Since Elon Musk's acquisition of Twitter, the company has increasingly disregarded or overlooked posts that violate the platform's hateful conduct policy. The changes have amplified concerns about the impact of toxic online speech. But does charged rhetoric online actually translate to crimes offline?

In a paper in the American Economic Journal: Applied Economics, authors Karsten Müller and Carlo Schwarz found evidence that higher Twitter use amplified an increase in anti-Muslim hate crimes after the 2016 presidential primaries—a period which saw an increase in inflammatory comments about Muslims by Donald Trump and his followers. 

The results come from event study regressions, based on county-level data from Twitter and the FBI. Figure 4 from the authors’ paper plots one of their main estimates.

Figure 4 from Müller and Schwarz (2023)

The dark blue line shows the estimates of the logarithmic differences in hate crimes between areas with more and less Twitter usage. The orange vertical line indicates the year leading up to the 2016 presidential primaries. The shaded light blue band represents 95 percent confidence intervals.

The chart reveals that before 2015, higher levels of Twitter use were not associated with more anti-Muslim hate crimes. But after the start of the campaign, when anti-Muslim sentiment increased on Twitter, a clear connection developed. The authors calculate that a one-standard-deviation increase in Twitter usage led to a 7 percent increase in anti-Muslim hate crimes per year in the post-campaign period. 

The researchers note that these results may be susceptible to confounding factors, but they get around this issue by developing a more suitable estimation technique, helping to isolate the true impact of online hate speech.