Four-factor structure of the correctional personnel rating scale.

• Author: Mitchell, Tom
• Position: Canada

Within corrections research, efforts to better understand the role of correctional officers (CO) has focused primarily on assessing their attitudes towards their job and offenders. For instance, Tellier, Dowden, Fournier, and Franson (2001) provide an excellent overview of scales to assess COs' orientation to work (Robinson, Porporino, and Simourd 1993), how they view offenders (Jurik 1985), and how they view the quality of their interactions with inmates (Gerstein, Topp, and Correll 1987).

Jurik (1985) found that after organizational factors were taken into consideration, minorities (Black and Hispanic COs) held more positive attitudes toward inmates than did whites. In the same study it was also found that COs who were more intrinsically interested in their job held a more favourable attitude towards inmates than those less career oriented. Similarly, Cullen, Lutze, Link, and Wolfe (1989) found that, in general, COs view their work from a human services and rehabilitation orientation.

However, to better understand COs' interaction with inmates, it is important to go beyond assessing attitudes to determine how they actually behave toward inmates. While early research in social psychology showed little correspondence between attitudes and behaviour (Wicker 1969), later evidence indicates that under certain conditions (e.g., when behaviour is unconstrained), attitudes can predict behaviour (Ajzen and Fishbein 1977). The primary objective of the current study was to refine an existing behavioural observational measure of COs' interaction with inmates that could be used for performance management and development.

Traditional performance measures used to evaluate correctional officer behaviours focus on broad job components. As a result, more specific CO job behaviours are seldom taken into account. Earlier research focused primarily on physical requirements for successful CO performance (Wilkins 1975). While there has been relatively little research to establish valid measures that predict correctional officers' job performance, recent evidence supports the use of personality measures, bio-data, and cognitive ability (Hunter and Hunter 1984; Pulakos and Schmitt 1996). Ross and McKay (1981) have argued that the selection and evaluation of COs should be based on specific behavioural skills that relate to effective CO performance. The Correctional Personnel Rating Scale (CPRS) is such a behaviourally based measure. It describes effective and ineffective behaviours couched in language that raters familiar with the job will understand (Szilagyi and Wallace 1980).

In the early stages of the CPRS scale development, COs identified job-relevant behaviours that were originally classified into six different categories. However, data on item analysis, reliability, and validity were insufficient (Willis, Jessup, Savage, Cooper, and Slesser 1979). Later, Willis, Savage, and Jessup (1979), on the basis of a rational analysis, grouped the 69 items of the CPRS into five dimensions: (1) routine job tasks, (2) leadership, (3) emotional control, (4) staff relations, and (5) resident relations. Although the CPRS is a rationally developed alternative method for assessing desirable and undesirable CO job behaviours (Willis, Savage, and Jessup 1979), the scale has not been examined closely to determine its psychometric properties.

Only two studies have used the CPRS to rate CO job behaviours for research purposes. Willis and colleagues (1979) had two correctional personnel specialists rate each officer, whereas Wahler and Gendreau (1990) had supervisors, co-workers, and inmates rate each officer. However, Wahler and Gendreau did not replicate the original study and the factors identified in the original scale. Rather than the five dimensions identified by Willis et al, in their factor analysis, Wahler and Gendreau found three factors: (1) responsibility/leadership skills, (2) behavioural deficit skills, and (3) inmate relations. However, their factor analysis was based on a relatively small sample size (N = 199) for all three subject groups. Likewise, the sample size was very small in the Willis, Savage, and Jessup (1979) study (N = 20), limiting the generalization of the findings.

In addition to the sample size considerations in the previous studies, there is also confusion about the actual number of factors that comprise the CPRS. This confusion may be due to the possibility that some of the items are flawed or unclear. The psychometric properties of the CPRS sub-scales have not been clearly established. As such, the scales are not likely to be generalizable to other populations of COs in other correctional facilities. Furthermore, it has been suggested that once the psychometric properties of the CPRS are clearly established, it would be useful to develop behaviourally oriented selection criteria for COs (Wahler and Gendreau 1985).

Therefore, the purpose of the present study was to develop content validation for the meaningfulness and clarity of the items and to assess the factor structure of the CPRS as a performance measure by using a larger and more diverse sample of COs than was used in the two previous studies.

Method

Participants

Two supervisor ratings were...