Past Research at ISERP
Election Campaigns and the Voter
by Robert Erikson (Political Science)
This project will complete our initial work which laid out a plan for analyzing the timeline of the political campaign, exploring the amount and nature of the change in election preferences reflect in trial-heat polls. One task was the measurement of the amount of actual volatility in preferences, adjusted for measurement error. The other was the accounting for this change, which we describe in the language of time-series analysis but analyze as varying cross-sections, i.e. across election years. In our analysis, we predict election outcomes form poll results, varying the date of the polls, and trace the variation in the statistical evidence as a function of the time lag between the poll and the election. The data for our analysis include virtually every national presidential poll conducted between 1944 and 1996.
Now we seek to analyze additional poll data, both at the aggregate and individual levels. With respect to the aggregate, we will analyze state-level poll data, for president and other offices, and also national data for Congressional office. With respect to the individual level, we will analyze presidential preference surveys from at least every month of each election year from 1952 through 2000. We ask: How much do campaigns actually influence voters and how long do the effects last? We intend to examine how variables, such as party identification, grow and recede in importance as the campaign evolves.
See Also
- Seed grant: Campaign Dynamics
Funded by National Science Foundation »
