Using Cell-Phone Mobility Data to Study Voter Turnout

• In Event: Geography and Spatial Analysis

Abstract

Scholars of voting behavior are often confronted with poor data availability or unhelpfully large units of aggregation for reported turnout (such as cities or states). We demonstrate a big-data solution to this challenge, using fine-grained cell-phone mobility data recorded on an hourly basis for over 43 million unique individuals on three recent election days in Tokyo, Japan. Our approach combines GPS data points and the geolocations of polling stations to uncover previously unavailable, individual-level (but anonymized for privacy) patterns of voting behavior. We first examine the validity of the GPS-based voter turnout data by comparing it to reported turnout at official administrative units of aggregation. We then demonstrate the substantive utility of our approach with an application exploring whether the extent of neighborhood-level damages sustained during the Tokyo firebombing in World War II relates to present-day patterns in voter participation, as hypothesized in an emerging literature on the long-run effects of political violence.

Authors

• Masataka Harada, Fukuoka University

• Gaku Ito, Hiroshima University

• Daniel M. Smith, Columbia University