ABSTRACT
Numerous datasets in policy analysis and public administration are pooled datasets in which an individual organization might be represented more than once because several years of data are included. This chapter shows that the basic logic of applying substantively weighted analytical techniques (SWAT) to pooled datasets and suggests a series of SWAT approaches that analysts of pooled data may find useful. In a change of pace, SWAT techniques will be used to analyze one of the most contentious and controversial policy questions faced by state policy makers —what can government do to control and reduce violent crime? The leap from the cross-sectional to the pooled case involves some modifications of the original SWAT process. All the approaches to generating estimators from pooled datasets are essentially efforts to get the error process to "behave." The chapter focuses on the five key policy variables—capital punishment, the number of executions, prison population, income equalization, and educational commitment.
