Reading the July/August issue of the Atlantic Monthly on the plane to California, where I took a much needed two-week vacation traveling up the Pacific Coast, I ran across this line in Joshua Green's piece about the horrendous state of modern political advertising:
There is ample scientific evidence that, despite widespread public distaste for them, negative ads are the most effective kind, because people are more apt to remember negative information than positive information.Green is echoing what is perceived to be a truism in American politics, one repeated often by pundits, consultants and politicians with the same certainty they used after 9/11 to say that George W. Bush was unbeatable. As that analogy suggests, however, the truth is far more complicated, and it's past time for people to understand this.
Green's logic – negative ads should be more effective because people are more likely to remember negative information – is plausible enough. Generally speaking, it's true that we obsess about the things that annoy us more than information that makes us happy. In fact, this is one reason people on both sides of the aisle think the mainstream press is biased against them: The stories that were critical of their side stick in their craw and seem more prevalent than those that are neutral or favorable (something academics call the hostile media effect).
But while this would lead us to hypothesize that negative ads are more effective, absent any actual evidence that it does we cannot be sure. In fact, despite numerous well-crafted studies by scholars from a variety of disciplines using a host of often creative methods, it is not at all clear that negative ads have the uniformly powerful effects most assume they do.
The problem begins with the fact that academics can't even agree what a negative ad is. Some define it as an ad that makes any negative claim about an opponent, or in the case of ads run by political parties and interest groups, a target candidate. This is obviously too blunt an instrument – an ad that makes nine positive claims and one negative is surely not the same as one that attacks nine times and ends by saying the sponsoring candidate loves puppies – so others have used a variety of other metrics. These include whether the majority of claims are attacks, whether "most" of the ad is attacking, or whether the ad is deceptive.
So pundits and political consultants can be forgiven for being confused. It's not as if the experts are helping them out much.
Perhaps the most important thing to understand is that simply talking about negative and positive advertising is misleading. As many in the press are starting to grasp, many ads – let's call them contrast ads – include both. In an ideal world – and assuming all of the information in the ad to be fairly and accurately presented – this is what we would want to see: Candidates not only telling you why not to vote for their opponent, but why you should vote for them. Studies show the obvious: These ads have more information in them than do pure attack ads or, especially, puffy advocacy ads.
A second problem is that not all attack ads are bad. If I accurately say that my opponent, say, sent young men and women off to die in a war based on false pretense and exaggerated claims, that's not only accurate but fair. If, on the other hand, I say that my opponent cast three different votes to cut funding for troops and weapons systems when in fact he cast one vote for a massive bill that happened to include some of those outlays, I'm being deceptive. Fair attack is necessary and useful in a campaign; lies are not.
But the question is, which kinds of ads are more "effective?"
Scholars are divided (in some cases bitterly) over whether negative advertising decreases or stimulates voter turnout overall. There is some evidence to support both the mobilization and demobilization arguments (though probably more on the side of a stimulation effect or no effect at all).
But let's face it, candidates don't really care about total voter turnout, they care about whether more people vote for them than their opponent. So what role do negative – or, for that matter, contrast and advocacy – ads have in helping a candidate win?
Some research shows that negative ads sometimes do work in damaging one's opponent, but that there can also be a boomerang effect if the ad crosses a line. Specifically, when an attack is seen by voters as fair it appears more likely to have the intended effect – that is, it "works" – whereas when it is unfair it doesn't. At least one study has shown that negative contrast ads hurt the opponent without hurting the candidate sponsoring the ad.
Similarly, in one of the better studies on the subject, University of Pennsylvania scholars Kathleen Hall Jamieson, Joseph Cappella and Dan Romer looked at the effects of all three types of ads in the 1996 presidential election. They found that purely positive ads increased the candidate's vote share (that is, the percentage of people who actually voted – what they really care about), but had no effect on overall turnout (that is, what do-gooders care about). Pure attack ads reduced both overall turnout and the vote share of the candidate sponsoring them. Contrast ads, on the other hand, increased both turnout and the candidate's vote share.
Their study did not break out the effects by partisans, so we don't know for sure who is being affected by these different ads. Other studies that have shown some evidence that negative campaigning in fact bucks up one's base while keeping fence-sitters at home – the Bush strategy this year – are equally hard to interpret because they didn't break out the effects of contrast ads from other ads that included attacks.
What does all of this mean for this year's presidential election? For one thing, it means journalists need to be more careful in discussing the purported utility of negative advertising. Let's be clear: I'm not saying there's no evidence for the claim of Green and others that attack "works." I'm saying that the truth is far too complicated to make blanket statements about its effectiveness.
The reality seems to be exactly what we would want it to be: Fair attacks help a candidate by hurting the opponent, while vicious and deceitful slurs can often give him a faceful of his own mud.
It also might help us understand the limited effectiveness of the gluttonous ad buys by the Bush campaign this spring. The ads seem to have been successful in planting the seeds that John Kerry is a flip-flopper – though this probably has as much or more to do with an uncritical press' regurgitation of the charge – but have not succeeded in opening up the kind of intimidating lead Bill Clinton did when he flooded the airwaves in the spring of 1996 with ads. (Most of them used contrast, by the way – the typical one would close, "Bob Dole: wrong in the past, wrong for our future. Bill Clinton: protecting our values.") Perhaps it's not a coincidence that Bush's ads are grossly misleading (to be charitable).
Polls show voters don't really have much of an opinion about Kerry, and at some point he will need to introduce himself to them more effectively than he has. (The early ads have left a lot to be desired, to say the least.)
But by the same measure, at some point Bush is going to have to stop relying on misdirection and mendacity in his. Truth should not be a stranger to political advertising, any more than it should to policy, presidential addresses, and press conferences.