mixed.
It appears that the variables brought to the medical
setting diminish in importance as factors related to the
medical training system increase in influence on specialty
choice. Preliminary data suggest that role models and the
opportunity to practice a role reinforce or refute the
beliefs and attitudes that students bring with them into the
clinical setting and consequently effect their perception of
the various medical specialties. However, the studies in
this area have largely used self-selected groups of subjects
who take an elective course in a clinical clerkship so that
the results are not generalizable.
Career factors have not received adequate attention by
researchers. Information on the career and personal goals
of young doctors could provide clues as to the needed
changes in the work setting to make shortage areas more
attractive career options.
In sum, what the literature seems to indicate is that
(a) results with sociodemographic variables are mixed and
therefore require further study; (b) even though there are
methodological problems in preliminary studies,
opportunities for interaction with role models and role
playing in the clinical setting holds promise for
influencing specialty choice but further study is needed;

-52-
and (c) insufficient data are available on career and
personal goals of medical students from which to make
important educational and policy decisions.

CHAPTER III
METHODOLOGY AND INSTRUMENTATION
Setting for the Study
The study was conducted at a large southeastern
university college of medicine, the University of Florida
College of Medicine. Unless otherwise specified, the
descriptive information pertaining to the setting for the
study was summarized from the College of Medicine catalog
(University of Florida, 1985). The University of Florida is
a public, state-funded institution established in 1906. The
College of Medicine, a component college of the J. Hillis
Miller Health Center of the University of Florida, admitted
its first class in 1956. College of Medicine educational
offerings include undergraduate medical courses leading to
an M.D. degree, graduate medical education experiences,
residency programs in various specialties and
subspecialties, graduate courses in the basic medical
sciences leading to a Ph.D., and postgraduate fellowships
in clinical and scientific areas.
Medical school faculty and students are housed in the
J. Hillis Miller Health Center and the adjoining Shands
Hospital located on the main campus of the University of
Florida and in the Gainesville Veterans Administration
-53-

-54-
Medical Center situated across the street from the Health
Center. In addition to these sites, the College of Medicine
has developed educational programs in community settings
where students can gain insight into the day-to-day problems
of minor and major illnesses as they occur in urban and
rural settings. While university medical centers such as
Shands Hospital play a unique and indispensable role in the
student's educational experiences, fewer than 5% of
physician/patient contacts occur in hospitals (White,
Williams, & Greenberg, 1961). The Association of American
Medical College's Panel on General Professional Education of
the Physician (GPEP) (1984) recommended that medical
students' general professional education be a balance
between acute illnesses and working with patients and
communities to prevent or ameliorate disease (p. 16). This
requires ambulatory settings for required clerkships.
The community health experiences available to University of
Florida medical students include preceptorships with
practicing physicians throughout the state of Florida, a
rural health clinic in Dixie County, Florida, and family
practice centers. The preceptors are primary care
physicians who have volunteered to accept students for a
specified period of time, usually two or four weeks of
training. Generally the students work in the physician's
office seeing patients with the physician. The student
accompanies the doctor on hospital rounds and becomes
involved in all medically related activities. Often the

-55-
student is lodged in the preceptor's home so that she or he
is provided an opportunity to view and understand non
medical aspects of the physician's roles and obligations to
the community as well as the usual medical responsibilities.
There are elective and required assignments in two
University of Florida affiliated family practice centers,
the Family Practice Center in Gainesville and St. Vincent's
Family Practice Center in Jacksonville. Here students work
with residents in family practice training and faculty
members to become acquainted with the health care delivery
in urban areas.
The four years of undergraduate medical education
leading to the M.D. degree at the University of Florida are
divided into three parts, Phases A, B, and C. During Phase
A, the first 12 months of study, students are presented the
core of basic medical science courses necessary for clinical
training. The course schedule includes biochemistry, human
genetics, anatomy, bacteriology, immunology, virology,
physiology, neuroscience, human organ system development,
and human behavior.
Phase B is divided into preclinical and clinical
experiences. The first 27 weeks of the second year students
take pathology, pharmacology, physical diagnosis, and
interviewing. A course in social and ethical issues is
offered in the last part of the 27-week block preceding the
clinical portion of Phase B.
The major portion of Phase B is the 12-month long
clinical clerkship period in which students rotate in groups

-56-
of 20 among the 6 clinical specialty services, pediatrics,
internal medicine, obstetrics/gynecology (ob/gyn), surgery,
psychiatry, and family medicine. Students work as members
of the medical patient care teams experiencing direct
patient care responsibilities.
Phase C occupies the last 15 months of medical school.
Students are required to complete four additional weeks of
surgery and four weeks of medicine clerkships. They review
clinical pharmacology, pathophysiology, and study infectious
diseases for an eleven week period of time. The last 10
months are reserved primarily as elective time which allows
students to select clinical activities of special interest
and interview at the various residency programs where
they are interested in applying for graduate medical
education.
Revision and Development of the
Questionnaire
Historical Background of the Questionnaire
The instruments used in this research project were
adapted from two survey questionnaires used to study
i
attrition from Florida's family medicine residency programs.
In the mid-to late 1970s these residency programs were
experiencing substantial numbers of residents who dropped
out after one year of training (Dallman, Crandall & Haas,
1980). In order to study the issue of attrition among its
residency programs, a cooperative project was begun in 1977

-57
by the Department of Community Health and Family Medicine at
the University of Florida, the Florida Council of Graduate
Education for Family Practice, and the Florida Academy of
Family Physicians (Dallman et al., 1980, p. 833). One of
the purposes of the residency study on attrition was to
delineate characteristics of two groups of residents, the
first group being those who after one year of training
planned to continue on to completion of the residency
program. The second group included residents who indicated
at the end of the first year of residency that they planned
to leave the residency program.
Dallman and his group (1980) mailed a questionnaire to
all incoming residents in Florida Family Medicine Residency
programs in August of 1977 and those entering in August,
1978. The questionnaire was designed to obtain data on each
resident's social, familial, and educational background, as
well as the resident's beliefs and attitudes concerning
family medicine and primary care, and the importance the
resident gave to financial security, autonomy, lifestyle,
and social position.
The second questionnaire was mailed to each resident at
the end of the first year of residency training. The same
attitude and opinion items were included on the second
questionnaire as appeared on the entering form except that
the background items were omitted on the second survey
(Dallman et al., 1980, p. 833).

-58-
Analysis of the residency survey data demonstrated that
a number of the questionnaire items discriminated between
dropouts and nondropouts at the end of one year of family
practice residency training. Dallman and his group (1980)
found that while sociodemographic characteristics such as
sex, marital status, age, race, citizenship, and hometown
location and size did not discriminate between the dropouts
and nondropouts, source of financial support did
differentiate the two groups. Those residents who were
their sole support were more likely to leave training than
those who drew financial support from spouses or others
(Dallman et al., 1980, p. 834). The researchers also found
that the rank order of career goals differed between the two
groups with dropouts more interested in economic security
and time for family and leisure activities. The attitudes
section showed a significant difference between the dropouts
and nondropouts also. Those who left the programs felt
curative medicine was more interesting than preventive or
health maintenance (Dallman et al., 1980, p. 834). Based on
the significant findings from the residency study and the
imperative need for understanding of the factors influencing
choice of family medicine and primary care specialties, it
was decided to introduce the questionnaire earlier in the
educational process to attempt to identify characteristics
of those likely to choose these specialty areas over others.
Thus, the present study concerning identification of factors
related to student choice of a medical specialty was begun.

-59-
Development of the Medical Student Questionnaire
The medical student questionnaire used in this study
(Appendices A and B) and first administered in 1983, was
essentially identical in content to the resident attrition
survey conducted in 1977 and 1978. The changes to the items
in the medical student questionnaire were principally to
correct for differences in level of training between
residency and undergraduate medical education. For example,
the section in the residency questionnaire entitled "Family
Practice Residency Content" allowed residents to address
components of the ideal family medicine training program.
In the medical student questionnaire this section was
changed to "Medical Education Content" and retained some of
the same choices of answers where appropriate.
The medical student questionnaire was designed to
gather information from a number of broad categories.
Section A, of the entering questionnaire, deals with
background data. This section includes aspects of personal
data, family, finances, education, and work experiences.
Section B addresses obligations to federal agencies for
loans and scholarships the student may have incurred and
plans for future style of practice, e.g., solo, partnership,
group. These two sections required the student to choose
between alternative answers by placing a check or an X in
the space beside the alternative of their choice. Some of
the questions required the respondent to write in a short

-60-
answer as when asked age and major subject of undergraduate
education.
The importance the student placed on financial
security, autonomy, lifestyle, social position in the
community, and position in the peer group was examined in
Section C, Career Goals. Students were asked to rank in
order of importance items that relate to career goals. The
most important goal was assigned the rank of 1, the second
most important goal received a rank of 2, and so forth with
no choice being used more than once.
Section D, Medical Education Content, lists eight items
pertaining to medical education. Students ranked the items
from 1 to 8 according to how important each was considered
to be in producing the best physicians.
Improving Health Status, Section E, lists five
different approaches to improving the health status of
people not receiving adequate health care. The student
ranked them in order of his/her opinion as to their
effectiveness with the rank of 1 equaling the most effective
and the rank of 5 being least effective in improving health.
The Attitudes Toward Primary Care and Family Medicine
section, Section F, contains 29 statements which reflect the
goals and philosophy of primary care and delineate family
medicine and primary care from specialty-oriented medicine.
The attitude section concentrates on the following areas:
(a) preventive medicine, (b) continuity of care, (c)
managing psycho-social problems, (d) ambulatory versus

-61-
hospital-based care, (e) broad spectrum of medical care,
(f) attitudes toward consultants and paramedical personnel,
and (g) perspectives toward research in primary care
settings.
Student respondents were asked to indicate whether
they strongly agree, tend to agree, are not sure, tend to
disagree, or strongly disagree with each of the 29
statements on attitudes and beliefs concerning primary care
and family medicine. A Likert scale of numerical values was
assigned to each response depending on the degree of
agreement or disagreement. Strongly agree received a value
of 1 while strongly disagree was valued at 5.
The last section of the entering survey of medical
students asked the respondent to indicate to the best of
his/her knowledge future plans for residency training. The
alternatives from which the student could choose were (a)
family medicine, (b) pediatrics, (c) general internal
medicine, (d) other primary care, (e) other specialty/sub
specialty, or (f) uncertain. Students answering under the
category "other primary care and other specialty/sub
specialty" were asked to specify the specialty area by
writing in a line beside the choice indicated. The last
section was included on all of the entering questionnaires
except the questionnaires for Rotation 2. It was
inadvertently omitted from this group's questionnaires.
Thus, 20 students did not have the opportunity to express
their specialty preferences. A comparison of the
sociodemographic characteristics of the students who did not

-62-
respond to this section revealed no significant differences
at £ .01 between those who responded and those who did not.
The follow-up survey questionnaire items were identical
to the entering survey with the exception of the deletion of
the first two sections containing social, personal, and
demographic data. The last section on future plans for
residency training was also not included in the follow-up
survey. The abbreviated follow-up survey included four
sections common to the entering questionnaire: Career
Goals, Medical Education Content, Improving Health Status,
and Attitudes Toward Primary Care and Family Medicine.
Administration of the Questionnaire
Sample
The sampling frame for the study included all medical
students in the third year of required clinical clerkships
in 1983. Third-year medical students were chosen because
they were further along in the professional and
socialization process and because fourth-year students are
geographically scattered due to the nature of the curriculum
which allows for externship at other sites and time out of
school for interviews for residency positions.
During the clinical portion of Phase B students rotate
in small groups through the individual clinical courses or
clerkships. These clerkships were scheduled in blocks of
eight weeks each. The Community Health and Family Medicine
clerkship was allotted six weeks of one of the eight-week

-63-
blocks sharing the other two weeks with a neurology
clerkship. The Community Health and Family Medicine
clerkship ran concurrently with the students' clinical
experience in medicine, pediatrics, psychiatry, surgery, and
ob/gyn.
In the Community Health and Family Medicine rotation
all clinical experiences were in settings outside of the J.
Hillis Miller Health Center and Shands Hospital. The six
weeks were scheduled as follows: four weeks in a family
practice center (either the Family Practice Center in
Gainesville or St. Vincent's Family Practice Center in
Jacksonville) and two weeks in a rural health clinic either
in Cross City or in Mayo, Florida. Participation in the
various community settings provided the students with first
hand experience in the medical and health problems as they
exist in the different communities.
The rotation year begins in March each year and
continues for 12 months. The schedule for the 1983-84
rotation year was as follows:
Rotation 1
Rotation 2
Rotation 3
Rotation 4
Rotation 5
Rotation 6
March 21-May 14, 1983
May 15-July 9, 1983
July 10-Sept. 3, 1983
Sept. 11-Nov. 5, 1983
Nov. 6, 1983-Jan. 14, 1984
Jan. 15-March 9, 1984
Approximately three weeks prior to the beginning day of
the rotation, an orientation meeting was held for the
students. The goals and objectives were explained, the
course syllabus was distributed, and the evaluation system
was discussed. The sequencing of the various clinical

-64-
experiences in Community Health and Family Medicine was
explained. A description of the sites was provided in the
syllabus and this was reiterated during the orientation.
Students were given the opportunity to ask questions about
lodging accommodations, travel arrangements, and any other
concerns. Assignment to the clinical sequence was by
lottery.
Administration of the Entering Survey
At the beginning of the orientation session, after all
students had arrived and taken a seat, the entering survey
questionnaire was passed out to each student. Attached to
the front of the questionnaire was a cover letter (Appendix
C) signed by the chairman of the Department of Community
Health and Family Medicine stating the purpose of the survey
and assuring confidentiality of responses. Students were
encouraged to participate but were told that the
participation was voluntary. The author thanked the
respondents for their help with the study.
The written information was verbally reinforced as the
questionnaires were distributed. Students were told (a) a
follow-up questionnaire would be mailed to them at their
campus mailboxes or at the clinical site at the end of the
rotation; (b) to put their name and social security number
on the questionnaire but that a code number would be
assigned when the data were entered in a computerthis
would maintain confidentiality while allowing the following

-65-
data to be collated with the entering data; and (c)
participation or nonparticipation would not affect their
rotation grade as those clinical faculty members responsible
for grading would not have access to individual's responses.
It required approximately 30 minutes for students to
complete the questionnaire.
Follow-Up Survey Questionnaire
The follow-up administration of the survey
questionnaire was accomplished at the end of the rotation
time period. The students were not told that the follow-up
questionnaire was identical to the entering survey with the
exception of the background data questions. The survey
questionnaire was included in the envelope with the student
evaluation of rotation forms. Two self addressed envelopes
were included for the respondent to return the evaluation of
rotation and the survey questionnaire. A cover letter
(Appendix D) stated that the College of Medicine required
student evaluation of rotation. It reminded students that
they had participated in the entering survey and asked their
cooperation in the follow-up survey.
A follow-up code sheet was used to record the name,
social security number, and code number for each respondent.
When the follow-up form was returned an X was placed in the
appropriate column. Those students who did not respond
within two weeks were sent a note reminding them to send the
questionnaire back. If another two weeks passed with no

-66-
response, another note (Appendix D) was sent with a copy of
the questionnaire. The envelopes enclosed in the student
correspondence were addressed to a campus mailbox and
therefore required no postage if mailed at a university post
office. Students returned the questionnaires in the
envelopes provided. Since most students were physically
located at the Health Center where access to campus mail
drop boxes were readily available, this was felt to be an
acceptable method of securing their return.
Residency Match
All students who enter the profession of medicine must
at some point make a secondary career choicethe type of
medical specialty they wish to practice. Training in the
medical specialty of choice is acquired through residency
programs. Students begin applying to residency programs and
interviewing at the various sites by the beginning of the
fourth year of medical school. The final selection is made
through a nationwide, computerized matching system, the
National Resident Matching Program (NRMP). A majority of
students participating in the match process are placed in a
residency program of their choice although it may not be
their first ranked preference. In January of their senior
year in medical school, students ranked the residency
programs in which they wanted to take their first year of
graduate training. The students' selections are matched
against the residency programs' choices of candidates.

Results of the match were received by the individual
students, the residency programs, and the medical schools
the following March.
Research Design
To investigate the research questions for this study, a
pre- and post-research design was used with multiple group
replication measures. The research procedure involved
observations taken immediately before and after a clerkship
and again after a passage of time. The design used intact
groups of third-year medical students previously formed into
rotation groups through a lottery system.
The design for this study was considered a variation of
the Recurrent Institutional Cycle Design, a strategy for
field research described by Campbell and Stanley (1963) as
appropriate to those situations in which some aspect of the
institutional process is on a cyclical schedule. In field
research in educational settings where restrictions prevent
an experimental design with control of who would be exposed
to the experimental variable and the effects of a global and
complex construct are sought, this design offers a measure
of strength over the one-group pretest-posttest design. The
fact that the treatment variable, the clerkship, is
continually being presented to a group previously exposed
and a group about to be exposed makes for some degree of
experimental control as elements of the cross-sectional and
longitudinal approaches are present.

-68-
The Recurrent Institutional Cycle Design (Campbell &
Stanley, 1963) as adapted for use in this study is shown in
Figure 1.
The design for this study required the collection of
preclerkship data from each medical student at the beginning
of the family medicine rotation. Following preclerkship
data collection, each rotation group received a learning
exposure to family medicine. A second administration of
the questionnaire occurred immediately following the
family medicine rotation. The final component of the
design involved residency match choices for each study
participant.
The cyclical schedule for this design was such that at
one and the same time a group which had been exposed to the
clerkship (X) and a group which was about to be exposed were
measured. As Figure 1 demonstrates, subjects in Rotation 1,
which began in mid-March, were completing the follow-up
questionnaire in mid-May at approximately the same time as
the subjects in Rotation 2 received the entering
questionnaire. Each subsequent rotation began a new cycle
as the previous rotation ended their clerkship. Students
ranked the residency programs for the National Resident
Matching Program in January of their senior year of medical
school. Residency match choice information was received at
the sponsoring institution in March, 1985.

I 1 = Orientation g
X = Family Medicine Clerkship U11
0 = Questionnaire, Entering
and Follow-Up
(3 = Ranking of Residency
Programs g
/\ = Resident Match I x _
Choice Results
0
12
- X -
10
0.
0,
- X -
S
E
Q
u
E
N
C
E
3-
2-
I-
- X -
0
0
3 u4
dZZ
0.
o
A
o
A
o
A
o
A
o
A
o
A

M
1983
i 1
A M
1 1 i 1 1
J J A SON D
1 1
J F
1
M
-4 1 1
A M J
i i 1 1 1
JASON
1
D
1 1 1-
J F M
1985
Phase B
Phase C
MONTHS
Figure 1. Data Collection Points

-70-
Data Analysis
One important criterion in the choice of which
statistical procedure to use to test a research hypothesis
is that the data meet the assumption of a particular level
of measurement (Marks, 1982). The elements of observation
in this study were considered nominal- and ordinal-level
data. When words or symbols are used to classify persons or
characteristics into groups, they constitute a nominal
scale. In this research, sociodemographic characteristics,
specialty preference groups, and specialty choice groups
were identified as being nominal-level variables. With
ordinal-level data, the classes or categories stand in
relation to each other so that rank ordering is permitted.
Students' rankings of career goals and their attitudes
toward primary care measured on a Likert scale resulted in
information contained in an ordinal scale.
Since nonparametric statistical tests are appropriate
for data inherently in ranks or numerical scores having the
strength of ranks, and some apply to nominal-level data as
well, techniques of inference drawn from this body of
methods were chosen for this study. The following
nonparametric statistical procedures were used to test the
10 null hypotheses:
1. Hypothesis 1, there is no association between
sociodemographic traits of medical students and their
specialty preferences in the third year of medical school,
was tested using the chi-square test (X^). The chi-square

-71-
test is a statistical technique that tests the difference
between what is observed and what might be expected under
some theoretical model. It can be applied to one-sample
cases, two-sample cases, or larger cases such as the three
subsamples of specialty preference contained in this
hypothesis. The three specialty preference response
variables were primary care, subspecialty, and undecided.
Chi-square is an appropriate statistical tool when
frequencies in classified categories constitute the data of
research such as the sociodemographic characteristics of
this study.
2. The chi-square test for two independent samples was
applied to Hypothesis 2 to determine if significant
associations existed between sociodemographic
characteristics of the medical students and their medical
specialty choices as evidenced by residency selection of
either primary care or nonprimary care specialties.
The chi-square test is valid only if at least 80% of
the cells in a contingency table have frequencies of at
least five. Where small cell frequencies are involved,
categories may be collapsed as long as it is done logically
and not merely to achieve significance or create bias. Some
of the sociodemographic characteristic categories were
collapsed for purposes of statistical analysis. When small
cell frequencies persisted, Fisher's exact test was used as
an alternative test of significance for 2x2 tables
(Champion, 1970). The Agresti-Wackerly method of computing

-72-
exact tests of independence was used for R x C cross
classification tables when R or C was greater than two
(Agresti & Wackerly, 1977).
3. Hypothesis 3 tested the null hypothesis that there
is no difference between medical students grouped by
sociodemographic characteristics according to their
attitudes toward primary care before or after a family
medicine clerkship. Individual attitude scores were formed
by summing each student's responses to the Attitudes about
Primary Care and Family Medicine section of the
questionnaire. Attitudes were measured before and again
after a family medicine clerkship. The Kruskal-Wallis one
way analysis of variance was used to determine whether or
not there was a difference in distribution of attitude
measurements between the groups of students identified by
their sociodemographic characteristics. The primary
assumption for this test is that the data be ordinal-level
data. The data to be ranked should be continuous rather
than discrete. The Kruskal-Wallis makes no assumptions
concerning distributions and it is appropriate for unequal
sample sizes, both large and small. Based on these
assumptions it was considered a suitable test for this
hypothesis. When there are more than five subjects in each
element or group, as occurred in this study, the Kruskal-
Wallis H value is treated as a X2 for interpretive purposes
(Champion, 1970, p. 188).
4. The Kruskal-Wallis one-way analysis of variance
test was again deemed the appropriate choice of statistical

-73-
analysis for Hypothesis 4 which sought to determine the
difference between the three specialty preference groups in
the third year of medical school and the two specialty
choice groups from the fourth year with respect to career
goal rankings. The career goal data met the primary
assumption of this procedure, namely that it be ordinal-
level measurement as well as the assumption that the samples
were independent.
5. Spearman's rho measures the degree of association
between two sets of ordinal-level data. Therefore, for
Hypothesis 5 it was considered an appropriate statistical
procedure to investigate whether or not medical students'
rankings of career goals were associated with their
attitudes toward primary care before a family medicine
clerkship and again, after the family medicine clerkship.
6. Hypothesis 6 was tested with the Kruskal-Wallis
technique. In the computation, the difference in students'
rankings of career goals before and after a family medicine
clerkship were obtained. These difference scores were then
entered into the Kruskal-Wallis test to decide whether the
medical specialty groups were from different populations.
Since the scores could be considered at least ordinal data,
the Kruskal-Wallis was appropriate to the data.
7. Hypothesis 7 tested whether the association between
medical students' third-year specialty preferences and their
fourth-year specialty choices was different from what might
be expected by chance. As the data were in discrete

-74-
categories at the nominal level, the chi-square test was a
suitable one.
8. Hypothesis 8, there is no difference in medical
students' attitudes toward primary care before and after a
family medicine clerkship, was tested with the Wilcoxon
matched-pairs-signed-ranks test. In this case, the medical
students acted as their own controls in a before-and-after
design. Another assumption is that the level of measurement
be interval-level. As the attitude scores before-and-after
the clerkship were ordered in a numerical scale, the
information was judged to be at least equal-appearing. A
definite advantage of the Wilcoxon test is that it considers
not only the direction of the difference in scores but also
the magnitude of the change. As Champion (1970) suggested
for sample sizes exceeding 25, an alternative formula, a
modification of the Z test, was used to complete the
Wilcoxon matched-pairs-signed-ranks test (p. 169).
9. The Kruskal-Wallis one-way analysis of variance
test was judged to be the proper procedure to test whether
attitude scores of one specialty preference group were
significantly different from the two other specialty
preference groups of third-year medical students. Since the
three groups of specialty preference could be considered
independent and the attitude scores represented at least
ordinal-level data, the Kruskal-Wallis test was considered
the correct method for Hypothesis 9.
10. Hypothesis 10 was also tested with the Kruskal-
Wallis test. The hypothesis was that there is no difference

-75-
between groups of medical students identified by their
fourth-year specialty choice with respect to their attitudes
toward primary care before and after a family medicine
clerkship. The two groups of specialty choice constitute
independent groups and the measure of attitudes before and
after the clerkship constitute at least ordinal-level data;
therefore, the Kruskal-Wallis was concluded to be the right
procedure.
When multiple statistical tests are conducted on data
as they were in this study, the chance of making erroneous
conclusions is increased and, therefore, the overall
significance level (OSL) must be set conservatively. The
significance level of _< .01 was used for hypotheses tested
in this study so that the OSL would not exceed .10.
Marks (1982) discusses two philosophies concerning the
significance level used for testing the null hypothesis (p.
22). One philosophy is to perform the statistical test
using a predetermined level of significance and to use the
test results only to see if the predetermined level of
significance is achieved. Another approach is to present
the attained significance level, decide whether or not the
evidence is sufficient to reject the null hypothesis, and
let others draw their own conclusions from the results. The
second rationale was used in presenting the results of this
study.

CHAPTER IV
PRESENTATION OF RESULTS
Sociodemographic Characteristics of
the Study Group
Data reported here were obtained from a survey of 109
medical students of the 1984-85 graduating class of a state-
supported medical school in the south/southeast. The data
are presented in Table 1. Some categories from the original
survey have been combined for purposes of statistical
analysis and ease in reporting. Students ranged in age from
22 to 35 with a mean age of 25 years. Males made up 73% of
the respondents; females, 27%. Whites were heavily
represented with 88% of the group while all other races
combined accounted for the additional 12%. The religious
preferences of the group were (a) 44% Protestant, (b) 22%
Catholic, (c) 11% Jewish, and (d) the remainder indicated no
religious preference. There was almost no variation between
the percentages of students indicating that religion played
a great deal of importance in their lives (37%), some
importance (35%), and very little to none at all (28%).
Eighty-two percent of the students were from the
south/southeast region of the country with about equal
representation from small towns, mid-sized cities, and
metropolitan areas. Forty-seven percent of the subjects
-76-

Table 1
Summary of Sociodemographic Characteristics of Subjects
Sociodemographic
Sociodemographic
Characteristics
n
Percent*
Characteristics
n
Percent
Sex
Marital Status
male
79
73.1
married
34
32.1
female
29
26.9
other
72
67.9
Race
Citizenship
white
94
87.9
U.S.
105
97.2
nonwhite
13
12.1
other
3
2.8
Hometown Location
Father's Occupation
south/southeast
89
82.4
professional/technical
56
52.3
not south
19
17.6
business/other
51
47.7
Hometown Size
Mother's Occupation
0-99,000
51
47.6
housewife
54
50.0
> 100,000
56
52.3
work outside home
54
50.0
Hometown Economic Base
Parents' Annual Income
farming
11
10.6
>$50,000
31
31.3
business
37
35.6
$0-50,000
68
68.7
heavy industry
light industry
mixed economy
5
2
49
4.8
1.9
47.1
Source of Present Support
parents
other
31
77
28.7
71.3
Level of Education-Degree
bachelors
98
90.7
Indebtedness
86
20
81.1
18.9
graduate
10
9.3
yes
no
Amount of Debt
$1000-5000
7
8.2
> $5000
7U
91.8
Sociodemographic
Characteristios
Religious Orientation
Protestant
Catholic
Jewish
other
no preference
Religious Importance
very much
some
a little/none at all
Parents' Religion
Protestant
Catholic
Jewish
other
no preference
High School Attended
public
private
College Attended
public
private
Undergraduate Major
science
other
n Percent*
46
43.8
23
21.9
12
11.4
8
7.6
16
15.2
40
37.4
37
34.6
30
28.0
50
47.6
30
28.6
11
10.5
9
8.6
5
4.8
83
78.3
23
21.7
76
71.0
31
29.0
90
86.5
14
13.5
I
I
Overall category may not equal 100% due to rounding

-78-
judged their hometown economic base as a mixed economy.
Thirty-six percent came from places dominated by a business
economy. Farming communities contributed 11% to the sample
and only 6% were from industrial areas.
When asked to indicate their parents' occupations while
they were growing up, 52% of the students checked physician,
health care provider, or other profession. Forty-eight
percent cited business, clerical/sales, skilled, unskilled,
farm/farm worker, and other as their father's occupation.
Mothers' occupations were equally divided between housewife
and working outside the home. Considering their parents'
annual income, 69% had parents with incomes that fell in the
$50,000 or less categories compared with 31% in the above
$50,000 stratum.
Seventy-one percent of the students indicated that
their principal source of support came either from
themselves, spouses, or a combination; parents supported
29%. The financial status is further clarified by the
information that 81% of the respondents reported some
educational debt. For those reporting indebtedness, 92% had
debts greater than $5000.
For the majority of the students, their education took
place in public high schools (78%) and public-supported
colleges (71%). As would be expected, 86% majored in
science with 91% obtaining a bachelor's degree and 9%
achieving a graduate degree.

-79-
Specialty Preferences
Prior to a third-year Family Medicine clerkship, the
medical students were asked to indicate their future plans
for residency training. General internal medicine,
pediatrics, family medicine, and other primary care
specialties were the most frequently chosen areas with 47%
of the group. About 24% expected to go into subspecialty
residencies and 29% were still undecided on a specialty at
that point. Table 2 shows the students' career plans by
their preferred specialties.
Residency Plans
The students' first choice of career plans by residency
choice are presented in Table 3. Primary care specialties
were the first choice of career activity of 66% of the
students. Thirty-six percent planned residencies in
internal medicine with 14% choosing family medicine, 7%
pediatrics, and 9% transitional residencies. Approximately
33% of the 1985 graduates elected subspecialty residencies.
The most frequently named subspecialty was obstetrics and
gynecology with 10% of the subjects. Slightly less than
4% opted for residencies in surgery, radiology, and
pathology.

-80
Table 2
Medical Specialty Preferences
of Third-Year
Medical
Students
Specialty Preference Groups
n
Percent
Primary care
34
47.2
Specialty-subspecialty
17
23.6
Undecided
21
29.2
Total
72
100.0

-81-
Table 3
Residency Choice of Fourth-Year Medical Students
Specialty Choice Groups
n
Percent
Primary Care
family medicine
15
13.8
pediatrics
8
7.3
internal medicine
39
35.8
transitional
10
9.2
Total
72
66.1
Subspecialty
ob/gyn
11
10.1
ER medicine
1
.9
anesthesiology
2
1.8
neurology
1
.9
dermatology
1
.9
pathology
4
3.7
radiology
4
3.7
radiation therapy
3
2.7
physical med. & rehab.
1
.9
surgery
4
3.7
plastic surgery
1
.9
neuro surgery
1
.9
orthopedic surgery
1
.9
psychiatry
1
.9
Total
36
32.9

-82-
Results of the Data Analysis for Testing
the Hypotheses
Hypothesis 1
There is no association between sociodemographic
characteristics of medical students and their medical
specialty preferences expressed in the third year of medical
school. The chi-square test was used to determine
separately for each sociodemographic variable whether
medical specialty preferences of third-year medical students
were dependent on sociodemographic characteristics. Medical
specialty preferences were categorized as primary care,
subspecialty, and undecided based on the students' responses
to the questionnaire item, Please indicate to the best of
your knowledge your future plans for residency training.
Since multiple statistical tests were performed, a
conservative overall significance level of £ = .01 was used
to decrease the probability of falsely rejecting the null
hypothesis. The results of the chi-square procedures used
to test the first set of variables related to Hypothesis 1
are reported in Table 4.
The overall significance level of < .01 was not
achieved in the comparison of sociodemographic variables and
third-year specialty preferences of medical students.
Therefore, the null hypothesis was not rejected.
However, one sociodemographic category, parents' annual
income, was significantly related to specialty preference at
the < .01 level. This test produced a chi-square value

-83-
Table 4
Association between Third-Year Specialty Preference
and Sociodemographic Characteristics
Sociodemographic
Characteristics
df
X2
Vaue
Significance
Sex
2
5.513
.06
Marital Status
2
.946
.62
Religious Orientation
8
3.097
. 94 +
Religious Importance
4
6.027
.20
Race
2
4.010
. 16 +
Citizenship
2
1.133
1.00 +
Hometown Size
6
3.134
. 79 +
Hometown Location
2
2.218
. 39 +
Hometown Economic Base
8
6.867
. 61 +
Father's Occupation
2
7.727
. 02 + *
Mother's Occupation
2
1.312
.52
Parents' Religion
8
7.291
. 51+
Parents' Income
2
15.864
.0004***
Source of Present Support
2
4.412
.11
Amount of Debt
2
1.780
. 56 +
Indebtedness
2
3.279
. 21 +
Public or Private High School
2
3.432
. 20 +
Public or Private College
2
5.080
.08
General Health
2
3.281
. 24 +
Family's Medical Care Provider
Family Practitioner/General
2
2.257
.32
Internist
2
4.331
.11 +
Pediatrician
2
1.678
.43
Other
2
.825
. 70 +
Work Experiences
Medical
2
3.531
. 19 +
Technical (not medical)
2
.756
.69
Helping Activities
2
.705
. 70 +
Public (sales, etc.)
2
3.309
.22
Teaching
2
1.957
. 37 +
Unskilled
2
.873
.65
Other
2
2.068
. 37 +
Level of Education
2
1.788
. 52 +
College Major
2
4.013
. 16 +
*£ < .05
**£ < .01
***£ < .001
+Agresti-Wackerly
Significance Level

-84-
significant at the .0004 level. It should be noted that the
association between father's occupation and specialty
preference attained a probability of .02.
Father's occupation. Student responses to the item,
indicate your parents' occupations while you were growing
up, were used to form two occupation groups for father's
occupation: (a) professional/technical and (b) business/
other. The professional/technical occupational group
included the students' responses to the categories of
physician, other health care occupation, and other
professional work. The group labeled as business/other was
composed of responses to occupational categories of
business, clerical/sales, skilled worker, farmer/farm
worker, and other. The two occupational groups were
compared with the three medical specialty preference levels:
(a) primary care, (b) subspecialty, and (c) undecided. The
chi-square statistic, X2 (d_f 2, N = 72) = 7.727, £ = .02,
indicated that the probability of the observed association
between medical specialty preference groups and fathers'
occupations being a function of chance was 2 in 100 or less.
Table 5 shows that for students whose fathers were in
professional/technical occupations when they were growing
up, 33% preferred primary care specialties, 26% indicated
subspecialty preferences, while 41% were undecided as to
their eventual specialty choice. Of the third-year students
whose fathers engaged in business and other types of work,
64% preferred primary care, 21% indicated subspecialty

-85-
Table 5
Association between Medical Specialty Preferences and
Father's Occupation and Parents' Annual Income
Specialty Preferences
Sociodemographic Primary Care Subspecialty Undecided
Characteristics n Row % n Row % n Row %
Father's Occupation
professional/
technical
13
33.33
10
25.64
16
41.03
business/other
21
63.64
7
21.21
5
15.15
Parents' Income
> $50,000
9
20.00
25
55.00
11
25.00
< $50,000
12
62.22
2
11.11
6
26.67

-86-
practice, and only 15% indicated that they had not made a
decision. Thus, the greatest differences were in the
primary care and undecided medical preference levels for the
two groups. Almost twice as many students whose fathers
were in business/other indicated primary care preferences as
did those from the professional/technical group. A much
larger percentage of those from the professional/technical
group indicated a delay in decisions on specialty choice.
The importance of this large percentage of undecided
students from the professional/technical background can be
seen in the results from testing Hypothesis 2 when the
actual specialty choice was made in the fourth year (Table
6). While the percent choosing residencies in subspecialty
areas increased, the most dramatic gain was made in the
professional/technical group where the percentage of
students selecting primary care went from 33.3% to 62.5%.
Parents' annual income. The findings related to
parents' annual income and third-year specialty preferences
were consistent with the data on fathers' occupations. The
parents' annual income question asked the students to choose
from among four income categories: (a) less than $10,000,
(b) $10,000 to $25,000, (c) $25,000 to $50,000, or (d) above
$50,000. Two income groups were formed from these four
categories for comparison with the three specialty
preference levels. The two income groups were divided
between responses above and below $50,000. The chi-square

-87-
Table 6
Association between Medical Specialty Choice of
Fourth-Year Students and Father's Occupation
Specialty
Choices
Father's Occupation
Primary Care
n Row %
Subspecialty
n Row %
Professional/technical
35
62.50
21
37.50
Business/other
36
70.59
15
29.41

-88-
statistic, X^ (d_f 2, N = 65) = 15.864, £ = .0004, led to the
conclusion that there was a difference in preference for
medical specialties of these third-year medical students
based on parents' income.
Of the students from the more affluent families (i.e.,
annual incomes greater than $50,000), 20% preferred primary
care, 55% indicated that they wanted to pursue subspecialty
practice, and 25% were undecided. Those students whose
parents' annual income was equal to or less than $50,000
chose primary care more often than any other of the
preference levels. Sixty-two percent chose primary care,
11% preferred subspecialties, and 27% were undecided. The
percentage preferring primary care was three times greater
among the lower income students. A wider difference between
the income groups occurred in their preference for
subspecialties. The percent of undecided students was
fairly consistent for both groups. These data are presented
in Table 5.
Hypothesis 2
There is no association between sociodemographic
characteristics of medical students and their medical
specialty choices as evidenced by fourth-year residency
selection. The procedure for analyzing this hypothesis was
the chi-square test of significance. The results of the
analysis for Hypothesis 2 are listed in Table 7.

-89-
Table 7
Association between Fourth-Year Specialty Choice and
Sociodemographic Characteristics
Sociodemographic
Characteristics
df
X2
Value
Significance
Sex
1
.377
.54
Marital Status
1
1.506
.22
Religious Orientation
4
1.640
.80
Religious Importance
2
1.219
.54
Race
1

.54 +
Citizenship
1

. 26 +
Hometown Location
1
.128
.72
Hometown Size
3
2.228
.52
Hometown Economic Base
4
6.326
.18#
Fathers Occupation
1
.782
.38
Mother's Occupation
1
2.667
*10#
.57#
Parents' Religion
8
3.046
Parents' Income
1
4.602
.03*
Source of Present Support
1
.090
.76
Indebtedness
1
.711
.40
Amount of Debt
1

. 67 +
General health
1

1.00 +
Public or Private High School
1
1.453
.23
Public or Private College
1
.153
.70
Family's Medical Care Provider
Family Practitioner/General
1
1.212
.27
Internist
1
.266
.61
Pediatrician
1
3.142
.08
Other
1
.000
1.00
Work Experiences
Medical
1
.000
1.00
Helping Activities
1
.000
1.00
Technical
1
.019
.89
Teaching
1
.026
.87
Public
1
.928
.34
Unskilled
1
3.157
.08
Other
1
.559
.45
College Major
1
7.338
.01**
Level of Education
1

. 30 +
*£ < .05
**£ < .01
+Fisher's Exact Test
#
Agresti-Wackerly Significance Level

-90-
In the comparisons of fourth-year medical students'
specialty choices with their sociodemographic
characteristics, the overall significance level of _< .01 was
not achieved. It was concluded that there was insufficient
evidence to reject the null hypothesis.
One sociodemographic variable, college major, was sig
nificantly related to the medical students' specialty choice
at the £ < .01 level of significance. A second comparison
using the sociodemographic variable, parents' annual income
reached a significance level of .03. For these two
variables, a discussion of the results is as follows.
Parents' annual income. The parents' annual income
category asked the students to indicate their parents'
combined income by checking one of four income categories.
As in Hypothesis 1, two income groups were formed from the
four categories for comparison with the two medical
specialty choice levels, primary care and subspecialty
practice. The two income groups were (a) greater than
$50,000 annual income and (b) annual income equal to or less
9
than $50,000. The resulting chi-square statistic was X (df
1, N = 99) = 4.602, £ = .03. The percentage of higher level
income students choosing primary care and subspecialty
residencies was nearly equal with approximately 52% choosing
primary care and 48% going with subspecialty residencies. A
considerably higher percentage (74%) of students from the
$50,000 or less income group chose primary care than chose
residencies in subspecialty areas (26%). The results are
presented in Table 8.

-91-
Table 8
Students and Parents'
Annual
Income and
College Maj
or
Specialty
Choices
Sociodemographic
Characteristics
Primary Care
n Row %
Subspecialty
n Row %
Parents' Annual Income
> $50,000
35
51.61
33
48.39
< $50,000
23
73.53
8
26.47
College Major
Science
65
72.22
25
27.78
Other
5
35.71
9
64.29

-92-
College major. The questionnaire provided space for
the students to write their major subject while in
undergraduate college. As shown in Table 8, the majority
of the students indicated science as a major subject. All
nonscience majors were included in the category labeled
other. Of the science majors, 72% chose primary care
residencies. The highest percentage, 64%, of the nonscience
majors, went into subspecialty residencies. The chi-square,
(d_f 1, N = 104) = 7.338, £ = .0067, led to the conclusion
that when college major was tested for independence with
specialty choice, there was a significant difference between
the science majors and nonscience majors with respect to
specialty choice in this group of students.
Hypothesis 3
There is no difference between medical students grouped
by sociodemographic characteristics according to their
attitudes toward primary care measured before or after a
third-year family medicine clerkship. Responses to 29
attitudinal statements were collected both before and after
a clerkship in family medicine. For purposes of analysis
the Likert scale rankings for each attitude statement were
ordered so that higher ranks indicated a favorable attitude
toward primary care and family medicine. To assess the
overall attitudes for individual students, the Likert scale
ranks were summed. The Kruskal-Wal 1 is one-way analysis of
variance was used to test this hypothesis at a significance

-93-
level of £ .01. Results of the sociodemographic groupings of
medical students for attitudes toward primary care in the
preclerkship sampling are summarized in Table 9.
Postclerkship results are given in Table 10.
The overall significance level for Hypothesis 3 was
established at £ .01. Since all of the tests did not reach
this significance level, the null hypothesis could not be
rejected. Those individual variables which did reach
the .01 level of significance or closely approximated it are
discussed.
Religious importance. The students were asked to
indicate "how much importance does religion have in you
life?" The levels of response--very much, some, and
little/none at allwere used to identify three categories
of students. When the three categories of students were
compared for differences in attitudes toward primary care,
the preclerkship data yielded an H value of 5.84 (elf 2, N =
102), £ = .05 (Table 9). The postclerkship test statistic
was H (df 2, N = 85) = 6.63, £ = .04 (Table 10).
Citizenship. When students from this country and
foreign students were ranked according to their
postclerkship attitude scores, the result was an H value of
4.63 with one degree of freedom and a significance level
of .03. This finding should be viewed with caution,
however, due to the small number of students in the
noncitizenship group.
Hometown location. The respondents were asked to
indicate the location of their hometown community in the

-94-
Table 9
Sociodemographic Groupings of
Medical
Students
and
Preclerkship Attitudes
Sociodemographic
H Significance
Characteristics
df
Value
Sex
1
2.02
.16
Marital Status
1
1.47
.22
Religious Orientation
4
.98
.91
Religious Importance
2
5.84
.05*
Race
1
.49
.49
Citizenship
1
1.52
.22
Hometown Location
1
4.45
.03*
Hometown Size
3
.43
.93
Hometown Economic Base
4
3.89
.42
Father's Occupation
1
.00
.98
Mother's Occupation
1
.37
.54
Parents' Religion
4
.29
.99
Parents' Income
1
1.72
.19
Source of Present Support
1
2.42
.12
Indebtedness
1
.76
.38
Amount of Debt
1
.36
.55
General health
1
.40
.53
Public or Private High School
1
.59
.44
Public or Private College
Family's Medical Care Provider
1
.01
.92
Family Practitioner/General
1
.70
.40
Internist
1
.68
.41
Pediatrician
1
1.44
.23
Other
1
.43
.51
Work Experiences
Medical
1
.44
.51
Helping Activities
1
.84
.36
Technical
1
.22
.64
Teaching
1
3.89
.05*
Public
1
7.52
.006**
Unskilled
1
.14
.71
Other
1
. 28
.60
College Major
1
.18
.68
Level of Education
1
2.17
.14
*£ < .05
**£ < .01

-95-
Table 10
Sociodemographic Groupings of
Medical
Students
and
Postclerkship Attitudes
Sociodemographic
H
Significance
Characteristics
df
Value
Sex
1
1.14
.28
Marital Status
1
2.62
.11
Religious Orientation
4
5.19
.27
Religious Importance
2
6.63
.04*
Race
1
.16
.69
Citizenship
1
4.63
.03*
Hometown Location
1
4.02
.05*
Hometown Size
3
3.84
.28
Hometown Economic Base
4
2.42
.66
Father's Occupation
1
4.06
.04*
Mother's Occupation
1
3.29
.07
Parents' Religion
4
6.52
.16
Parents' Income
1
4.19
.04*
Source of Present Support
1
8.27
.004**
Indebtedness
1
.13
.72
Amount of Debt
1
3.01
.08
General health
1
.16
.69
Public or Private High School
1
.00
1.00
Public or Private College
1
.83
.36
Family's Medical Care Provider
Family Practitioner/General
1
3.73
.05*
Internist
1
1.41
.23
Pediatrician
1
1.44
.23
Other
1
.26
.61
Work Experiences
Medical
1
1.10
.29
Helping Activities
1
.00
.96
Technical
1
.08
.77
Teaching
1
9.19
.002**
Public
1
2.09
.15
Unskilled
1
.43
.51
Other
1
1.27
.26
College Major
1
.35
.55
Level of Education
1
.97
.32
*£ < .05
**£ < .01

-96-
precollege years. The greatest number of students'
hometowns were in the same geographic area as the medical
school, south/southeast. When students were categorized
into groups according to whether their hometown was
south/southeast or another location, differences between the
two samples were found with respect to attitudes toward
primary care. The preclerkship data yielded an H value of
4.45 (d_f 1, N = 103), £ = .03 (Table 9). The postclerkship
test statistic was H (d^f 1, N = 86) = 4.02, £ = .05 (Table
10). In both pre- and postclerkship data analysis the
students from the south/southeast had lower mean rank scores
indicating lower attitude scores.
Father's occupation. Father's occupation while growing
up, classified as professional/technical or business/other,
was the basis for identification of two samples of students
which were examined for differences in attitudes toward
primary care. Postclerkship, the observed H statistic for
the two samples was 4.06 with 1 degree of freedom and
significance level of .04. The mean rank of 48.9 was higher
for the business/other group than the professional/technical
group's score of 38.1 indicating a higher attitude score for
the business/other student group.
Source of present support. Responses pertaining to the
major contributor to the students' present support were
categorized as (a) parents and (b) other. When students
from these two groups were analyzed for differences or
similarities in their attitudes toward primary care in the

-97-
postclerkship period, the observed test statistic was H =
8.27 (d_f 1, N = 86), £ = .004. The mean rank score for the
group supported by parents was 31.06 and 48.31 for the other
group. Thus, it was concluded that students whose parents
were their major source of support had significantly lower
attitudes toward primary care than did students who derived
their major support from other sources.
Parents' income. Closely related to the source of
support was the parents' combined annual income item. Two
income categories were used in the comparison analysis with
attitudes. The income categories were (a) students whose
parents had annual earnings of $50,000 or less and (b) those
whose parents earned over $50,000. The preclerkship data
produced a significance level of .19. However, the
postclerkship test statistic was H (df 1, N = 79) = 4.19,
£ = .04. The mean rank score of the $50,000 or less group
was 43.38 while those from families with an excess of
$50,000 annual income had a mean rank score of 31.76. From
these data it was concluded that students from the higher
income group had less positive attitudes toward primary care
than those from the $50,000 or less income group.
Family medical care provider. In response to the
question Who provided primary care for your family? students
had the choice of (a) family/general practitioner, (b)
pediatrician, (c) internist, and (d) other. No differences
in preclerkship attitude scores were noted among the groups.
Postclerkship attitude scores were different (£ = .05) for

-98-
students whose medical care was provided by a family
practitioner/generalist from those students who did not
check this alternative answer. The Kruskal-Wal 1 is test
statistic was H (d_f 1, N = 86) = 3.73. Based on the mean
rank score, students who were cared for by family/general
practitioners held more positive attitudes toward primary
care.
Work experience. From a list of job descriptions, the
students were asked to check the areas in which they had had
work experience. Preclerkship attitude scores of students
who had dealt with the public either in sales or public
relations were different from the ones who did not engage in
this type of work at a significance level of .006. The test
statistic was H (df 1, N = 103) = 7.52.
The attitude scores of students who had work experience
as teachers were significantly different from those who had
not participated in activities in both the preclerkship
samples and the postclerkship measurements. Preclerkship,
the observed H of 3.89 (d£ 1, N = 103), was significant at
the £ = .05. Postclerkship, the test statistic of 9.19 (df
1, N = 86), had a probability of .002. The mean rank scores
for those with teaching experience was higher in both
preclerkship and postclerkship measurements indicating
higher attitude scores.

-99-
Hypothesis 4
There is no difference between groups of medical
students identified by medical specialty preference in the
third year or medical specialty choice in the fourth year
with respect to their rankings of career goals. Just prior
to a family medicine clerkship and again, after the
clerkship, students ranked 2 sets of career goal phrases.
Each set contained six goal phrases which might be achieved
through a career in medicine. The students were asked to
rank the importance of the goals by putting a "1" next to
the most important goal, a "2" next to the second most
important goal through "6" for the least important. The
Kruska1-Wal1 is one-way analysis of variance by rank was used
to test the null hypothesis. Students' rankings of each
career phase were tested using a significance level of
< .01.
The three medical specialty preference groups, primary
care, subspecialty, and undecided, were not differentiated
by the rankings they gave to the goal statements before the
family medicine clerkship. The results of the preclerkship
goal rankings and specialty preference groups are found in
Table 11.
When the preclerkship career goal rankings were
analyzed with the fourth-year specialty choice groups, the
primary care specialty group and the subspecialty group of

-100-
Table 11
Results of Preclerkship Career Goal Rankings and
Three Preference Groups
Career Goals4" N = 72
H £
Set 1
a. helping to solve community's medical
and social problems
7.77
.02
b. gaining respect and friendship of
other physicians
2.36
.30
c. developing long-term relationships
with patients
2.85
.24
d. contributing to medical knowledge
through research
1.17
.56
e. assuring financial security for
self and family
1.06
.59
f. allowing me to live in preferred
area and style
3.15
.21
Set 2
a. providing me with capital for
investment
1.50
.47
b. being acknowledged by other
physicians as a leader in health
care for my patients
1.06
.59
c. having time for family and non
professional activities
.53
.77
d. gaining trust and confidence of
my patients
4.55
.10
e. making me a respected member of
my community
.28
.87
f. actively conducting research in my
areas of interest
1.77
.41
+Paraphrased
from original questionnaire

-101-
students were not significantly different in their goal
orientation (Table 12).
After the family medicine clerkship, students ranked
the career goal phrases in the same manner as they had been
asked to do prior to the clerkship some 2 months previous.
These postclerkship rankings were compared for the two
fourth-year specialty choice groups, primary care and
subspecialty. The two groups were not distinguishable based
on their goal rankings (Table 12).
Based on the analysis of the data, it was concluded
that the medical specialty preference groups and the fourth-
year specialty choice groups were not dissimilar in the
relative importance they assigned to the career goal
phrases. The null hypothesis could not be rejected.
Hypothesis 5
There is no relationship between medical students'
rankings of career goals and their attitudes toward primary
care before or after a third-year family medicine clerkship.
This null hypothesis used the data in the respondent-
reported section of the questionnaire entitled Career Goals
and the section labeled Attitudes about Primary Care and
Family Medicine. Respondents were surveyed before and after
a third-year clerkship in family medicine. Students rank-
ordered two sets of six career goal phrases on a scale from
1 to 6 with 1 being the most important goal you wish to

Table 12
Results of Preclerkship and Postclerkship Career Goal Rankings and Two Specialty
Choice Groups
Preliminary Goal
Rankings
Follow-Up Goal
Rankings
Career Goals"1"
H
£
H
£
Set
a.
1
helping to solve community's medical and
social problems
.11
.74
1.14
.29
b.
gaining respect and friendship of other
physicians
.42
.52
3.38
.07
c.
developing long-term relationships with
patients
.05
.82
.15
.70
d.
contributing to medical knowledge through
research
1.58
.21
.05
.82
e.
assuring financial security for self and
family
.06
.84
1.19
.28
f.
allowing me to live in preferred area
and style
.09
.77
.81
.37
-102-

Table 12
Continued
Preliminary Goal
Rankings
Follow-Up Goal
Rankings
Career Goals4"
H
£
H
£
Set
2
a.
providing me with capital for investment
.23
.63
.01
.93
b.
being acknowledged by other physicians as
a leader in health care for my patients
.80
.37
.53
.47
c.
having time for family and non-professional
activities
2.30
.13
.11
.74
d.
gaining trust and confidence of my
patients
.17
.68
.06
.81
e.
making me a respected member of my
community
1.33
.25
. 26
.61
f.
actively conducting research in my areas
of interest
.54
.46
.09
.77
+Paraphrased from original questionnaire
-103-

-104-
achieve through a career in medicine and 6 being the least
important goal. A summative score for each student was
derived from their responses on the 29 attitude statements.
A high numerical score indicated a favorable attitude toward
primary care and family medicine. The statistical index
used for finding the relationship between the data sets was
the Spearman's rho correlation coefficient. The overall
significant level was set at £ .01. Although some variables
reached significance, the overall significance level
established prior to the analysis was not achieved.
Evidence was considered insufficient to reject Hypothesis 5.
The results of this analysis are found in Table 13.
Significant correlations occurred in two of the items
included in the first set of career goal statements and
attitudes expressed before and after the clerkship. Two
correlations reached significance among the second set of
six career goal rankings and attitude scores. The direction
as well as the degree of the relationship was considered.
There was an inverse relationship between the rankings of
the career goal statements and the attitude scores.
Therefore, when the variables were highly related in a
negative direction, the career goal was very important to
students with a highly favorable attitude toward primary
care and family medicine. Conversely, when the direction of
the correlation was positive, the career goal was ranked low
in importance to those with highly favorable attitudes and
more important by those students who have lower attitude

Table 13
Correlation Coefficients for Career Goal Rankings and Attitude Scores
Preliminary Goal Follow-Up Goal
Rankings Rankings
Preliminary
Follow-Up
Follow-Up
Career Goals
Attitudes
Attitudes
Attitudes
Set
1
a.
helping lead my community to the
solution of social and medical problems
-.11
r1

1
0.22*
b.
gaining respect and friendship of other
physicians
.04
.08
.003
c.
developing long-term relationships with
patients
-.27**
-. 15
-.17
d.
contributing to medical knowledge through
research
.07
i

o
-. 16
e.
assuring financial security for self and
family
.03
.04
.22*
f.
allowing me to live in preferred area
and style
. 28**
.29**
.31**
-SOI-

Table 13
Continued
Preliminary Goal
Rankings
Follow-Up Goal
Rankings
Career Goals+
Preliminary
Attitudes
Follow-Up
Attitudes
Follow-Up
Attitudes
Set
2
a.
providing me with capital for investment
.19
. 10
.34***
b.
being acknowledged by other physicians as
a leader in health care for my patients
.003
.10
i

o
kD
c.
having time for family and non-professional
activities
-.05
-.06
.06
d.
gaining trust and confidence of my
patients
-.14
-.13
-.13
e.
making me a respected member of my
community
.05
.007
.09
f.
actively conducting research in my areas
of interest
-.05
-.06
-.31**
*£
**JD
k k p
.05
.01
.001
+Goal
phrases
paraphrased
from original questionnaire
-901-

-107-
scores. For those items which either reached significance
or were < .05, the results were as follows:
In the follow-up data, the career goal "helping lead my
community to the solution of social and medical problems"
was significantly correlated with follow-up attitude scores
(Set 1, Goal a, Table 13). The correlation coefficient
between these two variables was -.22, jd = .04. The
direction of the correlation indicated that students who
ranked this goal as among the more important goals had more
favorable attitudes toward primary care and family
medicine.
When the career goal "developing long-term, intensive
relationships with patients" was correlated with follow-up
attitudes, the result was an r value of -.27, jo = .008 (Set
1, Goal c, Table 13). The direction of the relationship was
interpreted to mean that students with a favorable attitude
toward primary care ranked this goal high in importance.
The follow-up rankings which students assigned to the career
goal "assuring financial security for self and family"
correlated with follow-up attitudes at a significance level
of £ = .04 (Set 1, Goal e, Table 13). The positive
direction of the correlation indicated that as the
importance of this goal increased the attitude scores
decreased.
The career goal "allowing me to live in the type of
area and lifestyle that I prefer" was significantly related
to attitudes before and after the clerkship (Set 1, Goal f,
Table 13). The observed r for the preliminary data was .28,

-108-
£ = .004. The preliminary goal rankings for this goal were
compared with attitude data obtained in follow-up, the r
value was .29, £ = .006. After the clerkship, the
correlation coefficient between career goal rankings and
attitude scores was .31, £ = .004. The sign of the
correlation coefficient was in a positive direction so that
the importance of this goal increased as the attitude
scores decreased. Conversely, students with favorable
attitudes toward primary care ranked this goal lower in
importance.
Students ranked the extent to which they perceived
"providing me with capital for investment" an important
product of their medical career (Set 2, Goal a, Table 13).
The correlation coefficient for the follow-up data was .34,
£ = .001. Based on the positive sign associated with the
correlation coefficient, it can be assumed that those
students who held this goal to be important had a lower
attitude score.
Following the family medicine clerkship, the importance
of "actively conducting research in my areas of interest"
was significantly related to attitude scores, r = -.31,
£ = .003 (Set 2, Goal f, Table 13). The direction of the
correlation suggested that as the importance of this goal
increased so did the positive attitude toward primary
care.

-109-
Hypothesis 6
There is no difference between the groups of medical
students choosing primary care and those choosing
subspecialties in the fourth year with respect to their
rankings of career goals before and after a family medicine
clerkship. Medical students ranked the personal importance
of two sets of six career goal phrases (total 12) prior to a
third-year family medicine clerkship and again, immediately
following the clerkship. The difference in the preclerkship
and postclerkship rankings served as the observations for
testing the independence of the two specialty choice groups,
primary care and subspecialty, which were identified from
fourth-year residency choices. The statistical procedure
used for the analysis was the Kruskal-Wallis one-way
analysis of variance. The overall significance level was
set at < .01.
This change in career goal rankings from preclerkship
to postclerkship did not distinguish the two fourth-year
specialty choice groups at the _< .01 significance level.
Hypothesis 6 was not rejected. Results by goal statement
are shown in Table 14.
Hypothesis 7
There is no association between medical students'
specialty preferences expressed in the third year and their
career choice decisions evidenced by fourth-year residency
selection. The last item contained in the questionnaire

-110-
Table 14
Comparison of Specialty Choice Groups Based on
Differences in Preliminary and Follow-up Career
Goal Rankings
Career Goals+
H
£
Set 1
a. helping to solve community's medical
and social problems
1.23
.27
b. gaining respect and friendship of
other physicians
4.38
.04
c. developing long-term relationships
with patients
.06
.81
d. contributing to medical knowledge
through research
1.44
.23
e. assuring financial security for
self and family
.00
1.00
f. allowing me to live in preferred
area and style
.80
.37
Set 2
a. providing me with capital for
investment
.02
.89
b. being acknowledged by other
physicians as a leader in health
care for my patients
.08
.78
c. having time for family and non
professional activities
.63
.43
d. gaining trust and confidence of
my patients
.55
.46
e. making me a respected member of
my community
3.59
.06
f. actively conducting research in my
areas of interest
.82
.36
+Paraphrased from original questionnaire

-Ill-
administered in the third year asked the medical students to
indicate to the best of their knowledge their future plans
for residency training. Their responses were categorized
into three specialty preference groups: (a) primary care
which included family medicine, pediatrics, and general
internal medicine and other primary care; (b) other
specialty/subspecialty; and (c) undecided. The fourth-year
specialty choices were derived from the residency match
information. These two specialty choice groups were
identified as primary care and subspecialty.
Eighty-five percent of the students who preferred
primary care in the third year selected residencies in
primary care while the remaining 15% chose subspecialties
instead. Of those initially indicating an interest in
subspecialty areas of medicine, 76% chose a subspecialty
residency. Fifty-seven percent of those undecided in the
third year ultimately selected primary care residencies.
Using Goodman and Kruskal's lambda statistic as a measure of
association, there was a 33% improvement in ability to
predict residency choices given that specialty preferences
in the third year was known. The results of the data
analysis are demonstrated in Table 15.
The chi-square test statistic produced by the analysis
of the data was X2 = 18.810 (df 2, N = 72) = £ .001. This
led to the rejection of the null hypothesis and the
conclusion that residency choices for this sample were
related to third-year preferences.

-112-
Table 15
Association between Third-Year Specialty Preferences
and Fourth-Year Specialty Choices
Fourth-Year Specialty Choices
Third-Year Specialty
Preferences
Primary Care
n Row %
Subspecialty
n Row %
Primary Care
29
85.29
5
14.71
Subspecialty
4
23.53
13
76.47
Undecided
12
57.14
9
42.86

-113-
Hypothesis 8
Hypothesis 8 examined whether or not a clerkship in
family medicine modified medical students' attitudes toward
primary care. Attitudes were measured before and after an
8-week family medicine clerkship in the third year of
medical school. Using a 5-point Likert Scale, the medical
students responded to 29 statements of attitudes and beliefs
about the practice of medicine and medical care. After each
administration of the questionnaire, an attitude score was
obtained for individual students. In order to test the
hypothesis, a difference score was calculated for each
subject by subtracting the preclerkship attitude score from
the postclerkship score. The Wilcoxon matched-pairs signed-
ranks test was selected to test the significance in the
difference scores between the two related samples. The
significance level was set at the _< .01 level. A Z score of
-2.008 resulted from the analysis. The probability
associated with the occurrence under a null hypothesis of a
Z as extreme as -2.008 has a probability of .0444. In as
much as this probability was greater than the established £
< .01, the decision was made not to reject the null
hypothesis and to conclude that a family medicine clerkship
does not affect attitudes toward primary care at the £ .01
level of significance.

-114-
Hypothesis 9
The objective for Hypothesis 9 was to test whether a
significant difference existed with respect to attitudes
toward primary care between primary care-oriented medical
students, subspecialty-oriented students, and students who
were undecided. Students were separated into the three
groups on the basis of their stated preferences for a
specialty at the initial administration of the
questionnaire. Information on attitudes toward primary care
was obtained from the 29 attitude statements in the last
section of the questionnaire. The Kruskal-Wal1is one-way
analysis of variance test was used to determine whether the
three groups came from the same sample with regard to their
attitudes. The significance level for the test was set at
£ .01. The resulting statistic was H = 4.57 (d_f 2, N =
70), £ = .10. Since the magnitude of the test probability
was larger than the previously set level of significance, it
was decided that the three groups came from the same or
identical samples with respect to attitudes and, therefore,
the null hypothesis could not be rejected.
Hypothesis 10
Hypothesis 10, there is no difference between medical
students choosing primary care residencies and those
choosing subspecialties with regard to attitudes toward
primary care before and after a family medicine clerkship,
was tested using the Kruskal-Wallis one-way analysis of

-115-
variance procedure and a £ .01 level of significance. The
subjects were distributed into two groups according to their
residency match choices of specialties. Data on their
attitudes were obtained from the 29 questions addressing
attitudes toward primary care and family medicine measured
before and after the family medicine clerkship. To test the
hypothesis, the change in each student's attitude score was
computed. The mean attitude score of 103.09 (SJO = 7.17) for
students choosing primary care specialties in the initial
measurement was very close to their mean score at follow-up
(M = 103.22, SD = 8.18). Those students taking subspecialty
residencies scored a mean of 100.36 (SD = 8.51) in the
initial sampling of attitudes toward primary care, while at
follow-up their mean score increased to 103.11 (SD = 10.69).
The comparisons of the two groups with respect to their
attitude change scores yielded an H value statistic of 5.7
with 1 degree of freedom and a probability of .02. The mean
rank score for primary care medical students was 37.47 and
51.25 for students choosing subspecialty residencies. Thus,
the subspecialty-oriented students' attitude scores improved
more from preclerkship to postclerkship. However, it was
determined that the two groups, primary care and
subspecialty students, were not different with respect to
their attitudes before and after the clerkship experience at
the < .01 level of significance. Therefore, the null
hypothesis was not rejected for Hypothesis 10.

CHAPTER V
SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS
Summary
The nation is faced with an oversupply of physicians
and an imbalance of physicians across specialty areas. A
number of undesirable consequences result from this
oversupply and maldistribution among doctors, the chief one
being spiraling health care cost. Although a large body of
research has been generated, unequivocal evidence regarding
the causes for the trend away from general or primary care
toward subspecialization has been elusive. Interventions
introduced at the residency level have met with only limited
success in producing the proper mix of primary care and
nonprimary care physicians for the needs of society.
Because the initial specialty choice decisions are made
during medical school, the study of variables at this level
has the advantage of providing information with the greatest
potential for influencing the decision process.
The purpose of the study was to identify factors
related to medical students' third-year specialty
preferences and fourth-year specialty choices including
their sociodemographic characteristics, the relative
importance placed on career goals, and their attitudes
toward primary care. In addition, associations among
-116-

-117-
the variables, sociodemographic characteristics, career
goals, and attitudes toward primary care were explored.
Another intention of the study was to examine the extent to
which a clerkship in family medicine affected the students'
attitudes toward primary care. Included in the aims of the
study was the investigation of the association between
third-year career preferences and fourth-year specialty
choice.
The sampling frame for the study included all medical
students in the 1985 graduating class of the College of
Medicine, University of Florida. The class contained 117
students of which 109 participated. The students were
sampled in their third year of medical school because they
were far enough along in the educational process to have
begun exploration of self and specialties, the family
medicine clerkship occurs in the third year, and the fourth-
year students are geographically scattered.
A multidimensional questionnaire was administered to
third-year medical students immediately before and after a
family medicine clerkship. The follow-up questionnaire was
identical to the entering form with the exception of
exclusion of the sociodemographic data section and the
question related to specialty preferences which appeared on
the initial questionnaire. On the initial survey form the
students indicated whether they preferred a primary care
specialty, a subspecialty area, or were undecided. The

-118-
results of the fourth-year residency match served as
evidence of the student's area of specialty choice.
The approach to examining the data was through
nonparametric procedures appropriate for the research model
and the nominal- and ordinal-levels of measurement. The
statistical tests included chi-square, Spearman's
correlation coefficient, Kruskal-Wallis one-way analysis of
variance, and Wilcoxon matched-pairs signed-ranks. Ten null
hypotheses were tested.
Overall, the sociodemographic characteristics of the
medical students in this sample were poorly associated with
third-year specialty preferences and fourth-year specialty
choices. However, some of the background variables did
achieve statistical significance. For example, parents'
annual income was related to third-year preferences
(£ = .0004). Students whose parents earned an annual income
of $50,000 or less were considerably more likely to express
a preference for the primary care specialties in the third-
year and to maintain that predilection into the fourth-year
residency choice than were students from the more affluent
familiesincome greater than $50,000.
Another measure of socioeconomic status, father's
occupation, was associated with third-year specialty
preferences (jd = .02). Students whose fathers engaged in
business or other nonprofessional types of employment were
more likely to favor a primary care specialty than those
with professional/ technical fathers. A much larger percent

-119-
of students from the latter occupation group indicated a
delay in specialty choice decisions in the third year of
medical school.
There was a significant difference between college
science majors and nonscience majors with respect to
specialty choice (£ = .0067). This association should be
viewed with caution, however, as such a high percent of the
students in this sample were science majors.
Scores on attitudes toward primary care differentiated
several of the sociodemographic groupings. Students
identified as having had work experience with the public
were distinguished by their preclerkship attitude scores
(£ = .006). Postclerkship (£ = .002) attitudes
differentiated students who had work experience involving
teaching. There were also differences in attitudes toward
primary care among those whose major source of present
support was their parents versus other financial resources
(£ = .004). Interpreting the mean rank score, students
whose major financial support did not come from their
parents, those who had work experience in teaching and in
work with the public had more favorable attitudes toward
primary care.
The relative importance students placed on career goal
statements did not distinguish between specialty preference
groups in the third year or fourth-year specialty choice
groups. Neither did the difference in goal rankings from

-120-
preclerkship to postclerkship successfully discriminate
between the two specialty groups.
Correlational relationships were found between
students' rankings of several career goal statements and
their scores on attitudes toward primary care. In
preclerkship measurements, students who had high scores on
attitudes toward primary care considered "developing long
term relationships with patients to be an important career
goal." The goal "allowing me to live in the preferred area
and life style" was significantly correlated with
preclerkship and postclerkship attitudes. The direction of
the correlation was interpreted as those students who scored
low on attitudes toward primary care considered this an
important goal. Postclerkship ranking of the goal
"providing me with capital for investment" also correlated
positively with low postclerkship attitude scores. A
negative correlation resulted from the postclerkship
rankings of the goal "actively conducting research in my
areas of interest" and attitude scores which indicated that
students who scored high on attitude scores considered this
an important career goal.
A statistically significant difference in students'
responses to the attitude items could not be demonstrated
from the pre- and postclerkship data (£ = .04). Nor did the
attitude measurements prove to be helpful in discriminating
career preference (£ = .10) and specialty choice groups
(£ = .02) .

-121-
In summary, evidence generated from the analysis of
data from this study was insufficient to reject all but one
of the 10 hypotheses. For the one null hypothesis
rejected, a significant association (£ < .01) was found
between medical students' third-year specialty preference
and their fourth-year residency selection (£ = .001).
Eighty-five percent of students who preferred primary care
specialties in the third year actually chose primary care
residencies in the fourth year. Similarly, students
indicating a preference for subspecialty practice maintained
their position in the fourth-year residency choice. Almost
one-third of the students were undecided as to their
specialty preferences when sampled in the third year,
however. The implications of these results are that the
majority of students are capable of making stable career
decisions at least as early as the third year. Those
medical educators and health policy makers interested in
impacting on the specialty decision process should consider
early interventions.
Conclusions
This investigation was initiated to identify factors
related to medical students' third-year specialty preference
and fourth-year specialty choice as well as the effects of a
family medicine clerkship experience on career goals and
attitudes toward primary care. The aim of the research was
to contribute to the base of knowledge in the area of

-122-
intraprofessional career decisions in medicine in an effort
to provide some direction for medical educators and policy
makers.
Consistent with previous research (Anderson, 1975; U.S.
DHHS, 1980), this study demonstrated the capriciousness of
sociodemographic characteristics in association with
specialty choices. An explanation for this may be the
selection process, medical students are a highly selected
group and, therefore, homogeneous in background
characteristics. In addition, the timing of this study was
such that students had been in the medical school
environment for two to three years which could have served
to further dilute the impact of the sociodemographic
differences on specialty preferences and choices.
Two measures of socioeconomic status, parents' annual
income and father's occupation, were associated with
specialty selection for the study group. Students whose
parents earned in excess of $50,000 annually and students
whose father's occupation was classified as
professional/technical were much more likely to choose
subspecialty practice than primary care. A large percent of
students who were undecided as to their specialty
preferences in the third year had fathers in the
professional/technical ranks. These findings support
Zuckerman's (1977) contention that economic affluence may
encourage specialization.

-123-
When sociodemographically derived groups of students
were differentiated based on their attitudes toward primary
care, those students who were supported by their parents
were found to hold attitudes dissimilar to students
dependent on other financial resources. The mean rank of
the parent supported group was lower indicating less
positive attitudes toward primary care. This finding is
consistent with the previously discussed measures of
socioeconomic index, parents' annual income and father's
occupation, which were associated with specialty decisions
and lends further support to the premise that economic
security plays a part in specialty decisions.
Students who had work experience in teaching activities
were distinguished from those who had not by their more
positive primary care attitudes. Teachers and others who
work with the public are considered to share some common
characteristics with primary care physicians. Namely, they
generally work with broad segments of society, must
communicate effectively to impart information to others and
are, usually, humanistically oriented. It was concluded
that this finding provided evidence of validity for the
attitude section of the questionnaire.
The relative importance medical students assigned to
career goal statements did not differentiate between third-
year specialty preferences nor fourth-year specialty choice
groups. One explanation for this finding may be the
proportionately large number of students in the undecided

-124-
group of third-year specialty preferences. The undecided
group contained a mix of students who would eventually align
themselves with either subspecialist or primary care.
Assuming that these two groups, primary care and
subspecialty, actually have different values, goals, and
attitudes, the undecided group would have reflected a
combination of these. There is no ready explanation for the
results obtained with the fourth-year specialty choice
groups on this variable. Perhaps the most plausible
explanation for these data is that this sample of medical
students was homogeneous with regard to the career goal
statements used in the questionnaire and what may have
differed was the students approach to attaining the goals.
Significant correlations occurred between several
career goal statements and attitudes toward primary care.
Students who held unfavorable attitudes toward primary care
gave higher rankings to the following statements: (a)
assuring financial security for myself and my family, (b)
allowing me to live in the type of area and lifestyle that I
prefer, and (c) providing me with capital for investment in
business, real estate, the stock market, etc. Students with
favorable attitudes toward primary care gave priority to
(a) helping lead my community to the solution of social and
medical problems, (b) developing long-term intensive
relationship with my patients, and (c) actively conducting
research in my areas of interest. The results followed a

-125-
predictable pattern suggested by reports in the literature
that students attracted to primary care specialties were
patient-care centered rather than interested in financial
security, social position, and professional status (Burkett
& Gelula, 1982; Collins & Roessler, 1975; Leserman, 1978;
Plovnick, 1980).
The discovery of the research interest aligned with
primary care was in conflict with the findings of Burkett
and Gelula (1982) that students attributed greater
importance to helping people than to scientific endeavors.
It could be hypothesized that the primary care grouping
which included family medicine, internal medicine,
pediatrics, and the transitional residency was sufficiently
large to accommodate those with research interests.
Residency selection not being the final career decision
point, students choosing primary care residencies may later
pursue an academic career or subspecialize in a primary care
area both of which include research potential.
Just as the values placed on career goals did not
discriminate specialty groups, neither did attitudes of
medical students gauged prior to the family medicine
clerkship and following the clerkship distinguish the
specialty groups. Changes in attitudes from before to after
the clerkship were analyzed. The conclusion drawn from
analysis of these data was that even though the clerkship
experience could not be shown statistically to modify

-126-
students' attitudes toward primary care, attitude scores
remained high for primary care students and even improved
for the subspecialty group. Thus, the clerkship could be
considered an important component of the training program to
maintain student interest in primary care and to improve the
understanding and attitudes of subspecialty students toward
the practice of primary care.
The significant association between medical students'
third-year preferences and their fourth-year choice of
residency was in contrast to what could be expected given
the exploratory state of the career decision process in the
third year. A high percentage (75%) of students preferring
primary care actually chose primary care residencies.
Similarly, students going into subspecialties were stable
from third to fourth year (85%). The undecided group in the
third-year split 57% to 43% in favor of primary care. Since
this group comprised almost one-third of the total number in
the sample, the direction of their choices was an important
factor in the results. The findings of this analysis
suggest that third-year preferences are fairly stable for
those who indicate a specialty area but the one-third who
are uncommitted are a potential pool through which a
substantial impact could be made on specialty
maldistributions

-127-
Implications of the Study
For those interested in increasing the pool of primary
care physicians, implications which may be drawn from the
findings of this study include the following:
1. Students whose parents' annual income is $50,000 or
less and those who derive their support from their parents
tend to be attracted to primary care specialties. Loans and
scholarships should be provided to encourage and support
students from less affluent families.
2. Medical students tend to be fairly homogeneous in
background characteristics. The admission criteria should
be generalized to gain more diversity in the
sociodemographic characteristics of medical students.
3. Nearly one-third of the students in this study were
undecided as to their specialty preferences in the third
year of medical school. Development of a formal structure
for facilitating career guidance of medical students is
essential so that career decisions are not based on limited
experience and information.
4. Clinical experiences in general primary care
specialties should be emphasized in the curriculum,
particularly early educational training activities in
ambulatory care centers in order to maintain interest in
primary care areas and to improve attitudes and knowledge
about primary care specialties.

-128-
The findings of this study have implications which
extend beyond medical education to general professional
education requiring field experiences. Such experiences
have the potential to maintain and/or improve students'
attitudes and should be planned accordingly.
Recommendations
Several factors lead to the recommendation for
continued, similar investigations. There is very little
research-substantiated data available concerning the
mechanisms through which medical education influences
specialty choice patterns and how the educational
environment and students' background, values, and attitudes
interact. Medical education is called upon to intervene in
a number of areas to meet societal needs and one of these is
the maldistribution across specialties. Educational
policies addressing this area of concern must rest on
sounder ground than is possible given the present
preliminary and relatively immature state of the research to
date.
Sociodemographic Chacteristics
Medical students are a highly selected group. Much of
the variation in their backgrounds is screened out beginning
as early as high school, again at the college level,
and

-129-
certainly by the medical school admissions process. Given
this highly discriminating process, it is still unclear as
to how such a homogeneous group translates into so many
specialties. Background characteristics have for the most
part been inconsistent predictors of specialty choice with
the exception of a few, isolated variables. One such
variable, sex of the students, has proved to be one of the
few constant sociodemographic measures associated with
specialty choice. Historically, women have been more
inclined toward primary care specialties. Even this
variable is circumspect given the present and impending
changes in women's roles in society, in general, and
medicine, in particular.
The numbers of women in this study were relatively
small. It is recommended that in data collection with
subsequent groups, this variable and variables related to
other minority groups be given attention.
The complete picture of cost factors and their
influence on specialty choice is not clear from the
literature. It can be assumed that economically hard-
pressed students are influenced in their choice of specialty
as a result of financial considerations. Not only is income
lost with additional years of graduate education but each
added year of residency may actually increase the
indebtedness of the trainee. The present study demonstrated
an association between parents' annual income and specialty

-130-
preference and choice. The cost of medical education is
increasing dramatically and research on these financial
aspects needs to be elaborated.
The cost of medical training is conceivably linked to
other background factors such as initial wealth, parents'
support, marital status, age, sex, and others. It is
recommended that the interaction of these background
variables be investigated. The continued study of
background characteristics is warranted in view of the
fact that no studies as yet have settled the issues with
respect to these variables and it is possible that with
increasing numbers of women and other previously less
frequent groups the overall picture of the profession may be
clarified.
Attitudes and Goals
It is not apparent from previous research how students'
goals and attitudes relate to their specialty choices or how
they relate to students' background characteristics and
educational experiences. In this study, two of the
sociodemographic characteristics of the medical students
were significantly associated with their attitudes toward
primary care even though the significance level of .01 led
to retention of the null hypothesis. For these subjects,
attitudes did not change sufficiently following an

-131-
experience in primary care to warrant rejection of the null
hypothesis. Even so, the findings are preliminary and
require further investigation. It is recommended that the
study of these variables be continued longitudinally to
obtain a larger sample of responses. This would possibly
produce stable and definitive conclusions. In addition,
factor analysis of the attitude statements from the
questionnaire may be possible with a larger sample size thus
reducing the number of questions and delineating the
questions most related to primary care choices.
Clerkship Experiences
In addition to the continued study of medical students'
individual characteristics, it is recommended that the
influence of educational experiences (i.e., family medicine
clerkships) continue to receive research attention.
Specialty choice is at least in part a time-dependent
process so that the time factor must be taken explicitly
into consideration. The time of exposure to family medicine
in relation to other clinical experiences is presumed to be
an important factor. It was not within the scope of this
study to examine subgroups of students based on their
rotation sequence. It is recommended that the application
of the questionnaire be continued immediately before and
after the clerkship so that the effects of the clerkship

-132-
on attitudes, goals, and specialty choice can be discerned
taking into account the rotation sequence of each
subgroup.
Students draw from role models and clinical experiences
in the career decision process. If these decisions are to
be made from a broad base of information of all specialties,
students need a clear idea of how each is practiced
including how primary care is conducted by practicing family
physicians. In this study, a subset of students spent two
weeks of the third-year family medicine clerkship under the
supervision of family physicians in private practice, living
with these physicians in their homes. The role-playing
opportunities in this experience and the role modeling of
the preceptors is very different from the other educational
activities in medical school. The power of such an
experience to exert an influence on students' attitudes
toward primary care and thus, their career choice deserves
to be explored. It is recommended that in subsequent
replications, the influence of this experience be
evaluated.
While the choice of a residency represents the decision
to specialize, in reality these decisions are prolegomenous
in some cases. For example the majority of students
choosing internal medicine residencies do not stay in
general internal medicine but rather proceed to subspecialty
areas after the first year. Even practicing physicians

-133-
sometimes change their specialties or limit their practice
to a specific area. In this study the groupings according
to specialty preference and choice were broad and may have
included many students who actually will enter subspecialty
training after the first year of residency. Continuing the
study longitudinally would increase the number of subjects
sufficiently to enable more definitive categorization of
groups. It is also recommended that the study of these
students be extended into the residency years and beyond to
obtain data which more closely approximate their actual
specialty practice.
Revision of the Questionnaire
It is recommended that the existing instrument be used
in replications with the following revisions to the
Background section:
1. Revise the item on Citizenship as it is a low-yield
item with less than 3% of the students indicating
citizenship other than U.S. Omit Racial Background
categories of American Indian, Mexican-American, and Puerto
Rican and add Hispanic when using the questionnaire
longitudinally with students at the institution of this
study. Remove the question Do you consider yourself in
excellent health? as 96% of students gave a positive
response. Such low yield questions lengthen the
questionnaire unnecessarily.

-134-
2. Revise the item pertaining to Finances to
accurately reflect current economic conditions. The dollar
amount of the categories of indebtedness are too low to
provide accurate data on students' educational debts.

APPENDIX A
MEDICAL STUDENT ENTERING QUESTIONNAIRE

A LONGITUDINAL STUDY OF FAMILY PRACTICE EDUCATION
SURVEY OF ENTERING STUDENTS
NAME:
SOCIAL SECURITY NUMBER:
STUDENT PROFILE
BACKGROUND
1. Personal Pata
a. Age
b. Sex: Male Female
c. Marital Status: Married Single Other
d. Religion:
(i)Orientation:
Protestant Catholic Jewish
Other (specify) No preference
(ii)How rtuch importance does religion have in your life?
Very ix>ch Same
A little None at all
e.Racial Background:
Black Oriental
American Mexican
Indian American
White
Puerto
Rican
f.Citizenship:
U.S. Canadian Other
specify
2. Environment
a."Hometown* comnunity (pre-college years)
(i) Location
Midwest Northeast West Coast
Rocky Mtn/ South/ Foreign
Southwest Southeast Country
(ii) Size (pooulation)
_ Less than 2,500 25,000 to 99,999
2,500 to 24,999 100,000 to 500.COO
__ Above 500,000
(iii)Economic Base
fanning mining business
heavy light mixed
industry industry economy
b.Family
(1) Size
Number of children (including yourself)
If more than 1, give your position, e.g., third of five
children
(ii) Indicate your parents* occupations while you were
growing up.
Father Mother
Physician
Other health care
occupation
Other professional work
Business
Clerical/Sales
Skilled worker
Unskilled worker
Farmer/Farm worker
Housewife
Other (specify)
(ill) Parents' religion write "P* and "T for
"Father" and "Mother" on appropriate lines.
Protestant Catholic Jewish
Other (specify) No preference
c. Finances
(1) Source of present support (major contributor):
Self Spouse & Self Parents
Spouse 50* 5<*> Other
(ii) Indeotedness (greather than $500)
res No
If "Tes", indicate by checking the appropriate
catesory- greater
Less than SI 000 51000 SSOOO than $5000
(iii) Parents' Comoined Annual Income: Check the
appropriate cagegory:
less than $10,000 to $25,XO Above
$10,000 $25,000 to $50,000 $50,000
d. Health and Medical Care
(1) Do you consider yourself in excellent general health?
Yes No
If "No", what major health problem(s) do you have?
(11) Who provided primary care for your *aaily? Check more
than one if applicable.
Family/General Practitioner Internist
Pediatrician Other (specify)
Education
a. High School: Public Private
b. College:
(1) Private Public
(11) Major subject
Minor subject (if applicable)
c. Level of Education:
(1) List all degrees obtained at college level and above:
(11) List study at college level and above that stopped
before degree was obtained:
Work Experience Check areas in which you have had experience:
a. Medical care setting (e.g. hospital orderly, health
aide, clinical laboratory assistant)
b. "Helping* activities (e.g., camp counselor, social
service. Peace Corps)
c. Technical work, not related to actual medical care.
(e.g., laboratory technitian, research assistant,
engineering assistant)
d. Teaching
e. Dealirag with public (e.g., sales, public relations,
press)
f. Unskilled work
q. Other (specify)
-136-

-137-
OBLIGATIONS 4N0 a LANS
1. Obligations to Federal Agencies
a. Do you expect to owe "time" to any organization in return
for scholarships or loans in medical school?
Yes No
E IMPROVING HEALTH STATUS
Below are Usted several different approaches to improving
the health status of people not no* receiving adequate
health care. Rank each of them 1 to 5 in order of your
opinion of their prooaole effectlveness, with 1 most
effective and 5 least effective. (Use each number only
once).
b. If yes, which of the following agencies?
Armed Services
National Public Health Service
Home town or state sponsored agency
Other (specify)
c. If yes, how much time?
1 year 2 years 3 years
4 years Other
(specify)
2. Plans for the Future
Solo practice
Two man partnership
Group practice
Emergency room
Academic medicine
Armed Forces
National Public Health Service
Other (specify)
;er GOALS
. Each of the phrases below describes an important goal that
some oeoole want to achieve througn a career in medicine.
Please ran* your personal goals by cutting a "l" next to
your most important goal a "2" next to your second most
important goal, etc. (Your least important goal will be
ranxed "6") DO NOT USE ANY NUMBER (1-6) MORE THAN ONCE!
a. Helping lead my conmunity to the solution of
social and medical proolerrs.
b. Gaining the respect and friendship of other
pnysicians in my community.
c. Developing long-term intensive relationships
with my patients.
d. Contributing to medicai e. Assuring financial security for myself ana my
farm ly.
f. Allowing me to live in the type of area and
lifestyle that I prefer.
2. Please ran* order the following list cf goals in the same
.manner (1-6). The goal of greatest personal importance to
you should be ranked "1".
b.
c.
d.
e.
f.
Providing me with capital for investment in
business, real estate, the stock mancet, etc.
Being acknowledged by other physicians as a
leader in health care for my patients.
Having adequate time for my family and for
non-professional activities.
Gaining the trust and confidence of my patients.
Making me a respected member of my community.
Actively conducting research in my areas of
interest.
a. Improving housing conditions
b. Raising the general level of education
c. Improving public transportation to health care
facilities.
d. Expanding medical services (e.g., more private
doctors, more clinics)
e. Providing more jobs for the uneroloyed
ATTITUDES ABOUT PRIMARY DARE AND FAMILY MEDICINE
(Circle the ouno'er of the response wmcn best
describes your own personal feelings toward each
statement.)
The diagnosis and treatment
of illnesses in hosoitalizeq
patients is more difficult
than similar problems in
ambulatory patients.
Doctors who have large
inpatient practices tend to
be better doctors than those
with small inpatient practices.
The family physician can provide
his greatest service in follow
ing long-term health and adjust
ment of patients rather than in
concentrating on the treatment
of their irmediate complaints.
A wide variety of proalems err-
comoassing all age groups is
interesting to me.
In medical practice today, there
is sufficient aopropriate
specialists so mat a family
physician snould not assume long
term responsibility for patients
with cnronic illnesses.
I feel that preventive medicine
is more important than curative
medicine.
I bel i ove that the social and
family environment of patients
is a major influence on their
state of health.
Medical education should concen
trate primarily on recognition
and treatment of specific
disease processes.
h
Il II II
11
12 3 4
12 3 4
12 3 4
12 3 4
3 4
12 3 4
12 3 4
12 3 4
D MEDICAL EDUCATION CONTENT
1. Each of the following items may be an appropriate part of
medical education. Please rank then from 1 (most) to ''8"
(least) according to how important you believe that each is
in producing the best physician.
9. A patients ability to pay
should influence quality of
treatment given.
10. Epidemiological and preventive
medicine research is interesting
to me.
b.
c.
d.
e.
f.
Coursework in the "basic sciences"
Clinical experiences with hospitalized patients.
Clinical experiences with ambulatory patients.
Coursework in medical ethics.
Training in the business aspects of medical practice
Training in public and coemunity health.
Research into the causes and treatments of disease.
Research into patient behavior and compliance with
medical care.
11. A primary care physician should
use consultants for managing
critically ill patients.
12. An emotional upset should be as
valid an excuse for missing work
as a bad cold.
12 3 4
12 3 4
12 3 4
12 3 4

-138-
A Longitudinal Study of Family 3,-actice Education
Survey of entering students (continued)
ATTIT'JDES ABOUT PRIMARY CARE AMP ~AMLY MEDICINE (continued)
feelings toward eacn statement.
(Circle the nunber of the response which best describes your own personal
The only way to practice good
medicine is to do a complete
history and physical examina
tion eacn visit.
Most patients with emotional
disorders cannot be helped
without spending a lot of
the doctor's time.
Health maintenance is not
as interesting to me as
curative medicine.
Disease prevention should be
the responsibility of public
health departments rather
than the personal physician.
In order to provide the best
care *or his patients, a doctor
should personally perform as
many of the direct patient-
contact services as possible.
Consultants play a secondary
role in the patients total
health care.
People with impaired mental
health are as likely to get
well as oeoole with impaired
pnysiciai health.
Except for certain diseases
specific knowledge pertaining
to disease prevention is so
fragmented that a pnysician
snould limit his efforts to
curative medicine.
2 3 4
2 3 4 5
2 3 4
3 4
21. The provision of services by
specialty oriented pnysicians
should be coordinated and
controlled by primary care
pnysicians.
22. There is no useful researcn
work that can oe done in
primary care practice.
23. Family practices are not as
profitable financially as
specialty practices.
24. The demands of a family
practice leave little
leisure time 'or family
life or recreation.
25. A comolete medical and social
data base on each patient is
important to me.
26. Most patients medical problems
involve aspects of their coping
mechanisms witn life's daily
challenges.
27. Physician's assistants will
play an imoortant role n future
primary care.
28. I would prefer to spend my time
dealing witn patients' medical
proolems rather than their social
or psychological preplans.
29. Physician's assistants snould
handle the acute minor illness
problems in primary care.
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
Please indicate to the best of your knowledge your future plans for
residency training:
Fanily Medicine
Pediatrics
General Internal Medicine
Other primary care (specify)
Other specialty/subspecialty (specify)
Uncertain
strongly
disagree

APPENDIX B
MEDICAL STUDENT FOLLOW-UP QUESTIONNAIRE

A LONGITUDINAL STUDY OF FAMILY PRACTICE EDUCATION
A FOLLOW-UP SURVEY
NAME: SOCIAL SECURITY NUMBER:
CAREER GOALS
1.Each of the phrases below describes an iooortant goal
that sone people want to achieve wrougn a career in
medicine. Please rank your personal goals by putting
a l next to your most important goal, a "2" next to
your second most important goai, ate. (Your least
imoortant goal will be ranxed "5"' X NOT USE ANY
NUMBER (1-6) MORE THAN ONCE!
a. Helping lead my community to the solution
of social and medical problems.
b. Gaining the respect anc friendship of
other physicians in my ccmnunity.
c. Developing long-term intensive relation
ships with my patients.
d. Contributing to nedicai knowledge through
research.
e. Assuring financial security for myself
and my family.
f. Allowing me to live in the type of area
and lifestyle that I prefer.
2.Please rank order the following list of goals in the
same manner (1-6). The goal of greatest personal
importance to you should be ranxed "1".
a.
b.
Providing me with capital for investment
in business, real estate, the stock market,
etc.
Being acknowledged by ttner pnysicians as a
leader in health care *rr my patients.
Having adeouate time *zr my amily and for
non-orofessional activities.
Gaining the trust and confidence of my
patients.
Making me a respected ^emoer pf my
comnuni ty.
Actively conducting research in my areas
of interest.
MEDICAL EDUCATION CONTENT
1. Each of the following items may be an aoorooriate part
of medical education. Please ranx them from "1 (most)
to 8 (least) according to how iaoortant you believe
that each is in producing the best physician.
a. Course work in the basic sciences-.
b. Clinical experiences with nospitalized patients.
c. Clinical experiences with ambulatory patients.
d. Course work in medical ethics.
e. Training in the business aspects of medical
practice.
f. Training in public and coemunity health.
g. Research into the causes ano treatments of
disease.
h. Research into patient behavior and compliance with
medical care.
O ATTITUPES ABOUT PRIMARY CARE AND FAMILY MEDICINE
(Circle the numoer or the response wnich oest describes
your own personal feelings toward each statement.)
o
a w
C V V
O il T5 31
U. L. CL. _
** o* "O C l/
1. The diagnosis and treatment I 2
of illnesses in nosoitalized
patients is more difficult
than in similar problems in
amoulatory patients.
2. Ooctors who have large 1 2
inpatient practices tend to
be better doctors than those
with small inpatient practices.
3. The family physician can provide 1 2
his greatest service in following
long-term health and adjustment of
patients rather than in concentrating
on the treatment of their immediate
complaints.
4. A wide variety of problems 1 2
enconoassing all age groups is
interesting to me.
5. In meo.cal practice today, there 1 2
are sufficient aoorooriate
soecialists so that a family
physician should not assune long
term responsibility for patients
with chronic illnesses.
6. I feel that preventive medicine 1 2
is more important than curative
medicine.
7. I believe that the social and 12
family environment of patients is
a major influence on their state
of health.
8. Medical education should concentrate 1 2
primarily on recognition and treat
ment of specific disease processes.
9. A patient's ability to pay should 1 2
influence quality of treatment
given.
10. Epidemiological and preventive 1 2
medicine research is interesting
to me.
11. A primary care physician should 1 2
use consultants for managing
critically 111 patients.
12. An fictional upset should be as 1 2
valid excuse for tsissing work
as a bad cold.
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
-140-
tend to
disagree

-141
A Longitudinal Study of Family Prictlce Education
A Follow-Up Survey
ATTITUDES ABOUT PRIMARY CARE AMO rAMILY MEDICINE (continued) (Circle the number of the response which best describes your own personal
feelings tcwaro eacn statement!
13. The only way to practice good 12345
medicine is to do a comolete
history and physical examina
tion eacn visit.
14. Most patients with emotional 12345
disorders cannot be helped
without spending a lot of
the doctor's time.
21. The provision of services by 12345
specialty oriented physicians
snould be coordinated and
controlled by primary care
pnysiclans.
22. There is no useful researcn 12345
work that can be done in
primary care practice.
15.Health maintenance is not 12345
as interesting to me as
curative medicine.
23.Family practices are not as 12345
profitable financially as
specialty practices.
16.Disease prevention should be 12345
the responsibility of public
health deoartments rather
than the personal physician.
24.The demands of a family 12345
practice leave little
leisure time for family
Ufe or recreation.
17. In order to provide the best
care for his patients, a doctor
should personally perform as
many of the direct patient-
contact services as possible.
18. Consultants play a secondary
role in the patients total
health care.
13. People with impaired mental
health are as likely to get
well as oeoole with impaired
pnysicial healtn.
20. Except for certain diseases
specific Knowledge pertaining
to disease prevention is so
fragmented that a physician
should limit his efforts to
curative medicine.
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
25. A comolete medical and social
data base on each patient is
important to me.
26. Most patients' medical proolems
involve aspects of their cooing
mechanisms with life's daily
challenges.
27. Physician's assistants will
play an important role in future
primary care.
23. I would prefer to spend my time
dealing with patients' medical
proDlems rather than their social
or psychiological problems.
29. Physician's assistants should
handle the acute minor illness
problems in primary care.
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
strongly
disagree

APPENDIX C
LETTER OF EXPLANATION ATTACHED TO
ENTERING QUESTIONNAIRE

DEPARTMENT OF COMMUNITY HEALTH AND FAMILY MEDICINE
COLLEGE OF MEDICINE
UNIVERSITY OF FLORIDA
(904) 392-4321
BOX J-222, MSB
J. HILLIS MILLER HEALTH CENTER
GAINESVILLE, FLORIDA 32610
MEMORANDUM
TO: Phase B Medical Students
FROM: William L. Stewart, M.D.
RE: Clerkship Evaluation
The Department of Community Health and Family Medicine is
conducting a longitudinal study of specialty choices. This
study will correlate certain attitudes with specialty choices.
The information the survey contains will not be used in any
way that will identify the individuals. The information will
be used as group statistics.
Your name and social security number will be replaced by a
code number for entry into the computer. The code number
is to enable the Department to identify the specialty choice
or residency choice. No one will have access to the code
except the investigators.
There will be an abbreviated follow-up form to be completed
after the Family Medicine Rotation. We would like to thank
you for your assistance in this study. If you have any ques
tions, feel free to ask us.
This survey is not mandatory, but we would appreciate your
participation. If you decide not to participate, we would
appreciate your notifying us of that decision. An envelope
is enclosed. Thank you.
-143-
EQUAL LMPLOYMENT OPPORTUNITY/AFFIRMATIVE ACTION EMPLOYER

APPENDIX D
LETTER OF EXPLANATION ATTACHED TO
FOLLOW-UP QUESTIONNAIRE

DEPARTMENT OF COMMUNITY HEALTH AND FAMILY MEDICINE
COLLEGE OF MEDICINE
UNIVERSITY OF FLORIDA
(904) 392-4321
BOX J-222, MSB
J. HILLIS MILLER HEALTH CENTER
GAINESVILLE, FLORIDA 32610
MEMORANDUM
TO: Phase B Medical Students
FROM: William L. Stewart, M.D.
RE: Clerkship Evaluation
The Department of Community Health and Family Medicine is
very interested in your feedback on the rotation so that
worthwhile learning experiences can be continued and those
which are not meaningful can be revised or deleted. You
are requested to complete the evaluation form attached to
this memo.
To assure your anonymity, do not sign your name to the
evaluation form. When you have completed the forms place
them in the envelope and seal the envelope. The sealed
envelope should them be sent to Darlene Hood, Phase B
Secretary. The envelope with your name in the upper left
hand corner will be discarded. Your name appears on the
envelope to enable her to follow-up on those who have not
returned their evaluation form. Faculty members responsible
for evaluation of students do not see the individual evalu
ations. Your grade will not be forwarded to Dean Hill's
office until the rotation evaluation has been completed and
received in the Phase B Secretary's office.
At the orientation to the Community Health and Family Medi
cine Clerkship you participated in a longitudinal study on
attitudes toward Family Medicine. Please assist us by com
pleting the follow-up portion which is attached to this memo.
Do not write your name or social security number on the form.
Return the completed form in the enclosed envelope. Thank
you for your participation.
WLS/mls
Enclosure
-14 5-
EQUAL EMPLOYMENT OPPORTUNITY/AFFIRMATIVE ACTION EMPLOYER

APPENDIX E
LETTER OF REMINDER SENT WITH SECOND
FOLLOW-UP QUESTIONNAIRE

DEPARTMENT OF COMMUNITY HEALTH AND FAMILY MEDICINE
COLLEGE OF MEDICINE
UNIVERSITY OF FLORIDA
(904) 392-4321
BOX J-222, MSB
J. HILLIS MILLER HEALTH CENTER
GAINESVILLE, FLORIDA 32610
Attached is the follow-up form for the longitudinal
study in Family Practice Education. I know how busy your
schedule is but it means alot to the validity of the study
to obtain the follow-up. Thank you for your help in this
regard.
Sincerely,
Margaret Duerson, M.Ed.
Curriculum Coordinator
MD/mls
Enclosure
-147-
EQUAL EMPLOYMENT OPPORTUNITY/AFFIRMATIVE ACTION FMPl nvFB

REFERENCES
Agresti, A., & Wackerly, D. (1977). Some exact conditional
tests of independence for R x C cross-classification
tables. Psychometrika, 4 2, 111-125.
Anderson, R.B.W. (1975). Choosing a medical specialty: A
critique of literature in the light of "curious
findings." Journal of Health and Social Behavior, 16,
152-162.
Association of American Medical Colleges. (1977).
Descriptive study of medical school applicants, 1975-76
(DHEW, Publication No. HRA 77-52). Hyattsville, MD:
Author.
Association of American Medical Colleges. (1984).
Physicians for the twenty-first century. Washington,
DC: Author.
Becker, H.S., Geer, B., Hughes, E., & Strauss, A.L. (1961).
Boys in white: Student culture in medical school.
Chicago: University of Chicago Press.
Bergquist, S.R., Duchac, B.W., Schalin, V.A., Zastrow, J.F.,
Barr, V.L., & Borowiecki, T. (1985). Perceptions of
freshman medical students of gender differences in
medical specialty choice. Journal of Medical
Education, 60, 379-383.
Brearly, W.D., Simpson, W., & Baker, R.M. (1982). Family
practice as a specialty choice: Effect of premedical
and medical education. Journal of Medical Education,
57, 449-454.
Burkett, G.L., & Gelula, M.H. (1982). Characteristics of
students preferring family practice/primary care
careers. Journal of Family Practice, 15, 505-512.
Califano, J.A. (1979). The government-medical education
partnership. Journal of Medical Education, 54, 19-24.
Campbell, D.T., & Stanley, J.C. (1963). Experimental and
quasi-experimental designs for research. Boston:
Houghton Mifflin.
-148-

-149-
Carline, J.D., Cullen, T.J., Dohner
Zinser, E.A. (1980). Career
second-year medical students:
Journal of Medical Education,
, C.W., Schwarz, M.R.
preferences of first-
The WAMI experience.
55, 682-691.
&
and
Champion, D.J.
research.
(1970). Basic statistics for social
Scranton, PA: Chandler Publishing Co.
Chandra, P., & Hughes, M. (1984). Factors affecting the
choice of anesthesiology by medical students for
specialty training. Journal of Medical Education, 59,
323-330.
Cole, D.R., Fox, T.G., & Liberman, J.A. (1983). Social
identities of medical students oriented toward careers
in family medicine. Journal of Family Practice, 17,
332-334, 338.
Collins, F., & Roessler, R. (1975). Intellectual and
attitudinal characteristics of medical students
selecting family practice. Journal of Family Practice,
2, 431-432.
Cuca, J.M. (1979). The specialization and career
preferences of women and men recently graduated from
U.S. medical schools. Journal of American Medical
Womens Association, 34, 425-435.
Dallman, J.J., Crandall, L.A., & Haas, W.H. (1980).
Factors affecting residency program dropouts: A
longitudinal study. Journal of the Florida Medical
Association, 67, 833-835.
Department of Community Health and Family Medicine. (1981).
A longitudinal study of family practice education:
Survey of entering students and follow-up survey.
University of Florida, Gainesville.
/
Dube, W.F. (1973). Women students in U.S. medical schools:
Past and present trends. Journal of Medical Education,
48, 186-189.
Fadem, B.H., Nicolich, M.J., Simring, S.S., Dauber, M.H., &
Bullock, L.A. (1984). Predicting medical specialty
choice: A model based on students' records. Journal
of Medical Education, 59, 407-415.
Friedman, C.P., Striter, F.T., & Talbert, L.M. (1978). A
systematic comparison of teaching hospital and remote-
site clinical education. Journal of Medical Education,
53, 565-573.

-150-
Funkenstein, D.H. (1978). Medical students, medical
schools and society during five eras: Factors
affecting the career choices of physicians 1958-1976.
Cambridge, MA: Ballinger Publishing Co.
Ginzberg, E. (1984). The monetarization of medical care.
New England Journal of Medicine, 310, 1162-1165.
Goldsmith, G. (1982). Factors influencing family practice
residency selection: A national survey. Journal of
Family Practice, 15, 121-124.
Gough, H.G. (1978). Some predictive implications of
premedical scientific competence and preferences.
Journal of Medical Education, 53, 291-300.
Gough, H.G., & Hall, W.B. (1977). A comparison of medical
students from medical and nonmedical families. Journal
of Medical Education, 52, 541-547.
Hadac, R.R. (1984). A longitudinal study of factors
associated with choice of medical specialty and
practice setting at the University of Washington
(Doctoral dissertation, University of Washington,
1983). Dissertation Abstracts International, 44,
1029A.
Hadley, J.
choice
(1977). An
Inquiry,
empirical model
14, 384-401.
of medical specialty
Harr
is, D.L., Coleman, M., & Malea, M.
participation in a family practice
student career decisions. J ourna 1
Education, 57, 609-614.
(1982). Impact of
track program on
of Medical
Harr
is, D.L., Kelley, K., & Coleman, M. (1984). The
stability of personality types and their usefulness
medical student career guidance. Family Medicine,
203-205.
in
16,
Herman, M.W., & Veloski, J. (1977). Family medicine and
primary care: Trends and student characteristics.
Journal of Medical Education, 52, 99-106.
Hutt, R. (1976). Doctors' career choice: Previous
research and its relevance for policy-making. Medical
Education, 10, 463-473.
Jacoby, I. (1981). Physicians manpower: GMENAC and
afterward. Public Health Reports, 96, 295-303.
Kindig, D.A., & Dunham, N.C. (1985). Physician specialist
growth into the 21st century. Journal of Medical
Education, 60, 558-559.

-151-
Lanska, M.J., Lanska, D.J., & Rimm, A.A. (1984). Effect of
rising percentage of female physicians on projections
of physician supply. Journal of Medical Education, 59,
849-855.
Leserman, J. (1978). The professional values and expecta
tions of medical students. Journal of Medical
Education, 53, 330-336.
Marks, R.G. (1982). Analyzing research data. Belmont, CA:
Lifetime Learning Publications.
Matteson, M.T., & Smith, S.V. (1977). Selection of medical
specialties: Preferences vs. choices. Journal of
Medical Education, 52, 548-554.
Mawardi, B.H. (1979). Physicians and their careers
(Monograph: Sponsor Series, for Case Western Reserve
University). Ann Arbor, MI: University Microfilms
International.
McCaulley, M.H. (1981). The Meyers-Briggs Type Indicator
in medical career planning. Gainesville, FL: Center
for Applications of Psychological Type.
McGrath, E., & Zimet, C.N. (1977). Female and male medical
students: Differences in specialty choice selection
and personality. Journal of Medical Education, 52,
293-300.
Naisbitt, J. (1982). Megatrends: Ten new directions
transforming our lives. New York: Warner Books.
Neuhauser, D. (1983). Twenty-first century medical
education: Economic, social, political and
technological implications for change. Speech
presented at the 94th Annual Meeting of the Association
of American Medical Colleges, Washington, DC.
Paiva, R.E.A., Vu, N.V., & Verhulst, S.J. (1982). The
effect of clinical experiences in medical school on
specialty choice decisions. Journal of Medical
Education, 57, 666-674.
Petersdorf, R.G. (1975). Health manpower: Numbers,
distribution, quality. Annals of Internal Medicine,
82, 694-701.
Plovnick, M.S. (1980). Medical students values, socializa
tion, and primary care career choices. Journal
of Family Practice, 11, 323-324.

-152-
Rakel, R.E., & Pisacano, N.J. (1984). The family
physician. In R.E. Rakel (Ed.), Textbook of family
practice (3rd ed., pp. 3-20). Philadelphia: W.B.
Saunders.
Rezler, A.G. (1974). Attitude changes during medical
school: A review of the literature. Journal of
Medical Education, 49, 1023-1030.
Robbins, L., Robbins, E.S., Katz, S.E., Geliebter, B., &
Stern, M. (1983). Achievement motivation in medical
students. Journal of Medical Education, 58, 850-858.
Roos, N.P., & Roos, L.L. (1980). Medical school impact on
student career choice. Evaluation and the Health
Professions, _3, 3-19.
Rosenblatt, R.A., & Alpert, J.J. (1979). The effect of a
course in family medicine on future career choice: A
long-range follow-up of a controlled experiment in
medical education. Journal of Family Practice, 8,
87-91.
Rosevear, G.C., Tickman, M.S., & Gary, N.E. (1985).
Relationship between specialty choice and academic
performance. Journal of Medical Education, 60, 640-
641.
Samra, S.K., Davis, W., Pandit, S.K., & Cohen, P.J. (1983).
The effect of a clinical clerkship on attrition of
medical students toward anesthesiology. Journal of
Medical Education, 58, 641-647.
Schroeder, S.A. (1984). Western European responses to
physicians oversupply. 2. Lessons for the United
States. Journal of the American Medical Association,
252, 373-384.
Skipper, J.K., & Gliebe, W.A. (1977). Forgotten persons:
Physicians' wives and their influences on medical
career decisions. Journal of Medical Education, 52,
764-765.
Steinwachs, D.M., Levine, D.M., Elzinga, D.J., Salkever,
D.S., Parker, R.D., & Weisman, C.S. (1982). Changing
patterns of graduate medical education. 2. Analyzing
recent trends and projecting their impact. New
England Journal of Medicine, 306, 10-14.
University of Florida, College of Medicine. (1985).
1986 College of Medicine catalog. Gainesville:
University of Florida.
1985-

-153-
U.S. Department of Health and Human Services, Health
Resources Administration. (1980). Summary report of
the graduate medical education national advisory
committee, Vols. I-V (DHHS Publication No. HRA 81-651).
Washington, DC: U.S. Government Printing Office.
Weisman, C.S. (1984). Gender composition of medical
schools and specialty choices of graduates. Journal
of Medical Education, 59, 347-349.
Weiss, K. (1982). Corporate medicine. 2. What's the
bottom line for physicians and patients? The New
Physician, _9, 19-25.
White, K.L., Williams, T.F., & Greenberg, B.G. (1961).
Ecology of medical care. New England Journal of
Medicine, 265, 885-891.
Zimny, G.H., & Shelton, B.R. (1982). Sex differences in
medical specialty preferences. Journal of Medical
Education, 57, 403-405.
Zuckerman, H.S. (1977). Evaluation of literature on career
choice within medicine. Medical Care Review, 34, 1079-
1100.
Zuckerman, H.S. (1978). Structural factors as determinants
of career patterns in medicine. Journal of Medical
Education, 53, 453-463.

BIOGRAPHICAL SKETCH
Margaret Craig Duerson was born in Louisiana.
Following primary and secondary education in Mississippi,
she returned to the Bayou state where she began lower
division studies at Louisiana State University in Baton
Rouge. She married Goodloe Kearney Duerson, a graduate in
engineering. For several years she devoted full time to her
growing family and community activities.
Resuming her education at the local community college
where the family had settled in Lake City, Florida, her
interest in health care and working with people drew her to
the nursing program. She received the Associate of Arts
degree and the Associate of Science in Nursing degree
graduating summa cum laude from Lake City Community College
in 1970.
Mrs. Duerson worked for three years as a critical care
unit nurse in the local veterans administration hospital.
Desiring to broaden her knowledge base, she continued upper
division nursing studies at the University of Florida where
she graduated with high honors earning a Bachelor of Science
in Nursing degree in 1974.
Upon graduation she joined the University of Florida
College of Medicine as coordinator of clinical research for
-154-

-155-
the Department of Surgery. While serving in this capacity,
she earned a Master of Education degree from the University
of Florida in 1978.
In 1982, Mrs. Duerson accepted her present position as
curriculum coordinator for undergraduate medical education
for the Department of Community Health and Family Medicine
at the University of Florida, College of Medicine.
Currently she is a member of the American Nurses'
Association, the Florida Nurses' Association, the Society of
Teachers of Family Medicine, Kappa Delta Pi, Phi Delta
Kappa, and Phi Kappa Phi.

I certify that I have read this study and that 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.
Ld /'
Jame¡a W. Hensel, Chairman
Professor of Educational Leadership
I certify that I have read this study and that 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.
Forrest W. Parkay
Associate Professor o
Leadership
Educational
I certify that I have read this study and that 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.
Small
Professor of Irttmuno.
Microbiology
Medical

I certify that I have read this study and that 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.
^ Cvv-A^-J Sty. l/l/l
Ronald G. Marks
Associate Professor of Statistics
I certify that
opinion it conforms
presentation and is
a dissertation for
I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the degree of Doctor of Philosophy.
as
This dissertation was submitted
the College of Education and to
accepted as partial fulfillment
degree of Doctor of Philosophy.
to the Graduate Faculty
the Graduate School and
of the requirements for
of
was
the
December 1986
Dean, College of Education
Dean, Graduate School



-109-
Hypothesis 6
There is no difference between the groups of medical
students choosing primary care and those choosing
subspecialties in the fourth year with respect to their
rankings of career goals before and after a family medicine
clerkship. Medical students ranked the personal importance
of two sets of six career goal phrases (total 12) prior to a
third-year family medicine clerkship and again, immediately
following the clerkship. The difference in the preclerkship
and postclerkship rankings served as the observations for
testing the independence of the two specialty choice groups,
primary care and subspecialty, which were identified from
fourth-year residency choices. The statistical procedure
used for the analysis was the Kruskal-Wallis one-way
analysis of variance. The overall significance level was
set at < .01.
This change in career goal rankings from preclerkship
to postclerkship did not distinguish the two fourth-year
specialty choice groups at the _< .01 significance level.
Hypothesis 6 was not rejected. Results by goal statement
are shown in Table 14.
Hypothesis 7
There is no association between medical students'
specialty preferences expressed in the third year and their
career choice decisions evidenced by fourth-year residency
selection. The last item contained in the questionnaire


Table 13
Correlation Coefficients for Career Goal Rankings and Attitude Scores
Preliminary Goal Follow-Up Goal
Rankings Rankings
Preliminary
Follow-Up
Follow-Up
Career Goals
Attitudes
Attitudes
Attitudes
Set
1
a.
helping lead my community to the
solution of social and medical problems
-.11
r1

1
0.22*
b.
gaining respect and friendship of other
physicians
.04
.08
.003
c.
developing long-term relationships with
patients
-.27**
-. 15
-.17
d.
contributing to medical knowledge through
research
.07
i

o
-. 16
e.
assuring financial security for self and
family
.03
.04
.22*
f.
allowing me to live in preferred area
and style
. 28**
.29**
.31**
-SOI-


-155-
the Department of Surgery. While serving in this capacity,
she earned a Master of Education degree from the University
of Florida in 1978.
In 1982, Mrs. Duerson accepted her present position as
curriculum coordinator for undergraduate medical education
for the Department of Community Health and Family Medicine
at the University of Florida, College of Medicine.
Currently she is a member of the American Nurses'
Association, the Florida Nurses' Association, the Society of
Teachers of Family Medicine, Kappa Delta Pi, Phi Delta
Kappa, and Phi Kappa Phi.


REFERENCES
Agresti, A., & Wackerly, D. (1977). Some exact conditional
tests of independence for R x C cross-classification
tables. Psychometrika, 4 2, 111-125.
Anderson, R.B.W. (1975). Choosing a medical specialty: A
critique of literature in the light of "curious
findings." Journal of Health and Social Behavior, 16,
152-162.
Association of American Medical Colleges. (1977).
Descriptive study of medical school applicants, 1975-76
(DHEW, Publication No. HRA 77-52). Hyattsville, MD:
Author.
Association of American Medical Colleges. (1984).
Physicians for the twenty-first century. Washington,
DC: Author.
Becker, H.S., Geer, B., Hughes, E., & Strauss, A.L. (1961).
Boys in white: Student culture in medical school.
Chicago: University of Chicago Press.
Bergquist, S.R., Duchac, B.W., Schalin, V.A., Zastrow, J.F.,
Barr, V.L., & Borowiecki, T. (1985). Perceptions of
freshman medical students of gender differences in
medical specialty choice. Journal of Medical
Education, 60, 379-383.
Brearly, W.D., Simpson, W., & Baker, R.M. (1982). Family
practice as a specialty choice: Effect of premedical
and medical education. Journal of Medical Education,
57, 449-454.
Burkett, G.L., & Gelula, M.H. (1982). Characteristics of
students preferring family practice/primary care
careers. Journal of Family Practice, 15, 505-512.
Califano, J.A. (1979). The government-medical education
partnership. Journal of Medical Education, 54, 19-24.
Campbell, D.T., & Stanley, J.C. (1963). Experimental and
quasi-experimental designs for research. Boston:
Houghton Mifflin.
-148-


-40-
Harvard experiment where the teaching of family medicine was
introduced in 1953. Rosenblatt and Alpert (1979) followed
up the first three cohorts of this early program to
determine the impact of the experience on subsequent career
choice, amount of family-oriented practice, and other
achievements. No statistically significant differences
could be found among the various groups, either within
cohorts or across time. Of critical importance to the
interpretation of the results of this very early attempt to
integrate family medicine into the curriculum is the fact
that at the time the program was introduced, Harvard had no
faculty in family medicine to serve as role models. In fact
in the latter years (cohort II and III) the directors were
two pediatricians, not family practitioners.
It is reasonable to believe that the availability of
role models and opportunities to practice the role will have
a beneficial effect on the numbers of students choosing a
specialty. In an effort to determine the beneficial and
detrimental parameters influencing the choice of family
practice as a career, Brearley and his colleagues (1982)
surveyed 134 first year residents of southeastern family
practice residencies on 18 curriculum components and
elements that influenced specialty choice. The positive
value of preceptorship experiences in family medicine during
the third and fourth years of medical school and association
with family physicians was striking. Detrimental influences
were peer group attitudes and the traditional curriculum


APPENDIX D
LETTER OF EXPLANATION ATTACHED TO
FOLLOW-UP QUESTIONNAIRE


-15-
and within their range of competence (Rakel & Pisacano,
1984 ) .
Clerkships are student-selected elective and required
clinical experiences occurring during the last two years of
medical school. Working as a member of the health care team
in the actual care of patients provides the student with the
opportunity to apply and practice newly acquired knowledge
and skills.
Preceptor refers to a licensed, practicing physician
who teaches medical students. This physician may be in
private practice or in an academic setting.
The terms occupation preference, choice, and attainment
are frequently used interchangeably. Vroom's distinction
between occupation choice, preference, and attainment as
explicated by Matteson and Smith (1977) are
Occupation preference is that occupation which at a
given time an individual would most like or prefer to enter.
Occupation choice is that occupation which a person
chooses to enter and then engages in behaviors to implement
that choice.
Occupation attainment is the occupation of which the
person is a member.
This study concerns itself with two points of
occupational decision, preference and choice.
Preference of specialty is a positive written or verbal
expression of a desire to pursue the particular area of
medical practice. The participants' preferences for a


Results of the match were received by the individual
students, the residency programs, and the medical schools
the following March.
Research Design
To investigate the research questions for this study, a
pre- and post-research design was used with multiple group
replication measures. The research procedure involved
observations taken immediately before and after a clerkship
and again after a passage of time. The design used intact
groups of third-year medical students previously formed into
rotation groups through a lottery system.
The design for this study was considered a variation of
the Recurrent Institutional Cycle Design, a strategy for
field research described by Campbell and Stanley (1963) as
appropriate to those situations in which some aspect of the
institutional process is on a cyclical schedule. In field
research in educational settings where restrictions prevent
an experimental design with control of who would be exposed
to the experimental variable and the effects of a global and
complex construct are sought, this design offers a measure
of strength over the one-group pretest-posttest design. The
fact that the treatment variable, the clerkship, is
continually being presented to a group previously exposed
and a group about to be exposed makes for some degree of
experimental control as elements of the cross-sectional and
longitudinal approaches are present.


-72-
exact tests of independence was used for R x C cross
classification tables when R or C was greater than two
(Agresti & Wackerly, 1977).
3. Hypothesis 3 tested the null hypothesis that there
is no difference between medical students grouped by
sociodemographic characteristics according to their
attitudes toward primary care before or after a family
medicine clerkship. Individual attitude scores were formed
by summing each student's responses to the Attitudes about
Primary Care and Family Medicine section of the
questionnaire. Attitudes were measured before and again
after a family medicine clerkship. The Kruskal-Wallis one
way analysis of variance was used to determine whether or
not there was a difference in distribution of attitude
measurements between the groups of students identified by
their sociodemographic characteristics. The primary
assumption for this test is that the data be ordinal-level
data. The data to be ranked should be continuous rather
than discrete. The Kruskal-Wallis makes no assumptions
concerning distributions and it is appropriate for unequal
sample sizes, both large and small. Based on these
assumptions it was considered a suitable test for this
hypothesis. When there are more than five subjects in each
element or group, as occurred in this study, the Kruskal-
Wallis H value is treated as a X2 for interpretive purposes
(Champion, 1970, p. 188).
4. The Kruskal-Wallis one-way analysis of variance
test was again deemed the appropriate choice of statistical


-100-
Table 11
Results of Preclerkship Career Goal Rankings and
Three Preference Groups
Career Goals4" N = 72
H £
Set 1
a. helping to solve community's medical
and social problems
7.77
.02
b. gaining respect and friendship of
other physicians
2.36
.30
c. developing long-term relationships
with patients
2.85
.24
d. contributing to medical knowledge
through research
1.17
.56
e. assuring financial security for
self and family
1.06
.59
f. allowing me to live in preferred
area and style
3.15
.21
Set 2
a. providing me with capital for
investment
1.50
.47
b. being acknowledged by other
physicians as a leader in health
care for my patients
1.06
.59
c. having time for family and non
professional activities
.53
.77
d. gaining trust and confidence of
my patients
4.55
.10
e. making me a respected member of
my community
.28
.87
f. actively conducting research in my
areas of interest
1.77
.41
+Paraphrased
from original questionnaire


-51-
nonprimary care specialists. The older, married students
from small towns are most likely to choose primary care
specialties. With all other background characteristics, the
results are mixed.
It appears that the variables brought to the medical
setting diminish in importance as factors related to the
medical training system increase in influence on specialty
choice. Preliminary data suggest that role models and the
opportunity to practice a role reinforce or refute the
beliefs and attitudes that students bring with them into the
clinical setting and consequently effect their perception of
the various medical specialties. However, the studies in
this area have largely used self-selected groups of subjects
who take an elective course in a clinical clerkship so that
the results are not generalizable.
Career factors have not received adequate attention by
researchers. Information on the career and personal goals
of young doctors could provide clues as to the needed
changes in the work setting to make shortage areas more
attractive career options.
In sum, what the literature seems to indicate is that
(a) results with sociodemographic variables are mixed and
therefore require further study; (b) even though there are
methodological problems in preliminary studies,
opportunities for interaction with role models and role
playing in the clinical setting holds promise for
influencing specialty choice but further study is needed;


-13-
5. The follow-up questionnaire was mailed to medical
students at the completion of the family medicine clerkship.
Mailed surveys often result in low return rates and even
lower percent usable data as some items are not completed.
6. Working in the real world of educational research
has the advantages of practical applicability and the
disadvantages of having to deal with many variables
simultaneously. If there is to be any application of the
method to this or other sites, the natural setting has
obvious advantages.
Assumptions of This Study
The following assumptions are implicit in the
limitations and delimitations of the study:
1. Participants responded voluntarily to questionnaire
iterns.
2. Responses were made in a nonthreatening environment
so they were truthful and those intended by participants.
3. Administration and collection of the questionnaire
did not influence the responses.
4. The participants were representative of previous
medical students at the same level of training and those
that follow at the University of Florida College of
Medicine.
5. Experiences while on the family practice rotation
were similar for all students.


-120-
preclerkship to postclerkship successfully discriminate
between the two specialty groups.
Correlational relationships were found between
students' rankings of several career goal statements and
their scores on attitudes toward primary care. In
preclerkship measurements, students who had high scores on
attitudes toward primary care considered "developing long
term relationships with patients to be an important career
goal." The goal "allowing me to live in the preferred area
and life style" was significantly correlated with
preclerkship and postclerkship attitudes. The direction of
the correlation was interpreted as those students who scored
low on attitudes toward primary care considered this an
important goal. Postclerkship ranking of the goal
"providing me with capital for investment" also correlated
positively with low postclerkship attitude scores. A
negative correlation resulted from the postclerkship
rankings of the goal "actively conducting research in my
areas of interest" and attitude scores which indicated that
students who scored high on attitude scores considered this
an important career goal.
A statistically significant difference in students'
responses to the attitude items could not be demonstrated
from the pre- and postclerkship data (£ = .04). Nor did the
attitude measurements prove to be helpful in discriminating
career preference (£ = .10) and specialty choice groups
(£ = .02) .


-112-
Table 15
Association between Third-Year Specialty Preferences
and Fourth-Year Specialty Choices
Fourth-Year Specialty Choices
Third-Year Specialty
Preferences
Primary Care
n Row %
Subspecialty
n Row %
Primary Care
29
85.29
5
14.71
Subspecialty
4
23.53
13
76.47
Undecided
12
57.14
9
42.86


-62-
respond to this section revealed no significant differences
at £ .01 between those who responded and those who did not.
The follow-up survey questionnaire items were identical
to the entering survey with the exception of the deletion of
the first two sections containing social, personal, and
demographic data. The last section on future plans for
residency training was also not included in the follow-up
survey. The abbreviated follow-up survey included four
sections common to the entering questionnaire: Career
Goals, Medical Education Content, Improving Health Status,
and Attitudes Toward Primary Care and Family Medicine.
Administration of the Questionnaire
Sample
The sampling frame for the study included all medical
students in the third year of required clinical clerkships
in 1983. Third-year medical students were chosen because
they were further along in the professional and
socialization process and because fourth-year students are
geographically scattered due to the nature of the curriculum
which allows for externship at other sites and time out of
school for interviews for residency positions.
During the clinical portion of Phase B students rotate
in small groups through the individual clinical courses or
clerkships. These clerkships were scheduled in blocks of
eight weeks each. The Community Health and Family Medicine
clerkship was allotted six weeks of one of the eight-week


Final
goes to my
whose love
and grow,
possible.
y, and most especially, my sincerest appreciation
husband, Kearney, my two sons, and my parents
and encouragement have enabled me to develop
Kearney's faith in me ultimately made this
ii i


-151-
Lanska, M.J., Lanska, D.J., & Rimm, A.A. (1984). Effect of
rising percentage of female physicians on projections
of physician supply. Journal of Medical Education, 59,
849-855.
Leserman, J. (1978). The professional values and expecta
tions of medical students. Journal of Medical
Education, 53, 330-336.
Marks, R.G. (1982). Analyzing research data. Belmont, CA:
Lifetime Learning Publications.
Matteson, M.T., & Smith, S.V. (1977). Selection of medical
specialties: Preferences vs. choices. Journal of
Medical Education, 52, 548-554.
Mawardi, B.H. (1979). Physicians and their careers
(Monograph: Sponsor Series, for Case Western Reserve
University). Ann Arbor, MI: University Microfilms
International.
McCaulley, M.H. (1981). The Meyers-Briggs Type Indicator
in medical career planning. Gainesville, FL: Center
for Applications of Psychological Type.
McGrath, E., & Zimet, C.N. (1977). Female and male medical
students: Differences in specialty choice selection
and personality. Journal of Medical Education, 52,
293-300.
Naisbitt, J. (1982). Megatrends: Ten new directions
transforming our lives. New York: Warner Books.
Neuhauser, D. (1983). Twenty-first century medical
education: Economic, social, political and
technological implications for change. Speech
presented at the 94th Annual Meeting of the Association
of American Medical Colleges, Washington, DC.
Paiva, R.E.A., Vu, N.V., & Verhulst, S.J. (1982). The
effect of clinical experiences in medical school on
specialty choice decisions. Journal of Medical
Education, 57, 666-674.
Petersdorf, R.G. (1975). Health manpower: Numbers,
distribution, quality. Annals of Internal Medicine,
82, 694-701.
Plovnick, M.S. (1980). Medical students values, socializa
tion, and primary care career choices. Journal
of Family Practice, 11, 323-324.


Page
V SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS.... 116
Summary 116
Conclusions 121
Implications of the Study 127
Recommendations 128
APPENDICES
A MEDICAL STUDENT ENTERING QUESTIONNAIRE 136
B MEDICAL STUDENT FOLLOW-UP QUESTIONNAIRE 140
C LETTER OF EXPLANATION ATTACHED TO ENTERING
QUESTIONNAIRE 143
D LETTER OF EXPLANATION ATTACHED TO FOLLOW-UP
QUESTIONNAIRE 145
E LETTER OF REMINDER SENT WITH SECOND FOLLOW-UP
QUESTIONNAIRE 147
REFERENCES 148
BIOGRAPHICAL SKETCH 154
v


CHAPTER IV
PRESENTATION OF RESULTS
Sociodemographic Characteristics of
the Study Group
Data reported here were obtained from a survey of 109
medical students of the 1984-85 graduating class of a state-
supported medical school in the south/southeast. The data
are presented in Table 1. Some categories from the original
survey have been combined for purposes of statistical
analysis and ease in reporting. Students ranged in age from
22 to 35 with a mean age of 25 years. Males made up 73% of
the respondents; females, 27%. Whites were heavily
represented with 88% of the group while all other races
combined accounted for the additional 12%. The religious
preferences of the group were (a) 44% Protestant, (b) 22%
Catholic, (c) 11% Jewish, and (d) the remainder indicated no
religious preference. There was almost no variation between
the percentages of students indicating that religion played
a great deal of importance in their lives (37%), some
importance (35%), and very little to none at all (28%).
Eighty-two percent of the students were from the
south/southeast region of the country with about equal
representation from small towns, mid-sized cities, and
metropolitan areas. Forty-seven percent of the subjects
-76-


-96-
precollege years. The greatest number of students'
hometowns were in the same geographic area as the medical
school, south/southeast. When students were categorized
into groups according to whether their hometown was
south/southeast or another location, differences between the
two samples were found with respect to attitudes toward
primary care. The preclerkship data yielded an H value of
4.45 (d_f 1, N = 103), £ = .03 (Table 9). The postclerkship
test statistic was H (d^f 1, N = 86) = 4.02, £ = .05 (Table
10). In both pre- and postclerkship data analysis the
students from the south/southeast had lower mean rank scores
indicating lower attitude scores.
Father's occupation. Father's occupation while growing
up, classified as professional/technical or business/other,
was the basis for identification of two samples of students
which were examined for differences in attitudes toward
primary care. Postclerkship, the observed H statistic for
the two samples was 4.06 with 1 degree of freedom and
significance level of .04. The mean rank of 48.9 was higher
for the business/other group than the professional/technical
group's score of 38.1 indicating a higher attitude score for
the business/other student group.
Source of present support. Responses pertaining to the
major contributor to the students' present support were
categorized as (a) parents and (b) other. When students
from these two groups were analyzed for differences or
similarities in their attitudes toward primary care in the


-2-
taken to correct the situation. Using the supply
assumptions of the GMENAC Report, Kindig and Dunham (1985)
projected physician specialist growth into the 21st century.
By the year 2020, the absolute number of physicians in
primary care will increase by 55% but the absolute numbers
in nonprimary care will increase by 111%. As these authors
observed, it is conceivable that with the aging of the
population and the competitive approaches to cost
containment the additional primary care physicians will be
needed but it is difficult to assume a need for an
additional 260,000 nonprimary care physicians.
Intervening in the situation while the numbers of
physicians yet to be trained is amenable to change is
essential. Two approaches toward correcting the problem
include (a) stringent government regulations to limit the
types and numbers of residencies which train specialists, to
bring reimbursement of specialists in line with primary care
physicians, and to restrict the number of graduates of
foreign medical schools and foreign physicians coming into
the country; and (b) working within the educational
environment to influence specialty choice decisions. Of the
two approaches, the latter seems the most viable solution in
a democratic society to meet health care needs and at the
same time allow individuals freedom in the specialty
decision process.
Until recently there has been little effort directed
toward assuring that the medical education system produced


-78-
judged their hometown economic base as a mixed economy.
Thirty-six percent came from places dominated by a business
economy. Farming communities contributed 11% to the sample
and only 6% were from industrial areas.
When asked to indicate their parents' occupations while
they were growing up, 52% of the students checked physician,
health care provider, or other profession. Forty-eight
percent cited business, clerical/sales, skilled, unskilled,
farm/farm worker, and other as their father's occupation.
Mothers' occupations were equally divided between housewife
and working outside the home. Considering their parents'
annual income, 69% had parents with incomes that fell in the
$50,000 or less categories compared with 31% in the above
$50,000 stratum.
Seventy-one percent of the students indicated that
their principal source of support came either from
themselves, spouses, or a combination; parents supported
29%. The financial status is further clarified by the
information that 81% of the respondents reported some
educational debt. For those reporting indebtedness, 92% had
debts greater than $5000.
For the majority of the students, their education took
place in public high schools (78%) and public-supported
colleges (71%). As would be expected, 86% majored in
science with 91% obtaining a bachelor's degree and 9%
achieving a graduate degree.


-81-
Table 3
Residency Choice of Fourth-Year Medical Students
Specialty Choice Groups
n
Percent
Primary Care
family medicine
15
13.8
pediatrics
8
7.3
internal medicine
39
35.8
transitional
10
9.2
Total
72
66.1
Subspecialty
ob/gyn
11
10.1
ER medicine
1
.9
anesthesiology
2
1.8
neurology
1
.9
dermatology
1
.9
pathology
4
3.7
radiology
4
3.7
radiation therapy
3
2.7
physical med. & rehab.
1
.9
surgery
4
3.7
plastic surgery
1
.9
neuro surgery
1
.9
orthopedic surgery
1
.9
psychiatry
1
.9
Total
36
32.9


-20-
from rural areas tended more often to prefer general
practice.
Closely related to choice of specialty is choice of
practice location. The relationship between practice
location and specialization becomes more apparent when one
considers that highly trained subspecialists would not
likely find sufficient patients or the sophisticated
equipment necessary for practice of their specialty in small
towns. In a study of students entering the University of
Washington School of Medicine in 1975, 1976, and 1977,
Carline and associates (1980) found that contrary to other
studies there were few differences in preferences for
specialties based on size of hometown, while students'
attitudes toward location of practice indicated a preference
for practice in communities similar in size to those in
which they were raised. Support for this finding was also
reported by Hadac (1984). In a review of the literature
associated with specialty and location choices, he found
that a preference for rural or small town practice was
associated with high school attendance in a town of similar
size.
A second variable of interest is socioeconomic status
or social class which has been commonly measured in terms of
the father's occupation and level of education. Studies of
the relationship between social class and choice of
specialty are mixed. Results of some studies demonstrated
an absence of a relationship or ambiguous results, while


CHAPTER I
INTRODUCTION
An important goal of this nation is to provide an
effective, efficient health care system for its citizens.
To accomplish and maintain such a system requires that
shortages and surpluses of health care providers
particularly physiciansbe avoided. When considering
supply and requirements for physician manpower needs, the
focus must not be limited solely to aggregate numbers but
must take into consideration the distribution among
specialties of physicians, also.
The Summary Report of the Graduate Medical Education
National Advisory Committee (GMENAC) (U.S. Department of
Health and Human Services, 1980) cited imbalances in
physician specialty areas as a major health care problem
facing the nation. There is general consensus that there
are too many physicians in some specialty areas and not
enough in others. Using projected manpower needs for 1990,
surpluses are expected to occur in nonprimary care areas
such as surgery, while many of the primary care specialties
will fall short of the projected requirements (Jacoby,
1981) .
Neither is the maldistribution across specialties
likely to abate in the near future unless some action is
-1-


-7-
9. There is no difference between groups of medical
students identified by specialty preferences with regard to
their attitudes toward primary care.
10. There is no difference between groups of medical
students identified by their fourth year specialty choice
with respect to their attitudes toward primary care before
and after a family medicine clerkship.
The .01 level of significance was used for rejection of
the null hypotheses.
Background and Justification
Achieving a better balance between physicians in
primary care specialties and those in other specialties has
become one of the major concerns of medical educators,
health care policy makers, and others vitally interested in
the proper functioning of the health care industry.
Approximately 80% of the physicians in this country are
specialists. When the ratio of generalist to specialist
(primary care to nonprimary care) is imbalanced several
consequences are likely to result: (a) cost of medical care
increases due to the high dependence on specialists for
primary care, (b) quality of health care may actually be
compromised from iatrogenic complications resulting from
unnecessary procedures and physicians practicing outside of
their areas of expertise in order to maintain their income,
and (c) doctors' job satisfaction could suffer as a result


-39-
Rezler (1974) concluded from his review that attitude
changes resulting from a special program will fade unless
they are reinforced by the total environment.
One suggested strategy for examining the influences of
the medical education environment on specialty choice is
based on the immediacy of their impact on the students (U.S.
DHHS, 1980). First-order effects have a direct influence on
the knowledge, skills, values, attitudes, and interests of
the students. The primary example of first-order effects
are the role models, teachers, and attending physicians.
Also included in this category are the opportunities for
role playing and practicing the acquired knowledge, skills,
values, and attitudes.
Second-order influences are filtered through the first-
order effects and are composed of the organizational and
institutional components which determine the type of agents
and opportunities available for the socialization process.
Third-order and all subsequent influences are those which
are further removed and are filtered through the first- and
second-order effects. The effects of these influences are
so indirect and confounded by other variables that it
precludes definitive conclusions (U.S. DHHS, 1980).
Following a logical sequence, first-order effects will
be considered first.
An exemplary example of how difficult it is to
introduce a major innovative experience into a traditionally
structured medical school curriculum was demonstrated in the


-108-
£ = .004. The preliminary goal rankings for this goal were
compared with attitude data obtained in follow-up, the r
value was .29, £ = .006. After the clerkship, the
correlation coefficient between career goal rankings and
attitude scores was .31, £ = .004. The sign of the
correlation coefficient was in a positive direction so that
the importance of this goal increased as the attitude
scores decreased. Conversely, students with favorable
attitudes toward primary care ranked this goal lower in
importance.
Students ranked the extent to which they perceived
"providing me with capital for investment" an important
product of their medical career (Set 2, Goal a, Table 13).
The correlation coefficient for the follow-up data was .34,
£ = .001. Based on the positive sign associated with the
correlation coefficient, it can be assumed that those
students who held this goal to be important had a lower
attitude score.
Following the family medicine clerkship, the importance
of "actively conducting research in my areas of interest"
was significantly related to attitude scores, r = -.31,
£ = .003 (Set 2, Goal f, Table 13). The direction of the
correlation suggested that as the importance of this goal
increased so did the positive attitude toward primary
care.


-86-
practice, and only 15% indicated that they had not made a
decision. Thus, the greatest differences were in the
primary care and undecided medical preference levels for the
two groups. Almost twice as many students whose fathers
were in business/other indicated primary care preferences as
did those from the professional/technical group. A much
larger percentage of those from the professional/technical
group indicated a delay in decisions on specialty choice.
The importance of this large percentage of undecided
students from the professional/technical background can be
seen in the results from testing Hypothesis 2 when the
actual specialty choice was made in the fourth year (Table
6). While the percent choosing residencies in subspecialty
areas increased, the most dramatic gain was made in the
professional/technical group where the percentage of
students selecting primary care went from 33.3% to 62.5%.
Parents' annual income. The findings related to
parents' annual income and third-year specialty preferences
were consistent with the data on fathers' occupations. The
parents' annual income question asked the students to choose
from among four income categories: (a) less than $10,000,
(b) $10,000 to $25,000, (c) $25,000 to $50,000, or (d) above
$50,000. Two income groups were formed from these four
categories for comparison with the three specialty
preference levels. The two income groups were divided
between responses above and below $50,000. The chi-square


-90-
In the comparisons of fourth-year medical students'
specialty choices with their sociodemographic
characteristics, the overall significance level of _< .01 was
not achieved. It was concluded that there was insufficient
evidence to reject the null hypothesis.
One sociodemographic variable, college major, was sig
nificantly related to the medical students' specialty choice
at the £ < .01 level of significance. A second comparison
using the sociodemographic variable, parents' annual income
reached a significance level of .03. For these two
variables, a discussion of the results is as follows.
Parents' annual income. The parents' annual income
category asked the students to indicate their parents'
combined income by checking one of four income categories.
As in Hypothesis 1, two income groups were formed from the
four categories for comparison with the two medical
specialty choice levels, primary care and subspecialty
practice. The two income groups were (a) greater than
$50,000 annual income and (b) annual income equal to or less
9
than $50,000. The resulting chi-square statistic was X (df
1, N = 99) = 4.602, £ = .03. The percentage of higher level
income students choosing primary care and subspecialty
residencies was nearly equal with approximately 52% choosing
primary care and 48% going with subspecialty residencies. A
considerably higher percentage (74%) of students from the
$50,000 or less income group chose primary care than chose
residencies in subspecialty areas (26%). The results are
presented in Table 8.


-134-
2. Revise the item pertaining to Finances to
accurately reflect current economic conditions. The dollar
amount of the categories of indebtedness are too low to
provide accurate data on students' educational debts.


-63-
blocks sharing the other two weeks with a neurology
clerkship. The Community Health and Family Medicine
clerkship ran concurrently with the students' clinical
experience in medicine, pediatrics, psychiatry, surgery, and
ob/gyn.
In the Community Health and Family Medicine rotation
all clinical experiences were in settings outside of the J.
Hillis Miller Health Center and Shands Hospital. The six
weeks were scheduled as follows: four weeks in a family
practice center (either the Family Practice Center in
Gainesville or St. Vincent's Family Practice Center in
Jacksonville) and two weeks in a rural health clinic either
in Cross City or in Mayo, Florida. Participation in the
various community settings provided the students with first
hand experience in the medical and health problems as they
exist in the different communities.
The rotation year begins in March each year and
continues for 12 months. The schedule for the 1983-84
rotation year was as follows:
Rotation 1
Rotation 2
Rotation 3
Rotation 4
Rotation 5
Rotation 6
March 21-May 14, 1983
May 15-July 9, 1983
July 10-Sept. 3, 1983
Sept. 11-Nov. 5, 1983
Nov. 6, 1983-Jan. 14, 1984
Jan. 15-March 9, 1984
Approximately three weeks prior to the beginning day of
the rotation, an orientation meeting was held for the
students. The goals and objectives were explained, the
course syllabus was distributed, and the evaluation system
was discussed. The sequencing of the various clinical


-97-
postclerkship period, the observed test statistic was H =
8.27 (d_f 1, N = 86), £ = .004. The mean rank score for the
group supported by parents was 31.06 and 48.31 for the other
group. Thus, it was concluded that students whose parents
were their major source of support had significantly lower
attitudes toward primary care than did students who derived
their major support from other sources.
Parents' income. Closely related to the source of
support was the parents' combined annual income item. Two
income categories were used in the comparison analysis with
attitudes. The income categories were (a) students whose
parents had annual earnings of $50,000 or less and (b) those
whose parents earned over $50,000. The preclerkship data
produced a significance level of .19. However, the
postclerkship test statistic was H (df 1, N = 79) = 4.19,
£ = .04. The mean rank score of the $50,000 or less group
was 43.38 while those from families with an excess of
$50,000 annual income had a mean rank score of 31.76. From
these data it was concluded that students from the higher
income group had less positive attitudes toward primary care
than those from the $50,000 or less income group.
Family medical care provider. In response to the
question Who provided primary care for your family? students
had the choice of (a) family/general practitioner, (b)
pediatrician, (c) internist, and (d) other. No differences
in preclerkship attitude scores were noted among the groups.
Postclerkship attitude scores were different (£ = .05) for


-132-
on attitudes, goals, and specialty choice can be discerned
taking into account the rotation sequence of each
subgroup.
Students draw from role models and clinical experiences
in the career decision process. If these decisions are to
be made from a broad base of information of all specialties,
students need a clear idea of how each is practiced
including how primary care is conducted by practicing family
physicians. In this study, a subset of students spent two
weeks of the third-year family medicine clerkship under the
supervision of family physicians in private practice, living
with these physicians in their homes. The role-playing
opportunities in this experience and the role modeling of
the preceptors is very different from the other educational
activities in medical school. The power of such an
experience to exert an influence on students' attitudes
toward primary care and thus, their career choice deserves
to be explored. It is recommended that in subsequent
replications, the influence of this experience be
evaluated.
While the choice of a residency represents the decision
to specialize, in reality these decisions are prolegomenous
in some cases. For example the majority of students
choosing internal medicine residencies do not stay in
general internal medicine but rather proceed to subspecialty
areas after the first year. Even practicing physicians


-50-
factors has yielded little help with regard to prediction of
specialties. Personality tests have distinguished between
psychiatrist and surgeons but have been less helpful with
regard to other specialties. Academic measures available on
admission become less useful over time being more predictive
of early medical school performance but not later clinical
abilities. Little difference has been found in recent
studies of academic performance between those choosing
primary care and others.
Recognizing the limitations of the current admission
data, a viable alternative to the present admission process
was offered in the Summary Report of the Graduate Medical
Education National Advisory Committee (U.S. DHHS, 1980):
An idea which has been put forth as a solution to
another problem, that of the inevitable
disappointment of qualified applicants not being
accepted to medical schools because of heavy
competition from many other equally qualified
applicants, might also be a solution to the issue
of personality screening for admission to medical
schools. The idea is that of a two stage
screening process, the first stage involving
screening on the basis of intellectual and
academic qualifications, the second involving a
random lottery. Such a process would assure that
acceptees would (a) be intellectually qualified,
(b) have an equal chance for selection at the
second stage (with "fate" making the final
decision and thus shouldering the blame for
nonacceptance) and (c) display personality traits
in proportions representative of those occurring
in the population of qualified applications.
(Vol. V, p. 17)
Sociodemographic variables have provided some limited
insight into specialty selection. Size of home community,
age, and marital status have been most consistent of the
background variables in distinguishing primary versus


APPENDIX A
MEDICAL STUDENT ENTERING QUESTIONNAIRE


DEPARTMENT OF COMMUNITY HEALTH AND FAMILY MEDICINE
COLLEGE OF MEDICINE
UNIVERSITY OF FLORIDA
(904) 392-4321
BOX J-222, MSB
J. HILLIS MILLER HEALTH CENTER
GAINESVILLE, FLORIDA 32610
MEMORANDUM
TO: Phase B Medical Students
FROM: William L. Stewart, M.D.
RE: Clerkship Evaluation
The Department of Community Health and Family Medicine is
very interested in your feedback on the rotation so that
worthwhile learning experiences can be continued and those
which are not meaningful can be revised or deleted. You
are requested to complete the evaluation form attached to
this memo.
To assure your anonymity, do not sign your name to the
evaluation form. When you have completed the forms place
them in the envelope and seal the envelope. The sealed
envelope should them be sent to Darlene Hood, Phase B
Secretary. The envelope with your name in the upper left
hand corner will be discarded. Your name appears on the
envelope to enable her to follow-up on those who have not
returned their evaluation form. Faculty members responsible
for evaluation of students do not see the individual evalu
ations. Your grade will not be forwarded to Dean Hill's
office until the rotation evaluation has been completed and
received in the Phase B Secretary's office.
At the orientation to the Community Health and Family Medi
cine Clerkship you participated in a longitudinal study on
attitudes toward Family Medicine. Please assist us by com
pleting the follow-up portion which is attached to this memo.
Do not write your name or social security number on the form.
Return the completed form in the enclosed envelope. Thank
you for your participation.
WLS/mls
Enclosure
-14 5-
EQUAL EMPLOYMENT OPPORTUNITY/AFFIRMATIVE ACTION EMPLOYER


-126-
students' attitudes toward primary care, attitude scores
remained high for primary care students and even improved
for the subspecialty group. Thus, the clerkship could be
considered an important component of the training program to
maintain student interest in primary care and to improve the
understanding and attitudes of subspecialty students toward
the practice of primary care.
The significant association between medical students'
third-year preferences and their fourth-year choice of
residency was in contrast to what could be expected given
the exploratory state of the career decision process in the
third year. A high percentage (75%) of students preferring
primary care actually chose primary care residencies.
Similarly, students going into subspecialties were stable
from third to fourth year (85%). The undecided group in the
third-year split 57% to 43% in favor of primary care. Since
this group comprised almost one-third of the total number in
the sample, the direction of their choices was an important
factor in the results. The findings of this analysis
suggest that third-year preferences are fairly stable for
those who indicate a specialty area but the one-third who
are uncommitted are a potential pool through which a
substantial impact could be made on specialty
maldistributions


-33-
sample size of this study precluded meaningful statistical
analysis but several important trends were identified using
three factors, Orientation to Patient Care, Orientation to
Work Conditions, and Orientation to the Profession.
Students choosing primary care careers scored higher on the
factor "Orientation to Patient Care," indicating more
concern for people and less orientation towards the
profession. More unsettling was his finding that over the
four years of medical school, primary care respondents
exhibited a shift in their attitudes away from concern for
patient care toward a somewhat greater self-concern.
Plovnick explained this trend as the general socialization
influence of medical school which might be changed by a
further differentiation of programs designed for primary and
tertiary care providers.
A widely used instrument for investigating personality
differences and career choices is the Myers-Briggs Type
Indicator (MBTI). Based on the personality theory of Jung,
it compares four categories on a continuum of opposites:
extroversion-introversion, thinking-feeling, sensing-
intuition, and judgment-perception. These type differences
are related to differences in preferences for activities,
interests, personal outlooks, and choices of occupation. In
the early 1950s, Myers tested over 5000 medical students.
Those still in practice were followed up in 1973 (McCaulley,
1981) and confirmed that their specialty choices were
similar to those chosen at graduation. Harris, Kelly, and


-80
Table 2
Medical Specialty Preferences
of Third-Year
Medical
Students
Specialty Preference Groups
n
Percent
Primary care
34
47.2
Specialty-subspecialty
17
23.6
Undecided
21
29.2
Total
72
100.0


-83-
Table 4
Association between Third-Year Specialty Preference
and Sociodemographic Characteristics
Sociodemographic
Characteristics
df
X2
Vaue
Significance
Sex
2
5.513
.06
Marital Status
2
.946
.62
Religious Orientation
8
3.097
. 94 +
Religious Importance
4
6.027
.20
Race
2
4.010
. 16 +
Citizenship
2
1.133
1.00 +
Hometown Size
6
3.134
. 79 +
Hometown Location
2
2.218
. 39 +
Hometown Economic Base
8
6.867
. 61 +
Father's Occupation
2
7.727
. 02 + *
Mother's Occupation
2
1.312
.52
Parents' Religion
8
7.291
. 51+
Parents' Income
2
15.864
.0004***
Source of Present Support
2
4.412
.11
Amount of Debt
2
1.780
. 56 +
Indebtedness
2
3.279
. 21 +
Public or Private High School
2
3.432
. 20 +
Public or Private College
2
5.080
.08
General Health
2
3.281
. 24 +
Family's Medical Care Provider
Family Practitioner/General
2
2.257
.32
Internist
2
4.331
.11 +
Pediatrician
2
1.678
.43
Other
2
.825
. 70 +
Work Experiences
Medical
2
3.531
. 19 +
Technical (not medical)
2
.756
.69
Helping Activities
2
.705
. 70 +
Public (sales, etc.)
2
3.309
.22
Teaching
2
1.957
. 37 +
Unskilled
2
.873
.65
Other
2
2.068
. 37 +
Level of Education
2
1.788
. 52 +
College Major
2
4.013
. 16 +
*£ < .05
**£ < .01
***£ < .001
+Agresti-Wackerly
Significance Level


-37-
socially correct-. Most students consider "helping people"
as the appropriate answer rather than "economic gain" when
asked the reason to pursue a career in medicine.
Disregarding the difficulties inherent in personality
testing, there is a major problem inherent in manipulating
specialty distribution based on personality characteristics,
namely, the legality of and societal distaste for doing so
(U.S. DHHS, 1980). The subjective interpretation of the
results of personality tests will certainly not go
uncontested by those denied admission were these tests given
specific weighting in the admission process. Society would
surely reject the routine use of personality tests as the
basis of admission due to the perceived subjectivity,
unreliability, and possibility of social regimentation
implied by such action. The information gained from
personality inventories and questionnaires is most
appropriately reserved for counseling students.
Factors Related to the Medical Training System
As Zuckerman (1977) pointed out, sociodemographic and
personality characteristics present at admission become less
useful over time in predicting specialty choice, while those
factors associated with the institutional environment take
on more importance. Students are presumed to come into the
medical training setting with certain motives, values, and
knowledge. These factors are subject to change or
reinforcement by the pressures of agents or conditions


-8-
of improper use or under utilization of their skills and
knowledge (Schroeder, 1984 ).
From the post-World War II period through the 1960s and
1970s, physician manpower policies were primarily concerned
with the overall number of physicians and their geographic
distribution. In the 1940s and 1950s task forces on health
manpower needs predicted a shortage of 50,000 doctors by
1980 (Petersdorf, 1975). Health manpower legislation of the
1960s was directed toward remedying this perceived shortage
by increasing medical school enrollments, recruiting
faculty, and expanding the existing facilities. The
preoccupation with increasing the aggregate number of
physicians was so intense that the trend toward increased
specialization went largely unnoticed. From 1931 to 1974,
physicians identifying themselves as general practitioners
decreased from 83% to 18% (Harris et al., 1982).
The health manpower legislation of the 1960s was so
successful, it is estimated that by the year 2000 there will
be 145,000 more physicians than needed to provide physician
services (U.S. DHHS, 1980, Vol. I). The problem of
oversupply of physicians is not limited to the United
States, however. Virtually every Western country is facing
a current or projected excess of physicians. In European
countries the surpluses have produced varying results from a
dramatic decrease in physician income to actual unemployment
for a sizeable number of medical doctors. While the United
States shares the problem of surpluses with other countries,
compared to its Western allies, this country stands alone in


-91-
Table 8
Students and Parents'
Annual
Income and
College Maj
or
Specialty
Choices
Sociodemographic
Characteristics
Primary Care
n Row %
Subspecialty
n Row %
Parents' Annual Income
> $50,000
35
51.61
33
48.39
< $50,000
23
73.53
8
26.47
College Major
Science
65
72.22
25
27.78
Other
5
35.71
9
64.29


APPENDIX E
LETTER OF REMINDER SENT WITH SECOND
FOLLOW-UP QUESTIONNAIRE


-19-
which shares a common approach or similar variable, rather
than summarizing the studies chronologically. The factors
will be categorized into four groups: (a) background
factors, (b) variables related to personality and attitudes,
(c) factors pertaining to the medical environment or
training system, and (d) career factors.
Background Factors
A large part of the literature on specialty choice
focuses on students' background characteristics including
age, sex, marital status, hometown, socioeconomic status of
family, and educational history. Two reasons for
researchers' seeming preoccupation with these variables is
assumed to be their accessibility from students' records and
their potential use in the selection process for admission
to medical school. What must not be forgotten, however, is
that medical students are a highly selected, homogeneous
group so that differences in these background
characteristics are small while specialties and
subspecialties are quite diverse making the background
variables poor predictors of specialty choice.
Both Anderson (1975) and Zuckerman (1977) in extensive
reviews of the literature reported evidence that the larger
the community of origin the greater the likelihood of the
student choosing specialty practice over general practice.
Hutt (1976) reported similar findings in Britain where those


CHAPTER V
SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS
Summary
The nation is faced with an oversupply of physicians
and an imbalance of physicians across specialty areas. A
number of undesirable consequences result from this
oversupply and maldistribution among doctors, the chief one
being spiraling health care cost. Although a large body of
research has been generated, unequivocal evidence regarding
the causes for the trend away from general or primary care
toward subspecialization has been elusive. Interventions
introduced at the residency level have met with only limited
success in producing the proper mix of primary care and
nonprimary care physicians for the needs of society.
Because the initial specialty choice decisions are made
during medical school, the study of variables at this level
has the advantage of providing information with the greatest
potential for influencing the decision process.
The purpose of the study was to identify factors
related to medical students' third-year specialty
preferences and fourth-year specialty choices including
their sociodemographic characteristics, the relative
importance placed on career goals, and their attitudes
toward primary care. In addition, associations among
-116-


-56-
of 20 among the 6 clinical specialty services, pediatrics,
internal medicine, obstetrics/gynecology (ob/gyn), surgery,
psychiatry, and family medicine. Students work as members
of the medical patient care teams experiencing direct
patient care responsibilities.
Phase C occupies the last 15 months of medical school.
Students are required to complete four additional weeks of
surgery and four weeks of medicine clerkships. They review
clinical pharmacology, pathophysiology, and study infectious
diseases for an eleven week period of time. The last 10
months are reserved primarily as elective time which allows
students to select clinical activities of special interest
and interview at the various residency programs where
they are interested in applying for graduate medical
education.
Revision and Development of the
Questionnaire
Historical Background of the Questionnaire
The instruments used in this research project were
adapted from two survey questionnaires used to study
i
attrition from Florida's family medicine residency programs.
In the mid-to late 1970s these residency programs were
experiencing substantial numbers of residents who dropped
out after one year of training (Dallman, Crandall & Haas,
1980). In order to study the issue of attrition among its
residency programs, a cooperative project was begun in 1977


-44-
findings of the tabulation procedure revealed 1,145
different career patterns for 2,514 students sampled.
Obviously, the rigid tracking system Zuckerman hypothesized
was not supported by the findings of the study.
Friedman, Stritter, and Talbert (1978) examined closely
three community hospitals and an academic teaching hospital
where their students had clerkships. Based on the amount
and types of clinical experiences provided, they concluded
that it was fallacious to try to dichotomize facilities.
Some community hospitals paralleled teaching hospitals on
many characteristics.
There are probably multiple opportunities and
combinations of experiences at sites which might be
considered primarily clinical or strictly academic. These
two studies highlight the imprecision of the criteria used
to evaluate medical schools, not only by researchers but
also by accrediting agencies.
Medical school administrators and government policy
makers advocate founding new departments, increasing full
time faculty, and increasing research resources in an effort
to interest students in a given specialty. How valid is the
assumption that such investments produce dividends? The
Canadian study by Roos and Roos (1980) offers some insight
into this question which may be applicable to the U.S.
situation. They collected aggregate data on four charac
teristics of Canadian medical schools: number of full-time
faculty, number of part-time faculty, research grants,


APPENDIX C
LETTER OF EXPLANATION ATTACHED TO
ENTERING QUESTIONNAIRE


I 1 = Orientation g
X = Family Medicine Clerkship U11
0 = Questionnaire, Entering
and Follow-Up
(3 = Ranking of Residency
Programs g
/\ = Resident Match I x _
Choice Results
0
12
- X -
10
0.
0,
- X -
S
E
Q
u
E
N
C
E
3-
2-
I-
- X -
0
0
3 u4
dZZ
0.
o
A
o
A
o
A
o
A
o
A
o
A

M
1983
i 1
A M
1 1 i 1 1
J J A SON D
1 1
J F
1
M
-4 1 1
A M J
i i 1 1 1
JASON
1
D
1 1 1-
J F M
1985
Phase B
Phase C
MONTHS
Figure 1. Data Collection Points


-65-
data to be collated with the entering data; and (c)
participation or nonparticipation would not affect their
rotation grade as those clinical faculty members responsible
for grading would not have access to individual's responses.
It required approximately 30 minutes for students to
complete the questionnaire.
Follow-Up Survey Questionnaire
The follow-up administration of the survey
questionnaire was accomplished at the end of the rotation
time period. The students were not told that the follow-up
questionnaire was identical to the entering survey with the
exception of the background data questions. The survey
questionnaire was included in the envelope with the student
evaluation of rotation forms. Two self addressed envelopes
were included for the respondent to return the evaluation of
rotation and the survey questionnaire. A cover letter
(Appendix D) stated that the College of Medicine required
student evaluation of rotation. It reminded students that
they had participated in the entering survey and asked their
cooperation in the follow-up survey.
A follow-up code sheet was used to record the name,
social security number, and code number for each respondent.
When the follow-up form was returned an X was placed in the
appropriate column. Those students who did not respond
within two weeks were sent a note reminding them to send the
questionnaire back. If another two weeks passed with no


-43-
design, the results are interesting enough to prompt further
research.
A conclusion to be drawn from the studies of first-
order effects is that clinical teaching faculty exert a very
strong role modeling effect on decisions of specialty
choice. Experiences during the clinical rotations are an
important component in the decision process, also, but it is
probably the interplay of the two which is critical. The
findings have implications for recruitment of students into
primary care specialties.
Second-order influences are those that indirectly
effect first-order factors. These might include financing
of medical education, the structure and organization of the
medical school, and research endeavors. Given the fact that
second-order must exert force through first-order, it is
hard to see how it would impact career choice among medical
students, but a number of researchers reported interesting
findings.
Zuckerman (1978) investigated the structural factors
within the educational milieu. He hypothesized these
factors linked together to form particular patterns which
resulted in different career outcomes. Based on structural
characteristics, 28 medical schools were dichotomized into
academic and clinical categories. Students were classified
according to academic standing. The type of residency
chosen was also categorized. Finally, the type of medical
practice entered was divided into four general areas. The


-68-
The Recurrent Institutional Cycle Design (Campbell &
Stanley, 1963) as adapted for use in this study is shown in
Figure 1.
The design for this study required the collection of
preclerkship data from each medical student at the beginning
of the family medicine rotation. Following preclerkship
data collection, each rotation group received a learning
exposure to family medicine. A second administration of
the questionnaire occurred immediately following the
family medicine rotation. The final component of the
design involved residency match choices for each study
participant.
The cyclical schedule for this design was such that at
one and the same time a group which had been exposed to the
clerkship (X) and a group which was about to be exposed were
measured. As Figure 1 demonstrates, subjects in Rotation 1,
which began in mid-March, were completing the follow-up
questionnaire in mid-May at approximately the same time as
the subjects in Rotation 2 received the entering
questionnaire. Each subsequent rotation began a new cycle
as the previous rotation ended their clerkship. Students
ranked the residency programs for the National Resident
Matching Program in January of their senior year of medical
school. Residency match choice information was received at
the sponsoring institution in March, 1985.


-25-
undergraduate grades in science and the MCAT scores as
criteria for admission is questionable. The Anderson (1975)
review of the literature reported mixed results in studies
that used MCAT scores to differentiate specialty choice.
Subsequent studies have not provided more consistent data.
Gough (1978) used four measures, the MCAT science
subtest scores, premed grade point average (GPA), a
preference for scientific subjects, and a composite of the
three, in this study of 1,135 graduates of the University of
California, San Francisco, Medical School. When these
factors were used to predict performance in medical school,
all four factors were significantly correlated the first
year of medical school but the following years became less
and less predictive of performance until by the fourth year
they were completely unrelated. With regard to specialty,
the scientifically oriented students entered specialties
such as surgery, anesthesiology, and pathology; those
ranking lower went into internal medicine, pediatrics, and
psychiatry.
In his review of the literature to 1977, Zuckerman
found that students planning to enter general practice
scored low on MCATs and had low academic standing in medical
school. Conversely, those who planned academic or research
careers had high MCAT scores and high academic standing.
Those planning to specialize and go into private practice
scored in the mid range on MCATs. Other studies published
the same year as the Zuckerman review reported (Association


A LONGITUDINAL STUDY OF FAMILY PRACTICE EDUCATION
A FOLLOW-UP SURVEY
NAME: SOCIAL SECURITY NUMBER:
CAREER GOALS
1.Each of the phrases below describes an iooortant goal
that sone people want to achieve wrougn a career in
medicine. Please rank your personal goals by putting
a l next to your most important goal, a "2" next to
your second most important goai, ate. (Your least
imoortant goal will be ranxed "5"' X NOT USE ANY
NUMBER (1-6) MORE THAN ONCE!
a. Helping lead my community to the solution
of social and medical problems.
b. Gaining the respect anc friendship of
other physicians in my ccmnunity.
c. Developing long-term intensive relation
ships with my patients.
d. Contributing to nedicai knowledge through
research.
e. Assuring financial security for myself
and my family.
f. Allowing me to live in the type of area
and lifestyle that I prefer.
2.Please rank order the following list of goals in the
same manner (1-6). The goal of greatest personal
importance to you should be ranxed "1".
a.
b.
Providing me with capital for investment
in business, real estate, the stock market,
etc.
Being acknowledged by ttner pnysicians as a
leader in health care *rr my patients.
Having adeouate time *zr my amily and for
non-orofessional activities.
Gaining the trust and confidence of my
patients.
Making me a respected ^emoer pf my
comnuni ty.
Actively conducting research in my areas
of interest.
MEDICAL EDUCATION CONTENT
1. Each of the following items may be an aoorooriate part
of medical education. Please ranx them from "1 (most)
to 8 (least) according to how iaoortant you believe
that each is in producing the best physician.
a. Course work in the basic sciences-.
b. Clinical experiences with nospitalized patients.
c. Clinical experiences with ambulatory patients.
d. Course work in medical ethics.
e. Training in the business aspects of medical
practice.
f. Training in public and coemunity health.
g. Research into the causes ano treatments of
disease.
h. Research into patient behavior and compliance with
medical care.
O ATTITUPES ABOUT PRIMARY CARE AND FAMILY MEDICINE
(Circle the numoer or the response wnich oest describes
your own personal feelings toward each statement.)
o
a w
C V V
O il T5 31
U. L. CL. _
** o* "O C l/
1. The diagnosis and treatment I 2
of illnesses in nosoitalized
patients is more difficult
than in similar problems in
amoulatory patients.
2. Ooctors who have large 1 2
inpatient practices tend to
be better doctors than those
with small inpatient practices.
3. The family physician can provide 1 2
his greatest service in following
long-term health and adjustment of
patients rather than in concentrating
on the treatment of their immediate
complaints.
4. A wide variety of problems 1 2
enconoassing all age groups is
interesting to me.
5. In meo.cal practice today, there 1 2
are sufficient aoorooriate
soecialists so that a family
physician should not assune long
term responsibility for patients
with chronic illnesses.
6. I feel that preventive medicine 1 2
is more important than curative
medicine.
7. I believe that the social and 12
family environment of patients is
a major influence on their state
of health.
8. Medical education should concentrate 1 2
primarily on recognition and treat
ment of specific disease processes.
9. A patient's ability to pay should 1 2
influence quality of treatment
given.
10. Epidemiological and preventive 1 2
medicine research is interesting
to me.
11. A primary care physician should 1 2
use consultants for managing
critically 111 patients.
12. An fictional upset should be as 1 2
valid excuse for tsissing work
as a bad cold.
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
3 4 5
-140-
tend to
disagree


Table 12
Continued
Preliminary Goal
Rankings
Follow-Up Goal
Rankings
Career Goals4"
H
£
H
£
Set
2
a.
providing me with capital for investment
.23
.63
.01
.93
b.
being acknowledged by other physicians as
a leader in health care for my patients
.80
.37
.53
.47
c.
having time for family and non-professional
activities
2.30
.13
.11
.74
d.
gaining trust and confidence of my
patients
.17
.68
.06
.81
e.
making me a respected member of my
community
1.33
.25
. 26
.61
f.
actively conducting research in my areas
of interest
.54
.46
.09
.77
+Paraphrased from original questionnaire
-103-


-4-
Is there a difference in career goals between groups
of medical students identified by third-year specialty
preferences or fourth-year specialty choice?
What is the relationship between the importance medical
students attach to career goals and their attitudes toward
primary care?
Is there reason to believe that career goal rankings
before and after a family medicine clerkship are different
for medical students choosing primary care from those
choosing subspecialties in the fourth year of medical
school?
Is specialty preference expressed in the third year of
medical school associated with medical students' specialty
choice decisions in the fourth year as evidenced by
residency selection?
Y Are medical students' attitudes toward primary care
changed after a clerkship in family medicine?
Is there a difference in attitudes toward primary care
between groups of medical students identified by third-year
specialty preferences of fourth-year specialty choice as
evidenced by residency selection?
Purpose of the Study
The primary purpose of the study was to identify
factors related to medical students' third- and fourth-year
specialty choices and to determine the extent to which a
clerkship in family medicine affected the students'


-127-
Implications of the Study
For those interested in increasing the pool of primary
care physicians, implications which may be drawn from the
findings of this study include the following:
1. Students whose parents' annual income is $50,000 or
less and those who derive their support from their parents
tend to be attracted to primary care specialties. Loans and
scholarships should be provided to encourage and support
students from less affluent families.
2. Medical students tend to be fairly homogeneous in
background characteristics. The admission criteria should
be generalized to gain more diversity in the
sociodemographic characteristics of medical students.
3. Nearly one-third of the students in this study were
undecided as to their specialty preferences in the third
year of medical school. Development of a formal structure
for facilitating career guidance of medical students is
essential so that career decisions are not based on limited
experience and information.
4. Clinical experiences in general primary care
specialties should be emphasized in the curriculum,
particularly early educational training activities in
ambulatory care centers in order to maintain interest in
primary care areas and to improve attitudes and knowledge
about primary care specialties.


-113-
Hypothesis 8
Hypothesis 8 examined whether or not a clerkship in
family medicine modified medical students' attitudes toward
primary care. Attitudes were measured before and after an
8-week family medicine clerkship in the third year of
medical school. Using a 5-point Likert Scale, the medical
students responded to 29 statements of attitudes and beliefs
about the practice of medicine and medical care. After each
administration of the questionnaire, an attitude score was
obtained for individual students. In order to test the
hypothesis, a difference score was calculated for each
subject by subtracting the preclerkship attitude score from
the postclerkship score. The Wilcoxon matched-pairs signed-
ranks test was selected to test the significance in the
difference scores between the two related samples. The
significance level was set at the _< .01 level. A Z score of
-2.008 resulted from the analysis. The probability
associated with the occurrence under a null hypothesis of a
Z as extreme as -2.008 has a probability of .0444. In as
much as this probability was greater than the established £
< .01, the decision was made not to reject the null
hypothesis and to conclude that a family medicine clerkship
does not affect attitudes toward primary care at the £ .01
level of significance.


-35-
Seventy-three percent of this group expressed interest in
primary care and 54% were interested in rural health. This
represents a higher percent than might be expected compared
to national statistics. Steinwachs and his colleagues
(1982) in their analysis of recent trends in graduate
medical education reported that approximately 55% entering
residency in 1980 chose primary care.
In a national survey of third-year students, Burkett
and Gelula (1982) asked students to assign relative values
to four motivating factors in their decision to enter
medical school. Those students indicating a preference for
primary care attributed greater importance to the desire to
help people than to the other three choicesdesire for
financial reward, desire for social status, and desire to
apply scientific ability. They also demonstrated a tendency
to consider the sociopsychological context of patient
problems and felt a need for change to improve health care.
According to Burkett and Gelula (1982), "the overall pattern
appears to confirm the notion of primary care as a 'person-
centered' health care field which attracts students who have
a more 'holistic' orientation" (p. 512).
Robbins et al. (1983) studied gender differences in
interest and motivation for a career in medicine. Not
surprising, both men and women scored highest on the medical
science category of the Strong-Campbel 1 Interest Inventory.
Both sexes were similar on a projective technique where
90% were judged to fear success. On the attitudinal


-121-
In summary, evidence generated from the analysis of
data from this study was insufficient to reject all but one
of the 10 hypotheses. For the one null hypothesis
rejected, a significant association (£ < .01) was found
between medical students' third-year specialty preference
and their fourth-year residency selection (£ = .001).
Eighty-five percent of students who preferred primary care
specialties in the third year actually chose primary care
residencies in the fourth year. Similarly, students
indicating a preference for subspecialty practice maintained
their position in the fourth-year residency choice. Almost
one-third of the students were undecided as to their
specialty preferences when sampled in the third year,
however. The implications of these results are that the
majority of students are capable of making stable career
decisions at least as early as the third year. Those
medical educators and health policy makers interested in
impacting on the specialty decision process should consider
early interventions.
Conclusions
This investigation was initiated to identify factors
related to medical students' third-year specialty preference
and fourth-year specialty choice as well as the effects of a
family medicine clerkship experience on career goals and
attitudes toward primary care. The aim of the research was
to contribute to the base of knowledge in the area of


-94-
Table 9
Sociodemographic Groupings of
Medical
Students
and
Preclerkship Attitudes
Sociodemographic
H Significance
Characteristics
df
Value
Sex
1
2.02
.16
Marital Status
1
1.47
.22
Religious Orientation
4
.98
.91
Religious Importance
2
5.84
.05*
Race
1
.49
.49
Citizenship
1
1.52
.22
Hometown Location
1
4.45
.03*
Hometown Size
3
.43
.93
Hometown Economic Base
4
3.89
.42
Father's Occupation
1
.00
.98
Mother's Occupation
1
.37
.54
Parents' Religion
4
.29
.99
Parents' Income
1
1.72
.19
Source of Present Support
1
2.42
.12
Indebtedness
1
.76
.38
Amount of Debt
1
.36
.55
General health
1
.40
.53
Public or Private High School
1
.59
.44
Public or Private College
Family's Medical Care Provider
1
.01
.92
Family Practitioner/General
1
.70
.40
Internist
1
.68
.41
Pediatrician
1
1.44
.23
Other
1
.43
.51
Work Experiences
Medical
1
.44
.51
Helping Activities
1
.84
.36
Technical
1
.22
.64
Teaching
1
3.89
.05*
Public
1
7.52
.006**
Unskilled
1
.14
.71
Other
1
. 28
.60
College Major
1
.18
.68
Level of Education
1
2.17
.14
*£ < .05
**£ < .01


-130-
preference and choice. The cost of medical education is
increasing dramatically and research on these financial
aspects needs to be elaborated.
The cost of medical training is conceivably linked to
other background factors such as initial wealth, parents'
support, marital status, age, sex, and others. It is
recommended that the interaction of these background
variables be investigated. The continued study of
background characteristics is warranted in view of the
fact that no studies as yet have settled the issues with
respect to these variables and it is possible that with
increasing numbers of women and other previously less
frequent groups the overall picture of the profession may be
clarified.
Attitudes and Goals
It is not apparent from previous research how students'
goals and attitudes relate to their specialty choices or how
they relate to students' background characteristics and
educational experiences. In this study, two of the
sociodemographic characteristics of the medical students
were significantly associated with their attitudes toward
primary care even though the significance level of .01 led
to retention of the null hypothesis. For these subjects,
attitudes did not change sufficiently following an


-58-
Analysis of the residency survey data demonstrated that
a number of the questionnaire items discriminated between
dropouts and nondropouts at the end of one year of family
practice residency training. Dallman and his group (1980)
found that while sociodemographic characteristics such as
sex, marital status, age, race, citizenship, and hometown
location and size did not discriminate between the dropouts
and nondropouts, source of financial support did
differentiate the two groups. Those residents who were
their sole support were more likely to leave training than
those who drew financial support from spouses or others
(Dallman et al., 1980, p. 834). The researchers also found
that the rank order of career goals differed between the two
groups with dropouts more interested in economic security
and time for family and leisure activities. The attitudes
section showed a significant difference between the dropouts
and nondropouts also. Those who left the programs felt
curative medicine was more interesting than preventive or
health maintenance (Dallman et al., 1980, p. 834). Based on
the significant findings from the residency study and the
imperative need for understanding of the factors influencing
choice of family medicine and primary care specialties, it
was decided to introduce the questionnaire earlier in the
educational process to attempt to identify characteristics
of those likely to choose these specialty areas over others.
Thus, the present study concerning identification of factors
related to student choice of a medical specialty was begun.


-3-
the proper mix of primary care and nonprimary care
physicians for the needs of society. When the medical
education system has been targeted, modification has
occurred at the residency level rather than undergraduate
medical education. Establishing policies and finding
solutions has been hampered because factors influencing
specialty choices are complex and not well understood.
Although a large body of research has been generated
the specialty choice research literature
suffers from major inadequacies in providing
guidance for policy-makers and only rarely can
provide unequivocal evidence regarding various
influences on physician specialty choice. (U.S.
DHHS, 1980, Vol. V, p. 2)
Statement of the Problem
The problem addressed in this study is the need for
research-substantiated information on the factors related to
student choice of medical specialties as stated broadly in
the following questions:
Are sociodemographic characteristics of medical
students associated with third-year specialty preferences or
choice of specialty as evidenced by selection of a primary
care or nonprimary care residency in the fourth year of
medical school?
Is there a difference between medical students grouped
by sociodemographic characteristics with regard to their
attitudes toward primary care and family medicine?


-137-
OBLIGATIONS 4N0 a LANS
1. Obligations to Federal Agencies
a. Do you expect to owe "time" to any organization in return
for scholarships or loans in medical school?
Yes No
E IMPROVING HEALTH STATUS
Below are Usted several different approaches to improving
the health status of people not no* receiving adequate
health care. Rank each of them 1 to 5 in order of your
opinion of their prooaole effectlveness, with 1 most
effective and 5 least effective. (Use each number only
once).
b. If yes, which of the following agencies?
Armed Services
National Public Health Service
Home town or state sponsored agency
Other (specify)
c. If yes, how much time?
1 year 2 years 3 years
4 years Other
(specify)
2. Plans for the Future
Solo practice
Two man partnership
Group practice
Emergency room
Academic medicine
Armed Forces
National Public Health Service
Other (specify)
;er GOALS
. Each of the phrases below describes an important goal that
some oeoole want to achieve througn a career in medicine.
Please ran* your personal goals by cutting a "l" next to
your most important goal a "2" next to your second most
important goal, etc. (Your least important goal will be
ranxed "6") DO NOT USE ANY NUMBER (1-6) MORE THAN ONCE!
a. Helping lead my conmunity to the solution of
social and medical proolerrs.
b. Gaining the respect and friendship of other
pnysicians in my community.
c. Developing long-term intensive relationships
with my patients.
d. Contributing to medicai e. Assuring financial security for myself ana my
farm ly.
f. Allowing me to live in the type of area and
lifestyle that I prefer.
2. Please ran* order the following list cf goals in the same
.manner (1-6). The goal of greatest personal importance to
you should be ranked "1".
b.
c.
d.
e.
f.
Providing me with capital for investment in
business, real estate, the stock mancet, etc.
Being acknowledged by other physicians as a
leader in health care for my patients.
Having adequate time for my family and for
non-professional activities.
Gaining the trust and confidence of my patients.
Making me a respected member of my community.
Actively conducting research in my areas of
interest.
a. Improving housing conditions
b. Raising the general level of education
c. Improving public transportation to health care
facilities.
d. Expanding medical services (e.g., more private
doctors, more clinics)
e. Providing more jobs for the uneroloyed
ATTITUDES ABOUT PRIMARY DARE AND FAMILY MEDICINE
(Circle the ouno'er of the response wmcn best
describes your own personal feelings toward each
statement.)
The diagnosis and treatment
of illnesses in hosoitalizeq
patients is more difficult
than similar problems in
ambulatory patients.
Doctors who have large
inpatient practices tend to
be better doctors than those
with small inpatient practices.
The family physician can provide
his greatest service in follow
ing long-term health and adjust
ment of patients rather than in
concentrating on the treatment
of their irmediate complaints.
A wide variety of proalems err-
comoassing all age groups is
interesting to me.
In medical practice today, there
is sufficient aopropriate
specialists so mat a family
physician snould not assume long
term responsibility for patients
with cnronic illnesses.
I feel that preventive medicine
is more important than curative
medicine.
I bel i ove that the social and
family environment of patients
is a major influence on their
state of health.
Medical education should concen
trate primarily on recognition
and treatment of specific
disease processes.
h
Il II II
11
12 3 4
12 3 4
12 3 4
12 3 4
3 4
12 3 4
12 3 4
12 3 4
D MEDICAL EDUCATION CONTENT
1. Each of the following items may be an appropriate part of
medical education. Please rank then from 1 (most) to ''8"
(least) according to how important you believe that each is
in producing the best physician.
9. A patients ability to pay
should influence quality of
treatment given.
10. Epidemiological and preventive
medicine research is interesting
to me.
b.
c.
d.
e.
f.
Coursework in the "basic sciences"
Clinical experiences with hospitalized patients.
Clinical experiences with ambulatory patients.
Coursework in medical ethics.
Training in the business aspects of medical practice
Training in public and coemunity health.
Research into the causes and treatments of disease.
Research into patient behavior and compliance with
medical care.
11. A primary care physician should
use consultants for managing
critically ill patients.
12. An emotional upset should be as
valid an excuse for missing work
as a bad cold.
12 3 4
12 3 4
12 3 4
12 3 4


-54-
Medical Center situated across the street from the Health
Center. In addition to these sites, the College of Medicine
has developed educational programs in community settings
where students can gain insight into the day-to-day problems
of minor and major illnesses as they occur in urban and
rural settings. While university medical centers such as
Shands Hospital play a unique and indispensable role in the
student's educational experiences, fewer than 5% of
physician/patient contacts occur in hospitals (White,
Williams, & Greenberg, 1961). The Association of American
Medical College's Panel on General Professional Education of
the Physician (GPEP) (1984) recommended that medical
students' general professional education be a balance
between acute illnesses and working with patients and
communities to prevent or ameliorate disease (p. 16). This
requires ambulatory settings for required clerkships.
The community health experiences available to University of
Florida medical students include preceptorships with
practicing physicians throughout the state of Florida, a
rural health clinic in Dixie County, Florida, and family
practice centers. The preceptors are primary care
physicians who have volunteered to accept students for a
specified period of time, usually two or four weeks of
training. Generally the students work in the physician's
office seeing patients with the physician. The student
accompanies the doctor on hospital rounds and becomes
involved in all medically related activities. Often the


-26-
of American Medical Colleges, 1977; Hadley, 1977) that
students with high undergraduate grade point averages and
high MCAT scores appear to be predisposed toward
subspecialty areas rather than primary care.
More recent studies of academic variables indicate that
unlike the students in years past who went into general
practice, students choosing primary care or family medicine
perform at least as well if not better than those tending
toward subspecialty areas. In a study of trends in senior
students' interest in primary care specialties at Jefferson
Medical College, Herman and Velosk (1977) found that
students interested in family medicine performed as well or
better on measures of basic science knowledge and subjects
emphasizing clinical information as those students
interested in all other specialties except internal
medicine. Similarly, Collins and Roessler (1975) found no
significant differences in intellectual characteristics as
judged by MCAT scores and undergraduate GPA between family
medicine oriented students and any other group in their
study of Baylor Medical College students. Results of a more
recent study of a large national sample of medical students
(Burkett & Gelula, 1982) added additional weight to the
previously cited findings. In a comparison made between
primary care versus nonprimary care fields, students in the
two groups were similar on the basis of measures of
premedical school academic performance and MCAT scores.


-49-
medical care is administered by primary care facilities
(Neuhauser, 1983). It follows then that there will be a
demand by the hospital corporations for more primary care
physicians to staff their hospitals. Subspecialists will be
used sparingly and only on consultation. This change in the
organization of the health care system is likely to have an
effect on specialization.
Some of the career factors discussed here are amenable
to manipulation. Working conditions can be changed to be
more responsive to the needs of those seeking primary care
specialties in order to increase their numbers. Societal
factors are not so easily defined and interventions are
difficult to implement. Economic incentives can be
addressed in several ways. Decreasing the discrepancy
between primary care and nonprimary care specialties would
help offset this factor. The area needs more study upon
which to base policy changes.
Summary of the Review of Literature
Although the specialty choice research literature is
voluminous, only rarely does unequivocal evidence of
specific influences on physician specialty choice surface.
The conceptual differences and methodological problems make
comparisons across studies difficult.
Since admission criteria have not been well defined,
the weight given to academic and personality variables by
admissions committees is unclear. The study of personality


-88-
statistic, X^ (d_f 2, N = 65) = 15.864, £ = .0004, led to the
conclusion that there was a difference in preference for
medical specialties of these third-year medical students
based on parents' income.
Of the students from the more affluent families (i.e.,
annual incomes greater than $50,000), 20% preferred primary
care, 55% indicated that they wanted to pursue subspecialty
practice, and 25% were undecided. Those students whose
parents' annual income was equal to or less than $50,000
chose primary care more often than any other of the
preference levels. Sixty-two percent chose primary care,
11% preferred subspecialties, and 27% were undecided. The
percentage preferring primary care was three times greater
among the lower income students. A wider difference between
the income groups occurred in their preference for
subspecialties. The percent of undecided students was
fairly consistent for both groups. These data are presented
in Table 5.
Hypothesis 2
There is no association between sociodemographic
characteristics of medical students and their medical
specialty choices as evidenced by fourth-year residency
selection. The procedure for analyzing this hypothesis was
the chi-square test of significance. The results of the
analysis for Hypothesis 2 are listed in Table 7.


-117-
the variables, sociodemographic characteristics, career
goals, and attitudes toward primary care were explored.
Another intention of the study was to examine the extent to
which a clerkship in family medicine affected the students'
attitudes toward primary care. Included in the aims of the
study was the investigation of the association between
third-year career preferences and fourth-year specialty
choice.
The sampling frame for the study included all medical
students in the 1985 graduating class of the College of
Medicine, University of Florida. The class contained 117
students of which 109 participated. The students were
sampled in their third year of medical school because they
were far enough along in the educational process to have
begun exploration of self and specialties, the family
medicine clerkship occurs in the third year, and the fourth-
year students are geographically scattered.
A multidimensional questionnaire was administered to
third-year medical students immediately before and after a
family medicine clerkship. The follow-up questionnaire was
identical to the entering form with the exception of
exclusion of the sociodemographic data section and the
question related to specialty preferences which appeared on
the initial questionnaire. On the initial survey form the
students indicated whether they preferred a primary care
specialty, a subspecialty area, or were undecided. The


-71-
test is a statistical technique that tests the difference
between what is observed and what might be expected under
some theoretical model. It can be applied to one-sample
cases, two-sample cases, or larger cases such as the three
subsamples of specialty preference contained in this
hypothesis. The three specialty preference response
variables were primary care, subspecialty, and undecided.
Chi-square is an appropriate statistical tool when
frequencies in classified categories constitute the data of
research such as the sociodemographic characteristics of
this study.
2. The chi-square test for two independent samples was
applied to Hypothesis 2 to determine if significant
associations existed between sociodemographic
characteristics of the medical students and their medical
specialty choices as evidenced by residency selection of
either primary care or nonprimary care specialties.
The chi-square test is valid only if at least 80% of
the cells in a contingency table have frequencies of at
least five. Where small cell frequencies are involved,
categories may be collapsed as long as it is done logically
and not merely to achieve significance or create bias. Some
of the sociodemographic characteristic categories were
collapsed for purposes of statistical analysis. When small
cell frequencies persisted, Fisher's exact test was used as
an alternative test of significance for 2x2 tables
(Champion, 1970). The Agresti-Wackerly method of computing


-42-
influence of particular events was more pronounced in the
clinical phase than in the basic science period and the
responses were more dichotomous on the evaluation scale.
Those students choosing specialties selected by small
numbers of students such as psychiatry and anesthesiology
perceived the faculty influence to be a more important
factor than did students choosing the popular specialties of
internal medicine, surgery, and family medicine. Another
finding which has implications for curriculum planners is
the influence of the sequence of the rotations. The order
of the clerkships was particularly relevant for the students
who indicated prior to the clinical rotations that they were
undecided on their specialty as most of them ended up
choosing a specialty they had encountered during the first
few clinical rotations.
Changes in understanding and attitudes resulting from
an experience with a particular specialty were studied by
Samra et al. (1983). Students participating in an
anesthesiology rotation completed an attitudinal
questionnaire containing statements related to
anesthesiology. They also indicated their career choices.
The clerkship did not have any immediate impact on the
students' choices of specialties but there was evidence of
change in perception of the image of an anesthesiologist to
that of a physician rather than a technician. The students
also perceived the role of the anestheiologist to be broader
after the clerkship. Although limited by the research


FACTORS RELATED TO STUDENT CHOICE
OF MEDICAL SPECIALTY
By
MARGARET C. DUERSON
A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
1986


TABLE OF CONTENTS
Page
ACKNOWLEDGEMENTS
LIST OF TABLES vi
ABSTRACT viii
CHAPTER
IINTRODUCTION 1
Statement of the Problem 3
Purpose of the Study 4
Hypotheses 5
Background and Justification 7
Delimitations, Limitations, and
Assumptions 11
Definition of Terms 14
Organization of the Study 16
IIREVIEW OF THE LITERATURE 18
Introduction 18
Background Factors 19
Personality and Attitude Factors 31
Factors Related to the Medical Training
System 37
Career Factors 46
Summary of the Review of Literature 49
IIIMETHODOLOGY AND INSTRUMENTATION 53
Setting for the Study 53
Revision and Development of the
Questionnaire 56
Administration of the Questionnaire 62
Residency Match 66
Research Design 67
Data Analysis 70
IVPRESENTATION OF RESULTS 76
Sociodemographic Characteristics of the
Study Group 76
Results of the Data Analysis for Testing
the Hypotheses 82
IV


-95-
Table 10
Sociodemographic Groupings of
Medical
Students
and
Postclerkship Attitudes
Sociodemographic
H
Significance
Characteristics
df
Value
Sex
1
1.14
.28
Marital Status
1
2.62
.11
Religious Orientation
4
5.19
.27
Religious Importance
2
6.63
.04*
Race
1
.16
.69
Citizenship
1
4.63
.03*
Hometown Location
1
4.02
.05*
Hometown Size
3
3.84
.28
Hometown Economic Base
4
2.42
.66
Father's Occupation
1
4.06
.04*
Mother's Occupation
1
3.29
.07
Parents' Religion
4
6.52
.16
Parents' Income
1
4.19
.04*
Source of Present Support
1
8.27
.004**
Indebtedness
1
.13
.72
Amount of Debt
1
3.01
.08
General health
1
.16
.69
Public or Private High School
1
.00
1.00
Public or Private College
1
.83
.36
Family's Medical Care Provider
Family Practitioner/General
1
3.73
.05*
Internist
1
1.41
.23
Pediatrician
1
1.44
.23
Other
1
.26
.61
Work Experiences
Medical
1
1.10
.29
Helping Activities
1
.00
.96
Technical
1
.08
.77
Teaching
1
9.19
.002**
Public
1
2.09
.15
Unskilled
1
.43
.51
Other
1
1.27
.26
College Major
1
.35
.55
Level of Education
1
.97
.32
*£ < .05
**£ < .01


I certify that I have read this study and that 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.
^ Cvv-A^-J Sty. l/l/l
Ronald G. Marks
Associate Professor of Statistics
I certify that
opinion it conforms
presentation and is
a dissertation for
I have read this study and that in my
to acceptable standards of scholarly
fully adequate, in scope and quality,
the degree of Doctor of Philosophy.
as
This dissertation was submitted
the College of Education and to
accepted as partial fulfillment
degree of Doctor of Philosophy.
to the Graduate Faculty
the Graduate School and
of the requirements for
of
was
the
December 1986
Dean, College of Education
Dean, Graduate School


-118-
results of the fourth-year residency match served as
evidence of the student's area of specialty choice.
The approach to examining the data was through
nonparametric procedures appropriate for the research model
and the nominal- and ordinal-levels of measurement. The
statistical tests included chi-square, Spearman's
correlation coefficient, Kruskal-Wallis one-way analysis of
variance, and Wilcoxon matched-pairs signed-ranks. Ten null
hypotheses were tested.
Overall, the sociodemographic characteristics of the
medical students in this sample were poorly associated with
third-year specialty preferences and fourth-year specialty
choices. However, some of the background variables did
achieve statistical significance. For example, parents'
annual income was related to third-year preferences
(£ = .0004). Students whose parents earned an annual income
of $50,000 or less were considerably more likely to express
a preference for the primary care specialties in the third-
year and to maintain that predilection into the fourth-year
residency choice than were students from the more affluent
familiesincome greater than $50,000.
Another measure of socioeconomic status, father's
occupation, was associated with third-year specialty
preferences (jd = .02). Students whose fathers engaged in
business or other nonprofessional types of employment were
more likely to favor a primary care specialty than those
with professional/ technical fathers. A much larger percent


-133-
sometimes change their specialties or limit their practice
to a specific area. In this study the groupings according
to specialty preference and choice were broad and may have
included many students who actually will enter subspecialty
training after the first year of residency. Continuing the
study longitudinally would increase the number of subjects
sufficiently to enable more definitive categorization of
groups. It is also recommended that the study of these
students be extended into the residency years and beyond to
obtain data which more closely approximate their actual
specialty practice.
Revision of the Questionnaire
It is recommended that the existing instrument be used
in replications with the following revisions to the
Background section:
1. Revise the item on Citizenship as it is a low-yield
item with less than 3% of the students indicating
citizenship other than U.S. Omit Racial Background
categories of American Indian, Mexican-American, and Puerto
Rican and add Hispanic when using the questionnaire
longitudinally with students at the institution of this
study. Remove the question Do you consider yourself in
excellent health? as 96% of students gave a positive
response. Such low yield questions lengthen the
questionnaire unnecessarily.


-22-
children. However, there was a pattern toward over
representation in some specialty areas for the physicians'
children, namely ophthalmology, otolaryngology, dermatology,
and surgery.
Mawardi (1979), in a longitudinal study of Case
Western Reserve University graduates, found that 39 of
the 135 physicians studied gave credit for their interest
in medicine to family members in the profession. The
most frequently mentioned influential relative was a
father. These 39 subjects cited 59 relatives who were
physicians.
Given the rising cost of medical education and the
concommitent decrease in financial aid available to
students, it could be hypothesized that students from
families with few financial resources will not apply to
medical schools. This could lead to a profession of the
socially elite. Based on the preliminary studies of
socioeconomic status and specialty choice, a preponderance
of students from affluent backgrounds will effect the mix of
specialist toward nonprimary care specialties.
Factors such as age and marital status have provided
some consistent clues to specialty choice decision. Using
the data from a longitudinal study of Jefferson Medical
College students from 1971 through 1975, Herman and Veloski
(1977) demonstrated not only a rising trend in the
proportion of the Jefferson students interested in family
medicine but those senior students expressing an interest in


-34-
Coleman (1984) demonstrated the stability of the MBTI from
entrance to graduation for the classes of 1980 and 1981 at
the University of Utah School of Medicine. In a discussion
of implications relevant to the findings of MBTI studies,
Hadac (1983/1984) observed that sensing types tend to be
denied admission to medical schools because they generally
score lower on entrance exams. When they are admitted, they
prefer primary care specialties, thus admission policies may
be eliminating those students likely to correct the
specialty maldistribution. Hadac cautioned that
psychological type alone cannot predict specialty choice for
all students. He stated, "Also, all personality types are
found in all specialties and there is not at present, and
may never be, strong evidence that any type should be
excluded from any specialty" (p. 46).
When considering the literature on personality and
attitudes, it was considered worthwhile to address
interests, values, and early orientations to medicine. The
criterion group in the Leserman (1978) study were first year
medical students at three North Carolina medical schools.
Data from this group of students
suggest that incoming medical students are
concerned with helping people but not necessarily
through political means, committed to some
geographic and specialty areas of patient need,
choosing medicine for reasons other than economic
rewards but not opposed to physicians' large
income and status and somewhat unaware of
discrimination toward women physicians and
patients. (p. 330)


-101-
students were not significantly different in their goal
orientation (Table 12).
After the family medicine clerkship, students ranked
the career goal phrases in the same manner as they had been
asked to do prior to the clerkship some 2 months previous.
These postclerkship rankings were compared for the two
fourth-year specialty choice groups, primary care and
subspecialty. The two groups were not distinguishable based
on their goal rankings (Table 12).
Based on the analysis of the data, it was concluded
that the medical specialty preference groups and the fourth-
year specialty choice groups were not dissimilar in the
relative importance they assigned to the career goal
phrases. The null hypothesis could not be rejected.
Hypothesis 5
There is no relationship between medical students'
rankings of career goals and their attitudes toward primary
care before or after a third-year family medicine clerkship.
This null hypothesis used the data in the respondent-
reported section of the questionnaire entitled Career Goals
and the section labeled Attitudes about Primary Care and
Family Medicine. Respondents were surveyed before and after
a third-year clerkship in family medicine. Students rank-
ordered two sets of six career goal phrases on a scale from
1 to 6 with 1 being the most important goal you wish to


-41-
which the authors interpreted as being in opposition to
students' interest in family medicine.
Closely associated with this study is the work of
Harris et al. (1982) who inquired into the question "Is
participation in a family practice track program associated
with career choice decisions in family practice as evidenced
by residency selection?" (p. 610). Each year for two years,
20 students who applied for an elective in family medicine
were randomly selected for the experience. The students who
participated in the elective selected family medicine
residency at a significantly higher rate than those who
expressed an interest in the track but were not selected for
the elective. Although the subjects may be considered a
biased group, the positive effect of experiencing such a
program merits consideration.
Data pertaining to influences of persons and
experiences on career decisions published the same year as
the Harris study (Paiva et al., 1982) extended the
information in this area. The 1980 and 1981 classes of
Southern Illinois University School of Medicine were sampled
at three points in the curriculum, at the end of the basic
science period, the end of the clinical science period, and
the end of the clinical clerkships. Analysis of the data
showed that approximately three-fourths of the students
reported that a faculty member had some or very much
influence upon their specialty decisions. The major
influence came from full-time clinical faculty. The


-129-
certainly by the medical school admissions process. Given
this highly discriminating process, it is still unclear as
to how such a homogeneous group translates into so many
specialties. Background characteristics have for the most
part been inconsistent predictors of specialty choice with
the exception of a few, isolated variables. One such
variable, sex of the students, has proved to be one of the
few constant sociodemographic measures associated with
specialty choice. Historically, women have been more
inclined toward primary care specialties. Even this
variable is circumspect given the present and impending
changes in women's roles in society, in general, and
medicine, in particular.
The numbers of women in this study were relatively
small. It is recommended that in data collection with
subsequent groups, this variable and variables related to
other minority groups be given attention.
The complete picture of cost factors and their
influence on specialty choice is not clear from the
literature. It can be assumed that economically hard-
pressed students are influenced in their choice of specialty
as a result of financial considerations. Not only is income
lost with additional years of graduate education but each
added year of residency may actually increase the
indebtedness of the trainee. The present study demonstrated
an association between parents' annual income and specialty


-45-
and residency programs. This information was compared with
first-year medical student preferences for specialties, the
students' fourth-year choices, and their careers four years
after graduation. The results suggest that hiring more
full-time faculty in a specialty area and encouraging more
research in one area than another does cause physicians to
enter one specialty over another. Full-time faculty and
research activities have more influence on career choice
than clinical inputs such as part-time faculty and residency
programs. In future planning, then, it would seem that
increasing efforts in these directions would encourage
students in the direction where the faculty and research are
located.
Concerned about the decreasing interest in family
medicine since the early 1970s, Goldsmith (1982) surveyed
135 medical schools with 71 returns on their administrative
structure and number of graduates choosing family medicine.
The data revealed that schools with departmental status had
a higher percentage of graduates in family medicine. The
explanation for the observed association may be that those
with departmental status can effect admission policies,
command more time in the curriculum, attract more faculty,
and participate more heavily in career counseling.
A survey of departments of anesthesiology supported
these findings (Chandra & Hughes, 1984). Staffing patterns
in the departments were significantly related to the number
of students choosing to specialize in anesthesiology.


CHAPTER II
REVIEW OF THE LITERATURE
Introduction
During the last 10 years researchers have examined a
wide range of factors thought to influence medical specialty
choice. Because of the importance of the problem and the
consequences of these intraprofessional career decisions,
medical educators, sociologists, psychologists, health
policy-makers, and others concerned with medical care in
general have attempted to understand why and how choice of
medical specialization occurs. The Graduate Medical
Education National Advisory Committee Report (U.S. DHHS,
1980), confirming the suspicion that the nation faced an
oversupply in some specialty areas and actual shortages in
other specialties, gave further impetus to the study of
variables associated with specialty choice.
This great interest in the career choice process of
young doctors has resulted in an extensive body of
literature. Literature reviews on career choice within
medicine were published in the mid-1970s. Due to the
comprehensive nature of these publications, this review will
examine the literature published since 1975.
Approaches to the study of the problem have been so
diverse, this review will consider related research together
-18-


xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EBQ9N92O7_0HBO9J INGEST_TIME 2014-10-09T22:14:47Z PACKAGE AA00025830_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES


-29-
choices as women ranked pediatrics second and men chose
internal medicine. Surgery ranked last for women and third
for men.
Similarly, Cuca (1979), based on plans for board-
certification of 1978 U.S. medical school graduates, found
that over half of both males and females preferred primary
care while twice as many women preferred pediatrics as did
the men. The category of surgery again evidenced dissimilar
trends for women and men.
The work of Zimny and Shelton (1982) supported these
earlier studies. Responses of 380 third year medical
students on the Medical Specialty Preference Inventory
showed no significant differences between male and female
students' preference for obstetrics, pediatrics, and
psychiatry. Men, on the other hand, had a slight preference
for internal medicine and surgery.
Bergquist and his colleagues (1985) asked the 1983
entering class of one medical school to select their
intended specialty choice from a list of 50 medical fields.
When the specialties were categorized into primary care,
surgery, medicine, and other, the results showed that women
selected specialties in primary care more often than men.
Men selected surgical categories significantly more
frequently than women.
On the theory that as the proportion of women in
organizations reach parity with men, women will be less
likely to be entrapped in limited or stereotypically female


-122-
intraprofessional career decisions in medicine in an effort
to provide some direction for medical educators and policy
makers.
Consistent with previous research (Anderson, 1975; U.S.
DHHS, 1980), this study demonstrated the capriciousness of
sociodemographic characteristics in association with
specialty choices. An explanation for this may be the
selection process, medical students are a highly selected
group and, therefore, homogeneous in background
characteristics. In addition, the timing of this study was
such that students had been in the medical school
environment for two to three years which could have served
to further dilute the impact of the sociodemographic
differences on specialty preferences and choices.
Two measures of socioeconomic status, parents' annual
income and father's occupation, were associated with
specialty selection for the study group. Students whose
parents earned in excess of $50,000 annually and students
whose father's occupation was classified as
professional/technical were much more likely to choose
subspecialty practice than primary care. A large percent of
students who were undecided as to their specialty
preferences in the third year had fathers in the
professional/technical ranks. These findings support
Zuckerman's (1977) contention that economic affluence may
encourage specialization.


-16-
primary care or nonprimary care specialty was solicited in
the initial administration of the questionnaire in the third
year of medical school prior to the family medicine
clerkship.
Choice of medical specialty requires that some action
be taken in the direction of or participation in a
particular area of specialty practice, for example, taking
steps to secure a residency in a particular medical
specialty area. For purposes of this study, specialty
choice was considered to be evidenced by the selection of
residency in a given specialty area in the fourth year of
medical school.
Organization of the Study
Chapter I contained an introduction, a statement of the
problem, and the purpose of the study. The research
hypotheses underlying the study were stated. The
limitations and delimitations of the project were explained
along with the definitions of words and phrases specific to
this study.
Chapter II is a review of the literature related to
medical specialty choice. Chapter III addresses the
research methodology and instrumentation. The data
collection procedures, the setting for the study, research
design, and data analysis procedures are explained.
Statistical results are presented in Chatper IV.


-38-
present in the environment. There is a strong feeling that
these structural influences and the socialization process of
the medical education system are key determinants of career
choice. Based on this premise a number of studies have
focused on the institutional environment and manipulation of
variables therein.
Early reports concerning the development of medical
students were generally qualitative but, nonetheless, paved
the way for later quantitative studies on institutional
influences. Becker et al. (1961), in a hallmark study of
the socialization process, addressed the observed increase
in cynicism of medical students during their training. He
argued that students enter with idealistic concerns about
the sick and serving mankind. From his observations,
students did not lose this idealism for broader social
issues. What some interpreted as cynicism was really the
response to pressures of day-to-day details with long hours
and intense studying which caused a narrowing of concerns
just to get through the rigors of medical school. This
short term effect disappeared but the long term
institutional effects did not. Becker thought there was
enough congruence between students chosen for admission
and the institutional values to effect long range
perspectives.
In a review of the literature published almost 20
years after the Becker book, cynicism continued to be
reported as a by-product of the medical school environment.


-99-
Hypothesis 4
There is no difference between groups of medical
students identified by medical specialty preference in the
third year or medical specialty choice in the fourth year
with respect to their rankings of career goals. Just prior
to a family medicine clerkship and again, after the
clerkship, students ranked 2 sets of career goal phrases.
Each set contained six goal phrases which might be achieved
through a career in medicine. The students were asked to
rank the importance of the goals by putting a "1" next to
the most important goal, a "2" next to the second most
important goal through "6" for the least important. The
Kruska1-Wal1 is one-way analysis of variance by rank was used
to test the null hypothesis. Students' rankings of each
career phase were tested using a significance level of
< .01.
The three medical specialty preference groups, primary
care, subspecialty, and undecided, were not differentiated
by the rankings they gave to the goal statements before the
family medicine clerkship. The results of the preclerkship
goal rankings and specialty preference groups are found in
Table 11.
When the preclerkship career goal rankings were
analyzed with the fourth-year specialty choice groups, the
primary care specialty group and the subspecialty group of


-48-
investigations he concluded that economic incentives and
ideology are more compelling than basic characteristics or
original plans in predicting career outcomes. Further,
societal forces prevailing at the time of the decision exert
significant influence. As Funkenstein (1978) observed, the
"Sputnik" era saw an increase in emphasis on mastery in the
sciences with increased resources available in scientific
subjects. Students admitted to medical school during this
period had heavy science background. The 1960s and 1970s
were periods of social conscience-raising and a time when
individualism was prominent. This was followed by decreased
funding in some areas of medicine so that careers in
academic medicine were not as attractive (Funkenstein, 1978,
p. 30) .
Society continues to be enamoured with technology and
is willing to reward those who provide it. High-tech in the
health field has brought us heart transplants, microsurgery,
and life-sustaining equipment (Naisbitt, 1982). It could be
hypothesized that high tech translates into highly
specialized physicians to administer this type of medical
care and a renewed interest in family doctors to balance the
high-tech with highly personal care (Naisbitt, 1982, p. 41).
On the other hand the organization of the health care
system is changing (Ginzberg, 1984). As large corporations
buy and operate hospitals, medical care has become a
businessa business for profit (Weiss, 1982). High
technology is expensive; subspecialists are expensive;
tertiary care hospitals are expensive. The most economic


Table 1
Summary of Sociodemographic Characteristics of Subjects
Sociodemographic
Sociodemographic
Characteristics
n
Percent*
Characteristics
n
Percent
Sex
Marital Status
male
79
73.1
married
34
32.1
female
29
26.9
other
72
67.9
Race
Citizenship
white
94
87.9
U.S.
105
97.2
nonwhite
13
12.1
other
3
2.8
Hometown Location
Father's Occupation
south/southeast
89
82.4
professional/technical
56
52.3
not south
19
17.6
business/other
51
47.7
Hometown Size
Mother's Occupation
0-99,000
51
47.6
housewife
54
50.0
> 100,000
56
52.3
work outside home
54
50.0
Hometown Economic Base
Parents' Annual Income
farming
11
10.6
>$50,000
31
31.3
business
37
35.6
$0-50,000
68
68.7
heavy industry
light industry
mixed economy
5
2
49
4.8
1.9
47.1
Source of Present Support
parents
other
31
77
28.7
71.3
Level of Education-Degree
bachelors
98
90.7
Indebtedness
86
20
81.1
18.9
graduate
10
9.3
yes
no
Amount of Debt
$1000-5000
7
8.2
> $5000
7U
91.8
Sociodemographic
Characteristios
Religious Orientation
Protestant
Catholic
Jewish
other
no preference
Religious Importance
very much
some
a little/none at all
Parents' Religion
Protestant
Catholic
Jewish
other
no preference
High School Attended
public
private
College Attended
public
private
Undergraduate Major
science
other
n Percent*
46
43.8
23
21.9
12
11.4
8
7.6
16
15.2
40
37.4
37
34.6
30
28.0
50
47.6
30
28.6
11
10.5
9
8.6
5
4.8
83
78.3
23
21.7
76
71.0
31
29.0
90
86.5
14
13.5
I
I
Overall category may not equal 100% due to rounding


-10-
schools to meet In order to continue to receive capitation
grant funds (U.S. DHHS, 1980, Vol. V, p. 1).
All of these early efforts toward correcting the
shortage of primary care physicians were directed at the
residency training period. Almost no attention was given to
understanding and influencing the career preferences and
choices made earlier in the medical education process.
Until recent years, research on career decisions at the
undergraduate medical education level had been principally
conducted by sociologists and psychologists with academic
interest in furthering the knowledge base in their own
discipline rather than an interest in obtaining information
for program evaluation and policy formulation. Influencing
career choice at the undergraduate medical education level
was sanctioned by Califano, Secretary of Health, Education,
and Welfare, in an address to the Annual Meeting of the
Association of American Medical Colleges in 1978:
I urge medical educators to exert leadership in
dealing with the dangerous and wasteful decline in
primary care physicians by encouraging course
offerings or other innovations that increase the
exposure of students to primary care settings and
heighten the appeal of primary care. (Califano,
1979, p. 22)
Medical schools accepted the challenge and expanded the
curricula to include programs in family medicine at the
undergraduate level of medical education as well as the
residency level.
Since the medical education environment is recognized
as one of the factors contributing to the specialty


-75-
between groups of medical students identified by their
fourth-year specialty choice with respect to their attitudes
toward primary care before and after a family medicine
clerkship. The two groups of specialty choice constitute
independent groups and the measure of attitudes before and
after the clerkship constitute at least ordinal-level data;
therefore, the Kruskal-Wallis was concluded to be the right
procedure.
When multiple statistical tests are conducted on data
as they were in this study, the chance of making erroneous
conclusions is increased and, therefore, the overall
significance level (OSL) must be set conservatively. The
significance level of _< .01 was used for hypotheses tested
in this study so that the OSL would not exceed .10.
Marks (1982) discusses two philosophies concerning the
significance level used for testing the null hypothesis (p.
22). One philosophy is to perform the statistical test
using a predetermined level of significance and to use the
test results only to see if the predetermined level of
significance is achieved. Another approach is to present
the attained significance level, decide whether or not the
evidence is sufficient to reject the null hypothesis, and
let others draw their own conclusions from the results. The
second rationale was used in presenting the results of this
study.


-153-
U.S. Department of Health and Human Services, Health
Resources Administration. (1980). Summary report of
the graduate medical education national advisory
committee, Vols. I-V (DHHS Publication No. HRA 81-651).
Washington, DC: U.S. Government Printing Office.
Weisman, C.S. (1984). Gender composition of medical
schools and specialty choices of graduates. Journal
of Medical Education, 59, 347-349.
Weiss, K. (1982). Corporate medicine. 2. What's the
bottom line for physicians and patients? The New
Physician, _9, 19-25.
White, K.L., Williams, T.F., & Greenberg, B.G. (1961).
Ecology of medical care. New England Journal of
Medicine, 265, 885-891.
Zimny, G.H., & Shelton, B.R. (1982). Sex differences in
medical specialty preferences. Journal of Medical
Education, 57, 403-405.
Zuckerman, H.S. (1977). Evaluation of literature on career
choice within medicine. Medical Care Review, 34, 1079-
1100.
Zuckerman, H.S. (1978). Structural factors as determinants
of career patterns in medicine. Journal of Medical
Education, 53, 453-463.


-150-
Funkenstein, D.H. (1978). Medical students, medical
schools and society during five eras: Factors
affecting the career choices of physicians 1958-1976.
Cambridge, MA: Ballinger Publishing Co.
Ginzberg, E. (1984). The monetarization of medical care.
New England Journal of Medicine, 310, 1162-1165.
Goldsmith, G. (1982). Factors influencing family practice
residency selection: A national survey. Journal of
Family Practice, 15, 121-124.
Gough, H.G. (1978). Some predictive implications of
premedical scientific competence and preferences.
Journal of Medical Education, 53, 291-300.
Gough, H.G., & Hall, W.B. (1977). A comparison of medical
students from medical and nonmedical families. Journal
of Medical Education, 52, 541-547.
Hadac, R.R. (1984). A longitudinal study of factors
associated with choice of medical specialty and
practice setting at the University of Washington
(Doctoral dissertation, University of Washington,
1983). Dissertation Abstracts International, 44,
1029A.
Hadley, J.
choice
(1977). An
Inquiry,
empirical model
14, 384-401.
of medical specialty
Harr
is, D.L., Coleman, M., & Malea, M.
participation in a family practice
student career decisions. J ourna 1
Education, 57, 609-614.
(1982). Impact of
track program on
of Medical
Harr
is, D.L., Kelley, K., & Coleman, M. (1984). The
stability of personality types and their usefulness
medical student career guidance. Family Medicine,
203-205.
in
16,
Herman, M.W., & Veloski, J. (1977). Family medicine and
primary care: Trends and student characteristics.
Journal of Medical Education, 52, 99-106.
Hutt, R. (1976). Doctors' career choice: Previous
research and its relevance for policy-making. Medical
Education, 10, 463-473.
Jacoby, I. (1981). Physicians manpower: GMENAC and
afterward. Public Health Reports, 96, 295-303.
Kindig, D.A., & Dunham, N.C. (1985). Physician specialist
growth into the 21st century. Journal of Medical
Education, 60, 558-559.


-28-
medical practice requires a high intellectual functioning,
there are no discernible differences based on scholastic
factors.
With the growing numbers of women entering medical
school, research on the effects of gender composition on
specialty choice has been heightened. Historically, women
have been underrepresented in medicine. Women constituted
r
only 7.7% of medical students in 1964-1965 (Dube, 1973).
Concurrent with the women's movement of the late 1960s and
the efforts to include women in occupations where they had
been formerly excluded, their representation in medicine
increased. Presently females constitute over 30% of the
medical student population. The percentage of women in the
entering class of 1991 is expected to be close to 45.3%
(Lanska et al., 1984). This change in gender distribution
has possible implications for health care delivery in
general and specialty choice in particular.
Several studies of sex differences related to specialty
choice found significant differences between male and female
medical students' propensity toward specialties.
Male/female differences in specialty choice, beliefs about
specialties, and personality characteristics were explored
by McGrath and Zimet (1977). Students at two state
university medical schools rank ordered their preference for
five major medical specialties. Both men and women listed
family medicine as their first choice but significant
differences occurred at the levels of second and third


-141
A Longitudinal Study of Family Prictlce Education
A Follow-Up Survey
ATTITUDES ABOUT PRIMARY CARE AMO rAMILY MEDICINE (continued) (Circle the number of the response which best describes your own personal
feelings tcwaro eacn statement!
13. The only way to practice good 12345
medicine is to do a comolete
history and physical examina
tion eacn visit.
14. Most patients with emotional 12345
disorders cannot be helped
without spending a lot of
the doctor's time.
21. The provision of services by 12345
specialty oriented physicians
snould be coordinated and
controlled by primary care
pnysiclans.
22. There is no useful researcn 12345
work that can be done in
primary care practice.
15.Health maintenance is not 12345
as interesting to me as
curative medicine.
23.Family practices are not as 12345
profitable financially as
specialty practices.
16.Disease prevention should be 12345
the responsibility of public
health deoartments rather
than the personal physician.
24.The demands of a family 12345
practice leave little
leisure time for family
Ufe or recreation.
17. In order to provide the best
care for his patients, a doctor
should personally perform as
many of the direct patient-
contact services as possible.
18. Consultants play a secondary
role in the patients total
health care.
13. People with impaired mental
health are as likely to get
well as oeoole with impaired
pnysicial healtn.
20. Except for certain diseases
specific Knowledge pertaining
to disease prevention is so
fragmented that a physician
should limit his efforts to
curative medicine.
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
25. A comolete medical and social
data base on each patient is
important to me.
26. Most patients' medical proolems
involve aspects of their cooing
mechanisms with life's daily
challenges.
27. Physician's assistants will
play an important role in future
primary care.
23. I would prefer to spend my time
dealing with patients' medical
proDlems rather than their social
or psychiological problems.
29. Physician's assistants should
handle the acute minor illness
problems in primary care.
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
strongly
disagree


-36-
questionnaire, males and females ranked the same reasons as
important for going to medical school although they ranked
them in different order. Males' order of choices were
interest in science, helping people, and having a career.
Women ranked helping people as first in importance, followed
by desire for a career, and interest in science. All
respondents listed a happy family life and job satisfaction
as important life goals. Men tended to consider a high
income as more important than the females. Although
attitudes toward chores were egalitarian, responses
indicated that in reality women performed most routine
household activities. The researchers pointed out that
women's value considerations may affect initial career
choices as to training programs and work opportunities.
Assessment of personality characteristics has been done
rather unsystematically using a variety of standardized and
nonstandardized instruments administered at varying times
throughout medical school. The characteristics observed
have reflected the special interests of the investigators
and have often measured vague concepts such as idealism and
cynicism. There has been no attempt to replicate previous
research in most cases.
A further difficulty with systematizing the assessment
of personality factors is that psychological tests can be
falsified if the respondent is motivated to do so. The
person can answer the items in the way that is believed to
be preferred by the investigator or is perceived to be


Table 12
Results of Preclerkship and Postclerkship Career Goal Rankings and Two Specialty
Choice Groups
Preliminary Goal
Rankings
Follow-Up Goal
Rankings
Career Goals"1"
H
£
H
£
Set
a.
1
helping to solve community's medical and
social problems
.11
.74
1.14
.29
b.
gaining respect and friendship of other
physicians
.42
.52
3.38
.07
c.
developing long-term relationships with
patients
.05
.82
.15
.70
d.
contributing to medical knowledge through
research
1.58
.21
.05
.82
e.
assuring financial security for self and
family
.06
.84
1.19
.28
f.
allowing me to live in preferred area
and style
.09
.77
.81
.37
-102-


-52-
and (c) insufficient data are available on career and
personal goals of medical students from which to make
important educational and policy decisions.


-9-
the higher ratio of specialists to generalists (Schroeder,
1984) .
Factors responsible for the trend toward increasing
specialization are difficult to determine. Rapidly
expanding technology and scientific knowledge may cause some
to conclude that the only way to achieve depth of knowledge
and skill in an area or to remain current in an area of
medicine is to narrow the field of interest. Petersdorf
(1975) commented on the possible causes:
Simplistical ly speaking, every man has an innate
desire for self-advancement, intellectually and
financially. In addition, subspecialists seem to
acquire prestige, particularly in medical schools.
More than anything else, however, the structure of
training programs in medical schools has led to
specialization. (p. 697)
To fill the void left with the demi se of the old
generalists' role, a new specialty, family practice,
developed specifically designed to deliver primary care.
The final approval of family practice as the twentieth
primary specialty was granted by the American Medical
Association in 1969 (Rakel & Pisacano, 1984). In an effort
to respond to the need for more primary care physicians and
aided by funding provided by the Comprehensive Health
Manpower Training Act of 1971, medical schools initiated
residency training programs in this new specialty. In a
further attempt to influence specialty distribution,
legislation was passed in 1976 which "set minimum
requirements for percentages of first year residency
positions in the primary care specialties" for medical


-32-
that those specialties where the choice is made later, based
on experience in the fields, are much less successfully
identified from personality characteristics. Hutt
commented, "The field of personality differences in relation
to specialty, important as they are, is not a promising one
for policy-makers who wish to put right particular
imbalances in the supply of doctors" (p. 466).
Though there have been problems in predicting specialty
choice from comparison of personality variables, it is
important to review the findings. For example, Collins and
Roessler (1975) compared four different career groups,
internal medicine, surgery, obstetrics-gynecology, and
pediatrics, with family medicine residents. One hundred
eighteen third-and fourth-year students at Baylor College of
Medicine participated in a battery of personality, vocation,
and attitude tests to investigate the relationship between
the test data and subsequent specialty choice. The Baylor
students who chose family medicine residencies were in most
instances significantly different from the other four
groups. They scored higher on the need for affiliation,
were less aggressive and less materialistic. The family
medicine students' scores were far below the population
average of 50 on materialism on the Birkman Vocational
Interest and Attitude Survey.
Plovnick (1980) also observed differences in attitudes
and values between students oriented toward primary care and
those choosing nonprimary care specialties. The small


Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy
FACTORS RELATED TO STUDENT CHOICE
OF MEDICAL SPECIALTY
by
Margaret C. Duerson
December 1986
Chairperson: James W. Hensel
Major Department: Educational Leadership
The purpose of this study was to identify factors
related to medical students' third- and fourth-year
specialty choices and to determine the extent to which a
clerkship in family medicine affected attitudes toward
primary care. A multidimensional questionnaire was
administered to third-year medical students immediately
before and after a family medicine clerkship. The
questionnaire addressed sociodemographic characteristics of
the medical students, their career goals, and attitudes
toward primary care. On the initial administration, the
students indicated whether they preferred a primary care
specialty, a subspecialty area, or were undecided on their
future plans for residency training. The results of the
fourth-year residency match served as evidence of the
students' choice.
Responses by 109 students revealed that (a) third-year
specialty preference was significantly associated with
fourth-year specialty choice of residency; (b) socioeconomic
v i i i


-107-
scores. For those items which either reached significance
or were < .05, the results were as follows:
In the follow-up data, the career goal "helping lead my
community to the solution of social and medical problems"
was significantly correlated with follow-up attitude scores
(Set 1, Goal a, Table 13). The correlation coefficient
between these two variables was -.22, jd = .04. The
direction of the correlation indicated that students who
ranked this goal as among the more important goals had more
favorable attitudes toward primary care and family
medicine.
When the career goal "developing long-term, intensive
relationships with patients" was correlated with follow-up
attitudes, the result was an r value of -.27, jo = .008 (Set
1, Goal c, Table 13). The direction of the relationship was
interpreted to mean that students with a favorable attitude
toward primary care ranked this goal high in importance.
The follow-up rankings which students assigned to the career
goal "assuring financial security for self and family"
correlated with follow-up attitudes at a significance level
of £ = .04 (Set 1, Goal e, Table 13). The positive
direction of the correlation indicated that as the
importance of this goal increased the attitude scores
decreased.
The career goal "allowing me to live in the type of
area and lifestyle that I prefer" was significantly related
to attitudes before and after the clerkship (Set 1, Goal f,
Table 13). The observed r for the preliminary data was .28,


-93-
level of £ .01. Results of the sociodemographic groupings of
medical students for attitudes toward primary care in the
preclerkship sampling are summarized in Table 9.
Postclerkship results are given in Table 10.
The overall significance level for Hypothesis 3 was
established at £ .01. Since all of the tests did not reach
this significance level, the null hypothesis could not be
rejected. Those individual variables which did reach
the .01 level of significance or closely approximated it are
discussed.
Religious importance. The students were asked to
indicate "how much importance does religion have in you
life?" The levels of response--very much, some, and
little/none at allwere used to identify three categories
of students. When the three categories of students were
compared for differences in attitudes toward primary care,
the preclerkship data yielded an H value of 5.84 (elf 2, N =
102), £ = .05 (Table 9). The postclerkship test statistic
was H (df 2, N = 85) = 6.63, £ = .04 (Table 10).
Citizenship. When students from this country and
foreign students were ranked according to their
postclerkship attitude scores, the result was an H value of
4.63 with one degree of freedom and a significance level
of .03. This finding should be viewed with caution,
however, due to the small number of students in the
noncitizenship group.
Hometown location. The respondents were asked to
indicate the location of their hometown community in the


-70-
Data Analysis
One important criterion in the choice of which
statistical procedure to use to test a research hypothesis
is that the data meet the assumption of a particular level
of measurement (Marks, 1982). The elements of observation
in this study were considered nominal- and ordinal-level
data. When words or symbols are used to classify persons or
characteristics into groups, they constitute a nominal
scale. In this research, sociodemographic characteristics,
specialty preference groups, and specialty choice groups
were identified as being nominal-level variables. With
ordinal-level data, the classes or categories stand in
relation to each other so that rank ordering is permitted.
Students' rankings of career goals and their attitudes
toward primary care measured on a Likert scale resulted in
information contained in an ordinal scale.
Since nonparametric statistical tests are appropriate
for data inherently in ranks or numerical scores having the
strength of ranks, and some apply to nominal-level data as
well, techniques of inference drawn from this body of
methods were chosen for this study. The following
nonparametric statistical procedures were used to test the
10 null hypotheses:
1. Hypothesis 1, there is no association between
sociodemographic traits of medical students and their
specialty preferences in the third year of medical school,
was tested using the chi-square test (X^). The chi-square


ACKNOWLEDGEMENTS
My deepest gratitude is extended to each of my
committee members. Special thanks go to my chairperson, Dr.
James W. Hensel, who provided me with unending support and
guidance. His was a warm and motivating force throughout
the project. As a role model he exemplifies the teacher I
hope to emulate as I interact with students and peers.
Each committee member played a unique and indispensable
role in assisting me in this endeavor. I would like to
thank Dr. Parker A. Small for his extensive, constructive
criticism and his encouragement to constantly strive for
excellence. I am grateful to Dr. Robert E. Jester and Dr.
Ronald G. Marks for their valuable comments and suggestions,
especially in the areas of research design and analysis. It
was my good fortune to have the advice and help of Dr.
Forrest W. Parkay.
The Department of Community Health and Family Medicine
made available to me information and resources during the
conduct of this study. For this, I am deeply indebted and
appreciative. I would also like to thank Robert Epting for
his help in executing the data analysis procedures. In
addition, support was provided by Health and Human Services
Grant number 2D15PE-84000-07.


-12-
different settings, generalizations beyond the sample
studied in the setting of interest are to be avoided. The
generalizability of this study was further limited since
subjects were not randomly selected but came from an intact
class. The students were randomly selected by a lottery
system into the six rotation groups, however.
2. A limitation of the study was a lack of control of
intervening variables. Specific events other than those
being studied may have occurred between the first and second
measurements to produce a change in the students' attitudes
toward primary care and family medicine.
3. In test-retest situations, exposure to the first
test may bias or sensitize the subject to the second test.
Since the questionnaire used to measure entering attitudes
was the same as the questionnaire used after the family
medicine clerkship, this source of error could have biased
the results of this study.
4. The design of the research instrument provided
limitations. The questionnaire items were forced choice,
closed-end, and very short answer format. The Attitudes
about Primary Care and Family Medicine section, Part F, used
a Likert scale with the following choices: strongly agree,
tend to agree, not sure, tend to disagree, and strongly
disagree. Variable interpretations of this and other
sections of the questionnaire could potentially distort the
study's results.


measures, parents' annual income greater than $50,000, (c)
students with experience as teachers were distinguished by
more positive attitudes toward primary care; (d) students
supported by parents scored lower on attitudes toward
primary care than those whose major means of support came
from other sources; and (e) the importance students assigned
to 12 career goals did not differentiate between third-year
specialty preference or fourth-year residency choice groups.
Based on their mean attitude scores, students choosing
primary care residencies held more favorable attitudes
toward primary care before and after a clerkship in family
medicine than subspecialty-oriented students but the
subspecialty scores demonstrated greater improvement in
attitudes.
These findings support several educational
interventions. Since the majority of students are capable
of making stable career decisions by the third year,
educational strategies likely to impact on these decisions
should be implemented early in the educational process. The
link between socioeconomic status and specialty choice
suggests that as tuition costs increase, adequate financial
assistance becomes critical for the less affluent students
who are most likely to fill the ranks of primary care
specialties. Early training activities in ambulatory care
settings should be fostered to maintain the interest of
students preferring primary care and to improve attitudes
and attract other students.
IX


-123-
When sociodemographically derived groups of students
were differentiated based on their attitudes toward primary
care, those students who were supported by their parents
were found to hold attitudes dissimilar to students
dependent on other financial resources. The mean rank of
the parent supported group was lower indicating less
positive attitudes toward primary care. This finding is
consistent with the previously discussed measures of
socioeconomic index, parents' annual income and father's
occupation, which were associated with specialty decisions
and lends further support to the premise that economic
security plays a part in specialty decisions.
Students who had work experience in teaching activities
were distinguished from those who had not by their more
positive primary care attitudes. Teachers and others who
work with the public are considered to share some common
characteristics with primary care physicians. Namely, they
generally work with broad segments of society, must
communicate effectively to impart information to others and
are, usually, humanistically oriented. It was concluded
that this finding provided evidence of validity for the
attitude section of the questionnaire.
The relative importance medical students assigned to
career goal statements did not differentiate between third-
year specialty preferences nor fourth-year specialty choice
groups. One explanation for this finding may be the
proportionately large number of students in the undecided


-79-
Specialty Preferences
Prior to a third-year Family Medicine clerkship, the
medical students were asked to indicate their future plans
for residency training. General internal medicine,
pediatrics, family medicine, and other primary care
specialties were the most frequently chosen areas with 47%
of the group. About 24% expected to go into subspecialty
residencies and 29% were still undecided on a specialty at
that point. Table 2 shows the students' career plans by
their preferred specialties.
Residency Plans
The students' first choice of career plans by residency
choice are presented in Table 3. Primary care specialties
were the first choice of career activity of 66% of the
students. Thirty-six percent planned residencies in
internal medicine with 14% choosing family medicine, 7%
pediatrics, and 9% transitional residencies. Approximately
33% of the 1985 graduates elected subspecialty residencies.
The most frequently named subspecialty was obstetrics and
gynecology with 10% of the subjects. Slightly less than
4% opted for residencies in surgery, radiology, and
pathology.


-57
by the Department of Community Health and Family Medicine at
the University of Florida, the Florida Council of Graduate
Education for Family Practice, and the Florida Academy of
Family Physicians (Dallman et al., 1980, p. 833). One of
the purposes of the residency study on attrition was to
delineate characteristics of two groups of residents, the
first group being those who after one year of training
planned to continue on to completion of the residency
program. The second group included residents who indicated
at the end of the first year of residency that they planned
to leave the residency program.
Dallman and his group (1980) mailed a questionnaire to
all incoming residents in Florida Family Medicine Residency
programs in August of 1977 and those entering in August,
1978. The questionnaire was designed to obtain data on each
resident's social, familial, and educational background, as
well as the resident's beliefs and attitudes concerning
family medicine and primary care, and the importance the
resident gave to financial security, autonomy, lifestyle,
and social position.
The second questionnaire was mailed to each resident at
the end of the first year of residency training. The same
attitude and opinion items were included on the second
questionnaire as appeared on the entering form except that
the background items were omitted on the second survey
(Dallman et al., 1980, p. 833).


-14-
6. Items on the questionnaire resembling attitude
statements concerning primary care and family medicine
were representative of generalizations of the domain of
interest.
7. Attitudes and beliefs are not open to inspection
but can be inferred from the answers to statements
pertaining to behaviors and positions toward certain
actions.
Definition of Terms
The terms used in this research study are defined as
follows:
A primary care physician is one who (a) is the
physician of first contact for the patient; (b) makes the
initial assessment and attempts to solve as many of the
patient's problems as possible; (c) coordinates the
remainder of the health care team, including ancillary
health personnel as well as consultants, that are necessary
in dealing with the patient's problems; (d) provides
continued contact with the patient and often his or her
family; and (e) assumes continued responsibility for his or
her care (Petersdorf, 1975). The primary care physician
administers a highly personalized type of care which
coordinates all of the health care needs of the individual
in sickness and in health. Few specialties consider this
comprehensive and continuing care to be their responsibility


-23-
family medicine tended to be slightly older than those
choosing other specialties.
Age and marital status may be interrelated in,that in
general, married students tend to be older. Assuming that
students who choose family medicine share common social
characteristics, Cole, Fox, and Lieberman (1983) examined
questionnaire data obtained from 429 students at a public,
northeastern medical school. The profile of the subgroup of
students who chose family medicine was disproportionately
older, married, and more often Protestant. It may be that
as with socioeconomic status, it is not the age or marital
status per se that effects specialty choice for the older,
married student but rather the economic restraints that make
long residency training difficult.
Hutt (1976) found similar outcomes in Britain.
Marriage had a precipitating influence on specialty choice.
Twice as many married men opted for careers in general
practice. Married women were more apt to take up careers
outside of hospitals or, if inside, to take those not
involving clinical duties.
In a unique approach to gain a better understanding of
physicians' career choices, Skipper and Gliebe (1977)
surveyed not only senior medical students on their plans for
a medical specialty but also asked the students their
perceptions of the influence of their spouses on those
plans. Going one step further, the researchers collected
data from 17 of the 21 wives of the married students. Due


-149-
Carline, J.D., Cullen, T.J., Dohner
Zinser, E.A. (1980). Career
second-year medical students:
Journal of Medical Education,
, C.W., Schwarz, M.R.
preferences of first-
The WAMI experience.
55, 682-691.
&
and
Champion, D.J.
research.
(1970). Basic statistics for social
Scranton, PA: Chandler Publishing Co.
Chandra, P., & Hughes, M. (1984). Factors affecting the
choice of anesthesiology by medical students for
specialty training. Journal of Medical Education, 59,
323-330.
Cole, D.R., Fox, T.G., & Liberman, J.A. (1983). Social
identities of medical students oriented toward careers
in family medicine. Journal of Family Practice, 17,
332-334, 338.
Collins, F., & Roessler, R. (1975). Intellectual and
attitudinal characteristics of medical students
selecting family practice. Journal of Family Practice,
2, 431-432.
Cuca, J.M. (1979). The specialization and career
preferences of women and men recently graduated from
U.S. medical schools. Journal of American Medical
Womens Association, 34, 425-435.
Dallman, J.J., Crandall, L.A., & Haas, W.H. (1980).
Factors affecting residency program dropouts: A
longitudinal study. Journal of the Florida Medical
Association, 67, 833-835.
Department of Community Health and Family Medicine. (1981).
A longitudinal study of family practice education:
Survey of entering students and follow-up survey.
University of Florida, Gainesville.
/
Dube, W.F. (1973). Women students in U.S. medical schools:
Past and present trends. Journal of Medical Education,
48, 186-189.
Fadem, B.H., Nicolich, M.J., Simring, S.S., Dauber, M.H., &
Bullock, L.A. (1984). Predicting medical specialty
choice: A model based on students' records. Journal
of Medical Education, 59, 407-415.
Friedman, C.P., Striter, F.T., & Talbert, L.M. (1978). A
systematic comparison of teaching hospital and remote-
site clinical education. Journal of Medical Education,
53, 565-573.


-114-
Hypothesis 9
The objective for Hypothesis 9 was to test whether a
significant difference existed with respect to attitudes
toward primary care between primary care-oriented medical
students, subspecialty-oriented students, and students who
were undecided. Students were separated into the three
groups on the basis of their stated preferences for a
specialty at the initial administration of the
questionnaire. Information on attitudes toward primary care
was obtained from the 29 attitude statements in the last
section of the questionnaire. The Kruskal-Wal1is one-way
analysis of variance test was used to determine whether the
three groups came from the same sample with regard to their
attitudes. The significance level for the test was set at
£ .01. The resulting statistic was H = 4.57 (d_f 2, N =
70), £ = .10. Since the magnitude of the test probability
was larger than the previously set level of significance, it
was decided that the three groups came from the same or
identical samples with respect to attitudes and, therefore,
the null hypothesis could not be rejected.
Hypothesis 10
Hypothesis 10, there is no difference between medical
students choosing primary care residencies and those
choosing subspecialties with regard to attitudes toward
primary care before and after a family medicine clerkship,
was tested using the Kruskal-Wallis one-way analysis of


DEPARTMENT OF COMMUNITY HEALTH AND FAMILY MEDICINE
COLLEGE OF MEDICINE
UNIVERSITY OF FLORIDA
(904) 392-4321
BOX J-222, MSB
J. HILLIS MILLER HEALTH CENTER
GAINESVILLE, FLORIDA 32610
Attached is the follow-up form for the longitudinal
study in Family Practice Education. I know how busy your
schedule is but it means alot to the validity of the study
to obtain the follow-up. Thank you for your help in this
regard.
Sincerely,
Margaret Duerson, M.Ed.
Curriculum Coordinator
MD/mls
Enclosure
-147-
EQUAL EMPLOYMENT OPPORTUNITY/AFFIRMATIVE ACTION FMPl nvFB


-47-
their male counterparts in (a) length of residency, (b)
employment settings, and (c) intended areas of specialties.
The women planned to spend fewer years in residency training
and wanted salaried practice options rather than private
practice styles. These first two factors could explain the
third--the tendency for women to participate at a higher
rate in primary care specialties as opposed to some of the
nonprimary care specialties which require long training
periods and provide less flexible hours.
This is not to say that males do not have personal
goals that include time for family, leisure activities, and
other pursuits and may consider these interests in making
their choices. Up to this point, however, these goals have
not been evidenced. It may be that as more families become
two career families and more couples share family
responsibilities, the males will express some of these same
needs and demands.
The financial compensation for specialists in
nonprimary care specialties is generally higher than primary
care specialties. Testing for economic incentives as
factors in career choice is difficult. As stated earlier,
people will not usually admit to what might be considered
materialistic motives.
Funkenstein had some success with economic factors in
relation to career choice. Funkenstein's (1978)
longitudinal, prospective study of Harvard medical students
covered a period of five eras from 1958-1976. From his


-85-
Table 5
Association between Medical Specialty Preferences and
Father's Occupation and Parents' Annual Income
Specialty Preferences
Sociodemographic Primary Care Subspecialty Undecided
Characteristics n Row % n Row % n Row %
Father's Occupation
professional/
technical
13
33.33
10
25.64
16
41.03
business/other
21
63.64
7
21.21
5
15.15
Parents' Income
> $50,000
9
20.00
25
55.00
11
25.00
< $50,000
12
62.22
2
11.11
6
26.67


Page
14 Comparison of Specialty Choice Groups Based
on Differences in Preliminary and Follow-up
Career Goal Rankings 110
15 Association between Third-Year Specialty
Preferences and Fourth-Year Specialty
Choices 112
v 11


-138-
A Longitudinal Study of Family 3,-actice Education
Survey of entering students (continued)
ATTIT'JDES ABOUT PRIMARY CARE AMP ~AMLY MEDICINE (continued)
feelings toward eacn statement.
(Circle the nunber of the response which best describes your own personal
The only way to practice good
medicine is to do a complete
history and physical examina
tion eacn visit.
Most patients with emotional
disorders cannot be helped
without spending a lot of
the doctor's time.
Health maintenance is not
as interesting to me as
curative medicine.
Disease prevention should be
the responsibility of public
health departments rather
than the personal physician.
In order to provide the best
care *or his patients, a doctor
should personally perform as
many of the direct patient-
contact services as possible.
Consultants play a secondary
role in the patients total
health care.
People with impaired mental
health are as likely to get
well as oeoole with impaired
pnysiciai health.
Except for certain diseases
specific knowledge pertaining
to disease prevention is so
fragmented that a pnysician
snould limit his efforts to
curative medicine.
2 3 4
2 3 4 5
2 3 4
3 4
21. The provision of services by
specialty oriented pnysicians
should be coordinated and
controlled by primary care
pnysicians.
22. There is no useful researcn
work that can oe done in
primary care practice.
23. Family practices are not as
profitable financially as
specialty practices.
24. The demands of a family
practice leave little
leisure time 'or family
life or recreation.
25. A comolete medical and social
data base on each patient is
important to me.
26. Most patients medical problems
involve aspects of their coping
mechanisms witn life's daily
challenges.
27. Physician's assistants will
play an imoortant role n future
primary care.
28. I would prefer to spend my time
dealing witn patients' medical
proolems rather than their social
or psychological preplans.
29. Physician's assistants snould
handle the acute minor illness
problems in primary care.
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
1 2 3 4 5
Please indicate to the best of your knowledge your future plans for
residency training:
Fanily Medicine
Pediatrics
General Internal Medicine
Other primary care (specify)
Other specialty/subspecialty (specify)
Uncertain
strongly
disagree


-17-
The final chapter, Chapter V, summarizes the findings
of the study, the conclusions, and the implications.
Recommendations for future research in the area of medical
specialty career choice are suggested.


-125-
predictable pattern suggested by reports in the literature
that students attracted to primary care specialties were
patient-care centered rather than interested in financial
security, social position, and professional status (Burkett
& Gelula, 1982; Collins & Roessler, 1975; Leserman, 1978;
Plovnick, 1980).
The discovery of the research interest aligned with
primary care was in conflict with the findings of Burkett
and Gelula (1982) that students attributed greater
importance to helping people than to scientific endeavors.
It could be hypothesized that the primary care grouping
which included family medicine, internal medicine,
pediatrics, and the transitional residency was sufficiently
large to accommodate those with research interests.
Residency selection not being the final career decision
point, students choosing primary care residencies may later
pursue an academic career or subspecialize in a primary care
area both of which include research potential.
Just as the values placed on career goals did not
discriminate specialty groups, neither did attitudes of
medical students gauged prior to the family medicine
clerkship and following the clerkship distinguish the
specialty groups. Changes in attitudes from before to after
the clerkship were analyzed. The conclusion drawn from
analysis of these data was that even though the clerkship
experience could not be shown statistically to modify


-82-
Results of the Data Analysis for Testing
the Hypotheses
Hypothesis 1
There is no association between sociodemographic
characteristics of medical students and their medical
specialty preferences expressed in the third year of medical
school. The chi-square test was used to determine
separately for each sociodemographic variable whether
medical specialty preferences of third-year medical students
were dependent on sociodemographic characteristics. Medical
specialty preferences were categorized as primary care,
subspecialty, and undecided based on the students' responses
to the questionnaire item, Please indicate to the best of
your knowledge your future plans for residency training.
Since multiple statistical tests were performed, a
conservative overall significance level of £ = .01 was used
to decrease the probability of falsely rejecting the null
hypothesis. The results of the chi-square procedures used
to test the first set of variables related to Hypothesis 1
are reported in Table 4.
The overall significance level of < .01 was not
achieved in the comparison of sociodemographic variables and
third-year specialty preferences of medical students.
Therefore, the null hypothesis was not rejected.
However, one sociodemographic category, parents' annual
income, was significantly related to specialty preference at
the < .01 level. This test produced a chi-square value


DEPARTMENT OF COMMUNITY HEALTH AND FAMILY MEDICINE
COLLEGE OF MEDICINE
UNIVERSITY OF FLORIDA
(904) 392-4321
BOX J-222, MSB
J. HILLIS MILLER HEALTH CENTER
GAINESVILLE, FLORIDA 32610
MEMORANDUM
TO: Phase B Medical Students
FROM: William L. Stewart, M.D.
RE: Clerkship Evaluation
The Department of Community Health and Family Medicine is
conducting a longitudinal study of specialty choices. This
study will correlate certain attitudes with specialty choices.
The information the survey contains will not be used in any
way that will identify the individuals. The information will
be used as group statistics.
Your name and social security number will be replaced by a
code number for entry into the computer. The code number
is to enable the Department to identify the specialty choice
or residency choice. No one will have access to the code
except the investigators.
There will be an abbreviated follow-up form to be completed
after the Family Medicine Rotation. We would like to thank
you for your assistance in this study. If you have any ques
tions, feel free to ask us.
This survey is not mandatory, but we would appreciate your
participation. If you decide not to participate, we would
appreciate your notifying us of that decision. An envelope
is enclosed. Thank you.
-143-
EQUAL LMPLOYMENT OPPORTUNITY/AFFIRMATIVE ACTION EMPLOYER


-131-
experience in primary care to warrant rejection of the null
hypothesis. Even so, the findings are preliminary and
require further investigation. It is recommended that the
study of these variables be continued longitudinally to
obtain a larger sample of responses. This would possibly
produce stable and definitive conclusions. In addition,
factor analysis of the attitude statements from the
questionnaire may be possible with a larger sample size thus
reducing the number of questions and delineating the
questions most related to primary care choices.
Clerkship Experiences
In addition to the continued study of medical students'
individual characteristics, it is recommended that the
influence of educational experiences (i.e., family medicine
clerkships) continue to receive research attention.
Specialty choice is at least in part a time-dependent
process so that the time factor must be taken explicitly
into consideration. The time of exposure to family medicine
in relation to other clinical experiences is presumed to be
an important factor. It was not within the scope of this
study to examine subgroups of students based on their
rotation sequence. It is recommended that the application
of the questionnaire be continued immediately before and
after the clerkship so that the effects of the clerkship


-24-
to their own career goals, the women preferred that their
husbands settle in a medium-sized city or suburban fringe
area. They also preferred group practice to solo practice
so that the husbands would have more time for family
activities. While the specialty choices expressed by the
wives was varied, some of the wives viewed family medicine
as less desirable because it would provide less free time
than other specialties. Researchers conclude that if
primary care specialties such as family medicine are to
appear attractive to wives of medical students as well as
the students themselves, there will need to be some attempt
to also convince the spouse of the good points of the
specialty.
The small numbers of subjects involved in the Skipper
and Gliebe (1977) study preclude definitive statements. It
does, however, suggest that as more marriages become two
career families, there will be pressures on career choices
to accommodate both husbands and wives.
Medical schools have the mixed blessing of having a
large applicant pool of highly qualified candidates from
which to choose. While medical schools do not publish the
admission criteria, it is generally known that students must
have demonstrated academic success by virtue of their
undergraduate grades and their scores on the Medical College
Admission Test (MCAT). Strong interest and a proven
aptitude for science are given more weight than other
factors in the selection process. This emphasis on


-31-
unique contribution to health care. As the number of women
physicians increases, medicine may take on some of the
characteristics common to female dominated occupations. The
changing gender mix of medical school classes merits
continued study.
Personality and Attitude Factors
It is not surprising given the homogeneity of medical
students' intellectual ability and some background
characteristics that the ability to associate specialty
choice with these variables has been inconsistent. Because
the sociodemographic data have been so capricious, some
investigators have sought to determine whether measureable
differences in personality variables are predictive of the
type of specialty chosen.
Zuckerman (1977) in a comprehensive review of the
literature found that psychological characteristics
distinguished psychiatrists and surgeons from other
specialty types but these factors have been less fruitful in
distinguishing other types of specialties. He concluded,
"The utility of personality characteristics to predict
variation in career choice appears to be limited" (p. 1083).
Hutt's (1976) review of previous research supports the
consistent findings of personality differences between
psychiatrists and surgeons. She observed that "there is
evidence that these two choices are made early and are more
consistently adhered to than others" (p. 466). It may be


LIST OF TABLES
Table Page
1 Summary of Sociodemographic Characteristics
of Subjects 77
2 Medical Specialty Preferences of Third-Year
Medical Students 80
3 Residency Choice of Fourth-Year Medical
Students 81
4 Association between Third-Year Specialty
Preference and Sociodemographic Character
istics 83
5 Association between Medical Specialty
Preferences and Father's Occupation and
Parents' Annual Income 85
6 Association between Medical Specialty Choice
of Fourth-Year Students and Father's
Occupation 87
7 Association between Fourth-Year Specialty
Choice and Sociodemographic Characteristics... 89
8 Association between Medical Specialty Choice
of Fourth-Year Students and Parents' Annual
Income and College Major 91
9 Sociodemographic Groupings of Medical
Students and Preclerkship Attitudes 94
10 Sociodemographic Groupings of Medical
Students and Postclerkship Attitudes 95
11 Results of Preclerkship Career Goal Rankings
and Three Preference Groups 100
12 Results of Preclerkship and Postclerkship
Career Goal Rankings and Two Specialty
Choice Groups 102
13 Correlation Coefficients for Career Goal
Rankings and Attitude Scores 105
vi


BIOGRAPHICAL SKETCH
Margaret Craig Duerson was born in Louisiana.
Following primary and secondary education in Mississippi,
she returned to the Bayou state where she began lower
division studies at Louisiana State University in Baton
Rouge. She married Goodloe Kearney Duerson, a graduate in
engineering. For several years she devoted full time to her
growing family and community activities.
Resuming her education at the local community college
where the family had settled in Lake City, Florida, her
interest in health care and working with people drew her to
the nursing program. She received the Associate of Arts
degree and the Associate of Science in Nursing degree
graduating summa cum laude from Lake City Community College
in 1970.
Mrs. Duerson worked for three years as a critical care
unit nurse in the local veterans administration hospital.
Desiring to broaden her knowledge base, she continued upper
division nursing studies at the University of Florida where
she graduated with high honors earning a Bachelor of Science
in Nursing degree in 1974.
Upon graduation she joined the University of Florida
College of Medicine as coordinator of clinical research for
-154-


-27-
Academic measures made later in the medical school
process have not proved much better in producing consistent
results. Two studies conducted in the same state but at two
separate medical schools used National Board of Medical
Examiners (N3ME) examination scores to study the
relationship to career choice. The NBME exam results are
available late in the third year or early fourth year of
medical school. One group of investigators who examined the
effects of NBME scores for 628 students at the University of
Medicine and Dentistry of New Jersey-New Jersey Medical
School, found that high scores on the NBME Part II subtests
in psychiatry, medicine, and surgery were associated with
selection of these specialties (Fadem et al., 1984).
Contrary to these findings, Rosevear, Tickman, and Gary
(1985) did not find a positive relationship between medical
students' career choice and their performance on the NBME
Part II subtest scores for 347 graduates (1979-1983) of the
University of Medicine and Dentistry of New Jersey-Rutgers
Medical School.
Factors predictive of specialty choice which occur
early in the educational process have the advantage of
providing more opportunity for impact and intervention.
Unfortunately, the early parameters such as undergraduate
grade point average, MCAT scores, and other cognitive
measures do not consistently discriminate between those
choosing primary care specialties and those choosing
nonprimary care specialties. It may be that because all


-110-
Table 14
Comparison of Specialty Choice Groups Based on
Differences in Preliminary and Follow-up Career
Goal Rankings
Career Goals+
H
£
Set 1
a. helping to solve community's medical
and social problems
1.23
.27
b. gaining respect and friendship of
other physicians
4.38
.04
c. developing long-term relationships
with patients
.06
.81
d. contributing to medical knowledge
through research
1.44
.23
e. assuring financial security for
self and family
.00
1.00
f. allowing me to live in preferred
area and style
.80
.37
Set 2
a. providing me with capital for
investment
.02
.89
b. being acknowledged by other
physicians as a leader in health
care for my patients
.08
.78
c. having time for family and non
professional activities
.63
.43
d. gaining trust and confidence of
my patients
.55
.46
e. making me a respected member of
my community
3.59
.06
f. actively conducting research in my
areas of interest
.82
.36
+Paraphrased from original questionnaire


-60-
answer as when asked age and major subject of undergraduate
education.
The importance the student placed on financial
security, autonomy, lifestyle, social position in the
community, and position in the peer group was examined in
Section C, Career Goals. Students were asked to rank in
order of importance items that relate to career goals. The
most important goal was assigned the rank of 1, the second
most important goal received a rank of 2, and so forth with
no choice being used more than once.
Section D, Medical Education Content, lists eight items
pertaining to medical education. Students ranked the items
from 1 to 8 according to how important each was considered
to be in producing the best physicians.
Improving Health Status, Section E, lists five
different approaches to improving the health status of
people not receiving adequate health care. The student
ranked them in order of his/her opinion as to their
effectiveness with the rank of 1 equaling the most effective
and the rank of 5 being least effective in improving health.
The Attitudes Toward Primary Care and Family Medicine
section, Section F, contains 29 statements which reflect the
goals and philosophy of primary care and delineate family
medicine and primary care from specialty-oriented medicine.
The attitude section concentrates on the following areas:
(a) preventive medicine, (b) continuity of care, (c)
managing psycho-social problems, (d) ambulatory versus


-152-
Rakel, R.E., & Pisacano, N.J. (1984). The family
physician. In R.E. Rakel (Ed.), Textbook of family
practice (3rd ed., pp. 3-20). Philadelphia: W.B.
Saunders.
Rezler, A.G. (1974). Attitude changes during medical
school: A review of the literature. Journal of
Medical Education, 49, 1023-1030.
Robbins, L., Robbins, E.S., Katz, S.E., Geliebter, B., &
Stern, M. (1983). Achievement motivation in medical
students. Journal of Medical Education, 58, 850-858.
Roos, N.P., & Roos, L.L. (1980). Medical school impact on
student career choice. Evaluation and the Health
Professions, _3, 3-19.
Rosenblatt, R.A., & Alpert, J.J. (1979). The effect of a
course in family medicine on future career choice: A
long-range follow-up of a controlled experiment in
medical education. Journal of Family Practice, 8,
87-91.
Rosevear, G.C., Tickman, M.S., & Gary, N.E. (1985).
Relationship between specialty choice and academic
performance. Journal of Medical Education, 60, 640-
641.
Samra, S.K., Davis, W., Pandit, S.K., & Cohen, P.J. (1983).
The effect of a clinical clerkship on attrition of
medical students toward anesthesiology. Journal of
Medical Education, 58, 641-647.
Schroeder, S.A. (1984). Western European responses to
physicians oversupply. 2. Lessons for the United
States. Journal of the American Medical Association,
252, 373-384.
Skipper, J.K., & Gliebe, W.A. (1977). Forgotten persons:
Physicians' wives and their influences on medical
career decisions. Journal of Medical Education, 52,
764-765.
Steinwachs, D.M., Levine, D.M., Elzinga, D.J., Salkever,
D.S., Parker, R.D., & Weisman, C.S. (1982). Changing
patterns of graduate medical education. 2. Analyzing
recent trends and projecting their impact. New
England Journal of Medicine, 306, 10-14.
University of Florida, College of Medicine. (1985).
1986 College of Medicine catalog. Gainesville:
University of Florida.
1985-


-98-
students whose medical care was provided by a family
practitioner/generalist from those students who did not
check this alternative answer. The Kruskal-Wal 1 is test
statistic was H (d_f 1, N = 86) = 3.73. Based on the mean
rank score, students who were cared for by family/general
practitioners held more positive attitudes toward primary
care.
Work experience. From a list of job descriptions, the
students were asked to check the areas in which they had had
work experience. Preclerkship attitude scores of students
who had dealt with the public either in sales or public
relations were different from the ones who did not engage in
this type of work at a significance level of .006. The test
statistic was H (df 1, N = 103) = 7.52.
The attitude scores of students who had work experience
as teachers were significantly different from those who had
not participated in activities in both the preclerkship
samples and the postclerkship measurements. Preclerkship,
the observed H of 3.89 (d£ 1, N = 103), was significant at
the £ = .05. Postclerkship, the test statistic of 9.19 (df
1, N = 86), had a probability of .002. The mean rank scores
for those with teaching experience was higher in both
preclerkship and postclerkship measurements indicating
higher attitude scores.


-119-
of students from the latter occupation group indicated a
delay in specialty choice decisions in the third year of
medical school.
There was a significant difference between college
science majors and nonscience majors with respect to
specialty choice (£ = .0067). This association should be
viewed with caution, however, as such a high percent of the
students in this sample were science majors.
Scores on attitudes toward primary care differentiated
several of the sociodemographic groupings. Students
identified as having had work experience with the public
were distinguished by their preclerkship attitude scores
(£ = .006). Postclerkship (£ = .002) attitudes
differentiated students who had work experience involving
teaching. There were also differences in attitudes toward
primary care among those whose major source of present
support was their parents versus other financial resources
(£ = .004). Interpreting the mean rank score, students
whose major financial support did not come from their
parents, those who had work experience in teaching and in
work with the public had more favorable attitudes toward
primary care.
The relative importance students placed on career goal
statements did not distinguish between specialty preference
groups in the third year or fourth-year specialty choice
groups. Neither did the difference in goal rankings from


Table 13
Continued
Preliminary Goal
Rankings
Follow-Up Goal
Rankings
Career Goals+
Preliminary
Attitudes
Follow-Up
Attitudes
Follow-Up
Attitudes
Set
2
a.
providing me with capital for investment
.19
. 10
.34***
b.
being acknowledged by other physicians as
a leader in health care for my patients
.003
.10
i

o
kD
c.
having time for family and non-professional
activities
-.05
-.06
.06
d.
gaining trust and confidence of my
patients
-.14
-.13
-.13
e.
making me a respected member of my
community
.05
.007
.09
f.
actively conducting research in my areas
of interest
-.05
-.06
-.31**
*£
**JD
k k p
.05
.01
.001
+Goal
phrases
paraphrased
from original questionnaire
-901-


APPENDIX B
MEDICAL STUDENT FOLLOW-UP QUESTIONNAIRE


-115-
variance procedure and a £ .01 level of significance. The
subjects were distributed into two groups according to their
residency match choices of specialties. Data on their
attitudes were obtained from the 29 questions addressing
attitudes toward primary care and family medicine measured
before and after the family medicine clerkship. To test the
hypothesis, the change in each student's attitude score was
computed. The mean attitude score of 103.09 (SJO = 7.17) for
students choosing primary care specialties in the initial
measurement was very close to their mean score at follow-up
(M = 103.22, SD = 8.18). Those students taking subspecialty
residencies scored a mean of 100.36 (SD = 8.51) in the
initial sampling of attitudes toward primary care, while at
follow-up their mean score increased to 103.11 (SD = 10.69).
The comparisons of the two groups with respect to their
attitude change scores yielded an H value statistic of 5.7
with 1 degree of freedom and a probability of .02. The mean
rank score for primary care medical students was 37.47 and
51.25 for students choosing subspecialty residencies. Thus,
the subspecialty-oriented students' attitude scores improved
more from preclerkship to postclerkship. However, it was
determined that the two groups, primary care and
subspecialty students, were not different with respect to
their attitudes before and after the clerkship experience at
the < .01 level of significance. Therefore, the null
hypothesis was not rejected for Hypothesis 10.


-89-
Table 7
Association between Fourth-Year Specialty Choice and
Sociodemographic Characteristics
Sociodemographic
Characteristics
df
X2
Value
Significance
Sex
1
.377
.54
Marital Status
1
1.506
.22
Religious Orientation
4
1.640
.80
Religious Importance
2
1.219
.54
Race
1

.54 +
Citizenship
1

. 26 +
Hometown Location
1
.128
.72
Hometown Size
3
2.228
.52
Hometown Economic Base
4
6.326
.18#
Fathers Occupation
1
.782
.38
Mother's Occupation
1
2.667
*10#
.57#
Parents' Religion
8
3.046
Parents' Income
1
4.602
.03*
Source of Present Support
1
.090
.76
Indebtedness
1
.711
.40
Amount of Debt
1

. 67 +
General health
1

1.00 +
Public or Private High School
1
1.453
.23
Public or Private College
1
.153
.70
Family's Medical Care Provider
Family Practitioner/General
1
1.212
.27
Internist
1
.266
.61
Pediatrician
1
3.142
.08
Other
1
.000
1.00
Work Experiences
Medical
1
.000
1.00
Helping Activities
1
.000
1.00
Technical
1
.019
.89
Teaching
1
.026
.87
Public
1
.928
.34
Unskilled
1
3.157
.08
Other
1
.559
.45
College Major
1
7.338
.01**
Level of Education
1

. 30 +
*£ < .05
**£ < .01
+Fisher's Exact Test
#
Agresti-Wackerly Significance Level


-30-
roles, Weisman (1984) found that, indeed, in the graduates
of 1970 through 1976 higher proportions of women in medical
schools were associated with smaller proportions of women
entering traditionally female specialties. There was only a
weak effect of gender on first-year residency choice.
There was disagreement in the studies as to whether
women were becoming less traditional in their specialty
choice. Both males and females expressed a strong interest
in the primary care specialties indicating some convergence.
However, there are also major differences in the factors
perceived to be important in choosing a specialty. Women
expressed a preference for patient contact specialties where
patients are involved in solving their own health problems
and were willing to forego career time for more family life
(Bergquist et al., 1985; Zimny & Shelton, 1982). Males
showed a high preference for complex medical problems and
technical procedures which would necessitate functioning in
highly specialized areas (Zimny & Shelton, 1982). Males
also expected to have higher incomes than women expected but
it is not clear whether this expectation dictated the career
choice or was a consequence of it (Bergquist et al., 1985).
The eventual impact on medicine of greater
participation by women is unclear. The fact that primary
care specialties are popular with women may help to correct
the problem of overspecialization but it is not the entire
solution to the national health problems. This solution
oversimplification of the problem and women's
would be an


-104-
achieve through a career in medicine and 6 being the least
important goal. A summative score for each student was
derived from their responses on the 29 attitude statements.
A high numerical score indicated a favorable attitude toward
primary care and family medicine. The statistical index
used for finding the relationship between the data sets was
the Spearman's rho correlation coefficient. The overall
significant level was set at £ .01. Although some variables
reached significance, the overall significance level
established prior to the analysis was not achieved.
Evidence was considered insufficient to reject Hypothesis 5.
The results of this analysis are found in Table 13.
Significant correlations occurred in two of the items
included in the first set of career goal statements and
attitudes expressed before and after the clerkship. Two
correlations reached significance among the second set of
six career goal rankings and attitude scores. The direction
as well as the degree of the relationship was considered.
There was an inverse relationship between the rankings of
the career goal statements and the attitude scores.
Therefore, when the variables were highly related in a
negative direction, the career goal was very important to
students with a highly favorable attitude toward primary
care and family medicine. Conversely, when the direction of
the correlation was positive, the career goal was ranked low
in importance to those with highly favorable attitudes and
more important by those students who have lower attitude


-61-
hospital-based care, (e) broad spectrum of medical care,
(f) attitudes toward consultants and paramedical personnel,
and (g) perspectives toward research in primary care
settings.
Student respondents were asked to indicate whether
they strongly agree, tend to agree, are not sure, tend to
disagree, or strongly disagree with each of the 29
statements on attitudes and beliefs concerning primary care
and family medicine. A Likert scale of numerical values was
assigned to each response depending on the degree of
agreement or disagreement. Strongly agree received a value
of 1 while strongly disagree was valued at 5.
The last section of the entering survey of medical
students asked the respondent to indicate to the best of
his/her knowledge future plans for residency training. The
alternatives from which the student could choose were (a)
family medicine, (b) pediatrics, (c) general internal
medicine, (d) other primary care, (e) other specialty/sub
specialty, or (f) uncertain. Students answering under the
category "other primary care and other specialty/sub
specialty" were asked to specify the specialty area by
writing in a line beside the choice indicated. The last
section was included on all of the entering questionnaires
except the questionnaires for Rotation 2. It was
inadvertently omitted from this group's questionnaires.
Thus, 20 students did not have the opportunity to express
their specialty preferences. A comparison of the
sociodemographic characteristics of the students who did not


-55-
student is lodged in the preceptor's home so that she or he
is provided an opportunity to view and understand non
medical aspects of the physician's roles and obligations to
the community as well as the usual medical responsibilities.
There are elective and required assignments in two
University of Florida affiliated family practice centers,
the Family Practice Center in Gainesville and St. Vincent's
Family Practice Center in Jacksonville. Here students work
with residents in family practice training and faculty
members to become acquainted with the health care delivery
in urban areas.
The four years of undergraduate medical education
leading to the M.D. degree at the University of Florida are
divided into three parts, Phases A, B, and C. During Phase
A, the first 12 months of study, students are presented the
core of basic medical science courses necessary for clinical
training. The course schedule includes biochemistry, human
genetics, anatomy, bacteriology, immunology, virology,
physiology, neuroscience, human organ system development,
and human behavior.
Phase B is divided into preclinical and clinical
experiences. The first 27 weeks of the second year students
take pathology, pharmacology, physical diagnosis, and
interviewing. A course in social and ethical issues is
offered in the last part of the 27-week block preceding the
clinical portion of Phase B.
The major portion of Phase B is the 12-month long
clinical clerkship period in which students rotate in groups


-11-
imbalances, it is appropriate to study the relationship
between the undergraduate educational process and specialty
and subspecialty choices. A knowledge of the variables
within the educational environment which influence and
predict specialty and subspecialty career choices would
provide guidance for medical educators, policy makers, and
others vitally interested in correcting the present medical
manpower imbalances and planning for future needs.
Delimitations, Limitations, and Assumptions
Delimitations of This Study
This research study was delimited to
1. Medical students who were in their third year of
training at one state-supported medical school in the
southeast in the years 1983-84 and who participated in the
residency match in 1984-85.
2. Factors affecting medical specialty choice at one
university medical school as represented by the group of
students sampled.
Limitations of This Study
There were several limitations that must be recognized
in conducting this study:
1. The results of this research project were
characteristic of this study only. Because of the unique
characteristics of samples and the variations inherent in


-74-
categories at the nominal level, the chi-square test was a
suitable one.
8. Hypothesis 8, there is no difference in medical
students' attitudes toward primary care before and after a
family medicine clerkship, was tested with the Wilcoxon
matched-pairs-signed-ranks test. In this case, the medical
students acted as their own controls in a before-and-after
design. Another assumption is that the level of measurement
be interval-level. As the attitude scores before-and-after
the clerkship were ordered in a numerical scale, the
information was judged to be at least equal-appearing. A
definite advantage of the Wilcoxon test is that it considers
not only the direction of the difference in scores but also
the magnitude of the change. As Champion (1970) suggested
for sample sizes exceeding 25, an alternative formula, a
modification of the Z test, was used to complete the
Wilcoxon matched-pairs-signed-ranks test (p. 169).
9. The Kruskal-Wallis one-way analysis of variance
test was judged to be the proper procedure to test whether
attitude scores of one specialty preference group were
significantly different from the two other specialty
preference groups of third-year medical students. Since the
three groups of specialty preference could be considered
independent and the attitude scores represented at least
ordinal-level data, the Kruskal-Wallis test was considered
the correct method for Hypothesis 9.
10. Hypothesis 10 was also tested with the Kruskal-
Wallis test. The hypothesis was that there is no difference


CHAPTER III
METHODOLOGY AND INSTRUMENTATION
Setting for the Study
The study was conducted at a large southeastern
university college of medicine, the University of Florida
College of Medicine. Unless otherwise specified, the
descriptive information pertaining to the setting for the
study was summarized from the College of Medicine catalog
(University of Florida, 1985). The University of Florida is
a public, state-funded institution established in 1906. The
College of Medicine, a component college of the J. Hillis
Miller Health Center of the University of Florida, admitted
its first class in 1956. College of Medicine educational
offerings include undergraduate medical courses leading to
an M.D. degree, graduate medical education experiences,
residency programs in various specialties and
subspecialties, graduate courses in the basic medical
sciences leading to a Ph.D., and postgraduate fellowships
in clinical and scientific areas.
Medical school faculty and students are housed in the
J. Hillis Miller Health Center and the adjoining Shands
Hospital located on the main campus of the University of
Florida and in the Gainesville Veterans Administration
-53-


-84-
significant at the .0004 level. It should be noted that the
association between father's occupation and specialty
preference attained a probability of .02.
Father's occupation. Student responses to the item,
indicate your parents' occupations while you were growing
up, were used to form two occupation groups for father's
occupation: (a) professional/technical and (b) business/
other. The professional/technical occupational group
included the students' responses to the categories of
physician, other health care occupation, and other
professional work. The group labeled as business/other was
composed of responses to occupational categories of
business, clerical/sales, skilled worker, farmer/farm
worker, and other. The two occupational groups were
compared with the three medical specialty preference levels:
(a) primary care, (b) subspecialty, and (c) undecided. The
chi-square statistic, X2 (d_f 2, N = 72) = 7.727, £ = .02,
indicated that the probability of the observed association
between medical specialty preference groups and fathers'
occupations being a function of chance was 2 in 100 or less.
Table 5 shows that for students whose fathers were in
professional/technical occupations when they were growing
up, 33% preferred primary care specialties, 26% indicated
subspecialty preferences, while 41% were undecided as to
their eventual specialty choice. Of the third-year students
whose fathers engaged in business and other types of work,
64% preferred primary care, 21% indicated subspecialty


-21-
other researchers found a positive relationship (Anderson,
1975; Zuckerman, 1977). Where positive associations between
social class and specialty choice were found, the higher the
socioeconomic status of the family, the more likely the
student was to choose specialization over general practice.
As Zuckerman (1977) pointed out "the economic security and
availability of financial assistance associated with higher
social class may encourage pursuit of the additional
training required for specialty practice" (p. 1082) .
Generally, the primary care specialties require fewer years
of residency training than the subspecialties and thus
decreased cost of education and more rapid entry into wage
earning status.
Medical students whose fathers and mothers are
physicians constitute an interesting group with regard to
social status, especially since they seem to be over
represented in medical schools based on the population
census. Gough and Hall (1977) observed that approximately
16% of students in American medical schools come from
medical families compared to a census estimate of 0.4%
physicians among employed adult males. In their study of
1195 students at the University of California, San
Francisco, School of Medicine, there were 162 from medical
families. These investigators noted that there was a
statistically insignificant shift away from family medicine
and general practice by the subgroup of physicians'


-73-
analysis for Hypothesis 4 which sought to determine the
difference between the three specialty preference groups in
the third year of medical school and the two specialty
choice groups from the fourth year with respect to career
goal rankings. The career goal data met the primary
assumption of this procedure, namely that it be ordinal-
level measurement as well as the assumption that the samples
were independent.
5. Spearman's rho measures the degree of association
between two sets of ordinal-level data. Therefore, for
Hypothesis 5 it was considered an appropriate statistical
procedure to investigate whether or not medical students'
rankings of career goals were associated with their
attitudes toward primary care before a family medicine
clerkship and again, after the family medicine clerkship.
6. Hypothesis 6 was tested with the Kruskal-Wallis
technique. In the computation, the difference in students'
rankings of career goals before and after a family medicine
clerkship were obtained. These difference scores were then
entered into the Kruskal-Wallis test to decide whether the
medical specialty groups were from different populations.
Since the scores could be considered at least ordinal data,
the Kruskal-Wallis was appropriate to the data.
7. Hypothesis 7 tested whether the association between
medical students' third-year specialty preferences and their
fourth-year specialty choices was different from what might
be expected by chance. As the data were in discrete


I certify that I have read this study and that 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.
Ld /'
Jame¡a W. Hensel, Chairman
Professor of Educational Leadership
I certify that I have read this study and that 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.
Forrest W. Parkay
Associate Professor o
Leadership
Educational
I certify that I have read this study and that 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.
Small
Professor of Irttmuno.
Microbiology
Medical


-128-
The findings of this study have implications which
extend beyond medical education to general professional
education requiring field experiences. Such experiences
have the potential to maintain and/or improve students'
attitudes and should be planned accordingly.
Recommendations
Several factors lead to the recommendation for
continued, similar investigations. There is very little
research-substantiated data available concerning the
mechanisms through which medical education influences
specialty choice patterns and how the educational
environment and students' background, values, and attitudes
interact. Medical education is called upon to intervene in
a number of areas to meet societal needs and one of these is
the maldistribution across specialties. Educational
policies addressing this area of concern must rest on
sounder ground than is possible given the present
preliminary and relatively immature state of the research to
date.
Sociodemographic Chacteristics
Medical students are a highly selected group. Much of
the variation in their backgrounds is screened out beginning
as early as high school, again at the college level,
and


-o -
specialty choices as evidenced by fourth-year residency
selection.
3. There is no difference between medical students
grouped by sociodemographic characteristics according to
their attitudes toward primary care measured before or after
a third-year family medicine clerkship.
4. There is no difference between groups of medical
students identified by medical specialty preference in the
third year or medical specialty choice in the fourth year
with respect to their rankings of career goals.
5. There is no relationship between medical students'
rankings of career goals and their attitudes toward
primary care before or after a third-year family medicine
clerkship.
6. There is no difference between the groups of
medical students choosing primary care and those choosing
subspecialties in the fourth year with respect to their
rankings of career goals before and after a family medicine
clerkship.
7. There is no association between medical students'
specialty preferences expressed in the third year and their
career specialty choices as evidenced by fourth-year
residency selection.
8. There is no difference in medical students'
attitudes toward primary care before and after a family
medicine clerkship.


-46-
Career Factors
The factors examined to this point have been intrinsic
characteristics of the students and the medical school
environment but there are potential extrinsic influences
worth consideration. Hutt (1976) wrote, "Curiously enough,
these have received considerably less attention than the
earlier groups although it is, of course, the career factors
which policy-makers could most easily use to alter the
distribution of doctors between specialties" (p. 468).
Extrinsic factors include such areas as working conditions,
economic incentives, and societal demands or needs.
There is considerable diversity between specialties
with regard to working conditions. Some specialties such as
general surgery, pediatrics, and obstetrics require long,
erratic hours with responsibilities at night and on
weekends. Other specialties are more amenable to regular 9-
to-5 schedules. Style of practice, solo, group, or
hospital, is related to specialty. For example, primary
care spcialties are principally ambulatory care practices
with little emphasis on hospital practice. Other
specialties such as surgery are principally hospital-based
practice.
As was noted earlier, women are willing to forego
career time for more family time (Bergquist et al., 1985;
Zimny & Shelton, 1982). They want more flexible hours and
will usually opt for specialties that allow for that time.
In Cuca's study (1979), women's career plans differed from


-5-
attitudes toward primary care. Observations obtained from
administration of a multidimensional questionnaire were used
to explore (a) the association between sociodemographic
characteristics of medical students, their career goal
rankings, and their attitudes toward primary care and the
students' specialty preferences or specialty choices; (b)
whether significant differences in sociodemographic
characteristics existed between medical students with
favorable and unfavorable attitudes toward primary care; (c)
the relationship between the relative importance students
assigned to career goals and their favorable or unfavorable
attitudes toward primary care; (d) the extent to which
attitudes toward primary care changed after a clerkship in
family medicine; and (e) the association between medical
students' career preferences in the third year and their
actual career choice in the fourth year as measured by
residency specialty.
Hypotheses
This study involved the testing of the following
hypotheses:
1. There is no association between sociodemographic
characteristics of medical students and their medical
specialty preferences expressed in the third year of medical
school.
2. There is no association between sociodemographic
characteristics of medical students and their medical


A LONGITUDINAL STUDY OF FAMILY PRACTICE EDUCATION
SURVEY OF ENTERING STUDENTS
NAME:
SOCIAL SECURITY NUMBER:
STUDENT PROFILE
BACKGROUND
1. Personal Pata
a. Age
b. Sex: Male Female
c. Marital Status: Married Single Other
d. Religion:
(i)Orientation:
Protestant Catholic Jewish
Other (specify) No preference
(ii)How rtuch importance does religion have in your life?
Very ix>ch Same
A little None at all
e.Racial Background:
Black Oriental
American Mexican
Indian American
White
Puerto
Rican
f.Citizenship:
U.S. Canadian Other
specify
2. Environment
a."Hometown* comnunity (pre-college years)
(i) Location
Midwest Northeast West Coast
Rocky Mtn/ South/ Foreign
Southwest Southeast Country
(ii) Size (pooulation)
_ Less than 2,500 25,000 to 99,999
2,500 to 24,999 100,000 to 500.COO
__ Above 500,000
(iii)Economic Base
fanning mining business
heavy light mixed
industry industry economy
b.Family
(1) Size
Number of children (including yourself)
If more than 1, give your position, e.g., third of five
children
(ii) Indicate your parents* occupations while you were
growing up.
Father Mother
Physician
Other health care
occupation
Other professional work
Business
Clerical/Sales
Skilled worker
Unskilled worker
Farmer/Farm worker
Housewife
Other (specify)
(ill) Parents' religion write "P* and "T for
"Father" and "Mother" on appropriate lines.
Protestant Catholic Jewish
Other (specify) No preference
c. Finances
(1) Source of present support (major contributor):
Self Spouse & Self Parents
Spouse 50* 5<*> Other
(ii) Indeotedness (greather than $500)
res No
If "Tes", indicate by checking the appropriate
catesory- greater
Less than SI 000 51000 SSOOO than $5000
(iii) Parents' Comoined Annual Income: Check the
appropriate cagegory:
less than $10,000 to $25,XO Above
$10,000 $25,000 to $50,000 $50,000
d. Health and Medical Care
(1) Do you consider yourself in excellent general health?
Yes No
If "No", what major health problem(s) do you have?
(11) Who provided primary care for your *aaily? Check more
than one if applicable.
Family/General Practitioner Internist
Pediatrician Other (specify)
Education
a. High School: Public Private
b. College:
(1) Private Public
(11) Major subject
Minor subject (if applicable)
c. Level of Education:
(1) List all degrees obtained at college level and above:
(11) List study at college level and above that stopped
before degree was obtained:
Work Experience Check areas in which you have had experience:
a. Medical care setting (e.g. hospital orderly, health
aide, clinical laboratory assistant)
b. "Helping* activities (e.g., camp counselor, social
service. Peace Corps)
c. Technical work, not related to actual medical care.
(e.g., laboratory technitian, research assistant,
engineering assistant)
d. Teaching
e. Dealirag with public (e.g., sales, public relations,
press)
f. Unskilled work
q. Other (specify)
-136-


-124-
group of third-year specialty preferences. The undecided
group contained a mix of students who would eventually align
themselves with either subspecialist or primary care.
Assuming that these two groups, primary care and
subspecialty, actually have different values, goals, and
attitudes, the undecided group would have reflected a
combination of these. There is no ready explanation for the
results obtained with the fourth-year specialty choice
groups on this variable. Perhaps the most plausible
explanation for these data is that this sample of medical
students was homogeneous with regard to the career goal
statements used in the questionnaire and what may have
differed was the students approach to attaining the goals.
Significant correlations occurred between several
career goal statements and attitudes toward primary care.
Students who held unfavorable attitudes toward primary care
gave higher rankings to the following statements: (a)
assuring financial security for myself and my family, (b)
allowing me to live in the type of area and lifestyle that I
prefer, and (c) providing me with capital for investment in
business, real estate, the stock market, etc. Students with
favorable attitudes toward primary care gave priority to
(a) helping lead my community to the solution of social and
medical problems, (b) developing long-term intensive
relationship with my patients, and (c) actively conducting
research in my areas of interest. The results followed a