Why Obama Won: Campaign Dogs that Didn’t Bite, Part 2

Following on my earlier post, I want to examine three more moments in the campaign that Dan Balz and Haynes Johnson see as significant. Indeed, I think many observers thought these moments “mattered.”

The first is Obama’s “bitter voters who cling to guns” comment. of April 6. Axelrod tells Balz and Johnson:

It was mildly damaging. But it came in the same period as the Reverend Wright stuff, which I think was much more damaging. At the end of the day…I think he fought his way through it and they overplayed it.

The second is the April 16th debate between Obama and Hillary, when Obama was asked about not wearing the flag pin, “bittergate,” Jeremiah Wright, and William Ayers.

The moderators [Gibson and Stephanopolous] drew considerable criticism for their questions, especially from Obama’s network of supporters around the country. But Obama suffered most from his performance…Given the state’s demographics and the solid working-class voters Clinton continued to attract, Obama was almost destined to lose Pennsylvania. But after the Wright controversy and the damage from his ‘bitter’ comments, the debate slowed his movement and turned the primary wholly in Clinton’s favor. (p. 209)

The third is Jeremiah Wright’s interview with Bill Moyers on April 25 and his National Press Club speech on April 28:

After Wright’s reappearance, Obama’s numbers plummeted in Indiana. (p. 211)

Here is a graph of the national primary polls from 2008. I have put vertical lines for a slew of events, including the major state primaries and Super Tuesday, the first Wright hullabaloo, the “bitter” comment, and the debate. There are two lines denoting the Wright’s two media appearances in late April.


The graph shows that Obama’s standing shot up after the earliest primaries, particularly after South Carolina (although the smoothed trendline somewhat obscures the apparent abruptness of the trend at this point). By mid-February, a sort of stasis emerged. This looks a lot like what Simon Jackman and Lynn Vavreck describe in the paper I discussed earlier: lots of momentum (Big Mo) replaced by little momentum (Slow Mo).

This stasis is important. Simply put, there are no meaningful changes in the candidates’ standings after about March 1. According to the smoothed trendlines, Obama’s standing ranges between 46% and 50%. Clinton’s ranges between 42% and 45%.

Obama is not “damaged” by the Jeremiah Wright scandal. By his comments about “bitter voters.” By the debate. By Wright’s reemergence. I said this before when Wright and “bitter voters” hit the news, but now let me shout it: ALL OF THESE EVENTS BASICALLY HAD NO IMPACT ON VOTERS NATIONWIDE.

The state-level polls that Balz and Johnson refer to tell a similar story, or non-story. Here is Pennsylvania, where Obama’s performance in the debate “slowed his movement”:


One problem is that Obama wasn’t “moving” before the debate. In the two weeks before the debate, there were 20 polls by 13 different pollsters. Those polls put Obama’s vote share between 37 and 45%. The trendline doesn’t move. It sits right at 41%. After the debate, you can see a little 1% downturn in Obama’s standing and a 2% upturn in Clinton’s. Perhaps that’s enough to persuade Balz and Johnson that the debate mattered. But if it mattered, the debate “turned” very little.

Finally, here are the Indiana polls. This graph begins later because there was only one poll before the end of March.


Because some points lie on top of each other, you cannot quite tell that there are three polls just before Wright’s April 25 interview and then three in between the interview and his speech on April 28. So the data is a little thicker than it appears. The question is whether Obama’s standing “plummeted” after Wright’s “reappearance.”

The trendline doesn’t conform to this interpretation. After Wright’s interview, Obama’s “plummet” was equal to about 2 percentage points. Clinton’s standing increases by this same amount. That doesn’t strike me as altogether momentous. Moreover, soon after Wright’s speech, Obama gained back what he had lost over the subsequent two weeks. So ultimately, it’s not clear that Wright’s “reappearance” had any systematic effect. At best there were only small and temporary consequences.

We are left, then, with a picture of the 2008 Democratic primary in which the vast majority of ink was spilled about events that mattered little, if at all.

[Many thanks to Chris Mitchell for putting together the polling data used in this and the previous post.]

3 Responses to Why Obama Won: Campaign Dogs that Didn’t Bite, Part 2

  1. James April 1, 2010 at 11:55 am #

    Thank you! Thank you for using real data to document what many observers *outside the beltway* have been saying. These minute-to-minute obsessions-predictions-analyses only matter in the beltway, where politics takes the place of baseball or basketball as the stuff of minor obsessions.

    This proves, of course, that the profession of the pompous political analyst/journalist is pretty much without measurable value.

    You hear that, Dan Balz and Dana Milbank?

  2. Ben Bishin April 2, 2010 at 10:24 am #

    These are terrific John. They also raise an interesting question. While it is understandable why Balz and Johnston want to attribute significance to these events (at a minimum it heightens the relevance of their horse race coverage), from reading the books (e.g., Game Change) it certainly seems as if the campaigns themselves see these things as important. Are Axelrod et al. misguided? or, perhaps, do they live in perpetual fear that these things could matter, and hence act as if they do in order to avoid making big mistakes?


  1. Further Thoughts On The Resonance And Relevance Of The Etch-A-Sketch Meme - March 23, 2012

    […] instance, when George Washington University political scientist John Sides looked at five events from the 2008 campaign identified by the journalists Dan Balz and Haynes Johnson in […]