Measurement Invariance of Cyberbullying Test Across Age, Sex and Self-Reported Emotional and Behavioral Problems

Despite the growing focus on cyberbullying victimization across scientific, clinical, and community spheres, an agreement on the most effective evaluation methodologies for evaluating its effects remains unclear. This research sought to examine the reliability, construct validity and measurement invariance of the Cyberbullying Test (CT) across age, sex, and self-reported behavioral and emotional problems, measured with the Strengths and Difficulties Questionnaire (SDQ). To accomplish this objective, we recruited 767 Romanian adolescents (aged: 11-18; M = 13.4 years). All analyses were performed using R with packages such as "lavaan" version 0.6.19. Confirmatory factor analysis (CFA) confirmed the three-factor structure of the measure (RMSEA = .001, SRMR = .025, and CFI = 1.00), while multigroup CFA supported measurement invariance across age, sex, and self-reported behavioral and emotional problems. This supports the comparison of means across sex, age, and the presence of self-reported behavioral and emotional problems. The main differences between the latent means were: (1) Females tend to engage less as cyberaggressors compared to boys; (2) Adolescents exhibit a higher frequency of involvement in cyberbullying as cybervictims or as cyberwitnesses compared to preadolescents and less as cyberaggressors, and (3) Those included in the group with high level of self-reported behavioral and emotional problems (compared to those with low levels of problems) are frequently involved in cyberbullying in all the three roles, with associated large effect sizes. Regarding reliability, Cronbach's alpha was .95 or higher for the three subscales. The present study establishes the CT as reliable and valid, advocating for its value in measuring cyberbullying implications among young people. Theoretical implications, evaluation considerations, and prospective directions are examined.