Introduction
Employment in academic veterinary medicine can be expected to include various combinations of clinical practice, research, other scholarly work, teaching, academic administration, and community outreach. There is a looming crisis, however, in academic veterinary medicine, with recent reports projecting labor shortages and numerous studies focusing on the challenges of recruiting veterinary faculty.1–5 For veterinarians who have completed a clinical residency training program, the decision to pursue an academic career versus entering private practice is a major inflection point. Given the importance of this career decision for the individuals involved as well as the veterinary profession and society as a whole, there is surprisingly little literature on what factors influence individuals in residency training programs to decide for or against a career in academic veterinary medicine. Understanding those factors may help leaders of academic programs prepare individuals for academic careers and enhance recruitment and retention efforts.
Incentives for pursuing a career in academic human medicine include the opportunity to conduct research, the variety of career opportunities, the chance to work in an intellectually stimulating environment, and the possibility of broadly influencing health through discovery and dissemination of new information.6 These factors have been examined for veterinarians in only a limited fashion previously. The challenges associated with recruitment and retention in veterinary clinical academic medicine have been described,1,5,7 but these studies have, for the most part, focused on individuals who completed specialty training and left academic medicine; however, 1 study2 did focus on veterinarians who chose to not pursue an academic career and their perceptions of why they made that decision. Stated concerns for veterinarians who did not choose an academic career included a lack of interest in research and concerns about compensation, excessive bureaucracy, the tenure process, and expectations for excessive committee work and service.2 How their perceptions were formed, how those perceptions interacted, and what the relative importance of those perceptions were to the individuals as they made their career decisions were not examined. We therefore felt that it was important to examine the perceptions and attitudes of veterinarians in clinical residency training programs to better understand factors that influence their choice of career, especially the decision to pursue a career in academic veterinary medicine versus private clinical practice.
Several models in the social sciences literature can be used to describe methods of career choice. Social cognitive career theory in particular proposes that individual personal traits and environmental factors interact to form learning experiences, which then determine the perceived confidence in one's ability to perform career-related tasks (self-efficacy) and the types of outcomes one expects from performing those tasks (outcome expectation).8–10 The objective of the study reported here was to use social cognitive career theory to identify factors that individuals in clinical residency training programs consider when making a choice for or against a career in academic clinical medicine. Secondary objectives were to identify relationships among factors influencing a resident's career choice and to develop a decision tree representing the factors influencing a resident's career choice. This investigation followed and was derived from a previous study11 that described the perceptions of individuals in clinical residency training programs related to their competency for an academic career as well as their training emphasis and mentoring during training.
Materials and Methods
Study population and survey development and distribution
The study population consisted of individuals who were in postgraduate clinical residency training programs overseen by recognized veterinary specialty organizations and members of the AVMA American Board of Veterinary Specialties. Residents were surveyed regarding their perceptions of factors related to their training program and their preparedness for an academic career with the Veterinary Academic Career Questionnaire; development and validation of this questionnaire has been described in detail previously.11 Data from the survey were then used to describe how respondents used these perceptions to make a decision about their subsequent career choice, particularly the choice to pursue a career in academic veterinary medicine versus private clinical practice. The study and questionnaire were evaluated by the Oklahoma State University Institutional Review Board, which granted an exemption for the research.
Questionnaire
The questionnaire consisted of 73 questions (Supplementary Appendix S1). Of these, 61 requested scaled responses reflecting perceptions of training, mentorship, self-efficacy, and personal and professional benefits and 12 sought information on personal characteristics (age, gender, ethnicity, and educational debt), employment, year of training, desired career following completion of residency training, interest in academic medicine, and various attitudinal factors.
Eighteen questions addressed respondents’ perceptions of the mentorship in various aspects of academic veterinary medicine that they received during their residency training. Responses consisted of a 5-point Likert-type scale, where 1 = no mentoring; 2 = occasional but inconsistent or ineffective; 3 = some mentoring but not consistently effective; 4 = frequent and effective but not optimal; and 5 = consistent, effective, and optimal. Sixteen questions addressed respondents’ perceived educational preparedness to enter academic medicine following their residency training, with responses consisting of a 5-point Likert-type scale, where 1 = unprepared; 2 = somewhat prepared but not adequately; 3 = adequately prepared for entry level; 4 = better than adequate but not fully confident; and 5 = fully confident that I would be able to perform satisfactorily. Sixteen questions asked respondents to rate the training emphasis on factors related to academic medicine on a 5-point Likert-type scale, where 1 = far too little or almost none; 2 = some but substantially too little; 3 = just about the right amount; 4 = a little too much; and 5 = way too much. Eleven questions asked respondents to rate the importance of factors influencing their choice to enter a career in academic medicine on a 5-point Likert-type scale, where 1 = not at all important; 2 = slightly important but not as much as other factors; 3 = moderately important and roughly equal to other factors; 4 = very important but not the only factor; and 5 = the sole factor in my decision. The questionnaire was administered online with commercially available software (Qualtrics Inc). Survey parameters were set such that an individual respondent's identification could not be determined and each person could take the survey only once.
Statistical analysis
Analysis of decision factors— Data were analyzed with standard software (SPSS version 26; IBM Corp). The effects of decision factors were initially analyzed from the perspective of previously described findings regarding veterinary academic career choice11 by grouping responses in the domains of mentoring, training emphasis, professional benefits and priorities, and educational preparedness. The importance of various factors that affected the individual's career choice were also evaluated; responses were categorized on the basis of desired career following residency training as academic medicine, private clinical practice, military service, industry, or other. Descriptive statistics were computed for each of the study's variables and examined for measures of central tendency; all variables were normally distributed. Associations between personal and professional characteristic and career choices were determined with c2 tests for categorical variables and Student t tests for continuous variables.
Principal component analysis with promax rotations was used to identify latent factors from the set of survey (scale) items.12,13 Promax rotation is a routinely used mathematical method whereby factor loadings are transformed to achieve a simpler structure and create a more interpretable solution for the data.14 Eigenvalues were calculated to represent each factor by dividing the sum of scaled scores for items clustering together by the total number of items representing each cluster.13 Items that displayed a minimum factor loading of 0.40 were retained for further analysis.15 The Cronbach α was calculated with standard methods as a measure of internal validity of the study questions.16 The Hotelling test was performed to determine the effect of respondents’ career choice on their perception of their residency programs. A scree plot was constructed as the eigenvalues were plotted against factor loadings to determine the number of meaningful factors to retain; effectively this method identified random error in the data.17
The importance of 11 personal and attitudinal factors influencing career choice was determined by having each respondent rate the importance of each factor related to their career decision and then rank those factors from most to least important. Responses were summarized as proportions and respondents’ rankings. Overall group ranking for each factor was determined by calculating a score for each factor and then arranging factors from the highest to lowest score. Factor scores were calculated as the sum of the values for each rank obtained by multiplying the number of responses for each rank by a weighting factor for that rank ranging from 11 (most important) to 1 (least important).
Career choice decision tree development— A classification and regression tree algorithm analysis was used to develop decision tree models incorporating factors that influenced respondents’ choice of whether to enter academic medicine or private clinical practice. This algorithm determined how independent continuous or categorical variables could best be combined to predict a binary outcome on the basis of if-then logic by proportioning each independent variable into mutually exclusive subsets on the basis of homogeneity of the data.18,19 For this analysis, the response (dependent) variable was the respondents’ choice to enter academic medicine or private clinical medicine; other career choices were not included owing to the small numbers of responses for other career choices. The classification and regression tree algorithm analysis was conducted with parent nodes defined at 20 respondents, child node defined at 5 respondents, and significance set at values of P < 0.05. The performance of the classification and regression tree model was expressed as overall accuracy of the model and accuracy for the prediction of a specific career choice (academic medicine or private clinical practice), both expressed as percentages.
There were no significant (P > 0.05) differences in age, gender, educational debt, and current year of training for early and late respondents. Thus, researchers concluded results were generalizable to the population of veterinary residents.
Results
Responses were obtained for 253 of the 1,053 surveys that were distributed, for a response rate of 24%. After removal of surveys with incomplete primary response data, 207 (20%) surveys remained in the study for analysis; not all respondents answered all questions.
Fifty-two of 206 (25%) respondents were male and 154 (75%) were female. Mean age was 34.7 years (median, 30.0 years; range, 25.0 to 62.0 years) for male respondents and 31.9 years (median, 31.0 years; range, 25.0 to 56.0 years) for female respondents; age did not differ significantly (P = 0.088) between male and female respondents. One hundred seventy of 206 (83%) respondents identified as Caucasian, and 36 (17%) identified as some other race-ethnicity, including Native American (n = 2), Pacific Islander (1), Asian (12), Hispanic-Latino (13), or multiple racial-ethnic categories (8).
Ninety-three of 194 (48%) respondents reported a desire to pursue a career in academic medicine, and 101 (52%) reported a desire to pursue a career in private clinical practice. Four respondents reported a desire to pursue a career in some field other than academic medicine or private clinical practice and were excluded from further analysis. Mean ± SD educational debt was $138,206 ± 151,698, with 60 of 186 (32%) respondents reporting not having any educational debt. Median educational debt did not differ significantly (P = 0.918) between those who preferred a career in academic medicine (median debt, $75,000) and those who preferred a career in private clinical practice (median debt, $100,000). Respondents’ perceptions of the importance of 11 factors in deciding to pursue a career in academic medicine were summarized (Figure 1), and the factors were ranked in order of the overall importance (Table 1).
Bar graphs of responses provided by veterinarians in clinical residency programs (n = 165) regarding the importance of 11 factors in influencing their choice to pursue a career in academic medicine.
Citation: Journal of the American Veterinary Medical Association 260, 5; 10.2460/javma.21.02.0075
Rankings of the relative importance of 11 factors influencing whether veterinarians in clinical residency programs (n = 165) would consider choosing to pursue a career in academic medicine following completion of their training program.
| Rank | Factor |
|---|---|
| 1 | Geography |
| 2 | Opportunity to influence the quality and impact of veterinary medicine |
| 3 | Flexibility in schedule (to include partial-week or partial-day scheduling) |
| 4 | Ability to balance work and family responsibilities |
| 5 | Compensation |
| 6 | Opportunity to influence the next generation of veterinarians |
| 7 | Opportunity to do creative or original work |
| 8 | Opportunity to travel and present information at meetings |
| 9 | Magnitude of my educational debt |
| 10 | Opportunity to engage in translational research that impacts human health |
| 11 | Opportunity to pursue academic veterinary leadership positions |
Principal component analysis of decision factors
Principal component analysis resolved 14 individual factors fitted to 4 subscales that each achieved an approximate simple structure. The Kaiser-Meyer-Olkin measure (0.88) verified sampling adequacy for the analysis. An initial analysis was performed to obtain eigenvalues for each factor. The scree plot revealed inflections that justified retaining 4 factors for 3 of the 4 subscales and 2 factors for the fourth subscale. The first subscale represented research, the second subscale represented clinical teaching, the third subscale represented classroom teaching, and the fourth subscale represented clinical practice. For all 4 subscales, the Cronbach α was > 0.70, indicating high reliability.20 Detailed analyses of survey items in the domains of educational preparedness, mentoring, and training emphasis were performed (Tables 2–4). Analysis of the 11 factors considered important in deciding to pursue a career in academic medicine yielded 2 subscales related to professional priorities and professional benefits (Table 5). For the 3 domains of mentoring, training emphasis, and educational preparedness, responses for each of the 4 subscales (research, clinical teaching, classroom teaching, and clinical practice) were compared between respondents who desired a career in academic medicine and respondents who desired a career in private clinical practice. Similarly, for the domain of importance, responses for the subscales professional priorities and professional benefits were compared between respondents who desired a career in academic medicine and respondents who desired a career in private clinical practice (Table 6).
Psychometric properties of items related to educational preparedness for veterinarians in clinical residency programs (n = 163) who responded to a questionnaire regarding factors influencing career choice following completion of their training program.
| Scale | Item | Factor loadinga | Mean (SD) score | Cronbach α |
|---|---|---|---|---|
| Research | 0.90 | |||
| Forming a research hypothesis and plan | 0.87 | 3.02 (1.20) | ||
| Writing grants | 0.85 | 2.22 (1.32) | ||
| Analyzing research data | 0.84 | 2.73 (1.16) | ||
| Designing descriptive or exploratory studies | 0.75 | 2.60 (1.20) | ||
| Presenting research findings | 0.66 | 3.27 (1.20) | ||
| Clinical teaching | 0.82 | |||
| Teaching students in small groups or “patient-side” | 0.88 | 4.23 (1.0) | ||
| Assessing student performance | 0.82 | 3.75 (1.14) | ||
| Clinical (practical) teaching | 0.71 | 3.90 (1.11) | ||
| Mentoring of advanced trainees and graduate students | 0.44 | 2.88 (1.29) | ||
| Classroom teaching | 0.87 | |||
| Preparing lectures, assignments, and examinations | 0.99 | 2.76 (1.26) | ||
| Developing course materials and presentations | 0.90 | 2.99 (1.24) | ||
| Delivering instruction to large groups in classroom settings | 0.72 | 3.26 (1.23) | ||
| Clinical practice | 0.73 | |||
| Accessing and critically reading research literature in your field | 0.77 | 3.88 (0.95) | ||
| Understanding the theory and empirical findings in your own clinical area | 0.76 | 3.70 (0.98) | ||
| Communication skills (expressing empathy, delivering feedback, reflective listening, etc) | 0.50 | 4.25 (0.91) | ||
| Clinical practice | 0.47 | 4.25 (0.96) |
Items were scored on a scale from 1 to 5, where 1 = unprepared; 2 = somewhat prepared, but not adequately; 3 = adequately prepared for entry-level; 4 = better than adequate, but not fully confident; and 5 = fully confident that I would be able to perform satisfactorily.
Loading represents the correlation between the item and overall scale.
Psychometric properties of items related to mentoring for veterinarians in clinical residency programs (n = 163) who responded to a questionnaire regarding factors influencing career choice following completion of their training program.
| Scale | Item | Factor loadinga | Mean (SD) score | Cronbach α |
|---|---|---|---|---|
| Research | Analyzing research data | 0.91 | 2.97 (1.23) | 0.88 |
| Forming a research hypothesis and plan | 0.87 | 3.04 (1.25) | ||
| Presenting research findings | 0.83 | 3.19 (1.31) | ||
| Writing grants | 0.76 | 2.23 (1.31) | ||
| Designing descriptive or exploratory studies | 0.74 | 2.73 (1.27) | ||
| Clinical practice | Communication skills (expressing empathy, delivering feedback, reflective listening, etc) | 0.84 | 3.72 (1.26) | 0.84 |
| Client relations and interactions | 0.79 | 3.79 (1.30) | ||
| Professionalism (dress, language, deportment, ethical and legal obligations of the profession, etc) | 0.79 | 3.79 (1.25) | ||
| Understanding the theory and empirical findings in your own clinical area | 0.73 | 3.74 (1.10) | ||
| Accessing and critically reading research literature in your field | 0.68 | 3.74 (1.15) | ||
| Clinical practice | 0.54 | 4.42 (0.84) | ||
| Classroom teaching | Preparing lectures, assignments, and examinations | 0.95 | 2.48 (1.35) | 0.90 |
| Developing course materials and presentations | 0.94 | 2.72 (1.31) | ||
| Delivering instruction to large groups in classroom settings | 0.78 | 2.51 (1.34) | ||
| Clinical teaching | Mentoring of advanced trainees and graduate students | 0.81 | 2.44 (1.38) | 0.80 |
| Teaching students in small groups or “patient-side” | 0.80 | 3.70 (1.24) | ||
| Assessing student performance | 0.80 | 3.09 (1.36) | ||
| Opportunities/benefits of an academic career | 0.55 | 2.69 (1.33) |
Items were scored on a scale from 1 to 5, where 1 = no mentoring; 2 = occasional, but inconsistent or ineffective; 3 = some mentoring, but not consistently effective; 4 = frequent and effective, but not optimal; and 5 = consistent, effective, and optimal.
Loading represents the correlation between the item and overall scale.
Psychometric properties of items related to training emphasis for veterinarians in clinical residency programs (n = 144) who responded to a questionnaire regarding factors influencing career choice following completion of their training program.
| Scale | Item | Factor loadinga | Mean (SD) score | Cronbach α |
|---|---|---|---|---|
| Research | Analyzing research data | 0.85 | 2.15 (0.80) | 0.84 |
| Presenting research findings | 0.84 | 2.49 (0.84) | ||
| Forming a research hypothesis and plan | 0.83 | 2.30 (0.82) | ||
| Writing grants | 0.7 | 1.81 (0.84) | ||
| Designing descriptive or exploratory studies | 0.61 | 1.91 (0.78) | ||
| Classroom teaching | Developing course materials and presentations | 0.84 | 2.12 (0.81) | 0.75 |
| Preparing lectures, assignments, and examinations | 0.78 | 1.91 (0.83) | ||
| Delivering instruction to large groups in classroom settings | 0.71 | 2.16 (0.85) | ||
| Mentoring of advanced trainees and graduate students | 0.48 | 1.96 (0.81) | ||
| Clinical practice | Accessing and critically reading research literature in your field | 0.80 | 2.84 (0.64) | 0.62 |
| Understanding the theory and empirical findings in your own clinical area | 0.74 | 2.74 (0.60) | ||
| Communication skills (expressing empathy, delivering feedback, reflective listening, etc) | 0.65 | 2.78 (0.79) | ||
| Clinical practice | 0.42 | 3.11 (0.57) | ||
| Clinical teaching | Teaching students in small groups or “patient-side” | 0.83 | 2.87 (0.68) | 0.64 |
| Clinical (practical) teaching | 0.73 | 2.75 (0.68) | ||
| Assessing student performance | 0.66 | 2.52 (0.84) |
Items were scored on a scale from 1 to 5, where 1 = far too little, almost none; 2 = some, but substantially too little; 3 = just about the right amount; 4 = a little too much; and 5 = way too much.
Loading represents the correlation between the item and overall scale.
Psychometric properties of items related to professional benefits and priorities for veterinarians in clinical residency programs (n = 163) who responded to a questionnaire regarding factors influencing career choice following completion of their training program.
| Scale | Item | Factor loading* | Mean (SD) score | Cronbach α |
|---|---|---|---|---|
| Professional benefits | Educational debt | 0.53 | 2.51 (1.28) | 0.7 |
| Geography | 0.62 | 3.18 (1.11) | ||
| Flexibility in schedule | 0.78 | 2.90 (1.13) | ||
| Compensation | 0.74 | 3.23 (1.04) | ||
| Balance work and family | 0.7 | 3.40 (1.05) | ||
| Professional priorities | Influence veterinary medicine | 0.8 | 3.37 (0.95) | 0.81 |
| Influence veterinarians | 0.7 | 3.38 (0.96) | ||
| Do creative or original work | 0.73 | 2.93 (0.99) | ||
| Pursue academic leadership | 0.71 | 2.36 (1.07) | ||
| Engage in translational research | 0.77 | 2.34 (1.15) | ||
| Travel and present at meetings | 0.5 | 2.49 (1.05) |
Items were scored on a scale from 1 to 5, where 1 = not at all important; 2 = slightly important, but not as much as other factors; 3 = moderately important and roughly equal to other factors; 4 = very important but not the only factor; and 5 = the sole factor in my decision.
Loading represents the correlation between the item and overall scale.
Comparison of scores assigned for various factors associated with career choice by veterinarians in clinical residency programs considering a career in private clinical practice versus academic medicine following completion of their training program.
| Domain | Private clinical practice | Academic medicine | P value | ||
|---|---|---|---|---|---|
| No. of respondents | Mean (SD) score | No. of respondents | Mean (SD) score | ||
| Mentoring | |||||
| Classroom teaching | 86 | 2.58 (1.22) | 75 | 2.57 (1.24) | 0.953 |
| Clinical teaching | 87 | 2.91 (1.12) | 75 | 3.08 (0.94) | 0.297 |
| Research | 87 | 2.70 (1.03) | 73 | 2.99 (1.05) | 0.795 |
| Clinical practice | 86 | 3.90 (0.95) | 75 | 3.95 (0.91) | 0.744 |
| Training emphasis | |||||
| Classroom teaching | 81 | 2.13 (0.61) | 71 | 1.93 (0.62) | 0.042 |
| Clinical teaching | 80 | 2.73 (0.60) | 73 | 2.69 (0.50) | 0.741 |
| Research | 78 | 2.09 (0.64) | 72 | 2.13 (0.62) | 0.698 |
| Clinical practice | 81 | 2.76 (0.50) | 73 | 2.80 (0.39) | 0.569 |
| Educational preparedness | |||||
| Classroom teaching | 86 | 3.05 (1.14) | 77 | 2.87 (1.12) | 0.347 |
| Clinical teaching | 86 | 3.52 (0.88) | 77 | 3.35 (0.96) | 0.260 |
| Research | 86 | 2.64 (1.01) | 77 | 2.83 (1.07) | 0.250 |
| Clinical practice | 86 | 4.07 (0.66) | 77 | 3.90 (0.88) | 0.871 |
| Importance | |||||
| Professional benefits | 87 | 3.23 (0.70) | 76 | 2.73 (0.83) | < 0.001 |
| Professional priorities | 87 | 2.58 (0.68) | 76 | 3.05 (0.77) | < 0.001 |
Decision tree analysis
The significant factors related to career choice were used to construct various decision tree models. The most accurate model had an overall classification accuracy of 82% and correctly classified a resident's career choice of private clinical practice 88% of the time and a resident's career choice of academic medicine 75% of the time. The model initially categorized respondents into 2 groups on the basis of whether the residency training program had had a positive or validating effect on their interest in a career in academic medicine or a negative effect, discouraging their interest in a career in academic medicine (Figure 2). Six independent variables were statistically significant and included in the model: respondents’ experience during residency training, the importance of professional benefits (educational debt, geography, flexibility in schedule, compensation, and ability to balance work and family), self-efficacy in research, training emphasis in classroom teaching, self-efficacy in clinical practice, and the importance of professional priorities (opportunities to influence veterinary medicine, influence future veterinarians, do creative or original work, engage in translational research, present research materials at meetings, and pursue academic leadership). A low importance of professional benefits, poor self-efficacy in research, minimal training emphasis in classroom teaching, low importance of professional priorities, and high self-efficacy in clinical practice had negative effects on desiring a career in academic medicine. A positive residency experience and high importance of professional priorities had positive effects on desiring a career in academic medicine.
Decision tree illustrating factors that influenced veterinarians in clinical residency programs (n = 189) to consider pursuing a career in academic medicine (AM) or private clinical practice (PCP) following completion of their training program. Importance of professional benefits, self-efficacy in research, training emphasis in classroom teaching, self-efficacy in clinical practice, and importance of professional priorities were scored on a scale from 1 (lowest) to 5 (highest).
Citation: Journal of the American Veterinary Medical Association 260, 5; 10.2460/javma.21.02.0075
Discussion
Individuals in clinical residency training programs have several career options following completion of their training, with the 2 primary options being academic medicine and private clinical practice. To our knowledge, the present study was the first in veterinary medicine to describe in detail the relationship of factors that influence an individual's choice to pursue a career in academic medicine or private clinical practice.
The present study was modeled on social cognitive career theory10 and derived from previous research in human and veterinary medicine that found that self-efficacy, mentoring, training emphasis, and personal and professional benefits were important in career decision-making.2,11,21 Social cognitive career theory has been used in many different disciplines to examine motivational behavior and career development, including human medicine and research training.9 Personal traits that, according to social cognitive career theory, can influence career decision-making include a person's social and educational background and experiences, age, physical ability, and sex, for example, whereas environmental factors can include the nature and type of training one receives, mentoring, observational learning opportunities, and encouragement from and interaction with others. The interaction of these factors forms the basis of an individual's decision for or against a particular career choice (in this case, academic medicine). For the present study, these interactions were examined through the technique of principal component analysis, which identified important discriminating factors that were then used to construct a decision tree describing the profile of factors that could be used to predict an individual's career choice.
The goal of principal component analysis is to reduce a large number of variables into a smaller number of uncorrelated variables that still represent most of the information in the original data.12,13 This reduces the complexity (ie, dimensionality) of the data and allows examination of relationships and relative contributions to major, generalizable themes. Analysis of data for the present study found that the examined variables could be reduced to 4 primary categories (research, clinical teaching, classroom teaching, and clinical practice) that were each composed of various factors. These categories had high (> 0.7) Cronbach α values, indicating robust reliability of the results. Exceptions were the Cronbach α values for clinical practice (α = 0.62) and clinical teaching (α = 0.64) within the domain of training emphasis. This likely reflected the fact that 2 items had low factor loading (clinical practice [0.42 factor loading] and assessing student performance [0.66 factor loading]), suggesting weaker correlations with the categories of clinical practice and clinical teaching, respectively. It was unclear whether this occurred because the questions regarding those items were poorly constructed or misunderstood by the respondent or because these items truly had little relationship to the overall category. Notwithstanding the lower factor loading for these items, correlations > 0.4 are often considered adequate,13 and the lower factor loading for these items should be considered valid representations of the question under study.
The decision tree developed in the present study represented the profile of perceptions used by respondents when deciding whether to enter academic medicine or private clinical practice following completion of their residency program. Each step in the tree indicates a decision point where 2 lines representing complimentary dichotomous options diverge. Findings of the present study suggested that a resident's perception of their training program, the importance the resident assigned to professional benefits, and the resident's perceived self-efficacy in research had the most impact in predicting whether that individual would pursue a career in academic medicine.
Previously reported findings for this same cohort of respondents found that 26% (50/193) felt that their interest in an academic career had increased during their residency, whereas 34% (65/193) reported that their interest in an academic career had decreased during their residency.11 This change in interest in an academic medical career has been highlighted in human medicine as well, with studies21–23 revealing a declining interest in academic medicine as residency training progresses. Numerous factors might influence a residents’ perceptions of their training program. On the basis of results of previous studies, the presence of a mentor, activities that improve research and teaching effectiveness and confidence, and encouragement to publish and present research findings have all been proposed as positively influencing an individual's desire for a career in academic medicine21,22,24–26 or affected their satisfaction with their training.27,28
We were not able to determine from our data which particular factors led to a training experience being perceived as good or poor by survey respondents, and this is an important knowledge gap to correct. Still, residency training programs that have a positive effect on interest in academic medicine presumably do not induce negative perceptions and value judgements about the programs. It is possible such programs attract or develop residents who place a higher value on opportunities to pursue academic leadership positions. A previous study11 reported that a substantial proportion of residents consider the setting for their training program (academic vs private practice) to be important in their decision regarding to which training programs (ie, academic or practice-focused) they would apply. On the basis of the decision tree developed in the present study, for individuals who consider other factors after opportunities to pursue academic leadership positions, compensation becomes an important factor in their decision on whether to pursue a career in academic medicine. Thus, it is important to note that even when a residency training program has a positive influence on a resident's interest in academic medicine, numerous other factors influence their decision to pursue an academic medicine career.
The role of compensation in faculty recruitment and retention has been a point of great concern and debate for several years. For most clinical disciplines, compensation in private practice is higher than that in academic medicine, and this fact, coupled with the high educational debt for many veterinarians, has led to the perhaps common perception that compensation is a major factor in the choice of an academic medicine career and an explanation for current challenges in faculty recruitment. In a study2 of board-certified veterinary clinical specialists who had specifically not chosen an academic medicine career, approximately 80% of 363 respondents indicated that they somewhat or strongly agreed with the statement that compensation in academic medicine was too low. In that study, a higher proportion of women than men (23.8% vs 14.9%) disagreed with the statement that compensation in academic medicine was too low. That study, however, did not attempt to determine how important the perception of low compensation was or what impact that perception had on the decision to not pursue a career in academic medicine. In the present study, compensation ranked fifth and magnitude of educational debt ranked ninth out of 11 factors related to career decision-making. Compensation and magnitude of educational debt were factors in the professional benefits subscale, but compensation was rated as only moderately important (mean score, 3.23 on a scale from 1 to 5) and educational debt was rated as only slightly important (mean score, 2.51). The importance of professional benefits was the second factor in the decision tree, after the impact of experiences during residency training. Notably, professional benefits also included the factors importance of balancing work and family, geographic location, and flexibility in work schedule, all of which were ranked higher by respondents than compensation in terms of importance to career decision-making. Further, professional priorities included factors such as the opportunity to influence veterinary medicine and the opportunity to influence future veterinarians, and professional priorities also had a substantial impact on career decision-making.
Research productivity during residency training (ie, number of published articles and abstracts) was found to be strongly positively correlated with entering a career in academic medicine for physicians in medical residency training programs,29–31 and authors have suggested that encouraging residents to publish may represent a way to motivate them toward a career in academic medicine.29 It is very possible, however, that the association between scholarly productivity during residency training and interest in an academic medicine career reflects a preexisting interest of the individual rather than an influence of the program. Understanding the effects of research exposure on subsequent career interests was beyond the scope of the present study, and further investigations are warranted.
An additional activity reported in 1 study29 that was strongly correlated with development of academic dermatologists was the number of presentations given at national conferences. These findings are in line with results of multiple other studies summarized by Tsoi et al,32 who stated that the bulk of the literature suggests that the earlier trainees are involved in research, the more likely they are to pursue careers in academic medicine. Publishing a manuscript is a requirement for certification by at least some specialty organizations, although the importance of and emphasis on presenting research findings at national meetings varies among programs. In the present study, we did not specifically investigate numbers of presentations or publications and therefore cannot make any specific comments about their impact on career decision-making for veterinary residents. However, our study did attempt to determine the importance of scholarly activity to respondents’ career decision-making process. The opportunity to do creative or original work was ranked by respondents as the seventh of 11 factors related to career decision-making, and the opportunity to travel and present information at meetings was ranked as the eighth of the 11 factors, while the opportunity to engage in translational research was ranked as the tenth. Notably, these personal viewpoints were not significant in our decision tree modeling, suggesting little motivation among respondents for scholarly work as a career. It was unclear, however, whether compelling additional scholarly activity in the form of manuscript preparation or presentations during a residency training program would have a positive effect on choosing a career in academic medicine. Previous research11 has suggested veterinary specialty trainees have a low self-efficacy (ie, confidence in ones’ ability to complete specific tasks) for research activities, and findings of the present study reinforce that conclusion. While it is hoped that activities that develop these skills may encourage individuals to pursue academic medicine careers, there is little evidence supporting that position. Efforts to improve student mastery of research techniques have been shown to improve research task self-efficacy25,33; however, the impact such efforts have on long-term retention in an academic career or career success remain unknown. Hence, selection of individuals with a desire for research activities may be a better strategy.
Several limitations must be considered in assessing the findings of the present study. Data collected in the present study were self-reported and therefore may have been influenced by social desirability bias. That is, respondents may have answered the survey in ways that highlighted or reflected their abilities in a manner consistent with expectations of their peers and supervisors. The effects of social desirability bias were likely mitigated by the facts that the survey was conducted online, rather than in-person, and that respondents were anonymous. Peers and supervisors were unlikely to be aware of a resident's participation, although it is still possible that bias may have been introduced in this manner. Further, it is unclear the extent to which scores may have been inflated, especially those related to self-efficacy. Overestimation of one's competence, particularly by inexperienced individuals, is consistent with the Bandura theory of self-efficacy and has been demonstrated in studies of advanced human medicine trainees.8,34 This effect was demonstrated in a previous study11 in which veterinary residents during the first year of their training self-reported an expectation of adequate competency in an academic medicine career of 78%.
The response rate of 20% for the present study was perhaps less than optimal, although there is no proven acceptable response rate that guarantees survey validity and the response rate for this survey was consistent with rates for similar surveys of physicians.35–37 Furthermore, response rate does not in itself indicate a lack of study validity.35 The greater risks to the validity of survey-based studies are biases induced by self-selection and nonresponse. Reasons for the low response rate to the present study's survey are purely speculative and cannot be determined from the data at hand. It is presumed that completing surveys is a low priority for busy residents in training. There are few methods that can be used to determine whether nonresponse bias has occurred; however, 1 method is used to compare responses of early versus late responders, with differences between these groups indicating some influence of nonresponse on the results.38 Our results did not support the presence of nonresponse bias, and we were unable to identify factors that would have biased the results. Therefore, we believe the results are valid and generalizable to the entire population. A final consideration is that the number of subjects in the terminal levels of the decision tree was small for some response items, weakening the strength of the conclusions.
Findings of the present study hold important implications for addressing the shortage of veterinarians entering academic medicine and support the notion that multiple, interacting factors contribute to a resident's choice to pursue a career in academic medicine. Among these are the perceived self-efficacy in research, amount and nature of mentoring received, nature of the training experience, and perceived personal and professional benefits of an academic medicine career. Academic medicine administrators should consider all of these factors to attract, encourage, and support residents who might have an aptitude and interest in academic medicine.
Supplementary Materials
Supplementary materials are posted online at the journal website: avmajournals.avma.org
References
- 1. ↑
Furr M. Factors associated with veterinary clinical faculty attrition. J Vet Med Educ. 2018;45(1):16–26.
- 2. ↑
Furr MO. Evaluation of factors reported by veterinarians who chose not to enter clinical academic veterinary medicine. J Am Vet Med Assoc. 2020;256(6):701–707.
- 3.
Hubbell JAE, Richardson RC, Heider LE. Workforce needs for clinical specialists at college and schools of veterinary medicine in North America. J Am Vet Med Assoc. 2006;229(10):1580–1583.
- 4.
Hubbell JAE. Veterinary teaching hospitals: current challenges and pathways for the future. J Vet Med Educ. 2008;35(1):62–65.
- 5. ↑
Adams SB, Schulz K, Hardy J, et al. Retention of surgery specialists in academia: a critical agenda. J Vet Med Educ. 2005;32(4):404–415.
- 6. ↑
Goldacre M, Stear S, Richards R, Sidebottom E. Junior doctors’ views about careers in academic medicine. Med Educ 1999;33(5):318–326.
- 7. ↑
Jelinski MD, Silver TI. The career path choices of veterinary radiologists. Vet Radiol Ultrasound. 2015;56(1):109–113.
- 8. ↑
Bandura A. Self-efficacy: toward a unifying theory of behavior change. Psychol Rev. 1977;84(2):191–215.
- 9. ↑
Bakken LL, Byars-Winston A, Wang M-F. Viewing clinical research career development through the lens of social cognitive career theory. Adv Health Sci Educ Theory Pract. 2006;11(1):91–110.
- 10. ↑
Lent RW, Brown SD, Hackett G. Toward a unifying social cognitive theory of career and academic interest, choice, and performance. J Vocat Behav. 1994;45(1):79–122.
- 11. ↑
Furr MO, Raczkoski BM. Perceptions of training, self-efficacy, and mentoring among veterinary clinical specialty trainees. J Am Vet Med Assoc. 2021;259(5):528–538.
- 12. ↑
Lever J, Krzywinski M, Altman N. Principal component analysis. Nat Methods. 2017;14(7):641–642.
- 13. ↑
Jolliffe IT, Cadima J. Principal component analysis: a review and recent developments. Phil Trans Royal Soc A. 2016;374(2065):20150202. doi: 10.1098/rsta/2015.0202
- 14. ↑
Finch H. Comparison of the performance of Varimax and Promax rotations: factor structure recovery for dichotomous items. J Educ Meas. 2006;43(1):39–52.
- 15. ↑
Costello AB, Osborne JW. Best practices in exploratory factor analysis: four recommendations for getting the most from your analysis. Pract Assess Res Eval. 2005;10(1):1–9.
- 16. ↑
Tavakol M, Dennick R. Making sense of Cronbach's alpha. Int J Med Educ. 2011;2:53–55. doi: 10.5116/ijme.4dfb.8dfd
- 17. ↑
Williams B, Onsman A, Brown T. Australian paramedic graduate attributes: a pilot study using exploratory factor analysis. Emerg Med J. 2010;27(10):794–749.
- 18. ↑
Siciliano R, Mola F. Multivariate data analysis and modeling through classification and regression trees. Comput Stat Data Anal. 2000;32(3-4):285–301.
- 19. ↑
Loh W-Y. Classification and regression trees. Data Min Knowl Discov. 2011;(1):14–23. doi: 10.1002/widm.8
- 21. ↑
Straus SE, Straus C, Tzanetos K. Career choice in academic medicine: a systematic review. J Gen Intern Med. 2006;21(12):1222–1229.
- 22. ↑
Leonard JC, Ellsbury KE. Gender and interest in academic careers by first- and third-year residents. Acad Med. 1996;71(5):502–504.
- 23. ↑
Cain JM, Schulkin J, Parisi VE, et al. Effects of perceptions and mentorship on pursuing a career in academic medicine in obstetrics and gynecology. Acad Med. 2001;76(6):628–634.
- 24. ↑
Sebastian M, Robinson MA, Dumeny L, et al. Training methods that improve MD-PhD student self-efficacy for clinical research skills. J Clin Transl Sci. 2019;3(6):316–324.
- 25. ↑
Lipira L, Jeffe DB, Krauss M, et al. Evaluation of clinical research training programs using the clinical research appraisal inventory. Clin Transl Sci. 2010;3(5):243–248.
- 26. ↑
Neacy K, Stern SA, Kim HM, Dronen SC. Resident perception of academic skills training and impact on academic career choice. Acad Emerg Med. 2000;7(12):1408–1415.
- 27. ↑
Pearlman SA, Leef KH, Sciscione AC. Factors that affect satisfaction with neonatal-perinatal fellowship training. Am J Perinatol. 2004;21(7):371–375.
- 28. ↑
Reck SJ, Stratman EJ, Vogel C, Mukesh BN. Assessment of residents’ loss of interest in academic careers and identification of correctable factors. Arch Dermatol. 2006;142(7):855–858.
- 29. ↑
Wu JJ, Ramirez CC, Alonso CA, Mendoza N, Berman B, Tyring SK. Dermatology residency program characteristics that correlate with graduates selecting an academic dermatology career. Arch Dermatol. 2006;142(7):845–850.
- 30.
Andriole DA, Jeffe DB. The road to an academic medicine career: a national cohort study of male and female US medical graduates. Acad Med. 2012;87(12):1722–1733.
- 31. ↑
Borges NJ, Navarro AM, Grover A, Hoban JD. How, when, and why do physicians choose careers in academic medicine? A literature review. Acad Med. 2010;85(4):680–686.
- 32. ↑
Tsoi M, Teitge BD, Madan CR, Francescutti L. Personal values influencing career path in academic medicine: perspectives of selected Canadian trainees. F100Research. 2018;5:1903. doi: 10.12688/f1000research.9026.2
- 33. ↑
Bierer SB, Prayson RA, Dannefer EF. Association of research self-efficacy with medical student career interests, specialization, and scholarship: a case study. Adv Health Sci Educ Theory Pract. 2015;20(2):339–354.
- 34. ↑
Biondi EA, Varande WS, Garfunkel LC, et al. Discordance between resident and faculty perceptions of resident autonomy: can self-determination theory help interpret differences and guide strategies for bridging the divide? Acad Med. 2015;90(4):462–471.
- 35. ↑
Nolte MT, Shauver MJ, Chung KC. Structure and establishing validity in survey research. Plast Reconstr Surg. 2015;135(1):216e–222e. doi: 10.1097/PRS.0000000000000794
- 36.
Sinclair M, Otoole J, Malawaraarachchi, Ledder M. Comparison of response rates and cost-effectiveness for a community-based survey: postal, internet and telephone modes with generic or personalised recruitment approaches. BMC Med Res Methodol. 2012;12(1):132. doi: 10.1186/1471–2288–12–132
- 37. ↑
Huang JY, Hubbard SM, Mulvey KP. Obtaining valid response rates: considerations beyond the tailored design method. Eval Program Plann. 2003;26(1):91–97.
- 38. ↑
Johnson TP, Wislar JS. Response rates and nonresponse errors in surveys. JAMA. 2012;307(17):1805–1806.
