|
PollyBio: Predicting Elections From Candidates' Biographies
|
Scott Armstrong and Andreas Graefe developed a bio-index model to predict the winner of U.S. Presidential Elections based on candidates’ biographies. Their model, named PollyBio,
uses 59 biographical cues that are expected to have an impact on the
chances of a candidate on being elected. Background
There is a large stream of research in the field of psychology that analyzes questions such as what makes people emerge as leaders? For example, meta-analyses found intelligence and height to have a positive impact on both leader performance and leader emergence. Such findings from prior research were used to identify and code the majority of variables in the bio-index. In addition, for some variables, the authors used common sense. For example, it was assumed that a candidates are more attractive to voters if they are married but not divorced.
Coding of 2012 Republican nominees
and historical candidates
from 1896 to 2008
(on Yes/No variables only)
|
|
In general, the model distinguishes two types of variables: - Yes / No variables (n=48): For this type of variable, candidates are assigned a score of 1 if they possess a certain attribute and 0 otherwise.
For example, a candidate
receives positive scores if he descends from a political family, is
first-born, single child, lost a parent in childhood, is married but
not divorced, has children, graduated from a prestigious college, held
political offices, has authored a book, or has military experience.
- Comparative variables (n=11): For this type of variable, the candidates of the two major parties are compared on the underlying attribute. The candidate who scores better than his opponent is assigned a score of 1 and 0 otherwiese.
For example, candidates who are taller, heavier, better-looking, more intelligent or have the more common name than their opponent receive a positive score while their opponent received a score of 0 for these variables.
After all variables have been coded, the total index scores for each candidate are calculated. Then, the candidate who achieves the higher
overall score is predicted as the election winner. Are you fit to be president?
Compare your biography to Obama's
and test your chances of winning the
2012 U.S. Presidential Election - or
find out whether your favorite politician
would have a chance. Click here. | |
Decision-making implications
In
using biographical information about candidates for generating the forecast, the model can be
useful to political decision-makers. The
bio-index model advises the challenging party (i.e., the Republicans)
to nominate the candidate with the highest index score relative to the
index score of the incumbent president (i.e., Obama). In particular, the model can
advise decisions such as - whether a candidate should run for office or
- which candidate a party should nominate.
Past performance
Armstrong and Graefe tested their model for the 29 U.S. Presidential
Elections from 1896 to 2008. The model failed only two times. For the remaining 27 elections,
the model correctly predicted the winner. The model wrongly predicted Ford to beat Carter in 1976 as well as Bush to defeat Clinton in 1992. This record of 93% correct
predictions compares favorably to other statistical models as well as
to polls and prediction markets. The table on the right shows the total index scores of the historical candidates on Yes/No variables for the 29 elections from 1896 to 2008 (as well as the index scores of the potential 2012 contenders). As can be seen, the election winners (shaded grey) generally tend to achieve higher scores on Yes/no variables than their opponent. Exceptions are Geroge Bush sen. and Gerald Ford who lost the election despite a very high bio-index. At the same time, candidates can sometimes make up for poor scoring on Yes/no variables if they beat their opponent on the comparative variables. This happened for Reagan in both the 1980 and the 1984 election. Forecast for 2012
With the bio-index, forecasts of the election outcome can be made as
soon as the candidates are known; they can be issued even before,
conditional on who is expected to be in the race. Thus, the model can
help the Republican Party to decide who they should nominate to run
against President Obama in 2012. The table on the right shows the total index scores on Yes/no variables (as far as information could be obtained) for potential Republican nominees (shaded red). As can be seen, the candidates that have been coded thus far obtained rather poor index scores. An exception is General David Petraeus whose index score is on a par with former presidents such as Clinton, Kennedy, Franklin D. Roosevelt, and... incubment president Obama. By comparison, with a score of 11, Ron Paul performs rather weak on the bio-index. For the final forecast it is necessary to also consider the candidates' scorings on the comparative variables, relative to incumbent president Obama. For each candidate, the table below shows the difference in index scores that was achieved compared to Obama. Accrding to the final bio-index score, Rick Perry scores highest with a total index score that is one point higher than Obama. Bio-index forecast for potential
2012 Republican nominees
|
| * Forecast used in the combined PollyVote.
(You can click on the links in the table to see the codings for each of the potential Republican nominees.) |
Bio-index forecast in the PollyVote
Until it is clear who will be the Republican nominee, we incorporate the forecast for the candidate that is currently seen as most likely to win the primaries in the combined PollyVote forecast. For this, we consult the PollyPrimaries forecast. The asterisk (*) in the table on the left shows who PollyPrimaries currently predicts to win the Republican nomination. The calculation of the bio-index forecast is done based on the vote equation that was
estimated from the results from the 29 elections from 1896 to 2008.
|
|
Written by Andreas Graefe
|
|
Saturday, 01 October 2011 05:17 |
|
Polly has updated the forecast from the bio-index method with new information on the facial competence of candidates. Perceived facial competence of candidates is one example of a biographical variable that prior research has found to have value in predicting election outcomes. Todorov, Mandisodza, Goren, and Hall (2005) presented 31 subjects with pictures of candidates running in U.S. House and Senate elections. Based on one-second exposures, the subjects rated each candidate’s competence. Subjects who recognized a candidate were excluded. For the three Senate elections from 2000 to 2004, the most competent-looking candidates won 71% of the 95 races. For the two House elections in 2002 and 2004, the most competent-looking candidates won 67% of the 600 races in their sample. In a similar study, Antonakis and Dalgas (2009) asked 684 university students and 2,814 children in Switzerland to rate pairs of black and white photos of candidates’s faces in the 2002 French parliamentary election. In both samples, the candidates who achieved higher ratings on facial competence won in 72% of the elections. Similarly, Armstrong, Green, Jones, and Wright (2010) found facial competence to be predictive for the outcome of the 2008 U.S. presidential primaries. In this study, university students from the US, Australia, and high school girls from New Zealand rated facial competence of 11 Democratic and 13 Republican candidates on a 10-point scale from 0 ("Completely incompetent") to 10 ("extremely competent"). Of the 11 Democratic candidates, Hillary Clinton ranked highest while Obama came in 3rd (with an average score of 6.8). Of the 13 Republican candidates, John McCain achieved the highest ratings (with an average score of 6.2). By comparison, Mitt Romney achieved an average score of 6.1 (0.7 points lower than Obama). Polly used the findings from this study to update the bio-index forecast. The latest forecast according to this model is that Obama would achieve 54.5% of the popular vote if running against Romney. |
|
|
Written by Andreas Graefe
|
|
Thursday, 18 August 2011 02:07 |
|
Here is the abstract and link to a new research paper that will be presented in a panel on Presidential Election Forecasting at the 2011 Annual Meeting of the American Political Science Association, Seattle, September 1-4. The paper presents forecasts of the 2012 U.S. Presidential Elections based on the bio-index model, which predicts the election outcome based on candidates' biographies. According to this model, Rick Perry has the highest chance of all Republican candidates to defeat Obama in a potential 2012 showdown.
Abstract: The state of election forecasting has progressed to the point where it is possible to develop highly accurate forecasts for major elections. However, one area that has received little attention is the use of forecasting as an aid to those involved with political campaigns. In the run-up to the presidential primaries, we use the bio-index model to test the chances of potential nominees to defeat President Obama in the 2012 U.S. presidential election. This model uses the index method to incorporate 58 biographical variables (e.g., age, marital status, height, appearance) for making a conditional forecast of the incumbent’s vote-share, depending on who is the opposing candidate. These variables were selected based on received wisdom and findings from prior research. For example, several studies found candidates’ perceived attractiveness or facial competence to be related to their chances of winning an election. The model is particularly valuable for making long-term forecasts of who will win an election and missed the correct winner only twice for the 29 elections from 1896 to 2008. Thus, the model can help candidates to decide whether they should run for office and can advise political interests in deciding whom to support in the primaries and caucuses. The forecasts from the bio-index model are compared to forecasts from polls and prediction markets. Reference: Graefe, Andreas and Armstrong, J. Scott, Who Should Be Nominated to Run in the 2012 Presidential Election? Long-Term Forecasts Based on Candidates' Biographies (2011). APSA 2011 Annual Meeting Paper. Available at SSRN: http://ssrn.com/abstract=1902852
|
|
Written by Andreas Graefe
|
|
Saturday, 30 July 2011 08:53 |
|
 Polly coded the biographical index for Obama and Herman Cain. According to the bio-index model, Cain has little chance to beat Obama in a possible 2012 showdown. While Cain achieves 15 points on the bio-index, Obama scores 21 points. This difference in index scores translates to a forecast of the popular vote-share of 55.9% for Obama and 44.1% for Cain. Thus, Cain is the weakest Republican candidate that Polly coded thus far on the bio-index.
The table below shows the respective codings for Obama and Cain. If you spot any mistakes in the coding, please let us know. And remember, you can use the Are you fit to be president? quiz to test the chances of your favorite candidate. |
|
Written by Andreas Graefe
|
|
Friday, 22 July 2011 01:04 |
|
Texas Governor Rick Perry has still not officialy annouced to run. Nevertheless, Rick Perry is now seen as most likely to win the Republican nomination at the intrade.com prediction market, although with a very small margin. Currently, the market gives Perry about 31% chance of winning the primaries, which is slightly higher than the market's forecast for Romney (30%).
This development has led Polly to update the bio-index forecast that is included in the combined PollyVote forecast. Each day, Polly calculates the bio-index forecast assuming that the candidate who is leading in the intrade market would win the Republican nomination. Until today, Romney was ahead of the field. Now, for the first time, Rick Perry has taken the lead. Thus, the bio-index forecast that goes into the PollyVote is calculated assuming that Rick Perry will win the nomination. As a result, the new bio-index forecast is that Obama would receive 49.7% of the popular vote. Inevitably, this has an effect of the Pollyvote forecast, which dropped to 51.0% for Obama. This forecast is another all-time low for Obama since the launch of the Pollyvote in January 2011. |
|
Written by Andreas Graefe
|
|
Wednesday, 20 July 2011 13:31 |
|
An article about the bio-index model was the most popular posting on the Houston Chronicle's Texas on the Potomac blog in the week ending July 9. Click here for the list of the top ten postings in that week. |
|
Written by Andreas Graefe
|
|
Tuesday, 12 July 2011 02:26 |
|
 Polly calculated the bio-index scores for a potential 2012 showdown between Obama and Rick Santorum, former U.S. Senator from Pennsylvania.
Santorum scores well on the bio-index. With a total score of 18, Santorum is only two points short of Obama, who achieved an index score of 20. In terms of vote-shares, the bio-index model predicts the following outcomes: Obama 52.2% vs. Santorum 47.8%. The good news for Santorum is that, among all Republican candidates that Polly coded and that have officialy announced to run, he performs strongest compared to Obama (see here for an overview of the scorings of all candidates thus far).The bad news is that nevertheless the bio-index model predicts him to lose against Obama. The table below shows the codings for the two candidates Obama and Petraeus. If you spot any mistakes in the coding, please let us know. And remember, you can use the Are you fit to be president? quiz to test the chances of your favorite candidate. |
|
Written by Andreas Graefe
|
|
Monday, 11 July 2011 04:04 |
|
 Polly calculated the bio-index scores for Jon Huntsman, former Governor of Utah and U.S. Ambassador to China. Given the information that is available to date, Huntsman achieves an index score of 17 points, which is five point lower than Obama's (22 points). This difference in index scores translates to the following vote-shares predicted by the bio-index model: Obama 54.7% vs. Huntsman 45.3%.
Thus, the bio-index model forecast predicts a clear victory of Obama. This forecast is in line with other indicators such as the latest RCP poll average, which shows that 50% of the voters would support Obama while only 36% would support Huntsman in a potential 2012 showdown. Similarily,the intrade prediction market sees only a 3.5% chance for Huntsman to win the 2012 U.S. Presidential Election. The table below shows the codings for the two candidates Obama and Petraeus. If you spot any mistakes in the coding, please let us know. And remember, you can use the Are you fit to be president? quiz to test the chances of your favorite candidate. |
|
Written by Andreas Graefe
|
|
Friday, 08 July 2011 08:21 |
|
 Polly calculated the relative index scores of Obama and Mitt Romney for the bio-index model. According to the bio-index model, Romney would have little chance of defeating Obama in 2012. Romney achieved a bio-index score of 18 points, which is four points lower than Obama's. This translates to a popular vote-share of 53.8% for Obama (vs. 46.3% for Romney).
If one looks only at the 49 Yes/No variables in the model, Romney achieves 14 points. In the last 29 elections from 1896 to 2008, only one candidate with a score that low was able to win the election: Ronald Reagan in 1980. For more information on the bio-index model and codings of other potential Republican nominees click here. The table below shows the codings for the two candidates Obama and Petraeus. If you spot any mistakes in the coding, please let us know. And remember, you can use the Are you fit to be president? quiz to test the chances of your favorite candidate. |
|
Written by Andreas Graefe
|
|
Thursday, 07 July 2011 03:50 |
|
 Polly calculated the bio-index scores for a potential 2012 showdown
between Obama and General David Petraeus. According to the bio-index, Petraeus would have a fair chance of winning the 2012 election: he achieved an index score equal to Obama's.
However, since the bio-index model incorporated an advantage for the incumbent in the case of a tie between the candidates, the model would still predict Obama to win 50.5% of the popular vote. That said, General Petraeus would certainly be within striking distance. In fact, he is the second strongest potential Republican candidate based on the bio-index - after Texas Governor Rick Perry. But thus far, neither Perry nor Petraeus are part of the race. Petraeus actually ruled out to run. The table below shows the codings for the two candidates Obama and Petraeus. If you spot any mistakes in the coding, please let us know. And remember, you can use the Are you fit to be president? quiz to test the chances of your favorite candidate. |
|
Written by Andreas Graefe
|
|
Thursday, 07 July 2011 03:44 |
Polly was provided with evidence that led her to code Texas Governor Rick Perry as athletic, which makes him gain an additional point on the index score. At the same time, Polly now questions that Perry is taller than Obama. If you know the answer to that question, please let us know. In sum, the bio-index scores for Obama and Perry remain unchanged. That is, Perry achieves an index score that is one point higher than Obama's.
|
|
Written by Andreas Graefe
|
|
Tuesday, 05 July 2011 02:21 |
 Polly calculated the bio-index scores for a potential 2012 showdown between Obama and Ron Paul. With a total index score that is six points lower than Obama's, the bio-index does not give much chance to Ron Paul for winning the election. Using the vote equation that was derived from historical data for the 29 elections from 1896 to 2008, the bio-index model predicts a popular two-party vote-share of 55.6% for Obama (vs. 44.4% for Ron Paul). Similarily, the intrade prediction markets sees only a 2% chance for Ron Paul to win the 2012 U.S. Presidential Election. |
|
|
|
|
|
|
Page 1 of 2 |
|
PollyVote 2012
Today's forecast of the
popular two-party vote
|
| Click here for more information about this forecast.
|
Fit to be president?
Compare your biography to Obama's and see your chances of winning the 2012 U.S. Presidential Election - or find out whether your favorite politician would have a chance. The test is based on the PollyBio model, which correctly predicted 27 of the last 29 U.S. Presidential Election winners. Click here to access the test.
|
