Association for Public Policy Analysis & Management

This study investigates the composition of the biomedical research workforce and how it has changed over time and then analyzes the factors responsible for the retention (or lack thereof) of scientific talent within this workforce. Understanding the biomedical labor force is crucial because research discoveries and innovations directly foster economic growth and affect public health and well-being, both in the United States and abroad. In the analysis, we pay particular attention to career trajectories for women and the factors associated with retention within the occupation of biomedical research given concerns about women’s progress in STEM more generally (Preston, 2004; Stephan and Levin, 2005; Hunt, 2016).

The analysis is conducted using restricted-use data from SESTAT, the the most comprehensive dataset on the scientific workforce in the United States, for the years 1993, 2003, and 2010. SESTAT is an integrated dataset comprised of data from the NSCG, NSRCG, and SDR. A particular advantage of SESTAT data is that they include scientists who received their PhDs abroad (the stand-alone SDR does not). They also include rich information about job preferences (e.g. pay and promotion, work-family balance) for all respondents and reasons for job change (where relevant). The restricted-use SESTAT data permit us to undertake longitudinal analysis. Among the distinguishing aspects of this work, we use an occupation-based definition of the biomedical workforce rather than using field of study (NIH, 2012). We further zero-in on those doing research as a primary or secondary activity. By using an occupation-based definition, we are able to look “backward” to field of training and thereby investigate the extent to which the biomedical “research enterprise” has grown increasingly interdisciplinary. Our definition of the workforce includes those with Bachelor’s and Master’s degrees as well as those with PhDs (though we restrict some analyses to the latter group). Finally, we look at biomedical researchers employed in all sectors, not just academia; our data show that 50% of these individuals are employed in government and industry combined.

We analyze cross-section trends for the period 1993 to 2010 and then exploit the longitudinal nature of the data to look at employment paths of those identified as biomedical researchers in 2003. Here we preview just some of the findings to date. We find that a larger fraction of female biomedical researchers have interdisciplinary training (as measured by trained in the social sciences) as compared to their male counterparts, and this difference has grown. Turning to some results from the longitudinal analysis, we find that even after a 7-year period (2003 to 2010), 65% percent of women switched out of research-oriented biomedical positions versus 58% of men. When the analysis is restricted to those with PhDs, we do not see this same gender difference. We are now conducting a more detailed analysis of factors associated with retention using OLS/logit. In doing so, we want to more fully investigate the role of field of training as well as exploit the rich information on job preferences and reasons for switches.