RESUMO
BACKGROUND: Bullying is more likely to happen in schools than in any other location. The purpose of this study is to use decision tree analyses to predict specific risk factors for bullying to identify areas of interest for school-based bullying prevention. METHODS: We obtained data from the 2013 National Crime Victimization Study (NCVS) School Crime Supplement. We used case-wise deletion to create a sample with completed data based on the measure of bullying (N = 4967). The dependent variable for this study was bullying. We used chi-square automatic interaction detection (CHAID) to uncover predictors of bullying victimization in schools. RESULTS: Results suggest that 21.7% of the participants were bullied during the 6 months prior to the survey. Being distracted in class and being involved in a fight were the top statistically significant variables for bullying victimization in schools. Fear of being attacked and seeing hate-related words or symbols in school were also strong predictors of bullying. CONCLUSIONS: Bullying victimization can often be predicted. Therefore, school personnel can implement programs and policies consistent with the Whole School, Whole Community, Whole Child (WSCC) model to improve the social and emotional climate in schools and proactively reduce opportunities for bullying victimization.