Loss of faculty to nonacademic careers or to other institutions poses significant costs to institutions of higher education. Such costs include the monetary and time expenses that institutions and individuals within the institutions incur to recruit, hire, and develop new faculty, as well as the loss of valuable human capital.1,2 The human capital argument is especially applicable in the case of loss of women faculty, because these faculty offer diverse perspectives that can inform research and foster innovation.3 Without the opportunity to thoroughly tap into the human capital of women faculty, institutions are at a competitive disadvantage in the global marketplace.4
Women faculty, compared with men faculty, are more likely to leave academe,5 and, when they do stay, they are much less likely to achieve the rank of full professor. Although the percentage of full professors in doctoral institutions who were women increased from 9% in 1988 to 17% in 2004, women continue to be underrepresented in the full professor rank compared with men (see Table 1).6 This trend occurs despite the fact that the percentage of doctoral degrees awarded to women rose steadily from 13% in 1969–1970 to 49% in 2005–2006.7 This loss of women along the traditional path from graduate school to full professor has been labeled the “leaking pipeline” and has been attributed to various factors, including aspects of organizational structure (e.g., formal policies and procedures) and climate (e.g., how women are treated, implicitly and explicitly, by others).8
Studies of climate and the status of women faculty in academic health centers verify that the “leaking pipeline” phenomenon is occurring in medical schools as well.9,10 Further, a recent analysis by the Association of American Medical Colleges confirms a “disproportionately high departure rate” of women from academic medicine.11
The statistics concerning women faculty in academic medicine mirror those of women faculty in doctoral institutions. For example, at U.S. medical schools in 2007, 17% of full professors, 32% of associate professors, and 41% of assistant professors were women.12 Results of a study of the promotion and tenure of faculty from 1969 to 2004 at the University of Minnesota Medical School showed that the same percentages of women and men progressed from assistant professor to associate professor, but a significantly lower percentage of women progress from associate professor to full professor. This was not a result of women taking longer to be promoted to full professor; in fact, the time to full professor was the same for both genders. It was a result of many more women than men leaving the medical school after having achieved the rank of associate professor.13 The findings of that study provided the impetus for our research question: What factors are associated with the exodus of women from academic medicine?
We chose to study various aspects of faculty satisfaction to investigate this question. We did so because faculty satisfaction is a strong predictor of intention to leave.5,14 Further, self-reported intention to leave is the best predictor of actual retention.15 The aspects of satisfaction that we chose to study reflect findings from previous research that satisfaction is predicted by such factors as collegiality, relations with the department chair, mentoring, level of influence in the department, salary, the promotion and tenure process, the intrinsic aspects of work itself, autonomy, stress, institutional recognition and support, department climate, and time constraints.5,16–19 Faculty who are satisfied with these factors are more likely to be satisfied overall and, thus, less likely to leave their current positions and academe altogether.
The data we used to investigate this question were part of the data collected in a larger climate study that we had undertaken to identify perceptions of variables associated with job satisfaction and retention for faculty in the University of Minnesota Medical School. In the present report, we focus on the gender differences in perceptions of those variables. The statistics on women faculty at the University of Minnesota Medical School mirror the national statistics for women faculty in medical schools, with women comprising 28% of full-time faculty and 12% of full professorships. Our hypothesis was that a combination of work–life factors and family–life factors serve as obstacles to satisfaction and retention for women faculty. This study was requested and supported by four parties: the University of Minnesota Center for Excellence in Women’s Health, the Office of Faculty Affairs, the Office of the Provost, and the dean of the University of Minnesota Medical School.
We developed a Web-based questionnaire to collect information on faculty perceptions about professional/collegial relationships, mentoring, obstacles to satisfaction, institutional policies, circumstances that contribute to departure, gender equality, family issues, work life, and demographics. Our questionnaire was adapted from one developed at Johns Hopkins Medical School.10 It was piloted, revised, and repiloted with a subgroup of faculty. The final instrument comprised 72 questions. This study was approved with exempt status by the University of Minnesota institutional review board.
In 2005, we electronically mailed an invitation to complete the questionnaire to the 615 full-time, paid faculty (both tenured/tenure-track and clinical scholars) at the University of Minnesota Medical School. The invitation included a link to the Web-based questionnaire, which took approximately 20 minutes to complete. Faculty were assured that individual questionnaire responses were anonymous, data would be made available only to persons conducting the survey or data analysis, and no reference would be made in oral or written reports that could link a respondent to specific responses. Reminders were sent to nonrespondents, who received up to three reminders at two, three, and five weeks after the initial invitation.*
The survey items predominantly used Likert-like response scales. Respondents were asked, for example, to indicate their levels of agreement with statements by choosing “strongly agree,” “agree,” “disagree,” “strongly disagree,” or “N/A” from the response options. Some questions used response options of “frequently,” “occasionally,” “seldom,” or “never,” and other questions used response options of “very satisfied,” “satisfied,” “neutral,” “dissastified,” “very dissatisfied,” or “N/A.” In addition, some items asked for “yes” or “no” responses, and some asked for specific information. An example is, “In the last one-year period, how many times were you invited to an informal, but important, work-related social gathering with your mentor(s)?” Participants also were asked to provide written comments. We used Statistical Package for the Social Sciences version 13.0 to analyze the quantitative data. For analysis purposes, we collapsed some response categories; for example, we collapsed “strongly agree” and “agree” responses into “agree,” and “strongly disagree” and “disagree” responses into “disagree.” Similarly, we collapsed “very satisfied” and “satisfied” responses into “satisfied,” and “very dissatisfied” and “dissatisfied” responses into “dissatisfied.” We computed means and percentages of, for example, “agreed” and “disagreed” responses or “yes” and “no” responses to items for all respondents and by gender subgroups. We made comparisons between subgroups using the chi-square test; for example, between the percentage of women versus that of men who “agreed” or responded “yes” to an item. We analyzed continuous variables using one-way analysis of variance and coded written responses for themes using NVivo.
Of the 615 faculty invited to complete the questionnaire, 354 (57%) responded. The race, gender, age, and rank demographics of the respondents mirrored those of the total University of Minnesota Medical School faculty. Women comprised 31% of the respondents. Of those who responded, 80 (38%) of the men were tenured full professors, compared with 23 (23%) of the women. With regard to productivity, both men and women with research responsibilities had published, on average, five articles in refereed journals in the year 2003–2004. Faculty of each gender worked approximately 57 total hours each week and, on average, devoted 24 of those hours to research and 11 to education. On average, men allotted 19 hours and women allotted 16 hours to clinical work, whereas women allotted five hours and men allotted three hours to university service. It is clear from these statistics that the men and women in our sample were equally productive by traditional measures and put in similar amounts of time at work.
Sixty-three percent of respondents reported they were satisfied overall with their current positions in the University of Minnesota Medical School. Further, there was no difference in overall satisfaction between women and men. The purpose of this study, however, was to investigate the nuances of satisfaction. In this vein, we analyzed the data to identify significant differences between the responses of women and men for each questionnaire item. Table 2 clusters those items into categories and subcategories, notes the number of items in each subcategory, and shows the percentage of items in each subcategory on which the responses of men and women differed significantly. Laying out the results in this fashion clearly identifies the areas in which there were the greatest differences between the perceptions of men and women faculty.
The results show significant differences (at the level of P < .05 or better) between the perceptions of women and men on items in several subcategories. In general, the differences on most items favor men with regard to organizational climate and family responsibilities; however, we explain directions of significance on individual items in detail in the respective sections of the remainder of this report. We pay particular attention to the subcategories that contained more than two items and that also had more than 20% of items with significant differences by gender: family situations, policies and choices related to family issues, obstacles to satisfaction, circumstances that can contribute to departure, career goals and plans, treatment of faculty with regard to gender equality, conditions in the unit (defined in the questionnaire as the division, department, center, or institute with which the respondent most identifies), and discrimination. These subcategories reflect areas related to work life, family life, and, often, the intersection of the two. We do not include the subcategory of number and type of mentors, because the only two items that showed significant gender differences were the number of female mentors whom a respondent had been formally assigned and the number of female mentors whom a respondent had at the time of the survey. In the following sections, we will discuss the findings related to each of the relevant subcategories in detail.
The data reveal numerous differences in the family situations of women and men faculty (see Table 3). The most striking difference is the number of hours per week spent on family and household responsibilities. Despite putting in an equal number of hours at work, women reported spending an average of 31 hours per week on family and household responsibilities, compared with men, who reported spending 19 hours per week. This difference may be attributable to the differences between women and men regarding the presence of a partner/spouse and the employment status of that partner/spouse. For all respondents, 19% of women did not have partners, compared with only 5% of men. Seventy percent of women had partners who were employed full-time, 5% had partners who were employed part-time, and 7% had partners who were not employed. In contrast, 36% of men had partners who were employed full-time, 26% had partners who were employed part-time, and 33% had partners who were not employed. In sum, only 12% of women, compared with 59% of men, had partners who were employed part-time or were not employed. Almost equal percentages of women and men had minor-age children, but significantly more men (39%) than women (22%) had children over 18 years of age. More women (16%) than men (9%) reported having no children. In theory, since women are less likely than men to have partners and are more likely to have partners who are employed, women are less likely to have help with family and household responsibilities from partners. In contrast, when men have partners who are employed part-time or are not employed, those partners may be more available to help with family and household responsibilities.
Policies and choices
We asked faculty about their interest in a part-time tenure track and about the medical school’s parental and pregnancy leave policies. Given the family situations of women compared with men in our sample, it is no surprise that a significantly higher percent of women than men said they were interested in a part-time tenure track (33% versus 14%). In terms of leave policies, 60% of women and 50% of men thought the time they took for parental leave was not adequate to meet the needs of their families. Women took an average of eight weeks, whereas men took an average of two weeks. Significantly more women (17%) than men (6%) thought their leave resulted in colleagues’ taking on additional clinical or other faculty work. Although the majority of faculty who responded to the question (73%) thought their colleagues’ attitudes toward their pregnancy or parental leave was supportive, significant differences in responses did exist between the genders. More women (88%) than men (61%) said their colleagues’ attitudes were supportive, whereas more men (32%) than women (4%) said their colleagues’ attitudes were neutral.
Obstacles to career success and satisfaction
Significantly more women than men indicated the following items were obstacles to their career success and satisfaction: absence of a part-time promotion track (22% versus 3%), meetings after 5:00 pm and on weekends (38% versus 18%), absence of on-site child care at work (23% versus 11%), absence of emergency child care at work (30% versus 16%), absence of on-site adult care at work (9% versus 2%), inadequate level of formal parental leave policy (19% versus 5%), and inadequate opportunities to pursue research (56% versus 42%). Many of the items that more women than men perceived as obstacles to professional success and satisfaction pertain to the intersection of work and family responsibilities.
Circumstances that contribute to departure
We asked faculty whether they were experiencing certain circumstances that might contribute to a decision to leave academic medicine or the institution for other career options. There were significant gender differences in the percentages who agreed for 6 of the 16 circumstances. These circumstances are conflicts of work and family responsibilities (32% women versus 18% men), inadequate formal parental leave policy (9% versus 2%), too stressful (42% versus 29%), uncertainty of external funding (68% versus 54%), inadequate opportunity to meet career goals (46% versus 32%), and inadequate opportunities to pursue research (53% versus 39%). These results show that, in our sample, women are more likely than men to perceive that they experience circumstances commonly associated with departure from academe.
In addition to gender differences found on items in the subcategories that directly reflect the intersection of work and family life, gender differences also existed on many items in four other subcategories. These subcategories are career goals and plans, treatment of men and women, conditions in the unit, and discrimination. Although the differences in these items may not relate directly to the intersection of family and work life, they are nonetheless the kinds of differences that are important to the satisfaction of women faculty. Responses to some of these items by our women respondents also could be an indirect function of the family situations for them, and perhaps for women faculty in general.
Career goals and plans
We asked faculty three questions about their understanding of the expectations for promotion and their career plans. Approximately three quarters of faculty of each gender reported they fully understood the expectations for promotion in the appointments they held. Significantly fewer women (71%) than men (86%) reported that they had a clear picture of where they wanted to be in their academic career in five to seven years. Further, 73% of men had well-defined plans for achieving their career goals, but significantly fewer women (51%) had such plans. This finding is important, as having clear career goals and a well-defined plan for achieving those goals is important to success. In fact, with regard to research success, an unpublished study of University of Minnesota Medical School faculty conducted by C.J. Bland in 2000 showed that faculty who had a plan for achieving their goals were more likely to be highly research productive than were faculty who did not have such a plan.
We asked faculty a set of 11 questions about how they perceive men and women are treated in the medical school. They could answer by selecting “men,” “women,” “both,” or “neither” for items such as “Some faculty have a condescending attitude toward ——.” The full set of questions and responses are shown in Table 4. There were significant differences in perceptions between women and men on all 11 of the items. In general, the results show that the greatest differences in perceptions were found on items that reflected attitudinal and behavioral aspects of the department based on gender (e.g., condescending attitudes, unequal attention to people in meetings, disrespectful behavior). The perceptions between genders were more similar, even though statistically significantly different, on items that reflected structural aspects of the department (e.g., access to research space, preferential treatment in recruitment and promotion, availability of research and scholarship support). On all items, women perceived more bias against women than did men. In many instances when women reported bias against women, men reported no bias against either gender.
Conditions in unit
Even though differences were found for only 4 of the 17 items regarding conditions in the unit and relationships with unit colleagues, the four items on which a significantly higher percentage of women than men agreed are items that are commonly reported as stressors for women faculty. Significantly more women than men perceived that they are expected to represent the point of view of their gender (25% versus 6%), are constantly under scrutiny by colleagues (28% versus 18%), work harder than they believe colleagues do in order to be perceived as a legitimate scholar (45% versus 21%), and that there are many unwritten rules about how one is to act with unit colleagues (55% versus 40%). Whether or not it is the case, women perceived that they have to put more effort into their work than do men in order to be accepted and successful.
The subcategory of discrimination is another area in which many of the items showed significant differences. We asked faculty whether they had experienced any discrimination in the last five years at the University of Minnesota in the areas of hiring, promotion, salary, space/equipment/other resources, access to administrative staff, and graduate student/resident/fellow assignments. If they answered “yes,” they were asked to note the bases of the discrimination. Overall, very few faculty reported experiencing any discrimination. Of those who did, significantly more women than men perceived gender discrimination in the following areas: promotion (10% versus 2%), salary (31% versus 1%), space and other resources (12% versus 0%), access to administrative staff (7% versus 0%), and graduate student/fellow assignment (5% versus 0%). Even though the overall number of faculty who reported experiencing discrimination was low, there was still a significant difference between the number of women and men who reported they experienced discrimination.
In addition to the quantitative measures, the responses to a qualitative question gives some insight into the perceptions of faculty about the gender climate at the university. The last item on the questionnaire asked faculty, “Are there any other things you would like to tell us about the gender climate at the university?” A total of 69 faculty (19%) responded to this question. The responses clustered into the following themes: there is no gender bias; there is gender bias; there is gender bias, but it is hard to address; and, leaders are essential for addressing and maintaining gender equity. Twenty-seven faculty reported that they perceived no gender-based preferences or biases in their departments. Seventeen of those faculty were men. Twenty-two faculty perceived gender bias, and nine of those faculty were men. Several faculty specifically used the word “subtle” to describe the bias they perceived. Usually, faculty who perceived bias did not describe blatant examples of gender discrimination, but they reported occasional, open discrimination in the context of consistent subtle advantages for men. An example of this perception is illustrated with the following comment from one man:
I have met two female faculty members who complained about the treatment that they received from members of our department. I think that young female faculty members have a lot of problems trying to balance a family life with the typical hospital schedule, no availability of child care, no opportunity to discuss family problems. Apart from that, I believe that there is a subtle gender preference towards males for all leadership positions, for all committees, etc. In short, I believe that the climate is less favorable for women and they are less likely to be heard than men in the U of MN.
Several faculty suggested remedies for the gender discrimination they perceived. The suggestions included having more women in senior and leadership roles and hiring more women on the tenure track. Other faculty mentioned including information on gender and climate in the annual reviews of department heads, as well as information that ensures that department heads abide by university rules and policies regarding equity. This type of qualitative information is valuable, because consistent subtle bias is likely to affect the career success and satisfaction of women, yet the existence of subtle bias may be more difficult to identify and quantitate.
Conclusions and Recommendations
The results of our study support our hypothesis that there are both work–life and family–life factors that serve as obstacles to the satisfaction and retention of women faculty. Many of these factors reflect the challenges women encounter that are attributable to subtle gender bias and the intersection of work and family life. These obstacles may contribute to the departure of women from academic medicine specifically, and from academe in general, before they reach the rank of full professor. The experiences reported by the women in our study are similar to those that women faculty encounter in other medical schools and in higher education institutions in general.9,10,20 Our study thus provides additional evidence that medical schools, academic health centers, and institutions of higher education stand to lose a valuable resource if they fail to address the gender climate and issues related to balancing work and family responsibilities.
To retain women faculty and benefit from their human capital, as well as profit from the initial financial investment in faculty, we maintain that institutions and individual departments must take action in two ways. First, and perhaps most important, institutions and departments must design and implement policy changes that provide support for faculty who must meet the demands of both work and family. Although it is the norms of society that have placed the majority of family responsibility on women, it is not reasonable for institutional policy makers to believe that society will change to meet institutions’ needs. Institutions, rather, must take steps toward developing and implementing policies that will accommodate faculty, particularly women faculty, when child care and elder care become more of a workload.
The women in our study clearly indicated ways in which policy change could help them balance work and family demands and, thus, provide them a better opportunity to persist and succeed in academe. Such policy changes include creation of a part-time tenure track, availability of on-site and emergency child care and on-site adult care, improved formal parental leave policies, and restructured meeting times that do not require evening or weekend hours. Many of these changes would benefit not only women, but all faculty. Academic health centers that have investigated and defined the obstacles to success for women faculty and implemented relevant policy changes have seen improvement in the satisfaction and retention of their women faculty.10
The second course of action that institutions and individual departments must take is to identify and address both blatant and subtle gender bias. According to our study, this bias may exist in, but may not be limited to, areas such as professional relationships, promotion, salary, and resource allocation. Institutions, and departments within institutions, should examine their current practices and make the necessary changes to improve the work climate for women faculty. Such changes should include inspection and adjustment of promotion, salary, and resource allocation procedures that may be biased subtly against women. In addition, institutions should offer quality mentoring programs to help women faculty articulate their career goals and plans and formally include them in networking opportunities. Actions along these lines can help to improve the satisfaction, success, and retention of women faculty.
How each institution addresses the climate for women faculty will depend on its unique circumstances. At the University of Minnesota, there are already many policies and programs in place to support faculty, particularly women faculty. For example, the tenure track allows for time off-track for either parent when a new child enters a family or when a faculty member needs to focus on an illness or care for an elder, and there is the University Office for Women and the University Women’s Cabinet that advises the president. Additional supports are provided in the medical school, such as funding support for one senior woman faculty member each year to attend the Executive Leadership in Academic Medicine program, funding support for women to attend the Association of American Medical College’s programs for junior and senior women faculty, the successful Department of Health and Human Services-funded National Center for Women’s Health, as well as the Building Interdisciplinary Research Careers in Women’s Health program that supports the research development of junior faculty who focus on women’s health. Since this study, the medical school has enhanced efforts in the following areas.
Part-time tenure track
The University of Minnesota comprises five campuses, 66,000 students, and 4,000 faculty who all work under the same tenure code. Any changes in the university tenure code are made by proposals brought by a major senate committee to the entire university faculty senate for a vote. This process has been started with regard to a part-time tenure track, but, if successful, it likely will take several years to be implemented. In the meantime, the medical school dean is encouraging department heads to help faculty who need to work less than 100% time to take advantage of the fact that the university considers any faculty member who is hired at 67% time to be a full-time employee and provides them full faculty benefits. Thus, this policy already allows faculty on the tenure track to work less than 100% time.
Greater focus on faculty affairs and faculty development
The position of associate dean for faculty affairs in the medical school was increased from 50% time to 100% time to improve attention to faculty support and, particularly, to allow more time to focus on policy issues. Further, the medical school recently developed and filled the position of assistant dean for faculty development (50% FTE).
Formal medical-school-wide mentoring program
One of the first responsibilities of the new assistant dean for faculty development was to develop and implement a medical-school-wide mentoring program. This work was facilitated by a policy recently passed by the academic health center that requires all junior faculty to have a mentor or mentoring team. While a formal and required mentoring program will help all faculty, it will be particularly helpful for women, who (if our study findings are representative) are less likely than men to have clear career goals and a plan for where they would like to be in their academic career in five years. Further, research shows that informal mentoring is less likely to occur spontaneously for women than for men.21
There are limitations to this study. Because the scope of the study was confined to only one medical school at a major research institution, our results may not be generalizable to other medical schools, other types of schools within institutions, or other types of institutions. Because of the study design, we are unable to determine causality. The findings that women are not as satisfied as men on certain items do not allow us to argue that those areas of dissatisfaction are the reason women faculty leave. In addition, our data are based on self-reports of experiences and perceptions; as such, respondents’ perceptions are based on their own constructions of reality. This study also had a low response rate at 57%, and we were not able to determine whether the attitudes of nonrespondents differed from the attitudes of respondents. There were, however, no major differences in the demographic information of respondents and nonrespondents. Despite these limitations, our study gave us valuable information regarding the experiences of women and men faculty and a foundation from which to recommend change.
In summary, the perceptions of work environments and family responsibilities by the women faculty we studied were sometimes significantly different from those of the men faculty studied. Women also reported that they experienced more obstacles to success and satisfaction related to family responsibilities, climate, and institutional policies. These differences may contribute to the underrepresentation of women in the full professor ranks and to women faculty leaving academe altogether. The good news is that by examining current practices and policies and making appropriate changes, institutions can play an active role in retaining and nurturing the wealth of human capital provided by women faculty.
We dedicate this report to our dear colleague and friend, Carole Bland, who passed away on August 23, 2008, from pancreatic cancer after a struggle marked by courage, dignity, and grace.
Carole devoted her extraordinary career to the professional development and productivity of faculty, administrators, and institutions of higher education. Her passion for cultivating faculty vitality—whether through mentoring, leadership development, or institutional change—was exceeded only by her vast scholarly contributions to the same field. She leaves behind a legacy of scholarship and programs that will continue to inform and inspire faculty and administrators for years to come. We are grateful for having had the privilege of working with her.