3. Mathematics Faculty

Robion Kirby claims (AWM Newsletter, Nov-Dec 1995) that "... there is, in print, nothing remotely close to evidence or argument that women are discriminated against in the math community" (emphasis added). I'm not sure what standard Kirby is applying here - mathematical proof? guilt beyond reasonable doubt? preponderance of evidence? a videotape of a meeting in which someone says, "we won't hire any women"? Two years ago I wrote an article entitled "Time for Advancement" (TFA) which (after being rejected by the AMS Notices) appeared (i.e., it is in print) in the Dec. 94 MAA Focus. The evidence there convinced many people, if not Kirby, that problems exist. Tables 1, 2, and 3 repeat and update some of that information. (All the information in these tables is computed from data that appeared in the AMS-IMS-MAA Data Survey reports published in the AMS Notices.)

Discussion of the presence or absence of women at top institutions often becomes a debate on whether or not a particular woman is good enough for department X. Therefore, I feel strongly that it is more productive to focus on the many group II and III departments who have no women, one woman, no tenured women, etc., etc., even though they can not reasonably claim there are none good enough. [Here, and in the tables, I use the AMS classification which breaks doctoral mathematics departments into groups I (top 39, based on the 1982 NRC survey), II (next 43), and III; groups IV and V designate departments of statistics and applied mathematics respectively, RI denotes a research institute, and M and B denote departments whose highest degree is a master's or bachelor's.] Table 1 shows that there is little difference in the percentages of women among tenured faculty at groups I, II, and III, but that the proportion more than doubles as one moves to M departments and jumps again at the B level.

Table 1: Women among Tenured Doctoral Faculty

	I	II	III	I-III	IV	V	M	B
1991	4	5	5.5	5	7	3	11	14
1992	4	5	5.5	5	6	6	12	15
1993	5	6	6.5	6	6	5	11	15
1994	4	6	7.6	6	7	5	13	15
# of Departments	39	43	88	170			250	1000
Approx. # Faculty	1450	1300	1700		700	250	3000	4000

Contrary to what is widely believed, this disparity in the representation of women faculty is not simply an historical artifact. For example, women who received PhD's as recently as 1994 or 1995 were almost twice as likely a men to obtain their first position in a bachelor's departments. (18% of women vs. 10% of men in 1994 and 16% of women vs. 9% of men in 1995.) With the high percentage of women receiving PhD's in the 1980's, is it really credible that so few group II or III department managed to hire and/or tenure women faculty? Could more have been hired without lowering standards? The data in table II suggest that for the past 10 years at least 15% of the pool was female. But even if it was only 12% and a department made 8 hires in that 10-year period, the probability of not hiring a women would be only 36%. Now, for a given department this is not prima facie evidence of blatant discrimination. But it does make me extremely skeptical about the frequent claim that departments which have assiduously tried recruit women were unable to do so.

It is sometimes claimed that the disparity in hiring is a consequence of the fact that women are significantly more likely to get degrees in statistics and slightly more likely to receive PhD's from group II or III departments than group I. However, in the 4-year period 1988-91 (the first for which for which this data is available) women did not even get their first position at doctoral institutions (i.e. I-III+RI) at the rate at which they received them from group I departments. From 1992-95 women do seem to have achieved parity for first positions at doctoral departments. However, this was accompanied by assertions that women were receiving preferential treatment and a disproportionate share of the jobs, although the data do not support such claims. It seems worth repeating a statement I made in TFA, noting that the data now available for 1994 and 95 seem to support my parenthetical remark about "statistical fluctuations".

In 1993, several things occurred simultaneously - the percent of women receiving PhD's from group I departments jumped from a previous high of 17% to 22%; the distribution of women receiving PhD's from groups I, II, and III was more uniform; groups II and III began to hire a few women from groups II and III as well as I, IV and V; women were hired by doctoral institutions at a rate slightly (very slightly) higher than that at which they received PhD's; and jobs were becoming increasingly scarce. The result was widespread claims that "women are getting all the jobs". When, for perhaps the first time in history, women finally began to get their fair share (any excess being well within the limits of statistical fluctuations), some men began to cry "foul".

After the article appeared, I received a lengthy e-mail message from a male mathematician who, after quoting the last sentence of this statement, said

Correct me if I am wrong, but I sense a feeling of, "So now you guys feel what we women have had to put up with for so long. I am not impressed with your complaint since I have seen a longer history of injustice."

I don't want to justify or ignore the injustice you have probably experienced, I only want to address the best way to correct it. Suppose a department store has a long history of overcharging women customers since that is what the male management wanted to do. But finally women and fair minded men rise to positions of power. Should they now overcharge men customers for a while, in order to balance the scales? While that may create better balance among victims, it does not restore the earlier female victims. It only creates more new victims of another type. Far better to strive for fair practices for all customers.

His analogy demonstrates the depth of his misunderstanding in interpreting "fair share" as "more than our share". A better analogy would be that the men, who as a result of price equalization now paid more than previously, began to complain about the "unfair" higher prices. An analogy that describes the current job situation would be that overall economic conditions changed so that both men and women now payed the same high prices that only women paid previously.

The data on first hires suggests that most of the substantial pool of women from the 80's did not not receive a first position conducive to research development. This may be why so few of them are now showing up among the tenured doctoral faculty. For example, Susan Landau (Mar-Apr 95 AWM Newsletter) tracked MIT PhD's from the period 1980-84. She found only two of fourteen women (i.e. 1/7) tenured at doctoral institutions (one in Group I and one in group III) although almost half the men had achieved this status, and one could hardly claim that women who received PhD's from MIT were not of high caliber.

Suppose we accept as a given that women who received PhD's before 1991 were discriminated against in the sense of not receiving equal treatment in their first position. Should we now give them preferential treatment even if it means lower standards? I feel that giving capable women who are not at top institutions research opportunities, such as an invitation to visit to one of the math research institutes or a targeted visiting professorship, is fully justified. However, hiring tenure-track faculty is another matter. What is needed is not differential standards, but a careful look at how we evaluate people and a recognition that there are many good women to be found struggling to do research at the less prestigious departments. Too often, departments confine their searches to the top institutions and react to promising candidates from elsewhere with " ... if she's really that good why is she at Southwest Mediocre State University". Changing that attitude to one of "if she can do that quality work at Mediocre State, just think what she might achieve here at Pompous Research University" would greatly enlarge the pool without changing genuine standards of faculty quality. There is a high correlation between faculty productivity and institutional quality. What is unclear is whether this is cause or effect. There may be a few Ramanujan-like geniuses who would do good work anywhere. But there is a much larger pool of promising people whose career development will depend upon the type of opportunities they receive.

Kirby has suggested, using a greatly oversimplified model, that implementation of affirmative action would result in distributing women so that most were among the weakest members of their departments. I disagree. The uncertainty in our ability to evaluate people and predict their future development is simply too great. Moreover, the type of department into which junior people are hired and the way they are treated by senior faculty can affect whether they flourish or stagnate.

I do not believe that we should hire women who are "less qualified" than the men in a department. But I think we need to examine what that means and just how we decide that "A" is better than "B". Do we count publications? Citations? Amount of research grants? Number of invitations to speak? There are administrators who have proposed (and even implemented) a point system for evaluating faculty (e.g., 10 points for publishing in a prestigious journals, 8 points for a lower-ranked refereed journal, 5 points for a conference proceeding, 2 points for an invited talk, etc.) Such a system would give a mechanism for producing a definitive ranking, but would it be meaningful? Shouldn't we consider the quality of a paper and significance of the results? An evaluation process that attempts to assess quality as well as quantity is necessarily subjective. The challenge is to be subjective without being unfairly, or even inadvertently, biased.

This may be particularly problematic in the group II and III departments where the mathematicians being evaluated are good, but not great. Candidates (whether for a new position or for promotion) are not going to be superstars, and detractors will always be able to find something to criticize. Colleagues must then assess, not just whether the criticism is legitimate, but whether the flaw is of sufficient importance to affect the recommendation and whether scrutiny of other candidates would also reveal deficiencies. This may not be easy and retrospective reflection on the cases when marginal candidates were given the benefit of the doubt as compared to those treated more harshly may be insightful and reveal hidden patterns of discrimination that are not evident in individual cases. The right to base decisions on subjective academic judgments is an important and valuable one which carries with it heavy responsibilities.

Next: 4. Conclusion Up: Thoughts on Affirmative Action Previous: 2. Education

AWM Newsletter.
Copyright ©1996 by Mary Beth Ruskai in form and content. Reprinted with permission
Brought to you by the Association for Women in Mathematics.
Comments: awm-webmaster@awm-math.org