dear steven fazzari

Hi, Steve, Fabio here. I recently read about how you are now the chair of the new sociology department at Washington University, St. Louis. It seems that you are getting advice from some excellent sociologists. Still, I wanted to offer a suggestion about how to build your program that I think has some merit  but that may not be obvious.

Here it is: build a program that, roughly speaking, is about 2/3 quantitative and 1/3 qualitative. However, don’t use the traditional criteria for “quantitative research,” which means anyone who does regression analysis or, as in economics, people who do research in theoretical statistics. Instead, the quantitative sector of the department should focus on unique and important quantitative types of data that sociologists are, or can be, good at. Roughly speaking, that means network analysis, social simulations, “big data,” and quantitative analysis of text. You might also toss in the experimenter or survey design guru.

Why? No one else is building such a program, but it would have a huge immediate impact on the profession of sociology. You would have an enormous first mover advantage. It also has other  benefits. For example, the graduate students would be immediately employable inside and outside academia; the faculty would be able to do some fundraising, though not as much as a demography center; and this sort of critical mass would increase the chance that WUSTL would be the origin of the next big quantitative advance in the social sciences.

The other 1/3 of the program should be filled with mid to late career qualitative scholars. You need this for a few reasons. First, sociology, especially the younger folks, has converged on the view that mixed methods is the way to go. So you will need top notch ethnographers, historical types, and interviewers to make sure that your PhD graduates have a proper view of sociology. Also, graduate students may opt for a qualitative PhD and you will need good faculty to make sure they don’t get lost in the cracks. The most important reason is that older scholars will be able to maintain a distinct identity and forge bonds in a program that is, by design, tilted in one direction.

As a well regarded private school, you might be tempted to mimic your peers and chase the “best people,” which means whoever recently graduated from high status programs with good publications. It’s not a bad idea, but you will directly compete with all the other top 2o programs that claims these graduates. Instead, you might consider a more focused mission that has a very specific, and achievable, intellectual goal. It’s worth a thought.

50+ chapters of grad skool advice goodness: Grad Skool Rulz/From Black Power 

Written by fabiorojas

October 28, 2014 at 12:01 am

Posted in academia, fabio, sociology

17 Responses

Subscribe to comments with RSS.

  1. Perhaps your advice will be better received if you address it to Dr. Steven Fazzari.



    October 28, 2014 at 12:10 am

  2. Got it, Mandy.

    Liked by 1 person


    October 28, 2014 at 1:13 am

  3. I’m not sure about your definition of “unique and important types of quantitative data that sociologists are, or can be, good at.”

    As far as I can tell, you’re basically putting in some other words for “what I do.”

    Is the best way to build a soc dept one where data is the criteria? What about ideas? Approaches? Insights? Creativity? Or, the old-fashioned idea of simply people that have and will do high quality research — even if the data is NOT “unique and important”?


    Not unique or important I guess

    October 28, 2014 at 6:51 am

  4. I really like this idea. It seems like a great way to get to answer important questions about human behavior. Keeping the qualitative people helps ensure that the program focuses on human behavior and not cool stats. BUT I’m an economist so my liking the proposal is probably a bad sign.



    October 28, 2014 at 3:03 pm

  5. @Not unique. There is more than one way to run a program, but methods is definitely a fair one. Chicago, Wisconsin, and Michigan did well by focusing on big N surveys a generation ago. The deeper point is comparative advantage – what can WUSTL do particularly well? If they focus on race, they will go head to head with Berkeley, Princeton, and others. If they do qualitative, they go head to head with Princeton (again), Northwestern, and Michigan’s historical part of the program. If they do population or health, they compete with UNC, Indiana, and Penn State. However, there isn’t a single program that is focusing on this combination of things AND there’s actually a real demand for folks doing this stuff.

    “Or, the old-fashioned idea of simply people that have and will do high quality research — even if the data is NOT “unique and important”?”

    Sounds fun in the abstract, but in practice you tend to fall into the rut of “let’s hire the recent top 5 program graduate with an ASR.” Then you have to compete with Princeton, Harvard, Berkeley and all the rest. So unless WUSTL lays down huge $$$$, they will probably lose that battle pretty quickly and get a predictable program in the 20-50 range.

    What I am suggesting is simple: don’t use cheesy mottos (“get creative people!”) and don’t compete in an already crowded space (e.g., race). Instead, think carefully about an under appreciated, but important, areas of social science and then do that well. You can actually win and you can build a great program.



    October 28, 2014 at 6:20 pm

  6. ^^^ Note that Princeton’s rise in the 1990s/2000s essentially followed that strategy. They over invested in two fields that weren’t fully appreciated at the time (culture and econ soc) and did remarkably well.



    October 28, 2014 at 6:22 pm

  7. Good points Fabio. Some respectful points of disagreement…

    I actually doubt any competitive program is really “run on [particular] data or methods.”

    Further, your arguments implies a) that your style quantitative work is really the most promising and b) there is a pool of faculty that do it and are hirable. I can’t evaluate B, but am initially skeptical that such a large labor pool exists. While I appreciate A is probably just a matter of you being enthusiastic about what you do, there are a lot of quantitative approaches that are rising besides just networks, big data, text and simulations. Why just your style and not some alternative? One can just as easily argue for genetic data or cross-national survey data or maybe even sequence analysis as all of those have gained a lot of momentum in recent years too. Further, any method or data tends to have an arc of enthusiasm, growth and plateau or decline (e.g. look at formal theory in poli sci vs 15-20 years ago). So, I’m skeptical that a methods/data focus for a whole dept is likely to be a sustainable strategy.

    My claim on creative/interesting/excellent people is a claim that emphasizing data/methods may come at the expense of what you present as cliche or conventional modes of evaluating faculty. If forced to choose – and such a choice seems plausible – which strategy is more likely to fail?

    Last, I would respectfully but strongly disagree with your last points. There are a great deal of excellent faculty outside the elite schools you mention. We are still coming out of a very long and deep recession and there are many many under placed people at early and mid career stages. In any economic climate, there are always outstanding scholars flying under the radar of elite places. That only elite schools are training the faculty you’d want is a hard claim to justify. *If* given resources and the opportunity to hire, any dept should be able to find 20 excellent sociologists at schools outside the top schools. One might even suggest elite pedigree is over-valued in our business and the “moneyball” strategy would find the under-valued but great people who have less elite pedigree. One has to find and recruit those people, but my claim is there is a lot of talent out there.

    Plus, what is so bad about being in a dept ranked 20-50 if that is where Wash U ends up? It seems our profession (not particularly you) could be a lot more careful with taken-for-granted elitism and rankings.


    still not unique or important

    October 28, 2014 at 8:08 pm

  8. Thanks for the response. I’ll make this brief since I think we’ve laid out the main points. 1. Why this bundle? Answer: specialization and comparative advantage. In any given time period, there are only a few fields that seem to be low hanging fruit. Aside from broadly articulated suggestions (“the best people”), can you name another field or approach that would be open to new entrants in sociology? Examples: In the 1970s, survey research was the low hanging fruit. In the 1980s, critical theory. In the 1990s, economic sociology. Or, given that WUSTL will probably pay good, but probably not great salaries, can you explain a concrete strategy for building a leading research program in 5 years or so?

    2. Why not settle for 20-50? Answer: WUSTL is usually ranked in the top 20 private schools in America. And if you toss in some elite public flagships, it’s rank would be about 20-ish. Thus, I don’t think that the deans are WUSTL are really going to pour a few millions bucks a year to get a school that is substantially below its poli sci program (#24 in the old NRC), econ (#29 old NRC) or business (# 22 in 2014 USNWR). So anything below #30 would be seen as under performing. 20s is appropriate given peer status and top 20 would make it the best social science program at the school.

    PS. I use old NRC since the new one is utterly confusing.



    October 28, 2014 at 8:23 pm

  9. PS. WUSTL actually has some top tend programs, like its medical school. This isn’t charity work for sociologists. They probably want a well known program.

    Liked by 1 person


    October 28, 2014 at 8:31 pm

  10. It’s worth mentioning that their political science department has already adopted a similar strategy. Along with Rochester, they are commonly recognized as one of the most heavily quant programs, especially in terms of graduate training. This might also push the department in the direction you suggest.



    October 28, 2014 at 9:26 pm

  11. I’m all for creating a sociology department that specializes in computational methods. One thing to consider though is infrastructure. Not every university has the right set-up, including providing server space, for this type of research. I know that at some schools grad students are forced to rent server space from Amazon because the computer science departments are unwilling to share. Having access to a high performance computing cluster would be really essential to getting something like this off the ground.

    Liked by 2 people

    brayden king

    October 28, 2014 at 10:59 pm

  12. Students can apply for education credits for AWS and farming this stuff out to AWS is probably a lot cheaper than building up on-site infrastructure, anyway. Not to mention, if you’re interested in training these students in computational methods that are applicable outside of the academy, navigating AWS/GCS/whatever is a useful skill to have.

    Liked by 3 people

    Trey Causey

    October 28, 2014 at 11:12 pm

  13. If you are building a program from the ground up, this makes sense in the abstract. But part of the reason sociology makes sense for Wash U is that it has a top-of-the-line social work program. At least, that was what they were using as a rationale in their press release. As a result, it makes sense to have people specializing in criminology/deviance, race, and health (or at least mental health). Maybe inequality too.

    But if the University of Rochester wanted to start a department…


    great idea, but...

    October 28, 2014 at 11:18 pm

  14. Go the methods route with a program that was closed down for being “too radical”? With such a suggestion, Gouldner’s ghost might visit you this Halloween.

    Liked by 1 person

    A spectre is haunting IU

    October 28, 2014 at 11:22 pm

  15. Ha!



    October 28, 2014 at 11:52 pm

  16. Fabio is right that WUSTL will need a brand. But I think this brand should be substantive rather than methodological/theoretical. Yale provides a decent example: they focused on building up a brand as historical sociology and have done quite well for themselves on that front with hires, attracting grad students, and decent placements.

    Topics or fields are much more likely to outlast fads than methods.



    October 29, 2014 at 2:37 am

  17. Perhaps WashU sociology could establish a unique identity/brand/direction/strategy by populating the new department with critical realists…


    keep smilin'

    October 30, 2014 at 1:45 am

Comments are closed.

%d bloggers like this: