Intimate partner violence: comparison between empirical data and stochastic agent-based modeling

Intimate partner violence (IPV) is a complex and preventable issue that involves not only victims and aggressors but also individuals that surrounding them (e.g., families, friends etc.), and the communities and society where they live. Literature review presents several theories that try to understand IPV dynamics. Recent promising prevention programs are the bystander interventions in which all individuals might play a positive role in reducing IPV. So far, a few studies investigated the IPV dynamics by means of alternative approach in the field of social psychology such as agent-based modeling (ABM). The current research attempted to examine the dynamically interactive processes within IPV system by means of different methods (e.g., qualitative, quantitative and ABM) and tried to compare their results, in order to obtain a broader understanding of this phenomenon and the factors that might decrease it. Applying a socio-ecological IPV bystander intervention model, the empirical studies presented in this thesis sought to describe (First qualitative study) and evaluate (Second quantitative study) the factors that could affect individuals' decision to help someone involved in IPV situations. Both studies use online instruments to collect data. For this reason, first of all, three preliminary studies was conducted in order to assess how individuals behave in online environments. The online qualitative study included 49 Italian university students and found an interconnected group of individual, relational and situational factors that a potential bystander considers during their choice to intervene towards a friend victim of IPV. In particular, participants were more willing to help if they perceived the situation as more severe and they preferred to intervene if they would be with a peer group and know the aggressor. Moreover, some gender differences appeared. The online quantitative study involved 1128 Italian, Brazilian, and French-Canadian university students and allowed us to explore macrosystem influence on bystander intervention. Participants reported their intent to engage in helpful bystander behaviors towards a friend and a stranger and they were more willing to help the first than the second. Multivariate analysis showed that factors, that affect both the intent to help a friend and a stranger, were peer helping norms, self-efficacy (i.e., generalized and specific to deal with violence as a bystander), knowledge/training about IPV, and female gender. By comparing the results among the subsamples, Italian university students reported less intent to help a friend and a stranger. Italian, Brazilian, and Canadian participants differed for a few factors. Peer helping norms, bystander self-efficacy to deal with violence and knowledge/training about IPV were among the most important factors that influenced the intent to help a friend and a stranger. The two stochastic agent-based models simulated the influence of individual (i.e., aggressiveness) and contextual (i.e., perceived violence and received informal social support) parameters on the evolution of a couple at risk of IPV. Some simulation results were in accord with previous studies. The first model suggests that when aggressiveness is supported by gender specific violence, a “clique” of similar violent behavior might arise in a society. The second model recommends the importance to provide informal social support to victim and aggressor regardless gender. To conclude, the comparison of the results point out that members of own informal social networks have a key role on IPV dynamics because they might support or reduce the problem depending on the social norms which they are bearing in their communities and society.