Measuring the strength of voter turnout habits. (April 2020)

• Record Type: Journal Article
• Authors: Cravens, Matthew D.
• Is Part Of: Electoral studies. Volume 64(2020)
• Publication Date: 2020-04
• Subjects: Voter turnout -- Habit -- Measurement -- Automaticity -- Self-identity -- Political development; Elections -- Periodicals; Voting -- Periodicals; 324.6
• DOI: 10.1016/j.electstud.2020.102117
• Languages: English
• ISSNs: 0261-3794

Abstract

Abstract: Voting in one election makes a person more likely to vote in future elections. Researchers often interpret the over-time turnout persistence as evidence of psychological habit formation. But there are few theoretically motivated or empirically validated measures of voter turnout habit. This study contributes by clarifying the concept of turnout habit and developing and validating a seven-item measure of the concept. The paper describes turnout habit as a durable disposition to vote determined by an ability to automatically initiate voting and self-identify as a frequent voter. The new measure is validated using U.S. and UK survey data. Turnout data are from both voter files and self-reports. Varied methodological approaches, including a confirmatory factor analysis and receiver operator characteristic curve analysis, provide evidence of the measure's validity. Habit predicts future turnout independent of election-specific considerations, self-predictions, and age. The habit measure discriminates between eventual voters and abstainers as well as respondents' own self-predictions. Self-report voting habit measures appear valuable for answering theoretical questions about why people vote and practical questions about how best to identify likely voters in survey data. Highlights: A seven-item measure of voter turnout habit is developed and validated. Mechanisms responsible for turnout habit formation are proposed. Habit predicts turnout in U.S. and UK samples independent of election-specific considerations, self-predictions, and age. Habit measures have advantages in predicting turnout over self-predictions.