Meta-Analysis of Non-Causal Descriptive Findings: Lessons from Three Studies

• In Event: Beyond ‘What Works:’ Advancing Understanding of Crime through Systematic Reviews

Abstract

Meta-analysis has been used extensively to synthesize findings from experiments, quasi-experiments, and non-experimental causal studies. However, much research is descriptive – showing a statistical distribution of a phenomena. In this paper, we describe how we meta-analyzed three of descriptive literatures showing crime concentration at places, among offenders, and among victims. The principle hurdle was to find a way to summarize findings from studies reporting results in numerous ways, and to find a suitable analogue to the effect size. We used visual binning and a logarithmic function. We describe the advantages and disadvantages of these approaches.

Authors

• John Eck, University of Cincinnati

• YongJei Lee, University of Colorado Colorado Springs

• Natalie Martinez, University of Cincinnati

• SooHyun O, University of Cincinnati