Use of spatial models for community program evaluation of changes in alcohol outlet distribution.

Alcohol outlets have geographic and spatial characteristics which are important for researchers to consider when planning and evaluating community prevention programs. Community-level data used in monitoring alcohol problems across community areas and over time exhibit spatial dependencies. Statistical procedures which depend on assumptions of independence may fail to give proper results in such a situation. Specific statistical techniques have been developed which adjust for the effects of spatial dependencies in measures across geography. This paper provides an example of the creation and use of computer Geographic Information Systems to display community alcohol outlets and alcohol-involved problems and the use of statistical analysis techniques which account for such spatial dependencies over time. This paper introduces the concepts, terminology, and justification for considering spatial analysis in community prevention planning, research and evaluation. The selection of a geographical unit of analysis will be discussed. Finally, as a demonstration, a variety of spatial statistics are applied to community spatial data for evaluation.