Do Voters Lie? The Role and Reliability of Fieldwork in the Modern Day Campaign
Reviewed by Chris Meyer
June 30, 2014
The ubiquity of polls has become a staple of modern political campaigns. As close races drag on, candidates increasingly rely on polls to bolster their public image and justify their optimistic belief in an electoral victory. But of course, this leaves us with a pressing question. How are we to trust these polls, especially when they are the product of person-to-person interactions between canvassers and voters? To what extent must we consider the awkwardness and discomfort that come with these interactions, and the potential they have to skew poll results? In his original examination of electoral canvassing, Swarthmore undergraduate Daniel Block investigates the microtargeting techniques of political campaigns, and uses these metrics to determine whether survey respondents exaggerate the truth when asked which candidate they are supporting.
Microtargeting is a campaign technique designed to target on-the-fence voters. In a tight race, campaigns do not want to waste resources on party loyalists or lost causes. Therefore, microtargeting compiles data such as past voting records, political contributions, and demographic information to assign voters a numeric ideology score. Canvassers can then target the middle scores, or the most promising leads. Block argues that voters with these middling scores that express strong support for a candidate are likely to be lying, and demonstrates this with data drawn from the Sean Patrick Maloney Campaign for Congress, which helped elect its candidate to Congress as the representative of New York’s 18th district.
Here, Block is able to raise the most interesting observations in his paper, namely that a purportedly objective process like political polling is still subject to the vagaries of person-to-person interaction. If a voter wants to get a canvasser out of their doorway as quickly as possible, they will be more likely to overstate their support for a candidate and avoid follow-up questions. Furthermore, they may simply want to avoid upsetting the canvasser, and will thereby respond based on unconscious cues from their interviewer. Block finds that, in the case of the Maloney campaign, microtargeting data provides sufficient evidence to suggest that respondents exaggerated their support for the candidate, a phenomenon that is attributable to any of the social dynamics illustrated above.
See the full paper here!