Individual Submission Summary
Share...

Direct link:

Do Surveys Other than ANES Mobilize the Electorate?

Sat, August 31, 8:00 to 9:30am, Hilton, Columbia 1

Abstract

Surveys are important instruments for understanding political behavior, but they may also affect the behavior of the respondents who we are trying to study. Warren Miller once said that the American National Election Studies (ANES) is “the most expensive voter mobilization project in American history.” One study of the mobilization effects of the ANES in 2012 (Jackman and Spahn 2018) has found a modest effect of those lengthy in-person interviews on turnout. We extend that work by examining the mobilization effects of a telephone and an internet survey on the likelihood of voting in a midterm election in a large, competitive state. While the findings of Jackman and Spahn are important, the ANES is uncharacteristic of most surveys in content, length, and mode, so it would be valuable to know whether surveys with more commonly used formats also mobilize voters.

To determine that, we estimate the effects of having participated in a survey in Florida in October 2017 on the likelihood of voting in the November 2018 midterm election. In Fall 2017, we obtained the September 2017 Florida voter file, and randomly assigned sampled registrants into either an internet or phone survey mode. In October 2017, we conducted a dual mode survey of (5,881) internet and (838) phone respondents focused primarily on the upcoming 2018 US Senate race between then-Governor Rick Scott and incumbent Bill Nelson. We will obtain the January 2019 voter file, and matching on the voter ID number, we will be able to determine whether respondents who were interviewed by phone and by internet were any more likely to have voted than registrants in the September 2017 voter file who we did not interview. Age, sex, race, and voter history are included in the voter file, so they can be included as controls.

Authors