We reconsider evidence from experiments that claim to show that using “house money” in standard public goods experiments has no effect on behavior. We show that it does have an effect when one examines the data using appropriate statistical methods that consider individual-level responses...