https://orcid.org/0000-0002-0751-7593
![Sample our Behavioral Sciences journals, sign in here to start your access, latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/110801/TOC-Subject-Sample-2021-Behavioral-Sciences.jpg"})
ABSTRACT
Drawing from the literature on the psychological harm of hate speech and the source magnification framework, we investigated the effects of hate messages from multiple sources on target group members in the context of Twitter. An online experiment conducted with Black Americans living in the United States provided evidence for source magnification of anti-Black messages. In the experiment, participants were exposed to one of three sets of tweets: hate tweets from multiple sources, hate tweets (of identical content) from a single source, and non-hate (control) tweets. The results showed that hate tweets from multiple sources, when compared to the identical hate messages from a single source and to the non-hate tweets, led to greater emotional distress, which, in turn, resulted in greater likelihood of attributing ambiguous social situations to racial prejudice. The findings provide initial evidence for source magnification of hate messages in the context of communication based on social media where simultaneous exposure to asynchronously posted hate messages from multiple sources is likely to happen. The findings also suggest that multiple-source hate messages, when compared to single-source hate messages, are more likely to indirectly result in potential relational harm.
Disclosure statement
No potential conflict of interest was reported by the authors.
Notes
1. This hypothetical example entailing job application–one of the situations presented in the Attribution to Prejudice instrument by Branscombe et al. (Citation1999)–was contextualized in terms of how job applicants with Black-sounding first names are less likely to form positive pre-interview impressions than Anglicized-named applicants (e.g., Watson, Appiah, & Thornton, Citation2011). We selected the name “DeShawn,” which is listed at the top of the 20 first names most commonly associated with Black Americans (Levitt & Dubner, Citation2005).
2. According to Hayes and Preacher (Citation2014), the sequential coding is useful when the categories of the experimental manipulation can be rank ordered reflecting the magnitude of the independent variable (e.g., none vs. moderate vs. high), which is true of the manipulation used in this study (i.e., control vs. single-source hate tweets vs. multiple-source hate tweets). We used the “mcx = 2” option of the PROCESS macro (Hayes, Citation2017, p. 563). For the control condition, both X1 and X2 were set to 0; for the single-source hate tweets condition, X1 was set to 1 and X2 to 0; for the multiple-source hate tweets condition, both X1 and X2 were set to 1.
3. The effect size for indirect effects was derived from partially standardized relative indirect effects as recommended by Hayes (Citation2017).
4. Given the inclusion of the two dummy variables (X1 and X2), we used the R2 change to test the overall direct effect of all experimental conditions on our outcome variable.
5. The indirect effect of the single-source hate condition (vs. control) on attribution to prejudice via emotional distress was significant, point estimate = 8.32 (effect size = 0.37), bootstrap SE = 2.51, 95% bootstrap CI = [3.60, 13.56]. However, the direct effect of the single-source hate condition (vs. control) on attribution to prejudice did not reach significance, point estimate = − 5.24, SE = 4.24, 95% CI = [−13.64, 3.17].