Can More Information Lead to More Voter Polarization? Experimental Evidence from Turkey

Many claim that increased availability of information via both old and new media drives political polarization, possibly undermining democratic institutions. However, rigorous evidence on this topic remains limited. I address this gap by conducting two experiments during a recent Turkish referendum that was on an important institutional change to weaken constraints on the executive. First, I designed a randomized door-to-door campaign. In this campaign, the opposition party gave uniform information on poor economic performance and increased terrorist activity under the incumbent's leadership to more than 130,000 voters. I show that voters, despite receiving the same information, diverged further in their vote choice on aggregate, leading to a significant increase in ideological polarization. The result is consistent with a model where polarization in vote choice is driven by differences in reaction to the same information and not self-segregation to different information sources, as others have assumed. The opposition failed to increase its vote share in this campaign, on aggregate, because it lacked the necessary data to target the subset of constituents that interpreted the information in their favor. In a second experiment with politicians, I confirm that the opposition had inadequate data on voters relative to the incumbent based on an analysis of roughly one million of their tweets. The evidence from both experiments, taken together, suggests that the incumbent party can exploit its access to higher quality data on voters to maintain its grip on power and advance an agenda that weakens democratic checks and balances.