Growing evidence suggests attrition from the physician–scientist pipeline.1,2 This trend appears particularly acute among female faculty; previous studies demonstrate that even promising female physician–researchers are less likely to advance in academic medical careers than their male peers.3–7 Prior work suggests an association between inadequate mentoring, faculty discontent, and the ultimate abandonment of academic medicine by men and women alike.8,9 Therefore, carefully characterizing what constitutes good mentoring in academic medicine—specifically, which aspects are associated with faculty satisfaction and whether there are differences by gender—is critically important.
The cohort of individuals receiving K08 and K23 career development awards from the National Institutes of Health (NIH)10,11 is an ideal population among which to explore issues of mentoring, gender, and career satisfaction. Because K08 and K23 award recipients are highly promising clinician–investigators, they are provided with resources (including salary support for protected time) so that they can prepare to ultimately apply for subsequent grant support from the NIH as independent researchers. Of note, the K08 and K23 award programs specifically encourage careful development of a mentoring plan as part of the rigorous application process.10,11 Surveying men and women about the mentoring they received during their K awards in order to explore differences in their experiences is particularly important because women may be less likely than men both to continue on in an independent research career and to transition from the K award to subsequent grants.3,4,7
Given growing interest in understanding the relationship between mentoring and career satisfaction in academic medicine,12–16 we conducted a nationwide survey study of current recipients of K08 and K23 awards, all of whom were required to develop relationships with designated mentors. First, we sought to examine the development and nature of these mentoring relationships, mentor characteristics, and extent of mentoring in different mentoring roles. Next, we aimed to evaluate whether there were systematic differences in those aspects of the mentoring relationship that related to the gender of the K award recipient, the gender of the mentor, or in the gender concordance between mentor and mentee or protégé. Finally, we sought to assess the associations between various aspects of the mentoring relationship and career satisfaction in this promising population of junior academic medical faculty.
Using the NIH RePORTER database,17 we identified 1,719 people who received new K08 and K23 awards in 2006 through 2009. After gaining ethical approval from the University of Michigan institutional review board (IRB), we conducted Internet searches and made telephone calls to obtain background information and current U.S. mailing addresses for these recipients. We obtained 1,708 valid U.S. mailing addresses.
Between August 2010 and February 2011, we mailed a survey questionnaire along with a $50 incentive to 1,708 individuals who received K awards in the years 2006 to 2009. We included a cover letter stating that this was an IRB-approved research study designed to assess the experiences of researchers who received K08 and K23 awards from the NIH. The cover letter explained the voluntary nature of participation, the risks involved and efforts to ensure confidentiality, and the source(s) of funding; it also provided contact information for the principal investigator (R.J.) and IRB. We conducted our mail survey following a modified Dillman approach,18 including targeted follow-up to nonrespondents, in order to reduce sources of error related to sampling and nonresponse.
On receipt of the completed questionnaires, we merged survey responses to data previously collected from RePORTER on K award type, award year, and characteristics of the recipient’s institution.
We designed the questionnaire after reviewing the relevant literature and considering other instruments used to determine outcomes of academic careers.12–16,19–29 We also engaged in detailed cognitive pretesting30 to identify problems with the survey questions that could result in response error (e.g., complicated instructions, vague wording, inappropriate assumptions). We conducted cognitive interviews with a small number of faculty members similar to those in our target population, using a think-aloud approach and verbal probing techniques. We then modified the survey questions on the basis of our findings.
The ultimate questionnaire included 173 items that assessed demographics (20 items), job and education information (6 items), time allocation (43 items), mentoring experiences (33 items), family responsibilities (20 items), and career satisfaction and work environment (51 items).
Development of mentoring relationships. Two questions addressed the ease with which respondents identified role models and developed mentoring relationships: (1) “How easy has it been for you to identify someone whose career could serve as a model for your own?” and (2) “How easy has it been for you to develop a relationship with a mentor?” We dichotomized the four response categories (“very easy” or “easy” versus “difficult” or “very difficult”) for descriptive analysis by gender.
Nature of mentoring relationships. The questionnaire asked respondents whether they had more than one designated K award mentor. It also asked how often they communicated in person, via telephone, or via e-mail with their primary K award mentor (dichotomized for analysis as at least once weekly versus less often) and how many hours in a typical month they spent meeting one-on-one with their primary K award mentor and other mentors (analyzed as a continuous variable). Another item asked, “Is your current relationship with your primary K award mentor closer to a student–teacher relationship or to a collegial relationship?” We dichotomized the five responses (ranging from mostly student–teacher to mostly collegial) for analysis (at least somewhat more collegial versus neutral or more student–teacher).
Mentor characteristics. Respondents reported the gender and race of their primary K award mentor. For analysis, we grouped race as white, Asian, or underrepresented minority in medicine (black, Latino, Native American).
Using five response categories, respondents rated the extent to which their primary K award mentor was fault-finding or judgmental, committed to mentoring, exploitative, patient, overprotective, controlling, available and accessible, manipulative, an important contributor to research in the field, and well connected to others of importance in the field. The negative mentoring characteristics were scored from 0 (not at all) to −4 (a lot), and the positive mentoring characteristics were scored from 0 (not at all) to 4 (a lot). We dichotomized the responses to each of the items in this battery (some, quite a bit, or a lot versus a little or not at all) for descriptive analyses by gender. In addition, we constructed two scales (behavior and prestige; see also Results) from these items for use in modeling the relationship between mentoring and career satisfaction.
Extent to which mentees perceived their mentors to have performed potential roles. We measured the extent to which K award recipients received certain types of mentoring using the question stem “Thinking about all of your mentors, how much have your mentors…,” followed by a table listing the following behaviors: served as role models, promoted your career through networking, advised about preparation for advancement (e.g., promotion, leadership positions), advised about getting your work published, advised about department/division politics, advised about obtaining the resources you need, advocated for you, advised about balancing work and family, taught you knowledge and skills, and modeled professional and ethical behavior. We dichotomized the four-point response scale for descriptive analyses by gender (0 “not at all” or 1 “a little bit” versus 2 “quite a bit” or 3 “a lot”). In addition, we constructed a scale from these items to measure the relationship between career satisfaction and the extent of mentorship in various mentoring roles (see also Results).
Satisfaction. Survey respondents indicated their level of satisfaction with both their primary K award mentor and with the mentoring they had received from all sources. We measured satisfaction with mentoring with five-point responses (0 = very dissatisfied to 4 = very satisfied) and dichotomized these for descriptive analysis by gender (very or somewhat dissatisfied versus neutral, somewhat satisfied, or very satisfied).
Survey recipients also indicated their level of satisfaction in various career domains, including the opportunity to collaborate with other faculty, the amount of social interaction with other members of their department/division, the level of funding for their research, their current salary, their sense of being valued for their research, their sense of contributing to developments in their discipline, the balance between their competing professional responsibilities, and the balance between their professional and personal lives. We created a scaled measure of career satisfaction from these items (see also Results).
Mentee, K award, and institution characteristics. We recorded respondents’ self-reported gender, race and ethnicity (which we then grouped as white, Asian, or underrepresented minority in medicine), degree (which we grouped as MD, MD/PhD, or non-MD), marital status (married/in a domestic partnership, single, or widowed/divorced), and parental status. We asked respondents to report their clinical specialty, which we then grouped into the following six categories, as we have described in detail elsewhere4,5,31: (1) medical specialties; (2) surgical specialties; (3) specialties caring for women, children, and families; (4) hospital-based specialties; (5) non-MD clinical specialties (which include pharmacists [PharmDs], veterinarians [DVMs], and many others); and (6) basic sciences. We also asked if the respondent’s research was laboratory-based and whether he or she spoke English as a native language. We relied on data merged from the RePORTER database17 for the respondent’s K award grant type, year of K award, and institutional characteristics. We grouped institutions so that all hospitals affiliated with a single university were considered to be a single institution. We then grouped institutions into four tiers containing roughly equal numbers of K awardees, based on the amount of total NIH funding received (hereafter NIH funding tier).
We restricted the current analysis to those respondents who reported receiving support from their K awards during the year prior to the survey in order to eliminate the few who may have changed careers, shifted to nonmentored grant support, or ceased to receive K award funding prior to having completed a typical term of mentored career development. We performed all analyses using the SAS System version 9.3 (SAS Institute Inc., Cary, North Carolina). We first conducted descriptive analyses and comparisons by gender using either the chi-square or Fisher exact tests for categorical variables, and the t test for continuous variables. For all comparisons by gender, we reported the P values after adjusting for the following variables: K award type, year of K award, institutional NIH funding tier, recipient race, degree, marital status, and parental status, whether the recipient’s research was laboratory based, and whether he or she spoke English as a native language.
We then sought to evaluate the associations between various aspects of mentoring and career satisfaction by developing a linear regression model of the satisfaction scale using the following covariates of interest: mentor gender, mentee–mentor gender concordance, multiple designated K award mentors, collegiality of the mentor–mentee relationship, frequency of communication between mentor and mentee, time spent meeting with mentor, the mentor behavior scale, the mentor prestige scale, and the extent of mentoring in various roles scale. The possible associations of these variables with the satisfaction scale were adjusted for the recipient characteristics of gender, race, degree, marital status, parental status, and specialty; for whether the recipient’s work was laboratory based; for whether the recipient spoke English as a native language; and for the K award type, year of K award, and institutional NIH funding tier.
We then constructed a best multiple variable model to explain the satisfaction scale. First, we modeled all covariates (adjustment covariates and covariates of interest) simultaneously. Second, we reduced the model removing insignificant adjustment covariates iteratively, until only significant covariates remained. Finally, we explored for statistically significant interactions between the respondent’s gender and the remaining covariates in the model.
For statistical inference, we considered test statistics with P values at or below 5% to be significant.
We received 1,275 completed ques tionnaires from the 1,708 individuals we contacted for a response rate of 75%. Our respondent sample did not differ significantly from the entire surveyed population of K award recipients by gender or K award year. However, a significantly higher proportion of non-MDs (81%, n = 220/272) responded to our survey than either MDs (74%, n = 754/1,024) or MD/PhDs (72%, n = 301/420; P = .02). A higher proportion of K23 recipients (78%, n = 645/831) responded than did K08 recipients (71%, n = 630/888; P = .002). Individuals at institutions with lower overall NIH funding were more likely to respond (80% [n = 322/401] from the lowest tier; 74% [n = 349/474] from the third tier; 73% [n = 353/486] from the second tier; and 69% [n = 236/340] from the top tier; P = .001). Of the 1,275 respondents, 1,227 (96%) reported actively receiving funding support from their K awards at the time of the survey and constituted the analytic sample.
The characteristics of the 559 female and 668 male K award recipients in the sample are detailed in Table 1. The majority of respondents held an MD degree, and the sample (n = 1,227) comprised the following specialties: 43% (n = 524) medical; 20% (n = 244) women, children, and families; 12% (n = 147) hospital based; 6% (n = 70) surgical; 3% (n = 32) basic science; and 17% (n = 210) non-MD clinicians. Of note, male respondents were more likely to hold MD/PhD degrees (28% of men [n = 190/668] versus 15% of women [n = 82/559]; P < .001). Women were more likely to hold non-MD clinical doctorates (25% of women [n = 138/559] versus 11% of men [n = 72/668]; P < .001). Men were more likely to have received K08 awards (60% of men [n = 399/668] versus 36% of women [n = 200/559]; P < .001). Men in this cohort were more likely than women to be married (92% of men [n = 616/667] versus 87% of women [n = 486/558]; P = .002) and to have children (82% of men [n = 545/667] versus 76% of women [n = 427/559]; P = .02). The majority of both men and women reported having a male primary designated K award mentor, although women were more likely than men to have a female mentor (29% of women [n = 159/550] versus 16% of men [n = 108/665]; P < .001).
Development of mentoring relationships
When asked how easy it was to develop a relationship with a mentor, relatively few respondents indicated difficulty, although women were slightly more likely to do so (21% of women [n = 117/552] versus 18% of men [n = 119/664]; P < .01). However, when asked how easy it was to identify someone whose career could serve as a model for their own, a substantial proportion of the respondents indicated difficulty, and women again were more likely than men to report difficulty (55% of women [n = 305/553] versus 40% of men [n = 269/667]; P < .0001).
Nature of mentoring relationships
Many respondents reported having more than one designated K award mentor, and women were more likely to do so (79% of women [n = 440/558] versus 61% of men [n = 410/668]; P < .001). Most respondents reported having a collegial relationship with their primary K award mentor. Overall, 68% of men (n = 455/668) and 62% of women (n = 348/559) reported a mostly collegial relationship with their mentors (P = .33). When asked how frequently they communicated with their primary K award mentor in person, via telephone, or via e-mail, the majority did so at least once a week (72% of men [n = 481/668] and 75% of women [n = 419/559]; P = .56). The mean number of hours that respondents spent in a typical month meeting one-on-one with a primary K award mentor was 3.2 for men and 3.0 for women (P = .35).
Primary K award mentor characteristics
Respondents generally reported that their primary K award mentors had positive characteristics, although a nontrivial minority did describe their mentors to be at least somewhat fault-finding (20%, n = 242/1,217), exploitative (17%, n = 200/1,212), overprotective (15%, n = 182/1,219), controlling (19%, n = 227/1,219), or manipulative (11%. n = 139/1,219). The vast majority indicated that their mentors were not only committed (89%, n = 1,086/1,221), patient (87%, n = 1,066/1,220), and available/accessible (88%, n = 1,078/1,221) but also important contributors to research (84%, n = 1,026/1,219) and well connected to others of importance in their field (89%, n = 1,085/1,220). Table 2 summarizes the perceived characteristics of primary K award mentors by gender of protégé, gender of primary K award mentor, and mentor–protégé gender concordance. Few gender differences were observed; however, protégés perceived female mentors to be more controlling (23% of female mentors [n = 62/267] versus 17% of male mentors [n = 164/942]; P = .014), whereas they perceived male mentors to be more available and accessible (89% of male mentors [n = 842/944] versus 85% [n = 226/267]; P = .035).
When all mentor characteristics (the items in Table 2) were considered together as a scale, the Cronbach alpha value was 0.73. However, when we iteratively removed items and recalculated the alpha value, a summary scale that excluded the final two items (i.e., an important contributor to research and well connected to others in the field) appeared more internally consistent. Therefore, two scales—following two latent constructs—resulted from the set of mentor characteristics that we examined. The first construct was mentor behavior and was measured by the first eight items listed in Table 2 (e.g., fault-finding/judgmental, patient, exploitative, available/accessible); the scale including these eight items had a Cronbach alpha of 0.78. This scale comprised three positive and five negative mentoring characteristics that were, as mentioned previously, scored from 0 (not at all) to 4 (a lot) for positive mentoring characteristics and 0 (not at all) to −4 (a lot) for negative mentoring characteristics; thus, in theory, the scale could vary between −20 and 12. In our study, the empirical distribution was skewed toward positive values. The second construct was mentor prestige and was measured by the last two items listed in Table 2; the scale for these two items had a Cronbach alpha of 0.87. Both of the component question scores ranged from 0 (not at all) to 4 (a lot), and the composite scale varied from 0 to 8 both in theory and in its empirical distribution.
Extent to which mentees perceived mentors to have played various roles
Respondents described a number of roles played by their mentors. The vast majority reported that their mentors had served as role models (76%, n = 933/1,225) or had modeled professional and ethical behavior (82%, n = 1,007/1,225). Most also reported that their mentors had taught them knowledge and skills (77%, n = 945/1,223), advised them about publication (75%, n = 912/1,224), or advocated for them (73%, n = 891/1,224). Respondents frequently reported that their mentors had promoted their careers through networking (62%, n = 757/1,225) or had advised them in areas such as obtaining resources (62%, n = 757/1,225), department politics (48%, n = 594/1,225), or preparation for advancement (58%, n = 710/1,223). Fewer respondents reported that their mentors had advised them about balancing work and family (22%, n = 273/1,224). Table 3 summarizes the perceived roles of primary K award mentors and the amount of mentoring given in each role by gender of protégé, gender of primary K award mentor, and mentor–protégé gender concordance. Those with female mentors were more likely to report receiving advice on work–life balance than those with male mentors (28% of protégés with female mentors [n = 74/266] versus 20% of protégés with male mentors [n = 192/946]; P = .008). Women with male mentors were most likely to report being taught skills and knowledge by mentors (83%, n = 322/389), followed by women with female mentors (77%, n = 121/158), men with male mentors (75%, n = 415/556), and men with female mentors (70%, n = 76/108). Overall, male K award recipients were slightly more likely to report that their mentors had served as role models (78% of male recipients [n = 517/667] versus 75% of female recipients [n = 416/558]; P = .02). Moreover, gender concordance was associated with the extent to which the mentor was perceived to serve as a role model (79% for female–female pairs [n = 125/158]; 78% for male–male pairs [n = 434/556]; 75% for male mentees with female mentors [n = 81/108]; and 73% for female mentees with male mentors [n = 285/391]; P = .05).
When considered as part of a scaled measure, all of these items seemed to measure the same latent construct—extent of mentorship in various mentoring roles—as reflected in excellent internal consistency (Cronbach alpha 0.91). When we removed individual items iteratively and recalculated the alpha value, the value varied between 0.89 and 0.90, further indicating that all items belong in the same scale. The items were scored 0 (not at all) to 3 (a lot), with the combined scale varying between 0 and 30 in theory. The empirical distribution did vary as expected and was approximately normally distributed.
Overall, few respondents were dissatisfied with the mentoring received from their primary K award mentor (10%, n = 122/1,220) or with the mentoring received from all other sources (10%, n = 128/1,221). We detected no significant difference in satisfaction with primary K award mentor by gender of mentor (P = .78), gender of mentee (P = .54), or gender concordance (P = .90). Women were slightly more likely to be dissatisfied with the mentoring they received from all sources (12% of women [n = 66/555] versus 9% of men [n = 62/666]; P = .04).
Figure 1 details the level of dissatisfaction expressed by respondents regarding a number of career domains. Dissatisfaction was most notable regarding the balance between professional and personal lives, about which nearly half of respondents expressed concern. Women were more likely than men to express their dissatisfaction with work–life balance (52% of women [n = 289/553] versus 40% of men [n = 268/663]; P < .001), the balance between their professional responsibilities (44% of women [n = 240/551] versus 32% of men [n = 211/659]; P < .001), their sense of being valued (31% of women [n = 169/554] versus 26% of men [n = 174/661]; P < .03), and the opportunity to collaborate with other faculty (14% of women [n = 77/555] versus 9% of men [n = 57/664]; P < .008).
When all eight career satisfaction variables (those that appear in Figure 1) were considered together as a scaled measure of career satisfaction, internal consistency was good (Cronbach alpha 0.79). If we removed one variable iteratively from the scale, the Cronbach alpha varied from 0.75 to 0.78, indicating that each of the questions appear to be part of the same latent construct and can be scaled together. In theory, the scale would be distributed between −16 and 16, following a normal distribution. The empirical distribution varied between −12 and 16 and approximated a normal distribution, thus confirming its suitability as an outcome variable for estimation using linear models.
In multivariable models, we found that several aspects of mentoring were significantly associated with satisfaction after adjusting for personal and institutional characteristics of the K awardee, as depicted in Table 4. These included the nature of the mentoring relationship (e.g., the collegiality of the mentoring relationship, time spent meeting with or communicating with the mentor), mentor characteristics (e.g., mentor behavior, mentor prestige), and the extent of mentoring in different mentoring roles. Mentor gender and gender concordance of the mentoring pair were not significant, nor was the number of K award mentors.
In the multivariable model presented in Table 5, in which backwards selection eliminated characteristics (as described in Method, above), the mentoring aspects that remained significantly associated with career satisfaction were the collegiality of the mentor–mentee relationship, mentor behavior, and extent of mentorship in various mentoring roles. Also independently associated with career satisfaction were mentee gender, specialty, and whether the mentee’s research was laboratory based. We evaluated for potential significant interactions between gender and other variables in the model, including the mentoring variables, and found none.
Discussion and Conclusions
In this study of junior faculty holding NIH K series career development awards, we found that a number of aspects of mentoring were associated with overall career satisfaction—including the nature of the mentoring relationship (e.g., collegiality), positive mentor behaviors, and the extent of mentoring in various mentoring roles. We found generally high satisfaction and positive mentoring experiences in this population with relatively few gender differences. However, our findings did indicate low levels of satisfaction and inadequate mentoring within the work–life balance domain, along with significant gender differences in this area.
Existing evidence supports the idea that mentoring relationships can affect career development and success in academic medicine.15 Previous research has shown that mentors influence important career outcomes, such as personal development and research productivity.23 They can assist their protégés in various areas of career development by serving as advocates; by providing opportunities for networking; by aiding in the development of important skills such as grant writing, data analysis, and manuscript preparation; and by offering personal advice and moral support.24–26 In addition, studies have identified a number of characteristics and behaviors that are desirable for mentors, including compatibility with their protégé, patience, generosity, and trustworthiness.21,24–26 Our study builds on this work by demon strating that multiple aspects of the mentoring relationship are important correlates of career satisfaction. Because mentor behaviors—particularly the level of collegiality perceived by the mentee—do appear to be directly associated with mentees’ career satisfaction, we believe that promotion of formal training pro grams that promote positive mentor behaviors may serve as a useful means to promote satisfaction and, in turn, retention in academic medical careers.
Prior research has suggested that women in academic medicine have unique needs and expectations that are not fully being met by current conceptualizations of mentoring32,33; therefore, investigators studying mentoring and satisfaction must give specific attention to the influ ence of gender. In this study, female respondents were less satisfied than men overall, and a substantial number of female respondents were specifically dissatisfied with the balance between their professional and personal lives. Of note, our respondents perceived female mentors to be more likely to give advice on work–life balance, especially when paired with a female protégé; still, mentoring in this domain was generally infrequent, even when both members of the mentoring dyad were female. Moreover, female mentors were perceived as being less available, and female protégés had greater difficulty in identifying role models. These findings are troubling because previous research suggests that women’s attrition from academic medicine may be partly due both to a dearth of positive role models who have successfully navigated the challenges of career and family and to a lack of sufficient advice in this important area.8 Our findings are especially compelling in light of social cognitive career theory,34 which focuses on how social cognitive mechanisms relevant to career development (e.g., self-efficacy beliefs, outcome expectations, goals related to balancing work and family) interrelate with other personal factors (e.g., gender), contextual factors (e.g., exposure to role models; cultural and gender role socialization), and learning/experiential factors (e.g., observing one’s own success or failure; observing the success or failure of role models) to either promote or hinder a person’s ability to progress in his or her career. Of note, not only women but also a substantial proportion of men in our study expressed dissatisfaction with work–life balance and received little mentoring in this domain. Thus, our findings add to the evidence suggesting a growing need for mentors and role models in academic medicine who are available and willing to share with their protégés—both male and female—their own personal experiences with managing the competing demands of work and home.8,16
Notably, our respondents perceived female mentors as less available and more controlling compared with male mentors. Such findings are worth considering in the context of prior research on unconscious gender bias.35–39 It is possible that similar behaviors in men and women are judged differently because of stereotypical cultural expectations35–37 and that gender bias may negatively affect the way successful female leaders are perceived.38,39 It could be the case that female mentors behaved similarly to male mentors but were judged more harshly (e.g., were rated as more controlling) because their actions were perceived as masculine and gender incongruent (e.g., women are not expected to be assertive). Additional research is necessary to explore this issue further. Educating both protégés and mentors—both male and female—about stereotypes and gender bias may be a crucial step toward “fixing the environment” rather than “fixing the women” in such circumstances.40,41
This study has a number of strengths, including both a large number of responses and a high rate of response from a population with established mentoring relationships, among whom the importance of different aspects of mentoring could be explored. Nevertheless, this study does have certain limitations. First, it relies on self-report for most of its measures. Although the questions we used have high face validity and were developed with standard techniques of survey design (including cognitive pretesting30), recall bias and/or other biases may have influenced participant responses. Second, like all survey studies, selection bias is possible. Although we achieved an extremely high response rate for a medical faculty survey (75%), and differences between respondents and nonrespondents were not marked, those who responded may have differed systematically from the overall target population. Nevertheless, it is unlikely that such differences would underlie either the associations between specific aspects of the mentoring relationship and satisfaction that we observed or the gender differences that we observed. Finally, the K award population may differ systematically from the broader population of faculty in academic medicine, limiting the ability to generalize, particularly regarding the overall level of satisfaction observed, beyond the group of individuals holding career development awards.
In sum, this study found strong associations between several aspects of mentoring and career satisfaction of K award recipients, indicating that those who have concerns about dissatisfaction with and attrition from academic medical careers would be well advised to target the training and development of mentors. Positive mentoring behaviors are common but not universal. Because these behaviors are not necessarily innate, they should be taught, like any important skill in medicine. In particular, because the greatest area of dissatisfaction we observed was in the domain of work–life balance, and because this was one of the only domains in which gender differences existed, specific attention towards developing mentoring skills in this area may be particularly useful in improving overall satisfaction and retention in academic medical careers for all, and for women in particular.
Acknowledgments: The authors wish to thank the K award recipients who took the time to participate in this study.
1. The Vanishing Physician–Scientist?. 2009 Ithaca, NY Cornell University Press
2. Zemlo TR, Garrison HH, Partridge NC, Ley TJ. The physician–scientist: Career issues and challenges at the year 2000. FASEB J. 2000;14:221–230
3. Ley TJ, Hamilton BH. The gender gap in NIH grant applications. Science. 2008;322:1472–1474
4. Jagsi R, Motomura AR, Griffith KA, Rangarajan S, Ubel PA. Sex differences in attainment of independent funding by career development awardees. Ann Intern Med. 2009;151:804–811
5. Jagsi R, DeCastro R, Griffith KA, et al. Similarities and differences in the career trajectories of male and female career development award recipients. Acad Med. 2011;86:1415–1421
6. Pohlhaus JR, Jiang H, Wagner RM, Schaffer WT, Pinn VW. Sex differences in application, success, and funding rates for NIH extramural programs. Acad Med. 2011;86:759–767
7. Martinez ED, Botos J, Dohoney KM, et al. Falling off the academic bandwagon. Women are more likely to quit at the postdoc to principal investigator transition. EMBO Rep. 2007;8:977–981
8. Levine RB, Lin F, Kern DE, Wright SM, Carrese J. Stories from early-career women physicians who have left academic medicine: A qualitative study at a single institution. Acad Med. 2011;86:752–758
9. Lowenstein SR, Fernandez G, Crane LA. Medical school faculty discontent: Prevalence and predictors of intent to leave academic careers. BMC Med Educ. 2007;7:37
12. Pearlman SA, Leef KH, Sciscione AC. Factors that affect satisfaction with neonatal–perinatal fellowship training. Am J Perinatol. 2004;21:371–375
13. Sciscione AC, Colmorgen GH, D’Alton ME. Factors affecting fellowship satisfaction, thesis completion, and career direction among maternal–fetal medicine fellows. Obstet Gynecol. 1998;91:1023–1026
14. Palepu A, Friedman RH, Barnett RC, et al. Medical faculty with mentors are more satisfied. J Gen Intern Med. 1996;11(4 suppl):107
15. 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
16. Levinson W, Kaufman K, Clark B, Tolle SW. Mentors and role models for women in academic medicine. West J Med. 1991;154:423–426
18. Dillman DA, Smyth JD, Christian LM Internet, Mail, and Mixed-Mode Surveys: The Tailored Design Method. 20093rd ed Hoboken, NJ John Wiley & Sons, Inc
19. Ramanan RA, Taylor WC, Davis RB, Phillips RS. Mentoring matters: Mentoring and career preparation in internal medicine residency training. J Gen Intern Med. 2006;21:340–345
20. Ramanan RA, Phillips RS, Davis RB, Silen W, Reede JY. Mentoring in medicine: Keys to satisfaction. Am J Med. 2002;112:336–341
21. 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
22. Pololi LH, Knight SM, Dennis K, Frankel RM. Helping medical school faculty realize their dreams: An innovative, collaborative mentoring program. Acad Med. 2002;77:377–384
23. Sambunjak D, Straus SE, Marusić A. Mentoring in academic medicine: A systematic review. JAMA. 2006;296:1103–1115
24. 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. 2009;25:72–78
25. Straus SE, Chatur F, Taylor M. Issues in the mentor–mentee relationship in academic medicine: A qualitative study. Acad Med. 2009;84:135–139
26. Cho CS, Ramanan RA, Feldman MD. Defining the ideal qualities of mentorship: A qualitative analysis of the characteristics of outstanding mentors. Am J Med. 2011;124:453–458
27. Kaplan SH, Sullivan LM, Dukes KA, Phillips CF, Kelch RP, Schaller JG. Sex differences in academic advancement. Results of a national study of pediatricians. N Engl J Med. 1996;335:1282–1289
30. Willis GB Cognitive Interviewing: A Tool for Improving Questionnaire Design. 2005 Thousand Oaks, Calif Sage Publications, Inc
31. 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
32. Robinson JD, Cannon DL. Mentoring in the academic medical setting: The gender gap. J Clin Psychol Med Settings. 2005;12:265–270
33. 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
34. Lent RW, Brown SD, Hackett G. Toward a unifying social cognitive theory of career and academic interest, choice, and performance. J Vocat Behav. 1994;45:79–122
35. Biernat M, Manis M. Shifting standards and stereotype-based judgments. J Pers Soc Psychol. 1994;66:5–20
36. Biernat M, Manis M, Nelson TE. Stereotypes and standards of judgment. J Pers Soc Psychol. 1991;60:485–499
37. Kobrynowicz D, Biemat M. Decoding subjective evaluations: How stereotypes provide shifting standards. J Exp Soc Psychol. 1997;33:579–601
38. Heilman ME. Description and prescription: How gender stereotypes prevent women’s ascent up the organizational ladder. J Soc Issues. 2001;57:657–674
39. Heilman ME, Block CJ, Martell RF. Sex stereotypes: Do they influence perceptions of managers? J Soc Behav Pers. 1995;10:237–252