Building entrance.
NSF funds one quarter of all U.S. federally funded research in the sciences. Credit: iStock.com/JHVEPhoto
Blue circle with the text Science Leads the Future, AGU Fall Meeting. Below the circle is the text #AGU22 coverage from Eos, also in blue

In 2020, geologist Christine Yifeng Chen was sitting on Zoom in her Asian Americans and Pacific Islanders in Geosciences group when she heard something startling: It was an open secret that Asian geoscientists had a more challenging time getting National Science Foundation (NSF) funding than other principal investigators (PIs). A recent Ph.D. graduate from the Massachusetts Institute of Technology and Woods Hole Oceanographic Institution, Chen remembered thinking, “Huh?”

Chen and her friend Sara S. Kahanamoku were pointed to a corner of the NSF website that hosts reports describing the funding outcomes of NSF’s grants by race and ethnicity back to the 1990s. After digging into the data, they discovered that the rumor was true, and they noticed that it wasn’t just Asian PIs who were funded less often, but other groups too.

“We also noticed that there was this pattern that white PIs were actually not only experiencing higher than average funding rates but that this advantage was, in fact, increasing over time,” said Chen, who is now a postdoctoral scholar at Lawrence Livermore National Laboratory.

Chen, Kahanamoku, and their colleagues found “pervasive racialized disparities” in NSF funding awards across 20 years of data, according to their recently published paper in eLife. “The prevalence and persistence of these racial disparities in funding have cascading impacts that perpetuate a cumulative advantage to white PIs across all of science, technology, engineering, and mathematics,” they wrote in the paper.

“NSF welcomes the attention this paper brings to questions about racial disparity in federal funding,” said an NSF spokesperson. “Racial disparity is one of the reasons NSF has been releasing federal-funding demographic data to the public every year since the 1990s.”

The average difference in the funding rate for racial and ethnic groups compared with the median applicant. The plot shows that white PIs have a funding rate 8.5% above average, whereas Black and African American PIs have a funding rate 8.1% below average. Gray dots show each year, and horizontal bars show the 20-year average. AI/AN stands for Alaskan Indian and Alaskan Native, and NH/PI stand for Native Hawaiian and Pacific Islander. Funding rates for multiracial PIs were available only between 2005 and 2019. Credit: Chen et al., 2022, https://doi.org/10.7554/eLife.83071, CC0 1.0

The Numbers

NSF receives tens of thousands of funding proposals annually and over the past 5 years has accepted about a quarter of them. The agency sends most proposals to outside experts for review, then considers the recommendations and issues funding awards. Applicants voluntarily select a race and ethnicity at submission, but external reviewers don’t see the answers.

In the latest analysis, proposals from white PIs were about 9% more likely to be funded than the average applicant between 1999 and 2019. And this advantage has grown. In 1999, proposals from white PIs had a 3% greater chance of being funded than those of the average applicant. In 2019, the most recent year analyzed, white PIs were 14% more likely to secure funding.

“There are people in the pipeline right now that are being affected by these [funding] decisions.”

In contrast, between 1999 and 2019, NSF was 21% less likely to fund proposals from Asian PIs than an average applicant. Native Hawaiian and Pacific Islander PIs and Black and African American PIs were similarly disadvantaged, landing funding 11% and 8% less often than average.

These statistics translate into differences in the funding rate among each group. Considering the number of proposals submitted by white PIs in 2019, this group should have received nearly 800 fewer awards than if it were funded at the average NSF rate for that year (27%). This “award surplus” is countered in part by a “deficit” of 432 awards for Asian PIs.

“There are people in the pipeline right now that are being affected by these [funding] decisions,” said Chen. NSF supports one quarter of all federally funded research in the sciences.

The racial stratification in funding rates is more apparent when comparing the types of grants awarded, the study finds. Black and African American PIs and Asian PIs had a lower success rate for research proposals than nonresearch proposals, though data distinguishing the two types of awards were available only from 2013 to 2019. Nonresearch awards support lab upkeep, teaching, or professional development and typically include less money than research awards.

This infographic depicts award surpluses and deficits calculated by Chen and colleagues. Credit: Chen et al., 2022, https://doi.org/10.7554/eLife.83071, CC0 1.0

The Geosciences Are Not Immune

Data from 2012 to 2016 reveal that white PIs in the geosciences received an average excess of 46 research awards annually.

White PIs in the geosciences had tailwinds, too.

Data from 2012 to 2016 reveal that white PIs in the NSF Directorate for Geosciences received an average surplus of about four dozen research awards annually.

Furthermore, white PIs had the highest submission rates from all NSF directorates, whereas Black and African American PIs had the lowest submission rates from all directorates, said Chen.

Geologists Rachel Bernard and Emily Cooperdock wrote in a 2018 Nature Geosciences comment that racial and ethnic diversity of geoscience Ph.D. graduates has changed very little, despite improvements in the gender balance over the same period.

Closing the Gap

“This is the first time to my knowledge, researchers have conducted such a comprehensive investigation into NSF-wide demographic data,” said Bernard, a geologist at Amherst College who reviewed the paper for eLife. The researchers of the latest work, which was also presented at AGU’s Fall Meeting 2022 in Chicago last week, did an “excellent” job, she said.

“While some people may disagree with the authors’ conclusions…you can’t argue with the data itself.”

“While some people may disagree with the authors’ conclusions—namely, that these disparities are a result of systemic racism within science and society as a whole—you can’t argue with the data itself,” said Bernard.

An NSF spokesperson told Eos that this year the agency is launching a new initiative that specifically addresses barriers and challenges that can impact competitiveness for external funding, called Growing Research Access for Nationally Transformative Equity and Diversity (GRANTED). The program will target PIs at less-research-intensive institutions with training, support research institutions to strengthen research services offered, and hold listening sessions and other meetings to improve NSF offerings.

The number of awards needed to bridge some racial disparities is small, Chen said. “If [NSF] just funded a small number of proposals, it would have a huge impact on reducing or even eliminating some of the disparities by funding rates.”

—Jenessa Duncombe (@jrdscience), Staff Writer

Citation: Duncombe, J. (2022), Twenty years of NSF funding show racial disparities, Eos, 103, https://doi.org/10.1029/2022EO220577. Published on 21 December 2022.
Text © 2022. The authors. CC BY-NC-ND 3.0
Except where otherwise noted, images are subject to copyright. Any reuse without express permission from the copyright owner is prohibited.