Anatomy of Risk: Cumulative Disadvantage and Risk Assessment Instruments

Abstract

One significant critique of the United States criminal justice system is the disproportionate rate at which people of color and people living in poverty are represented. Proponents of the use of risk assessment instruments in criminal justice system decision-making claim the tools are likely to reduce the effects of implicit biases and stereotypes in decision-making. On the other hand, critics have suggested that the use of risk assessment instruments in sentencing may increase the racial and socioeconomic disparities that they are ostensibly positioned to address. This integrative literature review examines risk domains commonly measured in risk assessment instruments and the potential for these tools to capture the effects of cumulative disadvantage. Seven risk domains are identified based on analysis by the Pennsylvania Sentencing Commission of 29 risk assessment instruments comprising a total of 125 risk factors. These domains include demographic factors, psychosocial factors, mental health, substance abuse, antisocial/psychopathic traits, criminal history, and dynamic factors. Research that assesses the effects of race/ethnicity and/or socioeconomic status in these domains is reviewed to examine the differential outcomes related to these statuses. In addition, research is analyzed and integrated as the risk domains relate to one another. This analytical approach is chosen based on suggestions that negative outcomes related to race and socioeconomic status are often cumulative in nature, thereby suggesting cumulative disadvantage models can provide insight into the odds ratios of various combinations of risk factors based on group membership. Extensive research in each of the identified risk domains highlights the frequently disadvantageous effects of race and socioeconomic status on domain-relevant outcomes (e.g., educational achievement, employment). Some research evidence also indicates that disadvantage in one domain often predicts negative outcomes in other domains. For example, mental health problems increase the likelihood of substance use disorder, and both of these exhibit negative effects on educational achievement and employment outcomes. The research in each of the seven domains is reviewed in a linear fashion as they might be expected to become relevant and salient over time and across the life course. The analysis and discussion of the research findings suggest that the accumulation of risk factors across time can occur independent of individual characteristics, in accordance with cumulative disadvantage theory. The magnitude of the effects of these disadvantages likely varies among groups. Further, the differential exposure to risk factors between groups suggests minorities and the poor are more likely to be disadvantaged in multiple risk domains, thereby increasing the likelihood they will encounter disadvantages that also increase the likelihood of (re)offending. The current research applies a cumulative disadvantage framework to clarify the intersectionalities of race, socioeconomic status, and recidivism risk factors, as well as the interactive relationship risk factors have with one another. The differential distribution of risk factors among groups suggests that the use of risk assessment categorizations in a punitive context may exacerbate racial and socioeconomic disparities in the justice system. Research evidence suggests policymakers should avoid the use of risk to increase punishment. Future research should focus on identifying constructs underlying each risk domain, apply empirical testing utilizing cumulative disadvantage models, and assess the variance in likelihood ratios of multiple risk factors by racial/ethnic group membership and socioeconomic status. Given the lack of research currently addressing the potential impacts of criminal justice decisions based on risk scores, policymakers are well-advised to use caution in the widespread implementation of risk assessment tools, particularly in contexts that are likely to increase the severity of punishment based on risk.