Race has no single, universally accepted definition because the term sits at an intersection of biology, history, politics, and everyday social experience. In genetics, the variation between any two people labeled as different “races” is, on average, only modestly greater than the variation between two people of the same group. The concept as most people understand it today traces back to European naturalists of the eighteenth century, but the boundaries they drew were shaped as much by the politics of colonialism as by observation of human bodies. That tension between the social weight of race and the messiness of the biology underneath it is what makes the term so difficult to pin down and so important to understand clearly.
How the Modern Idea of Race Took Shape
Historians generally mark two turning points in the story of race as a scientific category. The first is the formal classification of humans into discrete types by European naturalists, often dated to Carl Linnaeus’s Systema Naturae in 1735. The second is the dismantling of those racial typologies after the Second World War, when population-based studies of human diversity began to replace the older, fixed hierarchies.1Europe PMC. Race and History: Comments from an Epistemological Point of View Between those two moments, the concept did enormous practical work: it provided intellectual cover for slavery, colonialism, and, in the early twentieth century, the eugenics movement. From roughly 1907 to 1939, prominent American scientists used biological ideas about heredity and race to push for restrictive immigration laws, institutionalization of people deemed “unfit,” and compulsory sterilization programs.2Europe PMC. U.S. scientists’ role in the eugenics movement (1907-1939): a contemporary biologist’s perspective
That history matters for how people argue about race today. When someone asks whether race is “real,” the answer depends heavily on what era’s definition they have in mind. The rigid typologies of the eighteenth and nineteenth centuries are thoroughly discredited. But that does not mean human biological variation is imaginary; it means the boxes we inherited are poor containers for the patterns that actually exist.
What Genetics Tells Us About Human Variation
The single most cited number in this debate comes from a landmark 1972 study by the geneticist Richard Lewontin. Looking across a range of genetic markers, he estimated that about 85% of total human genetic diversity exists within any given population, roughly 8% exists among populations within the same conventionally defined race, and only about 6% falls between racial groups.3The Royal Society. Celebrating 50 years since Lewontin’s apportionment of human diversity In plain terms, two people from the same village can be more genetically different from each other than either is from someone on a different continent. That finding has been replicated with much larger datasets and more sophisticated tools over the past five decades, and the basic proportions hold up.
At the same time, when researchers use clustering algorithms to sort people by genetic similarity without telling the software anything about race or geography, the clusters that emerge tend to map loosely onto continental regions. The small discontinuities in genetic distance that exist, often coinciding with oceans, mountain ranges, and deserts, give software enough signal to group people in ways that resemble familiar continental labels.4PLoS Genetics. Clines, Clusters, and the Effect of Study Design on the Inference of Human Population Structure The researchers who documented those clusters were careful to note that this should not be taken as evidence for any particular concept of “biological race.” Both gradual clines and discrete clusters are useful scientific tools depending on the question, and debating the existence of biological races in the abstract is, as they put it, “largely orthogonal to the question of scientific utility.”
So there is real, measurable geographic structure in human genetics. But it is gradual, continuous, and shaped by migration and isolation over tens of thousands of years. It does not fall neatly into the three, four, or five categories that social definitions of race typically use.
Why Skin Color Is a Misleading Marker
Skin color is the trait most people reach for when they think about racial categories, yet it turns out to be one of the worst guides to overall genetic relatedness. Skin pigmentation evolved in direct response to ultraviolet radiation. Near the equator, strong UV drove the evolution of dark, melanin-rich skin that protects against DNA damage. At higher latitudes, weaker UV created pressure for lighter skin that allows enough UVB through to synthesize vitamin D.5PubMed Central. Human skin pigmentation as an adaptation to UV radiation Multiple research groups have confirmed that UV radiation is the primary selective force shaping human skin color.6PubMed Central. The evolution of human skin pigmentation involved the interactions of genetic, environmental, and cultural variables
The problem this creates for racial classification is that similar UV environments produce similar skin colors in populations that are not closely related. Sub-Saharan Africans, Melanesians, and Aboriginal Australians all have dark skin, but they are among the most genetically distant human groups on the planet. Grouping them together by appearance would be like sorting animals by whether they fly and concluding that bats and sparrows are close relatives.7American Journal of Human Genetics. Human Population Genetic Structure and Inference of Group Membership Adaptive traits like skin color reflect the environmental pressure that shaped them, not the broader pattern of genetic differentiation across the genome. And different adaptive traits carve up human populations in contradictory ways: a classification based on skin color will not match one based on lactose tolerance or resistance to malaria.8PubMed Central. Biological races in humans
Some combination of visible traits will distinguish most populations from one another, but those distinctions are shallow. They reflect local adaptations, not deep genetic boundaries.9American Anthropologist. Human Races: A Genetic and Evolutionary Perspective
When Self-Reported Race Matches Genetic Ancestry, and When It Does Not
If race were simply a stand-in for genetic ancestry, you would expect a person’s self-identified race to line up neatly with what their DNA shows. For some groups it does; for others, the correspondence breaks down in revealing ways. In a study of a multiethnic population in New York City, self-identified White individuals had an average genetic ancestry that was about 75% European, and self-identified Black individuals averaged about 78% African. But self-identified Hispanics showed a much more complex picture, with ancestry estimates splitting roughly into 29% European, 26% African, and 45% Native American.10PubMed Central. Comparing genetic ancestry and self-reported race/ethnicity in a multiethnic population in New York City
A larger analysis from the GENCOV study found high concordance for some groups: about 96% of self-reported White/European individuals were genetically estimated as European, and 95% or more of East Asian and South Asian individuals matched their self-reported categories. But for Latin Americans, only about 58% matched a single ancestry cluster, and for Middle Eastern and Ashkenazi Jewish individuals, the genetic ancestry estimation tools performed poorly, partly because those groups lack good reference populations in the databases.11Genetics in Medicine Open. P709: Self reported vs genetic ancestry from the GENCOV COVID-19 genomic sequencing study
Perhaps more striking is the social logic behind self-identification. Research has found that people who are genetically admixed tend to self-identify with the minority group rather than the majority one. Even individuals with primarily European genetic ancestry who also carry modest African ancestry are likely to identify as African American. By contrast, self-identification as White tends to rise only when European ancestry is very substantial. And self-reported Native American identity follows a pattern unlike any other group: it was actually positively correlated with European genetic ancestry, with only a small minority of people who identified as Native American showing significant Native American genetic ancestry.12PubMed. The Complex Relationship of Genetic Ancestry With Self-Reported Race/Ethnicity These patterns underscore that racial identity is shaped by social rules, family stories, and cultural belonging at least as much as by biology.
Race in Medicine
Medicine is where the race-definition question becomes most urgent, because the answer affects individual patients. For decades, clinical algorithms in areas like kidney function, lung capacity, and obstetric risk included race-based correction factors. A Black patient’s kidney function score, for example, was adjusted upward on the assumption that Black people have higher muscle mass, which meant a sicker patient could appear healthier on paper. Critics argued that this practice masked the real harm caused by lifelong exposure to social and economic disadvantage.13PubMed. Racialized algorithms for kidney function: Erasing social experience Researchers and clinicians have begun replacing race-based adjustments with race-neutral algorithms in areas like pulmonary function testing and obstetrics, with early results showing that diagnostic accuracy can be maintained. But the outcomes vary across clinical contexts, and in some cases it remains unclear whether the new approaches reduce or inadvertently reinforce disparities.14PubMed Central. Clinical Algorithms and the Legacy of Race-Based Correction: Historical Errors, Contemporary Revisions and Equity-Oriented Methodologies for Epidemiologists
There is, however, a context in which population-level genetic data genuinely matters for treatment: pharmacogenomics, the study of how genetic variation affects drug response. Variants in genes that control how the body absorbs, distributes, metabolizes, and excretes drugs differ in frequency across populations. A study of over a thousand such genetic variants across 19 populations found wide differences both between and within continental groupings, with African-ancestry populations showing the greatest diversity in drug-metabolism genes.15The Pharmacogenomics Journal. Pharmacogenomics, ancestry and clinical decision making for global populations The frequency of a variant that influences warfarin dosing, for instance, ranges from very common in some groups to nearly absent in others. These are real differences, but they track with geographic ancestry, not with socially defined racial categories. The push in genetics is increasingly to move from race as a proxy to ancestry as a more precise tool.16The American Journal of Human Genetics. Guidance on use of race, ethnicity, and geographic origin as proxies for genetic ancestry groups in biomedical publications
How Race Gets Under the Skin Socially
If much of what differs between racial groups is not baked into DNA, how do the stark health disparities between them arise? One answer is what researchers call “weathering,” the idea that chronic exposure to social and economic disadvantage accumulates in the body over a lifetime. A systematic review found consistent support for this hypothesis across multiple biological measures, including stress hormones, markers of inflammation, and even telomere length, a cellular indicator of aging.17PubMed Central. The Weathering Hypothesis as an Explanation for Racial Disparities in Health: A Systematic Review
The numbers are blunt. In a study of nearly 2,700 participants, Black individuals had a “biological age,” calculated from physiological markers, that was on average about 2.6 years older than their actual age. White individuals, by contrast, had a biological age roughly 3.5 years younger than their chronological age, meaning Black participants were weathering about six years faster. Lower socioeconomic status and more depressive symptoms were more strongly associated with accelerated aging in Black participants than in White ones.18PubMed Central. Racial differences in weathering and its associations with psychosocial stress: The CARDIA study Chronic inflammation appears to be one key biological mechanism through which cumulative stress translates into worse health outcomes.19PubMed Central. How does race get “under the skin”?: inflammation, weathering, and metabolic problems in late life
Weathering is an important piece of the race-definition puzzle because it shows how a category that is primarily social can produce biological consequences. The health gap between racial groups is real, but its main driver is lived experience, not genetic programming.
Why Census Categories Keep Changing
The categories available on the U.S. Census illustrate how unstable racial definitions are. The choices have changed in almost every decade since the Census began, shifting from a historical orientation around controlling people of color and protecting the political interests of White populations to a more contemporary focus on tracking inequality and distributing resources equitably.20Sociology of Race and Ethnicity. Race Counts: Racial and Ethnic Data on the U.S. Census and the Implications for Tracking Inequality Starting with the 2020 Census, the way the Census Bureau collects and codes race and ethnicity data changed substantially. Researchers have found that the change created a significant data discontinuity, affecting a wide range of demographic, socioeconomic, and mortality statistics that rely on Census data.21SSM – Population Health. Data impacts of changes in U.S. Census Bureau procedures for race and ethnicity data In practical terms, that means trend lines comparing racial health disparities or poverty rates before and after 2020 may not be comparing the same groups at all.
This is not a flaw unique to the 2020 Census. The categories have always been moving targets. “Hispanic” was not offered as an ethnicity option until 1980. People from the Middle East and North Africa have been classified as White throughout the Census’s history, despite many not identifying that way. Each revision reflects the political and social priorities of the moment, which is exactly what you would expect if race is fundamentally a social category rather than a natural kind.
How Children Learn to See Race
The fluidity of racial categories also shows up in how people perceive faces. Research on White children found that those who thought of race in more fixed, essential terms showed a specific pattern in face memory: they remembered White faces better than both racially ambiguous and Black faces, collapsing ambiguous faces into the “other” category. Children who did not think of race as fixed, by contrast, remembered White and racially ambiguous faces equally well, and only showed worse memory for Black faces.22PubMed Central. Essentialist thinking predicts decrements in children’s memory for racially ambiguous faces The boundaries between racial groups become more rigid in the mind as a child absorbs the idea that race is a deep, unchangeable essence. In other words, how sharply you draw the lines between races is partly a product of how you were taught to think about them.
Ancestry in Forensics and Consumer Genomics
Despite the problems with race as a biological category, genetic ancestry estimation has become a practical tool in forensic science and the direct-to-consumer DNA testing industry. Forensic labs can use panels of ancestry-informative markers, specific genetic variants whose frequencies differ sharply between continental populations, to estimate a person’s likely geographic origins from a DNA sample. One validated assay using just 34 such markers could assign individuals to African, European, or East Asian ancestry groups with very low error rates.23PubMed. Inferring ancestral origin using a single multiplex assay of ancestry-informative marker SNPs Forensic anthropologists also use skeletal measurements to estimate ancestry as part of identifying unidentified human remains, since accurate ancestry estimation helps narrow the search through missing-persons databases.24Europe PMC. Metric Methods for the Biological Profile in Forensic Anthropology: Sex, Ancestry, and Stature
In research genetics, ancestry-informative markers have also been used to map genes associated with diseases that differ in frequency between recently admixed populations, such as prostate cancer and hypertension in African Americans, whose ancestry draws from both African and European populations.25American Journal of Human Genetics. The Unfolding of Genetic Ancestry: A Review The key distinction here is that these tools estimate ancestry, not race. An ancestry-informative marker tells you something about the geographic origins of a stretch of DNA. It does not tell you what social group a person belongs to, how they identify, or how they move through the world. This is the distinction that the National Academies of Sciences called for in 2023 when they recommended that genetics research move away from using race, ethnicity, and geographic origin as proxies for genetic ancestry groups.26The American Journal of Human Genetics. Guidance on use of race, ethnicity, and geographic origin as proxies for genetic ancestry groups in biomedical publications Some anthropologists have argued that professional organizations need to update their own official statements on race to account for the growing use of “ancestry” as a framing concept, ensuring that the shift in terminology is understood as substantive rather than cosmetic.27PubMed Central. Anthropologists’ views on race, ancestry, and genetics