Being the aggressor in dreams has been investigated relatively sparsely. On the one hand, these kinds of nightmares may represent continuity of aggression in waking-life, but on the other hand, being the aggressor in dreams may compensate failing (suppressed) aggression from waking-life. Two subtypes of aggressor nightmares should thus be distinguished: Those in which the dreamer is the primary aggressor and those, in which the dreamer reacts with aggression to a threatening event. We hypothesized that nightmares in which the aggressive event was caused by the dream-self may be related to waking-life aggression according to the continuity hypothesis, while in nightmares, in which the aggressive event by the dream-self was the response to threats caused by other dream-characters, the dream-self may be reacting in self-defense. Thus, participants who dreamt that the aggressive act was initiated by the dream-self should be more aggressive in waking-life (supporting continuity) than those who dreamt that the offending act was initiated by other dream characters, and the dream-self was only reacting (supporting compensation). Waking-life aggression was obtained with the Buss and Perry Aggression Questionnaire. Results showed that dreamers who dreamt that their dream-self initiated the aggressive act were more aggressive in waking-life than those who dreamt that other dream characters had initiated the aggressive act. Nightmares with aggression done by the dream-self thus support the continuity hypothesis of dreaming while aggressor nightmare with reactive aggression of the dream-self support the compensatory hypothesis of dreaming.