Tuesday, May 12, 2009

Works Well With Others

Recently I perused a list of graduate degrees awarded in the 2008-09 academic year at my university. The list is incomplete because it only contains awarded degrees; there are also quite a few degrees that are pending but will soon be awarded at the official conclusion of the academic year.

Even so, this list, though incomplete, is fascinating.

The list contains the names of advisors. At first glance I had the impression that more science/engineering PhD students were co-advised (as compared to having one advisor) than were students in the humanities and social sciences, but once I added up the numbers and calculated the %s, I could see that my impression was unduly influenced by the fact that there are many more PhDs in science/engineering than in the other fields. In fact, I was surprised at how few PhDs were awarded in the humanities, but this may in part be a function of the incomplete nature of the list.

I went through the list twice because the first time I only kept track of the numbers of co-advised vs. single-advisor students, but as I did this I started to get the impression that women faculty made up a high percentage of the co-advisors. I didn't trust my impression so I went through the list again and kept track of the gender of the advisor(s).

The % of women advisors of PhD students matched almost exactly the % of women faculty at the university in each of the subdivisions I tracked: science/engineering, humanities, and social sciences. Therefore, although the list was incomplete, I think it is has enough data to provide some potentially interesting insights into advising trends.

Here is what I found: Although 10% of single-advisor science/engineering PhD students had women faculty as advisors, ~40% of the co-advised students had one woman advisor; I don't think any were co-advised by 2 women faculty.

Although the total number of PhDs is less in the humanities and the social sciences, the co-advised students in these fields were twice as likely to have a female advisor as were the single-advised students in each field. And, in contrast to science/engineering, there were examples of students co-advised by two women faculty.

I have no idea what this means (if anything) about women or men as advisors or about student preferences for advisors. To the extent that these data mean anything, possibilities include: Women tend to collaborate more? Women have a greater tendency to lack the confidence to be the sole advisor of PhD students? Women don't tend to have their ego tied up in whether they are sole or co-advisor of their PhD student but men have a greater tendency to want to be the sole advisor? Women seek ways to maximize their time by sharing responsibilities with others? If a student has one female advisor, they also want another advisor? None of the above? Some fraction of some of the above? Did I miss some obvious possibilities?

Some of my PhD students are co-advised and some aren't. The decision is always based entirely on the research interests of the students, so I never looked at the issue with my Genders Lenses on before.

It would be interesting if there were a rigorous study of a larger database on the issue of graduate student co-advising and gender. Does such a thing exist?

27 comments:

  1. One possible explanation, which hardly counts for the whole difference, is that women who get pregnant, or plan to get pregnant, while advising a student, might involve a second advisor for the sake of the student?

    ReplyDelete
  2. Complicating factor in studies of co-advisors: If they are in different departments, one of them might be a co-advisor in name only (usually the one from the department that the student is enrolled in) and the other might be the person whose lab the work is actually done in.

    I've seen both sides of this. My advisor was in a different department, since I was doing interdisciplinary work, and so my dissertation signature page lists two co-chairs for the committee. In reality, only one of them played a major role in the work. I'm now advising an M.S. student from another department, but somebody from that student's home department will be listed as committee co-chair.

    I've heard of studies showing that women are more common in interdisciplinary fields, so that might be the main factor here. If you're a student from department A looking for an advisor in department B, the person in that department that you're most likely to work with is going to be somebody who does interdisciplinary work, and that population skews female relative to the rest of the academic community.

    ReplyDelete
  3. With a little more information about the data, one could ask (and perhaps answer) many interesting questions. I sense a good project for the statistics class...

    ReplyDelete
  4. Is there a correlation between gender of the student and gender of the advisors or co-advisors? Can another possibility be that a (female) student has a male supervisor, and would also like to have a female (co-)mentor?

    ReplyDelete
  5. That is really interesting! I know at our university, a PI isn't allowed to be the sole supervisor of a PhD student until they have a masters student successfully defend. If that's the case at your university as well, then perhaps that is also a factor?

    I never realized so many students had co-supervisors! It'd be interesting to see if it's prevalent at other universities as well.

    ReplyDelete
  6. I'd be interested in seeing how many of those co-advisers were husband/wife teams who run a research group together. I have worked in several labs of that sort and they tend to co-advise their students (although one professor generally takes the lead role, depending on his or her research specialty and how it relates to the student's work).

    ReplyDelete
  7. You have to be careful to take into account the underlying probabilities. Suppose 10% of the faculty are women. Then in the absence of any bias in choosing advisors, 10% of single-advisor students will have a female advisor, and 1 - (1-0.1)^2=19% of the co-advised students will have at least one female advisor, and only 0.1^2=1% will have two female advisors. So the approximate doubling of the proportions is to be expected, as long as the proportion of women is not too high. If 50% of the faculty are women, 75% of co-advised students will have at least one female advisor. Of course, a 10% vs 40% might be more than chance.

    By the way, what do you think is the reliability of being able to tell one's gender based on his/her name, especially among university faculty where there are lots of foreigners? Not that it would bias the results, unless the co-advisors are more likely to have "funny" names.

    Aniko (who is often thought to be a man based on her name)

    ReplyDelete
  8. Another interesting aspect to consider is whether the female co-advisors are less likely to have tenure. If that is the case, they may seek co-advisors who have tenure in the event that they don't receive it themselves in an effort to protect the student.

    ReplyDelete
  9. Were the women more junior than the men faculty, in general? That could account for some of the difference.

    ReplyDelete
  10. Many women students may feel that they would benefit from having a female co-adviser. For example, if anything bad (as in harrassment bad) happened with their male adviser, they would have a woman in authority to turn to. Can you look at your lists and see if the students who are co-advised by a woman + man team are more often women? That would be my prediction if my hypothesis is true.

    For the same reason I made a point of choosing women on my committee (obviously they are also highly qualified to be on my committee and have contributed a lot), although I have only a single male adviser.

    ReplyDelete
  11. I'm coadvised by an untenured female professor and a senior, tenured male professor. I initiated the arrangement because a former boss (also co-advised by this male professor) spoke favorably about being exposed to multiple professors and multiple groups.

    As a data point, though, the more junior of my advisors is fine with the arrangement but would prefer to advise alone, so that her contributions in advising and graduating students are distinct when her tenure case is considered.

    ReplyDelete
  12. Interesting post. I have co-advisors who are married to each other, but I am the only student in our labs who has this (though most of the other students have both on their committee). The factor leading to this arrangement for me was a teaching assistanceship. As graduate student enrollment rises, our department limits the number of TA's in a PI's lab to ensure continued support for students throughout their degree. I wanted to teach and the co-advisory allowed that. I wonder what other extraneous factors could be influencing co-advisorship of graduate students?

    ReplyDelete
  13. I have co-advised two PhD students. One was doing a co-major, so he had an advisor from each major.

    The other co-advising was because the lead researcher on the proposal that kicked off the project demanded that one of other faculty members co-advise the student on my component of the project, and I am certain it's because I am a woman.

    ReplyDelete
  14. A few years back I had a PhD student from Asia. I was the primary advisor for this student and published numerous papers with him. Six months prior to his defense he informed me that he was going to add a co-chair advisor to his committee. The faculty was a senior male professor and had very little previous input to the student's research (plus in general has published only a few papers in the area of specialty that the PhD was in compared to myself). I assumed that this had something to do with the student seeking a faculty position back in Asia (which he is currently in) and a male advisor would look just a little bit more respectable; when I asked about this I never got a satisfactory answer. I still collaborate and have a very strong professional relationship with this past student I assume the addition of the co-chair has nothing to do with my advising abilities...not to mention the same year my students nominated me for an advising award.

    I think the saddest part of the story is that under the rules of my institution only one chair can be listed and without my knowledge the dean (male) of my college listed the male faculty as the main chair of the committee on the final paper work. It all got straightened out in the end, but I was quite appalled at the dean's assumption.

    Maybe students feel that if they have a female advisor their career benefits more from including a male faculty (even one that is less published then the female). I wander if this is related with particular cultures? Male students? My discipline? Just my two cents...

    ReplyDelete
  15. I was co-advised by two tenured male professors. In my case, joint funding was the main reason for the arrangement. Since women are (generally) less generously funded, or PIs in funding proposals, sharing a student may be more necessity than choice.

    ReplyDelete
  16. In my institution there are no "co-advisors" in the biomedical sciences. It is thus much simpler to look at gender differences in thesis advising. Looking at graduate student selection of thesis advisors among those potential advisors who were junior faculty, I discovered that the male assistant professors averaged twice as many grad students as the female assistant professors. I think this is more likely to happen with junior faculty than senior faculty who have more solid credentials. Does anyone else have information on the subject of gender bias in students' choice of thesis advisors?

    ReplyDelete
  17. I actually find these comments (and assumptions therein) to be at least as interesting as your data. Women PIs are less funded? Women PIs more interdisciplinary? Coadvisors more likely to be married? Female students like female advisors? Female advisors more likely to be pregnant? Are there any data for any of this? (Ok, I don't need data for that last one, though I sure don't see pregnant faculty in science departments very often, and my experience has been that pregnancy lasts a lot less time than the average Ph.D.).

    I met a student this morning who is being coadvised by two faculty who despise each other. I wonder how often that happens, and if it ever ends well....

    ReplyDelete
  18. Are the co-advisors who despise each other being selfless and putting aside their hostility for the sake of the student? I hope so, but it would be difficult to do for years on end. I have served on committees with people who despise each other, and that is difficult enough for the student (and other faculty).

    ReplyDelete
  19. The funding issue, at least, has been studied. Check out "Beyond Bias and Barriers" (The National Academies Press, 2007).

    And I can say, as a female graduate student, the sex of my advisor factored into my decision at least somewhat. She does awesome science, and that was my most important deciding factor. However, I was also excited at the chance to work for a successful female scientist. You don't run into that very often in my field.

    ReplyDelete
  20. Can another possibility be that a (female) student has a male supervisor, and would also like to have a female (co-)mentor?That is part of the reason for the co-advising situation I am in (I am a female student). It also has to do with research interests (both advisors are a good match, but in different areas) and the fact that the female advisor is not tenured and thus couldn't be the main advisor on her own.

    But I am very pleased about the situation. If I had only one advisor, I would want it to be a woman. It would be irritating to have to balance this with finding the right match in research interests, so I am very glad that I can have two.

    ReplyDelete
  21. Women PIs more interdisciplinary? .... Are there any data for any of this?The statement that women are more likely to work in interdisciplinary fields is often repeated. I don't know the source data, but I've seen it in multiple articles (which doesn't mean much). I don't know whether that reflects an age effect (there are more female professors than there used to be, so many of them are young, and young professors are more likely to work on interdisciplinary problems than their senior colleagues). If it's just an age effect, then comparing people who got their Ph.D.'s around the same time would not show a gender difference with regard to interdisciplinary work.

    ReplyDelete
  22. I agree- these comments are fascinating! When I first read this post, my first thought was, "well, of course the reason must be because of the best interest of the research project." Clearly there are often other contributing situations/motivations.

    ReplyDelete
  23. As a male ABD in the sciences, I would say that the female professors can often be more approachable and/or have better overall social skills. Students may pick a female advisor more often simply by not picking the most asperberger-ish (more likely) male professors in their department.

    ReplyDelete
  24. In fact, I was surprised at how few PhDs were awarded in the humanities, but this may in part be a function of the incomplete nature of the list.I'm not sure if anyone has mentioned this or not but ... IIRC, most Ph.D. candidates in the humanities have to pay their way, whereas Ph.D. candidates in the sciences can find funding for their work. That may be the reason you'll find less Ph.D. candidates in the humanities.

    ReplyDelete
  25. I met a student this morning who is being coadvised by two faculty who despise each other. I wonder how often that happens, and if it ever ends well....Avoiding this situation was a major factor in how I put together my thesis committee. One particular junior faculty member of my department has enormous expertise relevant to my project, but also is not very good at departmental politics and has a history of rubbing people the wrong way, including the department chair and my advisor. He has become rather bitter and it tends to rub off on everyone around him, so that his entire group is infused with a kind of myopic cynicism. He is not on my committee.

    If there were a woman with relevant research interests anywhere in the two institutions I am affiliated with, I would have been very interested in having her as a committee member, in order to have a female mentor, but there are exactly zero, so my committee is composed entirely of men. This is fine, since they are all excited about my research and can help me in various ways. Still, I hope eventually to also find a female mentor through collaborations or maybe as a postdoc advisor.

    So that's one more data point.

    ReplyDelete
  26. This is a very interesting topic. The institution where I did my PhD now requires that you have at least 2 advisors.

    I think the original thinking may have been that if you have trouble with one of them, you can go the other for help. Also sometimes one of the advisors is just so busy that you need the extra to get by! I knew of some cases where one advisor was swapped out for another as well.

    As an aside I was supervised by first one female (principal) advisor and one male advisor, then by three males after my principal went to industry. She needed two people to replace her! I think it was mainly for paperwork issues.

    ReplyDelete
  27. Although the findings indicated that "there are many more PhDs in science/engineering than in the other fields," findings did not indicate the % of female PhDs in science/engineering. If the ratio to males are much less, then it may be fair to conclude that there is not sufficient females to advise the number of students.

    ReplyDelete