Testing Causal Models of Delinquency

Abstract

Researchers studying the etiology of delinquency often use correlational data in an attempt to test causal theories. This article shows that several specific methodological problems have hampered much of the research in this area. These problems include confusion over the definitions of measurement model and causal model, use of improper methods for testing each sort of model, and misunderstanding of the conditions under which data collected at a single time period can be used to draw conclusions about causal relationships between variables. Data presented in several studies are reanalyzed here to illustrate the use of methods that can help to solve these conceptual and methodological problems. It is concluded that we may know less than we thought we did about the causes of delinquency.