Emotional Content in Autobiographical Memory through an Attachment Theory Framework

The current study investigates the relationship between individual differences in attachment style and the recall of autobiographical memories. According to attachment theory, affect regulation strategies employed by individuals high in attachment anxiety and high in attachment avoidance are likely to influence how information about the past is recalled. This study examines how attachment anxiety and attachment avoidance relate to the presence of negative emotions in autobiographical memories of upsetting events with important relationship figures (i.e., mother, father, or roommate). Participants included 248 undergraduate students ranging from ages 18-22 that attend a public university in the northeast. As hypothesized, individuals with an avoidant attachment expressed less sadness in their responses to the written narrative task, especially when prompted for memories involving their primary caregiver. Contrary to the hypothesis, anxiously attached individuals did not display higher levels of worry/fear emotions in their responses to the written narrative. Attachment anxiety was related to some differences in emotional content; however, this varied by relationship partner. The results provide evidence linking attachment style to emotion selection and retrieval in autobiographical memories of ‘upsetting’ events. Implications for close relationships and therapy are discussed.

Computer Simulation of Crowd Dynamics and Destructive Crowd Behavior

The social processes that lead to destructive behavior in celebratory crowds can be studied through an agent-based computer simulation. Riots are an increasingly common outcome of sports celebrations, and pose the potential for harm to participants, bystanders, property, and the reputation of the groups with whom participants are associated. Rioting cannot necessarily be attributed to the negative emotions of individuals, such as anger, rage, frustration and despair. For instance, the celebratory behavior (e.g., chanting, cheering, singing) during UConn’s “Spring Weekend” and after the 2004 NCAA Championships resulted in several small fires and overturned cars. Further, not every individual in the area of a riot engages in violence, and those who do, do not do so continuously. Instead, small groups carry out the majority of violent acts in relatively short-lived episodes. Agent-based computer simulations are an ideal method for modeling complex group-level social phenomena, such as celebratory gatherings and riots, which emerge from the interaction of relatively “simple” individuals. By making simple assumptions about individuals’ decision-making and behaviors and allowing actors to affect one another, behavioral patterns emerge that cannot be predicted by the characteristics of individuals. The computer simulation developed here models celebratory riot behavior by repeatedly evaluating a single algorithm for each individual, the inputs of which are affected by the characteristics of nearby actors. Specifically, the simulation assumes that (a) actors possess 1 of 5 distinct social identities (group memberships), (b) actors will congregate with actors who possess the same identity, (c) the degree of social cohesion generated in the social context determines the stability of relationships within groups, and (d) actors’ level of aggression is affected by the aggression of other group members. Not only does this simulation provide a systematic investigation of the effects of the initial distribution of aggression, social identification, and cohesiveness on riot outcomes, but also an analytic tool others may use to investigate, visualize and predict how various individual characteristics affect emergent crowd behavior.