We Have No Standard for Measuring Recidivism

The following article by County Attorney Bill Montgomery appeared in The Hill on November 15, 2017


In recent years, federal legislation aimed at reforming the criminal justice system, such as the Sentencing Reform Act and Corrections Act, has not passed Congress despite having bipartisan support. Numerous states have also considered and passed similar legislation. 

A common objective of the legislation is to reduce recidivism based on the belief that doing so will also reduce incarceration and its associated costs. The logic is compelling and I think that is correct. However, an important question in the criminal justice reform debate remains largely unanswered: How do we — and how should we — measure recidivism? 

After learning that the recidivism rate for prisoners released from the Arizona Department of Corrections is approximately 40 percent after the first year of release, I wondered how my state, Arizona, compares to other states. Research revealed that there does not appear to be a national consensus on how we currently measure or should measure recidivism. 

The Bureau of Justice Assistance website contains state recidivism rates in the state criminal justice profiles section. However, the bureau website only appears to have data from states on the three-year recidivism rates, and it also appears that states are reporting data from different years and using different definitions for recidivism making any relative comparisons difficult, if not impossible.  

This inconsistency in the data demonstrates the need for agreement on a clear and consistent definition for recidivism. With the understanding that no definition is perfect, the Uniform Crime Report illustrates the value of standard definitions for measuring criminal justice system outcomes. A uniform definition and set of measures for recidivism would improve data collection consistency, provide for relative comparisons and better inform public dialogue regarding the issue.

According to bureau data, the national average three-year recidivism rate is 37.6 percent, with Oklahoma claiming the lowest recidivism rate of 21.2 percent and Delaware claiming the highest recidivism rate of 67.3 percent. Comparatively, Arizona has a three-year recidivism rate of 42.4 percent, which is slightly above the national average. Using these numbers, it appears that a reasonable goal for Arizona would be to reduce the three-year recidivism rate by at least 5 percentage points, which would place us at or below the national average. 

Unsatisfied with knowing only how Arizona compared to other states at the three-year mark, I attempted to determine our standing over other time periods. Unfortunately, I was unable to find sufficient data to compare Arizona’s one-year and five-year recidivism rates with other states. 

The lack of national data for recidivism rates at one, five, seven and even 10 years, is a significant problem because it inhibits the ability of states to set reasonable goals for recidivism reduction and compare program effectiveness over time. It does not make sense to measure rates at only one period of time, three years, if the objective is long-term recidivism reduction. 

Consistent data collection over time will more accurately reflect the reality of recidivism in each state. That in turn will enable each state to set reasonable short and long-term goals for recidivism reduction, help identify successful strategies and help ensure accountability when state goals are not achieved. 

While states must have the ability to set their own goals reflecting the realities of crime in each jurisdiction, the federal government can assist them by collecting and publishing reliable and uniform data.  

Uniform definitions and data regarding recidivism should also enable us to accurately measure the success of different types of evidence-based practices and programming. This information should ensure that the most effective and efficient practices are made known and receive the most support. 

For these reasons, criminal justice reform legislation, such as the Sentencing Reform and Corrections Acts, should require the federal government to collect uniform data and publish recidivism rates for state criminal justice systems over various time periods, such as one, three, five and seven years. Doing so will help establish national benchmarks, assist states with data collection efforts, set reasonable recidivism reduction goals and allocate appropriate resources to achieve them.

Bill Montgomery is the Maricopa County Attorney. He is a West Point graduate and a decorated Gulf War veteran.