Background: The reported prevalence of bully-victims and aggressive/provocative victims varies quite considerably in previous research, and only a few studies have reported prevalence rates across grades. There is also a lack of detailed analyses of the extent to which victims are also bullies, and bullies are also victims. Aims: To study the prevalence of male and female bully-victims across grade/age and to establish the degree of overlap or relative size of the bully-victim group by relating them to all victims, all bullies and all involved students. Samples: Participants in Study 1 were 5,171 pupils in grades 5-9 from 37 schools. Study 2 comprised 12,983 pupils in grades 4-10 from 66 schools. Methods: The Olweus Bully/Victim Questionnaire was administered to the pupils in their ordinary classrooms. Results: The prevalence of bully-victims was low and mainly declined across grades. There were far more boys than girls in the bully-victim group. Bully-victims resembled victims only (pure victims) in terms of age-trends and bullies only (pure bullies) in terms of sex composition. The overlap of bully-victims with the total victim group was fairly small (10-20%) in all grades. In primary grades, bully-victims constituted about 30-50% of the total bully group, whereas in higher grades these proportions were considerably lower. Conclusion: Our analyses and overview of previous research suggest that the bully-victim group is small, and that the large variations across studies are mainly due to differences in choice of cutoff point. Bully-victims should generally be seen and treated statistically as a distinct subgroup. The relatively larger proportion of bully-victims in lower grades should be explored further.