PhD cohorts in academia

A common discussion about biomedical PhDs is whether we (as a society) are appropriately setting up our students for future career success. The general narrative is that we (as a whole) tend to train them for an academic system, even though the vast majority of biomedical PhDs will leave academia during their careers.

I’m completely in agreement with this narrative. I think we essentially have an “opt out” system, where the default training is for continuation in an academic system, and people need to defy the path that is implicitly dreamed up for them. It certainly doesn’t help that the vast majority of PhD mentors are (1) biased for and largely only knowledgable of academic systems (because that’s where they are), (2) exhibit survivorship bias, since they represent the most motivated individuals in the academic research (and not reflective of a full spectrum of the PhD trainee pool), (3) come from a different time when the academic climate was different, and (4) may even be conflicted, as *perceived* reputation as a great Principal Investigator / Scientist is associated with the number of other great scientists you help train (successful people anywhere are “good” for reputation, but I think there’s still a bias for successful people in academia).

I’m all for an “opt in” system for academia, where we teach students to be critical, logical thinkers, writers, presenters, and project managers during their PhDs, but doing so while consistently framing these skills in their potential applications in non-academic sectors. My general opinion is that PhD students should only do Postdocs if they (1) need additional training to pivot to a career of their choice, or (2) literally cannot imagine doing anything else other than academic research, thus fully “opting in”. Yes, people are in charge of their own fates and ultimately this responsibility falls to them, but as a society I think we may as well optimize situations we have control over and make it easier for more people to more easily get to where they will be most happy and productive.

At the global level, these stories are based on abstract generalities from (very useful and needed) studies surveying the general PhD-earning population; again, very useful for informing society, but less relevant when considering a particular individual at a particular point in time. Let’s pretend I was starting a PhD right now. I could use current me as a template, but that analysis suffers from small sample sizes (what happened to current me could have been massively influenced by random-chance events of large effect sizes that would not occur again, and thus not “reproducible” if repeated many times).

Seems to me that a happy medium could be seeing what happened to the cohorts of PhD students in my graduate program following their PhDs. I started with a directory of students I as given around when I started the program. To this list of names, I wrote down when they graduated (based on dissertation defense announcement emails). For the vast majority, since we all grew up in the internet age, I was able to get dates of job changes based on LinkedIn profiles. I ended up with 50 names spanning Harvard Virology incoming years 2000 through 2007, with about 40 data-points out to year 7, and 30 data-points out to year 10 (some people took longer to graduate, so fewer data-points for them following PhD so far).

About 25% of the trainees left academia right after graduating. It’s largely a one way street (most people leave academia, and once they leave, very few people return), so following that initial set of people that leave, there seems to be a pretty steady, slow decline of individuals in academia. The first decade after the PhD will certainly be the most dynamic, as people realize they aren’t able to find jobs as professors or permanent staff, so I imagine the curve will flatten out somewhat at some point (when the only people remaining are essentially tenured professors who will likely remain in academia through most of their lives), though I don’t have data far out enough to see this yet.

So there you have it. If I was starting the same grad program right now, I should go in knowing that — with no other prior information about myself specifically — there is a roughly 60% chance of me being in government, industry, communication, or education (High school or lower grades) within the decade. After a quick google search, these numbers seem reasonably in line with more global numbers of ~ 60% and ~ 70% seen in other studies. Arguable, graduating from Harvard Virology may give some more advantage / leverage in following one’s career of choice, but that doesn’t seem to result in drastic differences from the general population in terms of staying in academia.

My opinion on this should be clear by now, but I still want to reiterate it: “alternative” careers in science are the norm and are incredibly important, fulfilling, and *tractable* career choices, and how we train students should reflect that as much as possible.