Would an “Etch-a-Sketch” Attack Actually Work?

Brendan Nyhan concludes a great post on the Etch-a-Sketch comment by saying:

We’re currently 158 days from the Republican convention, and “Etch a Sketch-gate” will likely prove to be just as inconsequential. By the time the general election rolls around, the incident will most likely be forgotten.

I agree and said as much on Twitter.  But Kevin Collins then asked if this “gaffe” could the kind of thing that voters remember because the campaign emphasizes it.  That’s possible.  Maybe the Obama campaign will emphasize it.  But “Romney as flip-flopping Etch-a-Sketch guy” may not be the most effective message.  Here’s why.

It’s true that more people perceive Romney as a flip-flopper than they do Obama:

About 48 percent say that the phrase “takes positions and sticks by them” doesn’t describe Obama very well.  But 61% say that of Romney.  So far, so good for Obama.

The question, however, is whether voters’ perceptions of “whose the biggest flipflopper” are as potent as their perceptions of other characteristics of the candidates.  Consider characteristics like “cares about people like me” or “cares about the middle class”—qualities on which Romney is also disadvantaged relative to Obama.   Which of these is the strongest predictor of vote intention in a Romney-Obama race?  Using two January YouGov polls, I estimated the impact of different characteristics—takes positions and sticks by them, is personally wealthy, cares about people like me, cares about the poor, cares about the middle class, and cares about the rich—while controlling for people’s party identification.  In particular, I took ratings of each candidate on each dimension and subtracted ratings of Romney from ratings of Obama.  So these measures capture the kinds of comparative assessments that we think voters might make.

Here is the upshot: when people evaluate Obama more favorably than Romney on the flip-flopper dimension, they are also more likely to prefer Obama in a head-to-head match-up.  But this apparent effect pales beside the effect of two other dimensions: cares about people like me and cares about the middle class.  (The other dimensions do not have statistically significant effects.)

Take a political independent who thinks that the term “takes positions and sticks by them” describes Obama “not very well” but describes Romney “somewhat well.”  Assume their evaluations on the other dimensions are equal to the average of everyone else in the sample.  Now pretend this person switches their view to one more favorable to Obama: “takes positions and sticks by them” describes Obama somewhat well but describes Romney not very well.  What would happen to their chance of voting for Obama?  According to the model, it would increase by 31%.

But the same shift in perceptions of who “cares about people like me” would produce a much larger shift in Obama’s favor—61%.  The same shift in perceptions of who “cares about the middle class” has a similar effect, increasing the chances of voting for Obama by 59%.

Several weeks ago, Helene Cooper of the New York Times wrote that the Obama campaign was considering precisely the question this analysis addresses:

The bigger conundrum for the Obama campaign is how to balance its two lines of criticism of Mr. Romney, particularly if he wins the nomination. Do they go the out-of-touch, protector-of-Wall-Street route or the flip-flopper route?

These results—preliminary to be sure—suggest that people prioritize the candidates’ empathy more than their tendency to flip-flop.  If the choice is between drawing Romney’s face on an Etch-a-Sketch or putting it on a $10,000 bill, I’d take the bill.

[Cross-posted at The Monkey Cage]

John Sides

John Sides is an associate professor of political science at George Washington University.