Friday, September 5, 2008

Polls Post the Republican National Convention

Today marked the first full day after the republican national convention, which means today's polling numbers are the first time we get a chance to see what kind of impact Palin's speech has had. We will need another day to see the effect of McCain's speech and will probably need to wait until next Tuesday to see what kind of bounce McCain/Palin get from the convention. 

So, for everyone who was either frantically excited or upset regarding Palin's speech you all appeared to have canceled each other out more or less. Of the four major polls (I don't use polls that are done by the DNC or RNC as they are partisan polls and notoriously unreliable) that were put into the field (CBS/Rasmussen/the Economist/and Gallup) we have a 1-6 point lead for Obama. While this clearly shows a reduction of the 6-10 point lead that Obama had coming out of the Democratic convection, this reduction is more of a statistical correction than a big bounce McCain/Palin. 

Without giving a long and complicated explanation of how modern polling is done, the basic gist is that polls are constantly in flux. Taking a look at the sample size on the polls done right after the Democratic convention and the standard errors, there was a good chance that these polls had a little inflation in them. So, if there had been no republican convention at all, Obama's numbers should have come down to about a 6 points lead. Thus, the fact that 36 million people saw Palin's speech (Obama had 37 million) and they only got what amounts to a two point bounce, most likely means that Palin's speech energized both bases at once. She may have won a lot of social conservative hearts, but she also made a lot of liberals mad and energized to work for Obama.

Anecdotally, such a reading of the poling numbers is supported by the fact that after Palin's speech McCain claims to have raised 10 million, while Obama claims they had an influx of donations amounting 9 million, without a huge effort on their part to solicit. 

In addition, it is important to note that the CBS poll is the only poll that has Obama and McCain tied in a national poll, while the other polls have the race a little more separated (remember that national polling is not the best measure of who will win the election since we have electoral votes). Also, when the stats on the CBS poll are examined in detail it is important to note that it has a party ID distribution that is not the same as the other three, which may have lead it to overstate McCain's strength. We will have to watch this poll over the next week, but I am willing to bet as the sample size is increased, we will see the race separate between Obama and McCain again.

Also, what is even more interesting in today's polling than the national polls are some polls of likely voters in swing states. First, while some had believed prior to Palin that Alaska might be in play this year with the republican senate scandals of good old Ted, polling now shows the McCain/Palin ticket to have a 19 point lead. However, there wasn't much movement in key swing states, such as Indiana (McCain still leads by about 3 points), Penn (Obama by 2 points), Ohio (Tie), Florida (McCain by 5 points), and Colorado (Obama by 2 points). The key take away point here is that Palin's speech didn't really do much to move anything where it mattered. 

As we get a little farther away from the Republican Convention we should be able to see exactly what the bounce will be, but if these numbers are any indications, it appears that all Palin has done is to increase support in already red states that McCain was going to win anyway. 

2 comments:

Dave said...

I think it's also entirely possible that while 36 million people tuned into Palin's speech, a good number of those might have been curious rather than interested - between many people not knowing who she was and late night comedy absolutely abusing her for the past week, there was a lot of buzz about the "hot librarian," as John Stewart dubbed her.

R.P. McMurphyDBB said...

I think that is a good point and if true lends even more support to the idea that in the end she isn't going to have an meaningful effect on the outcome of the election.