``Validating Self-reported Turnout by Linking Public Opinion Surveys with Administrative Records.''



It is widely known that the self-reported turnout rates obtained from public opinion surveys tend to substantially over-estimate the official turnout rate. For example, in recent presidential elections, the self-reported turnout rate from the American National Election Studies (ANES) has consistently exceeded the actual turnout rate by approximately 15 percentage points. However, scholars sharply disagree on what causes the bias of self-reported turnout. Some blame misreporting due to social desirability, others attribute the bias due to non-response or other factors such as the quality of voter records. While in recent years it has become possible to validate self-reported turnout by directly linking surveys with administrative records, most existing validation studies rely on proprietary merging algorithms with limited scientific transparency and yield conflicting results. To shed a light on this debate, we apply a canonical probabilistic record linkage model, implemented via the open-source software package fastLink, to merge two major election studies - the ANES and the Cooperative Congressional Election Survey (CCES) - with a national voter file of over 180 million records. For both ANES and CCES, fastLink successfully produces validated turnout rates close to the actual turnout rates. Using these merged data sets, we find that the bias of self-reported turnout originates primarily from misreporting rather than survey non-response. Our findings suggest that those who are educated and interested in politics are more likely to over-report turnout. Finally, we show that fastLink outperforms a proprietary algorithm. (Last Revised, March 2018)
Our method is used to validate the self-reported turnout in the 2016 American National Election Study. Our turnout validation data are available at the ANES website.

  Related Paper and Software

Enamorado, Ted, Benjamin Fifield, and Kosuke Imai. ``Using a Probabilistic Model to Assist Merging of Large-scale Administrative Records.''
Enamorado, Ted, Benjamin Fifield, and Kosuke Imai. ``fastLink: Fast Probabilistic Record Linkage.'' available through The Comprehensive R Archive Network.

© Kosuke Imai
 Last modified: Thu Mar 1 08:04:35 EST 2018