Variations in the Two-Step Norm Elicitation Procedure

The two-step norm elicitation procedure describes a commonly used tool for measuring normative expectations in an incentivized way. This study tests some of its design features to determine whether elicited beliefs and related behaviours vary depending on i) the time of elicitation (before vs after the decision), ii) incentivizing vs not incentivizing a question about normative expectations, and iii) questioning subjects on their beliefs about the action of interest alone or combined with an alternative action. An online experiment is conducted via Prolific comprising a dictator game and the elicitation of fairness beliefs. A pretest reveals that applying role uncertainty does not alter beliefs and behaviours compared to a baseline treatment without it. Subsequently, three treatments are implemented. Contrary to previous studies, results indicate that varying the time of elicitation does not significantly alter the money-share decision. However, incentivizing the question about normative expectations significantly increases the fit with the actual majority norm. Finally, asking about a fair share and an unfair share instead of only about fair sharing does not alter personal normative beliefs or normative expectations, but it increases the empirical expectations that other dictators have provided a fair share.