« Microfoundations of slow European growth | Main | Markets in everything, principles of arbitrage edition »
Politically incorrect paper of the month
Many studies have shown that women are under-represented in tenured ranks in the sciences. We evaluate whether gender differences in the likelihood of obtaining a tenure track job, promotion to tenure, and promotion to full professor explain these facts using the 1973-2001 Survey of Doctorate Recipients. We find that women are less likely to take tenure track positions in science, but the gender gap is entirely explained by fertility decisions. We find that in science overall, there is no gender difference in promotion to tenure or full professor after controlling for demographic, family, employer and productivity covariates and that in many cases, there is no gender difference in promotion to tenure or full professor even without controlling for covariates. However, family characteristics have different impacts on women's and men's promotion probabilities. Single women do better at each stage than single men, although this might be due to selection. Children make it less likely that women in science will advance up the academic job ladder beyond their early post-doctorate years, while both marriage and children increase men's likelihood of advancing.
Here is the NBER version, here is a non-gated version. Alas, I have not had time to read this piece, although I know and respect the work of Shu Kahn, one of the authors.
Addendum: Matt Yglesias comments.
Posted by Tyler Cowen on November 16, 2006 at 06:20 AM in Education, Science | Permalink
Comments
So, does this mean Larry Summers was right?
Posted by: EclectEcon at Nov 16, 2006 10:04:46 AM
Matt Yglesias's comment strikes me as apparently unfalsifiable cant that merely incorporates the latest results into the same creaky old framework a la epicycles. One wonders what it would take to falsify the claim that the difference is due to "structural sexism". If it turned out that women had brains the size of a pea and were as intelligent as squirrels then one might see the following argument from Matt Yglesias:
Here, as in much of life, women and men are now allowed to compete on "equal" terms. The terms, however, were set up long ago -- by men -- before that was the case, operating under the implicit assumption that the competitors would be men, who have big brains and are intelligent. So things work out nicely if you're a man, or if you're a rare woman with a brain the size of a man's, but not so well if you're a woman who -- like most women -- has a brain the size of a pea.
Posted by: Constant at Nov 16, 2006 10:12:11 AM
Answering EclectEcon: actually it doesn't seem to have much to do with Larry Summer's hypothesis. His hypothesis involves variability of intelligence. But this result here involves the choice to have children. It might in fact be that these researchers are, by invoking children to explain a difference, ruling out intelligence as a factor. But I'd have to read the paper to know for sure.
Posted by: Constant at Nov 16, 2006 10:20:17 AM
Since a substantial subset show no earnings differential even when uncontrolled for productivity differences, it suggests that they might even be overpaid.
So if using all their variables and leaving out intelligence is sufficient to eliminate the measured differential then invoking intelligence is not needed to make the case.
Posted by: ghost at Nov 16, 2006 10:27:30 AM
Actually, Constant, Summers had four different hypotheses, and he considered that natural differences hypothesis to be the least likely to be true. Summers also included the differential burdens of child care as another hypothesis, and considered that more likely. Looks like he was right.
Summers' only mistake, IMHO, was apologizing.
Posted by: Keith at Nov 16, 2006 10:36:54 AM
Keith - thanks, I didn't know, yet again falling victim to the distortions of the media. Then - wow, political correctness really is strong among a certain crowd, considering how Summers was vilified for something he mentioned only as an unlikely hypothesis.
Posted by: Constant at Nov 16, 2006 10:51:12 AM
Summers was also lionized in some quarters for something "he mentioned only as an unlikely hypothesis." Interesting that.
Posted by: Barbar at Nov 16, 2006 11:06:11 AM
Hmm, interesting piece. So single women outperform single men early on, but having children is detrimental to promotion for mothers while increasing the likelihood that fathers will be promoted.
I'm having a hard time seeing how this piece is "politically incorrect;" it confirms what everyone already knows, that the academic system of promotion is terribly biased against motherhood (and therefore not unlike upper management in the corportate world, although I'd guess that fewer women in the corporate world are *explicitly told* by their supervisors to avoid having children, if only because of the possibility of lawsuits).
Contrast, Summers was taken to task for two things: first, bringing up fairly loopy intelligence theories in the same breath as structural inequalities such as the one highlighted in this paper. One doesn't need to be president of Harvard U to see that the effects of bias against (for instance) motherhood are so overwhelming as to render tenuous (at best )data on differences in the tails of the intelligence distribution moot. One need only be sentient, and spend some time in the academic (or corporate) world. Second, he presented the facts on the ground as if they were value-neutral; basically he said "academic life is tough, especially if you're a woman and want to have kids. A lot of women want to have kids, so no wonder they're not getting promoted. Oh well.", and left it at that. People were justifiably outraged that the president of Harvard (!) didn't have something more enlightened to say about structural inequities than "oh well."
Posted by: rb at Nov 16, 2006 11:14:41 AM
I somewhat suprised that so many here think that asking women but not men to remain childless to succeed in science is reasonable. It is true now, and has been true throughout history that women to compete professionaly with men need to have a higher level of commitment and dedication. This study does not really offer any new insight but does show along with the comments here that not that much has changed.
Posted by: joan at Nov 16, 2006 11:35:55 AM
I recall a "study" years ago of top scientists and engineers. The goal was to determine what ages they were when they did their best work.
The study backtracked work that was recognized as determined by their having been awarded top prizes such as the Nobel Prize etc.
I think physicists had a peak at 26, chemists 28 and electrical engineers at 35-36.
Since tenure decisions are partly based upon the research scientists DO if women are at home for the peak years that could contribute to under-representation later.
Posted by: SciGuy at Nov 16, 2006 11:39:28 AM
Summers had four different hypotheses, and he considered that natural differences hypothesis to be the least likely to be true.
Comments at Yglesias's point out that this is false.
Posted by: Matt Weiner at Nov 16, 2006 11:46:33 AM
I somewhat suprised that so many here think that asking women but not men to remain childless to succeed in science is reasonable.
And who says that anyone is asking women to remain childless? All the paper is pointing out is correlations. That doesn't mean that there was somebody in some office who asked a woman to remain childless and who passed her over when he discovered she had a child. It could mean, instead, that on average, women who had children spent less time in their professional field than did men who had children.
Posted by: Constant at Nov 16, 2006 11:57:14 AM
Constant,
Your comments are entirely sound:
"And who says that anyone is asking women to remain childless? All the paper is pointing out is correlations. That doesn't mean that there was somebody in some office who asked a woman to remain childless and who passed her over when he discovered she had a child. It could mean, instead, that on average, women who had children spent less time in their professional field than did men who had children."
And they also need to be carefully digested by "rb" who also completely misunderstands the meaning of the statistical results by saying:
"I'm having a hard time seeing how this piece is "politically incorrect;" it confirms what everyone already knows, that the academic system of promotion is terribly biased against motherhood (and therefore not unlike upper management in the corportate world, although I'd guess that fewer women in the corporate world are *explicitly told* by their supervisors to avoid having children, if only because of the possibility of lawsuits)."
As Milton Friedman (1912-2006) was fond of saying:
"Nothing could be further from the truth!"
Posted by: jim at Nov 16, 2006 1:38:24 PM
Seems to me to be more of a marriage partner problem, not an employer caused problem. You have babies and continue work and research and are productive at it, you do well. You have babies and stop work or research, or you continue working but your productivity goes way down, you don't do well.
The former describes men who have babies, the latter women who have babies. How in blazes is this a problem with how employers set things up a long time ago? It is an issue with how parents choose to divide child rearing responsibilities. It is nobody else's business.
Posted by: happyjuggler0 at Nov 16, 2006 1:46:22 PM
"How in blazes is this a problem with how employers set things up a long time ago?"
Um, read the addendum: http://www.matthewyglesias.com/archives/2006/11/incorrect/
Posted by: Barry at Nov 16, 2006 2:31:08 PM
Barry,
I did read the addedum, that is why I phrased it that way. Matt Yglesias is implying that it is the fault of employers that far more families with children set up as "Man works, Woman rears child" instead of Woman works, Man rears child. I call BS, and reiterate that how families choose to structure their relationships is none of the employers business.
Having a baby and working is not a problem. Having a baby and spending time and energy on the baby that one needs to spend on research instead is not some vast conspiracy against women with babies, it could easily apply against men if their family was structured as the man spending all his time and energy on the kid instead of helping his career.
Or is it a case of it being unfair that one needs to produce actual work in order to succeed?
Posted by: happyjuggler0 at Nov 16, 2006 2:51:40 PM
>Matt Yglesias is implying that it is the fault of employers that far more families with children set up as "Man works, Woman rears child" instead of Woman works, Man rears child. I call BS, and reiterate that how families choose to structure their relationships is none of the employers business.
This article gives a fuller discussion of what Yglesias means. It doesn't imply that it is the fault of the employers.
Posted by: Joe O at Nov 16, 2006 3:34:07 PM
I recall a "study" years ago of top scientists and engineers. The goal was to determine what ages they were when they did their best work.
The study backtracked work that was recognized as determined by their having been awarded top prizes such as the Nobel Prize etc.
I think physicists had a peak at 26, chemists 28 and electrical engineers at 35-36.
Not to get away from the main point of the discussion or anything, but it strikes me as beyond absurd that people in two important fields of learning reach their peaks at such young ages. Many athletes don't reach their peaks that early!
Posted by: Peter at Nov 16, 2006 3:50:23 PM
Speaking as an evolutionary psychologist, the discussion seems to neglect the fact that - like every animal on the planet - male and female humans are psychologically different.
We should not be expecting men and women to exhibit the same behaviour: and of course they don't. And the behaviors differ in biologically-predictable ways. Why be surprised or appalled at this? Sex differences are not due to sexism, they are fundamental. Sex differences in pay (in markets) are explicable on the basis of sex differences in temperament interacting with economics.
Most people take this for granted. Only a certain type of intellectual affects to find male-female differences evidence of discrimination.
The topic of sex differentials in pay is pretty well sorted-out, IMHO, by various applications of Gary Becker's reasoning. This is made accessible in an excellent recent book called Why Men Earn More by Warren Farrell. For an academic treatment see work by Kingsley R Browne of Wayne State University Law School.
Posted by: Bruce G Charlton at Nov 16, 2006 3:56:22 PM
I have to wonder if the marriage advantage for men is correlation, rather than causation. One hypothesis: socially inept men are both less likely to get married and to make the social connections among fellow faculty that make tenure more likely.
"Not to get away from the main point of the discussion or anything, but it strikes me as beyond absurd that people in two important fields of learning reach their peaks at such young ages. Many athletes don't reach their peaks that early!"
In what sport do athletes peak later than 35-36? Every sport I can think of has an athletic peak in its 20s. In women's tennis and gymnastics, it may be even younger.
Posted by: Ted at Nov 16, 2006 3:56:30 PM
Actually, this paper proves a theory on the major difference between men and women.
Women generally prefer to marry (or just mate with) men more successful than themselves. If men would simply give up their careers to raise children, this wouldn't be a problem.
Posted by: Xmas at Nov 16, 2006 6:42:26 PM
Joe O,
Thanks for the link. I also reread Iglesias' post yet again for the third time and decided that it was ambiguous enough to let it mean whatever the reader wanted it to mean, kind of like Bill Clinton's campaign utterances weren't actually promises but sounded that way to his listeners.
I honestly though don't see how he thinks university administrations can help to mitigate the situation, unless it was his intention to mean that such administrations make your linked article mandatory reading by students and possibly faculty.
For clarity, I agree with virtually everything that Linda Hirshman wrote, but she did it far more gracefully and edifyingly than my pugilistic post managed to do.
Posted by: happyjuggler0 at Nov 16, 2006 7:08:11 PM
Not to get away from the main point of the discussion or anything, but it strikes me as beyond absurd that people in two important fields of learning reach their peaks at such young ages. Many athletes don't reach their peaks that early!
In what sport do athletes peak later than 35-36? Every sport I can think of has an athletic peak in its 20s. In women's tennis and gymnastics, it may be even younger.
I was referring to physics and chemistry, with peaks at 26 and 28 respectively. People who are into baseball statistics - and there is no sport remotely as statistically analyzed as baseball - say that hitters peak at 27, pitchers a year or two later. Basketball players seem to peak around the same age.
Admittedly, there probably are no sports in which athletes peak at the same time as electrical engineers (35-36).
Posted by: Peter at Nov 16, 2006 10:40:30 PM
Chess players peak at 36, according to a book I read in the 1980s. Golfers probably peak in their early 30s, and it's not unusual for some to peak substantially later -- Ben Hogan and Vijay Singh had their best years (1953 and 2004, respectively) in their early 40s.
Posted by: Steve Sailer at Nov 16, 2006 10:56:15 PM
Children are consumption goods. On average, women are more into them than men (and, for my point, it does not matter if this is biological or socially constructed or both). I suspect that men with boats or other time-intensive hobbies do relatively poorly as well. Why should policy favor one consumption good - children - over others - boats?
Put differently, Matthew Y is correct that the process is "biased" against people who spent a lot of time on something other than research. Seems like that is a good thing.
Jeff
Posted by: Jeff Smith at Nov 17, 2006 8:19:47 AM
Surely childeren are consumption good. Why would anyone argue against it. You just walk in a store and buy a boy or a girl (you can even specify the age you want). I think they sell it in the same store with boats! Continueing the same line of logic men can purchase this a good as well as woman. So men should start purchasing babies and woman will enjoy their boats.
Posted by: vg at Nov 17, 2006 12:27:07 PM
Jeff,
I hope you didn't share your comment with your wife. Your marriage may not last very long. :-)
But more seriously, I thought you shared Heckman's view that early investments in children are important for their developnment. If so, then society has a strong reason for preferring "consumption" of children to consumption of boats. If women have a comparative advantage in producing children's human capital, then they will (under reasonable assumptions) spend more time than men do in childcare. If they do not derive more consumption benefit from childcare, then it is not unreasonable to argue that having them bear the cost of childcare represents a form of discrimination (and, yes, I understand that within family transfers might mean that they do not actually bear this cost, but I don't take it as a matter of religion that they do not).
More importantly, I think you miss the import of Matthew Y's point. Institutions are not immutable and there is little reason to believe that the institutions in academia are socially efficient. Unless you buy the argument rhat "if it isn't efficient, it wouldn't exist" (the most bastardized version of structural functionalism), you will have a hard time making the argument that the tenure clock for almost every academic field in almost every university should be six or seven years.
If institutions are neither immutable nor efficient, then we must admit the possibility that there are equally good institutions that would affect women with children less adversely. Or there may be institutions that would be less good for some groups but would improve the position of married women with children. The issue is similar to the usual question in discrimination cases as to whether some practice has disparate impact on men and women (or blacks and whites) and, if so, whether the practice is a business necessity, has a business justification or neither. I'm not saying that such institutions exist, but I don't take it as obvious that they do not.
Incidentally, as an economist you are on particularly weak grounds with your argument since economics is one of the few fields in which women without children do worse than men without children. In other words, the summary of the Ginther/Kahn paper does not apply to economics, leading to the supposition that economists may be particularly desperate to not see discrimination where it exists.
Posted by: not-chicago at Nov 21, 2006 8:55:15 PM
