Self-Signaling Versus Social-Signaling in Giving

I investigate the relative importance of social-signaling versus self-signaling in driving giving. I derive specific qualitative predictions about how the response of an image-motivated dictator to a change in the probability that her choice will be implemented depends crucially on the information available to the relevant observer. A probabilistic dictator-game experiment tests the joint, relative, and independent effects of self-signaling and social-signaling. The results provide little evidence of self-signaling, but stronger evidence of social-signaling, particularly in a large subsample that excludes likely `selfish types'.