I have a new post at 538 on this question. Based on some initial descriptive evidence from The Gamble, Lynn Vavreck and I are skeptical. We will do more to nail this down in the book, but at this point we can discern no shift in assessments of Romney that coincided with Obama’s ad blitz in May and June. And the fact that the effects of ads decay rapidly suggests that, even if such a shift had taken place, it wouldn’t have persisted until November.
Politico’s Glenn Thrush objects:
Phil Klinkner notes that if Obama’s early ads were somehow preventing movement toward Romney, there perhaps should have been movement toward Romney where there were no Obama ads—i.e., in the non-battleground states. But there wasn’t, as my post shows.
However, it’s important to be clear on the underlying counterfactual to help clarify causal claims, like Thrush’s claim that Obama’s early ads caused the polls to remain more stable than they would otherwise. I don’t know what Thrush’s precise counterfactual is, but here are two possibilities and my view of them:
1) What if Romney and his allies had advertised at the same level, but Obama had run no ads at all? If that had happened, I would have expected Romney to gain in the polls, given the large number of ads he was running. My 538 post acknowledged this, at least implicitly, by noting the likely folly of unilateral disarmament.
Again, the point here is not that ads never matter. It’s just unlikely that Obama’s early ad binge was that effective. The polls didn’t change, as you might expect if these ads were so numerous or so devastating. And there’s not a good reason, in my view, to think that polls would have changed absent this binge, as long as Obama was airing ads at a level comparable to Romney’s.