Division, Extremism, and Ideology at the National Science Foundation

The language and tone of the recent U.S. Senate Committee on Commerce, Science, and Transportation report, “D.E.I. Division. Extremism. Ideology. How the Biden-Harris NSF Politicized Science,” is very partisan, which makes the report less persuasive than it otherwise might be. Still, the report identifies and criticizes a growing failure of objectivity by the National Science Foundation (NSF).

The committee parsed the keywords and phrases in the proposals funded by the NSF between January 2021 and April 2024. The report shows a more than 9,000 percent increase in the share of new NSF grants focusing on “diversity, equity, and inclusion” (DEI) initiatives:

While only 0.29% of all grants with start-dates in 2021 centered on DEI initiatives, by 2024, more than a quarter (27%) of new grants pushed far-left perspectives. Redirecting funding to these subjective, ideologically based projects was deliberate. Beginning in 2021, the White House and NSF created scientific integrity policies to require that agencies “[i]ncorporate [Diversity, Equity, Inclusion, and Accessibility] considerations into all aspects of science planning, execution, and communication.

[RELATED: A Great Trump Appointment: Jay Bhattacharya (Part 1)]

Over ten percent (3,483) of grants awarded during this interval went to projects that the committee classified as focusing on status, social justice, gender, race, environmental justice, or a combination of these foci. The “status” category, the largest, is defined by grants describing “persons based on their membership in a population deemed underrepresented, underserved, socioeconomically disadvantaged, or excluded.” Status projects investigate the role of science research as a mechanism for oppressing identity groups.

The Senate Committee’s methodology for classifying these NSF awards was intended to produce powerful and significant results, keeping the probability of any flavor of misclassification low. The procedure relies on rules for parsing keywords and phrases from project descriptions to avoid Type I errors in classifications, that is, to try and prevent classifying projects into one or more of the five DEI categories based on merely casual or incidental use of keyword language.

The Senate report includes several excerpts from grant summaries intended to promote skepticism about these projects’ merits. Still, the committee does not claim to have reviewed these studies for scientific quality. The committee’s premise is that the words used to describe projects and make the case for funding them are likely to encapsulate what the projects are about.

Unfortunately, the report does not list the 3,483 grants the committee classified as DEI-focused, which makes it impossible to verify the committee’s classification results. Based on an N-gram analysis of 32,198 NSF grants identifying the most frequent terms appearing in project descriptions, the committee identified 699 DEI-centric keywords and phrases. They mapped each to one of the five DEI project categories. A search of NSF projects that included the phrase “environmental governance,” perhaps the most value-neutral of the 38 DEI keywords and phrases associated with the committee’s environmental justice category, turned up the project NSF Award Search: Award # 1425883 – Collaborative Research: The Policy Geography of Environmental Risk, which appears to be a credible social science project.

It is unclear whether the NSF belongs in the social science business. Consider a thought experiment. Extract from our lives the fruits of the last century of medical, physical, natural, and engineering science research. Our existence would be very different, much less comfortable, and relatively short. Instead, extract the last century of social science research from our lives. What would change? Something would, but it is hard to say what.

Setting aside the larger question of whether or not the NSF should be funding social science, including education and human resources projects, nothing in the project abstract for Award # 1425883 positions it as the sort of DEI project that concerns the Senate Committee. It is unclear how the committee classified this project, but based on the methodology summarized in the report, it may have been flagged and placed in the environmental justice category. If so, Type I errors are likely more frequent in the report than the committee presumes. Even so, Type II errors are likely very infrequent: The committee’s methodology would rarely miss projects the committee intended to identify as DEI-focused.

[RELATED: Ranga Dias Deals Another Blow to Scientific Integrity]

For most of my 35-year, full-time research university faculty career, the competition for NSF funding set one of the academy’s gold standards for merit-based recognition of scientific research, along with support provided by the National Institutes of Health (NIH), Defense Advanced Research Projects Agency (DARPA), Department of Energy (DEO) and the Office of Naval Research (ONR). It is disheartening and unnerving to learn that more than a quarter of the most recent grants awarded by the NSF are ideologically driven awards.

In 1997, the NSF formalized broader impacts as an explicit proposal review criterion but left panel reviewers the latitude to define the relative weight that should be attached to broader impacts relative to a research proposal’s intellectual merit. “Broader impacts” was a euphemism for impacts across identity groups, and when serving as an NSF referee, I always set the weight for broader impacts to zero. I stopped accepting opportunities to serve on NSF review panels circa 2013 when the agency made the broader impacts criterion a mandatory element of the review process. I also stopped submitting NSF proposals. What matters in science and engineering is the work, not who does it, not where the investigator does it, and not which group, if any, is targeted to benefit from it. Proceeding otherwise to try and make a case for new research is an exercise in fiction writing, and fiction is not my genre.

Perhaps the 9,000 percent increase in the share of funded proposals incorporating the Senate Committee’s DEI keywords and phrases describes otherwise high-quality work that has been cosmetically framed in terms of the NSF’s favorite jargon to make the proposed research more competitive in the review process. Alternatively, the vision and insight embedded in the NSF’s latest social engineering priorities could have elicited from prospective investigators novel work that would not otherwise be proposed. My thinking, though, is that the agency is most likely eliciting and rewarding gibberish.

Follow James E. Moore on X.

Cover by Jared Gould using image by fadi on Adobe Stock (Asset ID#: 918256019) and NSF Logo by MichianaSTEM on Flickr