The CBS News Horserace blog posted a faulty analysis of the exit polls from election day. In the analysis, their polling expert, Jennifer De Pinto, wrote:

As voters left the polls on Election Day, many were asked how they would have voted if the election match-up were between Hillary Clinton and John McCain rather than Barack Obama and McCain. 52 percent said they would have backed the former Democratic candidate; 41 percent would have voted for McCain, wider than Obama’s 7-point margin over McCain.

Interestingly, 16 percent of McCain voters said they would have voted for Clinton, the Democrat, if she had been her party’s nominee.

Now if De Pinto had simply explained the problems with using the exit polls and stated that we couldn’t draw full conclusions about the final vote if Hillary had been the nominee, then it would have been an interesting analysis of those folks who voted on election day who think that they would have voted for Hillary Clinton instead of John McCain. But by implying that these numbers show that Hillary Clinton would have won the election by a wider margin, a narrative has been created that others are repeating, including by David Shuster on MSNBC Thursday afternoon.

The problems with such a conclusion?

1. This analysis has Hillary Clinton with 52% of the vote and McCain with 41%. In the actual vote count, Obama has 53% to McCain’s 46%, so the question is what has happened to the 5% difference in the McCain vote? We don’t know from the post whether these people are simply undecided or voting for a third party. The undecideds could split evenly or go more for one candidate or another, so we really don’t know the margin.
2. These numbers are based on Exit Polls from election day. Over 31 million people early voted, or about 25% of the total of total presidential votes. We know that the early vote favored Obama by wider margins than of those who voted on election day. How do we know? A CBS News poll that shows the early vote went for Obama by a wide 57-38%. These more loyal Obama voters were not part of the Exit polls and their responses could easily change the outcome.
3. We don’t know where those Hillary Clinton crossovers live. If these voters are in California and other states where Obama won by double digits, it certainly would not have affected the outcome. Would Hillary Clinton have won North Carolina or Indiana? I would be skeptical, especially because of the next point.
4. Republican turnout was lower than expected. For example, Obama actually received roughly the same number of votes in Ohio as did John Kerry, but McCain received over 300,000 fewer than George W. Bush did in 2004. It is feasible that with Hillary Clinton at the head of the Democratic ticket, Republicans would have been more motivated to vote — and vote against a Clinton.
5. Hillary Clinton, having ended her quest at the end of the primary process, did not have to endure several months of negative commercials by the Republicans and associated groups. We do not know what kind of effect this propaganda would have had.

Now I want to emphasize that had Hillary Clinton been the nominee I think she would have won as well, and I for one would have voted for her. But this is about accuracy in reporting and poll analysis, and in this instance the CBS Horserace blog fails.

WHAT'S NEXT?

Subscribe via RSS or Email  ::     ::  Leave a comment

Bookmark to: