Narcotics and crime: A causal modeling approach

The present study examines the relationships between narcotics use and criminality utilizing a confirmatory structural equation modeling approach. Latentvariable and manifest-variable models are used to represent causal relationships between narcotics use and criminality in both the postaddiction and the pretreatment periods of the addiction career. Latent-variable models include a preexisting deviance construct as an exogenous factor, and manifest-variable models include a measure of drug trafficking, or dealing, as an influencing variable. Both inclusions are made as an attempt to minimize specification error in the models. The results are consistent in showing that, while simultaneous relationships between narcotics and criminality are clearly demonstrable, the prediction of either variable from the other across time cannot be demonstrated. It is concluded that causal relationships between narcotics and criminality are probably not characterized by an appreciable or identifiable time lag. Other theoretical considerations suggested by the modeling results are discussed, such as the role of idiosyncratic approaches to the economics of maintaining addition.