Fuzzy Math About Negative Ads

Referring to an exchange between McCain and Obama about negative ads, Time's Karen Tumulty asks and answers: “Who's more negative? It's a draw.”

Tumulty based her conclusion not on a qualitative assessment of the candidates' ads, but on a statement from Professor Ken Goldstein of the University of Wisconsin Advertising Project, which counts ads and assesses them as “positive,” “negative,” or “contrast.”

The Project doesn't differentiate between true and false negative ads, or assess degrees of negativity -- “John Smith's tax plan is wrong for America” is treated the same as “Jane Smith is a terrorist.” That's a fundamentally flawed approach to assessing which campaign is “more negative.” And, given the nastiness of McCain's ads, not to mention their frequent inaccuracy, it's an approach that favors McCain.

But, fine. We'll play along.

Tumulty's conclusion is an awfully generous reading of Goldstein's statement, which noted that 47 percent of McCain's general election campaign ads have been negative, compared to 35 percent for Obama. (27 percent of McCain's ads and 25 percent of Obama's have been contrast ads.)

Not only is Tumulty's conclusion generous to McCain, Goldstein seems to be as well.

According to Goldstein, 39 percent of Obama ads have been positive, 35 percent negative, and 25 percent contrast.

According to Goldstein, 26 percent of McCain ads have been positive, 47 percent negative, and 27 percent contrast.

So, that's pretty clear: a significantly larger portion of McCain's ads have been negative. A significantly larger portion of Obama's ads have been positive.

Yet Goldstein asserts:

But, Obama has aired over 50,000 more ads than McCain. So, hasn't he simply aired more of everything - including negative ads - than McCain has this year, or than anyone in history, as McCain may have alleged?

If one just looks at pure airings of negative ads, McCain has aired more than Obama. If one allocates contrast ads as half positive and half negative or considers contrast ads as negative - as the Advertising Project does - the tone of the McCain and Obama campaigns has been absolutely identical.

It isn't entirely clear what Goldstein is trying to say, but it appears he's saying that the two campaigns have run the same number (not percentage) of negative ads (including contrast ads.)* And that because they've run the same number of negative ads, their “tone” has “been absolutely identical.”

That seems reasonable enough -- until you remember that the campaigns have run positive ads, too. And that Obama has run many more positive ads than McCain. If they've run the same number of negative ads, and Obama has run many more positive ads, obviously their tone has not been “absolutely identical.” McCain's has been more negative.

So it may be accurate to say they've run the same number of negative ads -- Goldstein's statement doesn't include the raw numbers necessary to make such a determination, but that's what he seems to assert. But it does not follow that their tone is identical; Goldstein's conclusion is simply illogical.

* If this isn't what Goldstein's saying, his conclusion that the “tone of the McCain and Obama campaigns has been absolutely identical” is self-evidently false, as Goldstein's own numbers show that a significantly larger proportion of McCain's ads have been negative.

UPDATE: In the comment thread attached to her post, Tumulty writes: “Ah, just saw the Media Matters piece on this ... I'll go with Wisconsin on this.”

Here's the thing: “Wisconsin's” conclusion is demonstrably false; fatally undermined by their own data. Choosing to “go with Wisconsin” on this means choosing to be wrong. The Wisconsin project's data indicates that a significantly higher proportion of Obama's ads have been positive, and Obama has run many more positive ads. Concluding from that that the “tone” of the two ad campaigns has been “identical” is simply wrong. This isn't a matter of interpretation. It's right there in the project's data. And in the definition of “identical.”

So the question now is: Why does Karen Tumulty prefer to stand with a false conclusion that makes John McCain look better?