An Analysis of the Type of Criminal Activity Using the Logit Model

Abstract

Using the logit model, the authors investigate the determinants of the na ture of criminal activity an individual will pursue after release from prison. The likelihood of conviction increases the younger the offender is at release, the more previous convictions he has, the more rule violations he committed during his last prison sentence, the longer the follow-up period, and if he is an alcoholic or user of hard drugs. Among all possibilities of conviction and no conviction, an individual has a greater probability of being convicted of a crime other than an offense against persons or property if he is supervised and married when released and was relatively mature in age when first arrested. A n individual whose previous imprisonment was for a crime against a person is more likely to be reconvicted for a similar crime than for any other type of offense. If the previous imprisonment was for a property crime, this seems to increase the probability of conviction for both property and person crimes, but to decrease the probability of conviction for other crimes.