Indicators of lone actor violent events: The problems of low base rates and long observational periods.

A burgeoning literature exists on indicators associated with lone-actor terrorism, spree shooters, mass murders and other forms of targeted violence. Such studies of low- likelihood, high-impact crimes largely suffer from 2 interrelated problems: low base rates and long observational periods. These studies largely fail to consider whether risk factors are driven by temporal-cohorts within the wider observation pool or are uniform across the expanses of time under consideration. This paper compares a cohort of violent lone actors (composed of lone-actor terrorists, and solo mass murderer attackers) from 1990 to 2005 with a cohort from 2006 to 2013. We found no significant differences in terms of sociodemographic variables across the 2 temporal periods. The 2006 to 2013 cohort was significantly more likely to use the Internet in their attack planning, have a history of previous imprisonment, engage in multiple attack methods (e.g., a bombing and a shooting), and target ordinary citizens rather than a political or military target for example. The results also indicate that the 2006 through 2013 period contains fewer offenders who (a) had previous military experience, (b) made verbal statements to family/friends/wider audiences about their intent and beliefs, (c) socialized face to face with members of a wider network, (d) experienced being degraded or the target of an act of prejudice or unfairness, (e) experienced a recent stressor, and (f) interacted face-to-face with others holding a similar grievance. The conclusion discusses the research and operational implications of these findings.