The survey was completed by 239 students, representing a response rate of 37 %. The response rates for the three medical schools were 28 % (Wright State University; WSU), 34 % (University of Kentucky; UK), and 38 % (University of Louisville; UL). Of those providing demographic information, 104 (44 %) were male; 111 were female (46 %). The mean age was 26.8 years (SD 2.3); 73 % were Caucasian. See Table 1 for demographics of the sample. Comparison of population and sample demographics found the sample mean ages for each school were representative of the entire third- and fourth-year classes (UK, sample age = 26.6, class age = 27.2; UL, sample = 27.0, class = 27.3; and WSU, sample = 26.4, class = 27.5). Regarding gender, males were less likely to respond in the UK and WSU samples (UK, sample males = 45 %, class males = 61 %; UL, sample = 44 %, class = 43 %; and WSU, sample = 44 %, class = 38 %). Race also had skewed response rates (UK, sample minority status = 18 %, class minority status = 27 %; UL, sample = 21 %, class = 20 %; and WSU, sample = 11 %, class = 26 %).
To assess the correlation of student interest in academic medicine with several practice setting-related factors we used the Goodman and Kruskal Gamma (γ). Goodman and Kruskal’s γ measures the strength of association when two variables are measured at the ordinal level with a minimal range of responses, i.e., Likert response items. Values range from −1 to +1 where −1 reflects perfect negative association and +1 reflects perfect positive association. A gamma of +1 means that a person who ranked above another person on one variable would always rank above them on the other variables as well. A value of −1 means that a person who ranked above another person on one variable would always rank below them on the other .
Student interest in academic medicine was significantly correlated with several practice setting-related factors: teaching opportunities (γ = 0.74), research opportunities (γ = 0.53), interprofessional practice opportunities (γ = 0.34), and administrative opportunities (γ = 0.27). Community service opportunities were also significantly related (γ = 0.16). Neither patient care opportunities nor job autonomy was associated with interest in academic medicine.
To assess the association, student interest in academic medicine was classified into two categories: (1) low interest (not at all, a little, and somewhat interested/important responses) and (2) high
interest (pretty and very interested/important responses). Of the students who had participated in research activities, 45 % expressed high interest in academic medicine, while only 27 % who had not engaged in research activities expressed high interests (χ
2 = 5.54, df = 1, p = 0.019). Students whose speciality choice had been influenced by an academic role model also showed an increase of high interest (48 %) in academic medicine over students without an academic role model (27 %), χ
2 = 9.62, df = 1, p = 0.002.
Lastly, we found a modest but significant positive correlation between accumulated debt and interest in academic medicine (γ = 0.16, p = 0.016). Regarding student demographics, we found no significant relationships between gender or race and interest in an academic medicine career. However, we did find a statistically significant association for older students, who reported more interest in academic medicine than younger students [older students (≥27 years), 53 %; younger students (<27 years), 37 %, p = 0.035].
The two open-ended questions that asked students to write down the three factors they would find most appealing and least appealing about an academic medical career yielded 693 responses, 380 ‘most appealing’ and 313 ‘least appealing’ responses; the coding process generated five domains: 1) professional (313 responses); 2) research (61 responses); 3) personal (140 responses); 4) teaching and mentoring (123 responses); and 5) patients/patient care (56 responses). We also identified sub-categories for each domain (Table 2). Student descriptions of appealing factors in the professional domain accounted for 46 % of the most appealing responses; appealing factors in the teaching and mentoring domain accounted for 27 % of the most appealing responses. Student descriptors of the least appealing factors of a career in academic medicine focused on the professional (45 %) and personal (33 %) domains.