|dc.description.abstract||Violence perpetration among adolescents is a global public health and social problem. The prevalence of
adolescent violence perpetration in Norway is lower than most countries in Europe. However, the
occurrence persisted significantly in less severe forms of perpetration. Few studies have investigated
regional variations in prevalence of adolescent violent behavior and its associates. Hence, the study
examined municipal-level variations in prevalence of violent behavior among Norwegian adolescents.
Based on a socio-ecological model, it investigated the variation in individual factors of health-risk
behaviors, familial predictors and municipal-level characteristics associated to variation in violent behavior
across municipalities. The study results may be an essential input for policy makers to design area-specific
policies and intervention strategies to reduce its prevalence.
The study particularly employed a cross-sectional design and used data from the Ungdata cross national
survey 2014-2016. It included adolescents from lower secondary school (N=119346) nested within 309
The multilevel (two-level) logistic regression result showed that 1.58% of the variability in presence of
violence lies between municipalities. The prevalence estimates of violent behavior indicated a small
variation between municipalities. All individual and familial factors were associated with the odds of the
presence of violent behavior, i.e. alcohol, cigarette, snus, hash, marijuana and cannabis use; depressive
symptoms; parental control and parents’ financial status. Of municipal-level factors, a high proportion of
higher education accomplishment was the only significant factor associated with violent behavior.
Municipal factors, to a slight extent, explained the small variation in violent behavior across municipalities,
but the individual and familial factors did not explain the differences. The multilevel logistic regression
model which included all levels of socio-ecological factors provided the best fit to predict violent behavior.
Considering the small variations in violent behavior across municipalities, the study suggests a greater
emphasis on violence prevention focusing on individual characteristics of health-risk behaviors and
microsystem influencers of familial predictors than of exosystem, i.e. municipal characteristics. Further
studies should examine longitudinal effects and other associates of violent behavior that were not
addressed in the study.||en