Experimental methods: Pay one or pay all

In some experiments participants make multiple decisions; this feature facilitates gathering a considerable amount of incentivized data over the course of a compact session. A conservative payment scheme is to pay for the outcome from every decision made. An alternative approach is to pay for the outcome of only a subset of the choices made, with the amount at stake for this choice multiplied to compensate for the decreased likelihood of that choice's outcome being drawn for payoff. This “pay one” approach can help to avoid wealth effects, hedging, and bankruptcy considerations. A third method is to pay only a subset of the participants for their choices, thereby minimizing transactions costs. While the evidence on differences across payment methods is mixed, overall it suggests that paying for only a subset of periods or individuals is at least as effective as the “pay all” approach and can well be more effective. We further the discussion about how to best choose an incentive structure when designing an experiment.