Accentuate the Negative

Do negative campaign ads work? In the latest University of Texas/Texas Tribune poll, we tested the effectiveness of two of the earliest "comparative" ads run in the governor's race to see if they persuaded voters. Apparently so.

Republicans Kay Bailey Hutchison and Rick Perry were the first two candidates in the top race on the ballot to run negative spots. When our poll went into the field at the beginning of the month, Republican Debra Medina and Democrat Farouk Shami weren't running any negative commercials, and Democrat Bill White wasn't running any television spots yet at all.

The pollsters split registered Republican voters into two groups and asked each to rank the two gubernatorial candidates, then watch an ad and then rank them again.

Group A saw Perry's "Bailout" ad.

Group B saw Hutchison's "Cha-ching" ad.

In both cases, the commercials influenced voters, drawing their preference away from the candidate being attacked. But it didn't work on everyone.

"The Kay supporters see a negative ad about Perry, there's a little movement. The Perry people, the same," said pollster Daron Shaw, a professor of government at the University of Texas. "But the people who are undecided are the more interesting ones. Their jump was bigger ... which, of course, is a big point." Undecided voters, after all, are the people the candidates are fighting over.

The changes weren't huge but were "statistically significant" — the ads worked. Neither commercial, by the way, is still in circulation; both of the candidates have moved into other positive and negative territory. But they do get at the question of whether the barrage of commercials you've seen and will continue to see until Election Day on March 2 does anything for the campaigns.

Campaigns typically don't do polls on their commercials, Shaw said, generally using focus groups to gauge the effectiveness of their commercials. They're cheaper and faster, and they let the campaigns target their commercials to particular groups — whether the people in question are Soccer Moms, Angry White Men or whichever demographic is in play in a particular election.

In our poll, respondents rated the two candidates on a scale of 0 to 100, with Perry at one end and Hutchison at the other. Perry's spot attacking Hutchison moved his folks two points in his direction, Hutchison's folks five points in his direction and undecided voters almost nine points in his direction. Hutchison's attack on Perry pulled his voters her way by three points, improved her score with her own folks by two points and pulled undecided voters almost six points her way.

"Both of them had an effect," Shaw said. "The effect was significantly more pronounced among undecided voters. And there seems to be a slight edge to the Perry ad."

The Internet survey of 800 registered voters was conducted Feb. 1-7 and has a margin of error of +/- 3.46 percent. The GOP primary questions (including the advertising questions in this story) have a +/- 5.12 percent margin of error; the Democratic primary questions have a margin of error of +/- 6.02 percent.