Diversity in science refers to the range of people, perspectives, and populations represented across scientific research, from who conducts studies and who participates in them to whose knowledge and health outcomes are prioritized. It spans multiple dimensions: the demographics of the scientific workforce, the genetic ancestry of people in medical databases, the cultural knowledge systems that inform research questions, and the composition of research teams themselves. Each of these dimensions shapes what science discovers, who benefits from those discoveries, and what gets overlooked.
Who Makes Up the Scientific Workforce
The U.S. scientific workforce does not reflect the country’s population. According to the National Science Foundation, Black or African American workers made up just 8% of STEM workers in 2021, compared to 11% of the total workforce. Hispanic workers held about 15% of STEM jobs while representing 18% of all workers. Women held STEM occupations at roughly three-fifths the rate of men: 18% of female workers were in STEM compared to 30% of male workers.
These gaps widen further up the career ladder. In the life and social sciences, women earned 53% of doctoral degrees between 2008 and 2010, yet only about 32% of assistant professors were women during the same period. That drop-off between earning a PhD and landing a faculty position has persisted for decades. In the mid-1990s, women earned 42% of doctorates but held just 28% of assistant professorships. The pipeline narrows at nearly the same rate, generation after generation.
Why Diverse Teams Produce Different Science
When researchers talk about diversity improving science, they’re pointing to something measurable. A large-scale analysis published in the Proceedings of the National Academy of Sciences found that underrepresented groups produce higher rates of scientific novelty. Women and nonwhite scholars were more likely to introduce new conceptual connections in their research than white male scholars. The more underrepresented a researcher’s gender or race was within their discipline, the more likely they were to generate these novel links.
There’s a painful flip side. Those novel contributions were picked up by other researchers at lower rates. Equally impactful work by women and nonwhite scholars was less likely to lead to successful scientific careers compared to similar work by majority-group researchers. The study’s authors called this the “diversity-innovation paradox”: the people who bring the most new ideas to science are the least rewarded for them. This dynamic means science loses not just fairness but actual discoveries when diverse researchers leave the field or their contributions go unrecognized.
Diversity in Who Gets Studied
Perhaps the most consequential dimension of diversity in science is who participates in research. Over 90% of participants in genome-wide association studies, the large genetic investigations that shape modern medicine, are of European ancestry. That figure has barely budged in a decade. This means the genetic risk factors, drug responses, and disease markers that doctors rely on are calibrated almost entirely to one population group. Medications may work differently, genetic tests may miss critical variants, and risk predictions may be inaccurate for the majority of the world’s population.
Clinical trials have a similar problem. In 2023, Congress passed legislation requiring pharmaceutical companies to submit diversity action plans for certain clinical studies to the FDA. These plans must outline how sponsors will enroll participants who reflect different age groups, sexes, and racial and ethnic demographics. The requirement is binding once finalized, a significant shift from the voluntary diversity goals that previously governed trial enrollment. The underlying concern is straightforward: if a drug is tested overwhelmingly in one demographic group, there’s no guarantee it will be safe or effective for everyone else.
Broadening What Counts as Knowledge
Diversity in science also extends to knowledge systems. Indigenous Traditional Ecological Knowledge, or TEK, represents generations of accumulated observations about ecosystems, species behavior, seasonal patterns, and land management. Researchers working in ecology, conservation, and land management have increasingly recognized TEK as a living body of evidence-based knowledge, not a historical artifact. It offers observations and relational frameworks for understanding land that fall outside typical Western scientific approaches.
Integrating TEK into modern science typically happens through formal partnerships between academic institutions and Indigenous communities. These collaborations aim to co-produce land management strategies that draw on both Western scientific methods and Indigenous wisdom. One model, the Earth Partnership-Indigenous Arts and Sciences program, brings together university researchers and local tribal nations to develop restoration and land management practices. The goal is reciprocity: Indigenous communities contribute knowledge and guide research priorities, while academic partners provide institutional resources and broader dissemination. These partnerships also address historical environmental injustices by giving Indigenous communities decision-making power over research that affects their lands.
Institutional Efforts to Close the Gaps
Major funding agencies have built specific programs around these issues. The NIH launched its UNITE initiative with four focus areas: elevating health disparities research, promoting equity in the biomedical research ecosystem, improving equity within the NIH’s own workforce, and making racial and ethnic equity data more transparent and accessible. One explicit goal is ensuring that race and ethnicity no longer predict who receives research funding or what staff positions NIH employees hold.
Concrete programs have followed. The Faculty Institutional Recruitment for Sustainable Transformation (FIRST) program funds institutions to hire diverse faculty in cohorts rather than one at a time, a strategy designed to change departmental culture rather than just add individuals. The Community Partnerships to Advance Science for Society initiative funds community-led health equity interventions. A harassment reporting portal was expanded to cover discrimination and hostile work environments at institutions receiving NIH grants. These are structural changes, not just statements of intent.
Scientific publishing is shifting too. Some journals, including Cell, now ask authors to complete a diversity and inclusion form at submission, prompting them to consider the diversity of their author lists and whether their reference citations reflect a range of contributors. Language model tools are being developed to help authors audit their citation patterns for bias. Citation practices matter because they determine whose work gains visibility, accrues prestige, and shapes the direction of future research.
What Diversity Changes About Scientific Outcomes
The practical stakes are clearest in medicine. When genomic databases skew 90% European, a genetic test for heart disease risk may miss variants common in African or South Asian populations. When clinical trials underrepresent older adults or people with multiple health conditions, approved treatments may cause unexpected side effects in the patients most likely to use them. When ecological research ignores Indigenous land management practices that sustained ecosystems for millennia, conservation strategies may be less effective.
Diversity in science is not a single issue. It operates at every stage of the research process: who asks the questions, who funds them, who participates, whose prior work gets cited, and whose knowledge frameworks are considered legitimate. Each of these stages introduces potential blind spots when any one group dominates. The cumulative effect is a scientific enterprise that knows certain populations and problems in extraordinary detail while remaining largely uninformed about others.