Almost politically acceptable criminal justice risk assessment

Abstract

Modern, algorithmic risk tools are trained in the sense that they inductively seek associations between predictors (e.g., prior record) and an outcome of interest (e.g., a rearrest). There is no model. 3 The associations are used to construct a measure of risk. Risk can be represented as a numerical score, a probability of an outcome, or a particular outcome class. Because criminal justice decisions are typically categorical (e.g., release on parole or not), often built into the algorithmic is machinery to translate a numerical score or probability into an appropriate categorical outcome. If not, some less formal means are required to generate an outcome class to serve as the forecast. A probability of a rearrest, for instance, is not actionable until it is high enough to warrant assignment to the outcome class of, say, "recidivist." Once training is completed, the risk algorithm can be used to forecast outcomes for new cases when their outcomes are unknown. Data for such forecasting must be collected and properly managed, as is done for the training data. The process can be demanding because the data used for forecasting must have the same predictors as the data used for training and be realized in the same fashion. Instructive forecasts might well be obtained, for example, from training data and forecasting data properly realized from the same jurisdiction, in roughly the same time period, and for the same criminal justice setting (e.g., parole decisions). And just as for the training data, there will usually be important data quality concerns. The forecasting data are provided to a trained algorithm that, in turn, produces forecasts for each case. Ideally, these are forecasted outcome classes such as a post-release arrest for a violent crime, a post-release arrest for a nonviolent crime, or no post-release arrest whatsoever. Often, the reliability of those forecasts also can be determined (Berk, 2018b). With the forecasts and (preferably) reliabilities in hand, decisions can be made and actions can be taken. The recent risk literature makes clear that machine learning, algorithmic methods can have demonstrable, superior performance compared with subjective or model-based methods (Berk, 2020; Berk & Bleich, 2014; Berk, Sorenson, & Barnes, 2016). It has been long known that even simple statistical methods produce better accuracy relative to subjective approaches (Dawes, Faust, & Meehl, 1989; Meehl, 1954), and as a mathematical matter, machine learning algorithms can adaptively find far more complex relationships in the data, should they exist, compared with conventional models such as logistic regression (Hastie, Tibshirani, & Friedman, 2009). When the relationships between predictors and an outcome are not complex, algorithmic methods properly applied will perform no worse than conventional models.

Crime, inequality and public health: a survey of emerging trends in urban data science

“…Besides traditional statistical approaches that target either correlational or causal outcomes, geospatial modeling, network science, agent-based modeling, and machine learning have been the four main areas on which scholars have focused their attention. Virtually every area of criminology and crime research has been—to some extent—explored by computational approaches: from white collar crime (Ribeiro et al, 2018 ; Luna-Pla and Nicolás-Carlock, 2020 ; Kertész and Wachs, 2021 ) to terrorism (Moon and Carley, 2007 ; Chuang et al, 2019 ; Campedelli et al, 2021 ), from illicit drugs (Mackey et al, 2018 ; Magliocca et al, 2019 ; Sarker et al, 2019 ) to organized crime (Nardin et al, 2016 ; Troitzsch, 2017 ; Calderoni et al, 2021 ), from gun violence (Mohler, 2014 ; Green et al, 2017 ; Loeffler and Flaxman, 2018 ) to cyber-crime (Shalaginov et al, 2017 ; Duxbury and Haynie, 2018 , 2020 ), from recidivism (Tollenaar and van der Heijden, 2013 ; Duwe and Kim, 2017 ; Berk and Elzarka, 2020 ) to predictive policing (Caplan et al, 2011 ; Mohler et al, 2011 ; Perry, 2013 ). Particularly, the dialogue between computational methods and the study of recidivism and predictive policing not only focused on technical innovations to optimize forecasting and predictive models, but also provoked vivid debates regarding critical issues of algorithmic accountability, fairness, and transparency (Lum and Isaac, 2016 ; Dressel and Farid, 2018 ; Richardson et al, 2019 ; Akpinar et al, 2021 ; Purves, 2022 ).…”

The Effect of Operational Strategy on Organizational Performance of Commercial Banks in Meru County, Kenya

“…According to the results, 83 questionnaires were filled which gave a return rate of 87.4%. [7] argued that return rates are termed acceptable, good and very good if 50%, 60% and 70% respectively are achieved and thus accepted for analysis and publication. The high return rate obtained is attributable to self-administration of the questionnaires by the researcher.…”

A call to dismantle systemic racism in criminal legal systems.

“…Psychologists have drawn attention to the need for careful consideration of inherent bias in some risk assessment tools, particularly those that are based almost entirely on one’s prior criminal record (e.g., Vincent & Viljoen, 2020). In one response to the issue, Berk and Elzarka (2020) recommended using algorithmic recidivism risk assessment tools based on predictors in samples of White offenders only. However, this approach frames the racial injustice inherent to existing risk assessment methods as a problem not of fundamental fairness but rather of political acceptability, doubles down on the assumption of White normativeness, and ignores the unique risk and resilience factors affecting outcomes for offenders from racial and ethnic minority backgrounds.…”