Are sentencing disparities unavoidable? Lessons from centralized judicial training

Sentencing disparities arise when judges impose different penalties in similar cases. Fundamentally unjust, such disparities have been a persistent target of reform. Instead of adopting numerical guidelines, Poland chose to centralize training and assessment of its judiciary in an effort to harmonize sentences. To measure the extent of sentencing disparities under the new system, we analyze sentences issued by trainee judges during a high-stakes judicial exam. This naturalistic approach allows us to mitigate the validity threats present in both observational and experimental vignette studies on sentencing disparities. Under controlled conditions, each trainee judge ( n = 248) spent six hours drafting a judgment based on an identical 100-page case file concerning a single offender charged with two counts. Despite receiving the same centralized training over the preceding three years, judges imposed markedly different penalties in the same case – ranging sixfold from four to 24 months. Sentences followed a distinctly multimodal distribution, with clusters of most common dosages, typically round numbers of months. Defying ordinal deserts, judges showed little agreement on which of the two counts deserves a harsher penalty. They also differed in how they combined penalties for two counts, in whether they applied ancillary orders, such as victim compensation, and in the extent of compliance with the prosecutorial sentence recommendation. This first evaluation of sentencing disparities after the reform yields disappointing results. Without structured guidelines, judges may continue to impose highly idiosyncratic sentences. • Centralizing the training of judges was meant to reduce sentencing disparities. • High-stakes judicial exam provides a naturalistic design to measure consistency. • Despite uniform training, judges imposed a wide range of penalties in the same case. • Ancillary orders and ordinal justice varied markedly across judges. • Combining penalties produced further inconsistency in sentence severity.