The two daily tracking polls, Rasmussen and Gallup, have had a rightward lean this campaign season. One of these polls has a history of weird results, especially polls released many months before election day. Indeed Rasmussen, which is the only poll to show Mitt Romney ahead at least often as President Obama, inexplicably revealed an Obama swing just before the start of the Republican convention. At the time, no other poll was showing such movement. Not surprisingly, their post-convention polling has Romney ahead. A big bounce! Even Nate Silver, who is very cautious about questioning polls said this:
the two-point lead that Rasmussen Reports had given to Mr. Obama before the convention was a bit unusual for them.
Okay, forget about Rasmussen for a minute. We have the other tracker to consider, the one Nate says he has a lot of respect for, Gallup. While their sampling technique has resulted in polling slightly less favorable to Obama than the average, the day-to-day swings have made sense, usually tied to actual campaign incidents. So if Romney gets a convention bounce, they should reveal it.
Details below.
Here are the most recent results of the 7-day Gallup tracker:
RELEASE 7-DAY RANGE O R
8/22/2012 08/15-21/2012 45 47
8/23/2012 08/16-22/2012 46 46
8/24/2012 08/17-23/2012 46 46
8/25/2012 08/18-24/2012 46 46
8/26/2012 08/19-25/2012 46 46
8/27/2012 08/20-26/2012 46 47
8/28/2012 08/21-27/2012 46 47
8/29/2012 08/22-28/2012 47 46
8/30/2012 08/23-29/2012 47 46
8/31/2012 08/24-30/2012 47 46
9/01/2012 08/25-31/2012 47 46
The last one listed there includes polling done on Friday, the day after the Romney and Eastwood appearances. It also includes 2 days of polling after Paul Ryan's speech and 3 days of polling after Ann Romney and Chris Christie.
First impression: NO BOUNCE!
President Obama maintains his 47-46 lead. However, we can't be sure there is no bounce because 1) there are days in the 7-day average from before the convention started and 2) the most recent day of polling, the only day after Romney/Eastwood, could show improvement for Romney which is just too small to move the 7-day average.
Here's an attempt at deducing the day-to-day polling over the last 10 days, a Romney "best case scenario" with the math still adding up:
Date O R
8/20: 46 48
8/21: 46 48
8/22: 45 47
8/23: 45 47
8/24: 47 47
8/25: 47 45
8/26: 47 45
8/27: 47 46
8/28: 47 46
8/29: 47 46
8/30: 46 47
8/31: 45 48
I think this is just about the best case scenario for Romney's August 31 polling, Romney +3. It could easily be the case that polling on the date was more like Obama 47, Romney 46 or something in between. But even a Romney +3 would have to be repeated for several days to move the Gallup 7-day average significantly. And before the full 7-day average is post-RNC, the Democratic Convention will have been underway for several days.
Romney's Gallup tracker "bounce" is going to be between 0 and +2. Historically, that will put it on the lower end of convention bounces.
One other note on the Gallup poll. The sum of Obama and Romney has been 93 for a week now after months of being mostly 92. The undecideds are slowing coming off the fence.