Predicting sentencing for low-level crimes: Comparing models of human judgment

Laws and guidelines regulating legal decision making are often imposed without taking the cognitive processes of the legal decision maker into account. In the case of sentencing, this raises the question of whether the sentencing decisions of prosecutors and judges are consistent with legal policy. Especially in handling low-level crimes, legal personnel suffer from high case loads and time pressure, which can make it difficult to comply with the often complex rulings of the law. To understand the cognitive processes underlying sentencing decisions, an analysis of trial records in cases of larceny, fraud, and forgery was conducted. Applying a Bayesian approach, five models of human judgment were tested against each other to predict the sentencing recommendations of the prosecution and to identify the crucial factors influencing sentencing decisions. The factors influencing sentencing were broadly consistent with the penal code. However, the prosecutors considered only a limited number of factors and neglected factors that were legally relevant and rated as highly important. Furthermore, testing the various cognitive judgment models against each other revealed that the sentencing process was apparently not consistent with the judgment policy recommended by the legal literature. Instead, the results show that prosecutors’ sentencing recommendations were best described by the mapping model, a heuristic model of quantitative estimation. According to this model, sentencing recommendations rely on a categorization of cases based on the cases’ characteristics.