ABSTRACT
To establish causality one must forge a chain of three links (association, causal order, and lack of spuriousness), and the possibility that an antecedent variable not yet considered may account for the observed relation makes the third link inherently weak. To establish noncausality, one has only to break any one of these links. Despite the greater ease with which noncausality may be demonstrated, many assertions of noncausality in the delinquency literature turn out to be invalid. Some are invalid because the authors misuse statistical tools or misinterpret their findings. A factor is a sufficient condition for delinquency if its presence is invariably followed by delinquency. Examples of sufficient conditions are hard to find in empirical research. The nearest one comes to such conditions in delinquency research is in the use of predictive devices in which several factors taken together are virtually sufficient for delinquency.
