Israeli regulations require teachers to subject pupils who participate in violent altercations to severe punishment. How teachers actually apply this policy, however, has not yet been researched. The current study investigates teachers’ intentions in disciplining students who brawl. Two hundred ninety-nine teachers read fictional vignettes about students who took part in fighting and are asked to describe the disciplinary measures that they would invoke. The teachers’ responses are quantified on the basis of a specially developed key that determines the severity of the steps that the teachers propose to take. Multi-level regression is utilized. The results reveal that students who explain their use of violence as a response to a violent provocation await only mild discipline if they can prove their claim; those who cannot prove it face severe punishment. Students who turn out to be the aggressors are punished even more severely. The research participants are inclined to punish students more harshly for involvement in retaliation than in involvement in self-defense. The differences in inclination to discipline students with and without attention deficit hyperactivity disorder (ADHD) are inconsequential. The discussion examines the importance of the need to prove guilt in weighing disciplinary action and the compromise between two needs—to discipline violent students and to treat self-defenders fairly—that teachers make when administering punishment.