STATISTICAL-METHODS FOR THE ANALYSIS OF CRITICAL LIFE EVENTS

Problems such as family disruption, mental illness, child abuse, and j uvenile delinquency are often conceptualized as episodic events that d efine the success or failure of social intervention. This article exam ines the methodological problems associated with the analysis of event data by comparing five alternative statistical specifications: (1) mu ltiple regression, (2) discriminant analysis, (3) logistic regression, (4) tobit regression, and (5) event history analysis. Each procedure is used to estimate risk factors for out-of-home placement among child ren in families that received intensive family preservation services. The accuracy and precision of models produced by each method are discu ssed. Implications for the analysis of event data are explored.