The trouble with the way we talk about race is that our biological differences are by degree rather by category. The borders of a country or continent are not magical lines that demarcate one genetically distinct population from another. “There are no firm and clear boundaries if you sample every grid on Earth,” Tishkoff told me. But because we lack a common vocabulary to talk about these differences between people by degree, we draw boundaries with our words and categorize them: Korean, Mongol, Asian.
Those boundaries will depend who is drawing them and where and when. What race, for example, are Mexicans? In the 1930 U.S. census, Mexicans were their own racial category. In 1940, a court ruled that Mexicans were not eligible for citizenship because they were not white (under a law at the time), so President Roosevelt decided to count Mexicans as white in that year’s census order to shore up Mexican relations. In 1980, the census began distinguishing between race and ethnicity, allowing respondents to choose among several races and answer yes or no on Hispanic or Latino ethnicity.*
Even though geneticists know how messy these racial categories are, the categories are still deeply rooted in biomedical research. The U.S. National Institute of Health, the country’s largest funder of biomedical research, requires researchers to collect data on the race and ethnicity of clinical research participants. So when scientists go to analyze their data, one of the things they can always do is look for differences between the races. The very act of collecting data defines the questions scientists do ask. “There’s this idea there that data collection is somehow a neutral activity,” says Sandra Soo-Jin Lee, a medical anthropologist and bioethicist at Stanford. “We should disabuse ourselves of it.”
Implicit in the requirement to collect race data is a belief that race must be biologically meaningful in health. And this ends up producing research that reinforces this belief. The emphasis on race, says Duster who is now at Berkeley, “is so deeply in the structure of genetic medicine now, you cannot disentangle.” A study might find, for example, African Americans have higher rates of diabetes, prompting headlines about racial disparities and even more research into the genetics of African Americans with diabetes. But the focus on genes in African Americans elides the fact that such differences might predominantly come from a disproportionate number of them living in poverty.
Genetics has allowed scientists to start probing exactly how much innate genetic differences between races do matter in health, but this has unintended consequences, too. Jo Phelan, a sociologist recently retired from Columbia, has devised studies seeing how simply reading a news article about racial differences in genetic risk for heart attacks reinforces the belief that whites and African Americans are essentially different. The problem is that these differences are statistical—a mutation may be more prevalent in African Americans but that doesn’t mean every African American has it. There is no gene or set of genes that consistently codes for black, white, or any other race.