Pollster.com

Articles and Analysis

 

A Hits B, B Hits A, C Wins?

Topics: 2008 , The 2008 Race

Read between the lines of some recent reports and it sure sounds as if negative attack ads will be airing soon. "By December," The Washington Post's Dan Balz writes this week, "it's likely that television ads in Iowa will be airing that directly attack her [Hillary Clinton]." Ken Wheaton of Advertising Age agrees that "chatter from inside the parties" means "the negative campaigning would start just around Christmas." Meanwhile, a clever attack ad by the Edwards campaign posted to YouTube wins plaudits as "a mini-Internet sensation." Edwards consultant Joe Trippi mocks the Obama campaign for failing to "take the gloves off" and says that "asking questions about a candidate's position on the issues is not attack politics, it's responsible politics."

So with speculation building about a coming negative ad war, here are some words of advice: In a multi-candidate primary, an attack ad strategy presents huge risks for the attacker, and not just in Iowa.

Consider the names and races on the following list. What do they have in common?

All eight emerged as "surprise" winners in competitive multi-candidate races. All eight surged to victory in the final days of their respective races as the more heavily favored candidates hammered each other with negative attack ads. In all but one of these races the pattern was the same: The two front-running candidate - let's call them Candidate A and Candidate B - ran attack ads against each other. In each case, the ads "worked," but in an unexpected way. They moved support away from both A and B to the benefit of a third Candidate C (the names listed above) that was able to communicate a mostly positive message that reached voters in the final weeks of the campaign. When this phenomenon occurred in the Iowa Caucus campaign four years ago -- a negative exchange between Dick Gephardt and Howard Dean that worked to the advantage of John Kerry and John Edwards -- the pundits dubbed it "murder suicide."

The one semi-exception to the typical pattern was the 1984 New Hampshire primary in which John Glenn ran television ads attacking Walter Mondale in the final weeks of the campaign. As far as I can tell from the historical record (and my own memory), Mondale never responded with counterpunch ads of his own. According to Jack Germond and Jules Witcover's Wake Us When It's Over, Glenn's attack ads raised Mondale's negatives and initially helped Glenn narrow Mondale's lead in New Hampshire. However, after Gary Hart's surprise second place finish in Iowa made him a viable challenger to Mondale, internal campaign polls in New Hampshire showed Glenn "fading rapidly" with all the benefit going to Hart.

Yes, as First Read reminded us recently, voters in focus groups "don't like negative ads" but "that doesn't mean negative ads don't work." True on both counts. Voters do not like negative ads, but they work best in general election contests when voters typically have only two viable choices. Two-way contests are essentially a "zero sum game." If A moves support away from B, it goes to A, and vice versa. However, when voters believe they have a real third choice, the negative ads sometimes work to the benefit of that third candidate.

Of course, as a wise Republican pollster put it to me recently, the "necessary corollary" of this pattern is that Candidate C needs to be more than a passive bystander. That candidate needs "to be enough of a force field to have some gravity of their own" to pick up the voters that become disillusioned with the candidates at the top of the ticket. The various Candidate C's above had achieved critical mass through a combination of many of the following: Enough paid advertising toward the end of the campaign to establish solid name recognition, a coherent and differentiating message, perceived success in campaign debates, late newspaper endorsements and last minute evidence of growing viability (usually in the form of last minute public polls showing the candidate gaining and "doing better").

What is clear about the presidential nominating races in both parties is that several candidates on each side are or may soon be positioned to play the Candidate C role.

Now for some cautions: First, keep in mind that the "negative attacks" that typically trigger this phenomenon are paid advertisements that reach a mass audience on broadcast or cable television, not criticism in debates or speeches. Most if not all of the C candidates listed above criticized their opponents in debates or public appearances. Negative broadcast ads are more likely to boomerang than other attacks (including those that appear only on the Internet) for two reasons: They reach politically inattentive voters that pay less attention to politics and are typically harsher in tone than the statements made in speeches or debates. As such, the gentle chiding in Chris Dodd's recent ads do not fit the bill.

Second, the list above amounts to a pretty small sample size, and most of the examples come from Democratic primaries (perhaps reflecting my own skewed experience as a Democratic pollster). Readers may know of exceptions to this pattern, and if so, I certainly urge them to leave comments below.

Third, it is worth noting that John Edwards, should he begin airing broadcast attack ads, is running behind both Hillary Clinton and a potential "candidate C" Barack Obama. My own sense is that only increases the risk for Edwards, but it certainly makes his situation a bit of a break from the typical pattern.

For all of these reasons, I hesitate to describe the "A hits B, B hits A, C Wins" pattern as an inviolable "rule." Sometimes, campaigns find a way to defy the conventional wisdom. My point, again, is that in a competitive multi-candidate primary, a candidate takes an enormous risk in embarking on a broadcast attack ad strategy.

Partly for that reason, I am guessing that the Edwards campaign has not yet committed to "going negative" with its Iowa television buy. I have no inside information, but Clinton's position in Iowa is weaker than elsewhere, most polls show Edwards within single digits of Clinton and Obama and the Edwards campaign has just started airing its positive ads. For all the bravado, his consultants know the history as well as anyone. Edwards pollster Harrison Hickman (my one-time employer) worked on the Glenn campaign in 1984. So I am guessing that the Edwards campaign is holding fire to see if the combination of positive television and attacks limited to speeches, debates, Internet ads puts them in position to win.

We will soon see.

 

Comments
Bobby Z:

Sorry to go off topic, but what is a New York (NY1) poll doing on the Nevada Republican primary page (here)?

____________________

me:

I thought Steve Murphy was the one who first called it murder suicide.

____________________

CalD:

The first person I ever heard refer to Iowa 2004 as a murder-suicide was Joe Trippi. I thought at the time that it was a remarkably candid and dead-on accurate assessment of what happened between Dean and Gephardt. It wasn't until much later that I realized he thought it was Gephardt who had done the murdering and then shot himself. When I did realize it I thought to myself, "He hasn't learned a damned thing."

But I also think that the Kerry campaign's effective use of micro-targeting to identify and turn out support in places no one else thought to look tends to be given way too little credit in general, as does the remarkably effective field operation they had put together under the radar, while all eyes were on the leaders. My guess is that Kerry was already on track to do much better than anyone expected, then when Dean and Gephardt imploded, he was well positioned to exploit the windfall. You know what Louis Pasteur said about chance and the prepared mind.

Anyway, getting back to the present, a paper I read a while back on negative campaign advertising (most of which can be found in this article) made what I though was a compelling case that for negative ads to work well, they must be above all be perceived as fair. Since Edwards is trying to outflank Clinton and Obama on the left, but was actually somewhat more conservative than either of them while in the senate and has been working in the financial industry since then, that could be a tough needle to thread.

It's also possible that attacks on Clinton, who has obviously weathered a lot of them in her time, could tend to harden and energize her support particularly if unleashed during the Christmas holiday and especially if she is able to characterize them as unfair in any way -- e.g., as hypocritical, distortions of her record, irrelevant to the discussion, acts of desperation, etc. One more possible dynamic is that I think at least some significant fraction of Obama's support may tend to be as much anti-Clinton as pro-Obama. If so, then Edwards going after Clinton more forcefully might tend to draw off some of Obama's support, perhaps setting off some friction between Obama and Edwards and setting up a second front.

____________________

CalD:

The first person I ever heard refer to Iowa 2004 as a murder-suicide was Joe Trippi. I thought at the time that it was a remarkably candid and dead-on accurate assessment of what happened between Dean and Gephardt. It wasn't until much later that I realized he thought it was Gephardt who had done the murdering and then shot himself. When I did realize it I thought to myself, "He hasn't learned a damned thing."

But I also think that the Kerry campaign's effective use of micro-targeting to identify and turn out support in places no one else thought to look tends to be given way too little credit in general, as does the remarkably effective field operation they had put together under the radar, while all eyes were on the leaders. My guess is that Kerry was already on track to do much better than anyone expected, then when Dean and Gephardt imploded, he was well positioned to exploit the windfall. You know what Louis Pasteur said about chance and the prepared mind.

Anyway, getting back to the present, a paper I read a while back on negative campaign advertising (most of which can be found in this article) made what I though was a compelling case that for negative ads to work well, they must be above all be perceived as fair. Since Edwards is trying to outflank Clinton and Obama on the left, but was actually somewhat more conservative than either of them while in the senate and has been working in the financial industry since then, that could be a tough needle to thread.

It's also possible that attacks on Clinton, who has obviously weathered a lot of them in her time, could tend to harden and energize her support particularly if unleashed during the Christmas holiday and especially if she is able to characterize them as unfair in any way -- e.g., as hypocritical, distortions of her record, irrelevant to the discussion, acts of desperation, etc. One more possible dynamic is that I think at least some significant fraction of Obama's support may tend to be as much anti-Clinton as pro-Obama. If so, then Edwards going after Clinton more forcefully might tend to draw off some of Obama's support, perhaps setting off some friction between Obama and Edwards and setting up a second front.

____________________

CalD:

Oops. Sorry for the double post. I got a server error the first time.

____________________

Jay Gold:

The murder-suicide phenomenon goes back as far as anybody can remember. I remember the 1968 New York State Democratic primary for the U.S. Senate. Joe Resnick, a Humphrey supporter, and Eugene Nickerson, a Bobby Kennedy supporter, killed each other off, allowing Paul O'Dwyer, a Gene McCarthy supporter, to win the nomination (and go on to be trounced by incumbent Republican Jacob Javits).

____________________



Post a comment




Please be patient while your comment posts - sometimes it takes a minute or two. To check your comment, please wait 60 seconds and click your browser's refresh button. Note that comments with three or more hyperlinks will be held for approval.

MAP - US, AL, AK, AZ, AR, CA, CO, CT, DE, FL, GA, HI, ID, IL, IN, IA, KS, KY, LA, ME, MD, MA, MI, MN, MS, MO, MT, NE, NV, NH, NJ, NM, NY, NC, ND, OH, OK, OR, PA, RI, SC, SD, TN, TX, UT, VT, VA, WA, WV, WI, WY, PR