Comments on: Barack Obama vs. Hillary Clinton March 4 Primary Early Call

By: Susan Murphy

Susan Murphy — Wed, 05 Mar 2008 01:01:04 +0000

What I think is cool about your theory, Chris, is that you are using freely accessible information to make your case. As tools become more sophisticated, will we really need CNN to “make the call” in situations like this?

A few years ago here in Canada, we had some issues with our federal election in that people in the Maritimes were posting their election results when their polls closed, when it was still 2 1/2 hours till the West Coast’s polls closed. This was despite a ban from Elections Canada forbidding anyone to post this information. Of course people found ways around it. How can this kind of information be controlled, really. Not easy.

I think your research is very interesting, and it will be interesting to see how it plays into the overall results…good work!

By: Joel

Joel — Wed, 05 Mar 2008 00:58:13 +0000

The honest truth is that if Google predicted winners through it’s trend tracking then Ron Paul would basically have already won the election. Using these trends and relating them with actual poll results does not work.

By: financialaidpodcast

financialaidpodcast — Wed, 05 Mar 2008 00:43:40 +0000

Chip and Dennis – honestly, this is all back of the envelope stuff. My thinking on the topic is that Obama is leading search volume and Clinton is paying, paying, paying via AdWords’ estimator. By no means is any of this scientific at all. That said, if people are searching for Obama and not Clinton, that means either they know nothing about Obama and are doing research, or they already have their minds made up about Clinton, or both, or neither.

If I had more free time, I’d do an analysis of Super Tuesday I to see what search volume correlated with what wins. In this case, we’re not claiming causation, just correlation.

By: Dick

Dick — Tue, 04 Mar 2008 21:30:06 +0000

Thanks. I think you are right on!

By: Chip Griffin

Chip Griffin — Tue, 04 Mar 2008 21:03:57 +0000

Further, I should add that by tracking search volume you are actually looking at who wants to know more about a candidate. There are plenty of subjects I may be curious about, yet that doesn’t mean I support them.

By: Dnenis McDonald

Dnenis McDonald — Tue, 04 Mar 2008 21:01:36 +0000

What did you say was the relationship between search volume and voting patterns?

By: Chip Griffin

Chip Griffin — Tue, 04 Mar 2008 20:59:51 +0000

The problem with your analysis is that Hillary was up in the polls at the start of the trend charts, but these charts don’t reflect that, nor do they reflect a significant change in the spread over time.