A View on the Job Market for Computational Social Science

and some professional news...

Beginning September 1, I will be Assistant Professor and hold the chair of Computational Social Science at the European Union Institute in Florence. I’m extremely excited; this position is an ideal fit for my research and sangiovese interests, an intellectual adventure I’m thrilled to be undertaking.

This is a unique position (in addition to the various ways in which European academia functions differently from in the “states”), so if you are someone I know and want to hear more details, check the end of the post.

You might note that it’s nearly June. I received and signed the formal job offer this week. This job market season has been extremely time-consuming and complicated. So I've had plenty of time to reflect on my experience.

The academic job market is maddening. The stakes are crazy high and the odds not very good — there’s a slim chance of getting anything, let alone having any agency over where you and your family live. So it’s impossible not to spend dozens of hours agonizing over every little detail of your job market presentation, memorizing questions to ask every faculty you’ll meet — each of these details can, theoretically, be pivotal in determining where you’ll spend the next few to fifty years.

On the other hand, these agonizing hours might have literally zero effect on your chance of getting the job because the committee has their eyes on someone else — and there’s no way to know in advance. This information environment can turn even the calmest person into a paranoiac conspiracy theorist, reading intentions into every offhand remark or the timing of emails. And I am not the calmest person.

The situation becomes worse for interdisciplinary searches, when all parties have less experience and thus fewer shared expectations about what to expect. Candidates don’t know what to prepare for and hiring committees have a less fixed set of criteria by which to evaluate.

This is the current situation for Computational Social Science. Nobody agrees on what it is, but many early-career social scientists say they do it, and many departments know they want to do it, too.

Given that there were so many open positions in this area, I strongly and wholeheartedly identified as a Computational Social Scientist. And some parties, at least, agree: I was given a total of five invitations for in-person interviews (“flyouts”), including at EUI, where CSS is my new official title. I’ve attended the excellent Summer Institutes in Computational Social Science, and presented at multiple iterations of the International Conference on Computational Social Science. In fact, I’m part of a keynote panel at IC2S2 this summer in Philly, and gave a keynote presentation at last year’s European Association of Computational Linguistics in Dubrovnik.

All this to establish some baseline credibility. If you read this blog, you have some idea of what I work on, but it’s difficult to understand the contours of academic disciplines from the outside. I am genuinely excited about CSS, and I’m writing this in the hope of further consolidating the field. Workshops, conferences, special issues and summer institutes — these are all important steps in legitimizing an academic field, and things seem to be going well for CSS on these dimensions.

(Speaking of special issues — check out the collab between ICWSM and the JQD:DM, published today!)

But faculty hiring is where the rubber hits the road. Open science, registered reports, replication projects, R packages, data donations, journal innovations — all of these exciting and promising innovations in social science ultimately live or die on the preferences of a handful of boomer faculty who might not have any idea what you’re talking about, and even less interest in finding out.

That’s not an offhand generational jab, by the way — the demographic reality of higher education is a first-order problem for the consolidation of CSS, or for any attempt at shifting the course of social science, really. It’s not that older faculty aren’t smart or that they don’t know math or coding — rather that the lack of mid-career faculty working at this intersection limits the circulation of the common standards and aspirations that makes a field gel.

This argument is consistent with the fact that CSS seems to be taking off faster in Europe than it does in the US. This is partly catch-up to the cutting edge; Computational Sociology for example seems to simply mean any quantitative work in a way that is largely taken for granted in the US. But CSS seems a buzzier buzzword across the pond, will real institutional buy-in. And I can attest that 4 of my 5 job interviews in CSS were in Europe. (I also had two interviews at US Political Science departments not explicitly in CSS, that I won’t discuss here.)

In fact, I was given a flyout for every CSS job to which I applied in Europe.

And yet my experiences and outcomes were wildly different at these different places. Most importantly, I got only one of four job offers — and at EUI, the job I ultimately was offered, I was ranked fourth.

So take what follows with a grain of salt. This is just my perspective, from the inside, but I think that the inside is what really matters to the lives and careers of other young researchers in CSS.

Grad students are often given the advice that “every talk is a job talk”—you should always put your best foot forward because someone in the audience might be on a hiring committee one day. But my experience this year makes me think that the converse is also true — “every job talk is a talk,” an opportunity to share your research with people and get to know more about their research. And as a meta-scientist, every job talk is an opportunity to learn about this crucial element in the sociology of science.

It’s easy to slip into abstractions when talking about fields and disciplines. “CSS” as some huge, smooth, seemingly well-defined concept. But the reality is that CSS isn’t that big, it’s still a heterogeneous and lumpy collection of countably many departments and relevant faculty. In a given job market season, there are only a handful of positions, and the map of the abstraction fails to describe the jagged and idiosyncratic social reality underneath.

Here’s my take on the field: CSS is modern quantitative social science, freed from the heavy weight of our inherited theories, disciplines and institutions. There is not yet a theoretical or methodological core — and all the better. From my perspective on the economics-y side of political science, it’s especially useful as a way to exclude game theory as the only formal meta-language. My particular interest is in the quantitative study of the internet and its effects on politics and society — but it’s a big tent!

Here, then, are some of the amazing and fantastical participants in the CSS Circus.

---

The first flyout interview invitation was to EUI, followed immediately by a flyout for a position at the department of Methodology at LSE. Later came one at ENSAE/Crest, and another at the Toulouse School of Economics. I’m grouping these four together because they were CSS searches in Europe.

The job ads follow:

EUI

The Department of Political and Social Sciences at the European University Institute seeks to appoint an Assistant Professor in Computational Social Science, broadly understood. The position is advertised at the rank of Assistant Professor to commence in the academic year 2024/25. We welcome applicants from candidates with a substantive research agenda either in political science or in sociology, and who integrate research ethics into their methods teaching and research.

LSE Department of Methodology

The Department is recruiting a new faculty position in Computational Social Science to join its growing research and teaching group in this area. Candidates at the Associate Professor level should hold a PhD and have extensive expertise in the development and/or application of computational methods in the social sciences. They should have an emerging international reputation as a scholar applying cutting edge computational social science methods and a track record of outstanding research. The should have a significant track record of excellence in teaching computational research methods for the social sciences at undergraduate or postgraduate level.

ENSAE Paris-CREST - Assistant Professor of Sociology

Sociologists at ENSAE Paris-CREST use empirical data and a diverse range of methods, including quantitative and mixed methods, to study a wide range of questions. We have a strong expertise in methodology and a rich practice with national and international datasets and digital data. Our current work addresses questions of social mobility and inequality; demography; gender, sexuality, and family; work, employment, and organizations; lifestyles and culture; economic sociology; environmental social sciences; migration and immigration; health inequalities; politics; computational social science. The position is open to all quantitative, mixed methods or computational social scientists, regardless of their preferred methods and research themes.

Toulouse School of Economics — Department of Social and Behavioral Sciences

I can’t actually find the job ad for this one — but it was explicitly targeted at Computational Social Science. I recall that it said that only candidates from outside of Econ would be considered, which is a great idea for trying to shift institutional direction — make it harder for existing power blocs to spin whatever new thing you’re trying to do into hiring someone they wanted to hire anyway, with a thin pretense.

The interview procedure for the first three positions was essentially the same—and very different from what I’d experienced in the US (the interview at TSE was more American-style). Rather than two full days of one on one interviews, the total face time at each place was under three hours:

• the standard job talk

• a formal interview with the search committee only

• a brief, informal chat with PhD students and junior faculty

The nominal justification for this procedure, as I understand, is fairness: the lengthy, social element of a US job interview allows for biases to creep in. By keeping the evaluation short and standardized, all candidates are treated the same.

My perception of the process is that it did not do my application any favors. I have done lots of different kinds of work throughout my career; my “profile” is somewhat difficult to categorize. I think I would have benefitted from more time to discuss different aspects of my work with different people — and I think that this is especially important for an unsettled field like CSS.

Of course, from a meta-science perspective, we have absolutely no idea about any of this. We don’t have theories, descriptive data, experiments — nothing. According to our own professional standards of evidence, we have no reason to think that one way is better than another. That’s fine! We still have to act in the world. But I think that metascience can help us act better.

Substantively, it seemed that every place had a different expectation of what “CSS” means — though again, it was hard for me to really figure this out, with less than 3 hours of face time.

LSE seemed focused on the rigor of the methods from a social science perspective, and of collaborating with people in disciplinary departments. ENSAE/CREST seemed to interested in expanding their teaching in CSS, and focused on that (as well questions about how I would fit in in a sociology department). TSE was very open-ended but with a clear Econ inflection — they were building up a new department of social and behavioral science and wanted people from outside of Econ to build it up.

EUI seemed to be thinking more big-picture; I was asked questions about where I saw the future of computational social science and was chided, after my job talk, for only having said the word “theory” once. Readers of my blog will hopefully find this ironic.

The result of all this? I was ranked third at LSE, third at ENSAE/CREST, fourth at EUI, and not first at TSE.

---

I was invited to give one CSS job interview in the US; it was also the only joint hire-CSS position:

Rochester (joint between Data Science and Political Science)—Computational Social Science

The University of Rochester seeks applicants for a tenure-track assistant professor in computational social science…Candidates should have strong skills in computational social science and/or data analytics, as well as a research agenda that uses these methods to analyze substantive issues in political science, economics, or linguistics.

These “joint hire” positions further complicate the problem of an emerging field like CSS; rather than one search committee, you have to satisfy two. My concern coming in was the Data Science side, and I tried to add “proof of work” details about the models and data I used.

This turns out not to have been the choke point. I was ranked “below the bar” by the Political Science department. That is, even if everyone else they interviewed decided to turn the job down, they still wouldn’t make me an offer.

That’s what happened, in fact. The search failed.

---

From what I can tell, I was too much of a “political scientist” at some departments, and not enough of one at others. My work was too theoretical at some departments, and too nitty-gritty methodological for others.

Knowing what I know now, I think I could’ve adapted my presentation to each of these departments in a way that would’ve significantly improved my chances—to a much greater degree than for a standard disciplinary hire in Political Science. The amount of heterogeneity is simply much higher in CSS at the moment than in settled fields.

Of course, maybe I’m wrong about this and my work just sucks. The job market certainly can engender that kind of thinking. So in addition to providing a survey of the various approaches to CSS hiring, I’m hoping that this post can attest to the difficulty of this industry.

I ended up with an amazing job that I’m super excited about — but just barely. And that came alongside a lot of uncertainty, failure and disappointment.

Details of my new position

Hello friends! This is all very confusing, and just to have my spiel in one place, here are some details.

First, I want to say that I appreciated my time at Penn State, that the people there are excellent and that I was treated very well. My main impetus for leaving was simply the location; I enjoyed the bucolic college town environment but the dead-center of Pennsylvania was not a good fit for my stage of the life course. My acquired fondness for the Nittany Lions will, I suspect, persist throughout my life.

Onto the new gig!

EUI is “an international postgraduate and post-doctoral research-intensive university and an intergovernmental organization with juridical personality, established by its founding member states to contribute to cultural and scientific development in the social sciences, in a European perspective.”

There aren’t any undergraduate students, only Master’s and PhD students. (Note that in Europe these are separate but add up to the standard five years of graduate study. Incoming PhD students must have an MA and then take only three years to finish, rather than restarting as is generally the case in the US.) There are also a great bunch of postdocs and visiting fellows at any given time, so the intellectual community is excellent. Teaching takes the form of graduate seminars, and there are relatively fewer contact hours than is standard in the US, so advising takes up a higher percentage of non-research activity.

Most unusually, there are not permanent faculty positions at EUI. Like all full-time faculty contracts, my position is for five years, with a possible extension of three additional years. This means, of course, that tenure doesn’t exist.¹

Even within this unique structure, my position is novel. EUI has very little “general funds” in the way that an American University or College does; kind of like the Germanic systemic, faculty are hired based on “chairs” that fund them and their research. Traditionally, EUI (or at least the Political and Social Sciences Department) had hired senior faculty to these chairs, equivalent to “full professor” positions. These tended to be established and even prominent faculty from within Europe, with the occasional later-career American.

But this meant that there simply wasn't a position analogous to a “junior faculty” or “assistant professor.” Until this cycle, when the SPS Department decided to split up a chair to hire two Assistant Professors. I’m very excited that Sascha Riaz will also be joining the department, as the Peter Mair Assistant Professor in Comparative Politics.

Although I was scheduled to go up for tenure at Penn State next year, it will never happen at EUI. This entails a non-trivial risk, career-wise: if I stay at EUI for 8 years, I’ll be 14 years post-PhD and still have the official rank of Assistant Professor, with no permanent position.

Fuck it, we ball.

Which is to say that I’m very interested in maintaining and expanding my professional network within Europe! I’m hoping to host as many workshops and talks at EUI as I can, once I figure out how things work, and please keep me in mind for any relevant events you’re hosting across the continent!

1. This has extremely *significant* implications for my matching tattoo with J. Hodgdon Bisbee.