Measuring social norms: direct and indirect methods and the role of financial incentives in reducing misreporting

We test two common features of the opinion-matching method to elicit personal beliefs and perceived social norms regarding female labor force participation among a sample of young Tunisian men and women. Our survey experiment contains two orthogonal treatment arms. In the first treatment arm, respondents were randomly assigned to answer a direct question about their personal beliefs regarding women working outside the home and an indirect questioning technique that provides more cover for the respondent: the list experiment. We find significant discrepancies between the two, particularly among male respondents – suggesting either biased personal beliefs in the direct question or non-strategic misreporting in the list experiment. It follows that, depending on which method is deemed more credible and which gender sample is considered, the prevalence of pluralistic ignorance varies widely in our setting. For the second treatment arm, we provided half of our sample with a financial incentive when eliciting perceived social norms regarding the appropriateness of women working. We find that financial incentives are necessary for both men and women to distinguish their personal beliefs from their perception of what others believe.