A public service announcement

At the request of Jim Humphreys, I’m making a little PSA about postdoctoral positions at UMass: there seems to have been some confusion over the labels used for the job searches on MathJobs.  The label “POSTDOC” is a position is statistics.  If you are not a statistician, and wish to have a job at UMass next year, make sure you applied for the Visiting Assistant Professor position, labeled “VAP.”

Just so there’s something for people to discuss: why is the mathematics community so bad at coming up with a unifying terminology for postdoctoral positions?  The academic ranks on the tenure track as named very consistently throughout the US in both public and private institutions, but postdoctoral ones are a very confusing mishmash.  You can be a “Fellow,” a “Instructor”, a “Professor” or even a “Member” with all kinds of possible adjectives without any kind of clear terminology behind them.  I’m sure this is mostly just for historical reasons, but is there any hope of rectifying the situation?

31 thoughts on “A public service announcement

  1. This can actually be a significant problem in practice. I had a conversation at tea last semester with several grad students at Columbia who were on the market this year, and several of them had missed a lot of positions to apply to because they didn’t realize that they were postdocs.

    Probably Mathjobs are the only people in a position to fix this, as schools are unlikely to change their naming conventions quickly. Perhaps a checkbox so that jobs are listed on the front page as T, P, or A depending on whether they’re tenure stream, postdocs, or “adjunct” type positions?

  2. I think is basically a good idea, and seems to be something of an oversight on the part of MathJobs, since EIMS (the predecessor of MathJobs, which is just for announcements) actually does sort jobs into “Categories” which include “Post-Doctorate” and “Non-tenure-track faculty.”

    The problem with this is that a number of jobs don’t necessarily fall into one of these categories or the other. Certainly at Oregon, for example, we’ve hired people to “Instructor” positions (which would tend to be considered on the adjunct side of the postdoc/adjunction gap) which the department has done their best to treat like postdocs, and I assume this happens at other departments as well.

  3. I dunno that the adjunct/postdoc distinction is really so fuzzy. What’s the teaching load? I think a postdoc means its teaching load should be 1/2 or less. A postdoc is a position that has relatively low teaching, lasts for 1-4 years, and only goes to people who have gotten their Ph.D. recently.

  4. That’s a good rule of thumb, and was born out by a quick look at MathJobs. On the other hand, what about a position which is 2-2? Is that, by definition, not a postdoc?

  5. Yeah, so certainly the 2/2 case is a little controversial. But yeah, I’d argue that by definition it isn’t a postdoc. (With the possible exception of *teaching postdocs* meant to prepare people to be TT profs at elite liberal arts colleges, which I don’t know much about.)

  6. Though I’m not directly involved in hiring at this point, I know the local complexities of justifying “postdocs” to uninformed administrators. Their
    experience of “postdocs” is usually limited to the lab sciences, where big outside grants fund these as non-teaching positions. In our case (and many others), the VAP is a regular faculty position with decent salary and benefits but no tenure-track. At first most were 2 year positions, then most became 3-year contracts (but some people got real jobs and left early). The teaching load at UMass has been uniformly two courses per semester, but usually the newcomer gets two sections (30+ students) of a lower level course with one preparation. More experienced people usually get a chance to do upper level undergraduate courses if they wish. The trade-off for 2+2 is to have absolutely no other responsibilities. Not ideal, but realistic given the strained finances as state funding disappears and tuition/fees rise dramatically.

  7. P.S. Quite a few of our recent VAP people have gone on to real jobs at good places, some have left academia altogether, etc. The UMass department also has several “Lecturers” (regular faculty status but with year-by-year renewable contracts that tend to go on a long time) to handle large courses at a more remedial level. The lecturers are usually good at what they do, whether or not you agree with their pedagogical approach, but they typically have dropped out of math Ph.D. programs. While they are super cost-effective for the administration, they make the regular faculty teaching look skimpy. Ideally the administration wants only tenure-track faculty, as long as they draw in million dollar grants (like the polymer scientistis) without any VAPs. It’s hard to explain all of this to paying undergraduates, I guess.

  8. Regarding teaching loads, my post-doc at Michigan in the late 90s/early 00s, which was labelled as an Assistant Professor, had a 2/2 teaching load, and (at least in the first year) each class was 4.5 hours a week. On the other hand, it was clearly a post-doc in the conventional sense. Thus I think that the idea of the post-doc label being restricted to 2/1 teaching or less is pretty unrealistic (although an appealing fantasy/source of envy for those whose post-docs involve heavier teaching loads).

  9. But isn’t the whole theory of a postdoc that in exchange for lack of job security, wild geographical instability, and harm to your spouse’s career, that the postdoc gets *something* in return?

    If not a lower teaching load, then what makes postdocs not exploitative? What makes them different from an adjunct?

    To be clear I’m not blaming individual depts here, once the postdoc system is in place on the margin having more of them is not very exploitative. But we can compare math to fields without postdocs, and ask whether the system is exploitative.

  10. Noah-

    Compared to adjuncting jobs, 2/2 is a lower teaching load. The Instructor positions at UO are pretty reasonable as adjuncting jobs go, but the teaching load is 4 classes per term.

    As for whether the postdoc system is *exploitative*, since when is doing a relatively cushy job for pay considerably better than the US average “exploitative”? Would it be less exploitative to have a norm in which the only way of improving your credentials for a long-term job was to stay in grad school longer or take adjuncting positions?

    I’m not sure I know of any academic discipline where the career track is as favorable for the employee as math (with perhaps the exception of economics); in essentially all other cases I know of you either have to spend a long time in grad school, with a very low chance of getting a job when you graduate or you face an exceptionally difficult climb in getting your own lab started, which you also have a very low chance of succeeding at. Which is far from saying it’s perfect in math, just that things are a rough road in any academic discipline. Fundamentally, the issue is just too many people chasing too few resources.

  11. Dear Noah,

    I appreciate the spirit of your previous question, but in practice people don’t necessarily get to pick their working conditions. On the other hand, the position I (and many others) had at U of M was clearly not an adjunct position: It was certainly not viewed in that way by either those holding it or by the department, and in terms of more objective measures: it was for three years, the recipients were chosen primarily on the basis of research rather than teaching, and the expected career path was that people would move from those positions to tenure-track positions. There was just quite a bit of teaching!

    Best wishes,


  12. I’m willing to admit that I was too aggressive in the line that I drew, and I should have kept 2/2 as within the allowed range. (Though this depends a lot on the actual nature of the teaching, running sections vs. running a whole course are very different.) Nonetheless, I stand by the spirit of my argument, which is that if there’s too much teaching then it shouldn’t be considered a postdoc.

    In reference to Ben’s comment, certainly large swaths of academia are worse, but I’m very open to the argument that large swaths of academia are unethical. Furthermore, I think the “too many people chasing too few resources” is a bit of a red herring when it comes to postdocs in mathematics. The limited resource here is TT jobs, and the existence of postdocs doesn’t increase the supply of TT jobs or the likelihood of getting one, it just means that you have to commit another 2-5 years before you know what you get! Furthermore, that’s 2-5 years with limited choice of where you live and for which you have nothing to show if you end up leaving academia (unlike a Ph.D. which can be useful for non-academic jobs).

    For example, I’m sure philosophy as a discipline is worse for the average Ph.D., but all my friends who are philosophers have had nice tenure track jobs for a few years by now. I’d happily sacrifice 10% of my chance at R1 TT job in order to know the answer 3-4 years earlier.

  13. Furthermore, I think the “too many people chasing too few resources” is a bit of a red herring when it comes to postdocs in mathematics.

    I’m failing to see why it’s a red herring; if the balance of candidates against job openings were more favorable to the candidates, than schools wouldn’t be able to afford to insist that you have some number of years of postdoc first. Or, more accurately, they couldn’t insist on a track record that almost no fresh Ph.D. could conceivably accumulate. Instead they would be trying to grab the best people before they did postdocs.

    But that’s not how things are; I’ve seen the idea of hiring a promising fresh Ph.D. mooted in a hiring meeting, and it was shot down essentially instantly.

  14. The other thing is, a hypothetical world where mathematicians don’t do postdocs is a hypothetical world where mathematicians get tenure a lot less often, due to people not panning out, so you wouldn’t really know after all.

  15. Certainly supply and demand is part of the issue. However, I’m more inclined to attribute this to things like status quo bias, low rate of unionization, or the distorting impact of the top few schools (who might hire away your recruits down the road).

  16. Unionization? Who’s going to unionize? Maybe you don’t realize this, but (at least in my limited experience) professors want to hire postdocs; they don’t think they’re doing someone a disservice by hiring them in such a job. On the whole, what I’ve heard is clamoring for hiring more.

    Status quo bias is an issue in the sense that postdocification is a stable equilibrium; it’s hard to fight as a candidate, and the system works extremely well for all the people who actually make the decisions (who all have permanent jobs). But fundamentally, I think is just the dynamics of a system where too many people are chasing too few permanent jobs, and you stand a decent chance of improving your credentials over time (as opposed to, say, athletics). If we didn’t do it with postdocs in mathematics, it would just be some other route, probably in part hanging out in grad school longer. If you look in this article there’s a table of completion times: in “math and physical sciences” 2/3 of the people who would eventually graduate were out after 6 years, in humanities less than half were. Of course, hanging out in grad school longer works better for some people’s personal situations, but making two or three times as much money (which you typically will as a postdoc) works better for others.

    If you want to blame someone for postdocs, blame the NSF; they’re the only single player who actually has the money to bring serious weight to bear on the situation, and they’re leaning pretty heavily in the direction of postdocs. They’re giving a big slice of people (you and me included), mostly people who were already pretty well positioned, a big leg up in the arms race, and thus creating the incentive for schools to wait around for us to finish up our NSF money rather than hiring us right after graduating. The same could be said of AIM and Clay. Of course, they could (rightly) come back that their concern is mathematics, not the personal lives of mathematicians, and from that perspective, the postdoc fellowships make a lot of sense.

  17. Noah – if you’re willing to admit that vast swaths of academia are unethical, then you also have to admit that vast swaths of just about everything is unethical. For example, Microsoft releases software knowing that if they tested it more they would have found and fixed a few more important defects. Indeed, despite the treatment of adjuncts, most of academia is still more ethical than just about everything else out there.

    Fundamentally, the problem is that the world can’t sustainably support 6 billion+ people on an American standard of living, and we’re finally running out of ethical ways to do better than everyone else.

  18. Noah – from the perspective (which I admire quite a bit and have a lot of sympathy for) of selling everything you own and giving the proceeds to the poor, just about everything in the world is an ethical cesspool.

  19. Modulo the recession (which is a problem in all fields), I don’t think mathematics suffers from a severe lack of jobs. It sometimes takes a few tries, and you might need to lower your sights, but as far as I can tell for the most part people tend to get jobs.

    Instead the problem is uncertainty. It is very hard for candidates to reasonably estimate their chances. It is very hard for schools to predict which candidates will be successful in the future. It’s these uncertainties combined with a wide range of desirability of permanent jobs which I think is driving things.

    Even if there were the same number of candidates as jobs, you’d still need to figure out which candidates got which jobs. The most desirable schools would wait to see who was successful before making hires. Then the next tier of schools wouldn’t want to make their decisions too soon. Throw in some external funding for postdocs, I think you could end up in a postdoc system without any scarcity of jobs.

  20. Noah – the number of jobs and the pigeonhole principle tell me that it’s not true that most people get jobs (even before the recession).

    I’m inclined to believe that we just haven’t seen the sectors where a lot of people don’t get jobs.

  21. Noah-

    I’m not sure what point you’re making at this point. It sounds like you agree with me; even if there are enough jobs out there to accomodate all the Ph.D.s in math (which as quasihumanist says, is false), essentially any pool you care to define that people are chasing after is smaller than the number of chasers, almost by definition (because, as you say, people are uncertain, and can always aim a bit higher).

  22. If it were the case that there was a wild mismatch between the number of jobs and the people looking for them, that would suggest different ethical dilemmas and solutions. For example, that we should be admitting fewer Ph.D. students. If instead the primary driver is schools wishing to reduce their uncertainty, then I think the solutions is that some of that benefit should be shared with the postdocs in the form of significantly reduced teaching.

  23. I think we are all quite ignorant about what happens outside the research institution market. (I might be a little less ignorant myself, but not by a lot.)

    It’s also quite hard to seriously gather data here, because there are a LOT of small hiring institutions, and many of the people in this market come from small PhD programs (and don’t do postdocs though many do VAPs).

  24. I apologize in advance for almost certainly misunderstanding some of this recent discussion (am severely underslept today): but from where I sit, as a lowly toiler (mathematically speaking), the problem is one of resources. I would *love* to have a postdoc to come and work here in Saskatoon, but I can’t magic up the financial support of thin air.

    Now in the large scheme of things, one might say “why should this bozo get a postdoc?” and I’d be hard pressed to deny the force of that argument; but this to me seems to support the idea that it’s a question of resources.

  25. Yemon- I think you seem to have misunderstood the conversation. My understanding of Noah’s argument is that it is essentially cruel and unethical to ask someone to root up their entire life and move to Saskatoon to talk to you for 2 years, and then kicking them to the curb. I suppose his argument is that you should instead work to increase the number of tenure track jobs in your department (which of course, may or may not be something you can influence).

  26. Re 24/25, Ben’s summary is more or less right, except stated much more vehemently than I’d state it. In particular, I don’t think it’s unethical if the postdoc has reduced teaching, and at any rate I think “cruel” is far too strong a word.

    23. I don’t have a whole lot of data from people outside of research schools, but I have talked to some people at lower tier math Ph.D. programs, and they said their classmates were largely successful in getting college teaching jobs that they were happy with. But those were pre-recession graduates.

  27. Well, four years ago someone offered me the chance to uproot my entire life and move continent, and it wasn’t like I had any guarantee of a future job when I did it. Ce qu’il faut faire… But point taken, and I apologize if I came across as dismissive.

    My guess is that certain places may use the teaching load as a way of paying part of the postdoc’s salary, to reduce the demands on departmental funds or the supervisor’s grant. But since I don’t know how the US model works, perhaps this does not apply.

  28. Well, yes, I definitely overstated the case a bit in 25. A more gentle stating of Noah’s point which I agree with is this: given the expectation that a student has to do a postdoc in order to become a research mathematician, we (by which I mean the community of permanently employed mathematicians) feel like we are doing recent graduates a service by providing them with employment as postdocs; however, our creation of these positions is exactly what generates that expectation in the first place, and thus creates an expectation which does not work well for many people. There is actually a serious, and I think underexamined, conflict of interest here between people who have permanent jobs and those who don’t.

  29. Well, four years ago someone offered me the chance to uproot my entire life and move continent, and it wasn’t like I had any guarantee of a future job when I did it.

    Indeed, and no one is saying you shouldn’t have. But it is a serious question whether we want to structure the field of mathematics so that people will have to do that (well, maybe not changing continents, of course, but almost certainly moving to a different city, usually not that close to where they were before) in order to be a research mathematician.

  30. Indeed, that’s an excellent way of putting it. Clearer than anything I’d said.

    To my mind the main problem with the postdoc system is that people who don’t eventually get permanent jobs that they’re happy with would be better off knowing that earlier. For example: because starting another career path earlier is better, because moving is expensive and annoying, and for partnered people because moving is typically bad for the partner’s earning potential.

    Paradoxically, the postdoc system might increase competition for TT jobs. The additional 2-5 years spent as a postdoc increases the desirability of a research academic job relative to a non-academic or teaching job. E.g. I’d be much less likely to try to shift careers to being a software engineer in the bay area as a 32 year old living in New York, than as a 29 year old in living in Berkeley with a spouse who already had a bay area job.

Comments are closed.