CLINICAL PSYCHOLOGY REVIEW | Volume 106 | https://doi.org/10.1016/j.cpr.2023.102355
Michael C. Seto, Carissa Augustyn, Kailey M. Roche, Gabriella Hilkes
ABSTRACT
This review provides an overview and update of empirical evidence for psychologically meaningful dynamic risk factors and protective factors for sexual offending. Using the review by Mann et al. (2010) as a starting point, we reviewed relevant literature that has appeared since this publication, focusing on meta-analyses, systematic and scoping reviews of dynamic risk factors, recent evaluations of commonly used dynamic assessment tools, and studies of dynamic risk and protective factors in community samples in addition to clinical or forensic samples. Two risk factors previously deemed promising by Mann et al. (2010), hostility towards women and dysfunctional coping (conceptualized as hostile masculinity and emotional regulation deficits, respectively, in this review), could now be considered supported using this review's criteria of three or more studies demonstrating an effect size of 0.15 or greater. No new risk factors were identified. We conducted a broader search of protective factors in recognition of the relative newness of this literature: Positive social support was the only protective factor identified as empirically supported. We also discuss situational risk and protective factors.
INTRODUCTION
Identifying empirically-supported risk and protective factors is the cornerstone of effective practices with people who have committed sexual offenses. Clinical decision-making relies on our understanding of characteristics and circumstances that increase someone's risk for criminal behavior (risk factors), as well as factors that may reduce or buffer against this risk (protective factors). Over the last several decades, there has been a growing literature on risk and protective factors, including the emergence of numerous assessment tools designed to aid the assessment process (Kelley, Ambroziak, Thornton, & Barahal, 2020; Tully, Chou, & Browne, 2013).
While this influx of information has been beneficial, it raises challenges for synthesizing and translating what we have learned in research into coherent and digestible information for practitioners. For example, while it is possible to curate a list of factors from the available literature, it is more difficult to discern which factors have better empirical support, or which factors should be prioritized in interventions. Like Mann, Hanson, and Thornton (2010), discussed in more detail later, we also think that understanding empirically supported and psychologically meaningful risk and protective factors can be helpful with case formulation and intervention planning.
As we learned in this review, there is also considerable variability in the literature regarding the terminology used for – and conceptualizations of – risk and protective factors for sexual offending. Similar factors are described using different terms and often assessed using different tools, and conversely, different factors are described using the same terms. This can serve as an additional barrier to understanding and consolidating available evidence.
The importance of understanding which risk factors are empirically supported is emphasized by the Risk-Need-Responsivity (RNR) framework for effective rehabilitation (Bonta & Andrews, 2016). Hanson, Bourgon, Helmus, and Hodgson (2009) demonstrated that sexual offending programs that adhere to RNR principles produce better outcomes than those that do not. The Risk Principle finds that interventions have more potential impact if they are titrated to the long-term risk of recidivism. Long-term risk of recidivism is best assessed by focusing on personal, static risk factors such as the person's age or criminal history. These risk factors are static because they do not change (criminal history) or cannot be changed by intervention (age). The literature on static risk assessment is expansive and has resulted in multiple validated risk assessment tools (e.g., Tully et al., 2013). The Need Principle finds that interventions have more impact if they target criminogenic needs (also referred to as dynamic risk factors) associated with the likelihood of recidivism. A key distinction between the Risk and Need Principles is that dynamic risk factors are potentially changeable, unlike static risk factors that help explain long-term risk of sexual recidivism. Dynamic risk factors can therefore be important targets for treatment and supervision professionals. Well-known examples of dynamic risk factors identified by Mann et al. (2010) include atypical sexual interests, excessive sexual preoccupation, offense-supportive attitudes and beliefs, and association with antisocial peers. Definitions and examples of these factors are discussed later in this review. Finally, the Responsivity Principle finds that interventions have more impact if they are tailored to the individual. This includes both general and specific responsivity. An example of general responsivity is that most people prefer and benefit more from interventions that are practical, problem-focused, and delivered in a warm and empathic milieu. An example of specific responsivity is a program that relies less on reading and uses visual and other aids for someone with lower literacy levels.
There have been several cornerstone papers that have been integral to our understanding of risk factors for sexual offending, including reviews by Hanson and colleagues (Hanson & Bussière, 1998; Hanson and Morton-Bourgon, 2004, Hanson and Morton-Bourgon, 2005). Each of these reviews utilized systematic review techniques to integrate information from multiple studies and provide a more coherent understanding of the strength of evidence for various static factors and factors that are potentially dynamic but measured at only one point in time. This cumulative knowledge about static risk factors is essential for identifying individuals who are higher in long-term risk to sexually reoffend, and thus require more intensive intervention. However, identifying static risk factors does not directly address what should be targeted in intervention.
The conceptualization and review of evidence offered by Mann et al. (2010) was an important contribution to the field's understanding of changeable, dynamic risk factors. These authors reviewed the clinical and forensic literatures and identified psychologically meaningful risk factors. They defined factors as psychologically meaningful if they appeared in theoretical or clinical models of sexual offending and there is good evidence they predict sexual recidivism. They defined risk factors as individual propensities that may change over time and can result in sexual offending in particular environments. Mann et al. (2010) identified a number of psychologically meaningful risk factors, further distinguishing between factors that were supported versus promising (see Table 1). Factors were considered supported if there were at least three studies that, when meta-analytically integrated, were predictive of sexual recidivism at an effect size of 0.15 or greater. Factors were considered promising if there was at least one study showing a statistically significant predictive effect along with other supportive evidence, such as a significant correlation between the factor and a validated risk measure. The authors also distinguished between factors that were not supported but had interesting exceptions (e.g., the overall effect was non-significant but a significant relationship was found in one credible study), factors that were worth exploring, and factors that were unrelated to sexual offending (only supported and promising factors are included in Table 1).
Neither Mann et al. (2010) nor the RNR framework directly considered protective factors, though there is emerging interest in the idea that protective factors can also be important targets for interventions. While there is some debate in the literature as to how to define and conceptualize protective factors (e.g., Cording & Beggs Christofferson, 2017), there is growing clinical interest in assessing for strengths that provide a protective buffer against the effects of risk factors and therefore reduce the likelihood of recidivism. Some of the potential benefits of protective factors that have been identified include shifting away from a deficit focus, improving the accuracy of assessments, and increasing the clinical rapport between clients and their assessors (Cording & Beggs Christofferson, 2017; de Vries Robbé, Mann, Maruna, & Thornton, 2015). However, these ideas remain largely untested as, unlike risk factors, there has been limited empirical research on protective factors in general, but especially in regards to sexual offending and adult populations.
Some protective factor candidates have been identified in recent sexual offending research, including research on a new tool that includes protective factors (SAPROF-SO, Willis, Thornton, Kelley, & de Vries Robbé, 2017–2020, Willis, Thornton, Kelley, & de Vries Robbé, 2021). Other examples of protective factors in lifespan research on high-risk delinquent boys were identified by Laub and Sampson (2009), including a stable marital relationship, military service, and stable employment. A notable contribution to our understanding of protective factors of sexual offending was offered byde Vries Robbé, Mann, et al. (2015). The authors reviewed the available theory and empirical evidence and developed a list of eight plausible protective domains for sexual offending: (a) healthy sexual interests; (b) capacity for emotional intimacy; (c) constructive social and professional support network; (d) goal-directed living; (e) good problem solving; (f) engaged in employment or constructive leisure activities; (g) sobriety; and (f) hopeful, optimistic and motivated attitude to desistance (i.e., not offending). It is important to note that much of the evidence for these factors stemmed from the authors' conceptualization of these factors as “healthy poles” to established risk factors (e.g., healthy sexual interests such as sexual interest in consenting adults was regarded as a healthy pole to the risk factor of paraphilic sexual interests such as pedophilia). As we will discuss, this differs from our conceptualization of protective factors and the criteria for evidence used in this review.
