Sunday, January 25, 2004

Mickey Kaus is touting blogger Chrisishardcore's estimates of the one day ARG tracking poll results. As I mentioned in an earlier post, these one day numbers are meaningless. Furthermore, Chrisishardcore's estimates don't match up with what ARG is reporting. According to ARG "Howard Dean is up 10 percentage points from his low on January 22, but he still trails John Kerry by 17 percentage points in the latest daily numbers. But Chrisishardcore estimates that on 1/24, ARG had Kerry with 42 and Dean with 20, a 22 point difference, not 17. Furthermore, Chrisishardcore estimates that on 1/22, Dean had 15 percent, so if his support increased 10 points, he'd have to be at 25, not 20. Finally, Chrisishardcore has Dean's lowest day as 1/23 when he was at 13 percent, not 1/22.

So to my warning about relying on one day tracking poll results, add a caveat about never ever relying on estimates of one day tracking poll results. If, as some said, the Zogby polls before the Iowa caucuses were "crack for the weak", then these estimates are "crank for wimps."

No comments: