Ecological Modeling of College Drinking
The overall goal of the project was to establish a new experimental paradigm for alcohol policy studies based in mathematical modeling. To accomplish this goal a multidisciplinary team of researchers with expertise in alcohol policy research, theoretical ecosystem modeling, computer modeling and simulation, and software engineering have been brought together with the goal of developing predictive process-based models of college drinking that will be used to test central study hypotheses in a multi-campus experimental system.
Hypothesis: Campus level measures indicating greater physical and social availability – in terms of fewer alcohol control policies, greater alcohol advertising, higher densities of alcohol outlets, and less alcohol policy enforcement – are associated with higher levels of individual drinking behavior and outcomes across study campuses even after controlling for individual level predictors of college drinking.
Four major objectives will be achieved in the process of model development and testing.
Objective 1: To generate a cognitive model of college drinking behavior that successfully captures the multilevel causal dynamic associated with individual-context mismatch that exists at many US colleges.
Objective 2: To formulate a set of deterministic models of college drinking that simply capture the multilevel, multidimensional dynamic evidenced in the cognitive models.
Objective 3: To transition the candidate models through tests of model type, model complexity and parameter values to arrive at the most robust models resulting from the transition process.
Objective 4: To test study hypotheses with the resulting process-based models using an experimental system that requires successful prediction of college drinking outcomes replicated across multiple colleges.