I have not seen enough evidence yet to say I've made a firm conclusion on this issue, but I think there may be a divide in the approval ratings pollsters are finding for President Obama depending on whether they're using IVR or live callers.
A few things have put this on my radar screen:
-If you look at Obama's national approval on Pollster.com's chart, most of the lower approval levels are coming from Rasmussen.
-While Civitas found Obama's North Carolina approval this week at 65/16, we only had it at 52/41 last week.
-The other places we've asked poll questions about Obama this month were New Hampshire and Texas. We'll release the Texas result later this week, but we found a net negative approval for him there. I haven't seen any live operator approval ratings for him there yet.
In New Hampshire we did not ask an approval question, but one about the stimulus. 50% said they supported it and 40% were opposed. I think that's a pretty good surrogate for Obama's approval since we found the exact same +11 for both the stimulus and his approval in North Carolina. Those are much more modest numbers than the Granite State poll, which found 66% approval and 21% disapproval for Obama in the state.
Is the Bradley Effect finally showing up? I doubt it. My guess is that if this really does prove to be a trend it's that Republicans are more likely to want to seem bipartisan to a live caller than on an automated poll. Our NC poll last week found a very strong partisan divide on Obama that does not seem to have been as pronounced on the Civitas poll.
Again the body of evidence is limited at this point but it's certainly something to watch.