Experimental Design
The experiment is implemented online on Amazon Mechanical Turk via Cloud Research. Before randomization, we collect the socio-demographic characteristics of respondents, as well as their perceived importance and knowledge of the policy issues used as main outcomes. Participants are then randomized into (i) treatment messages, each consisting of written policy recommendation on two policy issues, and (ii) control messages, where the same policy recommendations are reported separately (i.e. from different sources). The issues bundled differ in ideological value, policy domain, and complexity. After reading the messages, respondents are asked to provide their policy preference of the two issues.