Skip to main content
. 2022 Dec 7;2(1):pgac281. doi: 10.1093/pnasnexus/pgac281

Fig. 2.

Fig. 2.

Multilevel binomial regression estimating the effects of moral words, moral-emotional words, and further controls on the likelihood of receiving hate speech. Shown are the coefficient estimates with 99% CIs. User-specific random effects are included.