Social interaction effects in the laboratory and society: Conformism and reciprocity in public good provision

People contribute more to public goods, the more others give ('crowding-in'). We investigate two possible causes of crowding-in: reciprocity, the usual explanation, and conformism, a neglected alternative. The issue is important since conformism has more scope to bring about endogenous social interaction effects outside the laboratory. We use Bardsley's (2000) Conditional Information Lottery design to control 'others' behaviour' without deception. We observe how contributions correlate within a group and between groups. Our main result is that conformism does cause crowding-in. The common presumption of reciprocity therefore deserves to be questioned, even in the laboratory.