The construct validity of the Bully/Victim Questionnaire (Olweus, 1996a) as a measure of bullying was examined. Although researchers have defined bullying and reactive aggression as two distinct types of aggression, this study examined how closely related they are using empirical data. Reports from students and teachers on measures of bullying and reactive aggression were compared for evidence of convergent validity (degree of similarity) and divergent validity (degree of difference). Comparisons of correlation coefficients indicate that convergent validity coefficients (0.26 to 0.85) are not higher than divergent validity coefficients (0.21 to 0.74). In addition, correlations between teachers and students differ. It is concluded that measuring bullying is problematic and that bullying and reactive aggression are highly correlated whereas teacher and student reports are not.