What's New on this Site
This page is currently inactive. For older news, see below.
- September 9, 2010 The race to forecast the U.S. midterm elections is on! Click here for Polly's latest prediction.
- September 8, 2010 APSA panel on the U.S. midterm election.The Monkey Cage blog has a summary of the forecasts for the U.S midterm elections presented at a panel of the American Political Science Association held in Washington, D.C., on September 4th. Also included are several of the power point slide shows or papers presented at the conference.
- February 11, 2008 Polly's new page for the 2008 election is online and will be cared for by Andreas Graefe. He is researcher at the Institute for Technology Assessment and Systems Analysis in Karlsruhe, Germany, currently on leave at the Wharton School. Visit: www.pollyvote.com
- January 19, 2008.Erikson and Wlezien apply damping to achieve substantial improvements in the accuracy of pre-election polls. â€œDampedâ€ polls provide more accurate forecasts than the Iowa Electronic Markets for the five U.S. presidential elections from 1988 through 2004. This "E&W poll damping" will be used in the construction of the Pollyvote for the 2008 election.
Traditionally, raw pre-election polls are interpreted as forecasts of election outcomes. However, these polls register preferences on the day of the poll but do not provide predictions of the election outcome because many things might change prior to the election. In contrast, prediction markets try to account for the changes that might occur. By comparing raw polls with prices of the Iowa Electronic Markets (IEM) for the five US presidential elections from 1988 through 2004, Berg et al. (2007) found that the IEM outperformed polls 74% of the time.
Erikson and Wlezien provide new evidence on using opinion polls to predict the outcome of elections. Longer-term forecasts involved more uncertainty, and uncertainty calls for more conservative forecasts. Thus, for a two-party vote, forecasts should be damped towards 50-50 -- and the longer the forecast horizon, the more damping needed. Erikson and Wlezien showed that damped polls were much more accurate than the original polls in forecasting the five US presidential elections from 1988 through 2004. Moreover, the damped polls were shown to be more accurate than IEM vote-share market prices and especially those from winner-take-all markets.
Reviewed by Andreas Graefe, January 19, 2008.Erikson R. S. and Wlezien C. (2007): â€œAre Political Markets Really Superior to Polls as Election Predictors?â€, Public Opinion Quarterly, forthcoming.
Further discussion of this paper and the question whether damped polls outperform prediction markets or vice versa can be found here.
- January 9, 2008 The second survey of Pollyâ€™s Experts Panel, completed before the Iowa Caucuses, generated the following nomination probabilities: Clinton (45%), Obama (37%) and Edwards (15%). On the Republican side, they are as follows: Giuliani and Romney, 25% each, Huckabee 20%, and McCain and Thompson 10% each. more ...
- December 13, 2007. At the APSA meeting which met in Chicago earlier last year, the Political Forecasting Group adopted a set of bylaws and elected its first cohort of officers. For more information on bylaws, click here. For information on Officers and At-Large Council Members, elected 31st August 2007, at APSA meeting Chicago, click here.
- October 24, 2007.The latest version of the paper "Predicting Elections From Politicians â€™ Faces" by Armstrong, Green, Jones, and Wright - Full Text.
- October 15, 2007.POLLYPRIZE 2008. IIF will make a $1,000 award to the author(s) of a model that best predicts the outcome of the 2008 American presidential election. more ...
- October 9, 2007. At a panel of the recently held annual meeting of the American Political Science Association in Chicago, a panel on Methods of Long-Range Election Forecasting included several papers with early forecasts of the outcome of the 2008 American presidential election. To access the papers, click here.
- July 7, 2007. Panels on Presidential Election Forecasting at the ISF 2007 International Symposium on Forecasting. A stream of three panels on U.S. presidential elections was included in the 2007 International Society of Forecasters â€œInternational Symposium on Forecasting,â€ held in New York City last June. To access the papers, click here.
- February 22, 2007. â€œRay Fair vs. James E. Campbellâ€. In a series of papers and e-mail communication, Professors Ray Fair (Economics, Yale) and James E. Campbell (Political Science, SUNY at Buffalo) recently discussed the formerâ€™s most recent paper, â€œPresidential and Congressional Vote-Share Equations.â€ They have kindly given Polly permission to post these exchanges, which are available by clicking here.
- January 14, 2007. Announcement from Michael Lewis-Beck, Associate Editor, IJF:
The INTERNATIONAL JOURNAL OF FORECASTING has taken a special interest in political forecasting, and in aid of that has appointed me as Associate Editor. I have been asked to organize two sessions on US presidential election forecasting, for the INTERNATIONAL SYMPOSIUM ON FORECASTING meetings, to be held in NYC, June, 2007. Please send me a proposal. The intention would be to PUBLISH the six papers selected in the April 2008 issue of the journal, assuming they met usual editorial conditions.
Dear Fellow Election Forecasters,
We are not seeking any particular methodological approach. In fact, all approaches are worthy of consideration, and a mix would make it more lively.
Officially, paper abstracts must be submitted by March 2, 2007, and official announcements of acceptance made by April 6, 2007.
Here is the link to the meeting: 27th Annual International Symposium on Forecasting.
This is a great opportunity to gain attention for our forecasting enterprise in a vital scientific community. Please let me know if you have any questions.
Mike Lewis-Beck Associate Editor, IJF
- January 3, 2007. Carl E. Klarner and Stan Buchanan won the Pollyprize competition. You can access the Panel of Judges report here, their paper here, and their data set here.
- 5 November, 2006.Will the Republicans retain their majority in the U.S. House of Representatives this year? Polly says...
- October 27, 2006. The Political Forecasting Group has won recognition by the American Political Science Association! For details, please click here.
- October 15, 2006. Will the Republicans retain their majority in the U.S. House of Representatives this year? Find out what Polly says.
- September 2, 2006. The new color scheme emphasizes the close relationship with of this page with the International Institute of Forecasters. The red also provides a high contrast with white text for the left-hand menu, which aids readability. The IIF logo has replaced our old logo. (As before, when you click on the logo, it will take you to the forecastingprinciples.com home page.) The addition of the third column makes the news more accessible as well as keeping the pressure on us to report the news. (You can help. Please send News items to Alfred CuzÃ¡n.) The new three-column layout also aids readability, as previously the lines of text were longer than studies suggest is optimal. We changed from Times New Roman to Georgia for text and to Verdana for subheadings and menu items because studies have shown that these fonts are a bit easier to read on screen. Georgia also does well in hard-print readability tests. Finally, technical changes have allowed for faster downloads. Please send suggestions that you might have about further changes.
- July 17, 2006. Recent research by Justin Wolfers and other scholars suggests that betting markets outperform the polls as election predictors. But in a recent paper, Robert S. Erikson and Christopher Wlezien conclude just the opposite. For a copy of the paper, click here
- April 18, 2006. Polly is happy to announce that on behalf of the International Institute of Forecasters, it will make a $1,000 award to the designer(s) of the best model for forecasting the results of this year's U.S. House of Representatives elections. For details on the competition, please click here.
- March 22, 2006. For â€œNews in Political Forecasting,â€ please click here. You will read about a panel on political forecasting that will be on the program of the American Political Science Association, which this year will meet in Philadelphia during the Labor Day weekend, as well as about other developments in political forecasting.
- February 9, 2006. Ben Franklin were alive today, how would he predict the next presidential election? He might suggest an index model that would work like this: count the factors favoring each candidate and the one with the most will win. In fact, using an index model called "The 13 Keys to the White House," American University professor Allan Lichtmann has correctly predicted all American presidential elections since 1860. In the current issue of Foresight, Scott Armstrong and Alfred CuzÃ¡n examine how this model performs in comparison with several other models. To access the article, click here.
- September 20, 2005. Alfred G. CuzÃ¡n and Charles M. Bundrick invite comment and criticism of a draft of â€œFiscal Policy: An Implicit Factor in the Trial-Heat and Time-for-Change Models?â€ For a copy of the paper, click here.
- August 12, 2005.The data for an updated, revised, and expanded estimate of James E. Campbell's state-level presidential vote model is posted on Polly's Data Page. There can also be found data sets for Erikson and Wlezien's "leading indicators" and Lewis-Beck and Tien "jobs model" presidential elections forecasting models.
- June 6, 2005. Text is now available for hree panels on presidential election forecasting included in the program of the 2005 International Society of Forecasters â€œInternational Symposium on Forecasting,â€ which will be held June 12-15 in San Antonio. One is a Roundtable on Ray Fairâ€™s Presidential Vote Equation, where Prof. Fair will respond to critiques of his model. Professors James E. Campbell, Richard Gleisner, and Joe Stone are among the presenters. For summaries of the presentations, click here. At two other sessions, several forecasting models that did relatively well in 2004 will be displayed, along with other papers. For papers by Professors Campbell, Ericson and Wlezien, Jones, Norpoth, and others, click here.
- March 24, 2005. The British Journal of Politics and International Relations is about to publish a collection of articles on election forecasting. Five articles apply different models to the forthcoming British election and a sixth casts a skeptical eye on the entire enterprise. Courtesy of that journal, you can read the abstracts by clicking here.
- January 6, 2005. Alfred CuzÃ¡n, Scott Armstrong and Randall Jones presented a paper on Polly at the Southern Political Science Association, New Orleans. To download â€œCombining Methods to Forecast the 2004 Presidential Election,â€ click here.
- January 2, 2005. Professor Abramowitz does a post-mortem on his 2004 presidential election forecast. His took the median spot out of the 10 included in Pollyâ€™s Table 1. To read this paper, click here.
- November 30, 2004. Professor Campbell does a post-mortem on his two forecasts for this yearâ€™s presidential election. One, which put Bushâ€™s share of the two-party vote at 52.8%, was the fourth most accurate out of the 10 included in Pollyâ€™s Table 1, after Wlezien and Erikson's, CuzÃ¡n and Bundrick's, and Lewis-Beck and Tien. To read his paper, click here.
- November 24, 2004. Professors Lewis-Beck and Tien do a post-mortem on their jobs model forecast for 2004. Their September forecast that Bush would take 50.2 percent of the two-party vote was off by only 1.3 percent. This was the third most accurate forecast of the ten included in Polly's Table 1 (which shows their August forecast), after those by Wlezien and Erikson's and CuzÃ¡n and Bundrick's. To read their paper, click here. For their data file, click here.
- November 22, 2004. Professors Wlezien and Erikson offer observations about the 2004 election and their forecast thereof, which came within less than half a percent point of the actual result. This was the most accurate prediction obtained with any of the forecasting models included in Polly's Table 1. Read more . . .
- November 18, 2004. We have made a major revision of the site, following the completion of the 2004 election. All materials relating to the Pollyvote of 2004 are now available fromPolly's Page or the Archives. Please note our plans for the future of this site on Polly's Home Page.
- November 5, 2004. At politicalforecasting.com we believe it is good practice to make it as easy as possible for scholars to replicate each other's work and mine a common body of machine-readable data. Accordingly, today we're introducing a Data page. It will serve as a clearinghouse for forecasters of elections and other political events to make their data available to fellow scholars. We begin by posting Data Set 1, a Microsoft Excel file which contains the values of the Pollyvote and its components, March 8 - November 1, 2004.
- October 29, 2004. Ray Fair issued a revised forecast based on the NIPA data released today. The actual values of Fair's three economic variables are as follows: GROWTH = 2.9 percent, INFLATION = 2.0 percent, and GOODNEWS = 2. Together they represent a minimal improvement in economic performance for the incumbents over the July economic forecast. When entered into Fair's presidential vote equation the result is a slight upward revision in Bush's share of the two-party vote, from 57.48 to 57.70 percent, or about what Eisenhower received in 1956. The fiscal model used by CuzÃ¡n and Bundrick incorporates Fair's two measures of economic growth (but not inflation). Accordingly, their revised forecast is also up trivially, from 51.1 to 51.2 percent for Bush. The principal reason for the disparity between Fair's forecast and that obtained with the fiscal model is that, unlike the former, the latter takes into account the negative effect on the vote produced by Bush's expansionary fiscal policy. See "Fiscal Policy: A Missing Factor in Ray Fair's Presidential Elections Forecasting Model?
- October 29, 2004. On second thought, we have removed Chappell et. al.'s forecast from the Quantitative Models component in the Pollyvote. Resting entirely on state-level polls, theirs is not a causal model. Like Andrea Moro 's, it is designed to calculate the probability that one or the other candidate will win a majority in the Electoral College.
- October 21, 2004. A new quantitative model forecast of the two-party vote going to President Bush issued by Chappell et. al. has been factored into the Pollyvote. For a brief description of the model, click here. To read a working paper describing their forecasting procedure, click here.
- September 7, 2004. Two more quantitative model forecasts have been factored into the Pollyvote, one each by Thomas Holbrook (University of Wisconsin, Milwakee) and Brad Lockerbie (University of Georgia). Their forecasts, along with others previously reported on this page by Alan Abramowitz(see August 12 entry) and Helmut Norpoth(May 13), as well as a revised one by Lewis-Beck and Tien (July 21), were presented at the American Political Science Association's Roundtable, "Forecasting the 2004 Presidential Election," Chicago, September 2. The Roundtable was chaired by James Campbell (University of Buffalo, SUNY), whose own forecast will be issued later this month.
- August 26, 2004. CuzÃ¡n and Bundrick compare Abramowitz's time-for-change model with the fiscal model and show that borrowing a measure of fiscal policy from the latter and adding it to the former improves the original.
- August 24, 2004. You can use Campbellâ€™s â€œtrial heatâ€ model to make your own prediction for this yearâ€™s presidential election. Just enter your guesstimates of President Bushâ€™s share of the two-party vote in the Gallup Poll published after Labor Day and the rate of GDP growth in the second quarter of the election year into the Campbell â€œtrial heatâ€ model election calculator. The result will be your very own forecast for November.
- August 12, 2004. Using a revised version of his time-for-change model, Alan I. Abramowitz issues a forecast for November, but expresses doubt about its accuracy. According to the model, Bush should take 53.7 percent of the two-party vote. But Prof. Abramowitz cautions that an electorate polarized over the war in Iraq may deal his model "its first erroneous prediction." See full text.
- August 2, 2004. The first Delphi survey of a distinguished panel of American politics experts has been concluded. The panel forecasts that Kerry will take 50.5 percent of the two-party vote (i.e., ignoring third-party shares). This means that although the panel believes that Kerry has a slight advantage over Bush, at this point the election is too close to call. We plan to administer at least two more surveys of this panel of experts before the election.
- August 1, 2004. Ray Fair has updated his forecast, as have CuzÃ¡n and Bundrick.
- July 29, 2004.Douglas Hibbs >uses his "Bread and Peace" model of American presidential elections to make a forecast for November, anticipating a Bush victory with about 53 percent the two-party vote. His forecast has been incorporated into the Pollyvote in Polly's Table 1. See the full text of the paper.
- July 21, 2004. Two new quantitative model forecasts have been factored into the Pollyvote, one by Christopher Wlezien and Robert S. Erikson, (â€œForecasting the Presidential Vote, 2004, the other by Michael S. Lewis-Beck and Charles Tien (â€œThe Jobs Model: Presidential Forecast for 2004â€). For their respective forecasts of previous elections, see our bibliography.
- May 13, 2004.Helmut Norpothforecasts a Bush victory against Kerry by 55 to 45 percent of the two-party vote. This forecast, first issued on January 29 of this year and reported in a paper presented at the Midwest Political Science Association last April, has been incorporated into the Pollyvote in Polly's Table 1.
- May 10, 2004. We are introducing a number of changes in our site. Polly's Page now displays a revised Table 1 as well as a new feature, Pollygraph 1, a visual representation of the Pollyvote, the Bush|Kerry Quotes in the Iowa Electronic Markets, and Bush's share of the two-party vote since March 8th. Polly's Menu has two new links, one to Polly's Tables and the other to the Pollygraphs.
- May 6, 2004. A new Pollygraph page has been created. Here the viewer will find graphs tracking presidential election variables. See Pollygraph 1.
- May 4, 2004. Forecasts for two quantitative models have been updated.
- April 23, 2004. History Prof. Allan Lichtman examines the direction of his The Keys to the White House for 2004, and forecasts President Bush's reelection.
- April 21,2004. â€œPollyâ€ is our site mascot and will add an interesting perspective (and vocabulary!) to our analyses of the upcoming election. The â€œPolly Sezâ€ page defines some of her favorite phrases and also gives readers the opportunity to contribute their own â€œPollyisms.â€ The "Pollyism of the Week" is also now available from the menu on Polly's Home Page.
- April 19, 2004. We're adding a new source of forecasts to Polly's Table. This is the mean of the average daily price of the Bush|Kerry contracts traded in the Iowa Electronic Markets(IEM) since March 8th. IEM is the creation of and operated by the faculty of the University of Iowa Business School for teaching and research purposes â€“ and as a forecasting tool. At IEM, Bush|Kerry contracts, that is, contracts on Bush winning against Kerry, will be paid off according to the percent of the two-party vote Bush wins in November. Thus, this market mimics the actual election many months before it is held. Polly will report the very latest latest quote (as of the date that Polly's Table 1 is updated with new polls) as well as the rolling average over the week prior to the day the latest polls is reported. This is approximately the period during which voters were being surveyed.