Citizens increasingly use the Internet to gather political information. Social media, in particular, has become a rich source, with people sharing, commenting on, and "liking" information they encounter. The cues available through these social activities may affect how people seek, evaluate, and respond when forming attitudes and evaluating candidates. We employ a novel experimental methodology where subjects respond to information as they do in social media environments. These responses are made available to other subjects in real time, facilitating studying dynamic social influences within the confines of a controlled environment. We use this software to investigate how social cues impact political information search and judgment. We hypothesize that social signals facilitate herd behavior in information seeking, focusing attention on certain content. We further hypothesize that social cues encourage attitudinal convergence, resulting in greater homogeneity in candidate preferences. Finally, we investigate whether social cues lead to greater polarization in candidate evaluations.
