Thursday, May 08, 2008

Innumeracy On the Front Page of the Washington Post

Did Rush Limbaugh Tilt Result In Indiana?
Conservative Host Urged 'Chaos' Votes

By Alec MacGillis and Peter Slevin
Washington Post Staff Writers
Thursday, May 8, 2008; Page A01

Even as Barack Obama's campaign celebrated Tuesday's primary results, aides charged yesterday that they would have had an even stronger showing were it not for meddling by an unlikely booster of Hillary Rodham Clinton: the popular conservative radio host and longtime Clinton family nemesis Rush Limbaugh.


It might be worth pointing out that according to exit polls, Clinton would have won in Indiana even if Republicans had not been allowed to vote in the Democratic primary.

Those looking for evidence of Limbaugh's influence pointed to Clinton's edge among Republicans in Indiana and North Carolina. In Indiana, 10 percent of Democratic primary voters described themselves as Republicans, a higher rate than in any state but Mississippi, and they went for Clinton by eight percentage points, according to exit polls.


OK using the fancy elite mathematical toll called multiplication, I conclude that, if the votes among Democrats and Independents had been equally divided between Clinton and Obama, Clinton would have won by 0.8%. Her actual winning margin was larger, so, according to exit polls, she would have won if the votes of Republicans had not been counted.

This seems relevant no ?

To argue that Rush won it for Clinton, one would have to argue that Obama would have won among Republicans by ten times (Clinton's margin of victory in percent - 8%) were it not for Limbaugh (I am using subtraction and my recollection of the margin, which Alec MacGillis and Peter Slevin helpfully give as an absolute number while they give self declared Republicans as a percent making it impossible to do the calculation without data not in the article such as the total number of voters, number of Republican voters or margin of victory as a percent of total voters). Certainly possible, but what is the reason to refrain from presenting the calculation (and maybe even compare the margin to Obama's overall average margin among Republicans before Limbaugh began his chaos campaign which Alec MacGillis and Peter Slevin helpfully describe without giving actual numbers (except for 7 and 8)

"By contrast, Obama won Republican voters, often by very large margins, in seven of the eight states where exit polls were able to report the group before the Texas and Ohio primaries on March 4, when Limbaugh first coaxed listeners to vote for Clinton."

OK not trusting my memory I ask the Google for Clinton's margin of victory as a percent. I think I recalled correctly, but the very last precincts to report went for Obama, because the final margin was 1.1 % so Obama would have had to win Republicans by 3% to tie Clinton.

That does mean that Clinton won because Republican voting behavior is different from what it was pre March 4. This could be Limbaugh or Republicans who thought up his scheme on their own or it could be that Republicans sincerely hate the person who they think will be the Democratic nominee, but there is quite a bit there.

No comments: