think on this.

Economic trends.  There is pretty good evidence that economic trends during the year before the election matter more than earlier trends.  Chris Achen and Larry Bartels have argued that the second and third quarters of the election year appear to matter most—at least if you focus on changes in personal income as the economic indicator.  So changes in July, August, and September are important.  Note, however, that this doesn’t mean we should hyperventilate about every monthly jobs report.  Most voters aren’t following the numbers that closely.  It’s more likely that a gradual accrual of good or bad economic news will filter down via news coverage.

Campaign advertising.  As I’ve noted before, campaign ads can have an effect, even in presidential races.  However, three caveats are important here, which speak to how one should follow the ads.  First, the effect of ads seems to emerge when one side is outspending the other by a significant margin.  How much of a margin is hard to say; let’s take 2-1 as a rough estimate, which corresponds to the apparently consequential imbalance in Bush and Gore ads in battleground states right before the 2000 election.  I’m not sure either Romney or Obama will muster that kind of advantage, even with the independent spending taken into account.  TBD.

The ground game.  To paraphrase a comment I heard from Sasha Issenberg, the problem with horse-race journalism is that it’s watching the wrong part of the horse.  It spends too much time pondering the (poorly understood) effects of messaging and not enough on the nitty-gritty activities of the campaigns on the ground.  Most importantly, we know from social science that get-out-the-vote drives can really work.  So if the election comes down to mobilizing supporters, the comments of Cory Booker on the Sunday morning shows will be far less important than whether Obama and Romney can use the accumulated data they have about voters to identify promising targets and then contact them in the most effective way.