*3.1. Descriptive Statistics*

Correlations, means, and standard deviations for online hate bystanders, perpetrators, victims, and toxic online disinhibition are shown in Table 1. All variables were significantly correlated with each other.

Overall 53.7% (*n* = 761) of participants reported that they observed at least one incident of hateful or degrading writing or speech online, inappropriately attacking certain groups of people or individuals because of their sex, religious affiliation, race, or sexual orientation. Regarding the frequencies, 46.3% (*n* = 655) reported they have never had observed online hate, 18.6% (*n* = 263) reported observing online hate very rarely, 16.8% (*n* = 238) occasionally, 10.2% (*n* = 145) frequently, and 8.1% (*n* = 115) very frequently. Concerning online hate perpetration, 11.3% (*n* = 160) of participants reported that they had posted at least item of one hateful or degrading writing or speech online, inappropriately attacking certain groups of people or individuals because of their sex, religious affiliation, race or sexual orientation. Furthermore, 88.7% (*n* = 1256) reported they have never had posted online hate, 7.6% (*n* = 104) reported posting online hate very rarely, 1.8% (*n* = 26) occasionally, 0.9% (*n* = 13) frequently, and 1.2% (*n* = 17) very frequently. Regarding online hate victimization, 16.9% (*n* = 240) of participants reported that they have personally been the target of hateful or degrading writing or speech online because of their sex, religious affiliation, race, or sexual orientation. Additionally, 83.1% (*n* = 1178) reported they had never personally been targeted by online

hate, 9.6% (*n* = 136) very rarely, 4.3% (*n* = 61) occasionally, 1.6% (*n* = 23) frequently, and 1.4% (*n* = 20) very frequently.

There was a positive correlation between age and observing online hate, *r* = 0.10, *p* ≤ 0.001, and posting online hate *r* = 0.10, *p* ≤ 0.001, but not with victimization through online hate. Girls ( *M* = 1.37, *SD* = 1.36) reported more often than boys ( *M* = 0.93, *SD* = 1.24) observing online hate online (*t* (1403) = 6.35, *p* < 0.001, Cohen's *d* = 0.35). Boys ( *M* = 0.26, *SD* = 0.73) reported more often than girls (*M* = 0.11, *SD* = 0.47) posting online hate (*t* (1208) = −4.42, *p* < 0.001, Cohen's *d* = 0.24). However, no sex differences were found regarding online hate victimization.

### *3.2. Association between Online Hate Bystanders, Perpetrators, and Toxic Online Disinhibition*

The overall model was significant, *F*(6, 1357) = 11.87, *p* < 0.001, *R*<sup>2</sup> = 0.19, indicating a large effect (Cohen's *f* 2 = 0.53). As Table 2 illustrates, there were statistically significant correlates of online hate perpetration. While controlling for participants' age, sex, migration background, and socioeconomic background, increases in being bystanders of online hate were positively related to being perpetrators of online hate (*b* = 0.08, *SE* = 0.19, *p* < 0.001). Toxic online disinhibition was positively associated with being perpetrators of online hate (*b* = 0.11, *SE* = 0.02, *p* < 0.001). Although age, migration background, and socioeconomic background were not significant predictors, online hate victimization (*b* = 0.16, *SE* = 0.04, *p* = 0.007) and sex (*b* = 0.16, *SE* = 0.05, *p* < 0.001) were significant predictors of online hate perpetration.

**Table 2.** Coefficients of the model predicting online hate perpetration.


Note: OHB = online hate bystanders; TOD = toxic online disinhibition; SES = socioeconomic status; \* 95% BCa = bootstrap confidence intervals based on 5000 samples.

As Figure 1 shows, significant moderation effects were found between bystanders of online hate and toxic online disinhibition when predicting online hate perpetration (*b* = 0.07, *SE* = 0.27, *p* = 0.007). Probing the interaction further revealed that bystanders of online hate reported more online hate perpetration when they reported higher levels of online disinhibition (*b* = 0.14, *SE* = 0.02, *p* < 0.001 at +1 *SD*) and less frequent online hate perpetration when they reported lower levels of toxic online disinhibition (*b* = 0.04, *SE* = 0.02, *p* = 0.029 at −1 *SD*).

**Figure 1.** Simple slopes equations of the regression of online hate bystanders on online hate perpetrators at high and low levels of toxic online disinhibition.
