Social sampling shapes preferences for redistribution: Evidence from a national survey experiment

Abstract

We offer experimental evidence for the effect of social sampling on redistributive preferences through a survey experiment using a probabilistic national sample in Germany. We primed respondents to think about different types of social contacts, in particular low- and high-income contacts. We find evidence for an indirect effect in which the priming task shapes preferences for redistribution through its effect on the respondents' estimates of their contacts' incomes. Respondents in the low-income (high-income) priming recalled social contacts with lower (higher) incomes, which in turn predict more (less) support for redistributive policies. The indirect effect of the low-income (high-income) priming is stronger among high-income (low-income) respondents, suggesting that our priming task elicited the social contacts whom the respondents, given their own incomes, are less likely to recall. We discuss the implications of these findings to our understanding of how social sampling shapes redistributive preferences as well as relates to social networks and ideology.