Simulating the Effect of Social Influence on Collective Decision-making: The Case of Educator Reporting of Child Abuse

[摘要] The dissertation is directed toward a simulation study of social influence in small, task-oriented, groups composed of education professionals of differing status who hold differing beliefs about what constitutes the reportable physical abuse of elementary students by parents. It is asserted on philosophical grounds that simulation allows the development and refinement of computational, process-based, models that reflect the stratified nature of social and educational reality. More practically, simulation makes it possible to trace socio-cultural processes over time rather than simply settle for an input-output analysis. The possibility of more adequate explanation, and transformation, of social and educational systems makes simulation relatively superior to other research methods.Persons are modeled as computational objects that participate in social relations. The simulation is based on the social-psychological theory of expectation states and is expressly framed to allow comparison of: (1) a static model and process model of social influence and (2) the social theories of Giddens and Archer regarding the timing of social tie formation. National data are used to formulate models of professional belief concerning N = 111 cases of reported physical abuse. The chief findings are: (1) as applied to the group task of child abuse reporting, the collective decision outcomes obtained from the two models of social influence disagree by a number greater than that expected by chance alone, and (2) as modeled, there is essentially no difference in the simulated collective decision outcomes generated under each theory concerning the timing of social tie formation.Several refinements are needed. First, it is important to construct a more adequate characterization of the relevant beliefs of education professionals, best done via the construction of an ethnographic decision model for each professional type. Second, given the importance of social influence, the simulation should be extended to incorporate socio-linguistic, especially argumentative, behavior. Third, it is important to extend the model to take into account the constraining power of belief with respect to social action. These extensions would add believability to the model and its outputs, thus enhancing its power to inform social and educational theory and practice.

[发布机构] the University of Pittsburgh

[效力级别] Education