Monday, February 21, 2011

Job Data

The actual, real-life purpose of The Grad School Experience image that I posted on Friday was an attempt-at-humor introduction to a discussion with my research group.

As part of this discussion about academia, careers, life etc., I presented some histograms that showed how many graduates from our research group were doing different types of jobs after obtaining a PhD or (graphed separately) MS. The categories for PhD graduates were Academia, Industry/Business, and Government; those 3 categories accounted for all PhD graduate students from this research group from the past 20 years. I then compared these data with those reported for PhDs in our general field of science, compiled from the NSF survey that tracks the careers of doctoral recipients.

The database for my research group represents the advisees of 4 faculty members in related fields from 1990 to 2010, and therefore consists of quite a few individuals. These data are not secret. My research group maintains an active, easily accessible directory of current and former graduate students and postdocs, including a listing of current employment. I don't think anyone had recently compiled it, though, so it was interesting to see and discuss the graphs.

There has been a lot of talk in the media/blogosphere about how graduate programs should show the current employment data of their alumni/ae so that prospective (and current) students will have a better idea of their chances for PhD-relevant employment, especially for academic jobs. The general idea has been that these statistics are grim, and therefore some potential grad students may be convinced to pursue a different education and/or career path.

But what if the data for a particular program show that 98% of graduates who wanted an academic job, no matter whether the PhD was obtained in 1990, 1999, or 2009, got an academic job? The danger there, of course, is that you will appear to be promising something that you (as an advisor or as a department) can't promise: that if a PhD graduate of that group wants an academic job, they will definitely get one.

Nevertheless, these data are real, the dataset includes a large number of individuals, and the results show that our graduates have been successful at obtaining academic jobs if they wanted that type of job. Perhaps owing to the nature of this particular research subfield, the % of graduates in academia is higher than the average for our general field of science. I have not yet broken out the data into finer-grained categories -- e.g., how many graduates from our group are at different types of academic institutions -- but that would be interesting to do as well.

It was also interesting to see that all of our former PhD students who are in non-academic employment sectors have careers that are relevant to their PhD training.

I think the data were useful to show, at least as a launching point for more in-depth discussions of career paths (academic or not) that have been taken by graduates of our research group. And I would go even further and say that these data were important to show because they indicate that getting a PhD in our research group/department/institution is worthwhile and is likely to lead to interesting and PhD-relevant career opportunities.

It is quite possible that not everyone will agree with me on that, so my question to readers is: Do you think that employment data such as these:

- should be shown whether or not they paint a grim or rosy picture of PhD-relevant employment opportunities: individuals can make their own interpretation and choices;

- should be shown if they indicate a slim chance of graduate degree-related employment, but not shown if they seem to give reason for optimism about employment (because that might be misleading and give false hopes and make the few who don't get their preferred job feel even worse?);

- should not be shown at all. Students and postdocs are responsible for educating themselves about career opportunities and/or it is irrelevant what other graduates of the same program have done in the past;

- other?

30 comments:

Alex said...

Absolutely show the data. If it's grim, they should know that and plan accordingly. If it's positive, well, grad students have so many other sources of misery, at least this takes one load off their shoulders.

I am keeping track of my undergraduate research students, senior project students, and Masters students (so far just 1, but hopefully more in the future). It's useful to be able to tell NSF "I have mentored a diverse and interdisciplinary group of N students from M different majors in the past Y years, including [insert # of women and under-represented ethnic minorities here] and [insert winner of a prestigious award here], and since graduating they have gone on to do the following...." If I'm showing this data to NSF, why shouldn't I be showing it to people who are thinking of working with me?

My only real beef with all of this is the focus on academia: So many people assume that the most successful undergrads are the ones who get PhD's (face it, the funding agencies LOVE to see American students go on to grad school), the most successful PhD students are the ones who become professors, and the most successful professors are the ones who are at R1 schools. I thought that the rest of the world spends money on academia because we provide training that has some sort of value in the outside world.

Kevin R. Covey said...

I think showing the data is a fantastic idea, no matter what it shows. It provides a concrete starting point for the important discussion of career paths, and either way the data leans, it helps the students cultivate realistic expectations for that particular institution and/or group. The discussion should also make the point that past performance is never a sure-fire indicator of future returns, but if the question is having students base their career expectations on some form of bona-fide data, or on their own unfounded expectations, I'll vote for data every time.

Anonymous said...

I work in the humanities, not in the sciences, and statistics are a lot worse than 98%. In my field, you'd have to be deaf, blind, and dumb to miss this, but students simply seem to think they will be the exception. So, more information might not help. On the other hand, some schools do a LOT better than others to place students in academic positions, and students deserve to have that information--certainly the fact that some institutions have a 90% placement rate and others less than 25% is a relevant difference that should be made known (and any responsible school ought to have that information publicly available).

Gears said...

I think it's a good idea to track where your students are and I think it's acceptable to show your current students that this is where your previous students went.

A more difficult thing to track would be the desired career track versus where they actually ended up.

For instance, showing that 50% of your students wanted academic jobs and 50% got academic jobs is great. Unless, it was the other 50% that got the academic jobs. I know that's a stretch but a category of "those who got the job they wanted" tells me a lot more than just an overall percentage.

Also, for those that want academic jobs, you might want to stress the trials and tribulations for them. A few weeks ago, I had a blog post (linky) stressing that you have to know what it takes to make it in academia and your CV has to show that. So many graduate students want to go into academia. But most of those have no idea about what it takes.

Plus, academic jobs are very fickle things and even if someone's fully qualified, it may not be in the cards for them due to timing. That can hamper your statistics...

mOOm said...

Students can't educate themselves if they can't access the data. If you have the data I think it is valuable to share it with the students whether the picture is good or bad. The tricky bit might be context. Preferences of your graduates will affect what career path they take and people who see the data need to understand that.

Pippin, the Gentle Pup said...

I think they should be shown. Interestingly, they are fairly similar to the data from my own department, which is a social sciences dept. frequently assumed to have grim job prospects. I was a grad student in the 90s and started paying careful attention to this kind of thing. In my experience, including in this terrible market, I'd say 95% of the people who want academic jobs--and are willing to move and wait a couple of years for them--get them within a couple of years. Not always pleasant as not that many get one in the first year and there are often several years of moving around, but what I tell prospective graduate students is that people who prioritize an academic job over other things (like where they live) generally get them. It's a good way to have the discussion of priorities and of how it's unlikely you can have exactly what you want in terms of "work/life" In most of the cases of my own students who decided not to pursue an academic job, it was because of family and a spouse perceived to be immovable. I haven't had a single student who pursued an academic job and didn't get one.

Anonymous said...

Should be shown- no matter the result. Possible before you even sign up for the program! I'm so impressed that you would choose to talk about this so openly with your research group. Please keep it up!

My adviser seems to think (although it's not clear- he is pretty guarded about what he thinks) those of us who don't stay on the academic track were a waste of training, even though he his most highly successful grads are not in academia. And if that's not actually the way he feels about it, I really wish that the person in my life who has been so formative of my science so far would want to be influential in my career search.

GradStudentAbroad said...

Show the data, please! I wish my department had data of this sort. It is very relevant to know what students from the same program have done in the past, since not all graduate programs will have an equal track record. Graduate students are adults and can draw their own conclusions, however, they will be able to draw better conclusions when they have real data, instead of just anecdata (or guesswork).

Giving people real information helps them calibrate their expectations realistically, making it less likely that they will either later be disappointed or that they will undersell themselves and fail to reach their potential.

Also, there is evidence that giving people accurate information about typical job prospects (and salary ranges) might reduce inequities that result from women (and maybe also minorities?) low-balling themselves. (See for instance the book Women Don't Ask, by Babcock and Laschever)

Brian Gilbert said...

I'd vote for the data being shown. Part of the discussion with prospective students/postdocs would need to reinforce that there is no assurance - ultimately their success in a job search depends largely on them. The mentoring they receive can help greatly though, especially in thinking about where they are likely to find the most success and satisfaction.

Anonymous said...

The breakdown of these data into "academic" versus industry / government is misleading. In FSP's case, perhaps not purposely so. As a "research associate" (POSTDOC) at a University, my graduate program proudly places me in the category of an academic position. In reality, a postdoc has a very small chance of achieving their (usual) goal of a TT faculty position.

unlikelygrad said...

I agree with most of the other posters that the data should be shown no matter what, and agree with Gears that you should show what percentage of students got the job they wanted.

Also, I think it would be interesting to see the overall data vs. data for students who have been out 5+ (and 10+, if you have that data). This would tell me if people "wandered" a bit before landing on the perfect job, if people get academic jobs (adjunct? visiting assistant professor?) that only last a few years before they end up in industry, etc.

Stu said...

I vote: show the data.

Lindsey Kuper said...

I'm curious -- although 98% of graduates who wanted an academic job got an academic job, you didn't mention what kinds of academic jobs those were. Are they tenure-track professor jobs, non-tenure-track jobs, postdocs? Those are meaningful distinctions.

Another thing: how do you determine who wants what kind of job? Is it self-reported or determined based on the positions they apply for? There exist people who might want a certain kind of job but assume they're not qualified and never apply.

Female Science Professor said...

The group talked informally about some of the issues raised by commenters. For example, of our graduates with PhDs in 2009-2010, some are already in faculty positions, and the others are in their first postdoc position (but most of them have been interviewing for tenure-track positions this year).

Our alumni/ae dataset shows the name, PhD date, and academic position (location, title) of each person. With a few exceptions (that 2%), all pre-2009 (and some 2009) PhDs who wanted academic positions are tenure-track or tenured professors.

I am not inclined to go any further with the data in terms of whether PhDs got the positions they wanted.

In my own case, I ended up at a type of institution far different from what I (thought I) wanted. So: did I get the job I (thought I) wanted? No. And that seems to have been a good thing in my case.

If we ask instead 'Are you happy in your job now?', even that is a meaningless question for the purposes of using this dataset as an indicator for current and prospective grad students. There are too many variables in the happiness equation, and only some relate to the academic job market. Do they like where they live? Do they like their department/colleagues? Are they in good health? Is their significant other happy with the location/their career? Do they have enough cats?

Miss MSE said...

I think that data would certainly be insightful for graduate students, and I would personally find it very helpful in setting my own expectations. If I see my professor has produced a number of people succeeding in what I want to do, I can feel confident that he will be able to help me achieve those goals. If that's not the case, I know I have to be more proactive in finding other mentorship that will help me reach those goals.

Female Postdoc said...

I think the information is important to share. If you feel it necessary, you can provide some disclaimer that the following information is a snapshot in time and is not meant to reflect any of the individual career paths or goals.

I think it is also important to list research group alums who are no longer in science, as their reasons for leaving are likely as varied and winding as any other person on the list.

Pippin, the Gentle Pup said...

Well, rarely enough cats.....

Thinkerbell said...

I think the more data is out there, the better informed students are. Which is important. Would it make a difference? I doubt it. If you really want a job in academica, you will convince yourself that if you just stick it out long enough, you will make it, no matter what the odds. Until you realize there are a lot of really good people left by the time you get to the top of the pyramid. What might be more of a help than discouraging or encouraging %%, is to see the full spread of different professions people ended up in. Non-tenure-track career choices deserve to be looked upon as favorably as the road to full professorship. More support to openly promote these, might actually motivate some people to investigate those possibilities more seriously and decide that they might be a better fit. Pippin e.g. mentioned that everyone who wants an academic job got it as long as they prioritize it over everything else. Which might not be the right life-choice for everyone. And frankly, it's little tidbits of information like that that students early in their career are not aware of as being major factors in your future career... not the statistics per se!

Female Science Professor said...

I hope it was clear from the post that the database includes all graduate students and postdocs from our group over the past 20 years; it doesn't exclude anyone.

The 4 science professors in our research group are all married with kids, so I presume that our advisees can reasonably infer that it is possible to have a career and a family.

RTG said...

Definitely show the data. The more information about PhD outcomes available to graduate students, the better. And as scientists, why would we want to hide data that doesn't conform to our assumptions about the job market.

In fact, if your group has much higher rates of *permanent* academic placement, that's good information for a graduate student to have. Perhaps it is something you and your colleagues are doing, or just your area of research. I think students would do very well to look at where prospective advisers' students are now, and factor that into their decision-making.

In general, departments, universities, and funding agencies do a terrible job of providing information about job outcomes and career paths. I applaud your efforts to track your own students, though overall that is not enough. U.S. funding agencies are obsessed with funding soft money positions without any regard to where people who have done three postdocs will end up. The incentives are dangerously skewed. But if we can't change policy, perhaps we can at least inform grad students to the point where they will know what non-academic career paths are out there (including those to which a PhD is relevant) and start exploring how to prepare for them early.

Ilya Gekhtman said...

So in mathematics, the Mathematics Genealogy Project:

genealogy.math.ndsu.nodak.edu/

lists (almost) every graduate student in the history of modern mathematics by their advisor, institution and year of graduation.

A simple google search will find any given advisor's/departments track record for placing students in academic jobs.
Maybe other fields should start such a database?

Anonymous said...

As a former member of your research group, I am very happy in my (industry) job.

MamaRox said...

Only show the data to incoming students if they're really spectacular.

If they're not near-perfect, how relevant are the data to me as an incoming student?

Yes, of course, I want to go to school somewhere where I believe I will have a good chance at getting a job later. If it takes me 7 or 8 years to finish my degree and then I'll need to do a postdoc or 3... it seems hard to reasonably predict the job market 7-14 years ahead.

Thus, bad % might be due to external factors. And, I wouldn't read too much into them.

BUT, your % are so good over 20 years! That may say something wonderful and positive about how the lab group prepares students in general.

gasstationwithoutpumps said...

I assume that FSP is asking about reports like
https://bme.soe.ucsc.edu/graduate/alumni
which list all graduates of a program and where they ended up.

(Note: such lists rarely discuss the students who quit, die, or are kicked out before finishing.)

Anonymous said...

It sounds like you have been extraordinarily successful in placing your academia-oriented trainees. Is this because

a) You are very, very careful about who you accept in to your lab?
b) You are a fantastic mentor who discourages those who, while they seemed awesome on entrance, did not really blossom from aiming too high?
c) You are just a phenomenal scientist who is so generous with her ideas and genius that it cannot help but rub off on all those who work with you?
d) ALL OF THE ABOVE.

No, seriously, I would like to know: where do you think this success comes from?

(I know of phenomenal labs at NYU and MIT in my field that are nowhere near 98% placement.)

Ms.PhD said...

These kinds of data should always be made public. Graduate programs should be held accountable. They should lose accreditation if they can't help people find relevant work.

Hooray for you that your rates are so high. I suspect that it is in part a positive feedback loop, where once you reach a critical mass, your former success stories are part of a network that also helps future students find relevant career opportunities.

It's too bad that not everyone can have access to great mentoring & career networks. And that we don't seem to have much to offer students and postdocs who didn't know how critical it was to join the right club from the very beginning.

Now I need to go look at what you posted Friday to see if I have any questions re: specifics or implementation. E.g. I've noticed that most places include postdoctoral positions as "employment" when they talk about graduate student careers. However, a postdoctoral position is NOT a career and may be in itself a dead end.

Also, I know quite a lot of PhD holders who ended up in "relevant" career paths that aren't so great, e.g. teaching the occasional community college course but not making enough money to survive; working as perma-docs under some other PI; or working as technicians.

Anonymous said...

Show the data but definitely add (either as data or through discussions) what those people have done to get to where they are - i.e. 98% of those that wanted academic jobs got academic jobs and this is how they were publishing (rate and example journals), how they were getting research money, teaching experience (if any), etc. Those that went into industry did so with X papers, high level skills in A,B, and C. I've definitely know grad students who've entered highly successful labs and assumed that they were all set - they could coast because all the students in front of them had gotten good jobs. They didn't always realize what those folks had done to get those good jobs - other than be smarty-pants undergrads that got into good labs. Although I'm sure as an adviser you'll push them to do these things, it's helpful to make this explicit early on. I think this is particularly true for students coming from smaller schools or from older undergrad advisers where the old-boy network had such a big role in finding jobs. I've been assured by new grad students that their admittance to the fancy dept we were in with a famous adviser was all they needed to get a great job. No matter how I tried to push the point that their performance was what would be evaluated, I have had some turn a deaf ear. Hearing that message from our adviser earlier would have helped some of them get on track sooner.

Anonymous said...

The fact that whether the data should be revealed at all is not dicated by transparency concerns, but by whether they will paint a rosy picture of your gang or not is very interesting.

It shows that academia is just that...a business cartel like so many else. As such, like the credit card business, there should be more govt requirements asking you to come clean.

prodigal academic said...

I definitely think the job data should be revealed. I wish I had access to that sort of information when I entered my program (back in the pre-Google dark ages). The truth is that most of the people in my PhD program who wanted academic positions at graduation got them (not all at highly ranked research universities, obviously). I don't think this is all that unusual, since most grad students in my field (in my experience) don't want to stay in academia.

I also agree with you that measuring aspiration vs placement is not useful. I NEVER wanted to be a professor when I was a student. Not when I entered my program, and definitely not when I graduated. But here I am, years later on the TT by way of a National Lab. In my experience, many students have no idea what jobs they really want to do until they have some actual experience under their belts.

Anonymous said...

Echoing the majority vote that the data should be shown. With the past students' permission, one could also show the exact jobs and positions they have.

It would also be helpful to show other statistics about the lab or program -- attrition rate, graduate funding policies, etc.