Together or apart: Modelling the inter-agency workings of emergency response multiteam systems

This thesis is concerned with the sub-optimized performance of emergency response systems in the UK. These emergency response systems come together during large scale civil emergencies to try and minimize the consequences of such events, with specific attention paid to protecting human welfare, the environment and the security of the UK. Such systems are comprised of individuals (referred to as agents throughout this thesis) from multiple agencies (i.e. fire service, health services, local authorities, private organizations, science advisors etc.) organized into multiple levels of command (i.e. operational at bronze level, tactical at silver level and strategic at gold level). In numerous past major incidents the emergency response system sub-optimized and did not perform as effectively or efficiently as it could. Inquests into these events have revealed that sub-optimization typically results from breakdowns in communication, collective understanding, coordination and decision making between the different agencies involved in the response. The aim of this thesis was thus to gain a greater understanding into why such sub-optimization occurs in emergency response systems – an organizational design I conceptualize as a multilevel multiteam system. Multiteam systems are a relatively novel concept to the organizational and management literatures, and thus our understanding of the functioning of such designs are currently still limited and worthy of further study. Computer simulation techniques were utilized within this thesis, specifically a relatively novel simulation technique known as agent-based modelling, in which agents with specific behavioural rules for acting and interacting are modelled with a view to determining the effect on aggregate level outcomes. I empirically tested the effects of theoretically derived generative mechanisms that could explain this system sub-optimization: social identity processes. These processes were isolated from the social identity approach (comprised of both social identity theory and self-categorization theory), which explains how people come to see themselves through their group membership, and interact with others on the basis of these memberships. The approach suggests that individuals have a bias towards favouring people within the same group, whilst treating those from ‘out-groups’ in a more derogatory fashion, and thus helps explain antagonism in intergroup contexts such as emergency response. Specifically, I considered how the level of commitment agents have to specific categorizations in conjunction with intergroup biases influence system-level communicative outcomes (specifically time taken, propagation and accuracy). The multilevel multiteam system design adopted in emergency response provides two salient groupings with which agents can categorize themselves that have not been considered in previous research: their originating organizational agencies (e.g. fire service, police service, local authority) and their level of command (e.g. bronze, silver or gold command), referred to in this thesis as horizontal categorizations and vertical categorizations respectively. It was found that high levels of commitment to horizontal categorizations and intergroup biases, both in isolation and in interaction, explain system sub-optimization in terms of communicative outcomes. Counterintuitively, it was also found that if agents had high commitment to their vertical categorization, then this could protect the system from sub-optimizing. The theoretical and practical implications of these findings are discussed including implications for designing interventions to prevent future communication breakdowns.