Details

Autor(en) / Beteiligte

• Giguère, Guy
• Lussier, Patrick

Titel

Debunking the psychometric properties of the LS\CMI: An application of item response theory with a risk assessment instrument

Ist Teil von

• Journal of criminal justice, 2016-09, Vol.46, p.207-218

Ort / Verlag

New York: Elsevier Ltd

Erscheinungsjahr

2016

Quelle

Applied Social Sciences Index & Abstracts (ASSIA)

Beschreibungen/Notizen

• Andrews, Bonta, and Wormith (2004) presented the LS\CMI as a fourth-generation instrument designed for the assessment of risk and needs of adults in custody. Now widely used, prior research has focused mainly on the examination of the predictive validity of the overall score provided by the instrument. In practice, however, parole/probation officer rely of the LS\CMI items for case planning and case management purposes. Yet, the psychometric properties of these items are not well known. The current study, therefore, aimed to investigate the internal structure of the LS\CMI using item response theory (IRT). To do so, prospective longitudinal data from a Canadian sample of 17,651 males (inmates, probationers) followed over a one-year period were utilized. All items of the LS\CMI were inspected using a series of IRT analyses. The findings show that, aside from static risk factors related to a person's criminal history, the majority of items reflective of a person's criminogenic needs have limited predictive value, thus raising questions about the use of this instrument for case management purposes. The findings also the instrument's varying psychometric properties across samples. In all, the study findings raise important and significant questions about the validity of the instrument. •Item response theory analyses were used to examine one section of the LS\CMI.•The majority of LS\CMI items shows limited discriminant power (recidivists vs. non-recidivists).•Items presenting the most desirable psychometric properties relate to a person's criminal history.•The findings show that the LS\CMI does not perform equally well with inmates and probationers.•The findings raise concerns over the use of the instrument across jurisdictions and settings.