Charness and Dufwenberg (2006) find that promises increase cooperation and suggest that the behavior of subjects in their experiment is driven by guilt aversion. By modifying the procedures to include a double blind social distance protocol we test an alternative explanation that promise keeping was due to external influence and reputational concerns. Our data are statistically indistinguishable from those of Charness and Dufwenberg and therefore provide strong evidence that their observed effects regarding the impact of communication are due to internal factors and not due to an outside bystander.