Impact of base rate information on estimated risk of recidivism of sex offenders in Japan

Abstract

ABSTRACT The purpose of this study was to examine the public’s estimates of recidivism risk among sex offenders and the ways those estimates are impacted by base rate information. Participants were 400 individuals (200 males and 200 females) in Japan recruited by a web-based research company. Each participant was allocated to either the experimental group with or the control group without base rate recidivism information and asked to estimate the recidivism rate of a sex offender after release from prison. The results show that the average estimated recidivism rates were 54.4% and 33.8% in the control and experimental groups, respectively, indicating a significant difference (Welch’s t (389.70) = 8.32, p < .001, Cohen’s d = .83). However, even the experimental group’s average estimated rate was much higher than the actual national base rate of 17.0%. Multiple regression analysis revealed that punitive attitudes toward crime and confidence in estimation significantly contributed to higher estimates of recidivism risk, whereas sex, age, and fear of crime did not. These results suggest that the public overestimates the recidivism rate of sex offenders and that simply presenting a base rate is not sufficient for risk communication. Practical implications and future research directions are also discussed.