This guest post is from @iGrrrl, a grant writing consultant. I think I first ran across her in the comments over at writedit's place, you may have as well. She brings a slightly different, and highly valuable, perspective to the table.
There was an old New Yorker cartoon of two people at a party, and one tells the other, "I'm a fiction writer in the grant-proposal genre." I hate putting fiction in grant applications, especially the type that will be due shortly in response to the Ginther report.
For those who have been worrying about their own grant applications, the Ginther report detailed the relationship of race and ethnicity to NIH grant funding at the R01 level, and NIH has created a few initiatives to try to change the pattern of lower success rates for African American applicants. In early April, the applications are due for the Building Infrastructure Leading to Diversity (BUILD) Initiative, which would fund large-scale projects within individual institutions, and the NIH National Research Mentoring Network (NRMN), which is designed to build a network of mentors. In other words, diversity interventions writ large, with millions of dollars behind them.
The NSF ADVANCE program started in a similar way, and some of the early ADVANCE projects included programs with a limited evidence base. The successes of components of these programs were variable, and the recent RFA included a social science research component to complement the required evaluation of ADVANCE-funded activity. It started with good intentions, and eventually became clear that more than good intentions were needed. Part of the NRMN announcement calls for pilot programs, but I would argue that in the first year, the network should do the social science work to consolidate the anecdotes African Americans tell about their mentoring experiences into hard data, so that the pilot programs can be based upon addressing the needs identified by African Americans who have been through, or training right now in, the current system.
At the recent AAAS meeting in Chicago there were sessions on building diversity in science. At one I learned that explicit bias has reduced in the last 30 years, but implicit bias hasn't. We think we have made progress, and that our conscious intentions are enough. But they clearly are not. Expecting trainees to overcome biased behaviors (to which the actors are blind) places an undue burden on those who are discriminated against. There are studies showing that education about implicit bias helps to reduce such biased behavior, but education attempts can also be done badly and backfire. As pointed out in a recent piece in Science by Moss-Racusin, et al., there is an evidence base now for doing intervention well. If NIH is putting money into large-scale intervention, I hope the existing science will be part of the applications, and expected by the reviewers.
I've spent a lot of my professional life working on exactly the kind of large, infrastructure-based grant application represented by the BUILD and NRMN programs. It is easy for PIs to make assumptions that interventions that sound good on paper will actually have any impact. My concern is that what will be proposed by the applicants to BUILD and NRMN may miss the strong social science work that exists, and that still needs to be done. In fact, some of the best research on effective mentoring is the business literature, a place where few biomedical scientists would think to look.
Grant applications shouldn't be pure fiction, but based on solid evidence. Every grant application represents a possibility, a reality, that could come to pass if the funds are awarded. In the mentoring literature, practices that improve the success for African Americans are often shown to improve the climate for everyone. There is an opportunity here for those in biomedicine to learn from other fields, to consider an evidence base that is outside their usual ken, and to improve the entire biomedical enterprise by improving the overall environment. I hope that those applying for BUILD and the NRMN include the social sciences, and even more importantly, include the voices and ideas of the very people these programs are meant to serve.
NIH has a long history of using dollars to encourage cultural change, with mixed results, because applicants can have varying levels of commitment to the NIH vision while being happy to take NIH dollars. The ADVANCE program at NSF had some hiccups as they worked out what worked to improve the climate for women in STEM. The leadership teams for BUILD and NRMN should include people with a deep knowledge of the research and scholarship on bias and on mentoring, and who can do the rigorous analysis of the current state of affairs for African Americans in biomedical science. I hope I'm wrong here in worrying that such people won't be included, but I've seen fiction in grant applications a few too many times.