Posted at 11:03 am

PhD Dropouts: Smart or Stupid?

What's up with PhD program attrition rates? Are doctoral programs simply accepting anyone, even nitwits who can't hack it and end up dropping out?

The data in the graph above comes from a study spearheaded by the Council of Graduate Schools [source]:

  • Physical Sciences doctoral candidates (including Chemistry, Computer Science, Math, Physics) drop out the quickest (8.2% after year one) and exhibit the highest overall dropout rate after ten years (36.9%)
  • Humanities (including English, History, and Philosophy) and Engineering doctoral candidates are the slowest to drop out (5.5% after year one) but after ten years, it's Life Sciences doctoral candidates who exhibit the lowest overall dropout rate (26.2%, vs. Humanities' 31.7% and Engineering's 26.7%)

But are these numbers necessarily alarming?  And does this mean that doctoral programs are admitting Zoolander-ian kids who can't read too good as some sort of cross-subsidy to ramp up Ramen noodle sales?

Doubtful. Certainly in the years leading up to 2009, there was intense industry interest in hiring individuals with exceptional quantitative skills. These are the kinds of folks who tend to be found studying things like Computer Science, Math, and Physics at the doctoral level. It doesn't take a PhD candidate in Math to figure out why someone might drop out of a doctoral program if a quant-based hedge fund in Manhattan wants to throw them a $250K base salary.

Since I'm too cheap to actually buy the full $30 analysis from CGS, I can only assume that their study wasn't ultra-rigorous with respect to people who might quit for a while but then return years later.  For example, a woman who lived in my college dorm left her doctoral program only to return a decade later to a different doctoral program. "Slightly different field of study -- both sciences, but Ecology and Evolution (then) to Genetics (now)," she told me over Facebook.

Yet another data point, one of my high school friends was a Math PhD dropout, but only for a while:

I was ABD in one area of math, left for five years to start a career and have kids (needed the health insurance and wasn't excited enough about my area of study). After having to teach a class in Discrete Math I decided math could be interesting again so I quit my job, sold my house, and moved my family so I could be full-time. Just got my PhD in May in a very different area of math.

He's now a tenure-track math professor.  Would this chart consider my dorm-mate a PhD dropout?  How about my high school friend? I'm not sure. (If anyone has read CGS' full analysis and can report on this, please log-in (you can easily link up with FB) and share your comments below!)

Conversely, while prospective math PhDs can find themselves batting away job offers sans doctorate, the Life Sciences doctoral candidates (Bio, Cell Molec Bio, et cetera) may have more confidence in their ultimate employability if and only if (or 'iff' for all you math nerds) they complete their degree.  Big pharma, small pharma, and medium-sized pharma are all ready to hire these folks, but only once they finish their PhD. So perhaps this market force helps illuminate why Life Sciences doctoral candidates are the least likely to drop out.

How do dropout rates compare by institution type?

The highest overall doctoral candidate dropout rate when segmented by institution type is exhibited by Physical Sciences candidates at public schools (40%), whereas the lowest overall doctoral candidate dropout rate is exhibited by Engineering and Life Sciences candidates at public schools (25%). The highest dropout rate at the private schools is exhibited by Humanities doctoral candidates, at 34%.

Any thoughts as to why this might be? Is there something about a doctoral program at a private institution that differs from that of a doctoral program at a public institution?

The final way in which CGS shares their results is by segmenting dropout rate by class size:

There's no clear relationship between class size and dropout rate, though Physical Sciences PhD candidates in classes with more than 15 students have the highest overall dropout rate when split out in this way. Engineering PhD candidates in classes with seven or fewer students have the lowest overall dropout rate ...

... but my sense is that overall, whether an institution is public or private or whether your class is big or small isn't necessarily what drives a doctoral candidate's decision to drop out. I'd be intrigued to see more research and analysis conducted around

  • market forces that shape completion and attrition rates, as alluded to above with quant funds and big pharma
  • whether dropout rates differ as a function of age at matriculation

Did you drop out of a doctoral program? Why?

What do you think informs a doctoral candidate's decision to walk away from their prospective PhD?

Are PhD dropouts, in your opinion, making a smart move, or a dumb one?

Login to comment on PhD Dropouts: Smart or Stupid?

Popular PhD programs at


To leave a comment, please register for
Login | Register

You may also be interested in one of these Sponsored Listings: