Polls, Polls, Polls

Yesterday, a slew of national and state polls were released to deduce the “convention” bounce. The broad picture is harder to measure than usual. Because the conventions were back-to-back, we don’t have a good reading of the pre-RNC state of the race. The pre-RNC polls were at the peak of the DNC bounce. Thus, it is unsurprising that the RNC bounce has been bigger (+6) than the DNC bounce (+4). Usually convention bounces dissipate over time, but not always. We need about a week or two to see if that happens. Nevertheless, there is good reason to be excited about some of the recent polling that shows Sen. McCain and Gov. Palin can win the race, and there is a majority in the country willing to support them. The futures markets have responded, giving McCain a 48% chance of winning which is his highest chance of the entire campaign.


The slew of FOX News/Rasmussen state polls are actually not great for Sen. McCain. These polls are taken at his convention bounce peak. The best news is the 51-44 lead in OH which even if it narrows is a good margin. The other states are all toss-ups at McCain’s peak: FL (48-48), PA (45-47), VA (49-47). The fact that FL is tied while OH is +7 means it is possible that FL will become a more closely contested state than OH in the end. At the least, this suggests that the McCain campaign will have to spend money in FL that they hadn’t as of late June (no TV ads by then). The VA and PA numbers aren’t bad, but if this is a peak they aren’t good either. The worst number of the 5 states polled is CO where McCain is losing 46-49 at his peak. With the emergence of energy as an issue, many people thought states like CO would move toward McCain. The addition of a hunter and fellow Westerner to the ticket was also thought to help. The one data point we have so far is that the most likely “tipping point” state is still in Obama’s column despite any convention bounce.

Interestingly in those 5 polls, the independent voters are all over the map: OH (+26), PA (-19), VA (+11), FL (-17), and CO (-10). Obama has less than 80% of Dems in FL (79%), OH (78%) and PA (74%). Obama gets 10% of GOPers in CO (10%) and PA (14%).

Besides the FOX News polls, several other state polls help define the current map. The most interesting data includes MI (Obama +1), Washington (Obama +4) and Virginia (McCain +2). The WA number is much closer than past polls and could add a new swing state for McCain to contest. The VA and MI polls echo the FOX News result that the swing state movement is less than the national movement. If McCain is ahead 2 or 3 today but losing MI by 1 and only winning VA by 2, that means the possibility of winning the national vote by 1-2% and losing the electoral vote is real. The one other state poll is a nice Oklahoma showing (McCain +33) that makes me proud to be a Sooner.

National polling analysis below the fold:


Let’s start with national polling. First, the USA Today (McCain +10) poll is a clear outlier. The rest of the national polls show a small McCain lead (tie, tie, +5, +1, +2, and +2). Assuming some dissipation, these numbers are nice but not a big lead. More important, Sen. McCain improved his image on a slew of indicators. McCain has a higher favorability rating than Obama now. He is more trusted on the economy than before, his lead among national security voters is higher, and he narrowed two big gaps: who can bring needed change and the excitement gap.

There are enough new national polls that I won’t discuss each individually, but I think there are two major possibilities to describe what is happening:

1) The race is a repeat of 2004 with one major difference. I’ve long argued that using 2004 as a base map is an error after the shift in partisan ID and the shift of the independent vote in 2006. Arguably, the conventions have moved the electorate closer to 2004 than 2006. We need more partisan ID polls to be sure but the most recent Rasmussen party ID poll shows the 10 point gap narrowing to 6 points. It was 6 points in 2006 and about 1 point in 2004.

The evidence of similarity is the large lead the GOP candidate has on national security and Iraq. This may surprise some liberals, but Sen. McCain has 20-40 point leads in the electorate’s view on national security, terrorism, and Iraq. Less than 50% believe Obama is qualified to be President, and that belief is based on his lack of any background with the military or foreign policy. More similarity is seen in the narrower leads the Democrat has on domestic issues: health care, the economy, and education.

The big difference is many more people care about the economy in 2008 than 2004. If 30-40% of voters were voting on national security issues, Sen. McCain would have a solid lead. But now 15-20% of voters are putting those issues on top. This shift toward a “pre-9/11” electorate makes the issue set friendlier to the Democrat.

If this model is correct, we are in for another narrow election depending on how independents swing in key states. Sen. McCain deserves a lot of credit if he has moved the election from a 2006 map back to a 2004 map. Independents fled the GOP in 2006, and Sen. McCain has succeeded in bringing enough back to make the contest as competitive as 2000 and 2004.

2) The race is similar to 2006 but Sen. McCain appeals across parties while Obama does not. That may sound similar, but it’s a different model. If we start with 2006 as a base map, then the GOP starts down 5 points. I presumed this model throughout this year. The party ID numbers (although better than six months ago) are still around where they were in 2006. The generic House ballot is similar to the 2006 numbers. The competitive House and Senate races are almost all in districts that would be safe under a 2004 map.

The reason we are in a competitive race under this model is that Sen. Obama loses more than 10% of the Democrats to Sen. McCain, and Sen. McCain has personal appeal to independents that overcomes the anti-GOP sentiment. In this case, the GOP is still going to lose badly in the House and Senate races but the Presidential race is close. The swing voters are Democrats and Independents who like both candidates, but dislike the Bush administration. A good portion of this swing group are concerned that McCain will be another Bush term and are concerned that Obama is woefully unqualified to handle major crises or issues of war and peace.

I continue to lean towards the second model. If the GOP really improved its image among non-partisan voters, we should see evidence of that in new Senate and House polls over the next few weeks as well as generic ballot numbers.

I expect more polling throughout this week as pollsters try to understand the overall outlines of the post-convention race. I’ll try to write up most of them as they come in.