The states should move away from collecting and using recidivism data as a single number that uniformly defines success or failure in sentencing and corrections policy, urged a recent report by the Urban Institute.
The report points out that recidivism, the most commonly used definition of correctional success, is “most frequently reported as a single statewide rate, which is too imprecise to draw meaningful conclusions and insufficient for assessing the impact of changes to policy and practice.”
“Making recidivism a meaningful performance measure demands that states employ a wide range of reoffending metrics. States should shift their thinking about recidivism from reporting one number to developing a portfolio of outcome measures that assess the impact of correctional interventions,” the report said.
The report, titled “Improving Recidivism as a Performance Measure,” analyzed the movements of the states that participate in the Justice Reinvestment Initiative to change the way they approach criminal justice policy development and adoption. These states “engage in a data-driven process that targets the drivers of the correctional population and prison costs,” and as a result, “more than 20 states have enacted reforms to reduce correctional spending so that a portion of the savings can be reallocated to programs that improve public safety,” the report found.
Citing various data from the states, the report outlined four steps that it suggests are necessary to make recidivism a meaningful performance measure:
The first is to define recidivism using multiple measures of success.
Recidivism reduction is the responsibility of multiple agencies and many different actors. States “should think of recidivism as a series of different performance indicators that must be carefully calibrated to the outcome they are intended to measure.”
In addition to return to custody, states should also include desistance, time to failure, crime severity and behavior change as indicators of success.
Second, states should develop protocols to ensure data collections are consistent, accurate and timely.
The current reporting on recidivism by the states is inconsistent at best. “Ongoing data collection and analysis is necessary for any performance measure.”
In the data collection process, states should assign unique identifiers and link data across criminal justice agencies, develop long-term records, collect contextual information and update changes in status.
Third, analyses of data should account for the underlying composition of the population.
Recidivism will be a more useful performance measure when we can compare recidivism outcomes across populations. As a necessary precursor, statistical techniques that account for the risk level of the population being studied must be employed for this comparison.
“Risk-adjusted” recidivism rates – expected recidivism rates compared with actual observed rate of recidivism – should be used in analyzing recidivism across populations, such as, for example, parolees and max-outs.
Last, states should package the findings on recidivism to maximize its impact, and get the results into the hands of decision makers.
Recidivism analysis “has to be packaged succinctly and disseminated to the right audience to influence the direction of policy and practice.
“To maximize the impact of recidivism as a performance measure, states should provide annual information to decision makers in all three branches of the government, report a variety of recidivism information to measure success and summarize findings for policymakers.”
