- Political forecasting
Political
forecasting aims at predicting the outcome of elections. Models include:Opinion polls
Polls are an integral part of political forecasting. However, incorporating poll results into political forecasting models can cause problems in predicting the outcome of elections. There are a few ways in which inaccurate election forecasts can be avoided.Averaging polls
It is well documented that combining poll data reduces the forecasting error of polls. [Alfred G. Cuzan, J. Scott Armstrong, and Randall Jones, [http://www.forecastingprinciples.com/PollyVote/index.php/pollys-archive/polly-2004.html "Combining Methods to Forecast the 2004 Presidential Election: The PollyVote"] ] Political forecasting models often include averaged poll results, such as the
RealClearPolitics poll average.Poll damping
Poll Damping is a method of discounting flawed indicators of public opinion when creating a forecasting model. Often, polls early in the campaign are poor indicators of the future choices of voters. By weighting the poll results from days closer to an election to a greater degree than previous polls, a more accurate prediction is established. [http://www.forecastingprinciples.com/PollyVote/index.php/who-is-polly/pollycomp-rcp.html#Campbell%20(1996) Campbell (1996)] [Campbell J. E. (1996), "Polls and Votes: The Trial-Heat Presidential Election Forecasting Model, Certainty, and Political Campaigns," American Politics Quarterly, 24 (4), pp.408-433.] demonstrates the power of poll damping in political forecasting.
Markets
Prediction Market s like theIowa Electronic Markets provide highly accurate forecasts of election outcomes. Comparing market forecasts with 964 election polls for the five US presidential elections from 1988 to 2004, [http://www.biz.uiowa.edu/faculty/trietz/papers/long%20run%20accuracy.pdf Berg et al. (2008)] showed that the IEM outperformed the polls 74% of the time. [http://www.biz.uiowa.edu/faculty/trietz/papers/long%20run%20accuracy.pdf] However, damped polls have been shown to outperform prediction markets. Comparing damped polls to forecasts of the Iowa Electronic Markets, [http://www.forecastingprinciples.com/Political/PDFs/EriksonWlezien_Markets_AAPOR_for_POLLY.pdf Erikson and Wlezien (2008)] showed that the damped polls outperformed both the winner-take-all and the vote-share markets.Regression models
Political scientists and economists have employed regression models of past elections to forecast the percent of the two-party vote going to the incumbent party candidate in the next election. Most models consist of between two and seven variables and are estimated over anywhere between scarcely over a dozen elections to close to twice as many. (By contrast, historian [http://www.forecastingprinciples.com/Political/PDFs/Keys_forecast_aug_2007_apsa_by_lichtman.pdf Alan Lichtman] uses 13 “Keys” to predict whether the incumbents will be reelected.) A common denominator across most quantitative models is at least one measure of economic conditions, although no two employ the same metrics. Also, most models include at least one public opinion variable, a trial heat poll or a presidential approval rating, although here again there is no unanimity on indicators.
Pollyvote
PollyVote is a political forecasting model which predicts The United States presidential elections.References
Wikimedia Foundation. 2010.
