Faculty Salary Equity: Still a Gender Gap?
John W. Curtis
By John W. Curtis, director of research and public policy, American Association of University Professors
When the American Association of University Professors (AAUP) added systematic collection of salary data by gender to its annual report on full-time faculty salaries in 1975–76, the overall average salary for women faculty members was 81 percent of that for men. In the 2009–10 report released in April, the proportion was…81 percent. So after more than thirty years of women’s increasing participation as faculty members in American colleges and universities, women’s salary disadvantage has not eased one bit. Gender inequity, it would seem, is alive and well in higher education….or is it?
The 2009–10 AAUP data indicate that women full-time faculty members earn less than their male colleagues at each of the traditional professorial ranks (professor, associate professor, and assistant professor), and overall in each institutional category (doctoral, master's, baccalaureate, and associate). Although women are approaching salary parity with men in community colleges, even there women have a slight overall disadvantage. The overall salary disparity between men and women is the product of both rank and institutional location: Women are more likely to hold faculty positions at lower ranks, and they make up a greater proportion of the faculty at the institutions that pay the lowest salaries. Moreover, the AAUP data do not even take into account women’s overrepresentation in part-time faculty positions, which pay piecework wages and rarely provide benefits.
When Martha West and I released the AAUP Faculty Gender Equity Indicators in 2006, we made the aggregate measure of full-time salary equity the fourth of our indicators. The other three indicators were employment status (full- and part-time), tenure-track status, and attainment of full professor rank. The report argued that “[t]he four indicators represent different aspects of the overall status of women faculty, which at current levels amount to a series of accumulated disadvantages…and the accumulated disadvantages of position are exemplified by the comparison of overall average salary in the final indicator” (2006, 7). It then presented analysis on each indicator for more than 1,400 colleges and universities. For nearly all institutions that granted degrees at the baccalaureate level and above, the overall average salary for women full-time faculty members was lower than that for men. Yet when faculty members tried to raise this point on their campuses, they frequently heard from administrators and skeptical colleagues alike comments such as “the overall averages don’t tell the whole story; you have to take other factors into account.” In other words, even though women faculty members earn less overall than their male counterparts, that is not conclusive evidence that any individual woman has been disadvantaged in her career.
To a certain extent, this response accurately reflects the caution against inferring the situation or beliefs of one individual from the characteristics of a category to which he or she belongs (or has been assigned). However, it also reflects the difficulty of turning academic sociological analysis into policy solutions. For several years, I have led summer workshops with faculty leaders grappling with these issues, and it is an opportune time to take stock of where the higher education community stands in terms of analysis and action on resolving gender pay inequities.
The Personal Is Methodological
A key insight informing an understanding of faculty salary equity is the distinction between an academic analysis of gender differences in earnings, and an equity study that is intended to pinpoint local inequities so they can be ameliorated. The typical quantitative journal article, using multivariate statistical analysis of data drawn from a large set of institutions, aims to specify and “explain” differences in earnings between men and women faculty members. The analysis will include measurements intended to quantify the impact of differences in background characteristics, academic productivity, and career “choices.” (Another article would be needed to explore why the concept of “choice” is so problematic in this context.) In these analyses, the available data are always somewhat limited, at least in part due to the use of questionnaires based on individual reporting. Researchers performing these quantitative analyses will construct a statistical model for how various factors might affect earnings with the objective of “explaining” the relative importance of different factors and determining whether there is still a difference by gender, even after “controlling for” all other factors.
Using the parlance of quantitative social science, even the most sophisticated analysis ends with “unexplained variance.” Toutkoushian and Conley’s comprehensive review and extension of various analytical models developed during the 1990s found that the “unexplained” salary gap between men and women faculty members remains at between 4 percent and 6 percent (2005). Porter, Toutkoushian, and Moore attempted to determine whether this gap is due to a difference in starting salaries, inequities in salary advancement, or both (2008). They found no statistically significant wage disadvantage at the time of hiring for women overall, indicating that the aggregate difference “is driven primarily by larger unexplained wage gaps for more senior faculty,” but they did find a troubling 9 percent gap for recently hired women faculty members at research universities (483). The situation at research universities is particularly important because of the large numbers of full-time faculty employed there and the relatively high potential salaries.
As we noted in the Gender Equity Indicators,
[a]lthough it is not appropriate to attribute this remaining differential to discrimination on the basis of this evidence alone, the statistical analyses clearly leave a series of questions unanswered: Why is the proportion of women faculty holding doctorates smaller than the proportion among men? Why are women less likely to obtain full-time tenure-track positions? Why are they less likely to be employed in research universities? Why do women faculty generally spend more of their time on student advising and committee service than do men? Why do positions in the disciplines in which women faculty are concentrated generally pay less? Why are women less likely than men to earn tenure and promotion to full professor? Why do they earn less on average at every rank than their male counterparts? (13)
Quantitative analysis will not provide answers to these questions in the foreseeable future, because the data simply are not being collected. But until we have answers, we are not likely to succeed in eliminating gender pay inequities.
In contrast to an academic analysis, an equity study is usually based on “nearly perfect” data from a single institution or system. Data are available for every faculty member, and each desired indicator can be measured exactly—although measurement of outcomes like “productivity” remains a fraught concept. The only restriction on the data is often the number of “cases,” meaning the number of individuals in the study population. This is especially limiting when there is either a small faculty (one hundred individuals or fewer) or a small number of individuals from minority categories. Because of these limitations, conducting a salary equity study that examines both gender and racial or ethnic factors can be next to impossible on a small campus.
In an equity study with essentially complete data, there is no such thing as “unexplained variance.” Differences in salary are actual differences between individuals, and the only question is whether those differences are “justified” (representing demonstrably relevant differences in work products, skills, or experience) or “impermissible” (the result of discrimination). Similarly, there is no such thing in an equity study as “statistical significance,” and this is often difficult for academics to accept. Any college student who takes a quantitative social science course learns that analysis is affected by sampling, probability, measurement error, and estimation. In an equity study there is no sampling—we have data on everyone—and there is no real measurement error, because researchers are working directly with payroll records and personnel files rather than with answers to a questionnaire. Too many campus equity studies are stymied by a misplaced finding of “no significant difference” in salaries between men and women faculty members after the relevant factors have been “controlled.” That’s the bureaucratic equivalent of telling a woman faculty member who earns less than her male colleague who joined the department two years later, “Sure, your salary is lower, but your career doesn’t really matter, anyway.”
In order to ensure that equity studies will truly serve their intended purpose, it’s important to keep a few principles in mind:
- Differences in salary are real and have a cumulative impact.
- The study process must be open and involve faculty in a meaningful way.
- The study should be repeated every three to five years, depending on the size of the faculty.
A study alone will not end inequities. The only truly effective remedy for inequity is the adoption of more standardized (and open) methods of determining initial salaries, increases, and special awards. As long as salaries are determined primarily by private individual negotiation or administrative discretion, inequities will reemerge.
Band-Aids and Consciousness-Raising
An abiding paradox in the history of attempted remedies for gender salary inequities is that the analysis of the problem is structural, yet the remedies are almost always individual. Multivariate quantitative methods, by their nature, measure differences in outcomes for groups or categories. The analytical nature of the process disguises the result somewhat, but the fact is that a finding of salary disadvantage for women means that all women faculty members are earning less than they should be, because they are women. Yet remedies for inequity are often underfunded, focus only on “fixing the worst cases,” or require individual faculty members to negotiate individual resolutions within the inequitable structure that failed them in the first place.
This type of piecemeal approach is one reason why we have made so little progress in the last four decades toward overall salary equity. It does not address the underlying problem. It belies the oft-heard contention that “things are getting better; it’s just a matter of time.” Maybe so—but analysis by Marschke and her colleagues (2007), though limited to one research university, indicated that at the rates of progress found there throughout the 1990s, it would take fifty-seven years for women to make up 50 percent of the full-time faculty. “Time” can move awfully slowly.
One reason that gender inequities are so intractable is, paradoxically, that we have made some progress. It is rare these days to hear blatant discriminatory views openly expressed in academia. Laws and regulations against discrimination are being enforced, even if the process can be painfully slow and discouraging. The structures and behaviors that produce inequities are much more nuanced, hidden, and subtle than open discrimination, which makes the resolution that much more difficult. As Virginia Valian argued in her 1999 book Why So Slow?, women faculty members are subjected to a “death of a thousand cuts” and the accumulation of slight disadvantages at every step of the way. Valian’s work amounts to good old-fashioned consciousness-raising, an attempt to help individuals—men and women alike—recognize the ways in which they allow their assumptions and subconscious prejudices to affect their judgment and keep barriers to women’s advancement in place. Joan Williams and her colleagues (see, for example, Williams and Bornstein 2008) have extended this line of analysis, integrating the insights of social psychological research with legal remedies to create a “carrot-and-stick” message for colleges and universities: Figure out how to eliminate bias against women (and caregivers more generally), or you will lose many talented faculty members and may end up in court.
In a further interesting twist on the consciousness-raising approach, Cress and Hart (2009) use a series of sports metaphors to illustrate the differences in lived experiences between men and women faculty members at the same institutions. They argue that such metaphors provide insight into the ways in which current structures form “modes of operation that focus on resources, reputation, and expertise at the expense of inclusion and innovation,” and in the process perpetuate disadvantages for women faculty members (483). They end with what should by now be a classic argument about the advantages of inclusion, similar to arguments advanced by other proponents of equity in academia:
[T]he rich array of talents and skills of faculty (as well as students) emerge when individuals are appreciated for their unique abilities and knowledge. Everyone is a valuable player of Team Academe. Perhaps the best solution is to create an entirely new field and a new game where all members of the faculty feel welcome. Until that time, truly leveling the playing field is an academic imperative. (485)
Women now comprise the majority of students and graduates in our colleges and universities, even among recipients of doctoral and professional degrees. As parents, mentors, and advisers, we tell these women they can “be anything they want to be.” But when they look to the faculty members helping them search for understanding and solutions to all the world’s problems, do they find role models who demonstrate that potential? Well… maybe they can be 81 percent of whatever they want to be. That’s close enough…right?
I like to think I’m a pretty enlightened guy. When I was in grad school in the mid-1980s—in sociology, no less—there was only one woman full-time faculty member in our department, a tenured full professor and established expert in her field. My fellow grad students, mostly women, would mutter that “she’d had to act like a man” to get where she was. Back then, I never quite understood what they meant. When I began working as director of research for the AAUP in 2002, a (woman) colleague asked me to compile some data disaggregated by gender. It hadn’t occurred to me to do that particular analysis, and I was one of those people who thought we had solved “all of that equity stuff” back in the 1970s. When I did put the data together, I was floored by the disparities that remained. Now, in all fairness, I had only been in the job for a month, and I’ve learned a little since then, but clearly some of us still need an occasional kick in the head (or at least in the seat of the pants) to get the message across: We haven’t reached equity yet. We need to keep working on this.
Let’s keep kicking, shall we?
American Association of University Professors. 1976. Nearly keeping up: Report on the economic status of the profession, 1975-76. AAUP Bulletin 62(2): 195-284.
———. 2010. No refuge: The annual report on the economic status of the profession, 2009-10. Academe 96(2): 4-80.
Cress, C. M., and J. Hart. 2009. Playing soccer on the football field: The persistence of gender inequities for women faculty. Equity and Excellence in Education 42(4): 473-88.
Marschke, R., S. Laursen, J. McCarl Nielsen, and P. Rankin. 2007. Demographic inertia revisited: An immodest proposal to achieve equitable gender representation among faculty in higher education. Journal of Higher Education 78(1): 1-26.
Porter, S. R., R. K. Toutkoushian, and J. V. Moore III. 2008. Pay inequities for recently hired faculty, 1988-2004. Review of Higher Education 31(4): 465-87.
Toutkoushian, R. K., and V. Martin Conley. 2005. Progress for women in academe, yet inequities persist: Evidence from NSOPF:99. Research in Higher Education 46(1): 1–28.
Valian, V. 1999. Why so slow? The advancement of women. Cambridge, MA: MIT Press.
West, M. S., and J. W. Curtis. 2006. AAUP Faculty Gender Equity Indicators 2006. Washington, DC: American Association of University Professors. www.aaup.org/AAUP/pubsres/research/geneq2006.htm.
Williams, J. C., and S. Bornstein. 2008. The evolution of “FReD”: Family responsibilities discrimination and developments on the law of stereotyping and implicit bias. Hastings Law Journal 59: 1311-58.