Discussion thread on placement

[jazzrap]

Guys, 

 

Hopefully my post will not start a ranking war. I apologize if it does. Perhaps there are already threads about schools where folks can discuss placement, but I still want to have a thread dedicated to this aspect of PhD programs, as it is the single-most important thing for our careers. This thread could help some of us fortunate enough to get into grad schools to have better educated choices to make. It might also be useful to prospective applicants. 

 

I want to start off with some of my observations made before and during this cycle. As any other applicant, I have been going to programs' websites to find information about placement, and trying to make an educated guess about what program has the better ability to land its ABDs into jobs. Thus far, these are my observations, and I really hope anyone here can comment on these things. 

 

First and the most obvious: a lot of programs that houses great faculty and seem to have wonderful resources do not participate in the "honest grad numbers" enterprise. My question is to what extent should a program's decision not to be a hundred percent transparent about its placement affect a prospective student's choices to attend. 

 

Second and still pretty obvious: some schools formally ranked between 10th and 15th that emphasize methods training have placement records better than any of the non-Stanford top fives (or top six). Their placement records are also better than many non-UCSD top tens. My question is why this is so. I have a few guesses, and hope you can comment on these.

 

a) a lot of the programs at the very very top tend not to require too much methods from their graduate students, and instead stress "methodological pluralism." Many PhD students coming without prior exposure to advanced statistics tend to try to avoid mathematical training, and their programs' structures are not preventing them from doing so. These people pass the comps, and go on the market.

 

Many top 10 and top 5 programs are pretty large, so faculty pay less attention to their students on average. This observation might be wrong, of course, because these programs may have larger number of faculty as well. However, I do feel that a relatively small program is better than a larger program in terms of sustaining a culture of working together to place well.

 

c) Some of the very very top programs might tend to admit students from diverse backgrounds. Some students have wonderful grades and have a lot of extraordinary experiences outside of the academia. Their SOPs are professional enough to pass the cut-off, yet not as "nerdy" as some other folks in their cohort. Maybe one third of them go on to become the greatest thinkers in their subfield, but a lot of them find themselves into a discipline that they do not particularly enjoy. These two thirds pass the comps, decide to stay, and go on the market without a frontier-driven paper. I know of one student from previous cycle who got into a top 5 in the US but instead joined a British program, with "having a better opportunity to work for the Economist" being one of the primary reasons. These people exist, and many of them do attend "top 5" or "top 10" schools . I am less confident about this theory, though.

 

d) Most importantly, as I found through some anecdotal evidence by some graduate students I know, is that top 5 or top 10 do not push students to work hard in comparison to some in the 10th-15th. Some of us lucky enough to be accepted into top 10 programs might decide to work hard anyways, and we will have pretty good endings. Others do not, and these folks go on the market as well. And top 5 or top 10 may not have intense job training, whereas one grad student from a top 15 literally told me that their DGS and Chair will tell job market candidates what kind of meals to order during office visits for job talks. 

 

In addition to my two observations and the guesses I made, I also hope to share with you one of my strategies to compare schools in terms of their records of placement. Hopefully people can come and correct me and offer their tips. One way to compare placement record is to find how many people who obtain TT jobs without first-authored publications. Say, if one school has a lot of these students, then this program must have a good ability to place. There are other programs that place their students very very well, but tend to place them through publications. I am not saying that we should go to the former kind and be lazy about research. My strategy is just to find out about programs' ability to place after controlling for the number of publications. This approach has a lot of caveats. One of them is endogeneity. More importantly, however, publications in general do not tend to be the sole decider on placement. Interesting dissertation presented in a job talk is much more important than a solo-authored piece in a top journal. Whether a student's dissertation is interesting cannot be observed, so I assume that these variables are randomly distributed, though they are not... 

 

I hope this thread has not offended anyone, and most importantly will not start a ranking war. Sorry if it does. 

 

One of the most urgent questions I and probably other applicants all have, is to what extent we should choose programs based on placement vis-a-vi fit or ranking. There is one school that provides me and other accepted applicants with information more thorough than the "honest grad numbers". From their information, calculated roughly by odds only, if anyone studying comparative politics goes on the market, he/she will have more than 70 percent chance to get a TT job as an ABD and 90 percent chance to get a TT job after a year of post-doc. NO kidding. There are better ranked schools that have much greater fit that also sent me offers. There must be other people here who have similar results and wonder whether they should attend based on record of placement. 

Edited February 25, 2014 by jazzrap

[cooperstreet]

One thing that you kind of hinted at, but didn't state explicitly is this: THEORY.

 

Any department that accepts theorists is going to fuckup their placement statistics. I think this explains Rochester and NYU having good placement records. They simply don't admit many theorists. And theorists don't get TT jobs.

[DKSL]

One thing that you kind of hinted at, but didn't state explicitly is this: THEORY.

 

Any department that accepts theorists is going to fuckup their placement statistics. I think this explains Rochester and NYU having good placement records. They simply don't admit many theorists. And theorists don't get TT jobs.

I wouldn't blame it all on Theory. Most schools explicitly state the subfield of their placements and there is still huge variation among top-20/25 programs within the subfields that aren't suffering as much as Theory in this job market. I personally have been examining IR placement primarily and comparative placement as a secondary concern (since my IR focus dips into comparative a bit) and I still have all of the concerns that  jazzrap has stated. 

[IRTheoryNerd]

a) a lot of the programs at the very very top tend not to require too much methods from their graduate students, and instead stress "methodological pluralism." Many PhD students coming without prior exposure to advanced statistics tend to try to avoid mathematical training, and their programs' structures are not preventing them from doing so. These people pass the comps, and go on the market.

 

I would argue too few top programs actually stress methodological pluralism---the farthest they go in terms of methodological pluralism is to assign George & Bennett alongside King, Keohane & Verba. The false dichotomy between qualitative and quantitative methods (not methodology) is alive and well as though it is the only difference in possible research techniques. Some PhD students coming with prior exposure to advanced statistics simply don't find the underlying logic and philosophical ontological assumptions of neopositivist methodology terribly useful in answering the sort of important questions they seek to ask. Insofar as this is the case, more advanced statistics requirements will not be a solution, but a waste of time for these students. Sure, it's nice to be able to arm yourself with counterarguments, so a few stats classes should be taken by all students. But a few philosophy of science and non-positivist methodology courses should also be a part of the standard curriculum. That would certainly go a long way in getting all of us to stop talking past each other, unfairly criticizing each others' work without any concept of internal validity, and allowing for a future US political science academe in which not only quantoids are welcome. I suppose that has little to do with the immediate concern of job placement, except this: students should focus on improving how they articulate their arguments for why non-positivist methodology make valuable contributions to the field, and engage with senior scholars (at conferences, within departments, on dissertation committees) who might push back against outside the box thinking. Find natural allies. Cultivate those relationships. Network the hell out of them. But don't compromise your research to the point that you barely recognize it as your own in order to land a job or appease a committee member whose raison d'être is to unreflectively prescribe positivist methodology for any and all projects, tone deaf to the actual research question itself, and convinced that anything that isn't hypothesis testing lacks scienticity.

[cupofnimbus]

There's a marked increase from prior years in the number of ABDs entering the job market early. And ABDs don't place as well, because it's a gamble for a school to hire someone who hasn't completed their dissertation. That's not to say it doesn't happen, but the numbers are against the ones who take that chance. Chances are good that you'll find a job of some kind when you finish your degree. Whether or not that's a TT job is yet to be seen (with the decline in TT jobs overall, that is a legitimate concern!).

 

But, rank of school doesn't guarantee anyone a TT job. Certainly, you see the highest ranked research institutions have the highest concentration of faculty from high-ranked schools, and if that's what someone's going for, then that's what they should do. If you can be satisfied with tenure-track and tenure at a mid-tier or unranked school, doing what you do, then it is not as much of a concern. This is a field where people care a whole lot about how they stack up in rankings. Personally, though, I'm not sold on it.

[cooperstreet]

I wouldn't blame it all on Theory. Most schools explicitly state the subfield of their placements and there is still huge variation among top-20/25 programs within the subfields that aren't suffering as much as Theory in this job market. I personally have been examining IR placement primarily and comparative placement as a secondary concern (since my IR focus dips into comparative a bit) and I still have all of the concerns that  jazzrap has stated. 

 Right but I think the theory students who go to Columbia, Princeton, Yale, etc., and then don't get jobs, will lower these schools placement rates. But schools that are lower ranked, like NYU and Rochester, don't have these theory students bringing their averages down. Sorry if I wasn't clear.

[IRToni]

One thing that you kind of hinted at, but didn't state explicitly is this: THEORY.

 

Any department that accepts theorists is going to fuckup their placement statistics. I think this explains Rochester and NYU having good placement records. They simply don't admit many theorists. And theorists don't get TT jobs.

 

In the same veign, I see that methods people have been placing really well recently, which also explains good placement records of some places.

 

A lot of it is, however, I believe idiosyncratic. UCSD places extremely well, for example, especially in IR, and from all accounts, this is in large part due to the subfield chair, who makes grad training a priority.

Edited February 25, 2014 by IRToni

[cupofnimbus]

In the same veign, I see that methods people have been placing really well recently, which also explains good placement records of some places. A lot of it is, however, I believe idiosyncratic. UCSD places extremely well, for example, especially in IR, and from all accounts, this is in large part due to the subfield chair, who makes grad training a priority.

 

Obviously, methods ties in with virtually everyone's research, but few identify strictly as methods. It's true that methods folks place extraordinarily well, but the Ns are very, very small.

[jazzrap]

Thanks people! I am really glad to see many folks are quite interested in discussing such an issue. 

[IRToni]

Would anyone be interested in trying a placement ranking (possibly by subfield)?

[jazzrap]

One thing that you kind of hinted at, but didn't state explicitly is this: THEORY.

 

Any department that accepts theorists is going to fuckup their placement statistics. I think this explains Rochester and NYU having good placement records. They simply don't admit many theorists. And theorists don't get TT jobs.

I agree that theory people might bring noise into the already noisy placement information. However, even taking theory into account, the two schools you just mentioned are still ahead of any of the non-Stanford top 5s in terms of placing their job candidates into TT jobs. 

 

In the same veign, I see that methods people have been placing really well recently, which also explains good placement records of some places.

 

A lot of it is, however, I believe idiosyncratic. UCSD places extremely well, for example, especially in IR, and from all accounts, this is in large part due to the subfield chair, who makes grad training a priority.

Thank you for bringing that up. Yes, among outstanding frontier scholars, some behave more like power brokers than others do, and some take much more time on grad mentoring and training than others do. However, at this point, it is very difficult, if not impossible, for incoming grad students to identify who have more connections in the discipline and put more time on advising. We only hear from second hand sources that some names are more powerful and "nice" than others, but for most senior professors, we don't know.  

[jazzrap]

In the same veign, I see that methods people have been placing really well recently, which also explains good placement records of some places.

Thank you for bringing that up. I guess my objection would be this. The methods placements probably do not explain why NYU and Rochester have better placements than those by the top 5. NYU and Rochester do not train as many methodologists as formal theorists. The market for formal theory is getting more and more abysmal, because econ candidates from a few schools like MIT and Chicago are coming into the battle. Methods placement come from Berkeley, Harvard, Princeton, WUSTL, and perhaps MIT recently. Taking all that into consideration, it will just make NYU and Rochester's placement record more impressive. 

[jazzrap]

Would anyone be interested in trying a placement ranking (possibly by subfield)?

I think you can manipulate the NRC ranking, though subfield information will be harder to obtain.

[AuldReekie]

Would anyone be interested in trying a placement ranking (possibly by subfield)?

 

I feels like there's an open placement project somewhere which aims to get full placement data from Poli Sci depts and publish them online. This strikes me as perhaps more valuable given the influence of individual advisors and the like..

[jazzrap]

I feels like there's an open placement project somewhere which aims to get full placement data from Poli Sci depts and publish them online. This strikes me as perhaps more valuable given the influence of individual advisors and the like..

Hi, I should have posted the link here:  I think this is the thing you mentioned. 

[AuldReekie]

Hi, I should have posted the link here:  I think this is the thing you mentioned. 

 

Aye, that's it, thanks! An up to date wiki (or similar) with full cohort placement details including principle advisor & subfield would be a good resource. Something like this would be extremely useful, especially if it was a requirement for programs like the April 15th deadline. Alas it's in the interests of many programs to obfuscate when it comes to placement.

[Orlien]

On the methodological pluralism point, from what I've heard, required methods training varies dramatically within the top 5. For example, I know that a large number of Princeton students (I think the majority of the non-theorists) take a rigorous three course sequence in quant methods.

 

I think we have to remember that we're working with a fairly small n here

[packrat]

 

 

Most importantly, as I found through some anecdotal evidence by some graduate students I know, is that top 5 or top 10 do not push students to work hard in comparison to some in the 10th-15th.

 

I don't know what anecdotal evidence you are working with here, but I've spent the past 7 years around a top-3 program and a top-10 program, and I can tell you that PhD students at both were working their a**es off. I've never heard any PhD student at a T-5 say they aren't working hard.

 

FWIW, I spoke to my POI at a T-3 yesterday, and she commented that in her opinion, outside of Stanford, there is not a big difference in job placement for any of the T-15 schools (at least in IR; I can't speak for other subfields). She said Stanford is really the only place that is consistently getting its graduates into reputable departments (though by looking here, you'll see that even at Stanford, they're scattered everywhere: http://politicalscience.stanford.edu/graduate-program/prospective-students/placement-record). She noted that especially within the top-5, besides Stanford it is a complete tossup vis-a-vis placement, with many factors coming into play that you cannot control as a prospective student (for example, what faculty will still be there 5 years later).  

 

I was a little surprised by that, as I'd always heard Stanford, Harvard and Princeton named as the "must attends." She really didn't seem to think so, even as a Princeton faculty member. 

 

Again, can't speak for other subfields, but I think the general consensus is that for the most part, the job market is out-of-your control and even if you're at Stanford, you need to be producing great research under committed faculty to even have a shot. 

[cupofnimbus]

Just throwing this one in here, too. This is another APSA report, addressing placement directly: http://www.apsanet.org/media/dsp/Snapshot%20Placement%202009%202012.UPDATED%2019%20JUNE.pdf

 

That is the short report from last year, though there is a longer version locked to APSA members.

[TheGnome]

Department placement figures are obviously very very important and useful, but they are still a proxy to get at the job market success of an X person with similar attributes to future you. That is what we all want to learn, right? This is why we often want to learn about the subfield placements as well as the overall placement records. We naturally think that -say- the IR placement record would be more relevant for us if our subfield is IR. 

 

One way to diminish some of this uncertainty is to check the placement records of your POIs or people with whom you think you are likely to work. Such data is seldom available through departments, so you kind of have to ask the person herself about her students. I want to suggest another way:

 

ProQuest has a dissertations and theses database. I don't know if the database includes all dissertations completed in the US, but I think it is fairly close (if not exhaustive) for the past 10-ish years for most of the top-50 or so institutions. The search engine allows you to search dissertations by adviser. Now this kind of data doesn't go too far back, but I would argue that the most important part is the post-2007 period anyway. Moreover, for the past 5 years or so, many records on dissertations also include committee members (so not just chairs), which is extra useful. Using this tool, you can compile a list of all the students who completed their dissertations under the POI(s) that you want to work with. Then, you can simply google these people, check their CVs and see where they initially placed at. It is not terribly hard to get at, if a bit time consuming.

 

A second step can be that if some of these people who were advised by your POIs have dissertations that are sort of in the same research area as your interests, you can just e-mail these people and ask about their experiences and suggestions. They may not reply or provide useful advice, but then again, maybe they will.

 

I found this to be a useful method, hope it helps you too.

[BigTenPoliSci]

I think it worth remembering that the single most important part of placement is you, the candidate. Attending a #4 program versus a #12 has a marginal effect on your job prospects, but the effect isn't as strong as a peer-reviewed publication and a really good dissertation.

[RWBG]

Thank you for bringing that up. I guess my objection would be this. The methods placements probably do not explain why NYU and Rochester have better placements than those by the top 5. NYU and Rochester do not train as many methodologists as formal theorists. The market for formal theory is getting more and more abysmal, because econ candidates from a few schools like MIT and Chicago are coming into the battle. Methods placement come from Berkeley, Harvard, Princeton, WUSTL, and perhaps MIT recently. Taking all that into consideration, it will just make NYU and Rochester's placement record more impressive. 

It's probably less about training capital-M Methodologists and more about training applied types who are methodologically sophisticated. The people we interviewed from NYU tended to use a fair bit of formal theory and statistics in their research, and different departments will do better jobs at training such people.

Beyond that, I don't have too much to add to this, except to note that it's going to be very difficult to get at some of the things you want to here. So many confounding factors, from selection effects, to variation in institutions' efforts to compile and publish data, to subfield variation, to variation in methodological approach to the discipline, to movement in faculty, etc. Best of luck though

[jazzrap]

I don't know what anecdotal evidence you are working with here, but I've spent the past 7 years around a top-3 program and a top-10 program, and I can tell you that PhD students at both were working their a**es off. I've never heard any PhD student at a T-5 say they aren't working hard.

 

FWIW, I spoke to my POI at a T-3 yesterday, and she commented that in her opinion, outside of Stanford, there is not a big difference in job placement for any of the T-15 schools (at least in IR; I can't speak for other subfields). She said Stanford is really the only place that is consistently getting its graduates into reputable departments (though by looking here, you'll see that even at Stanford, they're scattered everywhere: http://politicalscience.stanford.edu/graduate-program/prospective-students/placement-record). She noted that especially within the top-5, besides Stanford it is a complete tossup vis-a-vis placement, with many factors coming into play that you cannot control as a prospective student (for example, what faculty will still be there 5 years later).  

 

I was a little surprised by that, as I'd always heard Stanford, Harvard and Princeton named as the "must attends." She really didn't seem to think so, even as a Princeton faculty member. 

 

Again, can't speak for other subfields, but I think the general consensus is that for the most part, the job market is out-of-your control and even if you're at Stanford, you need to be producing great research under committed faculty to even have a shot. 

Thanks, although I would say that some schools still perform consistently much better than others within top 15s. 

[jazzrap]

I think it worth remembering that the single most important part of placement is you, the candidate. Attending a #4 program versus a #12 has a marginal effect on your job prospects, but the effect isn't as strong as a peer-reviewed publication and a really good dissertation.

Yes! I guess methods training being equal, I will go to a program with better fit, because this way I will make the most of the "me" factor. 

Edit: previous post was confusing. 

Edited February 26, 2014 by jazzrap

[jazzrap]

It's probably less about training capital-M Methodologists and more about training applied types who are methodologically sophisticated. The people we interviewed from NYU tended to use a fair bit of formal theory and statistics in their research, and different departments will do better jobs at training such people.

Beyond that, I don't have too much to add to this, except to note that it's going to be very difficult to get at some of the things you want to here. So many confounding factors, from selection effects, to variation in institutions' efforts to compile and publish data, to subfield variation, to variation in methodological approach to the discipline, to movement in faculty, etc. Best of luck though

Yes, I agree.