2012 popular vote: forecasts and predictions vs. results

Here’s what I posted up on Monday about the various election forecasts and predictions:

  • Political science forecasting models. October’s issue of PS: Political Science and Politics contained nine statistical forecast models (not including state-level models) of President Obama’s two-party vote total for the 2012 election, made anywhere from 2-10 months before the election. They ranged from a low of 46.9% to a high of 53.8%. The mean of all nine was 50.2%.
  • As of 8:00 PM on Election Eve, Nate Silver at FiveThirtyEight is predicting Obama will receive 51.4% of the two-party popular vote.
  • As of 9:30 PM on Election Eve, Simon Jackson’s model at Pollster.com is giving Obama 50.6% of the two-party popular vote.

The final popular vote tally, as of noon on Wednesday: 51.1%.

Not a bad showing for the statistics geeks, all around.

I should note that the political science forecasting models include only a handful of “election fundamental” variables in their predictions (economic growth, incumbency, foreign conflict, presidential approval, e.g.). These models do not include things like who the candidates are, the conventions, the debates, the campaigns, etc. From a political science perspective, the difference between what the fundamentals would predict and the actual outcome is what can likely be attributed to the campaign events and the candidates. In this case, it’s looking like the campaign ultimately made a difference of… about 0.9% overall (51.1% final outcome minus 50.2% average forecast model prediction). 

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s