Common measures of cultural attitudes, such as those constructed from the World Values Survey, are characterized by substantial within-country volatility. This volatility is at odds with the notion of culture adopted in economics: a set of slow-moving traits that determine preferences and expectations transmitted from one generation to the next via family or social interactions. The insufficient persistence of survey proxies for such traits may compromise empirical studies of culture as a determinant of economic outcomes. We illustrate this point via a thorough replication, using the most recent WVS waves, of analyses carried out previously for regions in Europe.