NCJ Number
227450
Date Published
2009
Length
27 pages
Annotation
This chapter demonstrates the importance of balancing treated and untreated study samples as well as the study population and the target population in order to accurately estimate effects of a crime-prevention intervention, generalize them to the study population, and extrapolate them to other target populations.
Abstract
The chapter describes the theory of causal analysis that has the broadest support within the community of social intervention evaluators interested in assessing causal claims. The theory can be made sufficiently general to cover both random control trials (RCTs) and alternative methods that may produce accurate evidence about the effects of an intervention. The chapter's first objective is to present a theory of causality that has become the received standard against which causal claims are assessed. The theory is known as Rubin's causal model (RCM). The RCM has a specific objective, i.e., to assess the magnitude of the effect of an intervention, which is referred to as the treatment effect. The size of the effect is important in justifying the costs of expanding an existing policy or reforming it. RCM has three key variables: potential outcomes, treatment or control condition, and the switch or mechanism by which individuals are allocated to either treatment or control conditions. RCM provides a formal basis to make causal inferences from study designs other than random assignment studies. It also provides guidance for addressing some of the common flaws that arise in many random assignment studies, as well as guidance on the reduction of bias in observational studies. Guidance is also provided for estimating causal effects for target populations that extend beyond the population from which the study samples were selected. In addition, it provides criteria against which evaluation findings can be assessed in the unbiased pursuit of accurate information for improving social conditions. 5 figures, 2 notes, and 38 references