The standard assumption of exogenous policy preferences implies that parties set their positions according to their voters' preferences. We investigate the reverse effect: Are the electorates' policy preferences responsive to party positions? In a representative German survey, we inform randomized treatment groups about the positions of political parties on two family policies, child care subsidy and universal student aid. In both experiments, results show that the treatment aligns the preferences of specific partisan groups with their preferred party's position on the policy under consideration, implying endogeneity of policy preferences. The information treatment also affects non-partisan swing voters.