Protective Factor Assessments: What Are We Measuring?—An Investigation of the Internal Validity of the Structured Assessment of Protective Factors for Violence Risk

• Author: Valerie Moulin,Milena Abbiati,Jacques Gasser,Philippe Golay
• DOI: 10.1177/0093854819901157
• Published date: 01 April 2020
• Date: 01 April 2020
• Subject Matter: Articles

CRIMINAL JUSTICE AND BEHAVIOR, 2020, Vol. 47, No. 4, April 2020, 383 –398.

DOI: https://doi.org/10.1177/0093854819901157

Article reuse guidelines: sagepub.com/journals-permissions

© 2020 International Association for Correctional and Forensic Psychology

383

PROTECTIVE FACTOR ASSESSMENTS

What Are We Measuring?—An Investigation of the

Internal Validity of the Structured Assessment of

Protective Factors for Violence Risk

MILENA ABBIATI

PHILIPPE GOLAY

JACQUES GASSER

VALERIE MOULIN

Lausanne University Hospital

University of Lausanne

The use of instruments measuring protective factors is on the rise in risk evaluation, management, and treatment planning.

Although assessment tools must have good internal validity to be generalizable and reproducible, little is known about the

internal validity of protective factor instruments. The present study evaluated the factor structure of the Structured Assessment

of Protective Factors for Violence Risk (SAPROF) via confirmatory and exploratory factor analyses of forensic risk assess-

ments of 143 individuals convicted of violent and sexual violent offenses in Switzerland. The SAPROF did not show any

differences by type of offenses. Although our confirmatory factor analysis did not confirm the classic SAPROF scale’s

structure, the results of the exploratory analysis were in line with previous theoretical research as it revealed a four-factor

model comprising Resilience, Reintegration, Treatability, and Living Conditions. These results have implications for the use

of the SAPROF in forensic risk assessments and treatment management plans.

Keywords: protective factors; expert forensic practice; risk assessment instruments; construct validity

Criminal justice systems are increasingly asking forensic health professionals to carry out

risk assessments in a wide range of criminal cases. The resulting reports may focus on a

variety of issues including a person’s potential for violence, the chances an individual will

reoffend, treatment needs, and the efficacy of interventions aimed at protecting public safety.

In many countries, including Switzerland, Germany, and Canada, these professionals are

AUTHORS’ NOTE: Correspondence concerning this article should be addressed to Milena Abbiati, Research

Unit of Legal Psychiatry and Psychology, Lausanne University Hospital, 1200, Lausanne, Switzerland; e-mail:

Milena.Abbiati@chuv.ch

901157CJBXXX10.1177/0093854819901157<italic>Abbiati et al.</italic> / Protective Factors: What Are We Measuring?Criminal Justice and Behavior

research-article2020

384 CRIMINAL JUSTICE AND BEHAVIOR

also expected to advise on mandatory treatment leading to indefinite post-sentence detention

(Delacrausaz & Gasser, 2012; Hardie et al., 2008).

Hundreds of general and specific psychology risk assessment instruments are now avail-

able for this purpose, and more than half the assessments conducted in any 12-month period

use such a tool (Neal & Grisso, 2014; Singh et al., 2014; Toop et al., 2019). Because psy-

chologists’ and psychiatrists’ recommendations are extremely influential on both the indi-

vidual and societal levels, any structured risk assessment instrument used in formulating

recommendations must have the best possible validity and predictive accuracy (Douglas

et al., 2017; Fazel & Wolf, 2018). Thus, validity, rather than time factors or personal prefer-

ences, must be the main criterion for choosing which instrument(s) to use when carrying out

criminal and clinical violence assessments.

RISK-BASED ASSESSMENTS

Unstructured professional judgments based on clinical evaluations and personal exper-

tise/knowledge (first-generation risk assessments) are now a thing of the past, as their pre-

dictive powers are no better than random judgments (Andrews et al., 2006). Second-generation

risk assessment instruments, which assess mainly static/unchangeable factors (e.g., gender,

history of violence, age at first arrest), and third-generation risk assessment instruments,

which take into account mostly dynamic/changeable factors (e.g., empathy, motivation for

treatment, alcohol addiction), have moderate to good validity for predicting antisocial and/

or violent behavior (Singh & Fazel, 2010; Yang et al., 2010) by adult justice-involved indi-

viduals (Campbell et al., 2009) or by people in psychiatric facilities (Eisenberg et al., 2019;

O’Shea et al., 2013).

More recently, fourth- and fifth-generation risk assessment instruments, which mainly

draw on the risk–need–responsivity (RNR) model (Andrews & Bonta, 2010), incorporate

case planning and risk management aspects, in addition to static and dynamic risk factors

(Andrews et al., 2006; Koetzle Shaffer et al., 2011; Wong & Gordon, 2006). They include

the Level of Service/Case Management Inventory (LS/CMI), the Correctional Assessment

and Intervention System (CAIS), Correctional Offender Management Profiling for

Alternative Sanctions (COMPAS), and the Risk Management System (RMS).

Administering these instruments at different points in time provides clinicians with infor-

mation on variations in dynamic factors, such as the general risk of recidivism (i.e., crimi-

nological needs), and therefore on the impact of treatment. They can then use this information

to adjust their interventions by better targeting responsivity needs. Evaluations of their

effectiveness in predicting violent behavior have given encouraging results, although such

studies have concerned a limited number of countries and do not provide enough informa-

tion on interrater reliability (Desmarais et al., 2016; Desmarais & Singh, 2013). Moreover,

further research is needed to clearly demonstrate the association between changes in

dynamic factors and changes in risk (Viljoen et al., 2018).

INCORPORATING STRENGTHS-BASED ASSESSMENTS

Research carried out in recent decades has highlighted the need for violence reduction/

prevention interventions not only to minimize risk factors but also to strengthen protective

factors. Protective factors are generally defined as strengths that counterbalance or weaken

risk factors and thereby help prevent individuals committing violence (de Vries Robbé et al.