Category — Surveys

Thanks to Olivier de Gaudemar at Insights 2.0 for finding and posting this clip of John Stewart grilling John Zogby over the polling disaster from the New Hampshire primaries a couple of weeks ago.

I think this is probably a market researcher’s version of hell: having John Stewart publicly questioning your data and methodology live on TV in front of millions of viewers.

Since 2007 is almost over, and I have yet to devote a blog post to any of Technorati’s hot topics, I might as well have a post about Republican Presidential candidate Ron Paul and online polling.

This chart from an AOL online straw poll is a window into the problems with online surveys. According to most other nationwide polls, Ron Paul is a distant 6th. In fact his awareness among the broader electorate is fairly minimal.

However, on the web, he is a powerhouse to be reckoned with, mostly due to his supporters. And while I believe engaged web users may be more likely to aware of Dr. Paul, the numbers still look a little fishy.

Which brings me back to sampling and online surveys in general. While methods like random digit dialing were able to provide telephone polls with method of avoiding selection bias in sample bases in the past, as AOL is finding out the hard way, surveys on the web still need a bit more work.