Victory probability mapObama lead over time

Monday, September 01, 2008

data integrity; Biden and convention bounces

I was surprised by the sudden uptick in Obama support, which appeared to happen a couple days before Obama announced Biden. I tracked it down to a data input error: Instead of typing in Rasmussen's recent TN poll as 60-35 McCain, I had typed it in as 60-35 Obama. So, I've been spending most of today creating some data integrity checks so that errors don't recur. I also went back through all the past data, and found that there were only a few errors, and they were generally immaterial.

My best evidence is that Obama's lead has grown by about 1.5% from mid-August to the close of the convention. It increased by about 0.5% in the days leading up to the announcement of Biden, another 0.5% right after that announcement, another 0.5% through the convention, and continues to trend upward. This isn't a humongous bounce, but it's about what I expected in today's 46-44-10 landscape.

The above chart shows, for each recent date in August, my best estimate of Obama's lead using only polls through that date.

I predict that McCain's bounce will be slightly larger than Obama's, and that Obama's lead will fall to about 1% after the Republican convention. The reasons are that McCain played his VP announcement excellently, and that the Republicans are getting great press by scaling down their convention during Gustav (much better than they'd be getting if George Bush and Dick Cheney were in St. Paul).

Although I think that McCain's inexperienced VP choice will be a liability for him in the long-run, Obama's response was uncharacteristically flat-footed. It's pretty clear to me that their response ad was a prefabbed ad that was intended for release regardless of who John McCain chose. They created an ad that avoided mentioned the name or sex of the running mate in the audio, and then just plugged in a couple shots of McCain and Palin once she was announced. As a result, the generic message of this generic ad ("VP pick is not a change") clashes with the obvious truth that McCain's pick is downright weird.

The Obama campaign has a couple big themes at its disposal -- change and judgment. As I just mentioned, "no change" is a hard message to make about Gov. Palin. Instead, the Obama campaign should have (and I think would have, if they had prepared properly) made an ad on the theme of McCain's judgment in choosing an inexperienced small-state governor.

Luckily for Obama, the McCain campaign seems to have prepared for his VP choice almost as little as the Obama campaign did. It appears that they didn't do due diligence on her state trooper scandal or her family situation, and they still have Alaska drawn without its panhandle on their U.S. map.

0 Comments:

Post a Comment

<< Home