Abstract
<p>People make many decisions in which increasing the benefits for some people reduces the benefits for others. What determines which lessons people learn from the outcomes of those decisions? We propose that selective attention can bias what people learn from such decisions. We tested this hypothesis across four preregistered experiments and an internal meta-analysis pooling data from 11,603 participants. In each experiment, participants made decisions in a social dilemma pitting the interests of a financially stable ingroup against those of a distant, impoverished outgroup. Critically, when participants subsequently learned about the consequences of their previous decision, an experimental manipulation selectively directed their attention to its effect on their ingroup, its effect on the outgroup, or irrelevant information. Across all experiments, directing attention to outgroup consequences in the previous decision increased subsequent donations to the outgroup, while directing attention to the previous decision’s ingroup consequences reinforced parochialism. These effects accumulated over repeated decisions, with learning trajectories diverging as a function of attention condition. The meta-analysis pointed toward meaningful prosocial learning effects of both outgroup-directed attention (d = 0.36 vs. control) and ingroup-directed attention (d=−0.21 vs. control), with a medium-sized intergroup contrast (d= 0.56). Empathic concern for the victim of a past decision partially mediated the effect of directing attention to consequences for the ingroup versus the outgroup. These findings demonstrate that prosocial learning is shaped by which consequences people attend to.</p>