Statistical Modeling for Public Health Law

In this guide, for which results are expected in mid-2010, David M. Murray,  Ph.D., the chair of the Division of Epidemiology and Biostatistics at the School of Publich Health of  The Ohio State University, and his co-authors, explain the methodological--especially statistical or modeling--aspects of public health law evaluation studies. Murray compares the main statistical modeling approaches for evaluation of national, state, and local laws that may affect health indicators for the population; Discusses advantages and disadvantages of modeling approaches (such as pooled cross-sectional time-series models with state and year fixed effects; hierarchical or random effects models; and state-specific Box-Jenkins ARIMA type models with results aggregated across states using inverse-variance weighting and similar methods); and makes recommendations for advancing the methodological (especially statistical or modeling) quality of public health law evaluation studies. The paper is aimed at experienced researchers and scientists who might not be aware of current advanced thinking regarding the optimal statistical models for law evaluation studies.