Where Have All the Tenured Women Gone?
Ingram Olkin, Stanford University
This has been a banner year for books, articles, editorials, panels, and discussions about women’s issues in industry, government, and academia. Larry Summers’ statement about women in science and mathematics was a catalyst for a flood of responses by women in academia. There were many seminars and editorials in universities across the nation. The recent book by Sheryl Sandberg, Lean In: Women, Work, and the Will to Lead, focused on the corporate world, and the recent Harvard Case Study on Gender Equity focused on inequities in academia (see The New York Times review, September 8, 2013).
The Mathematical Association of America (MAA) provided an analysis on the number of women authors in three of their journals (see Focus Oct/Nov 2012). The proportion of women members in MAA is about 25%, and the proportion of women authors in these journals is 72/583 = 12%. However, information about submission rates has not been gathered. I suggest editors of statistics journals begin to collect such data so as to try to determine the reason why women are not more visible as authors.
The JSM meeting in Montréal made it abundantly clear that women are under-represented in named lectures. Of the four named lectures (i.e., Wald, Rietz, Neyman, Fisher), the seven medallion lectures, and the two invited lectures, none were women. The Caucus for Women in Statistics took note of this and collected signatures in a petition calling for change. Terry Speed wrote an impassioned op-ed article in which he described areas of inequity.
This story is not new. Just a year ago, Amanda Golbeck noted the under-representation of women speakers at JSM 2012 (see “Where Are the Women in the JSM Registration Guide,” July 1, 2012, Amstat News).
My focus here is on the scarcity of tenured women in departments of statistics (excluding biostatistics) at universities offering doctorates. In 1995, I collected data from eight private universities—Carnegie-Mellon, Chicago, Duke, Harvard, Penn, Rice, Southern Methodist, and Stanford—on the number of tenured women in the statistics department. The results in increasing order were:
0/12, 0/11, 0/7, 0/7, 1/12, 1/10, 1/9, 2/15,
with a composite of 5/83 = 6%. The data from 21 public universities that offered doctorates—Arizona State, Berkeley, Colorado State, Davis, Florida State, Illinois, Iowa, Iowa State, Michigan, Minnesota, North Carolina, North Carolina State, Ohio State, Penn State, Purdue, Riverside, Rutgers, South Carolina, Texas A&M, Washington, Wisconsin—show:
0/24, 0/18, 0/14, 0/14, 0/12, 0/11, 0/9, 0/8, 0/8, 1/21, 1/16, 1/14, 1/14, 1/10, 1/9, 1/7, 2/23, 2/17, 2/4, 3/11, 4/32,
with a composite of 20/243 = 8.2%. (Note that the method of counting the number of faculty was often unclear. In two cases, there were two figures from 0/12 to 1/17 and from 0/8 to 1/13. This changed the composite to 22/173 = 8.1%.)
As in the case of private universities, this is not a cheerful outcome. But when one adds the fact that the pool of women doctoral students is probably in the order of 40%–50%, this becomes an embarrassment for our profession.
The Conference Board of Mathematical Sciences conducts a survey every five years on undergraduate programs in the mathematical sciences (including statistics) in the United States. Although data on doctoral programs is not included, there is data on the faculty. In 2005, the proportion of tenured women in doctoral statistics departments was 79/525 = 15%, and, in 2010, it was 95/560 = 17%. Furthermore, the number of tenured men and women under the age of 45 was 18% for men and 5% for women. Thus, the early generation of men and women is highly unequal. This is an important age group because it often serves as a role model.
To me, the picture is bleak. Of course, the next question is why this is happening. Often, the pipeline is cited as the reason, but with the high number of doctorates awarded to women, this reason is faulty. The MAA article includes a discussion of what is called “implicit bias” that notes that “even well-meaning individuals can display bias.” Consider the following statement made either at admission of students, hiring of faculty, or giving an award: Let’s choose the best candidate. At face value, this is a most reasonable statement. However, there lurks in the shadow the definition of “best,” which often means “just like us,” and, of course, the “us” are men. There are also examples, all too numerous, of “explicit bias.” These have been documented in many articles.
Women faculty members in a department of mathematical and statistical sciences are often isolated, with perhaps only one or two colleagues. They generally have difficulty in generating collaborative research. The National Science Foundation could help this group by providing summer support for them to visit statistical centers that have strong summer research activities. This is most profitable for faculty in their third or fourth year. Such visits permit them to present their work to a wider audience and interact with a more established statistical faculty, which they may not have in their home institution.
There is an incentive program in the United Kingdom called Athena SWAN (Scientific Women’s Academic Network) that offers bronze, silver, and gold awards to university schools or departments that increase equity by increasing the number of women students or activities that improve gender equality, or for achievement in equality in career progress in STEMM (science, technology, engineering, mathematics, medicine).
I would argue that the situation will not change by much at the departmental level. We have had ample opportunity for change, but it has not happened. A perusal of the composition of tenured faculty over the last 20 years will provide ample evidence of our falure at bringing women into statistics departments. I believe the university administration at the dean’s or at the provost’s level needs to step in by offering incentives such as positions for women (and, of course, for other minorities). There are some success stories. For example, North Carolina State University Dean Dan Solomon made a special effort to increase the number of women in the statistics department. However, it may take a more dramatic effort in the form of an incentive, such as extra travel or departmental research funds for each new tenured woman faculty member. The cultures at universities are varied and each administration needs to determine what type of incentive might succeed in increasing diversity in statistics departments.