Ph.D. Completion and Attrition: Analysis of Baseline Data (2008) [pdf] | Hacker News Reader

Ph.D. Completion and Attrition: Analysis of Baseline Data (2008) [pdf]

google.com

41 pointsvinchuco30 comments

Mods, can you please update the link to go directly to the pdf (http://www.phdcompletion.org/resources/cgsnsf2008_sowell.pdf) rather than going through the google redirect?

dang, do you hear this?

Although these general numbers won't be a surprise to someone who has experienced a PhD program or is well informed about them, I do think the detailed data is useful.

Congress recently considered legislation that would automatically grant a green card to anyone who has completed a PhD in certain STEM fields (not all branches of science qualify). An MS would also count under some circumstances. The logic behind this is that the people who receive these degrees are very talented, there is a terrible shortage of them, and that it makes no sense to force them to leave the US.

I do agree with point one and three. People who get PhDs in hard sciences and engineering are very talented, and it is truly bizarre to force them to leave the US. I'm also in favor of the green card, because truth is, people with these degrees who want to stay tend to be able to, and I think it would be better to grant them full and free membership in the US labor market, rather than forcing them to work under more restrictive conditions.

However, I also realize what this legislation means - that PhD programs, which I consider to be fairly abusive of their students, would now have an amazing weapon to coerce in their arsenal - they would now be institutions with the power to bestow green cards. A RAND study found that the aversion to STEM graduate degrees among people who already have US residency and work rights is economically rational when you consider other options available to elite students. [1] The "shortage" is simply a rational response to long completion times, high attrition rates, and relatively poor employment prospects (again, compared to the outcome for shorter, more certain degrees with vastly lower attrition rates). Keep in mind as you read this PhD data that elite law and medical degrees typically have attrition rates of below one half of one percent. Pretty much everyone who wants to finish does. I understand there are some differences here (half a JD isn't useful, half a PhD in CS could be). But we're talking about a stupendous difference in attrition rates, in a field where congress is taking legislative action to correct a severe shortage that supposedly threatens US competitiveness in high tech!

Before we grant this power to universities, they need to answer some tough questions. Why are general attrition 50-100 times higher than they are in professional degree programs? Why is it so much higher for women and some minorities? What are you going to do to remedy this?

[1] http://www.rand.org/pubs/issue_papers/IP241.html

I would respond with a question, why do you think it's appropriate to compare a vocational training program such as an MD/JD/etc to an academic research program? They have very different goals.

It may in fact not be a valid comparison. But if we're going to speak of a "shortage" of people going into graduate STEM programs, we can't discuss that in a vacuum - we need to consider the other options available.

Here's a link to the RAND study (should have included this in the original post)

http://www.rand.org/pubs/issue_papers/IP241.html

Which does contain a discussion of whether the comparison is appropriate, as well as a (short) justification of focusing on the doctoral level programs. Ultimately, though, their justification is largely market based. A couple of relevant passages:

"Alternatively, if the rewards of other careers—perhaps clinical medicine, law, or business—are higher and are growing relative to S&E, and if it is instead the costs of training for a job that are growing for S&E, then there is no shortage of scientists and engineers in this important fifth sense. Indeed, in this latter case, the “shortages” thatothers discern may well look more like discouraging surplusesto young people considering career choice."

"More young people today may arguably enjoy doing science or engineering than actually plan to prepare for such careers. Instead they may choose a professional degree, but only reluctantly. In a market economy, even one characterized by rigidities, regulations and unequal opportunity, qualified people tend toward career paths whose rewards and satisfactions are becoming more attractive and/or whose preparatory costs are becoming less onerous."

As an african-american math ph.d. dropout, nice to know that 63% of people like me also didn't finish. Oh wait, it's awful. At least I'm not alone though.

That certainly jumped out, as well as the much higher female drop out rates. Shameful!

Edit: To be clear, I put the vast majority of the blame on the advisor, not the student. I would not have made it without the remarkable support provided to me by my mentor. Conversely, I saw much stronger researchers leave for lack of a good advisor.

   I put the vast majority of the 
   blame on the advisor

One of the social functions of a PhD supervisor is a a'lightning rod' that enables the failed candidate to blame the supervisor, rather than themselves. Highly convenient ...

Two different words are being used here, but they are not interchangeable. The words are mentor and advisor. You are correct in that the Ph.D. is supposed to be capable of original research. An advisor who is a mentor is one to takes the raw talent of a beginning candidate and grows it into an independent researcher. Don't underestimate your importance here.

Seems to me that the entire blame should go to the system itself. The system is broken, and everything you're seeing here is a symptom of this disease: lack of good advisors, exploitation of adjuncts, lack of tenured positions, etc. But there doesn't seem to be much agreement about what exactly is wrong with the system and what can be done to fix it.

Or maybe research level mathematics is actually really difficult.

There's difficult because "the topic is hard" and difficult because "there are unnecessary social barriers put up" and I've seen a lot more people affected by the latter than the former.

(source: am a math phd)

I agree there may be some social barriers, but, for instance 4 of the 10ish math students that started the phd program in pure math at the same as me were filtered out by the written exams, another was pulled to industry in her second year, and another left due to (I think) not being able to finish their research before his funding ran out (had less due to entering with a masters). I could be forgetting someone, but I think I am the only one on the pure side, finishing at the 5 year mark in my program, and I put myself through some serious sleep deprivation to do it. The applied students seemed to do slightly better.

This doesn't have anything to do with mathematics, it has to do with the university system. It isn't just math departments that are having these problems, they're across the board.

What made you drop out specifically? Wish there was at least one black fields medal recipient :'(

I failed the preliminary qualifying exams. So technically I flunked out, or was "asked to leave".

the opportunity cost of doing a PhD is, more and more, becomingly increasingly beyond what is tolerable for many academics.

tenure lines are evaporating, adjuncting is a nightmare, postdoc comp is pitiful... Meanwhile PhD students are competing for ever smaller slices of ever smaller pies in terms of fellowships.

meanwhile, universities are being taken over by MBAs and running them like corporations. many people (not just PhD students but staff, etc.) eventually throw in the towel and say 'fuck it, i'll just go work for an actual corporation where i'll make a decent salary.'

Working in the commercial sector for a PhD isn't "throwing in the towel".

Academia has always been political and competitive just from the limited economics of it, so don't construct an MBA scapegoat for "ruining" it.

i think the argument that the economics of building a career in academia have gotten worse over time is a defensible one. the data concerning tenure line trends ([0], [1]) don't necessarily tell the entire story, but it's rather compelling in and of itself.

[0] http://www.aaup.org/sites/default/files/Faculty_Trends_0.pdf

[1] http://agb.org/trusteeship/2013/5/changing-academic-workforc...

as for throwing in the towel, i didn't mean that in a disparaging sense. just that, for some, leaving academia is a decision made with a measure of regret. for some, it can involve a sober reassessment of the opportunity costs involved -- sometimes these costs only unfold/show themselves along the way -- in which one simply can no longer justify (to oneself) continuing on one's current path.

There's definitely a little bitterness as a molecular biology PhD right now. The senior and emeritus faculty got in during a wave of setting up departments in the 60s and 70s. Jobs really were almost growing on trees, and the career trajectories they took are just so beyond what's possible today, well it does get a little depressing. That said, I think most people recognize that that was a singular era.

What universities are you talking about, being taken over by MBA's?

In my experience, it's pretty much academic faculty with decent track records that move on to the uni Dean/President path.

While 2008 is somewhat old data, a cumulative 10 year phd completion rate of 57% sounds about what I would expect. That implies a cumulative attrition rate of 43%, which puts it within striking distance of say, the 75-80% attrition rate associated with Navy SEAL BUD/S.

The attrition rate jump around years 2-3 is almost certainly due to failing the qualifying exam, which in my experience is by far the largest eliminator.

At least at some schools, a spike at 2 years or so can also come from "mastering out". In my PhD program we get an MS after completing the required classes just by filing a form (and I think this is common). Someone who figures out quickly that they don't like research all that much is well advised to at least make it to that point.

True! Though not all schools are like that. My brother's made you do an extra masters thesis on top of the dissertation. Unsurprisingly, most PhDs then just have the doctorate, and not both. I also concur that the drop rates at 2 years are due to failing the quals. The more interesting part is then the 3+ year rates, as those are really the people that dislike research and not just the ones that can hack it.

Yes, mastering out is a good thing; they've accomplished something and didn't waste extra time on what might be a futile endeavor. In my program you had to throw together a thesis, but it was something available to anyone and not hard to do. I know a lot of 4th and 5th year students who wonder whether they should have done that, I know I did.

They're data only shows attrition after 3 years, but it doesn't change much between 3 and 10 years in, so it definitely looks like the attrition comes at the beginning of the program for most people.

Interesting that Computer Science is the lowest of all the engineering and science degrees. Any idea why that is? (Higher level of programming required as a base skill? Too many different skills to master? Too many of the top undergrads go to industry?)

I assume you are looking at the table on slide 7. I think that is the percentage of all who completed relative to those who started it, so whether many CS undergrads go intro industry (which is most likely the case) or not is irrelevant. However, I can offer anecdotes from my experience (recent PhD in CS). First, it is usually easy to find good jobs in the industry while you are doing your PhD. I think half of students in my department that did machine learning and computer vision didn't finish their PhD and went off to work at Google and the likes. It is very tempting, because the hours in a PhD program are long, stress is high and money is never good. And many CS PhD end up working in the industry anyway. Second, writing a lot of code isn't necessary if you're doing pure theory --- but even then you might be making small prototypes or using proof assistants, so a no-programming research in CS is rare. Whether you need to master many new skills depends on your background and what you want to do for your dissertation.

Thanks for the reply.

To clarify on going to industry... If you are a top student with a bachelors in a field like Physics or Biology - there's aren't as many post-undergrad options as the top CS student. Therefore, the top Physics and Biology majors are more likely to head to grad school. Those top students are more likely to finish on time.

Again - just a hypothesis.

I think your point about mid-Phd CS folks having options is stronger. If you're 2 years into a Biology Phd program, your options are limited. If you're 2 years into a CS Phd program, the sky is the limit.

I hear you on CS theory. My impression during undergrad was that the grads were doing a lot of programming in the coursework, and many beyond.

My wife completed her major comps today. We are separated and will likely be divorced soon. But looks like she'll finish. Yay...

<austinPowers> What does all mean Basil? </austinPowers>

Reply on news.ycombinator.com