Most of us today don’t know the name W.E.B. Du Bois. But we should. Du Bois’s pioneering sociological work on race (books like The Philadelphia Negro and The Souls of Black Folk) and his leadership at the forefront of civil rights struggles (including a founder of the National Association for the Advancement of Colored People in 1909) earn him praise as one of the most important Americans of the 20th century.

One of Du Bois's early battles concerned the use of race in scientific studies. Just as biology and anthropology were beginning to untangle the riddles of human heredity in the opening decades of the 20th century, Du Bois worried that science would be put in service of demarcating the color line that he was fighting so hard to erase. His "cure" for American racism "was knowledge based on scientific evidence"—and in much of his early work he set out to utilize the scientific method to show that the concept of race was no more than an unscientific expression of America's racial mores.

"The world was thinking wrong about race, because it didn't know it. The ultimate evil was stupidity," Du Bois would write.

Du Bois argued that Americans talked about race the way they did because they didn't know any better ("they are not up to date on scientific advances regarding the understanding of human diversity").

He attacked the idea of whiteness, citing anthropological evidence that showed that Europeans were a diverse population.

He attacked the idea of a discrete black race, writing that "the human species so shade and mingle with each other that … it is impossible to draw a color line between black and other races."

Finally, Du Bois drew on data—from the census, from public health, from previously published studies (some of which made conclusions in support of the race concept) —to show that racial disparities in health outcomes did not have biological roots. If social and economic class were the dominant factors in how population groups were identified, Du Bois believed, the role of race in predicting health disparities "would be almost entirely eliminated." Ultimately, poverty and "the conditions of life," as he called them, were the real causes of health disparities.

What Du Bois understood, decades before population genetics and evolutionary biology provided the proof, is that human sub-populations (races) are much more diverse than they were once thought to be. As a result, race is an imprecise measure for delineating major groups.

More than 100 years later, race remains a tool that is used regularly in both genetic and public health research, although generally without the overt racism that once characterized the scientific study of human difference. Yet the use of race in science is still controversial—and it raises two specific concerns that echo Du Bois's:

First, racial science continues to fuel racist beliefs. The idea that the peoples of the world can be organized into biologically distinctive groups, each with their own physical, social, and intellectual characteristics, is an idea still embraced by some. You need look no further than a book published earlier this year by former New York Times science reporter Nicholas Wade, who wrote (among other absurd claims) that Jews “have adapted genetically to a way of life that requires higher than usual cognitive capacity” and that we can associate traits like aggression, tribalism, and an ability to function well in the modern capitalist world with specific racial groups. This is nonsense, and even a group of prominent geneticists and evolutionary biologists whose work was cited by Wade publicly rebuked him.

Second, despite the acknowledgement by many natural and social scientists that race is an imperfect way to understand human genetic diversity, some clinicians and drug companies hope to exploit racial differences in the development of personalized medicine—an emerging field that seeks to tailor treatment and drugs to the patient. Because it is currently too expensive to sequence everyone's genomes, race is being used to make best guesses about an individual's disease risk and how his or her genes influence response to drug treatments. Because genetic differences can cluster in populations, the belief is that this can help clinicians and drug companies make medical decisions based on race.

The genome pioneer J. Craig Venter and his colleagues have written that the best choice for personalized medicine is to look directly at an individual’s “genomic sequence instead of relying on a patient’s appearance or self-identified ethnicity.” Ultimately, race based predictions, because of the heterogeneous nature of all racial groups, are merely predictions. Venter illustrated this in a 2011 paper that used the genome sequences of two self-identified white men—himself and geneticist James Watson (of Watson & Crick fame). Venter found that while he had a gene variant that allowed his body to be an extensive metabolizer of certain drugs, including codeine, antipsychotics and antidepressants, his colleague did not. Watson’s genetic variant is indeed rare among people who refer to themselves as white; it shows up more frequently in East Asian populations. If race alone had been used to assess Watson’s ability to take advantage of these drugs, he would have received improper care.

More than a century after Du Bois's critique of racial science, race remains a problematic and highly charged concept. At its worst, racial science reinforces racist beliefs and has the potential to harm human health.

This blog is adapted from my book Race Unmasked: Biology and Race in the 20th Century, published last month by Columbia University Press.

Read more about The Public's Health.