Based on the latest data from the Wisconsin Advertising Project, Obama's spending advantage is starting to come into play. In the week from September 28 to October 4, Obama-Biden spent $16.2 million on ads in 15 swing states to McCain-Palin and the RNC's $9.5 million.
Is this spending making a difference? The answer is a qualified yes.
To figure out the impact of Obama's TV ad spending edge in various states, I updated this spreadsheet to aggregate a number of data points, including RCP averages from various points in the race, 2004 results, and aggregate ad spending as well as ad spending per voter per week.
In states where Obama has an above average spending advantages (more than 1.7 to 1), his RCP average lead has moved up 3.61 percent versus in 2.44 percent in states where the candidates were more stalemated in terms of ad spending. A similar shift of about 1.2 points was found when comparing these polling averages to July 1 or the first date we had polling for a state, when ad spending was a fraction of what it is today.
With 15 states, this kind of analysis is plagued by a small-n factor in which one state can throw the calculus off dramatically. It's also worth pointing out that state-specific RCP averages can be less reliable than their national counterparts.
In these cases, I find it useful to build typologies state by state. Overall, I found 5 states (Florida, Missouri, New Hampshire, North Carolina, and Virginia) in which a massive Obama spending edge correlated with an above-average shift to Obama since September 28, versus one state that defied this pattern (Indiana). Remove Indiana and the ad spending effect rises to 1.6 points.
As for the rest of the states, all but one are grouped as battlegrounds in that both candidates were competitive at a 2-to-1 spending ratio or lower, and these generally resulted in more muted shifts or contradictions between the July and September deltas.
The odd man out is Minnesota in which McCain is actually outspending Obama and has held the line relatively well there. This is the one example of an ad effect working in his direction.
Here's a scatterplot of Obama's ad spending edge in dollars per voter last week, mapped against the change in the RCP average from July 1st and from September 28th:
So, the difference between outspending your opponent 3.6 to 1 and 1.3 to 1 could be about 1.2 points -- maybe 1.6. But I wouldn't write home about this until we have several weeks of ad spending correlated with corresponding before and after poll data.
It's also worth noting that this cash advantage is not recreatable in every state. There are still lots of states where McCain will compete aggressively. That means saturation advertising on both sides, making it physically impossible to maintain a 3-to-1 cash advantage. And even if it were possible, it wouldn't be as effective since McCain has established a baseline presence.
Obama's spending advantage is also mostly in Republican states that are not likely to be tipping points that have seen little McCain advertising. These uncontested situations In these states are ripe for moving poll numbers. Obama is advertising at normal levels and McCain is on the air little or not at all. However, these will be wasted dollars if Obama doesn't succeed in flipping these states or realign them to the point where they are electoral vote #270.