Social Norms in Survey Experiments: Personal Beliefs versus Normative Expectations

In Experimental Economics, coordination games are used to elicit social norms as incentivized beliefs about others' beliefs (Normative Expectations). Conversely, representative surveys, such as the World Values Survey, elicit social norms as Personal Beliefs that are independent of others' beliefs. We design a survey experiment based on a representative sample of the Italian population (N = 1,501) to compare these two measures of social norms, with attitudes toward gender roles as a working example. We obtain the following results. At the aggregate level, agreement ratings from the two measures follow a similar pattern but differ significantly in magnitude, with norms based on Personal Beliefs depicting a more progressive view of gender roles. The individual-level analysis accounts for this result: most respondents report Personal Beliefs that are more progressive than their perceptions of others' beliefs, regardless of the accuracy of those perceptions. This behavior is positively associated with holding a university degree and may reflect genuinely progressive views, a bias toward appearing progressive, or an ongoing shift in social norms. Finally, we compare the explanatory power of the two measures of social norms in analyzing a set of women's labor market outcomes, showing that Normative Expectations outperform Personal Beliefs when the norm is ``strong.''