Abstract: Political knowledge is crucial to the functioning of democratic political systems, and therefore represents an important topic of study for political scientists. Attempts to measure determinants of political knowledge, however, have long relied on self-reports of key causal variables (such as media exposure) and often must contend with concerns regarding endogeneity. To address both these issues simultaneously, we conduct a panel survey that spans the 2015 UK election to measure changes in political knowledge and rely on a non-self-reported measure of media exposure. More specifically, we analyze respondents Twitter feeds to objectively measure exposure to political information, including the source of that information. We show that information from media accounts tends to increase knowledge of factual questions, and that information from politicians tends to increase knowledge of party platforms. We also examine issues specifically relevant to the 2015 UK election, and find that information from incumbent parties improves estimates of the state of the economy while tweets by opposition parties diminish them, and that information from UKIP about immigration tends to inflate beliefs about the number of immigrants to the UK.

Download the paper