A new study published in Psychological Science has found that racial minorities face challenges when building connections on social media, and these challenges persist across political affiliations. The researchers conducted a large-scale field experiment on Twitter (also known as X) and observed that users were generally less likely to follow Black accounts back than White accounts, regardless of political orientation. This study highlights the widespread nature of racial discrimination in online interactions.
Racial discrimination is a pressing societal issue, affecting various areas of life from hiring practices to social interactions. However, much of the existing research on racial discrimination in social settings is based on surveys, which may lead participants to answer in socially desirable ways, masking real attitudes.
Recognizing the limitations of survey-based studies, the research team aimed to examine discrimination in a more natural setting. They hypothesized that people might demonstrate bias in subtle, often unconscious ways when there is less pressure to appear unbiased—such as when reciprocating social ties on social media.
The researchers also wanted to explore whether conservative and liberal users differed in their likelihood to reciprocate social ties with accounts of different races, especially given past studies that suggest political differences in racial bias. In doing so, the study sought to reveal real-world behavior in an online context that people frequently use but which hasn’t been deeply studied for these behaviors.
“Various studies using survey-based methods have found that conservatives display greater racial prejudice and discrimination than liberals, with more recent work suggesting that these differences have widened in the past decade,” explained study author Krishnan Nair, a postdoctoral research associate at the University of Illinois at Urbana-Champaign.
“However, survey-based work involving sensitive topics can be plagued by social desirability bias and self-promotional motives. Indeed, racial prejudice and discrimination is widely viewed as socially unacceptable, with this being especially the case on the political left. Accordingly, we wanted to investigate whether the liberal/conservative gap in racial discrimination existed when participants are unaware that they were being observed by others.”
To explore racial discrimination on social media, researchers set up a field experiment on Twitter. They created 18 research accounts that differed based on race (Black or White) and political orientation (liberal/Democrat, conservative/Republican, or neutral). These accounts represented generic male profiles and were designed to look authentic, complete with a small follower base and content consistent with each political leaning. For instance, a conservative account might retweet content from conservative news outlets, while a liberal account might share posts aligned with liberal perspectives.
The study targeted a politically balanced group of Twitter users who were randomly selected based on their engagement with either right-leaning or left-leaning content. These users, totaling nearly 6,000, were not informed they were part of a research study to preserve the naturalness of their behavior.
The central measure was whether Twitter users would follow back the research accounts that followed them. Follow-backs are an everyday way of connecting on Twitter and are largely private, limiting any concern about users’ self-image or social pressure. To ensure that all accounts appeared similarly active and to avoid bias in follow-back probability, researchers adjusted each account’s follower count periodically. They also implemented additional controls to prevent interactions between users in different experimental conditions.
The researchers found that Twitter users were less likely to follow back accounts with Black profile photos than those with White profiles, indicating a bias in forming social ties based on race. This trend was observed across the political spectrum, challenging the notion that racial discrimination on social media varies significantly by political alignment.
Both conservative and liberal Twitter users were found to discriminate similarly, suggesting that racial bias in online tie formation may be a more universal phenomenon than previously thought. In other words, while past survey-based research indicated that conservatives might hold stronger racial biases than liberals, this field experiment did not reveal significant differences in how each group treated Black versus White accounts in a less monitored online setting.
“While there were reasons to expect that the liberal/conservative gap in racial discrimination would be smaller when people are unaware they are being observed, the fact that we detected no differences between the two groups was still somewhat surprising to us,” Nair told PsyPost.
The researchers also tested for any potential “partisan effect.” They hypothesized that people might be more inclined to reciprocate social connections with profiles they assumed to share their political views and that such shared political orientation could moderate racial bias.
However, the data did not support this idea. Both conservatives and liberals showed a similar likelihood to follow back profiles with aligned political views, but this tendency did not reduce or increase their racial biases. Thus, shared political orientation did not play a role in alleviating racial bias in the context of forming social ties on Twitter.
The findings provide evidence “that racial discrimination in private behavior is not restricted to one part of the political spectrum,” Nair said. “It is bipartisan.”
Interestingly, in the study’s additional analysis on users’ blocking behavior, researchers found no significant tendency to block Black accounts compared to White accounts. This pattern suggests that the bias observed in follow-back behavior may be more about whom users choose to connect with rather than whom they actively avoid or disengage from.
The findings shed new light on subtle forms of racial discrimination in social media interactions. But the study has limitations. One limitation is that the findings are primarily relevant to politically engaged Twitter users, as the sample was selected based on engagement with political content. Future research could broaden the sample to capture a wider range of users and explore whether the observed biases persist in other online communities.
Additionally, the study focused exclusively on male profiles. Different racial biases may be at play for female profiles, and exploring gender differences could provide a more comprehensive understanding of discrimination on social media.
“Our research examined the treatment of Black males versus White males,” Nair said. “Thus, we can’t be certain of the extent to which our findings extend to the treatment of women and/or those belonging to other racial groups. To us, these are interesting avenue for future research.”
“I want to continue to study the extent to which liberals/Democrats and conservatives/Republicans are similar or different in their attitudes and behavioral tendencies, and the implications of these differences for understanding behavior in both the business context, and in the context of electoral politics. I also want to thank my co-authors Mohsen Mosleh and Maryam Kouchaki, without whom this project would not have been possible.”
The study, “Racial Minorities Face Discrimination From Across the Political Spectrum When Seeking to Form Ties on Social Media: Evidence From a Field Experiment,” was authored by Krishnan Nair, Mohsen Mosleh, and Maryam Kouchaki.
