In criminal-justice reform, reducing crime without incarceration—and rehabilitating the already-incarcerated so they don’t return to prison—is something like the Holy Grail. Our new Manhattan Institute brief takes a hard look at one oft-touted intervention: cognitive behavioral therapy, or CBT.
CBT is a way of teaching individuals to change problematic thinking patterns, specifically those that might lead to unchecked actions and reactions. It’s been applied to treat numerous mental disorders—like generalized anxiety disorder and obsessive-compulsive disorder—and justice-related psychological problems including impulsiveness, drug addiction, and sexual offending. CBT programs are offered in many prisons nationwide, and sometimes to high-risk members of the general community.
The treatment method has been the subject of numerous studies and literature reviews, which provide hope—but also a reality check. Those reviews suggest that CBT can reduce recidivism by perhaps a quarter; if 40 percent of offenders, on average, will be reincarcerated, giving those offenders CBT could reduce that proportion to 30 percent. That is an impressive accomplishment, yet it’s clearly far from a panacea, and it comes with a number of caveats.
First, CBT studies often have methodological limitations. For example, they typically follow treatment recipients over a short time frame, such as a year, meaning that, for therapy-receiving inmates, we don’t know whether CBT prevented or simply delayed future criminal behavior. Additionally, some studies are not true experiments, meaning that positive outcomes attributed to CBT could have been caused by other factors, such as by an individual’s motivation to change or by the receipt of social services. This appears to be a serious flaw in the empirical case for CBT; a 2021 review in Lancet Psychiatry, which surveyed only randomized trials of carceral CBT programs, found that the therapy had no effect on recidivism in larger studies.
Another limitation is that CBT studies often evaluate the therapy against alternative versions of itself or against nothing at all. They fail to consider, in other words, whether non-CBT interventions—such as jobs programs, income supports, education, residential care, or adherence to psychiatric medication—could be equally or more successful than CBT in reducing criminal behavior. The Lancet Psychiatry review suggested as much, speculating that the therapy may have had no statistical impact because it didn’t address former inmates’ housing, employment, and financial difficulties.
Second, the rise of alternative CBT has muddied the empirical picture. While CBT, at its best, promotes personal responsibility, distinct “curricula” and therapeutic approaches have proliferated in recent decades, some deviating to include “processing past trauma” through “past-focused treatment.” Unsurprisingly, some applications of CBT—like those that are faithful to its key principles and keep individuals more regularly participate in lengthier treatment—seem more effective than others.
Further, it’s unclear what makes CBT, in any variation, “work”—or not work. For example, having a skilled administrator seems to matter, but being skilled doesn’t necessarily mean being a psychiatrist rather than a social worker, cyber therapist, or even witch doctor. This unknown matters, because the routine, real-life administration of CBT tends to be less careful than the version found in research or demonstrations.
Given this uncertainty, CBT programs should continue with careful evaluation, while those empirically proven to work should be expanded in a manner faithful to the treatment model. Policymakers must also consider the difficulty of establishing a program’s effectiveness, which researchers of the University of Chicago Crime Lab have demonstrated.
The lab, which has studied several CBT programs in a manner both more recent and more rigorous than previous literature reviews, drives home the reality that a program’s effectiveness often exists in the eye of the beholder. Consider, for example, an initiative called READI Chicago, which handles some of the city’s hardest cases—adults at extremely high risk of committing serious violence—and offers both CBT and subsidized employment. A recently published study, conducted by lab-affiliated researchers and others, estimated that the program cost $52,000 per participant.
That study found that READI Chicago had no measurable impact on the study’s main metric (an index combining arrests for various violent crimes). However, if one focuses on shooting-and-homicide arrests and uses a somewhat forgiving definition of statistical significance, that grave outcome declined a whopping two-thirds among program participants.
This confusing mix of promise and disappointment, alas, is the kind of thing that one tends to find when reviewing efforts to change human behavior. Some treatments can help and can be worth the money. But there are no silver bullets. The best that policymakers can do is to study what works, expand the best programs, and reform the rest.
Photo: LPETTET / E+ via Getty Images