What Internet Polls Are Good For


In 2002, when the Modern Library announced its list of the 100 greatest novels, it invited online readers to submit their picks as well. Two groups rose to the challenge: the Randians and the Scientologists. Where the official top three consisted of Ulysses, The Great Gatsby, and A Portrait of the Artist as a Young Man, the people's house picked Atlas Shrugged, The Fountainhead, and Mitt Romney's fave, Battlefield Earth.

That's the way Internet polls work. When the Modern Library turned its attention to nonfiction, The Virtue of Selfishness finished first and Dianetics was the runner-up; John Lott's More Guns, Less Crime made the top 10. (Some folks seemed to be voting methodically for everything in the Laissez Faire Books catalog.) In 1996, the Libertarian Party's presidential nominee, Harry Browne, did very well in various online surveys before wrapping up the election with a tiny 0.5% of the popular vote. And just this week, as I noted yesterday, one poll pronounced Fred Thompson the winner of the latest Republican debate even though Thompson wasn't actually on the stage.

This is why, despite all the emails I've received urging me to write about Ron Paul's strong performance in the Internet polls, I haven't been covering it. I like Paul, but Internet polls are meaningless as a measurement of anything but the enthusiasm of a candidate's supporters. I don't think, as some do, that Paul's performance is purely a product of cheaters spamming sites with multiple votes. There has been some of that, but the congressman does well even when the multi-voters are ferreted out and their ballots removed from the results. I just don't think it means a lot to win one of these contests.

But I have to laugh when the creators of these unscientific surveys try to find ways to discount Paul's wins without admitting the polls themselves are near-useless. When it became clear that Paul was doing well in Fox's text-messaging poll after the debate Tuesday night, for example, Fox host Carl Cameron suggested the congressman's supporters were gaming the system. He did not pause to ponder the point of offering a system so easily gamed. Nor did he admit that if the votes for Paul didn't mean much, the same was true of the remainder of the results.

Now the superhawks at Little Green Footballs have taken Paul out of their post-debate poll because—well, I'll let the site itself explain it:

They aren't "cheating," as in voting multiple times, but they have sent out emails and posted the link to our poll at several spots on the web, urging people to go vote for Paul. The end result is the same—the poll results are skewed, and it's not an accurate measure.

Internet polls are not scientific anyway, but when the gaming is this obvious I'm not going to let it slide, or to let our poll be misused by supporters of a man who, in my opinion, is nuts.

Unlike Fox, LGF is frank about the scientific quality of its survey. What it doesn't recognize is that the Paul backers' "misuse" of the poll is arguably the only legitimate use it has. I will say it again: Internet polls are meaningless as a measurement of anything but the enthusiasm of a candidate's supporters. If you're kicking out the enthusiasts, what's the point?

The Modern Library's polls might not tell us anything about what books are worth reading, but they do tell us a lot about which authors have especially devoted followings. These presidential polls won't tell us who's going to win the primaries, but they do give a good sense of who will still be running after the first few primaries have passed. You shouldn't expect Ron Paul to be his party's nominee next year. But you can expect him to be a thorn in the side of the big-government candidates for a long time yet.