Accurately Measuring Workforce Diversity: Why It Matters

stethescope and graph

By: Misty L. Heggeness, PhD, MPP, MSW, chief, Longitudinal Research, Evaluation, and Outreach Branch within the Social, Economic, and Housing Statistics Division, U.S. Census Bureau. At the time this work was completed, she was a labor economist, Division of Biomedical Research Workforce, Office of the Director, National Institutes of Health.

In 2000, while working as a social work intern in the Minnesota Department of Human Services, I participated as a staffer on two advisory committees. One examined the state of affairs of African American children in the child welfare system; the other examined best practices for handling child welfare cases when American Indian children were involved. What struck me about both committees was the passion of the stakeholders and the earnest effort of employees of the state to mend relationships and improve community interactions with the goal of bettering the child welfare system for children of color in Minnesota. I was charged with examining the data and wound up calculating representation ratios for children in the child welfare system. I will always remember my struggles to clean the data and make sure our measurements were as accurate and precise as they could be.

In our article in this month’s issue of Academic Medicine, this story has come full circle. My prior experiences calculating representation ratios in a state child welfare system were put to use helping the National Institutes of Health (NIH) examine issues of measurement and representation within its own community. We used a relevant labor market definition to calculate representation ratios that showed over or under representation of diverse subgroups at each stage of the pathway to becoming a biomedical researcher (see Supplemental Digital Appendix 1). We were interested in generating accurate estimates of the diversity of the biomedical workforce and wanted to better understand areas where NIH programs and policy could have an influence.

For example, if there is a large discrepancy in the representation of women transitioning from high school to college but little discrepancy in the representation of those women who reach careers as advanced degree biomedical scientists and those who receive NIH funding, that gives us relevant information from a policy perspective about how and where to intervene to improve women’s representation in biomedical science. Focusing resources on increasing women’s participation in independent funding awards, when their representation in this group is already what we can expect, is potentially not the best use of resources because the problem of underrepresentation begins at an earlier stage.

Our results are interesting for two reasons. First, we showed that when you use an inaccurate reference group for understanding representation, you get direly different results (see Figure 3). Measurement matters. If we are going to talk about diversity within our scientific community, we need to accurately measure it. Second, our results showed that NIH and associated institutions have been making ground in advancing a diverse pool of candidates into training and early career funded opportunities, although more work is needed to advance these candidates into the independent investigator (RPG/R01) pool.

In some sense, our results are nothing new. Anecdotal stories and some limited evidence (from Moss-Racusin et al.; Lautenberger et al.; and Ginther et al.) tell us that women, for example, are fully represented in medical training but do not funnel equally into leadership roles at medical schools and, similarly, into independent research roles. In light of these findings, our careful calculation and estimation of diversity measurement within the NIH-funded workforce is important for two reasons: (1) it allows us to quantify the extent of the issue and track and monitor trends, and (2) it informs policymakers and leaders within the biomedical community and helps them in making data driven decisions related to policies and programs. In all, my experience with this research brings back memories of my work over a decade ago in Minnesota–a healthy policy environment where honest, caring individuals come together from both the community and the government to collaborate to make informed decisions based on data and monitoring that is guaranteed to benefit future generations.

Related Posts

One Comment

  1. Laurie
    October 10, 2016 at 6:56 PM

    I definitely think that if studies are being done, they need to be done accurately so we know proper statistics. Great information, thanks for sharing!

%d bloggers like this: