News blog

Study disputes racial bias in NIH grant-making

A new study challenges the notion that there is racial bias in grant-making at the US National Institutes of Health (NIH), a major concern for the agency since a 2011 study, published in Science, found evidence of such bias.

The original study, led by Donna Ginther, an economist at the University of Kansas in Lawrence, found that, after controlling for other factors such as publication record and educational background, black applicants were 10% less likely than whites to land an NIH award (see ‘Black applicants less likely to win NIH grants‘). Agency director Francis Collins called the findings “unacceptable” and chartered a working group of his advisory committee to address the problem.

In December, Collins implemented many of that group’s recommendations, launching a ten-year, US$500-million initiative to provide grant support and mentoring to minority undergraduates, as well as a study-section programme piloting anonymized applications (see ‘NIH tackles major workforce issues‘).

Now, writing in the Journal of Informetrics, Ge Wang, until this week the director of biomedical imaging at the Virginia Tech–Wake Forest University School of Biomedical Engineering and Sciences in Blacksburg, challenges the idea that there is a problem — at least, a problem with bias in NIH study sections. (Wang has just moved to the Rensselaer Polytechnic Institute in Troy, New York.)

Wang and his colleagues applied a mathematical analysis to a random sample of 40 black faculty members in both clinical and basic sciences at the top 92 US medical schools. They were paired with 80 white faculty members using criteria matched for gender, degree, title, specialty and university. The authors found that the black scientists were less productive.

They also identified a subgroup of 11 black faculty with NIH funding and paired them with 11 white faculty members. They found that the black faculty outdid white counterparts in both number of NIH-funded projects and funding totals when compared to scientists with similar productivity levels. “In contrast to the [Ginther et al] Science paper,” the authors conclude, “our results suggest that there is no significant racial bias in the NIH review.”


  1. Report this comment

    Kausik Datta said:

    There appears to be a major problem with the way this News Report has been written. As stated in this report:<br />
    <em>Wang and his colleagues applied a mathematical analysis to a random sample of 40 black faculty members in both clinical and basic sciences at the top 92 US medical schools.</em><br />
    This sentence does not represent the design of the study correctly.<br />
    <br />
    Consider Wang <em>et al.</em>‘s statement in the cited paper:<br />
    <em>This study targeted the top 92 American medical schools ranked in the 2011 US News and World Report, from which 31 odd-number-ranked schools were selected for paired analysis (schools were excluded if they did not provide online faculty photos or did not allow 1:2 pairing of black versus white faculty members).</em><br />
    There is a lot of difference between 31 and 92. The study’s rationale for this shrinkage in the final pool may be reasonable, but the fact remains that the final pool from which the sample was drawn is a much smaller pool, which has a chance to bias the data and/or affect the generalizability of the conclusions. The same objection applies to the way 40 Black American samples were drawn from a pool of 130 Black American faculty members.<br />
    <br />
    Also stated in the News Report:<br />
    <em>The authors found that the black scientists were less productive.</em><br />
    This simple – or perhaps I should say ‘simplistic’ – statement, in absence of further elaboration, doesn’t adequately convey the full import of the analysis. When the analyzed data from a study reveal observations such as:<br />
    <em>…the analysis shows the male investigators were statistically more productive than the female colleagues, and the black faculty members statistically less productive than the white colleagues.</em><br />
    It is an important indication that the phenomenon of ‘stereotype threat’ must be considered.<br />
    <br />
    Wang <em>et al.</em> have not discussed the implication of their data in greater details, apart from saying that it contradicts Ginther <em>et al.</em>‘s conclusions. However, the very existence of situation in which only about 10% of Black American faculties can get funded by the NIH [I quote]:<br />
    <em>Among the 130 black samples in the initial list, 14 faculty members were funded by NIH during the period from 2008 to 2011.</em><br />
    … should give policymakers and regulators a pause, and make them reconsider the question whether some inherent imbalance and group (race/gender) disparities continue to exist in the current policies, as well as larger questions about the participation of disadvantaged groups in STEM education and research in this country.

Comments are closed.