The tactic common to all previous studies of the criminal behavior of mentally ill persons is the comparison of arrests for a selected group of mental patients (number of arrests per number of mental patients in the sample) to the general population's rate (number of arrests per total population). This approach has produced evidence that the rearrest rate of mental patients is higher than in the past and that the correlates of arrest in the mental patient population are probably the same as those in the general population. The major limitation of this approach is its failure to address the relationship between mental illness and criminality at the individual level of analysis. By using a model which estimates and controls for both prevalence (the proportion of the group arrested) and intensity (the number of arrests per active individual), a more accurate assessment of the criminal activity of mental patients in relation to criminal offenders or the general population can be obtained. 29 references.