**5. Limitations**

There are several limitations of this research requiring some discussion. First, the cross-sectional nature of the study's design limits the ability to draw any causal conclusions and temporal ordering of the main study constructs. Future research would benefit from longitudinal studies. Second, the data were exclusively collected through self-reports. Therefore, the observed relationships might be inflated due to shared method variance. In addition, measuring online hate solely through self-reports may affect adolescents' reports of these experiences. For example, some adolescents who perpetrated online hate might choose not to answer honestly for fear of consequences (i.e., restricted ICT use). Adolescents who think that they have a legitimate opinion or are just joking around might not recognize that they perpetrate online hate and consequently they might also underreport online hate. There is some research suggesting that people who recognize racism are more likely to perceive it as a deviant and

negative behavior [6]. More specifically, the items used to measure online hate refer to "inappropriate attacks" which might also lead to underreporting as adolescents might not perceive such behavior as inappropriate. On the other hand, it might be that some adolescents overreport perpetrating online hate to appear tough. Follow-up research should apply a multi-informant approach. Third, we did not control for involvement in other forms of cyber aggressions (i.e., cyberbullying, trolling), ICT access, time spent online, or online activities, all of which may have an impact on online hate perpetration. Future research should include theses control variables. Finally, we relied on single item measurement for the assessment of online hate. Follow-up studies should try to include validated scales to overcome concerns with one-item measurements (i.e., low content validity, sensitivity, and lack a measure of internal consistency reliability).
