Self-signaling and social-signaling in giving

Can a Bayesian signaling model explain patterns of giving driven by self-image concern as well as those driven by social-image? I experimentally test the predictions of such a model about how potential givers will respond to a change in the probability that their choice will be implemented. A self-signaler is predicted to respond with increased giving, but he predicted response of a social-signaler is increased giving, no change, or decreased giving depending on the information available to an outside observer. The experiment thus presents a test of the joint, independent and relative effects of social- and self-signaling in giving. The results provide little evidence of self-signaling, but stronger evidence of social-signaling.