Context Matters: Affirmative Action, Public Health, and the Use of Population-Level Data
By Wendy E. Parmet, Elaine Marshall & Alisa K. Lincoln*
Last June, in Students for Fair Admissions (SFFA), the Supreme Court ruled that universities could not consider race in admitting students. In support of that decision, the Court dismissed the relevance of data about the varied experiences of racial groups, insisting that admissions decisions must be based solely on the experiences and merits of individual applicants. The Court’s rejection of group-level data evinces a critical misunderstanding about the uses and limits of such data that, if applied more broadly, portends troubling implications for health equity and health policy.
The defendants in SFFA, Harvard University and the University of North Carolina, claimed that race-based admissions helped to create a diverse class of applicants. In rebuffing that argument, Chief Justice Roberts, writing for the majority, argued that the diversity rationale falsely assumed that individuals share the experiences and assumptions that are common to their racial group.
Quoting a previous affirmative action opinion, Roberts stated that the Court has “long held that universities may not operate their admissions programs on the ‘belief that minority students always (or even consistently) express some characteristic minority viewpoint on any point on any issue.’” He explained: “The Government must treat citizens as individuals, not as simply components of racial, religious, sexual or national class.” Nevertheless, Roberts wrote, applicants could discuss, and universities could consider, how individual applicants “overcame racial discrimination,” as long as that was “tied to that student’s unique ability to contribute to the university.” But Roberts warned, “a student must be treated based on his or her experiences as an individual – not on the basis of race.”
Roberts’ insistence that universities focus only on an applicant’s own experiences with racism rather than population-level statistics is akin to the concept of the “ecological fallacy,” which cautions about the application of group level data to individuals within the group. For example, if the average income in a neighborhood is $30,000, we cannot assume that any one person in that neighborhood makes $30,000. There may be some people who make more and some people who make less. Perhaps no one makes exactly $30,000.
Roberts is thus correct in insisting that universities cannot be certain about the experiences or views of any one applicant of color based on what statistics tell us about the experiences of applicants of color as a group. Any one student of color may have had experiences more in common with white applicants than other students of color.
Nevertheless, Roberts and the majority misunderstand the ecological fallacy in thinking that robust population-level research is of little value to understanding the circumstances facing applicants. In fact, such data can tell us a lot about the probabilities faced by individuals within the group. For example, although data about the average income in a community does not tell us the income of any one individual, it does tell us something about the probability of their having a particular income. Thus someone who lives in a neighborhood with an average income of $30,000 is far less likely to earn $1 million than someone who lives in a community where the average income is $1.2 million. Further, group level data can open our eyes to the context experienced by individuals within the group. So, for example, even an individual who earns $100,000 per year is less likely to have access to excellent schools and other top notch social services if they live in a neighborhood with an average income of $30,000 than if they live in a neighborhood with an average income of $1 million. Context matters.
In her dissenting opinion in SFFA, Justice Jackson seemed to recognize the relevance of population-level data to college admissions, writing, “on average, white families with college degrees have over $300,000 more wealth than black (sic) families with college degrees.” While this does not mean that the families of every individual Black applicant whose parents have college degrees have $300,000 less in wealth than white applicants with college-educated parents, it does tell us that white applicants whose parents have college degrees are likely to come from wealthier families than Black applicants with similarly-educated parents. In other words, just as a high SAT score may be relevant to determining whether an applicant can succeed in college even though it cannot predict any one applicant’s future GPA, a person’s race offers an admissions committee relevant (though not conclusive) information about the odds of an applicant’s wealth. It also provides some insight into the environment that applicants of different races are likely to experience.
Using a second example related to health, Justice Jackson quoted Lee C. Bollinger and Geoffrey Stone who noted, “Black Americans experience the highest rates of obesity, hypertension, maternal mortality, infant mortality, stroke, and asthma.” Again, such data cannot predict whether any individual Black American will experience any of the conditions Justice Jackson cited, but it does tell us that Black Americans experience these illnesses at higher rates than white Americans. This important information regarding health disparities would be lost if health policymakers followed the SFFA majority in ignoring population-level statistics.
The SFFA Court’s rejection of population-level data also forces individual applicants to prove that they have in fact experienced what the data tells us they are more likely than members of other groups to experience. In effect, the Court treats the exception as the rule, compelling applicants whose experiences with racism were within the norm to emphasize how racism uniquely affected them. Using the SAT example again, this is akin to asking a student with a 1600 SAT score to explain how that score will predict their own success. More problematically, it treats racism as an unusual occurrence in the lives of applicants of color, rather than a common one. By so doing, it robs students of the ability to situate their own experiences within the context of population-level data. Indeed, how can a school assess “the courage” needed to overcome racism unless it understands what racism is and how pervasive it can be? Likewise, how can a school appreciate the grit and determination showed by first generation students unless it understands the barriers that such students as a group generally face? If every case is reduced to its own particularities, important meaning is lost.
Applied to public health, the SFFA Court’s rejection of population-level factors would also lead us to overlook the factors that cause the most illness on a population level. To see this, consider smoking. We now know that cigarettes are the number one cause of lung cancer. Yet, if you only considered each patient with lung cancer as an individual case, it would be hard to see smoking’s effect on lung cancer, as in each individual case, other factors (genetics, exposure to pollution) might also play a role. Only by comparing the rates of lung cancer in two populations, one that smokes and the other that doesn’t, does smoking’s powerful impact become readily apparent. Yet, if we followed the Court and rejected population-level data, we could not see smoking’s significance or work to reduce its toll.
The push and pull between focusing on data pertinent to a discrete individual versus population level research that contextualizes experience is pervasive in both legal analysis and public health research and policy. Both the individualistic and population-based approaches have their limitations; each also offers important insights. Courts cannot do justice, nor can public health save lives, if only one type of approach is considered.
* This essay was written with the support of the Robert Wood Johnson and W. K. Kellogg Foundations.
Wendy E. Parmet is George J. and Kathleen Waters Matthews Distinguished Professor and Professor of Law at Northeastern University School of Law and Director of Northeastern’s Program on Health Policy and Law.
Alisa K Lincoln is the Director of the Institute for Health Equity and Social Justice Research and Professor of Sociology and Health Sciences at Northeastern University.
Elaine Marshall is a Postdoctoral Research Fellow for Salus Populi, and the Institute of Health Equity and Social Justice Research at Northeastern University.
