Abstract
Background:
Although women comprise an increasing proportion of US medical school faculty, they are underrepresented at higher ranks. Lack of effective mentoring may contribute to this disparity. We examined the role of academic rank, research focus, parenting, and part-time work on mentoring importance, needs, and gaps.
Methods:
In 2009, women faculty members of Harvard Medical School and Harvard School of Dental Medicine were invited by e-mail to participate in a 28-item structured questionnaire. Descriptive statistics and adjusted logistic regressions were used to identify relevant themes.
Results:
Of the 1179 women faculty who responded, 54% had a mentor, and 72% without a mentor desired mentoring. The most important mentor characteristic identified was availability. Respondents endorsed most mentoring areas as important (range 51%–99%); 52% of respondents identified mentoring gaps (area important and unmet) in developing and achieving career goals and negotiation skills. Interest in mentorship for skills needed for advancement (research and lecturing skills and getting national recognition) was significantly associated with lower rank. Assistant professors were most likely to identify mentoring related to writing and publishing articles, whereas associate professors identified program development/strategic planning as important. Faculty who are parents identified gaps in finding collaborators and balancing work and family life.
Conclusions:
This survey identified a desire for both comprehensive and targeted mentoring to address gaps that varied by faculty rank, research focus, parenting, and work time status. Strategies to enhance mentoring should address career stages and include a structured framework for assessing mentoring gaps.
Introduction
Women faculty comprise an increasing proportion of U.S. medical school faculty but continue to be underrepresented at the highest academic ranks, with women making up only 19% of the professorial rank. 1 Women scientists and medical faculty have been noted to experience a number of barriers to career development and promotion, including lack of effective mentoring, lack of skills that could be affected by mentoring, and issues related to childbearing and childcare. 2 –8 Some studies have suggested that female physicians are less likely than their male colleagues to have a mentor. 9 While investigators have been challenged to define the quality and role of mentoring in successful careers, studies have found positive impacts of mentoring on career development and research productivity. 10 –13 Faculty and professional organizations have voiced strong support for enhanced mentoring. 6,14
Over the past two decades, many medical schools, including ours, Harvard Medical School (HMS), have initiated offices of faculty development, career development workshops, and mentoring programs. Despite robust programs, faculty members have continued to express concern about access to mentoring in general, 2,8 although the specific unmet needs have been largely undefined. In a 2003 unpublished survey of all faculty at our institution, 51% of women faculty and 42% of men faculty reported having at least one mentor, and of those without a mentor, 65% and 37%, respectively, reported that they wanted a mentor (M. Connelly, personal communication), suggesting gender-specific needs for effective mentoring. For this reason, we undertook a structured survey of women faculty in order to understand better the specific characteristics and components of mentoring desired, especially as they related to having children, working part-time vs. full-time, research focus, and academic rank.
Materials and Methods
A 28-item questionnaire regarding experience with mentoring and mentoring needs was developed by members of the Joint Committee on the Status of Women from the HMS and Harvard School of Dental Medicine. The initial objective of the survey was to provide data about mentoring gaps and mentoring needs in order to create a matched mentoring program for women faculty. As there was not an existing survey that would address these questions specifically, we developed a new survey instrument with questions designed for faculty at all ranks. Before implementation of the survey, it was piloted by members of the mentoring subcommittee, which included 16 faculty members at all levels of rank. Survey development included input from faculty with expertise in survey design and methodology, who participated in the adaptation and iterative testing of the instrument. After approval by the HMS Committee on Human Studies, an invitation to participate in the study was sent three times via e-mail between June and October 2009 to 3516 female faculty members of HMS (98%) and School of Dental Medicine (2%) with active e-mail addresses. Participants gave informed consent by voluntary completion of the questionnaire through an anonymous link to an online survey tool (SurveyMonkey.com, LLC, Palo Alto, CA).
Demographic information requested included age, race/ethnicity, marital status, number of children, degree, academic rank, career stage, and percent of time spent on clinical, research, teaching, and administrative work. Participants were queried about their current mentoring situation, their current mentoring needs, traits and characteristics they felt were important in a mentor, and obstacles they had experienced in reaching their career goals. The mentoring areas included in the survey were developing and achieving career goals, clinical skills, research skills, teaching skills, administrative skills, finding collaborators, writing and publishing articles, obtaining grant funding, presentation/lecturing skills, program development/strategic planning, networking, getting national recognition, shifting careers, negotiation skills, resolving work conflicts, balancing work and family, and time management/stress management. For each area, survey participants were asked whether or not each area was met (yes, no) and important to them (yes, no). If a participant endorsed an area as important and also indicated that the need was not met for that area, we coded this area for that participant as a mentorship gap. The proportion of faculty with a mentorship gap was examined across the sample and with respect to potential influencing variables.
Respondents from all four academic ranks, including full professors, were included in the descriptive analyses and the information about obstacles to reaching career goals and important traits in a mentor. In the analyses examining factors influencing mentoring needs, importance, and gaps, we excluded the 48 participants at the full professor rank as there were relatively few respondents
Four primary characteristics were tested for association with demographic, academic, and mentoring variables, including academic rank, majority of time spent on research (research focus), parenting status and age group of youngest child (categorized as no children, children ≤5, children 6–15 years old, and children >15), and work status (full-time vs. part-time). Because the instructor rank is the first appointment of most faculty (91.5%) at HMS and can be considered the prefirst promotion rank similar to the assistant professor rank at many other academic centers, we further categorized instructors as being new to rank (<5 years) or longer in rank (≥5 years). Chi-square tests were used to compare categorical demographic or academic variables between groups, and t tests or analyses of variance (ANOVAs) were used to compare continuous demographic and academic variables between groups. Bivariate chi-square tests were performed to test the association between each primary predictor and each of the mentoring needs met, importance of mentoring areas, and mentorship gaps. For associations found to be significant at the bivariate level, we performed logistic regressions using each mentoring area as the outcome and rank, parenting/child age group, research focus, work status, or time in rank among instructors as the primary predictor. We then fit adjusted models that included other potential predictors or covariates to the logistic regression to determine whether or not the bivariate relationship remained after adjustment for other variables. Variables included in the multivariable models were current mentoring (yes, no), self-identified career stage (entry, mid, senior), work status, sharing of household tasks, number of active mentors, and age. No correction was made for multiple comparisons.
Results
Participant characteristics
Of the surveys sent, 1179 responses (33.5%) were received. The respondents were similar to the HMS women faculty by race/ethnicity. Survey respondents were 79% white, 3% black, 2% Hispanic, and 15% Asian, and HMS women faculty were 78% white, 3% black, 3% Hispanic, and 16% Asian. The percentage of respondents by rank were 5% professor, 13% associate professor, 28% assistant professor, and 53% instructors (55% of instructors with data were <5 years), which is comparable to the entire HMS women faculty, comprising 4% professor, 10% associate professor, 22% assistant professor, and 64% instructors (63% of instructors were <5 years). Respondents to the survey had a median age of 44 years (25th, 75th percentile: 38, 52) (Table 1). Most participants had children and were married or partnered; 17% worked part-time. Although just over half of female faculty (54%) indicated they currently had a mentor, with a mean of 1.2 mentors per faculty member; 72% of those without a current mentor indicated a desire for current mentoring.
N/A, not available.
Obstacles to career advancement and mentor characteristics desired
The most common obstacles identified by women faculty in reaching career advancement goals were insufficient mentorship (39%), insufficient financial support for research (35%), clinical workload (34%), and insufficient departmental or institutional support (33%). Faculty at the full professor rank were more likely to cite obstacles of insufficient administrative support (71% among full professor vs. 50% among all other ranks, p=0.004), insufficient departmental or institutional support (69% vs. 39%, p<0.0001), and gender (44% vs. 22%, p=0.0006) than those at lower rank. Instructors were more likely to cite clinical load (47% vs. 33% at higher rank, p<0.0001). The most important characteristics of a mentor identified were availability (71%), program development and strategic planning experience (54%), clinical experience (41%), and teaching experience (41%). In addition, women from underrepresented minorities (URMs) were more likely to indicate gender (p=0.03) and race (p<0.001) as important characteristics of a mentor. Specifically, 24% of black and Latina women believed the gender of a mentor was very important compared to 14% among Asians, 9% among white, and 11% among other race/ethnicities. Likewise, 9% of black and Latina women thought the race of a mentor was very important compared to 4% of Asians, 1% of white, and none of other race/ethnicities.
Mentoring areas: Importance, needs met, and mentorship gaps
Among instructors, assistant professors, and associate professors, the majority of participants endorsed the importance of all the mentoring areas (range 51%–99%), with the exception of shifting careers (33%). However, less than half the participants indicated that most of these areas were met (Fig. 1). The highest proportion of unmet needs was seen in areas of program development/strategic planning, shifting careers, and negotiation skills. Mentorship gaps (important and unmet) were identified among a number of areas; developing and achieving career goals and negotiation skills were each endorsed by 52% of participants.

Results from a survey of women faculty: percentage of faculty with mentoring needs met, percentage of faculty indicating importance of mentoring areas, and percentage of faculty identifying mentoring gaps.
Academic rank
The demographic and academic characteristics of instructors differed from those at higher ranks (Table 1). Instructors tended to be younger, less likely to be married or partnered, and were more often Asian, Latina or black, or other nonwhite race/ethnicity than assistant or associate professors. The prevalence of part-time work also differed by rank, with instructors, then assistant professors, more likely to work part-time compared with associate professors. Assistant and associate professors indicated several mentoring areas were met more often than did instructors. Having mentoring needs met on research skills (p=0.007), presentation and lecturing skills (p=0.05), writing and publishing articles (p=0.03), obtaining grants (p=0.01), and getting national recognition (p=0.001) remained significantly associated with rank even in adjusted models. In all cases, needs were met least often for instructors and increasingly met for assistant professors and associate professors, respectively. The rate of importance among certain mentoring areas also differed by rank. Writing and publishing articles remained significant in the adjusted model (p=0.01), with assistant professors most often identifying this area as important. Associate professors indicated program development/strategic planning as important more often than assistant professors and instructors, respectively (p=0.05). Finally, the distribution of mentorship gaps appeared to differ by rank for several mentoring areas. Even after adjustment, mentoring on publishing articles remained significantly associated with rank (p=0.003), with assistant professors identifying gaps most often compared to instructors and associate professors.
Among instructors, those in rank for ≥5 years indicated several mentoring areas met at significantly different levels from those newer to rank. Newer instructor faculty were more likely to have needs met in developing and achieving career goals, research skills, writing and publishing articles, obtaining grants, presentation and lecturing skills, networking, and resolving work conflicts compared to those in rank longer, although none was significant in adjusted models. Those newer to rank identified negotiation skills as important more often even when including additional covariates and predictors in the model (p=0.02). Those in rank longer identified gaps in developing and achieving career goals more often, whereas those newer to rank identified getting national recognition more often, although neither remained significant in adjusted models.
Faculty with a research focus
Because there is a continuum of time devoted to research activities among academic faculty, we identified those who spent >50% of their time on research. These faculty were more likely to have mentors (74% vs. 52%, p<0.001) and more likely to have ranks of assistant and associate professor (37% vs. 28% for assistant professor and 19% vs. 13% for associate professor, p<0.001) compared to those who spent less time on research. They were less likely to work part-time (5% vs. 19%, p<0.001) and more likely to have children aged ≤5 years (38% vs. 27%, p=0.01). Whereas the proportion of white faculty members was similar by this grouping of time spent on research (78% in both groups), a greater number of researchers were Asian (58, 18%, vs. 57, 14%) and a smaller number were Latina or black (7, 2% vs. 27, 7%) (p=0.009)).
In adjusted models, research faculty had mentoring needs met more often in research skills (p=0.003), obtaining grants (p<0.001), presentation and lecturing skills (p=0.02), and balancing work and family (p=0.02). Similarly, in adjusted models, researchers considered research skills (p=0.006), finding collaborators (p=0.03), and obtaining grant funding (p<0.001) important more often. Those with a research focus indicated clinical skills (p<0.001) and teaching skills (p=0.006) as important less often. Of note, those who spent <50% of their time in research more commonly identified gaps related to mentoring on research skills (p=0.04), presentation and lecturing skills (p=0.01), and balancing work and family life (p=0.03) in adjusted models. We also examined the data by degree (M.D.s, Ph.D.s, and M.D./Ph.D.s) as a proxy for likely research trajectory (although Ph.D.s may be clinical psychologists); the only mentoring gap that remained significant in adjusted models was mentoring for administrative skills, which were indicated as a gap more often by M.D./Ph.D.s.
Part-time and full-time work status
Faculty working part-time were older (median age 48 years compared to 42 years among full-time, p<0.0001), had been in rank longer (8.7 years vs. 5.5 years among full-time, p<0.001), were significantly less likely to indicate current mentorship (31% vs. 62%, p<0.001), and had fewer mentors (mean number of mentors 0.8 vs. 1.3 among full-time, p<0.001). The part-time faculty who did not have a mentor were also less likely to indicate a desire for mentorship (72% vs. 82%, p=0.02).
Parenting
When parenting status groups were compared, faculty with children ≤5 years were the most likely to report having current mentoring compared to those with children aged 6–15, those with children >15, and those with no children (p<0.0001) Among those with no mentor, desire for mentoring also differed between groups (p<0.001). Those with children <15 and no children reported similarly high rates of desire for mentoring (85% among those without children, 87% among those with children ≤5, 81% among those with children 6–15), whereas those with older children (>15) were less likely to wish for mentoring (63%). Twenty-eight percent of faculty with older children were working part-time, and part-time work was less common among those with children ≤15 (17% for those with children ≤5 and 18% among those with children 6–15) and no children (12%). Rank also differed significantly, with those without children and those with children ≤5 being more likely to be at the instructor rank (p<0.001). Faculty with children were older than those without children (p<0.001). No difference was seen by race and ethnicity.
Only mentoring for writing and publishing articles (p=0.04) remained significantly associated with parenting and child age group after other predictors and covariates were included in a multivariable model; those with no children or children aged ≤5 had the lowest rates of needs met. In adjusted models, only the importance of balancing work and family life (p=0.02) differed significantly by parenting status, with parents of children ≤5 indicating importance most frequently. Mentorship gaps differed significantly by parenting/child age groups in bivariate comparisons, but only finding collaborators (p=0.04) and balancing work and family life (p=0.02) remained significant in adjusted analyses. Faculty members without children and those with children ≤5 were most likely to identify gaps in finding collaborators, whereas those with children 6–15 years old were least likely. Faculty without children and with children <15 were most likely to identify gaps in balancing work and family life.
Discussion
This survey of women faculty of a large medical school affiliated with an academic medical center identified several themes of mentoring and faculty development to be addressed on an institutional and individual basis. Particular themes were identified commonly by all faculty, and others were more specific to groups or cohorts of women faculty members. One of the most notable findings was that there were mentoring gaps and a perceived low rate of effective mentoring across all faculty surveyed. Although 54% of women faculty reported having a mentor, a significant number of faculty reported that their mentoring needs were unmet. Moreover, the mean total number of mentors per faculty members was small (approximately 1 per faculty), even when current research has emphasized the advantage of multiple mentors and developmental networks. 15,16 Faculty surveyed identified lack of mentorship as an obstacle to career advancement. Prior studies have demonstrated that academic faculty frequently cite mentoring as critical to their success. 10 –13 In a systematic review of mentoring of trainees and junior faculty in academic medicine, Sambunjak et al. 11 noted the limitations in prior studies but also the positive influence of mentoring on personal development, self-confidence, career choice, and research productivity. Other investigators have found that mentoring increases the sense of institutional and departmental support for faculty careers. 13,17
Prior work on defining barriers to advancement of women in academic medicine has indicated that differences in the availability of effective mentoring is a possible contributor to the unequal representation of women at higher academic ranks. 5 Career success is impacted by lack of skills that could be affected by mentoring, such as work-life balance, decision making, national recognition, and negotiation. 3,4,7 Having a mentor and availability of that mentor are perceived as important for career success. Therefore, academic medical institutions should begin by addressing the apparent lack of availability of mentorship and lack of designated time for mentoring. This could begin by providing institutional support for mentorship training programs for all faculty in the practices of effective mentoring, collaborative mentoring, developmental networks, and communication strategies. This could take place in the form of formal sessions as well as simulation exercises and self-learning modules, with refresher sessions and ongoing assessments for both mentors and mentees. In addition, it could involve specific time allocation to faculty for mentoring purposes in the form of their own training as well as in the provision of active mentorship. As mentorship is typically an expectation above and beyond faculty responsibilities, this would better incorporate mentoring into current job expectations and would increase the likelihood it would take place and occur successfully, with enhanced academic productivity. Including mentoring in the criteria for academic promotion as well as institutional recognition of mentoring as an element of citizenship and annual evaluation could also reinforce the importance of mentoring. Recognition of mentoring is likely to increase the number of senior faculty members willing to spend time mentoring more junior faculty. Lastly, appreciation for mentorship could be made more apparent by establishment of an award program to honor outstanding faculty mentors.
There were several mentoring areas in which mentorship gaps (indicated by a faculty member rating an area as both important and unmet) were identified at high rates by all faculty, such as identifying career goals and negotiation skills. Identification of these gaps suggests some specific areas for future interventions. Mentors may assume that junior faculty have already established career goals, and this need may not be addressed as a key step in the mentoring process. It is logical, however, that if this need is not first addressed and readdressed periodically, progression through career stages and academic advancement would likely be slow, undirected, and unsatisfying for both mentor and mentee. As respondents indicated that many areas of mentoring were important, it will be critical for the mentor and mentee to establish priorities and high impact domains of particular salience for the individual mentee by using these areas as a checklist. Negotiation skills may not be immediately identified as necessary for advancement in academic medicine. 18 Mentorship on negotiation skills could be addressed both through formal workshops for mentees and in individual meetings that take into account the unique circumstances of each faculty member's career trajectory, department, or institution. Mentors coaching mentees through difficult negotiations and participating in case simulations can help reinforce the acquisition of needed negotiation skills for junior faculty. In addition, training mentors on how best to educate mentees in negotiation techniques may be helpful as well.
This survey identified mentoring importance, needs, and gaps that differed by certain faculty characteristics. Although addressing general mentoring needs for faculty is important, mentors should also be knowledgeable about mentoring needs and gaps that are associated with faculty rank, clinical and research trajectory, part-time or full-time work, and parenting status. Making mentors aware of gaps specific to these characteristics should allow the needs to be better addressed. Additionally, training or workshops directed at faculty with one of the characteristics associated with particular gaps (e.g., instructors, primary researchers, part-time faculty members) could be focused on the specific gaps identified by this group. Characteristics identified in this survey that were associated with mentorship needs and gaps included faculty rank, proportion of time spent on research, parenting, and part-time work.
Faculty rank was associated with different views of mentoring needs and areas of importance. Assistant professors identified writing and publication of articles and associate professors identified program development/strategic planning as key areas for mentoring. Instructors had their mentoring needs met at lower rates for all mentoring areas than did assistant and associate professors. It is, therefore, important to address mentoring needs early on and then adapt mentoring style and focus as faculty members advance. The skills needed for career advancement (e.g., research skills, presentation and lecturing skills) were recognized as particularly salient for early career faculty, while later mentoring in other skills becomes important but may remain unmet (e.g., negotiation, time and stress management, and networking). Faculty at rank of instructor for ≥5 years had significantly lower rates of current mentoring and higher rates of desiring a mentor and assigned different levels of importance to several mentoring areas than those at rank for a shorter interval. Thus, it is important to identify faculty members who have been at the entry level rank for longer to determine if the career trajectory or goals are different from those of other faculty or if lack of mentoring is playing a role in this trajectory.
Faculty who self-identified as being more research focused and spending more time on research were more likely to have mentoring needs met. However, those who were spending less time in research remained interested in developing research skills and finding collaborators. This suggests that faculty engaged in research for a smaller proportion of their time may require a different mentoring focus than those participating full-time in research. Even for faculty with a research trajectory, a previous study at our medical school of grant applications found that women faculty compared to men faculty requested fewer dollars and fewer years of funding and applied to foundations rather than the National Institutes of Health (NIH), with the difference particularly related to academic rank. 19 Thus, institutions and mentors may need to provide additional advice on budgets to ensure adequate direct and indirect funding for projects as well as the investigator's career advancement. Instructors were more likely to be Asian, Hispanic or black, or other nonwhite race/ethnicity than faculty at higher ranks, emphasizing that mentoring programs should address the needs of racially/ethnically diverse faculty. Individuals of these URMs noted that the race/ethnicity or gender of mentors was particularly important.
The independent effects of parenting vs. academic rank and age are difficult to tease out in our sample, as these three characteristics were so closely related. Not surprisingly, those with children identified higher importance of and mentorship gaps in balancing work and family life. Interestingly, it was not those with young children who were most likely to work part-time. Those who worked part-time were more likely to have older children than those working full-time and also tended to be older. Working part-time, however, was associated with a greater proportion of mentorship gaps in some areas, specifically mentoring in research skills and finding collaborators, as well as greater importance assigned to clinical and teaching skills. Working part-time has previously been found to be associated with receiving less mentoring and research support. 20 Mentors can help these faculty define achievable goals and career trajectories. Consistent with prior studies, mentoring programs should address varied needs among faculty with respect to academic track selection, 21 creating flexibility in schedule and career structure 9,22,23 supporting faculty at different levels in their careers, and incorporating new models of mentoring, such as multiple mentors and developmental networks from the business world. 15,16,24
There are several limitations to this study. One limitation was the response rate, although the total number of participants was large. The relatively low response rate could cause bias in interpreting the results of the survey, as the participants self-selected to complete the survey may be different from those who did not complete the survey. For example, if only the happiest faculty responded, we have underestimated mentoring gaps, whereas if only the most unhappy faculty responded, we have overestimated mentoring gaps. Because the survey appears to consist of a representative cross-section of HMS women faculty (based on data from the population of HMS women faculty), this is less likely the case. Additionally, as only women faculty were surveyed, gender differences could not be evaluated. Our results may not be generalizable to other medical schools with different promotion or mentoring programs, although there likely are comparable issues related to rank (e.g., assistant professor rank is often the entry rank at other institutions) and other characteristics we studied. The small proportion of dental school faculty members in this sample precluded a comparison of mentoring needs between dental and medical school faculty. We also did not have data to explore the specific meaning of mentor for individual faculty. Finally, we lacked sufficient power to further analyze race and ethnicity differences across the various groups.
Conclusions
This survey found that just over half of women faculty had a current mentor and that most desired mentoring and endorsed general and specific aspects of mentoring. They identified important mentorship gaps that needed to be addressed. Thus, identifying both the importance of mentoring and unmet mentoring needs of individual faculty members through checklists and discussions and understanding the impact of academic rank, duration at rank, research focus, parenting status, and work time on these issues has the potential to facilitate enhanced mentoring efforts. Additionally, recognizing that these issues are likely to evolve as a faculty member progresses through her career is critical. Addressing these needs and gaps by taking into account developmental and situational stages in careers and training mentors to ask specifically about these areas has the potential to create more successful frameworks for structuring mentoring initiatives.
Footnotes
Acknowledgments
We thank the Mentoring Subcommittee of the Harvard Medical School and Harvard School of Dental Medicine Joint Committee of the Status of Women (JCSW) for their many hours in identifying unmet needs in mentoring, designing this survey, obtaining IRB approval, and giving feedback on the implications of this study. We also thank Chris Warfield, B.S., M.S., for her role in survey design and the Office for Faculty Affairs for their support of JCSW projects and help in disseminating the survey. For a copy of the survey, please contact Carol Warfield (cwarfiel@bidmc.harvard.edu).
Disclosure Statement
No competing financial interests exist.
