So perhaps “front-loading” the ads, as Mr. Axelrod described it on Politico.com, wasn’t as effective as many believe. In fact, the polls throughout this time period resemble this description:

What’s been interesting to watch is that our data has been remarkably consistent really from last spring forward, and our battleground polls really didn’t fluctuate much. There were times when it would dip to where we had a two-point lead in the battleground states. There’s one poll over the course that we had a one-point lead. By and large, we’ve been three and four points ahead in the battleground polls.

Interestingly, that description was also provided by Mr. Axelrod.

To be sure, my claim is not that campaign advertising in the 2012 presidential election did not matter at all. As I noted earlier, an advertising advantage — if it is substantial enough and close to the election — can sway votes. So it would be folly for a candidate to unilaterally disarm and run no ads whatsoever. Moreover, this initial analysis cannot speak to other kinds of counterfactuals, including ones that the Romney camp is still pondering like, “What if we had run more positive ads?”

But the political science research and this initial evidence from 2012 suggests that the Obama campaign’s blitz of early advertising did little apparent damage to Mr. Romney in the minds of voters.

Political ads are for defense, not for offense.