Biomedical research scientists send proposals to the National Institutes of Health in the hopes of being funded. A recent study of this process, published in Science by the University of Kansas’s Donna Ginther and her colleagues, revealed that proposals from black applicants are significantly less likely to be funded than proposals from white applicants. This disparity was apparent even when controlling for the applicant’s educational background, training, publication record, previous research awards and employer characteristics.
The authors conclude that racial bias is not a likely explanation for these findings because the race of the applicants is not provided to the reviewers. In an accompanying article in Science, several prominent black biomedical scientists also express doubts about racial bias, concluding that the NIH peer review grades only the science. But what, aside from bias, can explain the racial discrepancy? The study’s lead author admits she has no idea. Understanding what causes bias is essential for developing a program to address it.
One possible explanation is that NIH peer review is structured to promote bias not so much against a racial group as against the unfamiliar and unconventional. Expert reviewers are asked to provide detailed assessments of long, highly complex, extraordinarily technical documents, and they are given little time to do it. The reviewers are usually conversant with the specific area of research that the proposal addresses, which means that they come to the application with preconceived notions. Short deadlines encourage them to rely on established knowledge and sensibilities. In this scenario, reviewers are more comfortable with proposals from scientists they are familiar with—scientists they either know or know of.
Black researchers, at least in the biomedical sciences, are often unfamiliar to reviewers, and their ideas may tend to be unconventional. This situation is in part because of their typical background. For instance, blacks and whites have different prevalence rates for some illnesses, such as end-stage kidney disease and malignant melanoma. Therefore, blacks may propose studies involving a different set of diseases than whites do.
Breaking into the ranks of funded investigators supported by the NIH is increasingly difficult, the data show. The average age of recipients of a first major grant from the NIH had climbed to 43 years in 2007, from 35 years in 1970. Black scientists also tend to make up smaller and smaller minorities in higher branches of science. In the period Ginther and her colleagues studied, blacks submitted 1.4 percent of total proposals compared with 69.9 percent for whites.
This statistic conforms with data collected by the National Science Foundation that indicate only 2.6 percent of doctoral-level biological scientists in the U.S. in 2006 were black. My sense is that the underrepresentation of blacks in biomedical research is even more definitive at the upper echelons: department chairs, research award winners, editorial board members, study section reviewers and members of the National Academy of Sciences. Because blacks have not shared proportionally in the power structure, it stands to reason that funding has been uneven, too.
NIH directors have recognized their failure to fund unusual proposals and have initiated awards, such as the NIH Director’s New Innovator Award and Pioneer Award Program, in response. These steps, though, have not gone far enough. One solution might be for the NIH to establish multiple, distinct mechanisms for making funding decisions. A lottery, for instance, would not result in racial disparity in grant awards. Neither would having rigorous sampling procedures for reviewers or peer review by crowdsourcing. Supplementing traditional peer review with new ways of screening grant applications may be the only way to eliminate the racial gap once and for all.