Asian American scientists reflect on NSF racial funding disparities
Jeemin H. Rhim, Rohini Shivamoggi, and V. Bala Chaudhary
Science is, fundamentally, our way of knowing. As scientists, we pride ourselves in developing ways to build knowledge in a manner that is free from bias. We train for years honing our skills in objective methods to generate new information about our world. Often, which questions scientists address are the result of decisions made around science funding. But what if funding decisions and hence the very knowledge produced by scientific research was the result of decades of systemic racial biases? What if which science gets funded is determined by a scientist’s ethnicity or race?
A new article published today in the journal eLife by Dr. Christine Yifeng Chen and colleagues describes years of racial disparities in funding rates indicative of systemic racism at the National Science Foundation (NSF), the core basic science funding agency in the United States. The authors used publicly available data from over 1 million proposals to analyze funding rates, award types, and race and ethnicity data voluntarily provided by principal investigators (PIs). They calculated annual overall funding rates for PIs and then showed how PIs of certain races and ethnicities were consistently funded either above or below the overall rate. Chen et al. demonstrate that, over a period of roughly 20 years, white PIs were consistently funded at higher rates compared to non-white PIs and that this disparity is increasing with time. The NSF is organized according to different disciplinary directorates (e.g., Biology, Geosciences, Math and Physical Sciences), but racial disparities held regardless of division. Finally, the study found racial stratification between award types; “research” proposals compared to “non-research” proposals (e.g., education and training grants, conferences and symposia) are funded at lower rates for non-white PIs, a disparity that is particularly low for Black/African American PIs. The study highlights severely problematic funding issues at the NSF, similar to those that have been highlighted in other U.S. science funding agencies.
One of the more striking results of the study, and the focus of this commentary, is the stark and consistently lower rate at which Asian PIs are funded. When data are aggregated, Chen et al. show that Asian PIs have funding rates 21.2% below the overall NSF funding rate. In fact, Asian PIs are funded consistently and substantially lower than annual average funding rates for the entire period that NSF data are available, regardless of directorate. However, unlike American Indian/Alaska Native, Black/African American, Hispanic or Latino, or Native Hawaiian/Pacific Islander PIs, Asian PIs are not considered underrepresented minorities (URMs) according to the NSF. In aggregation, and relative to the U.S. population of Asian Americans, Asian PIs are not underrepresented in most STEM disciplines. Still, lower funding rates for Asian PIs indicate that, despite not being considered URMs by the NSF, Asian PIs face significant obstacles to recognition in science communities as well as systemic bias, discrimination, and racism in science workplaces. These racist biases result in consistently lower NSF funding for Asian PIs, which has severe negative consequences for science career longevity and success.
It can be tempting to examine the data and graphs in Chen et al. and compare funding disparities across racial and ethnic groups only to engage in a “who has it worse” competition. But Chen and colleagues rightly rebuke “divide-and-conquer” narratives that position Asians against other communities of color. These myths — for example, that Asians are professionally excellent elites or that policing will solve anti-Asian violence — are not only incorrect, but are also motivated by anti-Black racism. The truth that they elide is that Asians, like other Black, Indigenous, and People of Color (BIPOC) communities, suffer because of white supremacy. Most BIPOC researchers were less successful in getting their proposals funded than you’d expect from the NSF’s average success rate. Chen et al.’s team examined over a million proposals; if there had been evidence to support the claim that Black, Indigenous, and Latinx researchers receive professional advantages at the expense of Asian researchers, they would have found it. Instead, they found that disparities in the NSF funding are accelerating the careers of white scientists at the expense of everyone else.
Similar attempts to pit BIPOC groups against one another are ongoing in recent arguments heard by the Supreme Court regarding the future of affirmative action in college admissions. Lawsuits brought by Students for Fair Admissions (SfFA) argue that Asians are treated unfairly in admissions, but primarily exploit Asian discrimination as a means of antagonizing other BIPOC communities. The US Department of Education already investigated Harvard in 1990 and found that Asian American applicants were disadvantaged mostly because of legacy admissions, which overwhelmingly favor white applicants. In this climate, we appreciate that Chen et al. delved into historical data to uncover the roots of Asian scientists’ low success rate. As their analysis shows, the solution is to address the overrepresentation of whiteness in educational institutions, and not to break solidarity with other minoritized communities. Asian Americans are not a pawn whose unfair treatment can be largely neglected and then opportunistically exploited to harm our allies.
Chen et al.’s finding comparing racial disparities between “research” and “non-research” proposals was particularly painful for us, as three Asian American women scientists who actively engage and find meaning in advocating for minoritized science communities. Diversity, equity, inclusion, accessibility, and justice (DEIAJ) work is vital but has costs for our careers and mental health. Although it is often perceived by departments as service work, DEIAJ work is valued less in promotion and tenure processes. We are not alone in accepting the cost of engaging with DEIAJ, though: workers belonging to systematically excluded groups are especially likely to work on DEIAJ, either voluntarily or because the inaction of institutional leadership shifted that burden onto them or both. As minoritized scientists, are we valued more for service than science? Are our roles in the scientific community limited to being confessionals or educational resources instead of peers with ideas worth hearing? Data in Chen et al. indicating that BIPOC researchers are more disadvantaged in getting their research proposals funded than their non-research proposals makes us feel that even the NSF trusts us less to do science than to do service.
Moving forward, it is our hope that the NSF will act with the same community-mindedness and empirical rigor we have come to expect from all NSF endeavors. We are especially concerned by Chen et al.’s observation that white researchers have been growing steadily more advantaged over the last twenty years, which gives us even more cause to reject the status quo, re-evaluate existing racial equity practices, and implement new mechanisms for regular assessment. As a scientific community, we can and should expect racial equity practices carried out by the NSF to be evidence-based, transparent, and accountable. We do not have time to waste on ineffective interventions, and the NSF should not have waited for a group of independent BIPOC researchers to document racial funding disparities and offer final solutions.
Chen et al. offer many insights for improving future data collection, many of which highlight the nuanced yet important patterns in their data. Importantly, if the NSF were to update their practices in data collection and analysis, care must be taken to (1) improve transparency, (2) disaggregate data, and (3) consider intersectionality to inform equitable resource distribution. Chen et al. elaborate on these suggestions in their Discussion as well as their response to reviewers (published by eLife alongside the article). We want to highlight a few of these action items from the perspective of Asian Americans and allies in solidarity with other minoritized groups.
1. Transparency - Chen et al.’s analysis would not be possible without publicly available data, such as the federally mandated annual reports on the NSF merit review process. Whether NSF should continue sharing this data with the public is not in question; how the data will be used should be re-examined. A fundamental question driving the design, implementation, and assessment of the data collection process should address why racial and ethnic data are being collected from PIs. If the answer is to improve progress toward equity, a mechanism for regular and transparent assessment — beyond the annual report of data — should be implemented for the merit review process. As Chen et al. discuss, the changes in data collection method can limit our ability to assess multi-year trends. Therefore, decisions as to what types of data to collect and how to collect them from PIs should be made carefully and frequently revisited for assessment. This does not necessarily mean, however, that changes should be made slowly.
2. Data disaggregation — Many readers may be surprised by the extent of the disparity for Asian researchers at the NSF, as evident in the overall funding rate and cumulative effect over time (e.g., Chen et al.’s Fig. 1–3; Fig. 7). Chen et al. provide two important reminders here. First, the patterns are nuanced, and disadvantages for Black/African American PIs are particularly pronounced when disparities across award types and directorates are taken into account. Second, meaningful disaggregation of data is essential for effective improvement of resource distribution at the NSF. In the response to reviewers, Chen et al. mention the potential artifact of the construction of the Asian category, which embodies diverse sub-groups with unique history, lived experiences, challenges, and needs. We believe that disaggregation of data for the Asian PI category will inform effective strategies for re-distribution of surplus awards for white PIs toward Asian PIs, the group with the largest absolute number of award deficits.
The history and evolution of the term “Asian Americans and Pacific Islanders (AAPI)” demonstrates the importance of data disaggregation. The term “Asian American” was inspired by political movements in the 1960s, including the Black Power Movement and protests against the Vietnam War. In 1968, student activists Emma Gee and Yuji Ichioka at University of California, Berkeley, established the Asian American Political Alliance as a unifying political identity for different groups of people of Asian descent on campus. Thus, the term “Asian American” symbolized a shared history of colonialism and immigration, experiences of labor exploitation and racism, as well as a common political agenda among people of Asian descent. In 1980, the U.S. Census Bureau grouped Pacific Islanders with Asian Americans and created the category “Asian and Pacific Islander (API),” which existed through the 1990s. In 1997, the U.S. Office of Management and Budget (OMB) directive separated API into “Asian” and “Native Hawaiian and Other Pacific Islander” categories, as the combined category (API) had not adequately addressed the specific needs of Native Hawaiians and other Pacific Islander groups for funding and resources. This was the last time the OMB made revisions to the Statistical Policy Directive №15, Race and Ethnic Standards for Federal Statistics and Administrative Reporting. Earlier this year in June 2022, the OMB announced its plan to review and revise Directive №15 no later than Summer 2024.
We call for a similar re-evaluation process within the NSF. We anticipate that further disaggregation of data for Asian researchers will reveal the vast diversity and inequity within the category and point us toward more effective strategies to close the award deficit gap for this group. In 2018, another analysis of AAPI representation in the NSF summary statistics also mentioned the lack of data with which to “account for the range of cultures, country of origins, mixed-race and personal backgrounds that Asian Americans represent,” which limits our ability to assess the unique challenges faced by different sub-groups of Asian researchers. Today, Asians are the fastest growing group of the undocumented population and the most economically divided racial or ethnic subgroup within the U.S. The disparities in socioeconomic backgrounds and educational history are reflected in the Asian demographics in academia, which cannot be discerned if we were to use an umbrella category of “Asian” for assessing inequity and resource allocation. Continuing to view Asians as a monolithic group will result in unintentional consequences of contributing to the Model Minority Myth, or the idea that “certain non-white minorities can serve as a successful assimilation ‘model’ for other non-white minorities.”
3. Consider intersectionality — Beyond the disaggregation of data within the Asian category, it is crucial for NSF to embrace an intersectional approach to data collection and analysis. As the feminist scholar Audre Lorde said, “there is no such thing as a single-issue struggle because we do not live single-issue lives.” Re-allocating resources toward the most marginalized groups will not only improve equity in funding practices but also have lasting impacts on shaping the future demographics of the NSF-funded fields of research. Studies have shown the importance of representation, relatable role models and community in shaping one’s scientific identity as well as in recruitment and retention of scientists from marginalized backgrounds, often with intersecting identities. Implementing an intersectional approach during data collection and analysis at the NSF will provide an essential first step for the scientific community to identify subgroups that are in most need of resource redistribution. According to Chen et al., some example data categories to analyze include (but are not limited to) “race, gender, disability, age, career stage, citizenship status, educational history, institution, and socioeconomic background.”
As the NSF reflects on ways to implement these action items, it will be important to consult with experts in relevant fields. For example, scholars in social sciences, history and/or Asian/Asian American studies would be far more equipped to suggest research-based strategies for survey design to improve data disaggregation and implement intersectional approaches. As Chen et al. conclude, progress toward an equitable future requires intentional and committed actions at a range of levels — including both funding agencies such as the NSF and the broader scientific community. We hope this study will inspire leaders and community members to take such actions to tackle the persistent and grand challenges of racial inequity in the sciences.
Jeemin H. Rhim is a Society of Fellows Postdoctoral Fellow at Dartmouth College. She is a geomicrobiologist who combines microbial experiments and isotope analysis to study the interaction between life and the Earth system. She is dedicated to promoting diversity, equity, inclusion, accessibility, and justice within the geosciences and broader STEM fields.
Rohini Shivamoggi is an atmospheric scientist who completed her Ph.D. at MIT. Her work focuses on hurricanes and she is dedicated to working on equity and effective mentorship in STEM.
V. Bala Chaudhary is an Associate Professor of Environmental Studies at Dartmouth College. Research in her lab examines plant-soil-microbial ecology at macrosystems scales and strategies for antiracist action in STEM.
All three authors acknowledge prior and current funding from the National Science Foundation. They share these perspectives as NSF beneficiaries and stakeholders in the spirit of collective community action to improve the future of science for all.
