how sociology professors choose their specialties – a guest post by james iveniuk
James Iveniuk is a doctoral candidate in sociology at the University of Chicago. He recently collected data on professors to understand how people choose their research specialty. He collected data on all professors at 97 ranked sociology doctoral programs in the US News & World Report. Click on this link: Iveniuk Discipline Analysis. Lots of fun results. In my view, this report supports the “Prada Bag hypothesis,” which suggests that the areas of cultural, politics, and historical are luxury items more likely to be found at higher ranked programs. Add your own interpretations in the comments.
50+ chapters of grad skool advice goodness: Grad Skool Rulz ($2!!!!)/From Black Power/Party in the Street!!
orgtheory.net 
Odd that Family isn’t one of the areas. How were they selected?
LikeLike
Philip N. Cohen
February 2, 2015 at 2:04 am
I think we should ask James, but it might be the case that it is lumped into demography. Aren’t family and demography considered highly interrelated fields?
LikeLike
fabiorojas
February 2, 2015 at 5:55 am
I notice that Economic Sociology is not on the list – is it no longer considered a legitimate specialisation?
LikeLike
Grad student
February 2, 2015 at 6:52 am
I had a similar reaction as Phil and Grad Student. Apparently, no faculty members in top 97 department list “inequality” as an area, which I find very hard to believe, or the grad student decided that inequality is the same as demography (or gender? or urban?), which it isn’t.
This makes me wonder what other data decisions were made and, less charitably, how many of them benefited the home team, even if unintentionally. It’s pretty easy to do. Step 1: lump together categories where Chicago doesn’t have many faculty, and differentiate the categories where it does. Step 2: create a measure of “best” that factors in the proportion of faculty in the department who study the topic. Step 3: construct a table (2b) that shows that Chicago is #1 in 4 of 5 categories.
Seems like a matrix of overlapping ASA section memberships would have been a more defensible way to construct categories.
LikeLike
krippendorf
February 2, 2015 at 12:02 pm
Hi all,
Sorry for taking a while to jump in here. I should say a few words about how these were chosen. Per Professor Krippendorf’s comments, these were not at all chosen systematically. There are lots of specializations that I did not look at, and the things that are listed here are absolutely not the only common tags. The list of tags that I chose to focus on were basically things that were of interest to me and my immediate peers. I absolutely did not intend to make Chicago look rosy, but that may be a side effect of my being situated at this university, for which I apologize. It was not my intention to offend, nor to stump for my institution. I am more than happy to share the data with anyone interested, if they are concerned about my results.
Further information: there were 289 professors who listed ‘inequality’ as an interest, there were 132 who listed ‘economic sociology’ as an interest, and there were 328 people who listed ‘family’ as an interest. In terms of what predicts membership in those three categories, Individuals who listed ‘inequality’ as an interest were more likely to be female (OR = 1.44, p<.01), more likely to be African-American (OR=1.66, p<.05), and more likely to be assistant professors versus other ranks (OR=1.43, p<.05). Individuals who study family are more likely to be female (OR = 3.03, p<.001), and more likely to be situated in Ohio versus other regions (OR = 2.75, p<.01). Individuals who study economic sociology are less likely to be female (OR = 0.47, p<.001), and more likely to be at highly-ranked departments (OR = 1.88, p<.001). I am perfectly happy to query more subfields throughout the day as people ask me questions, in order to assuage bias in the selection of topics.
More context: I really just did this for fun. This is not at all my area of expertise (see my page at UChicago for my main research interests). I have far, far too much on the go right now to make this a long-term project, but I also didn't want these data and these analyses to go nowhere. As I said, I would be very happy to share the data with anyone who is interested, and I can also share the Stata code that produces these results.
best,
James
LikeLike
James_Iveniuk
February 2, 2015 at 2:46 pm
Also, on the issue of lumping: Someone would actually have to say ‘family’ on their webpage to get the tag. So I didn’t lump ‘family’ with demography unless the person said on their page that they were interested in family.
LikeLike
James_Iveniuk
February 2, 2015 at 2:48 pm
I don’t have anything to say about which specialties were listed and which ones weren’t. But standardizing by the number of faculty seems like a really weird choice. Having more faculty, all else being equal, is a good thing. The only defensible reason for that measure is that it may be a proxy for graduate student cohort size, but if that’s the case, he should have standardized it by cohort size. There are definitely departments which admit small numbers of students but have a lot of faculty, and vice-versa. I think Neal Caren has that data.
In other words, the fact that Wisconsin has 57 faculty (that can’t be right; he must be counting courtesy appointments) shouldn’t matter. But the fact that Wisconsin had an average of 27 students in an incoming cohort over that time period should matter.
Otherwise, I think this is really neat and useful.
LikeLike
Guest
February 2, 2015 at 2:52 pm
To Guest: I would have loved to standardize by cohort size, but I had no idea how I would get those data, short of calling every department. Plus there was the added issue of finding out what the cohort size was during the year that each person started his or her PhD, which was a horrifically daunting task for me, given I was approaching this topic rather casually. But I agree, my solution is inelegant. And the 57 faculty at Wisconsin were indeed perplexing to me. Here they are, if you would like to look:
Brian Christens; Lewis Friedland; Douglas Maynard; Michal Engelman; Jane Collins; Jenna Nobles; Chad Alan Goldberg; Erik Nordheim; Adam Gamoran; Marcy Carlson; Ivan Ermakoff; Alice Goffman; Erik Olin Wright; Daniel Kleinman; Felix Elwert; Alberto Palloni; Joseph Conti; Gary Sandefur; Michael Bell; Noah Feinstein; Robert Freeland; Gary Green; Monica White; John DeLamater; Cecilia Ford; Monica Grant
Karl Shoemaker; Gay Seidman; Jason Fletcher; Sida Liu; John A Logan; Laura Senier; Leann Tigges; Michael Massoglia; Myra Marx Ferree; Geoffrey Borman; Jack Kloppenburg; Joshua Garoon; Joel Rogers
Pamela Oliver; Jess Gilbert; Pamela Herd; Mustafa Emirbayer; Eric Grodsky; Joan Fujimura; Christine Schwartz; James Montgomery; Nora Cate Schaeffer; Joseph Elder; Katherine Curtis; Chaeyoon Lim; Randy Stoecker; Ted Gerber; Michael Thornton; Samer Alatout; James Raymo; Sara Goldrck-Rab
I would especially appreciate some input from someone (grad student or faculty) at Wisconsin to help me to understand why their faculty seems so enormous. It may be that most of these should be dropped.
LikeLike
James_Iveniuk
February 2, 2015 at 3:06 pm
Cool analysis James.
How do you deal with affiliated/courtesy faculty? At some places, like Northwestern and Stanford that have a lot of sociologists scattered across campus, this matters a great deal.
LikeLike
brayden king
February 2, 2015 at 3:36 pm
Dear Dr. King,
I tried to avoid courtesy/affiliated faculty, although as ‘Guest,’mentions above, the website at Wisconsin may have confused me.
I have just looked you up in the dataset at you are not there. This is probably because on the Northwestern site, you are not listed under the ‘faculty’ tab but rather the ‘Affiliated and Joint Appointments’ tab ( http://www.sociology.northwestern.edu/people/faculty.html#joint ). I was never quite clear what these additional tabs meant, and so I excluded those individuals since I could not anticipate how involved they would or would not be in the department. This was shaped by my experience with courtesy faculty at UChicago; some of these affiliated faculty were not at all sociologists and had almost no involvement in the department. Since I was unsure about how to distinguish ‘involved sociologists’ from others, without reading all their CVs, I erred on the conservative side.
best,
James
LikeLike
James_Iveniuk
February 2, 2015 at 3:49 pm
Well, this is fun. Things that jump out at me:
1) Overrepresentation of health and demography among assistant professors. I assume that’s because departments are hiring in fundable areas these days, not because the health and demography folks aren’t getting tenure.
2) Every grad school applicant should be reminded, repeatedly, that top-10 departments only place 2-4 students a year in departments with PhD programs.
3) Unsurprisingly, the gender breakdowns line up pretty well with ASA section memberships: https://familyinequality.files.wordpress.com/2011/06/sociologysections.jpg.
4) A quarter of PhD departments have no demographer? That’s hard to believe. Is that because of places like Berkeley with freestanding demography departments?
5) Woo hoo, SUNY Albany makes the top ten for places to study demography and urban sociology. Give us fifteen minutes and somebody will slap it on the front of a web page!
LikeLike
epopp
February 2, 2015 at 4:07 pm
Hi Epopp,
1) That was my thought on the health finding as well. If you need to get tenure, there is just so much more grant money for health. So I am sure that there is a lot of pressure to make what you do about health/aging/etc.
2) Yes, agreed. Given the figures here, I wouldn’t be surprised if that number drops off precipitously after one is no longer looking at the top 10.
3) That’s reassuring!
4) I’m not sure why that is. Here are the departments that are coded as having no self-identifying demographers:
Miami; Delaware; Kentucky; Arizona; UC Santa Cruz; Colorado State; Arizona State; Teachers College Columbia; SUNY Buffalo; Boston College; Northeastern; UC San Diego; New School; Loyola Chicago; Nebraska Lincoln; Southern Illinois Carbondale; Boston University; Brandeis; Tulane; American University; Purdue; Kent State; SUNY Binghamton; Howard
I may have missed something here, and so I would value the input of the community.
5) Woo? I think?
best,
James
LikeLike
James_Iveniuk
February 2, 2015 at 4:24 pm
We need a name for that thing that is the number of steps between any potentially interesting analysis of academic sociology and the inevitable (usually useless) ranking of departments.
LikeLike
Philip N. Cohen
February 2, 2015 at 4:27 pm
James, it makes sense to err on the conservative side, I suppose, but doing so misses out on the diversity of sociology that exists at many of the country’s top universities (and within our discipline!) That decision may also reflect your own bias about what constitutes sociology. At Chicago, for example, there are a number of great sociologists in the business school, including Ron Burt. Over the years a number of students from the Chicago sociology department have wandered over there and benefited from that relationship. Business schools aren’t the only place where you’ll find sociologists, of course. As others have mentioned, demography departments are sometimes freestanding. Other culprits include public policy schools and criminology.
LikeLike
brayden king
February 2, 2015 at 4:45 pm
Yes I completely agree with you. It hurt me to leave out people like Ron Burt, but once I had formulated the policy I stuck to it, in order to streamline data collection and focus the analysis. If I had more time I would go back and do it in the way you suggest, if only because then I could run the analyses with and without the affiliated faculty and see if anything changed. I hope my decision doesn’t irredeemably distort the findings here.
LikeLike
James_Iveniuk
February 2, 2015 at 5:02 pm
I find it interesting that most discussion is about coding issues (which are obviously important) and less about the picture of the profession that this analysis provides.
LikeLike
fabiorojas
February 2, 2015 at 5:07 pm
Does the analysis provide useful information about the picture of the profession, given the coding issues people have brought up?
LikeLike
Grad Student II
February 2, 2015 at 5:14 pm
GR II: I think so. The demographics by themselves are telling and recoding probably won’t change many answers (high status schools consume culture).
LikeLike
fabiorojas
February 2, 2015 at 5:16 pm
Just echoing what Fabio is saying: Yes, I don’t think the demographics would change very much (e.g. coefficients for ‘gender’ in Table 2), and certain results, like the clustering of ‘health’ in assistant profs, seem revealing. The results for ‘cultural sociology’ may also be interesting here, since it would suggest that this label is clustered among younger scholars and relatively high-ranking schools, even though working on ‘culture’ broadly is not associated with being at a high-ranked school. To me that suggests a movement precipitated from the top, but Fabio may have more thoughts on that topic.
LikeLike
James_Iveniuk
February 2, 2015 at 5:39 pm
I like the project, but +1 on the questions about methodology. Could you explain how people were coded into different fields? For example, as far as I can tell, Philip Cohen (picked because he commented first) doesn’t list any areas of interest on his website or CV. Maryland has six different areas that they group people in on the website, but these (like “Gender Work and Family”) don’t all fit neatly into topics. In contrast, Northwestern groups people into about 35 different categories, with some folks (hi Jeremy!) listed in 8 different slots, but he only lists 5 on his personal website, and the two lists don’t match up cleanly.
I ask because I think there’s possible at least two different things to explain. First, where are the people who are studying certain things, like networks? Second, who calls themselves, or is called by their department, a scholar of networks? Sometimes, departments try to pitch themselves as good in a couple of key areas and group faculty, while other departments seek to show they cover everything, so they highlight variety. Some departments put some thought into this presentation, while others just fill out whatever template the web designer gives them. I suspect all these processes are associated with departmental status.
LikeLiked by 1 person
neal caren
February 2, 2015 at 7:07 pm
Hi Dr. Caren,
I will try to describe everything that I did in detail. I realize now I should have provided a fuller description in the original document, in order to assuage these concerns earlier.
I’m looking at Philip Cohen’s webpage right now, and as you say, he doesn’t have a list of stated interests. However he does provide several paragraphs of text describing his work. For every scholar, I drew upon the following three areas to code their interests: their list of interests on the webpage, their list of interests on their CV, and the blurb that they provide for themselves on their webpage. Not everyone has all three of these things, but I used every bit of information I could in order to code the person’s stated interests. So although he doesn’t have a box that says ‘demography’ next to his name, he says that he publishes in demography journals, and looks at demographic patterns in inequality. Therefore I would have (and did) code him as a having a ‘demography’ interest. I tried to be sensitive to canned categories as I went, since, as you say, some people get tagged with these labels by their departments. So if someone is tagged as ‘race, class, and gender’ but they don’t mention anything about gender, I would not label them with ‘gender’ in the dataset (in fact, there is a separate tag called ‘race&class&gender’ for such scholars). I was particularly sensitive to this issue because, to continue the previous example, having ‘gender’ anywhere on your webpage does not necessarily mean anything. Therefore wherever there was an ‘X and Y’ label, where it was obvious that this was a compound designation, I would not break this up into constituents, but instead join the two with an ampersand, in order to be able to separate them again later if I thought it might be a good idea later. These result do not do that – people who study ‘race, class, and gender’ were not coded as ‘gender.’
Of course I think you are right to distinguish those two things. I was motivated to collect these data basically in order to be able to easily find people who work on some topic, and who I believed had some expertise in that area. So I tried to only label someone as something if, upon reading through their summary, I felt confident that they had done something in that area.
Although, to turn the conversation back towards the results, yes, it absolutely seems plausible that higher-status departments would try to label their scholars as particular high-status types. In this case, the discussion becomes less a conversation about the location of scholars with particular interests, but rather the prestige of the labels themselves. So yes, I’m unable to distinguish those two processes, but it at least seems plausible to me that they would track in the same direction, and that one would not suppress the other. They might upwardly bias the status results, though. Ideally I would be able to compare, for each scholar, his or her self-reported stated interests with his or her web presence.
best,
James
LikeLiked by 1 person
James_Iveniuk
February 2, 2015 at 8:31 pm
Thanks. That’s helpful. This would seem quite a daunting task, and I don’t envy someone trying to figure out the distinction between “sociology of culture” and “cultural sociology” based on someone’s blurb.
LikeLike
neal caren
February 2, 2015 at 8:58 pm
This is not about me, but FYI I think you may be looking at my profile page from the pop center. My department page says “My research and teaching concern the sociology of families, social demography, and social inequality.”
Whatever happened to the areas of interest listed for every faculty member in the ASA Guide to Departments? That seems like a good source.
LikeLiked by 2 people
Philip N. Cohen
February 2, 2015 at 9:01 pm
Whoops. The Maryland page should really signal that the names are links.
LikeLike
neal caren
February 2, 2015 at 9:06 pm
Hi Dr. Cohen,
Sorry about that. In the original sampling I started from each sociology department’s page and then looked at the list of people there, page by page. I wouldn’t have looked at the pop centre page when I coded your areas of interest. That was just what came up first when I googled your name, which was never what I did when I constructed the data set. And I’m not sure about the ASA Guide to Departments. I’m sad to say I haven’t opened one. Perhaps if I ever revisit the topic I could use it.
best,
James
LikeLiked by 1 person
James_Iveniuk
February 2, 2015 at 9:08 pm
James: You can get the incoming cohort sizes at the NRC rankings.
I would estimate about 15 of those faculty at Wisconsin are actually in other departments, and they just have a courtesy appointment in sociology.
Also, I do think it’s a cool analysis.
LikeLike
Guest
February 2, 2015 at 11:51 pm
The reason people keep mentioning methodological issues is because they influence what conclusions you can draw. It’s an interesting analysis but I wouldn’t take it too seriously.
LikeLike
Anon
February 2, 2015 at 11:55 pm
Wading in late, “Wisconsin” graduate sociology is jointly run by two departments (Sociology in the College of Letters and Science, and Community and Environmental Sociology–the former Rural Sociology department–in the College of Agriculture and Life Science whose combined faculty in the 1980s were indeed about 60, but today the two departments combined have 48 full time faculty members. There are also people with part-time appointments and courtesy joint appointments. Ours is the only program in the country in which these two departments run one fully-integrated graduate program. From the point of view of the PhD program it is all one program, so the 48 number is solid. Among the joint appointments, some are actually very involved in graduate training, including a couple of people with 0% appointments in sociology, and others less so. By subtraction,9 of the 57 you count are not full-time in the sociology programs, but I count at least 4 of these who are very active in graduate training, giving 52 as probably the best number.
LikeLike
olderwoman
February 3, 2015 at 3:57 am