Bullying is commonly defined as repeated aggressive behavior in which there is an imbalance of power or strength between the two parties (Nansel et al., 2001 ; Olweus, 1993). Bullying behaviors may be direct or overt (e.g., hitting, kicking, name-calling, or taunting) or more subtle or indirect in nature (e.g., rumor-spreading, social exclusion, friendship manipulation, or cyberbullying ; Espelage & Swearer, 2004 ; Olweus, 1993 ; Rigby, 2002). In the last 10 years, studies on bullying have focused on different integrative approaches of this complex problem. One of the most investigated approaches is the social- ecological perspective which takes account of reciprocal interplay between individuals involved in the bully/victim continuum and his complex contexts (Bronfenbrenner, 1979 ; Espelage & Swearer, 2004 ; Olweus, 1993). Bullying does not occur in isolation. This phenomenon is encouraged and/or inhibited as a result of the complex relationships between the individual, family, peer group, school, community, and culture. This study examined an ecological perspective on bullying behaviors. The aim of this study was to investigate differences between individual characteristics, family, peer, school and neighborhood contexts of victims, bullies and noninvolved children. A total of 880 primary school children (10 to 16 years old) participated in the investigation during one school semester. For testing the differences between groups we used the one way ANOVA for independent samples. Overall, the results of this study suggested statistically significant differences between bullies, victims and noninvolved children for individual characteristics and all aforementioned contexts. On individual characteristics there were statistically significant differences in empathy level, where bullies had lower levels of empathy than victims and noninvolved children, impulsivity, where bullies had higher levels of impulsivity than victims and noninvolved children, and time spent on media, where bullies had more time spent on media than victims and noninvolved children. However, there were no statistically significant differences in sex and age between these three groups. In family context, especially on parents’ behavior, we also found statistically significant differences. Bullies and victims had parents who used more negative discipline and psychological control in raising their children and showed less acceptance of their children than parents of noninvolved children. Bullies also had parents who gave them less autonomy and less supervision than parents of victims and noninvolved children. The parents’ positive discipline and permissiveness were not statistically significant. In peer context, there was only one statistically significant variable ; peer acceptance where noninvolved children were the most accepted by peers, bullies were a little less accepted and victims were the least accepted by peers. The difference in the number of friends was not statistically significant. In the school context, all measured variables were statistically significant. Noninvolved children have better school grades than victims and they also feel more safety in school than victims and bullies. For bullies, the school climate was perceived as the most negative, for victims it was less negative, and for noninvolved children it was positive. For the last investigated context, neighborhood, we found statistically significant differences. The bullies perceived the neighborhood as the most dangerous ; by the victims it was perceived as less dangerous, and for noninvolved children as the least dangerous. We can conclude that there are major differences in individual characteristics as well as in multiple contexts between children with a different bullying status. Knowing these differences, we can direct our efforts at developing focused intervention programs for all children involved in bullying behavior.