There is no single definition of recidivism. However, all definitions share three common traits. Each has a starting event, such as release from custody, program completion, or placed on probation. Next, each has a measure of failure following the starting event, such as a subsequent arrest, a subsequent arrest for a violent crime, a conviction resulting from a subsequent arrest, or a new commitment resulting from a subsequent arrest. Finally, each has a recidivism window (e.g., 6 months, one year, two years, three years, etc,) beginning with the date of the starting event. Put together, an individuals is said to have recidivated if the individual has a failure event within the recidivism window. In contrast to an individual, a group of persons can have a recidivism rate, normally defined as the percentage of persons who recidivated (i.e., the percentage of all persons in the group who failed within the recidivism window).
Logically, recidivism rates will vary for the same group depending on the nature of the recidivism measure. For example, in most instances, recidivism rates will increase as the recidivism window increases. In contrast, recidivism rates will decrease as the failure event becomes less common; for example, given that some arrests do not result in conviction, a recidivism rate based on the failure event of conviction will be lower than will a recidivism rate based on the failure event of arrest.
And when the measure of recidivism is held constant, recidivism rates are likely to vary with the nature of the group studied. For example, given that males in general have a higher arrest rate than do females, it is likely that a recidivism rate based on re-arrest will be greater for males released from prison than for females released from prison. Similarly, using a recidivism measure based on the failure event of a new commitment to a jail or prison, it is likely that persons released from prison will have a higher recidivism rate than persons arrested for the first time.