Tuesday, April 13, 2010

Astrophysics as a Career: Where Do Professors Come From?

My last post in this series tackled the question of whether your PhD institution determined your career trajectory. The paper I cited by Gibson et al indicated that overall the effect was smaller than one might imagine, but that if a permanent position was the goal a PhD from a prestigious university does help. They did this by tracking the graduates of several PhD-granting institutions - something I’ll call forward career tracking. In this post I’m going to present some research that I have done recently that tries to address the same question by what I’ll call reverse career tracking.

Reverse career tracking means that I have found the PhD institution for faculty members at randomly selected universities and college in the US. To do this I randomly selected colleges and universities from the Carnegie Classifications of Institutions of Higher Education. To help provide a somewhat even sample I selected physics departments from the various Carnegie classifications to match the distribution of faculty between doctoral, master, and bachelor granting physics departments, as reported by the American Institute of Physics.

I then used department webpages and AIP surveys to find the PhD institution for each faculty member in each department. I did not include individuals that earned PhDs (or equivalents) from foreign institution, faculty members with Masters degrees, or those for whom no information could be found or multiple institutions were listed. I then took that data and found how many current faculty taught at universities in one of four categories (again using the Carnegie classifications): Doctoral-Very High Research Activity (DV, previous known as R1), Doctoral-High Research Activity (DH), Masters (M), and Bachelors (B). Universities with no physics department were counted with the bachelors granting departments as long as they had a major that included significant physics content and at least one identified professor of physics.

Finally I assigned each PhD source institution a score based on an average of 101 minus their rank according to the 2009 US News & World Report of graduate program in physics and 101 minus their 1997 National Research Council rank. Programs ranked in the top 100 of one system but not the other were given a score of 1 in the missing ranking system. Thus the higher the score the better ranked the institution. Overall these two rankings correlated extremely well, however the NRC rankings had an average ranking 13 points higher than the US News rankings, with most of the discrepancy coming in for rankings above 40 in either system. The correlation is shown here:
Finally I binned the data into average ranking ranges and did a standard least-squares linear fit to the binned data, including calculating the uncertainties in the best-fit parameters. In graphical form, the results are:

For DV faculty there is a strong preference for faculty to have come from high ranking institutions, in fact the slope of the linear fit is positive at the 8-sigma level. However as we look at the linear fits to the DH and M, we see that while they all have positive slopes, they are all consistent with no slope at the 2-sigma level. The linear fit to the B data is consistent with zero at the 1-sigma level.

It’s no surprise that having a PhD from a big name university helps one get a job at a big name university. Over 50% of faculty at top research universities received their PhD’s from institutions ranked in the top 15. However there is little to no evidence that getting a PhD at a highly ranked institution matters in getting a faculty job at anything except a highly ranked PhD institution. With almost half of physics and astronomy faculty in the US at Masters and Bachelors granting departments, if you want to be a professor and you are willing to take on some additional teaching responsibilities and give up some prestige, where you do your PhD matters little.

The Astrophysics as a Career Posts:

1 comment:

  1. Fascinating, Nick. I'm seriously geeking out, even while I recognize this information is very useful to me! Thanks.


To add a link to text:
<a href="URL">Text</a>