This report presents a system dynamics simulation model for the estimation and projection of the prevalence of cocaine use in the United States.
System dynamics models are developed and tested through a sequence of steps and according to principles intended to maximize their realism and their usefulness. The steps include problem definition, system conceptualization, model formulation, pattern replication testing, sensitivity testing, feedback analysis, and policy analysis. Technically, a system dynamics model is an interconnected set of difference equations that approximate a real-world system that operates in continuous time. Development of this model was guided by a variety of information sources, both numerical and descriptive. Starting with an orientation that was largely economic and supply-side, the focus turned increasingly to sociological, demand-side variables, until price was finally abandoned as a factor that explained historical trends; supply was viewed as being driven directly by demand during the historical period. The current model provides a good fit for most national indicator data for 1976-1987 regarding cocaine demand and supply as well as cocaine-related attitudes, morbidity, and mortality. It provides user population estimates broken down by product form (powder or crack), intensity of use (social vs. compulsive), and recency of use (past month, past year but not past month, ever but not past year). It also provides projections for every model variable through 1992. Sensitivity testing of the model has identified key points of leverage and uncertainty in the "cocaine system," with possible implications for both policymaking and targeted data collection. 34 references, 2 tables, and 39 figures
Downloads
Similar Publications
- A Scalable and Memory-Efficient Algorithm for De Novo Transcriptome Assembly of Non-model Organisms
- Coupling Gas Chromatography (GC) and Vacuum Ultraviolet (VUV) Spectroscopy for Forensic Applications
- Handwriting Evidence Evaluation Based on the Shape of Characters: Application of Multivariate Likelihood Ratios