Tuesday, 7 January 2003

Statistical noise = fact

Daniel Drezner posts that John Zogby's overhyping his own numbers on the Democratic contenders. Trying to read anything into single questions in a survey with under 500 respondents is problematic at best, and it's downright foolhardy to draw any conclusions out of marginals that show virtually everyone in the poll in a statistical dead heat. Meanwhile, Jacob T. Levy speculates that Richard Gephardt will bomb spectacularly in New Hampshire.

Ah well, at least Atrios isn't yet predicting that we won't have a 2004 election. Maybe next week.