In this paper, we explore whether a deliberative condition affects democratic preferences of ordinary citizens. We run a population-based vignette and conjoint experiment, testing which governance schemes – representative democracy, direct democracy, deliberative democracy, and stealth democracy as well as combined variants - exhibit more legitimacy and which do not. The basic setup of the experiment consists of randomly assigning survey respondents in Germany into three groups: (1) a deliberation group getting vignettes and a conjoint on governance schemes including information about the pros and cons of different governance schemes; this group has the possibility to deliberate about their democratic preferences in an asynchronous chat; (2) a first control group getting vignettes and a conjoint on governance schemes including information about the pros and cons of different governance schemes but with no possibility for discussion; and (3) a pure control group getting only the vignettes and the conjoint. The vignettes and the conjoint will present a story on a local construction project, namely the building of an asylum home. We expect that the deliberative condition induces preference clarification, a better alignment of basic politics preferences with the various governance schemes (consistency), and better knowledge of the pros and cons of the various schemes.
