< Previous20 PERSPECTIVESVOLUME 46, NUMBER 2 RISK NEED ASSESSMENT Contemporary tools are often referred to as “Risk- Need Assessments” (RNAs). This naming mechanism reflects tools with combined features, measuring both risk factors as well as criminogenic needs (Hamilton et al., 2017). Risk factors are personal characteristics and circumstances that are present in an individual and are predictive of future criminality (Andrews et al., 1990). Risk factors can be static (unchanging) or dynamic (changeable). However, dynamic factors may also represent criminogenic needs and become the target of case management and programming referral (Andrews et al., 1990; Hamilton et al., 2017). For example, a common static factor is age at assessment, as, of course, an individual’s age will increase with time but cannot change via intervention, while a common dynamic factor, current drug use, can change over time and may be influenced via programming. Notably, contemporary RNAs are composed of two tool components–a risk assessment and a needs assessment, where risk assessments use any item (static or dynamic) that improves the prediction of recidivism. However, needs assessments include only a subset of dynamic items that can be altered through intervention. Risk assessment scoring is commonly a summary of all items, regardless of content or type. Threshold scores are then set for the assessment to determine an individual’s risk category (i.e., high, medium, or low risk). These levels are either determined by the developer when the tool is created or can be modified to fit the agency/population needs. These risk levels are then used by agencies to prioritize resources, including provision of transitional living services, cognitive-behavioral programming, or even level of community supervision (Montford & Hannah-Moffat, 2021). This prioritization is not only beneficial to agencies with limited resources, restricting supervision and programing to higher-risk individuals, but it also prevents lower risk individuals from being “over programmed.” This element of modern assessment has been used to help identify participant eligibility for such interventions as diversion, administrative community supervision, kiosk reporting, and even early release from incarceration for those deemed lower risk (Hamilton, Duwe, et al., 2021; Taxman et al., 2013). Risk Versus Need Assessment Risk-needs assessments have evolved in many important ways. There have been nearly 40 years of historical assessment development, evolving from a “gut feeling” to the addition of items relevant to case management. The first few risk assessment tools only contained static, demographic, and historical items to predict risk. From that beginning of using tools that assess individuals on fewer than a dozen criminal history and demographic indicators, we have advanced to using tools that include additional indicators to assess both static and dynamic risks and potential responsivity barriers to be considered when assigning programming (Andrews, Bonta, & Wormith, 2006). However, unlike the risk portion of a tool, a “needs assessment” is more strategic in item content, including only changeable, dynamic measures that are considered “criminogenic” (or correlated with recidivism) and changeable in nature (Hamilton et al., 2017). Unlike the risk score, needs are divided by domains to produce sub-scores, where item responses collectively indicate higher/lower levels of concern for a given programming target. Combining both immediate and long-term needs, modern assessments allow case managers to understand the totality of an individual’s needs and provide interventions in a logical sequence. For instance, for those beginning community supervision, needs such as employment and housing must be resolved more quickly and represent “stabilizers” needed prior to addressing longer term programming needs (Taxman et al., 2013). However, for incarcerated individuals, an alternative sequence of programming, such as one focusing on criminal thinking and substance abuse prior to addressing reentry needs, may prove beneficial and more feasible. By pairing risk and needs scoring, case managers can also identify cohorts of individuals needing different treatment intensities. For example, the Correctional Service of Canada groups inmates into high- and moderate-risk cohorts, where those in the high-risk cohort receive a greater intensity (number and duration of sessions) of cognitive-behavioral programming as compared to moderate-risk cohort participants. Further, routine reassessments allow dynamic scales to provide more immediate feedback of reduced needs, providing an assessment of programming progress and changes over time. Criminogenic Risk Factors As the generations of assessment evolved, the addition of dynamic items was perceived as a positive contribution to the field (Taxman et al., 2013). Yet, development of the “needs assessment” has been an afterthought 21 AMERICAN PROBATION AND PAROLE ASSOCIATION RISK NEED ASSESSMENT for many developers and researchers. Beginning with Andrews and Bonta’s development of the Central Eight (Education/ Employment, Family/Marital, Leisure/ Recreation, Companions, Alcohol/Drug Problem, Pro- criminal Attitude/Orientation, and Antisocial Pattern), these domains were developed as clusters of similar items representing a consensus of research findings at the time, and that research is now over three decades old (Andrews & Bonta, 2010). While utilized widely—and even being adapted for assessments of juvenile populations—when implemented many of these domain scales have weak empirical support or represent ineffective predictors of recidivism (Caudy et al., 2013). As additional generalized assessment tools were developed (COMPAS, ORAS/OYAS, STRONG-R, PACT/YASI, SPIn, and WRNA), an array of items and alternate domains have been created, expanding the selection available for correctional agencies to adopt. With the field’s renewed focus on providing rehabilitative services and the reduced use of incarceration/detention, there is an expanding demand to evaluate, update, and recalibrate needs assessment elements of current RNAs, improving their case management utility (Sullivan & Childs, 2021). Protective and Responsivity Factors Related to the development of needs, research on resiliency has expanded the conversation regarding the types of items to be included in an RNA. The needs assessments in many contemporary tools focus on what individuals lack (i.e., housing, employment, or sobriety). In contrast, a recent focus on strengths-based assessment seeks to develop instruments that measure protective factors, specifically adding assessment items that measure strengths or acquired abilities that diminish the effects of a criminogenic risk (Rennie & Dolan, 2010). Examples are having completed a criminal thinking program or having measurable skill improvements in communication, problem-solving, or emotional regulation, all of which may make an individual less likely to re- offend. Assessing protective factors can also reduce the likelihood of inflating someone’s risk score, or over- classification, further individualizing a client’s treatment dosage (Cording & Beggs Christofferson, 2017). However, while exciting and innovative, this area of assessment is still relatively new, and we can expect conversations regarding it and the development of such scales to grow in the coming years. It is also worth noting that responsivity factors,1 representing the last R in the RNR model, have largely been ignored when discussing modern assessments. Briefly, responsivity factors are not crime-producing indicators, but they do represent potential barriers to programming. Some of these factors include education, language proficiency, mental health symptoms (i.e., anxiety, paranoia, dementia), trauma history, medical needs, and intoxication issues (Jung & Dowker, 2016; McDougall et al., 2014). Although these are not considered risks or needs, responsivity factors can substantially affect a person’s ability to participate in treatment. To make programming accessible to those who need it, it is important to address responsivity factors before treatment begins. For example, those who speak English as their second language may not be comfortable conversing about their needs in English, reading programming materials, or writing in a workbook, but these are common requirements of cognitive- behavioral approaches. RNA developers have also begun including responsivity items as unscored items. These identifiers assist case management teams in tailoring program cohorts to best fit the needs of the individual to the program group. For example, identifying individuals who are more likely to be disruptive or lack motivation for participation can help direct treatment selection in terms of utilizing a particular clinician and/or placing the individual with participants with similar characteristics. Gender also plays a role, as specified programs (such as Moving On & Beyond Trauma) have demonstrated a greater effectiveness in female correctional populations (Montford & Hannah-Moffat, 2021). Contemporary Concepts in Risk and Needs Assessment As assessments have become commonplace in correctional settings, a variety of advancements have been developed to improve instruments’ accuracy and specificity. Specifically, efforts to localize assessments as well as to consider individual characteristics such as gender, racial/ethnic identity, and age have been added to the design of assessment tools in order to describe risks and needs more accurately. In this section, these topics are outlined, and their contribution to the future of RNA is considered. 22 PERSPECTIVESVOLUME 46, NUMBER 2 RISK NEED ASSESSMENT Localization Most contemporary assessments used today were developed to be applied “off the shelf,” meaning that a given tool may have been developed for a probation population in Canada, a parole population in New York, or a detention population in Florida, but it is nonetheless assumed to function similarly everywhere. To assess how well a tool predicts for their population, agencies may complete a local validation of an off-the-shelf tool. Typically, local validations demonstrate shrinkage, where the off-the-shelf version of the tool does not perform at the same level as advertised. Recent research has indicated the importance of incorporating local variations (Duwe, 2014; Hamilton et al., 2017) and the need to update or improve an assessment’s items, responses, and scoring to account for distinctions within an agency’s population through a process called localization (Hamilton, Kigerl, et al., 2021). This often means altering a weighted tool, scoring all items “0” for the absence of an attribute and “1” when it is present (e.g., history of drug use), and recalibrating presence to a score of 2, 5, or 10, depending on the item’s importance for the local population. Unfortunately, assessments are commonly adopted by agencies, and used for many years, without using their local population’s data to adjust tool content. This results in less accurate classifications of risk and, as described, may produce a mis-assessment of needs and incorrect program placements (Caudy et al., 2013). Fortunately, when an assessment is localized, results become substantially more accurate and have the potential to improve prediction beyond the tool’s original findings (Hamilton, Kigerl, & Kowalski, 2021). More specifically, localization uses an agency’s assessment responses and recidivism data, entering their population’s scores into a statistical model. Based on the local population’s patterns of responses, predictive items are selected, and those with greater importance are given a larger weight. In addition, those found to be uncorrelated with recidivism are removed, potentially reducing the size of the tool and time needed for completion. Although certain items are universally predictive (e.g., age at assessment), others, such as number of prior incarcerations, might be irrelevant for one agency (diversion) while vitally important for another (parole). Further, populations may change, and following years of use of a tool some periodic updates can help combat shrinkage of accuracy over time. While many agencies have sought to adjust risk category thresholds to provide a better fit for their population, only recently have researchers and practitioners began a process of validating and updating assessments, creating localized versions (Duwe, 2014; Hamilton et al., 2016; Hamilton et al., 2019). We further advocate localization as a developing best practice, adopting new versions of tools over time, as this will enable agencies to provide an improved link between provided programs and individual need. Gender Responsivity Gender responsive assessments are a key component to understanding an individual’s specific pathway to recidivism (Hamilton, Kigerl, & Kowalski, 2021; Hamilton et al., 2019). Historically, the assessment creation process has ignored, or tried to remove, the influence of gender on recidivism prediction (Rettinger & Andrews, 2010). A primary critique of many contemporary assessments is that correctional populations are comprised of mostly males, which increases the likelihood that items specifically predictive for females are removed or scored to have less importance during assessment development (Hamilton et al., 2017). The result has been the over-classification of women by many classification tools and RNAs (Montford & Hannah-Moffat, 2021). This highlights the importance of localization and using assessments that are validated with the population of intended use. While some tool developers will claim their tools are “gender neutral,” a growing body of research has identified several variations in the types of items and scoring of risk and need factors applicable for women. Current literature highlights female-specific needs as histories of trauma, relationship issues, mental illness, drug use, self-efficacy, poverty, and parental issues (Van Voorhis et al., 2010). These differential needs suggest that gender affects not only risks but also criminogenic needs and the types of programs that need to be developed for female populations to reduce their specific risk2. Race/Ethnicity Considerations Following a ProPublica article depicting potential sources of bias in assessment classification (Angwin et al., 2016), researchers have sought ways of examining race/ethnicity disproportionality in assessment tools that may lead to biased risk assessment scoring and classifications. Recent studies have identified a 23 AMERICAN PROBATION AND PAROLE ASSOCIATION RISK NEED ASSESSMENT disproportionate pattern in which non-white individuals are commonly scored and classified as higher risk than white individuals, suggesting bias (Angwin et al., 2016; Campbell et al., 2018; Miller et al., 2021; Onifade et al., 2009). However, when examining the accuracy of tools across race/ethnicity sub-groups, bias is not identified. While new findings are quickly emerging, there is an early consensus that using a standardized risk assessment tool reduces the likelihood of bias as compared to returning to first generation assessments or “gut feelings.” Nonetheless, even though assessment tools have the potential to reduce disproportionality, criminal history items used as both predictors and outcomes of assessment tools contain inherent sources of bias. Specifically, people of color more frequently reside in areas of disadvantage with increased police presence, enforcement, and rates of conviction. The disproportionality ingrained into assessment measures has led many to suggest the need to explore new metrics and update current models to remove sources of bias (Hamilton et al., 2019). Two recent studies have suggested methods for removing criminal history items from tools, with one advocating for increasing the number of needs assessment indicators as a method to reduce classification disproportionality and, in turn, classification bias (Butler et al., 2021; Miller et al., 2021). Similar to the localization methods described previously, Butler and colleagues (2021) utilized a sample of youth assessed using the same RNA. By removing criminal history and other items found to be correlated with race/ ethnicity—the items with the greatest potential for bias— they developed a new version of the original RNA. As a result, the remaining items, mostly dynamic needs, were allowed to supplement, or carry more statistical weight, in predicting recidivism, creating a tool that was equally predictive when compared to the original tool. In this way, reducing the use of disproportionate criminal history measures via increased use of needs items has the potential to decrease, or remove, bias while more accurately identifying the programmatic needs of the population assessed. Specifications for Youth Assessments Unfortunately, research on youth assessments commonly lags behind that done on adult tools. Some recent findings have demonstrated, however, that contemporary tools are in substantial need of modification and advancement (Hamilton, Kigerl, & Kowalski, 2021; Mei et al., 2021). Regarding risk, Hamilton and colleagues (2021) identified that youth tools implemented off the shelf could have their predictive accuracy improved to the tune of 10% with modern updates such as localization, adding elements of gender responsivity, and predicting more specified outcomes. They noted that many agencies (i.e., Washington, Delaware, Florida, Iowa, Nebraska, and Maryland) are currently in the process of creating state-specific versions of some tools. All in all, youth risk assessments are like their adult counterparts in being overdue for revalidation. Mei and colleagues (2021) described a sheer lack of research on youth needs, noting that many assessment developers failed to consider the specific needs of youth, creating only slightly modified versions of their adult tools. Moreover, youth need scales have not been sufficiently examined, and several contemporary tools have tried, and failed, to confirm that assessment domains are measuring youth needs or validly predicting recidivism (Andrews & Bonta, 2010; Farabee & Zhang, 2007; Mei et al., 2021; Schmidt et al., 2005). Regarding youth needs assessments, special care should be taken to link scored domains to the evidence-based programs available to the agency. While Lipsey’s Standard Program Evaluation Protocol (SPEP) has given agencies a baseline regarding how to evaluate the effectiveness of youth programming, the assessment tools used to make programming recommendations go largely unevaluated. For example, domains that are generally tied together for adults (i.e., education and employment) should be redesigned for youth and probably separated. Further, many needs domains require re-framing to suit integrated programs designed to address multiple needs simultaneously (i.e., Functional Family Therapy, or Multi- Systemic Therapy). Thus, when updating youth tools, it is necessary to understand how their programming needs may be distinct and to consider adjusting needs domain scales to inform the evidence-based recommendations commonly/currently available. An RNR Update Example While many agencies may view their assessment as stable, prescribed, and even pitched as a “valid” or even “the most validated” tool, these instruments are built on decades-old research and are in need of reevaluation. Using a somewhat extreme example, just as the medical field no longer recommends leeches as a common or effective treatment to relieve infections, we 24 PERSPECTIVESVOLUME 46, NUMBER 1 should continually update and not use ineffective tools now that our knowledge of the justice system and the populations it serves has expanded. More specifically, the justice system is constantly evolving, with governments enacting new statutes and agencies implementing new policies, with resulting fluctuations in both crime rates and sentence lengths. Still, many contemporary tools are stagnant and slow to adapt. Some development efforts have sought to break new ground, advancing both the science and application of assessment tools. As an example, the Washington State Juvenile Court Administrators–Risk Assessment (WAJCA-RA) was developed in 1997. Made non- proprietary in 1998, yet relatively unchanged since its creation, the instrument has been adopted as both name brand tools (i.e., the PACT, YASI, Back on Track) as well as state-specific versions. Gathering 10 states worth of assessment and recidivism data, Hamilton and colleagues (2021) developed local versions of the tool, selecting and weighting assessment content to improve prediction and develop gender-responsive risk models. To update the needs assessment, advanced statistical methods were employed to redesign tool domains. Efforts were made to both create and empirically validate needs domains, identifying their correlation with recidivism and ensuring that reductions in needs is translating to reductions in risk (Mei et al., 2021). Six domains have specifically been related to youth recidivism (Education, Associations, Family, Alcohol & Drug, Mental Health, and Cognitions & Behaviors), with a global needs score to be used by case managers when considering programming recommendations (see Figure 1). While the six needs domain scores operate similar to other tools, the global score offers the ability to classify individuals as high, moderate, or low need by providing a single summary score combining scores across all domains. This novel assessment of overall need coincides with the RNR model, providing case managers a more expedient and simple method of identifying individuals that are both high-risk and high-need and hence the intended target of programming (Bonta & Andrews, 2016). The culmination of this work outlined the development of the Modified Positive Achievement Change Tool (M-PACT), which is currently implemented, or in the final development stages of implementation, in several state juvenile agencies. While updates were specific to this one tool, authors sought to outline a process for which tools can be similarly updated as an agency best practice. Conclusion Modern risk and needs assessments have become vital to effective rehabilitative practices. In this article, we have discussed how assessment development has adapted the scoring, domains, items, and responses to align with the RNR model. The field continues to advance, attempting to adapt to contemporary issues, such as Global Factor EducationAssociationsFamily Alcohol & Drug Mental Health Cognitions & Behaviors Figure 1. M-PACT Needs Assessment25 AMERICAN PROBATION AND PAROLE ASSOCIATION RISK NEED ASSESSMENT accuracy, responsivity, and bias that can be improved through research and updated versions of assessments designs. As a recent example, we detailed the application of the MPACT, adapting an off-the-shelf tool to the local populations and specified needs of youth. While RNAs provide a powerful tool that predicts risk of recidivism, we direct readers to consider how their tool is designed to function and how it, or any other tool, may be used to best inform supervision and case management needs of the population needing to be served. As discussed, there are many considerations for stakeholders when adopting and implementing new tools. Specifically, one should ensure items and responses have been properly adapted to the populations to be assessed, with an understanding that off-the-shelf versions may differ with regards to gender, race/ethnicity, and age. Second, after implementation, revalidation plans should be developed to assess the performance of the tool with local data. Where findings suggest the need for modification, agencies should explore updated versions, requirin`g assessment developers to adapt to the agency’s needs. Furthermore, updated tools should be able to both outline the criminogenic needs of the population and make necessary case management connections, guiding programming recommendations that are both available and provided by the agency. It is through these customization efforts that further reductions to the correctional population can occur, reducing recidivism and focusing limited resources on those with the greatest needs. References Andrews, D. A., & Bonta, J. (2010). Rehabilitating criminal justice policy and practice. Psychology, Public Policy, and Law, 16(1), Andrews, D. A., Bonta, J., & Hoge, R. D. (1990). Classification for effective rehabilitation: Rediscovering psychology. Criminal Justice And Behavior, 17(1), 19-52. Andrews, D. A., Bonta, J., & Wormith, J. S. (2006). The recent past and near future of risk and/or need assessment. Crime & Delinquency, 52(1), 7-27. doi: 10.1177/0011128705281756 Angwin, J., Larson, J., Mattu, S., & Kirchner, L. (2016). Machine bias. ProPublica, May, 23(2016), 139-159. Bonta, J., & Andrews, D. A. (2016). The psychology of criminal conduct. Routledge. Campbell, C., Papp, J., Barnes, A., Onifade, E., & Anderson, V. (2018). Risk Assessment and Juvenile Justice. Criminology & Public Policy, 17(3), 525-545. Caudy, M. S., Durso, J. M., & Taxman, F. S. (2013). How well do dynamic needs predict recidivism? Implications for risk assessment and risk reduction. Journal of Criminal Justice, 41(6), 458-466. Cording, J. R., & Beggs Christofferson, S. M. (2017). Theoretical and practical issues for the measurement of protective factors. Aggression and Violent Behavior, 32, 45-54. Duwe, G. (2014). The development, validity, and reliability of the Minnesota Screening Tool Assessing Recidivism Risk (MnSTARR). Criminal Justice Policy Review, 25(5), Farabee, D., & Zhang, S. (2007). COMPAS validation study: First annual report. Los Angeles, CA: Department of Corrections and Rehabilitation. Hamilton, Z., Kigerl, A., Campagna, M., Barnoski, R., Lee, S., Van Wormer, J., & Block, L. (2016). Designed to fit: The development and validation of the STRONG-R recidivism risk assessment. Criminal Justice and behavior, 43(2), 230-263. Hamilton, Z., Campagna, M., Tollefsbol, E., Van Wormer, J., & Barnoski, R. (2017). A more consistent application of the RNR Model. Criminal Justice And Behavior, 44(2), Hamilton, Z., Duwe, G., Kigerl, A., Gwinn, J., Langan, N., & Dollar, C. (2021). Tailoring to a mandate: The development and validation of the Prisoner Assessment Tool Targeting Estimated Risk and Needs (PATTERN). Justice Quarterly, 1-27. Hamilton, Z., Kigerl, A., & Kowalski, M. (2021). Prediction is Local: The benefits of risk assessment optimization. Justice Quarterly, 1-23. Hamilton, Z., Kowalski, M. A., Kigerl, A., & Routh, D. (2019). Optimizing youth risk assessment performance: Development of the Modified Positive Achievement 26 PERSPECTIVESVOLUME 46, NUMBER 2 RISK NEED ASSESSMENT Change Tool in Washington State. Criminal Justice And Behavior, 46(8), 1106-1127. Jung, S., & Dowker, B. A. (2016). Responsivity factors among offenders [Article]. Journal of Offender Rehabilitation, 55(3), McDougall, A., Dyck, H., Macaulay, A., Wershler, J., Canales, D. D., & Campbell, M. A. (2014). The responsivity principle of offender case management and the case of the long forgotten “R”. Psynopsis: Canada’s Psychology Newspaper, 36(1), 18-19. Mei, X., Hamilton, Z., Kowalski, M., & Kigerl, A. (2021). Redesigning the Central Eight: Introducing the M-PACT Six. Youth Violence and Juvenile Justice, 15412040211014264. Miller, W. T., Campbell, C. A., & Larnell, T. (2021). Bias detected? An examination of criminal history using the OYAS-DIS for girls and black youth. Journal of Ethnicity in Criminal Justice, 19(2), 101-119. Montford, K., & Hannah-Moffat, K. (2021). The veneers of empiricism: Gender, race and prison classification. Aggression and Violent Behavior, 59, 101475. Onifade, E., Davidson, W., & Campbell, C. (2009). Risk assessment: The predictive validity of the youth level of service case management inventory with African Americans and girls. Journal of Ethnicity in Criminal Justice, 7(3), 205-221. Rennie, C. E., & Dolan, M. C. (2010). The significance of protective factors in the assessment of risk. Criminal Behaviour and Mental Health, 20(1), 8-22. Rettinger, L. J., & Andrews, D. A. (2010). General risk and need, gender specificity, and the recidivism of female offenders. Criminal Justice And Behavior, 37(1), 29-46. Schmidt, F., Hoge, R. D., & Gomes, L. (2005). Reliability and validity analyses of the youth level of service/case management inventory. Criminal Justice And Behavior, 32(3), 329-344. Sullivan, C. J., & Childs, K. K. (2021). Juvenile Risk and Needs Assessment: Theory, Research, Policy, and Practice. Routledge. Taxman, F. S., Pattavina, A., Caudy, M. S., Byrne, J., & Durso, J. (2013). The empirical basis for the RNR model with an updated RNR conceptual framework. In Simulation strategies to reduce recidivism (pp. 73-111). Springer. Van Voorhis, P., Wright, E. M., Salisbury, E., & Bauman, A. (2010). Women’s risk factors and their contributions to existing risk/needs assessment. Criminal Justice And Behavior, 37(3), 261-288. Author Bios: Zachary Hamilton is an Associate Professor of Criminology and Criminal Justice and Associate Director of the Nebraska Center for Justice Research at the University of Nebraska - Omaha. His research on risk and needs assessment led to the development of the Static Risk Offender Needs Guide – Revised (STRONG-R) and the Modified Positive Achievement Change Tool (MPACT). These assessments identify the supervision level and programming needs for juveniles and adults, which are currently used in more than a dozen states. He has published over 50 journal articles, chapters, and books on risk and needs assessment, evidence-based practices, and program efficacy. Baylee Allen is a research assistant and doctoral student at the University of Nebraska- Omaha School of Criminology and Criminal Justice. Her research interests in risk and needs assessments and case management practices comes from working for the Nebraska Department of Correctional Services for several years. Her work in case management and quality assurance showed her the power that quality case management services can have with offender populations. Her goal is to make prison a safer place to live and work. Addison Kobie is a current doctoral student and research assistant at University of Nebraska-Omaha, where she studies Criminal Justice and Criminology. She has a M.A. from Sam Houston State University in Criminal Justice and Criminology. Her research interests include juvenile delinquency, corrections, and risk assessments. Recently, her work has been focused on risk and need assessments with special populations.27 AMERICAN PROBATION AND PAROLE ASSOCIATION RISK NEED ASSESSMENT PPPS WEEK AD AMERICAN PROBATION AND PAROLE ASSOCIATION Pretrial, Probation, and Parole Supervision Week www.appa-net.org/PPP-Supervision-Week | #pppsweek Sponsored by28 PERSPECTIVESVOLUME 46, NUMBER 2 RISK NEED ASSESSMENT TALKING ABOUT THE WAY WE TALK: Understanding Assessment Tool Communication to Improve Core Correctional Practices KIMBERLY R. KRAS, SHANNON MAGNUSON, & FAYE S. TAXMAN29 AMERICAN PROBATION AND PAROLE ASSOCIATION RISK NEED ASSESSMENT Risk/Need Assessments (RNAs) are an established evidence-based practice (EBP) and widely used in community supervision as well as the criminal-legal system more broadly (pre-trial release, parole processes, etc.). RNA tools help criminal justice system officials (pre-trial officers, corrections staff, community corrections officers) assess justice-involved individuals’ risk for recidivism, identify areas of need that when addressed may mitigate this risk, and use this information to develop a case plan for appropriate services and interventions to guide an individual’s time under correctional supervision toward successful completion of probation or parole supervision. However, there is a growing tension about how RNA tools are built by researchers and used by practitioners. For example, advocates for system-impacted people note that researchers and RNA developers build both pretrial and post-conviction tools with potentially problematic and racist arrest data (Freeman et al., 2021; Pre-Trial Justice Institute, 2020). There is also concern the tools increase probation and parole revocation rates and exacerbate racial and ethnic disparities when agencies use risk scores to drive contact standards. This latter concern taps into another growing tension about the use of “risk” language and how community corrections staff talk about and use the concept of risk in practice (Rudes et al., 2016). More specifically, at times actuarial risk and perceived risk may be conflated. Actuarial risk scores typically come from an array of indicators including static risk (things someone can’t change, like criminal history or age) and dynamic risk (things someone can change, like substance misuse or employment), also known as criminogenic needs. Perceived risk reflects one’s viewpoints about who might be a “risky” person, based on implicit bias or prior experiences. Prior experiences with certain “types of people” or “types of situations” can create a “cognitive shorthand” which encourages individuals to overestimate the risk of individuals they are currently working with. In community corrections settings, this lends itself to staff perceiving individuals as “high risk” or “risky” when RNA tool results would indicate otherwise. However, both perceived risk and RNA tools use the language of “risk,” complicating how we think about these concepts in everyday conversations and creating a barrier to parsing out the differences. Importantly, using the language interchangeably can influence how staff use RNA results, what information they include in the case planning process, how they broker resources, and ultimately what they report to the court (Steiner et al., 2011). These tensions are important, and we believe the tools themselves need continued evolution, including reconsidering whether measuring risk is even necessary. We also recognize actuarial assessments that predict an outcome (e.g., arrest or community supervision completion) are an EBP guiding day-to-day community supervision, and the use of these actuarial assessments is a critical step in the supervision process. To that end, it is imperative we continue unpacking how and why we engage in this practice. In this article, we contend that communicating information to individuals on community supervision about the goals of using assessments and the results of these assessment tools is an essential practice. Whichever assessment tool is used, communicating why it is used, how it is used, and the results comes with important opportunities to impact the individual experience and influence outcomes, especially success. This article will examine the micro-processes of communicating the assessment in practice, and how best to incorporate a person-centered and restorative approach. We contend that effective communication of the purpose, function, and results of assessments will result in a more transparent, equitable, and humanizing community supervision experience. Communication in Community Supervision Practice In criminology and criminal justice, researchers studying the use of assessment tools in practice have largely ignored the importance of communicating assessment information. Even less is known about if and how practitioners are attending to issues of communication. Drawing on research about an officer’s interactional style provides some knowledge about communication strategies. For example, the work of Kennealy and colleagues (2012) suggests the dual role that officers play in community supervision, and a style that is “firm, fair, and caring” is most effective at achieving individual TALKING ABOUT THE WAY WE TALK: UNDERSTANDING ASSESSMENT TOOL COMMUNICATION TO IMPROVE CORE CORRECTIONAL PRACTICESNext >