Predicting The Electoral Vote
David Moore, author of The Opinion Makers: An Insider Reveals the Truth Behind the Polls, is a former Vice President of the Gallup Organization and Managing Editor of the Gallup Poll. He has been guest blogging here all week.
PREDICTING THE ELECTORAL VOTE....As we all know, the national polls don't predict the presidential winner. George W. Bush lost the popular vote in 2000 but won the electoral vote. So, while the national polls are interesting, in a close race we know that it's the electoral vote predictions that give us the better insight.I've looked over the web, going to places I know, but also googling "electoral map 2008 president" to find sites I wasn't so familiar with. I've chosen a few for comparison, mostly the media sites, though two others that appear to be quite popular.
As one who regularly blogs on pollster.com, naturally I looked there first, but I found many others I do not regularly visit with very useful information. Here are the ones whose data are included in the table below (they are listed in order of the "overall Obama lead" — with Pollster showing the largest overall lead and NPR showing the smallest):
- Pollster.com (Pollster)
- Los Angeles Times (LAT)
- RealClear Politics.com (RCP)
- New York Times (NYT)
- CNN
- NBC
- National Public Radio (NPR)
The top three results favoring Obama were all produced by sites that make projections based on statewide polls. (The Los Angeles Times notes that there is no polling for Washington, D.C., but that its three electoral votes are counted for Obama.)
The other four media sites showing the smaller overall leads apparently base their projections on their in-house experts. NPR is explicit about that, while the NYT, CNN and NBC sites imply it (unless, of course, I've missed some important detail). It would appear that the media experts are especially cautious, suggesting a closer race than the more empirically-oriented sites.
Some of the most interesting differences and similarities among Pollster.com, the Los Angeles Times, and RealClearPolitics.com:
Minnesota — Pollster and LAT say this is a solid Obama state, while RCP says it's a toss-up.
Ohio and Michigan — seen as toss-ups by RCP, while Pollster has both leaning toward Obama, and LAT has Michigan for Obama, with Ohio as a toss-up.
New Mexico — leaning Obama say LAT and Pollster; a toss-up says RCP
Florida, Missouri, Virginia and Colorado — all three sites say these states are toss-ups.
There are, of course, other differences worth examining. It's worthwhile to visit the sites.
One of the most interesting charts I've seen is provided by LAT, which shows the electoral vote calculationsover time. Through most of the primary season, while Obama and Hillary Clinton were still contesting the nomination, McCain was consistently ahead in the electoral vote total. Toward the end of May, Obama caught up, and in mid-July surged to his highest LAT lead, before falling back somewhat.
The fluctuation in the electoral vote totals — from a 65 vote lead by McCain up to a 108 vote lead by Obama, down to the current Obama lead of 77 votes — should warn us that many voters have yet to make up their minds. Let's not take these results too seriously at this point in the campaign. (Two other sites that seem to provide empirically-based projections are 270towin.com, and electoral-vote.com, showing Obama with overall leads of 54 and 85 votes respectively.)