Ideology and Vote Choice in U.S. Mayoral Elections: Evidence from Facebook Surveys

Michael Sances

Are city elections about policy? While some may see the answer to this question as obvious, it turns out to be a rather tricky question to answer. Some urban politics scholars conclude city elections are apolitical, given voters have already “voted with their feet” by deciding where they wish to live; others argue urban elections are about conflicts between social groups. Yet it has been hard to answer this question definitively, given the difficulty of surveying local voters. At the same time, evidence is growing that city officials behave ideologically when they make policy decisions, and that city policies overall tend to reflect what voters say they want. These recent findings regarding city responsiveness, as well as advances in survey recruitment technology, prompt me to take a fresh look at the role of policy voting in city elections.
Using Facebook advertisements targeted to particular local audiences, I recruit samples of hundreds of voters during the 2015 mayoral elections in Memphis and Nashville, TN, as well as an additional sample of over 1,000 voters from smaller cities in Illinois. These data allow me to estimate the relationship between voter ideology and candidate choice, adjusting for other observable and unobservable determinants of voting. Across a variety of samples and estimation strategies, I find ideology has a strong influence on local vote choice.
First, in cross-sectional analyses, I find that ideology is a powerful predictor of vote choice in Memphis and Nashville, controlling for demographic characteristics such as race. This provides suggestive evidence that ideology matters; however, given I am working with observational data, alternative explanations are possible. Thus, to further disentangle ideology from other factors, I next conduct an analysis of voter learning in the Nashville mayoral election. Using a panel design where the same respondents are interviewed before and after the election, I show that voter knowledge of candidates’ ideological positions increases significantly over the course of the election campaign. Further, this learning causally impacts voters’ choices, with liberals (conservatives) who learn becoming more (less) likely to vote for the liberal candidate.
Of course, it may be that ideology operates differently in large cities, or in cities in Tennessee. To test the generalizability of these results, I field a survey experiment to voters in smaller Illinois cities, presenting them with hypothetical candidates for village and city mayors. As in the big-city studies, I find these small-city voters also weight ideology heavily in their voting decisions, more than any other factor.
These results show that electoral accountability for policy is indeed possible in both large and small U.S. cities, and offer an explanation for recent evidence of municipal responsiveness. Yet they also showcase the potential of using geotargeted online surveys to study local political behavior – a subject which scholars know surprisingly little about compared to national-level behavior. A recent count of election-related articles published in major journals over 20 years shows less than 1% focus on local elections. This is due in part to the relative difficulty of collecting original survey data from local contexts. By introducing a new, relatively low-cost method of collecting geographically targeted data, my article breaks down barriers that have long kept the study of local elections separate from research conducted at other levels of government.
Advertisements

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s