Bayesian Analysis of Children?s Moral Judgments on the Trolley Cases

The current project tested two competing views — with particular emphasis on the extent to which these competing accounts could explain an interesting and robust pattern of findings concerning the classic 'trolley problem'. Specifically, this project tested aspects of Greene's dual-process model (Greene, Sommerville, Nystrom, Darley, & Cohen, 2001) and Mikhail's (2000) theory of universal moral grammar. We also extended the research of Pellizzoni, Siegal, and Surian (2010), who utilized a sample of young children in an attempt to test Greene's theory that moral judgments are driven by differences in the directness of harm. The current study found little evidence suggesting that children between the ages of 3 and 5 consistently aligned their moral judgments with the contact principle. Furthermore, there was also little evidence that children in this age group conformed their judgments to the doctrine of double effect, which would have supported Mikhail's account. Finally, we also investigated candidate predictors of the tendency to conform to these heuristic principles. This revealed that inhibitory control negatively predicted the tendency to align one's judgments with the contact principle. This seems to support Greene's theoretical view. In addition, we found that participant age was a negative predictor of the tendency to align one's judgments with the doctrine of double effect — an unexpected result. This project adopted a Bayesian approach because of its overall flexibility and the ease with which missing data can be handled.