This is a post by Executive Director Gary McDowell.
In a preprint posted on bioRxiv, Chris Pickett (of Rescuing Biomedical Research), Adriana Bankston (a policy activist at FoR) and myself argue that the NSF’s Survey of Graduate Students and Postdoctorates in Science and Engineering, known as the GSS, is not suitable to be used as an indicator of the number of biological sciences postdocs in the U.S., despite its current role as the standard reference in discussions of postdoc numbers.
This paper began in a response to Garrison et al., 2016, “Biomedical science postdocs: an end to the era of expansion,” which argued that biomedical science postdoc numbers had declined from 2010, presumed to be due to people choosing to leave biomedicine. As I have discussed before, placing a straight line along the data in the paper’s Figure 1 from 1979 to 2008 shows a clear linear increase in the number of postdocs, and that the data deviates from 2008-2013 away from this line and, importantly, above it, before returning to the line in 2013*.
Therefore, it first appeared to us not that there had been a decline in the number of postdocs beginning in 2010, but a bubble from 2008 to 2010 that corrected from 2010 to 2013. Indeed, anecdotally, many of my peers were extremely confused by the original premise of the paper – that people have been leaving academia in their droves against the continuing growth of the number of PhDs awarded, since 2010. However, proposing that postdocs stayed on for longer while the economy recovered from 2008 to 2010 yielded far more productive conversations, with people citing examples of people in their labs who had done just that. Encouraged, we thought we were on to something, and began to dig further into the NSF’s data.
But we realized that this “postdoc bubble” could not reliably be given as an alternative explanation for this data, because we believe the underlying noise of the data to be so great as to prevent us from making any claim about the trend in the numbers at all.
The biomedical community is already widely aware, from the Biomedical Workforce Working Group Report of 2012, that the number of postdocs in the U.S. is vastly underestimated, largely due to the inability to define and count postdocs at institutions. However, it had been assumed that trends could at least be discerned from data collected year to year in a relatively standard fashion.
We believe this is not the case. In our preprint, we highlight that departmental rearrangements and institutional reforms, and wildly fluctuating numbers being reported by a number of institutions, combined with the separation of biomedical postdocs into “biological sciences” and “health sciences” categories by the NSF (which can be affected by a department at in institution being renamed, and moving the entire department into a new category) render the GSS an unsuitable indicator of trends in postdoc populations, and therefore no conclusions should be drawn about their number from this data.
Repeatedly, for decades, numerous calls have been made for institutions to define and track postdoctoral researchers. In my personal opinion, that no-one knows how many postdocs there are, while they are a near-essential requirement on the path to tenure-track faculty positions, is one of the greatest embarrassments to the U.S. biomedical enterprise. The value that this enterprise places on postdocs is highlighted by the decades-long failure to count them, and the number of problems that have been identified that could at least begin to be addressed by accurately tracking them, such as diversifying the faculty, and where postdocs actually go (both career-wise and location-wise), continues to grow. Calls for more scientists in the U.S. are often used to justify increased federal funding for science, and yet the U.S. taxpayer currently has no way of knowing what their investment in postdocs has yielded, in terms of producing scientists that they are told the country desperately needs.
There needs to be a standard definition of the postdoc; and there needs to be a standard way of counting them. Efforts are under way to define and track postdocs, and we support the adoption of a unified definition of a postdoc across all institutions, consolidation of postdoc titles and the creation of an index to better assess biological sciences postdoc trends, as described in the preprint. These are some of the most basic, and oldest, recommendations that are made in this arena, and there is no reason that institutions cannot track their scientists as well as they track their lawyers and medical doctors.
Please feel free to comment on the preprint or here, and let us know what you think!
*the linear trend then continued along this straight line for the 2014 data, but not for 2015, when it fell below the trendline.