Taking down the straw man: a reply to Webster and Doob.

• Author: Blanchette, Kelley
• Position: Response to Cheryl Marie Webster and Anthony N. Doob, Canadian Journal of Criminology and Criminal Justice, vol. 46, p. 395, July 2004

Introduction

The Custody Rating Scale (CRS) was found by Blanchette, Verbrugge, and Wichmann (2002) to be a valid measure of custodial risk for federally sentenced women offenders. CRS validation includes, but is not limited to, its role in predicting institutional adjustment, escape, and reoffending. Nevertheless, in the July 2004 issue of this journal Webster and Doob (2004) question the validity of the CRS for women, primarily because Blanchette et al.'s (2002) study revealed that one of its sub-scales does not predict institutional misconduct. The fact that scores on the Institutional Adjustment (IA) sub-scale of the CRS correlated positively and significantly with violent (r = 0.39) and non-violent (r = 0.47) incidents while in custody indicates that it is a valid measure of institutional risk. The fact that the Security Risk (SR) subscale did not correlate to institutional misconduct is not surprising, as that sub-scale was designed to measure public safety risk. By setting up this straw man, Webster and Doob (2004) demonstrate that their views are misdirected and reflect a narrow understanding of the application and validation of security classification instruments. The framework for our discussions is to challenge the "empirical investigation" of the CRS presented by Webster and Doob, who provide no new empirical data on the instrument per se and draw some erroneous conclusions. Rather than entering into a theoretical debate about security classification, we have chosen to take down their straw man by presenting the legal and empirical underpinnings of the CRS, offering new validation data on the instrument, and addressing other contentious issues.

Legal and empirical underpinnings

Unfortunately, Webster and Doob are misinformed when they state that the CRS "constitutes the foundation of the security classification system in Canada" (2004: 396). In fact, it is s. 30 of the Corrections and Conditional Release Act (CCRA) that legally prescribes how security classification decisions are made. Furthermore, in accordance with ss. 17 and 18 of the Corrections and Conditional Release Regulations (CCRR), the Correctional Service of Canada (CSC) is mandated to take the following factors into consideration in assigning a security classification of maximum, medium, or minimum: (a) seriousness of the offences committed by the inmate; (b) any outstanding charges; (c) the inmate's performance and behaviour while under sentence; (d) the inmate's social, criminal, and, where available, young offender history; (e) any physical or mental illness or disorder suffered by the inmate; (f) the inmate's potential for violent behaviour; and (g) the inmate's continued involvement in criminal activities. In the field of corrections, the purpose of security classification is to protect the public, employees, and offenders through appropriate placement of the inmate.

As is true of most classification instruments, CRS development began with a review of existing instrumentation being used in other jurisdictions and proceeded to empirically test common factors used to assign custody levels (Canada, Solicitor General 1987). Not surprisingly, certain factors were common to all these scales: severity of current offence, sentence length, violent criminal background, and various indices of behaviour prior to incarceration. More importantly, the background documentation for the CRS includes descriptions of classification models, two-dimensional in nature, that take into consideration both institutional risk and public risk. In these classification models, scores are integrated (in a matrix fashion) to arrive at a final custody designation.

In a similar way, the CRS was designed with two dimensions: an Institutional Adjustment (IA) sub-scale (5 items: history of involvement in institutional incidents, escape history, street stability, alcohol/drug use, age at time of sentencing) and a Security Risk (SR) sub-scale (7 items: number of prior convictions, most severe outstanding charge, severity of current offence, sentence length, street stability, prior parole and/or mandatory supervision/statutory releases, age at time of first federal admission). A statistical weighting scheme is used to score the CRS items, and items within each sub-scale are summed to provide a total sub-scale score. Cut-off values are applied to the total sub-scale scores and yield a security level designation of minimum, medium, or maximum. CRS designations are based on the higher of the two ratings (IA and SR). It is important to highlight that the CRS yields a recommended security designation, based on objective behaviourally based...