When an unidentified body arrives in the laboratory of Allysha Winburn, a forensic anthropologist at the University of West Florida, it’s her job to study the bones to help figure out who the person was when they were alive—to give the biological remains a social identity. “We have this vast population of possible missing persons the [remains] could match, and we need to narrow down that universe,” she says.
She measures the length of the limb bones to estimate height and examines the bones’ development to estimate age at death. She studies the shape of the pelvis for clues to the person’s likely sex. And, until recently, Winburn measured features of the skull, such as its overall length and the width of the nasal opening, to do what forensic anthropologists call ancestry estimation. By statistically comparing the measurements with those from skulls with known identities, she could predict the continental ancestry—and the commonly used racial categories that may correspond to it—that a person likely identified as when alive. In other words, she could predict whether they identified as Black, white, Hispanic, Asian, or Native American.
But Winburn, who is white, is now questioning whether she should continue to do so. And she’s not the only one: Over the past year, debate about ancestry estimation has exploded in U.S. forensic anthropology, with a flurry of papers examining its accuracy, interrogating its methods, and questioning its assumptions. A committee of the American Academy of Forensic Sciences’s standards board is now hammering out a new standard that would, if adopted, direct professionals away from racial categories and toward more specific social and biological populations, such as Japanese or Hmong instead of Asian. It will likely come to a vote by the end of the year, with potential implications for the hundreds of U.S. forensic anthropologists, as well as the more than 20,000 missing and nearly 14,000 unidentified people in the United States, and their loved ones.
“This debate has been percolating for decades,” says Shanna Williams, a forensic anthropologist and anatomist at the University of South Carolina School of Medicine, Greenville. Now, “There’s more of a willingness to challenge the status quo.”
Two biological and forensic anthropologists ignited the latest round of debate in the summer of 2020, in a letter to the Journal of Forensic Sciences. Jonathan Bethard of the University of South Florida and Elizabeth DiGangi of Binghamton University argued that ancestry estimation remains dangerously tangled up in its racist roots. Many of today’s forensic anthropologists were trained to identify race using the same techniques earlier generations of scientists employed to argue for biological differences and hierarchies among races in the “race science” of the 19th and 20th centuries. Today’s anthropologists now know those scientists were wrong both biologically and ethically. But translating skeletal data into race, a socially determined category, still reifies the erroneous notion that race is biological, Bethard and DiGangi argue. “Ancestry estimation is race science, pure and simple,” says DiGangi, who is Black and biracial.
The pair also wrote that the criminal justice system might give less attention to remains when they are classified as members of marginalized groups. [“We can’t] assume that our work is not harmful,” says Bethard, who is white. The pair urged forensic anthropologists to stop performing ancestry estimates and study whether the practice results in discrimination.
The letter triggered an explosion of debate. “It took a lot of courage for them to put that letter out,” says Kyra Stull, a forensic anthropologist at the University of Nevada, Reno. “It’s been a catalyst for a lot of positive change.”
But Stull disagrees with Bethard and DiGangi, co-authoring a response to the article that argues ancestry estimates are important tools. “Right now, the reality is that social race is part of how people identify in the United States, and that follows them [in death],” says Stull, who is white. Race is included in missing persons reports, police case files, and in almost every other description of a person. “That has to change for [ancestry estimation] not to be useful.”
“Many of the unknown individuals that come to us are from disenfranchised populations,” adds Williams, who is Black. “When you take off the table a parameter that could help somebody get home to their family … then it’s not the greater good.”
The fact that ancestry estimation sometimes works “does not in any way, shape, or form mean that [races] are biological categories,” stresses Agustín Fuentes, an anthropologist at Princeton University who is Hispanic and white. There’s no checklist of skeletal, physical, or genetic traits shared by all people of a certain race; in fact, there’s far more variation within racial categories than between them.
At the same time, “We’re all a product of our environment, evolution, and history,” says Ann Ross, a forensic anthropologist at North Carolina State University. People who share both deep evolutionary history and more recent social contexts, such as an industrial lifestyle or a history of discrimination, tend to also share some biological traits, including similar cranial measurements. She and some other anthropologists wonder whether current ancestry estimates are accurate enough to be helpful.
Winburn and a colleague did a study to try to find out. Among about 250 resolved cases in which forensic anthropologists offered an ancestry estimate, they correctly identified a person’s social race about 90% of the time, the team reported in April in the Journal of Forensic Sciences. But when anthropologists identified someone’s ancestry as “mixed” or “other,” they were wrong 80% of the time. Thirteen percent of unidentified people in the United States are listed as likely belonging to “multiple races”; another 21% are classified as “uncertain.” Based on her research, most of them are likely to be people of color, Winburn says. “Who are we serving and who are we failing with these continental ancestry estimates?” she asks.
Some forensic anthropologists who want to jettison the five commonly used racial categories say skeletal features could let them make more specific and useful distinctions. For example, cranial measurements can distinguish between Maya groups from Guatemala and Mexico and differentiate each from non-Maya people, according to a February study in the American Journal of Physical Anthropology. Such specificity could help identify people who die crossing the U.S.-Mexico border, the study argued.
The United States’s traditional racial categories don’t reveal such potentially informative variation—they obscure it, Ross says. People from any population in Mexico, Guatemala, and the rest of Latin America—as well as many populations in the United States—are all classified as “Hispanic,” a label Ross, who is from Panama, calls “biologically meaningless.” The proposed standard, which Ross is helping draft, will reflect that concern, promoting the use of “population affinity estimates” that refer to more specific social and biological groups.
Winburn supports the change to population affinity, but says more research is needed to ensure it lives up to its potential. For example, many of the field’s reference samples are themselves categorized by social race or continental ancestry and will have to be refined for population affinity studies. “We’re going to have to get creative,” she says.
In the meantime, Winburn won’t provide ancestry estimates until and unless she can take a population approach that doesn’t focus on reductive racial categories. She no longer wants to risk using a tool that she thinks may do more harm than good.