The coevolution of bullying and friendship networks and the moderating effects of classroom bullying popularity norms were examined in a sample of 965 students (52.1% boys) in 22 fourth- and fifth-grade classes. Longitudinal social network analysis showed that children were more likely to bully their friends\' victims (bully influence effect) and to be bullied by their friends\' bullies (victim influence effect); two children bullying the same child were likely to be friends (bully selection effect), and two victims bullied by the same child were likely to be friends (victim selection effect). Bullying popularity norms served as moderators, and the bully selection effect was significant weaker in the context of low bullying popularity norms. This study adds understanding of bullying as a group process and provides implications for preventing school bullying.

Although a number of studies have been conducted on the subject of bullying, there is an absence of a standard instrument which can map the bullying network and identify the groups involved in it. The current study aimed to provide a standard instrument for screening bullies and victims in the classroom, which has validity and reliability indicators to map the network of relationships between the bullies and victims. In this regard, 337 high school students (176 males and 161 females) were selected via cluster sampling and, afterward, the Screening Instrument for Bullies and Victims in Classrooms (SIBVC) was used in the sample group. To examine the concurrent validity, Matson Evaluation of Social Skills with Youngsters (MESSY) and Adolescent Peer Relations Instrument (APRI) were used. The results depicted that the individuals\' scores of bullying and victimization in SIBVC are positively and negatively related to their scores in MESSY subscales, respectively. Also, individuals\' scores of bullying in SIBVC is positively related to bullying scores and negatively related to victimization scores in APRI. Cronbach\'s alpha method was used to determine the reliability of SIBVC, and the results showed that in all classes, the obtained alpha was higher than .72. Two weeks after the first run, the instrument was reimplemented and the reliability coefficient was (p < .001, r = .97 for bullying and r = .96 for victimization). The results depicted that SIBVC benefited from favorable validity, and it was able to identify the bullies and the victims quite well in the classrooms; also, SIBVC scores had favorable reliability over time. Criteria such as the use of dyadic nomination and network mapping among the groups involved in bullying are the benefits of SIBVC, which make it a useful and new instrument in bullying research.