Estimating the causal effects of policy information on voter turnout: An Internet based randomized field experiment in Japan

Abstract

Political scientists since Anthony Downs (1957) have hypothesized that more policy information leads to a higher voter turnout. To empirically test this hypothesis, we conducted an Internet-based randomized field experiment during Japan’s 2004 Upper House election. Japan’s 2004 election is ideal for testing our hypothesis because political parties prepared formal “manifestos” that include proposals for major policy issues. We find that voters are less likely to abstain when they receive policy information about both ruling and opposition parties through their official party websites. The information effects are larger among those voters who were planning to vote but were undecided about which party to vote for. Our findings also shed light on the role of the Internet as an important source of information for voters. Methodologically, this paper illustrates how to design and analyze Internet-based randomized field experiments, which are becoming increasingly common in the social sciences. Our experimental approach avoids the problem of endogenous information acquisition, which is inherent when using observational studies to estimate the causal effects of information on voting behavior. Furthermore, we employ a randomized block design to ensure efficient randomization, and apply a Bayesian statistical model to account for noncompliance and nonresponse, the two prevailing problems of field experiments.