Path Dependency in Jury Decision-Making

A large behavioral economics literature is concerned with cognitive biases in individual and group decisions, including sequential decisions. These studies primarily find a negative path-dependency consistent with mechanisms such as the gambler's fallacy or contrast effects. We provide the first test for such biases in group decision making using observational data. Specifically, we study more than 27,000 verdicts adjudicated sequentially by over 900 juries for high-stakes criminal cases at London's Old Bailey Criminal Court in the 18th and 19th centuries. Using jury fixed effects to account for heterogeneity in their baseline propensity to convict, we find that a previous guilty verdict significantly increases the chance of a subsequent guilty verdict by 6.7-14.1%. This positive autocorrelation, which contrasts previous studies, is (i) robust to alternative estimation strategies, (ii) independent of jury experience and (iii) driven by the most recent lag and pairs of similar cases. Potential explanations of such positive path dependence include sequential assimilation effects, which may reflect a jury's desire to be internally consistent when deciding comparable cases and short-term 'emotional' impacts of the characteristics and/or outcome of one case on another. As in modern-day jury studies, our results highlight the possibility that factors independent of the facts and evidence of the current case affect jury behavior.