Monday, January 26, 2004

Chrisishardcore defends his estimates of the one-day results fromt the ARG tracking poll:

ARG's numbers are ultimately probably going to be wrong, so are Zogby's and so are everyone's. With two days left until the primary these 3 day tracking polls won't be able to accurately show late surges -- and that's what I'm trying to do.

But as I said in my earlier posts, even if you see a surge in the last day of a 3-day tracking poll, this could just be a routine blip in the margin of error--and blip are quite common with a margin of error of approximately +/- 7 percentage points. In other words, that late surge might turn out to be nothing. I'm not saying that Dean, or any other candidate, is or isn't surging. What I am saying is that absent some surveys with large samples, we just don't know. If Chrisishardcore's analysis turns out to be correct, it's more from luck than anything else.

No comments: