ABSTRACT
In line with self-control theory and routine activities theory, several previous studies have shown that low self-control and involvement in certain online activities are associated with cybercrime victimization. However, these studies were based on cross-sectional, observational data and therefore it remains unknown whether these findings indicate causal effects on cybercrime victimization or simply reflect spurious associations. In this study, a longitudinal dataset (2008-2016) drawn from a large representative sample of Dutch households was used to estimate the effect of low self-control and fifteen online activities on five types of cybercrime victimization: online harassment, online fraud, hacking, virus infection, and malware infection. In addition to standard cross-sectional analyses, quasi-experimental research designs (i.e., fixed effects panel models and discordant sibling models) were used to control for hidden bias. Results from the cross-sectional analyses show that low self-control and time spent on certain online activities were significantly related to cybercrime victimization. However, after applying the quasi-experimental research designs most significant associations disappeared, suggesting that these associations do not reflect causal effects but are rather the consequence of unobserved confounders. It is recommended that future studies on cybercrime victimization and offending use quasi-experimental research designs that control for hidden bias.
