Youth Risk/Need Assessment: An Overview of Issues and Practices

2. Historical Development of Risk/Need Assessments

The current emphasis on risk/need assessments evolved out of an influential professional ideology-and now deeply embedded practitioner-driven research agenda that embraced the rehabilitative ideal and its post-Martinson revival-and out of a long-standing organizational commitment to intervention and programs. This research rejects the popularized nothing works claim, seeks to determine what works, and strategically deploys effective, targeted correctional interventions. This approach endorses the use of science to resolve crime-related problems (Cullen and Gendreau 2001). Notwithstanding the scepticism within criminology and other social sciences about our ability to make accurate and reliable predictions of dangerousness and recidivism, Canadian correctional researchers maintain that there is a consistent relationship between the type and number of needs offenders present and the likelihood of recidivism. Further, Motiuk argues that the combined assessment of both risk and need will improve our ability to predict who is likely to reoffend and who is not (Motiuk 1993). These risks and needs are identified through research that estimates their predicted likelihood of future criminal involvement. Researchers' attempts to identify and quantify risk/need have resulted in the reconceptualization of needs and certain social structural barriers as risk factors.

The analysis of risk factors is linked to the identification of criminogenic needs factors that have a role in preventing, rather than simply predicting, offending. Actuarial tools are now being used to classify prisonersin terms, not just of their security risks, but also in terms of their criminogenicneeds. Risk assessment tools play a central role in matching "levels of treatment service to the risk level of the offender" (Andrews, Bonta and Hoge 1990). Much of this work is undertaken in Canada by psychologists working in the correctional field (Andrews et al.1990; Andrews, Zinger, Hoge, Bonta, Gendreau and Cullen 1990). It is closely tied to their view that only specific types of treatment, targeted to particular groups of offenders, can reduce reoffending. Risk/need classification, therefore, results in a security classification, as well as an allocation of level of treatment or supervision. More recently, the terminology has changed with criminogenic risksbeing referred to as static (i.e., unchangeable) factors (e.g., age or offence history), and criminogenic need as dynamic factors, which can be modified by treatment programs.

Extensive literature on the benefits of assessing both risk and need exists. Such assessment practices are believed to enhance the accuracy of clinical decisions, and to allow for targeted interventions, better classification, program evaluation, standardization, and efficient resource allocation (Andrews and Bonta 1998; Loza and Simourd 1994). The use of risk and need evolves out of a critique of static actuarial models of risk prediction in correctional literature on assessment and classification, where, in the early 1980s, psy-professionals raised a host of concerns about the use of static risk models.[4] This critique generated new ways of understanding risk and knowing the offender, and reasserted the premise that offenders can change if knowledge of their needs is integrated into assessment technologies.

Assessment tools and more general classification practices that combine risk and need are euphemistically referred to as third-generation risk assessments.[5] These third-generation tools are believed by many to be better clinical assessment tools and predictive devices than earlier first- and second-generation risk assessments. The first-generation risk assessment relied primarily on the unstructured clinical judgment of skilled practitioners. This tool was discredited because of its subjective, unempirical qualities and for its poor predictive accuracy.Second-generation risk assessments, which have garnered the most recent attention, are those that stress static historic factors, such as age, number and type of convictions, sexual offending, and relationship to victim. Examples of these tools include the Salient Factor Score (used in the United States), the Statistical Inventory on Recidivism (SIR) (used in Canada), and the Risk of Reconviction (used in the United Kingdom).[6] These tools were seen as more objective, empirically sound actuarially, and as having considerably better predictive accuracy than previous methods (Andrews and Bonta 1998). As noted above, these second-generation assessments have been extensively criticized for their rigidity and prohibitive reliance on static offence-based risk criteria.

The rigid knowledge of risk contained in second-generation risk tools produced a fixed risk subject (Hannah-Moffat 2002), who was designated to a particular risk category (high, medium, or low), based on accumulated historical factors that, for the most part, could not be changed. This conceptualization of the offending subject naturally limited practitioners and prescribed little by way of intervention, other than incapacitative measures; thus providing little guidance to correctional administrators and limiting the scope of correctional intervention. More abstractly, such understanding of risk was predicated on the implied failure of rehabilitative interventions and the tacit understanding of incapacitation as a preferable penal strategy. This logic contributed to what penal scholars have dubbed the post-welfare era of hyper or mass incarceration.[7]

At the same time that static risk logic was being mobilized to legitimate and inform penal policies, practitioners and correctional researchers were engaged in forms of knowledge production that challenged this seemingly dominant understanding of risk, and reasserted the importance of rehabilitative programming. For instance, Don Andrews (1989), a leading proponent of the what works movement and author of dominant assessment tools, indicates, "past (second-generation) assessments of risk fail to prescribe interventions, and ignore the fact that, once in the correctional system, offenders are subject to events and experiences that may produce shifts in their chances of recidivism" (p. 5). That is, lower-risk cases may remain low risk throughout their period of supervision, or they may move into higher-risk categories. On the other hand, higher-risk cases may remain high risk or they may move in the direction of lower risk. Andrews argues, "improving the accuracy of prediction risk assessments is contingent upon a determination of the characteristics of offenders and their circumstances that are subject to change during the sentence, and establishing which of those changes actually indicate an increased or a reduced chance of recidivism" (pp. 5-6). This knowledge, Andrews contends, requires researchers and practitioners to look beyond risk factors that cannot be changed, such as criminal history, to changeable dynamic factors, or criminogenic need factors.

Using the insights of meta-analysis, correctional researchers argued that the absence of dynamic variables or needs, such as employment, marital/family relationships, associates, antisocial attitudes, personality traits, substance abuse, and other theoretically[8] relevant items that were statistically shown to be correlated with criminal conduct, were a limitation of earlier tools (Andrews and Bonta 1998). This powerful critique of the first- and second-generation risk assessments led to the assimilation of needs into traditional risk assessments that in turn, increased practitioners' confidence in their ability to predict recidivism and design targeted interventions. Guided by the notion that "prediction should provide utility" (Andrews and Bonta 1998:225), a third generation of risk assessment evolved. Thethird-generation risk assessment is distinctive because it purports to objectively and systematically measure static and dynamic risk or criminogenic needs factors. A fourth generation of risk assessment is envisioned that will include the identification and measurement of key responsivity characteristics for treatment matching (Andrews and Bonta 1998). The most recent version of the Youth Level of Service/Case Management Inventory (YLS/CMI) and planned[9]revisions ofLevel of Service Inventory - Revised (LSI-R) integrate responsivity factors into the assessment of risk and need.

Clearly, the new convention in risk assessment and classification is to use strategies and tools that "systematically bring together information about an offender's history and needs to develop a treatment plan and assign levels of supervision" (Bonta 2002:1). Andrews and Bonta's (1998) principles of risk, need, responsivity, and professional discretion illustrate this new risk-informed managerial logic of penal governance. The quadrangle of risk, need, responsivity, and professional discretion are identified as four principles of classification/assessment.

The risk principle is an endorsement of the debated premise that criminal behaviour is predictable and that treatment services need to be matched to an offender's level of risk. Thus, offenders who present a high risk are those who are targeted for the greatest number of interventions. The needs principlepertains to the importance of targeting criminogenic needs and providing treatment to reduce recidivism. Through such tools, need is explicitly linked to rehabilitation; criminogenic needs/dynamic risk factors are rehabilitative targets. However, treatment often means cognitive behavioural interventions that claim to teach and not treat,as previous rehabilitative connotations suggest. The responsivity principle, which refers to the "[delivery] of treatment programs in a style and mode that is consistent with the ability and learning style of the offender" (Andrews and Bonta 1998:245) expands this premise. Offenders are human beings and the most powerful influence strategies available are behavioural/social/learning cognitive behavioural strategies. Finally, the principle of professional discretion strategically reasserts the importance of retaining professional judgment, provided that it is not used irresponsibly and is systematically monitored. Here, the term of professional includes a host of practitioners (or para-professionals) with little to no professional training in risk assessment and, in the most extreme cases, correctional officers or parole supervisors.

Given the rapid growth of risk/need assessment tools and their increased use at various stages of the criminal justice process, few international researchers have critically assessed the impact of this trend and or collected data on how these tools impact decision-making. Existing critical international literatures use varying theoretical approaches to raise concerns about a wide range of risk-based practices deployed in criminal justice and mental health systems. Some of these concerns include: due process, justice and proportionality (Hudson forthcoming; 2001; Rose 1998), moral and political dimensions of risk (Ericson and Doyle 2003; Gray, Laing and Noaks 2002; Stenson and Sullivan 2001), gender, racial and culture discrimination (Hannah-Moffat and Shaw 2001), the targeting of marginalized populations and the redistribution of resources based on risk profiles (O'Malley 1999; Rose 1998; Silver and Miller 2002), and the tenuous relationship between risk and rehabilitation or what works initiatives (Brown 1996; Hannah-Moffat, 2002; Kemshall 1998; Robinson 2002; 1999; 1996). This literature offers an interdisciplinary approach to the study of risk and risk-based governance. By and large these article raise theoretical questions about the increased used of risk based technologies and the potential impact of this trend. This literature examines specific operational questions about how, or even if, risk determinations ought to guide decision-makers, but does not examine whether these tools are valid and reliable predictors.


  • [4] See Rose (2002) for a discussion of psy-professionals and para-professionals.
  • [5] Bonta (1996) offers a comprehensive review of the correctional classification literature in which the evolution of risk assessment is categorized into three generations.
  • [6] Andrews and Bonta (1998) provide a detailed description of these tools and their development.
  • [7] For more detailed discussion of this era, see special issue of Punishment and SocietyVol 3:1.
  • [8] The underlying paradigm is social learning theory.
  • [9] Interview transcript - D. Andrews, 2003.
Date modified: