Arguing for Criminal Justice Reform: Examining the Effects of Message Framing on Policy Preferences

AbstractAcross the U.S., policy makers are enacting criminal justice reform while limiting many of those reforms to low-level, non-violent offenders. Given the power the public wields in shaping policy, it is necessary to consider which arguments for reform are most effective and who is viewed as most deserving of those reforms. The current study finds that varying the argument for reform, such as highlighting racial inequities, does not affect support for reform policies, regardless of the type of offender. Additionally, although respondents generally supported reform policies, results revealed significantly less support for implementing those reforms for violent offenders. Finally, findings indicate that individuals who believe violent crime is more of a Black phenomenon are less likely to support reform. These findings demonstrate that, to enact broader reform, it is necessary to address the public's preexisting opinions about crime more so than debate which rhetorical arguments should be used.Keywords: public opinionreformracial biasviolent offenders AcknowledgmentsI thank Natasha Frost and Kevin Wozniak for their helpful feedback on earlier versions of this manuscript.Disclosure statementNo potential conflict of interest was reported by the author(s).Notes1 Analyses also conducted without excluding inattentive participants. Results from ANOVA and regression tests yield results consistent with those described in the manuscript.2 This number does not include the approximately 4 million people on probation or parole, which is also part of the carceral apparatus. Given the content of the vignettes, the materials in this study only described the prison and jail population.3 To determine whether randomization was efficacious, I report the frequency of each covariate by condition, and then conduct chi-square tests to assess whether a particular covariate is statistically different across the cells. As seen in the Appendix, the covariates are not statistically significant and, therefore, one can assume that random assignment was effective.4 Similar outcomes emerged when evaluating the interaction between message frame and racialized beliefs about who is committing drug crimes.Additional informationNotes on contributorsAdam DunbarAdam Dunbar is an assistant professor in the Criminal Justice department at the University of Nevada, Reno. He researches the intersection of race and the criminal justice system, focusing specifically on how attitudes about race, culture, and crime can help explain racial disparities in policing and punishment. His work has been published in the Journal of Experimental Criminology, Race and Justice, and Psychology, Public Policy, and Law.