A new Yale-led study shows that Americans who self-report having positive attitudes toward non-Americans actually often demonstrate implicit anti-immigrant biases.
These automatic negative impressions of non-Americans correlate with anti-immigrant voting patterns in ballot measures held across the United States over the past 28 years concerning anti-immigrant policies, such as proposals to restrict immigrants’ access to public services and drivers’ licenses, according to the study.
“Anti-immigrant bias is often discussed in a politicized context that assumes that conservatives tend to hold negative feelings toward immigrants while progressives typically have more positive views of non-Americans,” said Melissa Ferguson, professor of psychology in Yale’s Faculty of Arts and Sciences and the study’s senior author. “We provide strong evidence showing that implicit anti-immigrant bias spans the demographical and political spectrums and that, at the county level, these automatic impressions predict the results of regional ballot measures that approved policies that clearly hurt immigrants.”
The paper, published in the journal Scientific Reports, encompasses nine different studies — separated into three sets — that examined people’s implicit (or automatically revealed) and explicit (or self-reported) evaluations of non-Americans and how their automatic appraisals affect real-world outcomes.
The first group of experiments examined people’s preferences for three labels widely used to identify non-Americans — “aliens,” “foreigners,” and “noncitizens” — relative their reactions to the label “American.”
According to the findings, 91% of participants demonstrated an automatic preference for the label “American” relative to the other three labels, which each generated equivalent levels of negativity. At the same time, study participants self-reported a slight but statistically significant explicit preference for the three non-American labels to “American.” In fact, only 33% of participants exressed an explicit preference for the label “American.”
The researchers reasoned that participants’ strong automatic preference for the “American” label over “non-American” labels could be specifically related to seeing the abstract labels in isolation without reference to individual people.
In the next batch of four experiments, however, researchers focused on how participants responded when the same labels were attached to individual people. In those experiments, participants were shown photos of two people’s faces, one labeled “American” and the other identified using one of the three labels for non-Americans. The majority of participants (61%) again demonstrated an implicit preference for the individuals labeled “American” relative to individuals associated with the other three labels; similar to the first set of studies, considerably fewer participants (44%) showed the corresponding explicit preference. This effect was consistent regardless of the race or gender of the individuals depicted.
“The fact that people are showing automatic negativity towards someone they know nothing about except seeing their picture and reading a label a few times is very telling,” said lead author Benedek Kurdi, assistant professor of psychology at the University of Illinois Urbana–Champaign, who conducted the research while a postdoctoral associate at Yale.
The effects remained consistent across participants of different genders and racial identities. Participants who identified as naturalized U.S. citizens — people who were not born in the U.S. but who have become U.S. citizens — exhibited the equivalent negative bias toward non-Americans as native-born citizens did. Ideology was the only variable by which participants’ automatic reactions differed. Participants who identified as politically conservative exhibited slightly higher levels of implicit anti-foreigner bias than self-identified progressives did, although progressives also demonstrated a significant degree of anti-foreigner bias.
“In other studies, such as ones that examine racial preferences, we find that people’s own identities strongly influence their attitudes,” Kurdi said. “That’s not the case with this study. We found that as long as you are a U.S. citizen, whether naturalized or native-born, you’re likely to show an implicit anti-foreigner bias.”
The last round of studies examined how implicit bias against non-Americans relates to real-world consequences. In the first experiment, the researchers gauged how participants’ pro-American/anti-foreigner implicit and explicit preferences correlated with their views on immigration policy. They also tested how one notable stereotype, which equates being white to being American and being Asian to being foreign, influences policy positions.
For the experiment, participants were asked to rate their support for 12 hypothetical immigration policies modeled after 18 actual ballot initiatives put before voters in the United States between 1994 and 2022. The researchers found that demonstrating an automatic pro-American preference predicted anti-immigrant policy positions.
In the final study, the researchers obtained county-level estimates of bias from 18 years of archival data collected by the Project Implicit educational website that measured the implicit white–American/Asian–foreign stereotype. They combined this data with anti-immigrant vote shares from the 18 ballot measures cited in the previous experiment, which were held in 10 states and concerned policy questions such as whether driver’s license applications should require proof of citizenship or immigration status.
The researchers found a significant correlation at the county level between the strength of the automatic “American-equals-white” stereotype and anti-immigrant vote shares. The effects were strong: A county with the lowest level of implicit anti-foreigner bias was predicted to have only 41% anti-immigrant votes, whereas a county with the highest-level of implicit anti-foreigner bias was predicted to have 54% anti-immigrant votes.
The result does not address individual behavior, as the bias data and voting data came from different individuals, but it does show that implicit stereotypes aggregated at the county level predict anti-immigrant voting patterns, the researchers said.
“In the case of the regional ballot initiatives, what people say they feel about the white/Asian stereotype doesn’t at all predict the vote outcomes,” Ferguson said. “However, their automatic reactions closely correlate with anti-immigrant voting patterns in these regional ballot initiatives.”
The study is coauthored by Keitaro Okura of Yale and Eric Hehman of McGill University.
