May 6, 2013
An early Mitt Romney campaign advertisement.
An early Mitt Romney campaign advertisement.

For those who worry a lot about the influence of money on elections, a study by political scientist Michael Franz should be reassuring. Franz found that in 2012, the effect of general election televised ads on vote choice was very, very, small. In fact, in the three elections for which Franz has done his study, the maximum effect he found works out to about six percentage points — if one party ran its maximum number of ads and the other ran its minimum. That’s in 2008; in the elections with incumbents on the ballot, that maximum effect is even smaller, down to about one percentage point in 2012.

Granted, if it’s Bush/Gore 2000, every tiny bit could flip an election. But for normal presidential elections, in battleground states where both candidates have plenty of money to spend, what this basically means is that there’s no effect. Even if one side had far more ads it would make only a very small difference, and in the real world they just cancel each other out.

The key to understanding this result, and what it says about other elections, is all about the information environment. Basically, we all know so much about the incumbent president that even a whole lot of ads aren’t going to move the needle much. But that doesn’t mean that ads are never important.

Think of it this way. Imagine you’ve moved to a new town, and you’re planning on going out to dinner. If you’ve decided on fast food, it’s not really going to matter very much how many McDonald’s or Burger King ads you’ve seen recently — you’ve probably eaten at both enough time to have established your preferences. The added information from the ads is tiny compared to what you already know.

But if you want a local place, you might well be influenced if you’ve seen a half dozen ads — maybe just billboards — for Hunan Gardens but haven’t seen anything for Szechuan Empire.  The information from the ads may be tiny, but it’s all you have.

So the less information there is otherwise, the more ads are likely to matter. The biggest piece of information for most of us is right there on the ballot: the party of the candidate (that’s sort of like knowing whether a restaurant serves Italian or Chinese; it’s enough in most cases, for most people, to make a decision, especially if it’s the only information we have). So in regular, partisan, general elections, we wouldn’t expect huge effects from ads; the more information available, the less the effect. In nonpartisan elections or primaries, or in general elections with very little other information, ads should be relatively more important.

So keep all of that in mind the next time you read news coverage of how important campaign ads are — or the next time you consider donating to a campaign. Money doesn’t dominate elections nearly as much as some of the news coverage would suggest — but in the right context, ads can still have very large effects.