Bundling violence: How do people trade-off combinations of violent acts?

Violence risk assessments are used in a variety of settings (e.g. forensic, psychiatric, public) to determine risk of violence of a given individual. The accumulation of data generated from risk assessments can be used by practitioners and policy makers to determine aggregate levels of violence so as to determine the demand of future services, particularly in the domain of violence reduction or prevention. For instance, in prisons risk assessments can reveal the extent to which, between prisons, and across time, violence is on the rise, so as to determine strategies to reduce violence. Often violent outbreaks occur in which different combinations of violence acts within each outbreak are observed. In order to determine future demands on prison services, what approach should be taken to assess if an outbreak in one prison is less than, equal to, or more violent overall than an outbreak in another prison? This is a particularly challenging question to answer because the aggregate score (total violence score) will significantly vary depending on how the severity of violent acts is taken into account; some risk assessments treat all violent acts as equally violent, some rank violent acts according to severity, and some use a weighted sum. In the present study we present a non-expert sample with putative violent bundles (combinations of violent acts), having taken place in a prison setting, in order to determine the most common way in which people intuitively aggregate violent acts.