Toward a predictive model of moral concern (Stage 1 Registered Report)

At the heart of many contentious debates—on the ethicality of performing an abortion, eating meat, or denying protection to refugees, for example—is the question of how much we should value the welfare and interests of different entities. Previous research has identified numerous characteristics that predict how much moral concern people show toward different entities (e.g., empathy, perceived similarity). However, integrating this quickly accumulating evidence has proven difficult and a general, psychological model of moral concern has not emerged. The present study (n = 800 U.S. participants) aims to address these issues. We first review the literature to identify hypothesized predictors of moral concern. We then use a machine learning approach to build an integrative, parsimonious, predictive model of moral concern. Our findings provide insights into (a) the collective predictive performance of the variables identified in previous studies and (b) which characteristics are most important for predicting moral concern.