Election Projections – a fun way to waste time

A lot of websites have joined the fun of predicting the 2004 presidential elections based on current state and national poll results, 2000 election results, and whatever else is determined relevant.  Some use simple polls, and some use complex formulas.  Some leave any close states as undecided, and some will predict a state on the basis of a 0.01% lead for one candidate.

Despite the differences, they all pretty much say that the election is much too close to call on June 22.

The Blogging Caesar’s (Scott Elliott) Election Projection probably uses the most complicated math of all the projections I found.  His formula uses the following three adjustments to the poll data: the Baseline Adjustment derived from the 2000 election result, the National Adjustment resulting from a compilation of various national polls, and the State Adjustment calculated from the results of state polls. The overall formula is too complicated for me to explain here, but the results are pretty interesting.  As of today, this methodology predicts an electoral tie, 269-269.  This tie includes projecting Florida for Bush due to a 0.01% adjusted lead there.  Also, the Blogging Caesar predicts an overall popular victory for Kerry of 2.2%

I commend the Scott on how thoroughly he explains his methodology and shows his base data.  There is a lot of interesting data here.  Also, I commend Scott in not letting his strong support for Bush affect his numbers.

Larry Sabato’s Crystal Ball does not have as much data, but he has  nice description of each state.  He predicts every state, no matter how close, but the methodology is not explained.  His current prediction is 274-264 for Kerry, with West Virginia being the advantage for Kerry here compared to the Blogging Caesar’s.

Another State-by-state poll projection has Kerry winning 275-259 (with New Hampshire’s 4 votes unallocated).  This site doesn’t have a pretty map, but it does show the poll numbers used to create the projection.  It has Kerry winning Ohio but losing Wisconsin as the difference with the Caesar.

MyDD :: President 2004 Analysis predicts a surprisingly strong win for Kerry, 301-237 (Florida and West Virginia being the difference with Caesar’s).  There’s a strong pro-Kerry bias throughout the site, but I think the prediction is pretty unbiased.  This site also shows an interesting table showing how various regions of the country have changed over the last few decades in their voting patterns.  For example, it shows explicitly how the Deep South has become more and more Republican, but California has gone from Republican to solidly Democratic.

The Cook Political Report’s Current Electoral College Outlook doesn’t explain its methodology, but it leaves nine states as toss ups: Florida, Iowa, Minnesota, Missouri, Nevada, New Hampshire, New Mexico, Ohio, and Wisconsin.  Outside those states, Kerry has a slight 228-211 lead (Bush has a slight 174-165 lead in solidly-held states).  The 99 Electoral Votes from the 9 undecided states will probably decide the election.

The Current Electoral Vote Predictor shows a large Kerry lead, 302-232 (with New Hampshire’s 4 votes again undecided).  It has Kerry winning Ohio and New Mexico but losing Wisconsin as the difference with the Caesar.  The data is available as a spreadsheet.

The Iowa Electronics Market takes a completely different approach.  It has a fake market, like a stock market, in which people place money on who will win.  This doesn’t predict the electoral college, but right now Bush has a 9% lead here.

The Poll of Polls doesn’t predict electoral votes, but it is a good compilation of other poll results, showing an average of those poll results over time.

In summary, none of the electoral predictors show a Bush victory, but none show an insurmountable lead for Kerry.  All say that the same group of states (Arizona, Missouri, West Virginia, Ohio, Florida, Nevada, New Mexico, New Hampshire, Iowa, Wisconsin, and Oregon will decide the election.

And none of this matters at all, because we still have four full months of campaigning and events to go.

Advertisements

About Lance Finney

Father of two boys, Java developer, Ethical Humanist, and world traveler (when I can sneak it in). Contributor to Grounded Parents.
This entry was posted in Politics. Bookmark the permalink.

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