Imagine you’re studying the impact of food consumption on health. You’re running a huge experiment — you rounded up 100,000 people, and measured all their vitals (weight, blood pressure, good cholesterol and bad, blood sugar, etc.).
You then divided them into two groups, strictly at random. Of course, some members of each group are healthy, others aren’t. Some have high cholesterol, others are low. Some eat healthy, others are junk food addicts. In short, each group looked like the population as a whole.
Then the experimental treatment: One group ate a single potato chip, the other had none. Each person was then retested.
{mosads}Would you expect to find that eating a single potato chip affected the health of your subjects?
I wouldn’t.
Would you then conclude from this massive, carefully designed study that eating potato chips has no effect on human health?
Why not?
In part, because, to paraphrase a famous potato chip ad, “Nobody eats just one!”
Potato chip eaters gobble up bags full each week. And at that volume they really are bad for you.
Say you took a large group suffering from infections and gave them one antibiotic pill and found nearly all of them still sick. Would you conclude antibiotics don’t work?
No.
Nobody would be content with just one pill.
Those still with me are no doubt wondering, “What the heck does this have to do with politics?”
It has everything to do with a comprehensive set of experiments analyzed by two distinguished scholars from Berkley and Stanford that has already been widely disseminated and wildly misinterpreted.
In short, as in my medical thought experiments, they demonstrated that a single piece of direct mail, or one canvass visit or a TV ad doesn’t persuade many people very much.
However, failing to emphasize this crucial limitation, they put their findings more broadly: “The best estimate for the persuasive effects of campaign contact and advertising — such as mail, phone calls, and canvassing — on Americans’ candidate choices in general elections is zero.”
The authors could have phrased their conclusion a bit more accurately this way: In the midst of substantial volumes of campaign-related communication, the incremental effect of a single piece of mail or a single ad or a single canvass visit is near zero.
Nothing really surprising there at all.
And precious little to support headlines like The Washington Post’s “The end of political campaigns as we know them? A new study suggests we’re doing it all wrong,” or Vox’s, “In general elections, campaigns’ attempts to win swing voters appear to not work at all.”
In the real campaign world, nobody gets just one ad or one piece of mail. To their frequent chagrin, voters are bombarded with a continuous barrage of communications.
Not one of them is likely to have a substantial impact on its own. But taken together there’s considerable evidence — though not of the highest quality — that such communications can change opinions or create views where none existed.
There are other limitations of the study, but it is important to acknowledge what this, and similar findings of “minimal campaign effects” going back to the 1940s, teach us.
First, persuasion is very difficult. It doesn’t happen very often and it doesn’t happen at a huge scale.
Be suspicious of consultants who regale potential clients with stories about ads that produced 20 point shifts in the horse race.
Second, volume counts. If you only have the budget for a couple of mail pieces or a few hundred points of radio or television you probably aren’t going to have much impact on the race. You may win, but it’s very unlikely to result from these sparse communications.
Third, as I’ve stressed here many times, partisanship matters — a lot. It’s one of the key forces limiting the persuasive impact of campaign communications.
Finally, careful targeting matters. The study finds persuasion is greatest when communications are highly targeted — not just to “independents” — but to individuals found to be sensitive to the specific message.
In the end, contra the Washington Post, campaigns aren’t “doing it all wrong.” But having a real impact is difficult and expensive.
Mellman is President of The Mellman Group and has helped elect 30 U.S. Senators, 12 Governors and dozens of House Members. Mellman served as pollster to Senate Democratic Leaders for over 20 years and as President of the American Association of Political Consultants.