The scientific weakness of the Zogby/Wall Street Journal/Battleground States poll, which I have posted about several times, is finally getting some scrutiny. The link is to a piece in the online version of the Columbia Journalism Review and mentions my criticism as well as that of Rutgers political science professor who actually knows something about statistics. My gripe in a nutshell: The sample is not random; Zogby solicits people to sign up on its Web site and polls that database.
So, guess who is signing up? Supporters of the candidates! U.S. Senate candidate Barbara Ann Radnofsky’s Web site appeals to supporters to sign up for Zogby. (Republicans allege that Bell is doing the same, but I couldn’t find any evidence on his Web site.) This explains why Kay Bailey Hutchison’s lead is so much narrower in the Zogby poll than in Rasmussen. Indeed, Radnofsky’s campaign manager signs his e-mails with such an appeal (see below). I don’t blame Radnofsky for trying to artificially inflate her numbers. I blame Zogby for devising an artificial poll.
Help Barbara Ann Win on November 7
2. Sign up for Zogby interactive polling at<http://interactive. zogby.com/ pollregistration /registration/>http://interactive. zogby.com/ pollregistration /registration/
3. Sign up for our campaign’s Yahoo user group at<http://groups. yahoo.com/ group/Radnofskyc ampaign/>http://groups. yahoo.com/ group/Radnofskyc ampaign/ to receive updates and action alerts for volunteers