Studies on race are a dime a dozen: researchers examine its relationship to everything from elementary school test scores to who's most likely to develop diabetes to which groups are overrepresented in ethnic militias to who Americans vote for, and we read about the results in news stories that are supposed to help us makes sense of the world.
But two Ivy League scholars say race is actually much more complicated than decades of social science research has acknowledged, and they're working to change that.
In their paper, "Race a Bundle of Sticks: Designs that Estimate Effects of Seemingly Immutable Characteristics," which will be published in the Annual Review of Political Science, Harvard's Maya Sen and Princeton's Omar Wasow explain that people who do quantitative research on race typically treat it as a single, fixed trait — what scientists call an "immutable characteristic."
Instead, they argue, quantitative researchers should acknowledge that any one person's racial identity is more like a collection of many different factors — from skin color, to neighborhood, to language, to socioeconomic status. With this insight, it becomes possible to study race not as a single, unchanging variable, but rather as a "a bundle of sticks" that can be pulled apart and carefully examined one by one.
The problem with oversimplifying race
Political science and other social science researchers often treat race as an "immutable characteristic" —something that we're born with and keep throughout life. That's how a lot of people do think about it. But this view makes studying the impact of racial categories on our lives really hard.
Why? First, it means race can't really be manipulated the way other variables can: you can't assign some people to be white, or black, or Latino for the purpose of a controlled experiment.
Second, because race is assigned at birth, it's tied up with a lot of other things in our lives. It can affect how a person is raised, how they're educated, the opportunities they have, and their outlook. This means that when researchers say they're looking at race, they're also getting all the stuff that goes hand in hand with it, making it impossible to tell whether they're seeing the effects of race itself or something that's a side effect of race or corollated with race.
Finally, racial labels can mean different things to different people. "Latino" can refer to a first-generation Mexican American from Los Angeles and a fourth-generation Puerto Rican from the Bronx. So research based on a racial category can capture broad swaths of people who have little in common except the way they choose to describe themselves, creating messy results.
How oversimplifying race makes for unhelpful research
When researchers ignore all the individual "sticks" and treat race as an immutable characteristic, it "misses a lot of the story," Wasow says.
Imagine a study that measures racial disparities in test scores by simply comparing results for black kids and white kids. It might tell us that there's a gap between these two groups. But it says very little about why. The lives of African American and white children tend to differ in so many ways that observing such a gap raises more questions than it answers.
"In a lot of traditional social science, by using a simple measure of race, we are often smoothing over all of this complexity and ignoring all of the ways that what we're calling race might be class, it might be neighborhood, it might be the experiences of our forbearers — it is really the composition of many, many experiences," Wasow says.
Worse, he says, these studies that simply point to race as a cause of different life outcomes miss the chance for policy interventions. For example, you can point to a black-white test score gap, but that doesn't tell you anything about what you might fix to close it. However, if you were to look at race in more detail and identified one "stick" that was responsible for black kids having lower test scores — say, access to better nutrition — that might provide something to work with in terms of solutions.
Here's another example: imagine a study that measures the impact of a campaign ad in which racial cues are being manipulated to change voting behavior. Wasow says this type of research can produce confusing results, too, if it doesn't look deeper into what exactly about race is being triggered by the ad and causing people to vote differently. Is it possible that it's a signal related to just one or two sticks, like socioeconomic status, or language? In the paper, he and Sen argue that researchers need to be precise about what type of cues are at work in their studies on bias, what "sticks" they relate to, and how these sticks are associated with race as a whole.
Why it makes sense to think of race as a "bundle of sticks"
"The core idea in the paper is that it is possible to study effects of race if we move away from a simple, fixed, biologically-determined model to a more complicated, fluid and socially-determined model," says Wasow.
The proposal is to take the idea that race is a social construct — which is gaining a cultural foothold as more people realize how complicated racial identity can be — and translate it to research.
Wasow and Sen propose that the "bundle of sticks" that make up racial categories includes things like societal values, status hierarchies, culture traits, physical attributes, diet, religion, ancestry, institutional power relationships, and education. Many of those traits are tightly linked to what we think of as race, but the key is that they can also be manipulated in the context of an experiment.
We could get better information on how race affects our lives, they argue, if researchers who study race would acknowledge this and carefully examine just one "stick" at a time: "Thinking about race as having constituent parts can clarify what precisely is being estimated when scholars attempt to understand how race and ethnicity operate in the world."
Here's a chart from their paper that contrasts the popular framework of looking at race with the constructivist/"bundle of sticks" one they embrace:
Which stick in the racial bundle is really affecting our lives?
Wasow and Sen write that, by focusing their studies on a single element (or stick), researchers can get much better information on the particular ways in which race as shapes life outcomes.
Remember when polls showed that African-Americans in California voted against the legalization of same-sex marriage at higher rates than whites? Wasow said examining the impact of the individual sticks that make up race helps revealed the actual factors that influenced opinions on the issue. In this case, religiosity was much more strongly associated with being against gay marriage than was race. Understanding the role that one stick played gives us much more accurate and richer information about voting behavior than simply saying, "black voters oppose gay marriage."
This "bundle of sticks approach" can also be useful when researchers are trying to determine how racial bias affects public opinion, political behavior, and hiring choices. Using signals of race — such as by varying a dialect over the phone when inquiring about housing or by varying the apparent ethnicity of a name on a resume when applying for jobs — they can measure levels of racial discrimination without needing to expose someone to the whole bundle that is race.
Wasow acknowledges that there are many important questions that are beyond the scope of this approach. "The framework we outline is particularly useful for scholars attempting to measure effects of race in experiments or studies that are like experiments. To study the role of race in historical processes or institutional bias where it may not be possible to run experiments, other methods may be more appropriate."
Either way, says Wasow, the key is recognizing that race is comprised of many parts, and that opens up the possibility of studying some individual "sticks" in isolation. This, he said, "helps us see what's actually doing the work," when it comes to how race affects our lives - which after all, is really what most of us want to understand.
Further reading: 11 ways race isn't real