Latent constructs in the measurement of risk and protective factors for violent reoffending using the HCR-20^v3 and SAPROF: implications for conceptualizing offender assessment and treatment planning

ABSTRACT

This paper describes a retrospective data analysis, examining the relationship between protective factors against violence, as measured by the Structured Assessment of Protective Factors for Violence Risk (SAPROF), and conceptually similar risk factors assessed in the Historical-Clinical-Risk Management-20 Version 3. The sample comprised 201 adult male violent offenders who completed a violence intervention programme in an Australian prison. Structural equation modelling revealed that conceptually similar risk and protective factors loaded onto correlated, but independent, factors. This suggests that these items from the two assessment tools may assess distinct latent constructs relevant to violence risk. SAPROF items did not, however, load onto the domains that are commonly targeted in violence treatment programmes. This study’s conclusion that assessment of protective factors may offer supplementary conceptual value in clinical practice with violent offenders is tempered by limitations, including the retrospective file review design, small sample, and lack of predictive validity of SAPROF ratings.

KEYWORDS:

• SAPROF
• protective factors
• violence
• risk assessment
• treatment

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

1 As Coupland (2015) highlights, many, if not most, of the items on the SAPROF could be subsumed under the central eight risk factors. These include: History of antisocial behaviour, Antisocial personality traits, Criminal thinking, Criminal associates, Substance abuse, Family/marital problems, Poor work/school performance, and Lack of prosocial leisure activities (Andrews & Bonta, 2006).

2 The HCR-20^V3 factors were: Chronic Antisociality (previous violence, age at first violence incident, psychopathy, early maladjustment, personality disorder, and prior supervision failure); Life Dysfunction (relationship instability, employment problems and substance use problems); Disagreeableness (lack of insight, negative attitudes, impulsivity, unresponsiveness to treatment and non-compliance with remediation attempts); and Destabilising context (plans lack feasibility, exposure to destabilisers, lack of personal support and stress). The fit of the model was modified by freeing some of the error covariances, χ² (16) = 39.664, p < .001; CFI = .983; TLI = .970; RMSEA = .040; SRMR = .049.

3 For example, the HCR-20^V3 item ‘Recent problems with instability’ is intended to capture instability in affect, cognition, and behaviour (Douglas et al., 2013). Arguably, ‘instability’ captures a much broader construct than the SAPROF item ‘Self-control’ and thus, self-control is unlikely to be directly opposite to and mutually exclusive from instability. Similarly, although an individual may possess self-control in some areas (e.g., substance use), they may simultaneously demonstrate instability, or a lack of self-control, in other areas (e.g., affective instability), suggesting that the definitions of these two items lead to the assessment of distinct constructs. The same argument can be made with regard to the SAPROF item ‘Motivation for treatment’. It is likely that this item represents an overarching construct that can be divided into various individual factors including motivation to comply with psychological treatment, in addition to responsiveness to treatment, and motivation to adhere with prescribed medications. Accordingly, the SAPROF item ‘Motivation for treatment’ is unlikely to capture exactly the same information as the HCR-20^V3 item ‘Future problems with treatment/supervision response’.