Engagement is, for sure, a slippery concept with no universally accepted definition. But Harris Interactive found that in 2004, declining voter engagement with John Kerry (essentially his favorability rating) and his ads during the final six weeks of the campaign predicted his downfall better than conventional poll numbers -- at least better than their own.
As the 2004 race drew to a close, the marketing researchers (as opposed to the political pollsters) at Harris Interactive were actually coming to a different conclusion. While Harris Interactive's online poll showed Mr. Kerry ahead in Florida, Ohio and Pennsylvania, engagement numbers in the separate Harris/ARF online poll indicated, correctly, that he'd only win Pennsylvania.
Harris had been collaborating with the Advertising Research Foundation to track consumers' emotional connections with candidates -- essentially their positive or negative feelings on a numerical scale -- and how ads affected those connections.
"We were able to -- the day before the Iowa [caucus] results came in -- know that Howard Dean couldn't win," said Marianne Foley, senior VP-portfolio development and research of Harris Interactive. "Even though the majority of our sample said they would vote for him ... you can't win when one out of four people hate you."
Mr. Kerry's favorability ratings headed south in September 2004, both in Harris Interactive's research and in most published national polls. Negative ads were hurting him badly, but not Mr. Bush's nearly so much as Mr. Kerry's own.
Negative not for everyone
The Harris/ARF survey found the infamous Swift Boat ad actually did more to increase negative feelings about Mr. Bush than Mr. Kerry. Several of Mr. Kerry's own attack ads hurt his favorability ratings more than the Swift-Boat ad studied. Harris' research shows his favorability started tanking in September, when his own negative ads started, rather than August, when Mr. Bush's began.
The lesson isn't that negative ads don't work, only that they don't work for all candidates, said Ms. Foley. Mr. Kerry had promised to run a more positive campaign and did, which helped boost emotional connections, she said. When he started going negative, he severed those connections.
By two weeks before the election, Mr. Kerry's positives fell below Mr. Bush's, and his negatives rose above Mr. Bush's. "Based on those numbers alone," Ms. Foley said, we said that Bush would have to win the election ... even though our pollsters got it wrong."
Harris hasn't yet begun tracking the 2008 election and isn't making predictions, but the emotional-connection factor doesn't appear to be on Ms. Clinton's side, even though the polls are.
Problems for Clinton
She leads the Democratic field and Republican front-runner Rudy Giuliani in almost every published national poll. But she also has by far the highest negative ratings of any candidate in three August polls.
Both Mr. Giuliani and Barack Obama have much better ratios of positives to negatives. But both have added 10 points or more to their unfavorable ratings since January.
The good news perhaps for Ms. Clinton is that, like Mr. Bush, her image could also let her go negative more easily, Ms. Foley said. Voters had come to expect Mr. Bush to run negative ads by 2004, so doing so didn't hurt him.
"People know [Ms. Clinton] is not going to soft-pedal what she thinks, so they're not going to perceive her as mean or lowdown or mudslinging." Likewise, she believes Mr. Giuliani's tough-guy image will let him go negative.