A large amount of research suggests that aggressive children differ from their nonaggressive peers in the way they process social information in conflict situations. Using Crick & Dodge's (1994) Social Information Processing following Provocation Model as a theoretical framework, this investigation was undertaken to explore adults' social information processing (SIP) under the influence of negative affect as well as assessing how specific subcategories of aggressive behaviour (reactive and proactive aggression) relate to different stages in this model. Pilot Study I (n = 10) and II (n = 13) were carried out to develop an anger provoking laboratory aggression paradigm (Hand Slapping Game). Hypothesising that responding in this paradigm would be associated with other well-established measures of aggression (e.g. AQ) convergent validity of this procedure was established. In Study I (n = 55) the paradigm was modified (Chopstick Game) which led to the successful induction of negative affect (which failed to occur in the Hand Slapping Game). It was hypothesised that the new paradigm would have convergent validity, that there would be sex differences in early SIP steps and that early SIP steps (e.g. intent attribution, response access) would be associated with trait/state anger. There was partial support for the convergent validity of the new paradigm, sex differences in early SIP steps and associations between early SIP steps and trait/state anger. In Study II (n = 62) it was predicted that the paradigm would have convergent validity, and that further early SIP steps (e.g. attention, goal formation) would be associated with trait/state anger and trait aggression. Convergent validity of the paradigm was confirmed and there was general support for a relationship between early SIP steps and trait/state anger and trait aggression. In Study III (n = 35) it was hypothesised that later SIP steps (e.g. outcome evaluation, self efficacy) would be associated with (pro-active) aggressive traits. Only limited support for the relationship • between later SIP steps and trait aggression was provided. Combined, these studies provide support for the presence of relationships of social information processing with • both trait anger/aggression and negative affect in samples of adults. The author concludes by discussing theoretical and treatment implications, highlighting limitations of the present investigation and making suggestions for future research.