Factors Affecting Postarrest Dispositions: A Model for Analysis

This paper provides an illustrative application of Goodman's log-linear analysis to the complex analytical problems encountered by criminologists in research on the relationship between social disadvantage and criminalization. Substantively, the analysis deals with the question of whether and to what extent such variables as an arrested individual's age, race, occupational status, and record of incarceration influence the type and severity of the case disposition, and, if such variables are found to have effects, whether that effect is direct or mediated through effects of the nature of the offenses charged. The results of the analysis suggest that assertions of legal system bias against the socially disadvantaged require better evidence than has previously been offered in their support. Further research on the relationship between social disadvantage and criminalization must recognize the complexity of that relationship not only in theory but equally in the design and execution of research. Log-linear analysis of multidimensional contingency tables appears to be an extremely useful tool for specifying the degree and nature of that complexity.