Why is a PhD this long and hard?

In most of the basic science departments in the United States (and in many other countries), a PhD takes between 5 and 7 years. That seems awfully long, doesn’t it? My five-year PhD was actually less than the average time taken to finish a PhD in my department. What’s interesting is that I’ve heard many PI’s say that it takes too long, and that “it used to take less time when we did our PhDs. I don’t know what’s changed” when in fact they are part of the “problem”. So, here are my thoughts on why it takes this long.

I don’t think it is simple. I certainly don’t think that the quality of students has decreased that much (at least in the premier research universities). The students I know have all been motivated, and have come in with some research experience (so weren’t complete newbies in the lab). They all work hard (most students work 6 days a week, 10 hours a day, juggling experiments with courses, journal clubs, assignments and whatever else). There may have been some wonderful students 20 or 30 years ago, but that probably doesn’t explain why PhDs have expanded from 3-4 years to the present 6-7 (that’s a doubling).

I would put down four broad reasons why it not takes so much longer to finish a thesis (which do overlap). The first is the structure or system itself, and the “requirements” for a PhD. Basically, there aren’t any clear expectations or requirements for what constitutes a PhD. When a student joins a program he/she isn’t quite clear on what all is required to have a thesis. Most good departments have some kind of unwritten “publication” rule, and students are expected to have at least a couple of first author publications in good journals (or one “stellar” paper). But that is a fuzzy rule, and much depends on how the projects unfold, and what their bosses themselves think is needed. I’ve known students who have graduated without a single first author paper and others who still are in grad school (even though they seem ready to graduate) after publishing some high profile papers. It’s a crapshoot. Part of it is because the expectations for the amount of data that goes into a paper has gone up, and substantially more work is required to make a complete paper. While some experiments are certainly easier due to easily available reagents, I don’t think the availability of improved reagents and tools is proportional to the amount of work that goes into a publication. A secondary factor might also be the massive increase in scientists, which has set the expectations or requirements from postdocs to be higher, and this trickles down to students as well. But a PhD is no longer only about coming up with a good hypothesis, and systematically thinking through and testing it, while acquiring a fairly thorough knowledge of the field.

The second is the more or less mandatory “rotations” that the student does before selecting a lab. The idea behind rotations is to give a student the opportunity to briefly work in 3 different labs for short periods of time, so that the student can figure out if s/he would be a good fit in the lab, if the PI wants the student, and if the research is exciting enough. However, most schools have approximately 3 month rotations (basically a full year of rotations). If the goal is only to give a student an opportunity to get a feel for a lab (without too many expectations of producing data), then 4-5 week rotations should be sufficient. Still, this is a more minor point, since I do think the rotations help students make better choices. However, there should be options for students with very clear research goals or ideas to avoid 3 mandatory rotations, particularly if they are sure they like the lab they have first rotated in, and feel it matches their goals.

A third (minor) reason PhDs drag on now is the non-research requirements that a student “has” to undertake. There is a substantial amount of mandatory coursework in the first two years, along with a few qualifier exams and whatnot thrown in. So, if you add that to the time spent in rotations etc, a student has barely done any serious experiments for about 2 years (discounting say a couple of months in summer). I actually liked courses, and took quite a few throughout my PhD, but I was able to multi-task, and get quite a bit of research done in the lab. But I do know that many students cannot focus on their research and simultaneously go through the grind of coursework and qualifiers. If I had to quantitate what this process does to the duration of a PhD, I’d say this adds another 6 months of time to the whole process (which isn’t that bad really, and can be useful if you get something out of the courses).

Which brings us to the final, and by far the most important factor that influences a PhD. The advisor. I have absolutely no idea what PhD advisors were like 25 years ago. But I do know that while there are still many good mentors out there (I had a great one), there are plenty of terrible ones. And a terrible mentor does not mean a terrible scientist (often it is quite the contrary). There certainly has been a huge explosion in the number of labs and PIs out there especially since the early nineties. Before that, there were perhaps a tenth as many scientists in research (particularly in the biological sciences). More importantly though, while I don’t know what the expectations from mentors were in the old days, I do know that there are NO real expectations from a PI as far as their graduate students go. Sure, there are department requirements and some committee meetings and suchlike. But those can be negotiated without much difficulty. There is no real incentive for investigators to actively mentor their students well, and importantly, there is no demerit if they are terrible. If an assistant professor is up for tenure, departments do look at her/his record with graduate students, and if they have managed to get out a couple of PhD students in their 5 years before tenure it is good for them. But, in most places, that is at best a secondary consideration for tenure. What matters is if the PI has managed to get a few grants, and a few high profile publications. After tenure, and particularly if the investigator is famous, there is no system in universities which really takes a look at how the students in that investigator’s lab fare. Neglecting a student isn’t noticed or penalized. As long as the investigator hasn’t done something seriously bad to the student, it’s all ok. This means a PI doesn’t really need to monitor a student’s progress, or sit down and think hard whether the student has a viable project or not, and can also tempt an investigator to make the student continue on a crazy (or dead) project far longer than they should. This is particularly true if the investigator doesn’t need to pay the student (who may be on some training grant or fellowship), or if the PI is so flush with funds that it doesn’t really matter. Finally, since there is no incentive for the investigators to get students out, they often appear to keep the students (particularly the productive ones) longer than they really need to stay, to get more out of them. That work might make the thesis thicker, but was it really needed for the thesis itself?

While this might sound like a critique of the graduate school system in the US, it is not. I have no hesitation in saying that the quality of PhD education in the US is far more substantial and comprehensive than anything else I’ve seen, particularly in the breadth of knowledge that is/can be acquired. I wouldn’t have come here if it wasn’t. However, while good, it certainly does drag on longer than it needs to (particularly for good, motivated students). Given that postdocs anyway not take much longer, at least the duration of a PhD can be kept to around 4 years. But that cannot happen if the key problem, that of the investigators themselves, isn’t tackled.

Many of the readers of this blog are (or were) grad students. What do y’all think?