November 3, 2016

Civic Engagement Strongly Tied to Local News Habits

Appendix A: Measuring local voting habits

The study suggests that being a dedicated voter in local elections is strongly connected to local news habits. To measure proclivity to vote, the survey used self-reports about frequency of voting in three types of elections: national elections, presidential primary elections and “local elections, such as for mayor or school board.” Overall, 27% of registered voters say they always vote in local elections.

Before settling on self-reported local voting behavior as the best method of measuring behavior, researchers also explored the possibility of measuring local voting behavior through the “voter file.” The voter file, gathered from publicly available individual voter lists from each state and standardized by a vendor, contains information on nearly every voters’ turnout history along with a variety of demographic information. The voter file does not indicate the candidates for whom a person voted, only whether he or she has a record of having voted in that election. Pew Research Center has suggested that incorporating voter file data in pre-election surveys may improve estimates of which respondents are most likely to vote.

However, the voter file poses several challenges for studying local voting. One of the biggest challenges is that local elections vary by jurisdiction: Elections for local, statewide and federal offices occur on the same date in some jurisdictions, while in other places they occur separately. Since the voter file records whether or not a person cast a ballot in a given election administration (rather than for a particular office or measure), in cases when multiple offices or measures are on the ballot, it is not possible to separate voting in elections for local offices from voting for statewide and federal offices using the voter file. Moreover, the voting records for some municipal elections are not reported to the state and thus will not be included in the state voter files from which the national file is compiled by the vendors. For these reasons, it was determined that the self-reported measure of local voting is a better-suited measure.

Finally, a common concern about self-reported measures of socially desirable behaviors like voting is that they may result in over-reporting. To minimize this, this study used a novel approach that asked respondents questions about voting in three types of elections, beginning with national and working down to local elections. Respondents thus had two opportunities to indicate that they were voters before being asked about local voting. In addition, the question text included information about the average rate of turnout for elections in each of the three types of elections – information designed to indicate to respondents that not voting in local elections is a common phenomenon. The full set of questions is available in the topline.