Diversity and Inclusion in Scientific Research: Progress and Challenges

Representation gaps in scientific research are not abstract equity concerns — they shape which questions get asked, which populations get studied, and which findings get treated as universal. This page examines how diversity and inclusion function within research institutions, what structural forces drive persistent gaps, and where measurable progress has emerged alongside stubborn resistance.

Definition and scope

Diversity in scientific research refers to the range of social identities, lived experiences, and perspectives represented among researchers, study populations, and institutional leadership. Inclusion is the related but distinct condition — the degree to which those represented individuals actually influence decisions, methodology, and institutional culture, rather than simply appearing in headcount statistics.

The distinction matters enormously. An institution can post diverse enrollment numbers while concentrating women and researchers of color in lower-ranked, lower-funded positions. The National Science Foundation's ADVANCE program, which has operated since 2001 with over $270 million invested through its first two decades, was specifically designed to address this gap between nominal representation and structural inclusion in academic STEM settings.

Scope encompasses three interlocking domains:

1. Workforce representation — who holds faculty positions, runs labs, leads research teams, and sits on grant review panels
2. Research content — which populations are enrolled in studies, which diseases are prioritized, which communities are treated as subjects versus co-investigators
3. Institutional infrastructure — mentorship pipelines, retention policies, harassment reporting mechanisms, and pay equity practices

The National Institutes of Health tracks workforce data through its ongoing monitoring of diversity supplements and training grants, providing one of the few longitudinal federal datasets on researcher demographics.

How it works

Research institutions generate diversity outcomes through the compounding effects of hiring pipelines, funding decisions, and informal culture — not through any single policy lever. A useful frame is the distinction between compositional diversity (headcount representation) and epistemic diversity (whether different perspectives alter research questions and methodology).

Compositional diversity is measurable and has improved in some fields. Women earned 53% of doctoral degrees in science and engineering in 2021, according to NSF's Women, Minorities, and Persons with Disabilities in Science and Engineering report. But faculty representation at the full professor level lags substantially behind those doctoral numbers, a pattern researchers call the "leaky pipeline" — a phrase that has itself attracted criticism for implying the problem is attrition rather than structural push.

Epistemic diversity produces different outcomes. When researchers with different backgrounds participate in designing studies, documented shifts emerge in which variables are measured, which comparison populations are selected, and how results are framed. The NIH's 2016 policy requiring the inclusion of females in preclinical cell and animal research directly addressed a decades-long assumption that male subjects produced universally generalizable data — an assumption with real clinical consequences for how drugs perform differently across sexes.

Funding structures also filter opportunity. Historically Black Colleges and Universities (HBCUs) and minority-serving institutions receive federal research funding at rates disproportionately lower than R1 research universities, which concentrates investigator development pipelines at a narrow set of institutions. The NIH UNITE initiative, launched in 2021, specifically identified structural racism within NIH's own grant review processes as a problem requiring institutional intervention.

Common scenarios

The gap between policy and practice appears most sharply in three recurring situations:

• Grant review panels: When reviewers share demographic and institutional backgrounds, unconscious affinity effects can systematically favor familiar research frameworks. A 2011 analysis in Science found Black applicants were 10 percentage points less likely than white applicants to receive NIH R01 funding after controlling for institutional prestige and prior training (Ginther et al., 2011, Science).
• Clinical trial enrollment: Historical exclusion of women, elderly patients, and minority populations from Phase III drug trials has produced FDA-approved treatments with efficacy data drawn from narrow demographic slices. The FDA has issued guidance on improving enrollment diversity, but reporting requirements remain inconsistent across therapeutic areas.
• Sexual harassment and hostile climate: The 2018 National Academies of Sciences, Engineering, and Medicine report Sexual Harassment of Women found that 50% of women faculty in academic science reported experiencing harassment — and that existing institutional reporting structures were poorly designed to address it without punishing the people who came forward.

These scenarios connect directly to research ethics and integrity frameworks, because exclusion is not merely an equity issue but a validity issue: non-representative data produces non-generalizable findings.

Decision boundaries

The clearest analytical distinction in this field separates representation mandates (required by funders or regulation) from institutional culture change (voluntary, slower, more durable). Federal funders including NIH and NSF now attach diversity plans to grant requirements — but compliance with a form and genuine structural change are different outputs.

A second decision boundary runs between individual-level interventions (mentorship programs, targeted fellowships, implicit bias training) and structural interventions (changing who sits on grant review panels, anonymizing applications, revising tenure-clock policies). Research from the Government Accountability Office and academic literature consistently finds that structural interventions produce more durable outcomes than individual-focused approaches, though both are present in most institutional programs.

Fields like biomedicine have further to go than others. The broadest scope of scientific research spans disciplines with very different starting points — environmental science and nursing have near-gender parity in graduate training, while fields like physics and computer science remain heavily male-dominated at the faculty level, according to NSF's 2023 S&E Indicators.