Avoiding information about adverse welfare consequences of self-interested decisions, orstrategic ignorance, is an important source of corruption, anti-social behavior and even atrocities. We model an agent who cares about self-image and has the opportunity to learn the social benefits of a personally costly action.  The trade-off between self-image concerns and material payoffs can lead the agent to use ignorance as an excuse, even if it is deliberately chosen. Two experiments, modeled after Dana, Weber, and Kuang (2007), show that a) many people will reveal relevant information about others' payoffs after making an ethical decision, but not before, and b)  some people are willing to pay for ignorance. These results corroborate the idea that Bayesian self-signaling drives people to avoid inconvenient facts in moral decisions.

Previous economic experiments on dual-process reasoning in altruistic decisions have yielded inconclusive results. However, these studies do not create a conflict between affective and cognitive motives, resulting in imperfect identification. We interact standard cognitive and affective manipulations in a giving task, and hypothesize that the affective manipulation has stronger effects when we simultaneously put the cognitive system under load. In line with earlier results, we find little evidence for dual-process reasoning in giving. Our independent treatment checks cast doubt on the effectiveness of standard treatment manipulations and show that both cognitive and affective manipulations consistently have opposite effects on the two sexes. We discuss the implications of our findings for economic experiments in this nascent research field.