Can PSM Strategies Replicate RCT Results in Criminal Justice Research?
Date
2019-02-27
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Justice Research and Statistics Association
Abstract
Identifying the best practices for reducing crime and victimization requires ample research with a strong methodology. Understood as the “gold standard” in research design, the randomized control trial (RCT) has been shown to provide reliable and valid findings. Despite their methodological strengths, however, RCTs are not always feasible in many criminal justice settings. As a substitute, analysts have increasingly had to rely on other quasi-experimental designs and statistical techniques to address their research questions. One of the more popular techniques is propensity score modeling (PSM), which is designed to simulate the effects of an RCT experiment. Despite its recent increase in use, a critical question remains: Can PSM methods replicate the results from RCTs? Drawing on a selection of 10 RCT databases from the National Archive of Criminal Justice Data (NACJD), we tested the reliability and validity of five commonly used PSM techniques by comparing the accompanied effect sizes to those from the original RCTs. This webinar describes how the presenters tested PSM and presents their meta-analytic findings of the comparison to RCTs. It concludes with a discussion about the implications of our findings for evaluation research.
Description
Webinar
item.page.type
Video
item.page.format
Keywords
Propensity Score Matching, Research Methods/Practices, Crime Statistics/Data