Validation and Application of the LS/CMI in Nebraska Probation

Abstract

This study examined the validity of the Level of Service/Case Management Inventory (LS/CMI), as probation officers in the state of Nebraska use the tool. Study 1 evaluated the predictive validity of the LS/CMI by examining 19,344 probationer records over a 5.5-year period (January 2007-July 2013), and found that the LS/CMI total risk score demonstrated moderate predictive validity. Consistent with past findings, logistic regression showed that the total risk score predicted recidivism (return to probation) differently for nonminorities than for minorities. Furthermore, minorities scored higher than nonminorities on seven of the eight criminogenic factors. Study 2, a true randomized experiment, explored probation officer bias as an explanation for these findings, and found that training increased officers’ ratings of scores in some LS/CMI domains and decreased ratings in others. Most importantly, there was no evidence that officers demonstrated racial bias in administering the LS/CMI survey when scoring Black versus White target clients.

Aboriginal and Torres Strait Islander perspectives on forensic risk assessment

“…Multiple explanations have been offered for these discrepancies in accuracy, including differences in offending base rates (Chouldechova, 2017 ), assessor bias (Jimenez et al., 2018 ; Venner et al., 2021a , 2021b ) and the item content of risk instruments (Shepherd & Lewis-Fernandez, 2016 ). As regards the content of risk instruments themselves, although widely used risk factors including the central eight are acknowledged as relevant to offending across cultures (Gutierrez et al., 2013 ), several researchers have questioned whether risk instruments that are normed on predominantly Western populations may include risk items that do not adequately reflect the experiences of BIPOC cultures and/or fail to capture culturally specific risk and protective factors (e.g.…”

Digital prediction technologies in the justice system: The implications of a ‘race-neutral’ agenda

“…The ways in which digital prediction technologies can amplify this racial ideology has not been theorized. Meanwhile, there has been an accretion of these technologies and mounting evidence that the commonly used versions are overpredicting the recidivism risks of racialized people (Angwin and Larson, 2016; Jimenez et al, 2018; Shepherd and Sullivan, 2016; but cf. Dieterich et al, 2016; Skeem and Lowenkamp, 2016).…”