Sentence Length and Recidivism: A Review of the Research

In response to increasing concerns about jail and prison overcrowding, many officials across the U.S. have undertaken different efforts aimed at reducing the prison population, including Los Angeles District Attorney George Gascón. In December 2020, Gascón introduced a policy to eliminate all prison sentence enhancements, thereby sharply reducing sentence lengths for offenders committing crimes with aggravating factors such as causing great bodily injury or use of a deadly weapon. In Special Directive 20-08, the new DA claimed that research and data show that longer sentence lengths are more likely to increase recidivism relative to shorter sentences.

While the directive cites one unpublished study in support of the finding, it is unclear whether this finding holds when considering the wider body of research on the topic. Namely, research findings can vary quite a bit across studies, so we argue that a critical examination of the research literature is warranted. When considering the research on the relationship between incarceration and recidivism, many studies compare custodial with non-custodial sentences, and fewer examine the impact of varying incarceration lengths on recidivism. Considering that everyone subject to a possible sentence enhancement will be serving some type of custodial sentence, the latter is more relevant concerning Gascón’s recent policy change. To assess the claims mentioned in Special Directive 20-08, my colleague Kent Scheidegger and I conducted a detailed review of the research on this topic.

UPDATE: The full-length version of this paper has been posted through the Social Science Research Network.

It is clear that the current literature on the impact of incarceration length and recidivism remains limited by important methodological considerations and inconsistency of findings across studies. Some studies suggest that longer sentences may provide additional deterrent benefits in the aggregate, though this effect is not always consistent or strong. In contrast, many of the studies found no recidivism effects. While none of the studies suggest a strong aggregate-level criminogenic effect, some found increases in recidivism for certain types of lower-level offenders. In sum, it seems that deterrent effects of incarceration may vary for different types of offenders. Rather than having a straightforward explanation,  the effect of incarceration length on recidivism appears too heterogenous to be able to draw universal conclusions. Thus, a deepened understanding of this relationship is warranted to reliably and accurately inform policy.

At present, it is clear that research has not fully unpacked the costs and benefits of incarceration, particularly regarding the impact on recidivism. Contrary to popular belief, identifying individuals more or less likely to recidivate remains a daunting and complex task. There are many factors that might impact how someone responds to incarceration, thereby potentially dictating whether the person experiences a deterrent effect versus a criminogenic effect. These factors are things such as age, offense history, family support, experiences in prison, and gang membership, among others.  Relatedly, it is difficult to know whether criminogenic effects exist, and if so, under what circumstances.

Evidence-based policy has become more popular in recent years, though it is clear that policymakers and legislators alike could strengthen their use of these approaches. Namely, legislation is too often based on knee-jerk reactions and selectively cited research rather than a critical examination of the full body of research as a whole. While reliance on empirical studies should be encouraged, it is just as important to critically assess the methodologies and limitations of the research as this may condition the actual policy relevance of the findings. In the same vein, no policy should be based on only one study, no matter how sound that study may be. It is important to consider whether the findings are consistent when replicated in different populations, contexts, and over time.

The U.S. criminal justice system has a lengthy history of rapidly implementing policy change without fully considering potential consequences, often resulting in damaging effects that are difficult to reverse. We believe that drastic changes in sentencing policy may too result in unintended consequences. Once again, this highlights the need to consider the totality of findings across research studies and the contexts to which they apply when enacting policy change. For more detailed information on the studies reviewed, see CJLF’s recently published working paper, Sentence Length and Recidivism: A Review of the Research. The full-length version can also be accessed through the Social Science Research Network.