|
Who will win the election if voters decide based on which candidate they expect to do better in handling the issues? Graefe and Armstrong addressed this question. They found that, historically, voters chose the candidate they expected to do the best job in dealing with the issues facing the country. The approach is based on the assumption that voters select the candidate they expect to perform best in handling the issues. Thus, it is assumed that - For the voter, it is not primarily important how the candidates intend to solve the problems or what policies they promise to pursue.
- Rather, the voters simply want the problems to be solved.
Graefe and Armstrong tested their approach for the 10 U.S. Presidential Elections from 1972 to 2008, analyzing data from 376 historical polls. Nine times their PollyIssues model correctly predicted the winner of the popular vote, with one tie. For the last three elections from 2000 to 2008, the method outperformed well-established election forecasting methods.
Abstract. Ideally,presidential elections should be decided based on how the candidates wouldhandle issues facing the country. If so, knowledge about the voters’ perceptionof the candidates should help to forecast election outcomes. By using the indexmethod, we developed the PollyIssuesmodel, which is based on the voters’ perception of which candidate will do thebest job in handling the issues facing the country. Using a simple heuristic,it correctly picked the winner for nine of the last ten U.S. presidential electionsfrom 1972 to 2008, with one tie. In predicting the two-party vote shares forthe last three elections from 2000 to 2008, on average, the model’sout-of-sample forecasts yielded lower forecasting errors than ten well-establishedregression models. Data files: View online or download as Excel sheet. This work has been presented at the 28th Annual International Symposium on Forecasting, Nice (France), 22-25 June, 2008.
|