Gender gaps plague careers in academic medicine as seen in differential rates in advancement,1–5 compensation,3,6 and productivity7,8 between women and men faculty. The work environment can be particularly complex for women faculty who struggle to balance professional and personal life demands9–11 and who are members of disenfranchised groups, such as racial–ethnic minorities12; foreign-born individuals; lesbian, gay, bisexual, and transgender (LGBT)13,14 people; and/or persons with disabilities.15
Mentoring is one promising strategy to address gender gaps in academic medicine, in part because women faculty report that insufficient mentoring is the most common challenge that impedes advancement.16,17 Mentoring is an intense developmental relationship18 that involves the reciprocal exchange of resources between a mentor and a mentee.19 The dyadic form of mentoring (i.e., a relationship between two individuals) remains common in academic medicine,20 and having more than one dyadic mentoring relationship is typical and encouraged.21–23 Effective mentoring includes mentoring with positive career and individual outcomes, such as promotions, greater salary growth, and career satisfaction.19,24,25
To promote effective mentoring for women faculty, a focus on mentor–mentee fit is needed. The goodness-of-fit model of mentoring suggests that the quality and value of mentoring are predicated on the degree of fit between mentor and mentee preferences, endowments, and the content of knowledge transmitted.19,26 Although prior studies have investigated desired mentor characteristics pertaining to the mentor’s personality, interpersonal abilities, and professional status,22,25–30 a lack of research on additional desirable mentor characteristics from the perspective of women faculty in academic medicine limits our ability to advance optimal mentoring through better mentor–mentee matching.
The topic of fit has been studied mostly at the stage of mentoring initiation. Mentorships are typically initiated on the basis of perceived similarity, identification, and interpersonal comfort between the mentor and the mentee.25,31–33 Underlying these psychological and interpersonal factors are the dynamics of similarity attraction (i.e., the tendency to be attracted to, like, and/or benefit from interactions with others we perceive as similar).34
Studies show that similarity can be based on sociodemographic factors, attitudes, interests, and beliefs or experience-based factors such as departmental affiliation and organizational setting.35–37 Although these studies on experience-based similarity were conducted outside of academic medicine, their results suggest that we should consider the importance of similarity in academic medical settings where contextual differences in demographics, promotion patterns, and expectations for productivity exist among departments.38 A mentor’s awareness of the departmental environment may influence his or her understanding or perception of mentee needs and the type of support offered.
This study addresses two research questions:
- Which mentor similarity characteristics do women faculty in academic medicine report as most important?
- Does importance of similarity differ among women faculty on the basis of current and prior mentoring, demographic and personal factors (race/ethnicity, foreign-born status, age, child care responsibilities), and career factors (rank, degree type)?
We investigate similarity in terms of medical institution (same medical school/academic medical center), department, personal and career interests, and race/ethnicity and gender. The importance of similarity is not assessed relative to a specific outcome but, rather, represents women faculty preferences in general.
Data and sampling
We used quantitative survey data from the Women and Inclusion in Academic Medicine study (WIAM).39 The aim of the WIAM study was to examine the characteristics and interrelationships of institutional, individual, and sociocultural factors that influence the entry, progression, and persistence of women faculty in academic medicine. The WIAM study was conducted by Converge: Building Inclusion in the Sciences, the research and evaluation arm of the Harvard Medical School Office for Diversity Inclusion and Community Partnership. Data were collected in 2012 using a purposive sample of 13 academic medical institutions based on geographic location, public versus private status, faculty size, Research Center in Minority Institution status, and research intensity (Table 1). All women faculty in clinical and basic science departments who had a valid e-mail address (N = 8,041) were invited to participate in the study via an e-mail that linked to the online survey. No material incentives were offered to encourage study participation. Across the 13 medical schools, 3,127 women faculty (39%) responded to the survey. Of the respondents, 3,100 met the inclusion criterion of being faculty at the rank of instructor or higher. The respondents closely represent the women faculty at the 13 medical schools in terms of rank and race/ethnicity breakdown. The Harvard Medical School Committee on Human Subjects approved the study.
Respondents were provided the following definition of a mentor: “an individual who holds a position senior to yours who takes an active interest in developing your career.” Using a five-point response scale ranging from “not at all important” (1) to “very important” (5), respondents were asked to indicate how important it is that a mentor (1) be at the same medical school/academic medical center, (2) be in the same department, (3) be of the same gender, (4) be of the same race/ethnicity, (5) have the same career interests, and (6) have the same personal interests as you. Responses to each of these items were used as single-item measures.
We coded current and prior mentoring using two survey questions: (1) Across your education, training, and employment, have you ever had a mentor? and (2) Do you currently have a mentor? Codes used included (1) have a mentor currently (reference), (2) had a mentor in the past but not currently, or (3) never had a mentor. We used data about multiple mentors and past mentoring to further describe those who reported currently having a mentor.
We measured race/ethnicity using self-identified race and ethnicity, coded as non-Hispanic white (reference), non-Hispanic black, non-Hispanic Asian, Hispanic, multiple races, and other/decline to answer. We measured U.S. born (reference) and foreign born as a dichotomous variable. On the basis of birth year, we grouped respondents into three age groups: (1) ≤ 44 years (reference); (2) 45 to 55 years; and (3) > 55 years. We asked whether respondents had child care responsibilities and again grouped respondents into three groups: (1) yes, within the past two years (reference); (2) yes, more than two years ago; and (3) no. Exploratory models included care for dependent adult(s), LGBT status, marital/partnership status, and disability status; however, these variables were not significant in bivariate or multivariable analyses and were omitted after assessing model fit.
Academic rank was coded as (1) full professor, (2) associate professor, (3) assis tant professor (reference), and (4) instructor. Degree type was categorized as (1) medical degree only (reference) (MD, MBBS, DO, etc.); (2) doctoral degree only (PhD, ScD, etc.); (3) medical and doctoral degree; and (4) master’s or bachelor’s.
The WIAM data were imputed to address missing data for items assumed to be missing at random. We used multiple imputation procedures to address missing data for items missing less than 30% and generated five completed datasets for analysis. To calculate point estimates and standard errors, we used Rubin’s40 rules to pool the results from the five imputed waves.
We assessed observation nonindependence within medical schools by calculating the intraclass correlation coefficient (ICC). The ICC for the six models ranged between 2% and 3%, indicating a marginal covariance by medical school. We used a set of dummy variables to control for the differences between medical schools to allow the simultaneous use of imputed data and ordinal logistic models.
The sample is described using summary statistics from the unimputed data. Numbers and percentages are presented for categorical values, and medians with interquartile ranges (IQRs) are presented for continuous variables. We tested the distribution of independent variables by academic rank using Pearson chi-square. Overlap among the categorical independent variables was assessed using Pearson chi-square, Spearman correlation, and variance inflation factor (VIF). Although the independent variables were related, the VIF was low and coefficients did not change if one or more variables were omitted from the models. We estimated the multivariable models with imputed data and used ordinal logistic regression. We estimated cross-product interactions with post hoc tests to investigate the differences among faculty by race/ethnicity and foreign-born status. All P values were two sided, and a minimum significance level of .05 was used. We used STATA version 13.1 (StataCorp, College Station, Texas) for the analysis.
Description of respondents
Table 2 shows characteristics of the respondents using the unimputed data. Study participants were predominantly white (n = 1,484; 68%); age ≤ 44 years (n = 1,045; 48%); assistant professors (n = 1,245; 41%) or instructors (n = 698; 23%); married or had a partner (n = 1,752; 79%); and currently care for dependent children (n = 1,220; 55%). One-quarter were foreign born (n = 558; 25%), with the highest percentages among Asian faculty (189/322; 59%) and Hispanic faculty (59/148; 40%). Over half of all participants currently have a mentor (n = 1,170; 53%), about one-third (n = 736; 34%) had a mentor in the past but not currently, and 13% (n = 290) never had a mentor.
Most independent variables were significantly associated with faculty rank (see Table 2). For example, of 465 participants with the rank of professor, 277 (77%) were white, 28 (8%) were Hispanic, 23 (6%) were Asian, and 10 (3%) were black. The majority of professors (n = 190; 53%) currently do not have a mentor but had a mentor in the past. Table 3 shows that the majority of all faculty in our study currently have a mentor. Among the 1,154 who have a mentor currently, 805 (69%) indicated that they currently have multiple mentors, and 56 (5%) reported no prior mentoring.
Importance of mentor similarity characteristics for women faculty
As shown in Table 4, on average, using the 1–5 rating scale, participants rated same medical institution (median = 4, IQR = 2) and same department (median = 4, IQR = 2) most important, and same race/ethnicity least important (median = 1, IQR = 1) among mentor similarity characteristics. Table 5 displays preferences for mentor characteristics among different women faculty participants.
Current and prior mentoring.
Compared with those with a current mentor, women faculty who had a past mentor but no current mentor had 22% lower odds of rating same institution (odds ratio [OR] = 0.78, 95% confidence interval [CI] = 0.63–0.98) or same department (OR = 0.78, CI = 0.65–0.92) very important and 31% higher odds of rating same gender very important (OR = 1.31, CI = 1.05–1.64). Compared with those with a current mentor, women faculty who never had a mentor were more likely to rate same gender (OR = 1.62, CI = 1.19–2.20) and same race/ethnicity (OR = 1.75, CI = 1.15–2.66) very important.
Demographic and personal characteristics.
Race and place of birth (U.S. vs. foreign born) were also associated with preferences for mentor characteristics. Compared with whites, blacks had 32% lower odds of rating same medical school very important (OR = 0.68, CI = 0.48–0.95), 50% higher odds of rating same gender very important (OR = 1.5, CI = 1.03–2.19), and over five times greater odds of rating same race/ethnicity (OR = 5.44, CI = 3.42–8.65) very important. Compared with whites, Hispanics (OR = 2.53, CI = 1.73–3.72) and Asians (OR = 1.37, CI = 1.07–1.76) were more likely to rate same race/ethnicity very important. The preference for same-race mentors was generally stronger among U.S.-born than foreign-born faculty. For example, compared with U.S.-born whites, the odds of rating same race/ethnicity as very important was six times greater for U.S.-born blacks (OR = 6.06, CI = 3.63–10.11) but only 3.5 times greater for foreign-born blacks (OR = 4.04, CI = 2.09–7.80). Other differences in preferences were also observed among faculty based on where faculty were born. For example, compared with U.S.-born whites, foreign-born Asians had 47% (OR = 1.47, CI = 1.41–1.90) greater odds of rating same department as very important.
Differences by age group were also observed. For example, compared with faculty whose age was ≤ 44 years, faculty older than 55 years had 22% lower odds (OR = 0.78, CI = 0.61–1.01) of rating same personal interest as important and 38% lower odds of rating same career interest as important (OR = 0.62, CI = 0.47–0.81).
Faculty of lower rank generally had greater preferences for mentors with the same career characteristics than did faculty at higher ranks. Compared with assistant professors, full professors (OR = 0.56, CI = 0.42–0.75) had 44% lower odds of rating same institution as important and 55% lower odds (OR = 0.45, CI = 0.33–0.62) of rating same department very important. Compared with assistant professors, the odds of rating same gender very important were 36% greater among instructors (OR = 1.36, CI = 1.09–1.68). The odds of rating same department very important were 33% lower for those with both medical and research doctorate degrees (OR = 0.67, CI = 0.47–0.95) compared with those who had a medical degree. Those with a doctorate degree alone had 19% higher odds for rating career similarity as more important (OR = 1.19, CI = 1.00–1.42) compared with those who had a medical degree alone.
Discussion and Conclusions
Finding a good fit between mentor and mentee is essential for effective mentoring. Given this fact and our limited knowledge about mentor preferences among women faculty in academic medicine, we investigated the importance that women faculty assign to mentor similarity with regard to various characteristics. While our findings underscore the importance that women faculty assign to mentor similarity, the variability in preferences based on mentee minority status, academic rank, and mentor exposure indicate that in designing mentoring programs, organizations should, from the beginning, seek direct input from individual mentees regarding their mentor preferences.
In general, same department and same institution were rated most important relative to other mentor characteristics. The importance given to same department and same institution may be attributed to contextual differences.38 We encourage greater accountability on all levels for access to “local” mentors, including from institutions (e.g., in the design of programs), departments (e.g., knowing and understanding their own unique contextual demands and challenges and mentor availability), and faculty (e.g., communicating their needs and preferences, especially through annual reviews). Such accountability measures should encourage faculty to build mentoring networks that consist of “local/internal” and “outside/external” mentors.
Our findings also document the preferences of diverse women faculty, and below we discuss these findings in terms of mentor preferences based on current and prior mentoring, as well as demographic, personal, and career characteristics. We also provide recommendations for mentor–mentee matching strategies.
Current and prior mentoring
The prevalence of mentoring is extremely variable, with some faculty having limited access to mentors, as demonstrated in this study and past studies.17 We found that faculty members’ preferences about mentors varied on the basis of prior mentor experiences. Faculty with no prior mentoring assigned greater importance to same race/ethnicity in the mentoring relationship than did faculty with current mentors. Those who never had a mentor may be making assumptions about benefits that might accrue to them in same race/ethnicity mentor–mentee relationships. Alternatively, for those who had past mentors, needs for race similarity might have been satisfied in prior mentoring relationships. This may also suggest that assumed difficulties associated with racial–ethnic incongruences may be mitigated with mentoring exposure.
Compared with faculty with a current mentor, participants without a current mentor (whether or not they had prior mentoring) viewed gender similarity as important. The issues of cross-gender mentoring may be more salient to those without a current mentor, who may perceive gender-related challenges1–11 for which they have inadequate support. Additionally, compared with participants with a current mentor, faculty with prior but no current mentors placed less importance on same department and same institution. This may be attributed to a shift in career interest, career focus, and/or needs that have changed since last having a mentor. This may be particularly relevant for senior faculty. In the matching process, prior and current mentor experiences should be ascertained to better understand and accommodate mentee preferences.
Demographic and personal characteristics
A review of mentoring programs at academic medical centers for faculty who are underrepresented in medicine (black, Hispanic, Native American) reports that matching is most often based on similarity in research interests and/or discipline.41 These findings may miss other unique needs of minority (black, Hispanic, and Asian) and foreign-born women faculty. Overall, our participants rated same race/ethnicity least important among mentor characteristics; however, it was more important for minority and foreign-born faculty than for their white and U.S.-born counterparts. This observation is consistent with relational demography theory (i.e., observable demographic factors will be more salient for those who are a numerical minority in an organization).42 Same-race matching should be strongly considered for minority and foreign-born faculty; however, availability of similar mentors within an institution or department may present difficulties for women of color. Building a diverse mentoring network that meets racial–ethnic similarity preferences may require reaching outside one’s department and institution. Additionally, unconscious bias training and cultural sensitivity training may increase all mentors’ sensitivity to cross-cultural differences and better enable mentors to meet the needs of diverse mentees.43
Compared with white U.S.-born faculty, U.S.-born black faculty in our study rated same-gender mentors more important and same-institution mentors less important. Prior research of obstetrics–gynecology residents found that, compared with other racial groups, African American women were more likely to have a female mentor. This trend was consistently observed over time.44 African American women, in particular, believed that same-gender mentors would be more understanding.44 Gender should be taken into account in mentor–mentee matching, particularly for African American women. Having mentors within the same department was important to foreign-born Asians in our study. This may reflect differences in background, training, and/or clinical versus basic science focus.
Matching based on personal and career interest may be less critical for older faculty compared with younger faculty, as suggested by our finding that faculty older than 55 years viewed these characteristics as less important. When matching mentors and mentees, one should be mindful that age does not equal rank. In our study, 29% of professors were under 55, and 20% of faculty older than 55 held the rank of either assistant professor or instructor. Needs may change depending on career- and life-stage,45 as well as personal, social, and environmental factors.46 Therefore, life- and career-stage should be considered in mentor matching. For example, younger faculty may prefer mentors whom they can emulate as they establish careers and/or families,16 while older faculty may be more focused on expanding or revising their career trajectories irrespective of rank.
Rank and degree type
In mentor–mentee matching and mentorship in general, there is a tendency to focus on junior faculty as mentees. In this study, full professors were least likely to have a current mentor compared with faculty at lower ranks. Both junior and senior faculty experience challenges23,47–49 and could therefore benefit from mentoring. As we found in our study, associate and full professors rated same-department mentors less important than did junior faculty, and full professors placed less value on mentors in the same institution than did junior faculty. This may relate to the scope of senior faculty’s work and service (e.g., greater national/international focus and/or taking on broader leadership assignments). These findings are consistent with past studies showing that the networks of senior faculty are less local or organizationally dependent,50 as senior faculty tend to source mentors outside their organization.50
We encourage organizations to consider specific efforts that target the mentoring needs of senior women faculty. This may include broadening access to external mentors and encouraging involvement in networks outside of the local institution. Senior faculty should also report their needs to department heads (e.g., during annual reviews or reappointment reviews) to bring attention to and/or greater accountability for mentoring of senior faculty. In our study, having a female mentor was considered important by instructors. Blood et al16 found that instructors have the greatest unmet mentoring needs. Although there is a limited number of female mentors in academic medicine,44,51 the need for female mentors among instructors could possibly be met because same gender was less important for faculty at higher ranks. These more senior faculty may not be as sensitive to cross-gender mentee–mentor matching. The preference of instructors for same-gender mentors should be assessed during the matching process and incorporated into discussions on mentoring as part of faculty development programs.
With regard to degree type, faculty with nonclinical degrees considered same career interest to be more important than did faculty with medical degrees. Same career interest is likely critical for faculty who are primarily engaged in research and for whom consideration of alternative career paths inside and outside of academic medicine may be a necessity. Particularly for nonclinical faculty, identification and matching require receptive mentors who share and/or can support mentees along career paths that reflect a wide array of interests.
In summary, we offer six recommendations for enhancing mentor–mentee fit:
- Institutions and/or leadership should be more accountable for access to “local mentors” for all women faculty, and for encouraging mentoring of senior women faculty.
- Prior and current mentor status of all women faculty should be ascertained prior to matching and/or during relationship initiation stages.
- The importance of same-gender and same-race/ethnicity matching should be assessed and discussed with women faculty, especially those at lower ranks or who are earlier in their careers, and with minority and foreign-born faculty.
- Mentor and mentee training programs should encourage mentor training in areas such as unconscious racial and gender bias.
- Age- and rank-related preferences should be treated as distinct.
- Identify and match mentors who share mentees’ intended career paths and/or who can support mentees along varied career paths, especially for nonclinical faculty.
Study limitations include reliance on cross-sectional and self-reported survey data, which could introduce common method bias.52 Although the participants in our sample had a roughly equal demographic composition to the entire population at their institutions, access to individual data of nonrespondents would have allowed us to better estimate any potential bias introduced by nonrespondents. The study focused solely on women in academic medicine. Given the lack of empirical studies on mentor preferences among academic medicine faculty in general, we suggest that future studies should include male faculty to allow comparison. The study highlighted what might be important mentor preferences for women faculty. More research is needed to fully understand why women faculty value certain mentor characteristics. Six respondents identified their race as American Indian or Alaskan Native, and three specified Native Hawaiian, limiting analysis by these racial groups.
A systematic review of mentoring in academic medicine found that insufficient evidence exists to provide guidance on mentor matching or selection of a mentor.30 Given the importance of mentor–mentee fit for effective mentoring,19,26 our study calls attention to the mentor preferences of women faculty in academic medicine specifically regarding the importance of mentor similarity. Our findings document the relative importance of mentors “in-place” (same department and/or institution) compared with other mentor characteristics as well as other preferences of a diverse sample of women faculty across 13 medical schools. These findings can be instructive in addressing ongoing challenges related to increasing faculty diversity in higher education. The results can inform preference-based matching in mentor program design and implementation. We acknowledge that faculty may have multiple mentors, therefore providing several opportunities to act on these recommendations. Taking preferences into consideration will improve goodness-of-fit for mentor–mentee dyads and ultimately should enhance the potential for optimized outcomes. Preference-based matching may provide an impetus for programs to be more explicit and transparent regarding the algorithms/criteria used in matching mentors with mentees. This could simultaneously inform research about mentoring preferences and the evaluation of mentor program effectiveness.
Acknowledgments: The authors gratefully acknowledge Michael Wake, MSW, MPH, for his contribution to the Women and Inclusion in Academic Medicine study, including general program management and data collection. In addition, they wish to thank Dr. Sharon-Lise Normand and Dr. Robert Wolf for their work on data imputation and statistical support, Dr. Stacy Blake-Beard for conceptual support, and Dr. Erica Warner for her analytical advice. Finally, they thank the participating site liaisons at the 13 medical schools and the study advisory committee for their support.
1. Carr PL, Gunn CM, Kaplan SA, Raj A, Freund KM. Inadequate progress for women in academic medicine: Findings from the National Faculty Study. J Womens Health (Larchmt). 2015;24:190–199.
2. Amrein K, Langmann A, Fahrleitner-Pammer A, Pieber TR, Zollner-Schwetz I. Women underrepresented on editorial boards of 60 major medical journals. Gend Med. 2011;8:378–387.
3. Ash AS, Carr PL, Goldstein R, Friedman RH. Compensation and advancement of women in academic medicine: Is there equity? Ann Intern Med. 2004;141:205–212.
4. LaPierre TA, Zimmerman MK. Career advancement and gender equity in healthcare management. Gend Mgmt Int J. 2012;27:100–118.
5. Settles IH, Cortina LM, Malley J, Stewart AJ. The climate for women in academic science: The good, the bad, and the changeable. Psychol Women Q. 2006;30:47–58.
6. Jagsi R, Griffith KA, Stewart A, Sambuco D, DeCastro R, Ubel PA. Gender differences in the salaries of physician researchers. JAMA. 2012;307:2410–2417.
7. Eloy JA, Svider PF, Cherla DV, et al. Gender disparities in research productivity among 9952 academic physicians. Laryngoscope. 2013;123:1865–1875.
8. Svider PF, D’Aguillo CM, White PE, et al. Gender differences in successful National Institutes of Health funding in ophthalmology. J Surg Educ. 2014;71:680–688.
9. Dyrbye LN, Shanafelt TD, Balch CM, Satele D, Sloan J, Freischlag J. Relationship between work–home conflicts and burnout among American surgeons: A comparison by sex. Arch Surg. 2011;146:211–217.
10. Shollen SL, Bland CJ, Finstad DA, Taylor AL. Organizational climate and family life: How these factors affect the status of women faculty at one medical school. Acad Med. 2009;84:87–94.
11. Buddeberg-Fischer B, Stamm M, Buddeberg C, et al. The impact of gender and parenthood on physicians’ careers—professional and personal situation seven years after graduation. BMC Health Serv Res. 2010;10:40.
12. Pololi LH, Evans AT, Gibbs BK, Krupat E, Brennan RT, Civian JT. The experience of minority faculty who are underrepresented in medicine, at 26 representative U.S. medical schools. Acad Med. 2013;88:1308–1314.
13. Eliason MJ, Dibble SL, Robertson PA. Lesbian, gay, bisexual, and transgender (LGBT) physicians’ experiences in the workplace. J Homosex. 2011;58:1355–1371.
14. Sánchez NF, Rankin S, Callahan E, et al. LGBT trainee and health professional perspectives on academic careers—facilitators and challenges. LGBT Health. 2015;2:346–356.
15. Steinberg AG, Iezzoni LI, Conill A, Stineman M. Reasonable accommodations for medical faculty with disabilities. JAMA. 2002;288:3147–3154.
16. Blood EA, Ullrich NJ, Hirshfeld-Becker DR, et al. Academic women faculty: Are they finding the mentoring they need? J Womens Health (Larchmt). 2012;21:1201–1208.
17. Sambunjak D, Straus SE, Marusić A. Mentoring in academic medicine: A systematic review. JAMA. 2006;296:1103–1115.
18. Kram KE. Mentoring at Work: Developmental Relationships in Organizational Life. 1988.Lanham, Md.: University Press of America.
19. Bozeman B, Feeney MK. Mentor matching: A “goodness of fit” model. Adm Soc. 2008;40:465–482.
20. Kashiwagi DT, Varkey P, Cook DA. Mentoring programs for physicians in academic medicine: A systematic review. Acad Med. 2013;88:1029–1037.
21. DeCastro R, Sambuco D, Ubel PA, Stewart A, Jagsi R. Mentor networks in academic medicine: Moving beyond a dyadic conception of mentoring for junior faculty researchers. Acad Med. 2013;88(4)488–496.
22. Mayer AP, Files JA, Ko MG, Blair JE. Academic advancement of women in medicine: Do socialized gender differences have a role in mentoring? Mayo Clin Proc. 2008;83:204–207.
23. Koopman RJ, Thiedke CC. Views of family medicine department chairs about mentoring junior faculty. Med Teach. 2005;27:734–737.
24. Ragins BR, Cotton JL, Miller JS. Marginal mentoring: The effects of type of mentor, quality of relationship, and program design on work and career attitudes. Acad Manage J. 2000;43:1177–1194.
25. Allen TD, Eby LT, Poteet ML, Lentz E, Lima L. Career benefits associated with mentoring for protégeé: A meta-analysis. J Appl Psychol. 2004;89:127–136.
26. Jackson VA, Palepu A, Szalacha L, Caswell C, Carr PL, Inui T. “Having the right chemistry”: A qualitative study of mentoring in academic medicine. Acad Med. 2003;78:328–334.
27. Hauer KE, Teherani A, Dechet A, Aagaard EM. Medical students’ perceptions of mentoring: A focus-group analysis. Med Teach. 2005;27:732–734.
28. Rabatin JS, Lipkin M, Rubin AS, Schachter A, Nathan M, Kalet A. A year of mentoring in academic medicine. J Gen Intern Med. 2004;19(5 pt 2):569–573.
29. Straus SE, Chatur F, Taylor M. Issues in the mentor–mentee relationship in academic medicine: A qualitative study. Acad Med. 2009;84:135–139.
30. Sambunjak D, Straus SE, Marusic A. A systematic review of qualitative research on the meaning and characteristics of mentoring in academic medicine. J Gen Intern Med. 2010;25:72–78.
31. Files JA, Blair JE, Mayer AP, Ko MG. Facilitated peer mentorship: A pilot program for academic advancement of female medical faculty. J Womens Health (Larchmt). 2008;17:1009–1015.
32. Okurame DE. Mentoring and preferences: A diagnostic survey for equal mentoring opportunity. Equal Oppor Int. 2008;27:519–536.
33. Ortiz-Walters R, Gilson LL. Mentoring in academia: An examination of the experiences of protégés of color. J Vocat Behav. 2005;67:459–475.
34. Byrne D. Interpersonal attraction as a function of affiliation need and attitude similarity. Hum Relat. 1961;14:283–289.
35. Eby LT, Allen TD, Hoffman BJ, et al. An interdisciplinary meta-analysis of the potential antecedents, correlates, and consequences of protégé perceptions of mentoring. Psychol Bull. 2013;139:441–476.
36. Harrison DA, Price KH, Bell MP. Beyond relational demography: Time and the effects of surface- and deep-level diversity on work group cohesion. Acad Manage J. 1998;41:96–107.
37. Zellmer-Bruhn ME, Maloney MM, Bhappu AD, Salvador R. When and how do differences matter? An exploration of perceived similarity in teams. Organ Behav Hum Decis Process. 2008;107:41–59.
38. Warner ET, Carapinha R, Weber GM, Hill EV, Reede JY. Considering context in academic medicine: Differences in demographic and professional characteristics and in research productivity and advancement metrics across seven clinical departments. Acad Med. 2015;90:1077–1083.
39. Hill EV, Wake M, Carapinha R, et al. Rationale and design of the women and inclusion in academic medicine study. Ethn Dis. 2016;26:245–254.
40. Rubin DB. Multiple imputation after 18+ years. J Am Stat Assoc. 1996;91:473–489.
41. Beech BM, Calles-Escandon J, Hairston KG, Langdon SE, Latham-Sadler BA, Bell RA. Mentoring programs for underrepresented minority faculty in academic medical centers: A systematic review of the literature. Acad Med. 2013;88:541–549.
42. Tsui AS, O’Reilly CA. Beyond simple demographic effects: The importance of relational demography in superior–subordinate dyads. Acad Manage J. 1989;32:402–423.
43. Prunuske AJ, Wilson J, Walls M, Clarke B. Experiences of mentors training underrepresented undergraduates in the research laboratory. CBE Life Sci Educ. 2013;12(3):403–409.
44. Coleman VH, Power ML, Williams S, Carpentieri A, Schulkin J. Continuing professional development: Racial and gender differences in obstetrics and gynecology residents’ perceptions of mentoring. J Contin Educ Health Prof. 2005;25:268–277.
45. Hall DT, Chandler DE. Career cycles and mentoring. In: The Handbook of Mentoring at Work: Theory, Research, and Practice. 2007:Thousand Oaks, Calif.: Sage Publications; 471–497.
46. Sambunjak D. Understanding wider environmental influences on mentoring: Towards an ecological model of mentoring in academic medicine. Acta Med Acad. 2015;44:47–57.
47. Carr PL, Ash AS, Friedman RH, et al. Faculty perceptions of gender discrimination and sexual harassment in academic medicine. Ann Intern Med. 2000;132:889–896.
48. Baldwin RG, Lunceford CJ, Vanderlinden KE. Faculty in the middle years: Illuminating an overlooked phase of academic life. Rev High Educ. 2005;29:97–118.
49. Leslie K, Lingard L, Whyte S. Junior faculty experiences with informal mentoring. Med Teach. 2005;27:693–698.
50. Peluchette JV, Jeanquart S. Professionals’ use of different mentor sources at various career stages: Implications for career success. J Soc Psychol. 2000;140:549–564.
51. Palepu A, Friedman RH, Barnett RC, et al. Junior faculty members’ mentoring relationships and their professional development in U.S. medical schools. Acad Med. 1998;73:318–323.
© 2016 by the Association of American Medical Colleges
52. Spector PE. Method variance in organizational research: Truth or urban legend? Organ Res Methods. 2006;9:221–232.