The Role of Geographic Shackles on Academic Careers of Women DRAFT, 7/13/05 by Dina Q Goldin, Ph.D. Computer Science & Engineering, University of Connecticut 1. Introduction The dire situation of women in academia, particularly in the sciences, has attracted much attention recently. The metaphor of the shrinking (or leaking) pipeline summarizes the problem, without specifying the mechanisms for the gradual loss of women along the way. This paper focuses on one of the factors that contribute to the shrinking academic pipeline, which is the greater lack of mobility by women. Women with family obligations are much more likely to be tied down geographically than their male colleagues in the same predicament. If their career demands it, men will pack up their family and move. A woman who will do the same is an exception rather than a rule. We shall be referring to this lack of mobility as "geographic shackles". Among the factors that contribute to the shrinking pipeline is the related problem known as the "two-body problem": women are more likely than men to have a partner with a career, necessitating a dual job search whenever they move. The two body-problem does reduce mobility, but the phenomenon of the geographic shackles is more general. In our informal surveys, we have found many shackled women who have no partner, or are the breadwinners in their families. Conversely, we have seen families move for the sake of a man's budding academic career even when his wife had to give up a well-paying job. The geographic shackles have been under-appreciated as a separate problem worthy of independent investigation. However, we show in this paper that they provide a unifying explanation for the many observed disparities in academic career success between the genders, such as the shrinking pipeline, the prevalence of childless women among the academic elite, the prevalence of women at lower-paying and less prestigious institutions, etc. Up to now, these disparities have received separate, and sometimes quite unconvincing explanations. In this paper, I focus on the academic pipeline only, consisting of Ph.D. candidates and recipients who have decided on an academic career in computing, or in other technical fields. This starting point allows us to assume that the people in question are willing and able to take on academic work, and are prepared to put in the required effort. Thus, gender issues related to the primary and secondary education in computing and sciences are not relevant to our discussion. While we acknowledge that other factors are at still at work, such as gender bias [Valian] or greater share of parenting responsibilities [Academe], we show in this paper that shackles constitute a very significant factor that should not be ignored. Outline: We explain further what we mean by "geographic shackles" in Section 2. In Section 3, we discuss their role in the shrinking of the pipeline. In Section 4, she show their role in some other observed statistical anomalies. We conclude with some suggestions in Section 5. Note that the explanations offered in this paper are by necessity of informal and anecdotal nature. Our aim is to raise awareness of this issue, to start a discussion regarding it, and to call for appropriate studies -- to confirm the presence of the geographic shackles, to understand their causes, and to ameliorate their effects. 2. The nature of the shackles It is not only husbands/partners that tie women to a place, but also other family members (children or elderly parents) as well as the support networks that women build around them to help with the care of their house and family, such as doctors, gardeners, or child care providers. Whereas the two-body problem makes it more difficult to find a new location that satisfices both careers, the geographic shackles make it more difficult to decide to look for a new location in the first place. Rather than exemplifying a lack of commitment or ability, or resulting from economic considerations, the shackles binding the academic women are a reflection of a different set of deep-seated priorities. For men, it is a matter of choice whether to give up a good career move and stay put, or to bite the bullet and move. By contrast, for whatever reason, shackled women seem to feel they have no such choice. Perhaps socialization plays a role here, but we suspect that the underlying causes are deeper, and closely connected to the innate differences of the two genders in coping with stress. These differences have now been widely recognized by the mainstream psychology research community; we are referring to what is known as fight-or-flight vs. tend-and-befriend phenomenon: women are much more likely than men to seek out and use social support in all types of stressful situations, including over health-related concerns, relationship problems and work-related conflicts. It is one of the most robust gender differences in adult human behavior [Azar] The causes of the shackles must be properly understood before they can be fully addressed. The above is only our conjecture, clearly indicating a need for basic research studies. 3. The shrinking pipeline among academic women Among the women in the academia, the shrinking takes place at different stages of their careers. If all goes well, they will find a job as an assistant professor after they graduate; they may have to find another position if their job does not work out for any reason; they may have the opportunity to take on visiting assignments that increase their visibility and help get promoted; finally, they may be lured away for a prestigious position as endowed chair or center director elsewhere. As we discuss in this section, the geographic shackles serve as a significant career impediment at each of these stages. 3.1 Choice of an Academic Position It was observed in [Report] that women apply on average to much fewer schools than men: 6 vs. 25. That report blamed the difference on women's lack of self-esteem. However, the report ignored the effect of geographic shackles: women tend to apply to those places which they can reach without moving. Luckily, women also tend to get more offers per application than men, thanks in part to the efforts of national organizations such as the CRA and the NSF which have been urging departments to increase their numbers of women faculty. As a result, the difference created by the geographic shackles at this first stage of the pipeline is not as noticeable. The same report considers "geographic factors" for new hires, but only as part of a decision whether to accept a job once offered. Since the geographic filtering takes place earlier, at the application level, it is no surprise that geographic factors did not turn up as criterion number one in this decision, either for men or for women. 3.2 Denial of Tenure It is well documented [Valian] that the tenure process in sciences and engineering is subject to gender bias. However, men are often denied tenure as well; tenure denial has been a fact of life since time immemorial. The difference between the genders stems less from the rate of tenure denial than from their differential ability to rebound from this setback, as explained by the geographic shackles. A worthy man who has been denied tenure for whatever reason will find a tenured position in another part of the country, whereas the geographic shackles will constrain a woman to look in the same geographic area. How many research universities in a given locale and in a given year will have an open tenured position in a given specialty? Given the low chance of finding an associate professor's position, academic women who have been denied tenure are often forced to either retreat in their academic career -- taking on jobs as instructors or assistant professors -- or abandon academia altogether. The following story, whose author asked to remain anonymous, is typical: Unfortunately, this is completely true in my own department. Our dean laid off both our untenured (women) assistant profs after 3 years of service due to "financial reasons". Both are tied to the area and are unable to search for positions beyond a limited geographical radius. One was able to find a visiting position, and the other a non-tenure track position. Without the geographical restrictions, I would have expected both to move immediately into tenure-track positions based on their qualifications. It is an open secret that the tenure process is arbitrary. Despite invariable objective measures, this process is subjective, where the objective criteria are systematically manipulated to justify gut decisions. The arbitrariness of the tenure process has a larger negative impact on women than on men: the geographic shackles will keep a woman from surviving an unjust decision by moving, as has always been done by the men in the profession. 3.3 Changing jobs The same report [Report] as mentioned above (Sec. 2.1) notes that when changing jobs, the picture changes from that for new hires (where the average number of applications was 6 for women vs. 25 for men). Now, women submit more applications than men: 5.6 vs. 5.1 on average. The report speculates that by the time they change jobs, "women have learned how to game the system more effectively or have gained more confidence in their abilities." When viewed side-by-side with the men's numbers, this explanation does not make much sense, for several reasons that we will not discuss here. But again, geographic shackles provide a much simpler explanation. The report failed to notice that a change of academic jobs may take place for two separate reasons, leading to two distinct populations of job changers. The first is due to denial of tenure (or some other serious lack of fit) at the current position that forces someone to leave. In this case, the number of applications is expected to be about the same as for a new hire, which was 6 vs. 25 (see Section 2.1). The second population of job changes are those who switch as a result of a better opportunity elsewhere, though they are happy in their current job. Typically, a professor is "invited" to apply for a tenured position which represents a good career opportunity -- perhaps at a much better school, or perhaps an endowed chair or a center director. For this type of job change, the statistical count will record just one application, unless the professor was lucky to be wanted by more than one place at once. Opportunities for a much better job are as likely to happen in California as in Massachusetts, and they normally entail moving. Geographically shackled, women are not open to such invitations, and therefore the second group of job changers, with one application each, will be predominantly men. It is not a surprise then that the average number of applications for job changers, when compared to the number of applications for new job hires, drops greatly for men but remains almost the same for women. There is no evidence that confidence plays a significant role here. 4. The condition of women in the academia In addition to the inexorable shrinking of the pipeline, the geographic shackles provide an explanation for other phenomena that have been observed with respect to women in the academia: they tend to earn less, even at the same position, and the ones who make it big are more likely to remain childless. The present section considers these issues through the prism of geographic shackles. 4.1. Working further and for less Often, women are forced to commute for 1 to 1.5 hours when the best available academic job is not close to home. Despite this inconvenience, they are more attached to the job than a comparable male, who is not held back by the shackles. Ironically, the women's unwillingness to move closer has been interpreted by tenure committees as lack of commitment to the job, making them more likely to be denied tenure. It is also known that academic women are more likely to take jobs further down the totem pole -- as instructors, or as faculty at non-research institutions. These jobs are lower-paying, and involve a different mix of teaching, service and research, but they usually do not involve a smaller workload. Why do they take them? Often, because of geographic shackles, they have no better choice. The following story is typical: I teach in a small liberal arts school and we have a disproportionate number of women in our science departments. Our pay scale is very low... so one theory is that men won't take the positions, but women, especially those whose husbands are working in the area, will. 4.2. Remaining childless It has been observed that women who are leaders in the academia are much more likely to be childless. An obvious explanation is that they are not burdened by child care. However, it has been shown [Valian] that child care issues find their solutions and do not by themselves result in academic women's loss of productivity. Once again, geographic shackles are to blame. They do not affect all women equally. Single women, as well as, to some extent, married women without children, are much more likely to be willing to pick up to follow their career. They can choose from a wider array of jobs after their doctorate; they can move when a given job does not pan out for whatever reason; they are open for prestigious positions as chairs or center directors no matter where those may open up; and they can accept high-profile visiting assignments that provide further boost to their careers. Also, without changing jobs per se, they can use outside job offers to gain negotiating leverage with their current institution; this tactic works only if the threatened job change is believable, so geographic shackles play a factor here as well. While any one missed opportunity may not signify a death blow to the academic career of a woman with shackles, her inability to seize any of those opportunities adds up over time, summing up to a career that is very unlikely to place her in a leadership position in her field, even if she survives to full professorship. 5. Conclusion The shrinking of the pipeline takes place at different stages of women's academic careers. They may not be able to find jobs after they graduate; they may have to leave academia or be sidetracked to non-tenurable positions when their job does not work out; and they are less likely to take prestigious positions as chairs or center directors when those open up. We have shown that the geographic shackles play a significant role in each of these stages. In addition, we have shown how the geographic shackles explain a range of observed phenomena related to the statistical disparity between the academic careers of men and women, from their lower salaries to the higher rate of remaining childless. The many individual stories of talented and capable but geographically shackled women are compelling, but it is the cumulative effect of the shackles on the population of academic women that really deserves attention. The obstacles generated by the geographic shackles in the case of individual academic women translate into significant statistical differences at the level of the population. For example, one experiment [Valian] set up a model that built in a tiny bias in favor of promoting men; after a while, 65 percent of top-level employees were male. Conversely, the "accumulation of disadvantage" for women creates very real job detriments. Like Occam's razor, the geographic shackles explain the statistical gender differences in the trajectories of academic careers without the need for assuming any additional differential factors with respect to ability, commitment, or household duties. While the presence of the shackles is easy to verify with simple statistical studies, basic research needs to be carried out before the problem can be fully addressed -- to understand the causes of the geographic shackles, to reduce the career impact on the women bound by them, and ultimately to minimize their prevalence. Though we expect that the shackles will always be with us to some degree, their current prevalence should not be taken for granted. For example, as increasing Internet connectivity and bandwidth makes it easier to develop support relationships and to maintain social connections over long distances, we might expect to observe that women will make increasing use of internet- based support and social networks that are independent of geography, thus loosening their geographic shackles. However, the academic establishment will also need to take active steps towards solving this problem; we propose a few possible suggestions next. One proposal is to change the perception that a woman's unwillingness to move represents a lack of commitment (see Section 3.1). Another, perhaps wishful thinking, is to change the tenure process so it is less arbitrary. As we discussed above (see Section 2.2), this arbitrariness has a larger negative impact on women than on men: the geographic shackles will keep a woman from surviving an unjust decision by moving, as has always been done by the men in the profession. But most importantly, we need to recognize that for a woman with a family, career decisions that entail moving involve family considerations to a degree rare for men. Therefore, her champions need to work with her whole family to arrive at a positive decision, rather than expecting her to convince the rest on her own. We also believe that raising the awareness about the shackles is an important contribution of this paper in its own right. When a woman intent on an academic research career forms a family, the need to anticipate mobility is often left unarticulated, even when other conditions of her profession, such as long hours, are well understood by both partners. By contrast, if she were an executive of a multinational company, the need for mobility would be acknowledged up front. Given this omission, geographic shackles are practically inevitable. Current efforts by organizations such as the CRA-W and the NSF have been focused on hiring more fresh women Ph.D.s and improving the environment in which they work to achieve tenure. This paper provides a new, additional, focus for the efforts of these organizations. As long as the role of geographic shackles remains under-appreciated and poorly understood, the situation for women in the academia will continue to remain dire. 6. References [Academe] Do Babies Matter, by Mary Ann Mason and Marc Goulden. AAUP Academe, Nov/Dec 2002 (part I) and Nov/Dec 2004 (part II). [Azar] A new stress paradigm for women, by Beth Azar. Monitor on Psychology, Volume 31, No. 7, July/August 2000. [Report] RECRUITMENT AND RETENTION OF FACULTY IN COMPUTER SCIENCE AND ENGINEERING, by John Stankovic and William Aspray. CRA report, 2003. [Valian] Virginia Valian, Why so Slow? The Advancement of Women. MIT Press, 1998. ======================================================================= Note: this is a draft of a paper that is intended for publication; for anyone intending to cite it, please watch this space for citation information