Does Science Promote Women?

Posted by Geoff Davis at 06PM on 11/20/06 | Categories: Women in Science | 6 comments

A few months back I tracked down some statistics on gender ratios in science and engineering departments . I knew things were skewed, but even so, I found the data surprising: men outnumber women in tenured and tenure-track S&E positions by more than 3 to 1. There has been a good deal of recent effort in figuring out why, and we'll be writing more about the subject over the next few months.

Take a look at the stats disaggregated by academic rank:

RankMaleFemale
All ranks77%23%
Full professor86%14%
Associate professor74%26%
Assistant professor64%36%

Is the skew in the top ranks simply the result of gender skew in the PhD pool decades ago? Or are women less likely to be hired into tenure track positions and promoted to tenure as a result of discrimination?

A new paper, "Does Science Promote Women? Evidence from Academia 1973-2001" by Donna Ginther and Shulamit Kahn provides some interesting answers.

Ginther and Kahn use data from the NSF's Survey of Doctorate Recipients (the SDR) to estimate the chances of obtaining a tenure track position and of subsequent promotions as a function of sex and other explanatory variables. The SDR is a huge, longitudinal survey that covers about 10% of all US PhD holders, and it provides a great resource for addressing these kinds of questions.

Their findings, in a nutshell:

  1. Overall, women are less likely to obtain tenure track positions in the sciences
  2. However, the gap is entirely explained (in a statistical sense) by marriage and children.
  3. Furthermore, there is little sex-related difference in the likelihood promotion to tenure or full professor.

Some key findings:

  • For women, having a child under age 5 lowers the probability of a tenure track appointment by 8%. Children in grade school have no effect.

  • In contrast, for men, children under age 5 have no effect, while each child in grade school increases the probability of a tenure-track job by 3%.

These results are consistent with Mason and Goulden's work. A few possible (non-mutually-exclusive) interpretations:

  1. The probability of a tenure-track position increases (up to a point) with time past PhD. Time off for maternity leave shifts the clock back for women. Perhaps the gap would be smaller if one compared the probability of tenure at 5 years for men with young children to, say, the probability at 5.5 years for women with young children.

  2. Several studies have documented that women scientists with young children spend fewer hours in the lab than their male counterparts, presumably because they bear the brunt of childcare responsibilities. Perhaps this results in fewer publications and consequently reduced competitiveness for tenure track positions. Interestingly, Anne Preston has found that "those male scientists who do childcare have similar impacts on their academic careers".

  3. There is often an implicit assumption that an academic position is the only desirable outcome for a PhD scientist. Perhaps women with children see that academia (particularly postdoc-docm) is not a particularly hospitable place and seek greener pastures in industry. Anne Preston has found that "a major reason that women leave science is because of childcare responsibilities."

A couple of other less central but still interesting results:

  • On average, a single childless woman is substantially more likely to get a tenure track jobs than a single childless man. The gap is 21% in the physical sciences, 11% in the life sciences, and negligible in engineering.

  • Marriage increases the probability of a tenure track job for both men and women, but it helps men much more than it helps women (22% increased probability for men vs. 5% for women). On the balance, men and women who are married and childless have about the same chance of obtaining a tenure track position.

The authors posit that these differences are the result of "selection effects", i.e. the populations of unmarried men and married men are differ in ways other than just marriage. They suggest that "'good catches' in the marriage market are correlated with 'good catches' in the labor market." They don't speculate on possible reasons for the gender asymmetry; I'd guess that it's partly due to the fact that that men are more likely to have stay-at-home spouses than women.

Marginal Revolution categorizes the paper as the "politically incorrect paper of the month". My take, however, is more in line with that of Matthew Yglesias: at least part of the reason for gender differences in the sciences is that academic careers are structured in a way that makes things particularly difficult for women who have young children early in their careers. On the plus side, once women achieve tenure-track positions, they appear to be just as likely as men to be promoted to tenure and to full professor (though, as Ginther has documented in prior papers, their pay lags that of their male counterparts).

The good news is that a few institutions are starting to rethink the way they handle faculty family issues and are making improvements. I'll talk about some innovations in a future post.