Following up on my earlier post about poll numbers, there’s one point I left implicit where I should have stated it explicitly.
I said before that partisan polls (ones commissioned by one of the candidates or the candidate’s party) are inherently suspect. The reason is not that these polls are necessarily flawed or cooked, though they can be. The problem with them is that a partisan poll only gets released when it’s good for the candidate who sponsors it.
So, hypothetically, a candidate could commission three polls, get a range of results and only release the one which shows him doing the best, a reading that’s likely an outlier and thus misleading.
Of late, we’ve been seeing lots of (D) partisan polls. And that’s, as I said, because the Dems are the ones who have an interest in showing how close a lot of these races are.
The point I didn’t mention is this: the normal response when one candidate puts out a poll favorable to him or herself is for the other side to go into the field and (if they can) and get a better number to release. And back in July they did just that. According to Roll Call (sub. req), the NRCC dropped $450,000 to conduct polls in 28 competitive districts. The article was dated July 31st. And the polls were conducted “over a two-week period this month.” In other words, the polling was almost certainly done at least a month ago.
So far as I can tell, I’ve seen few if any of those polls. And it’s not hard to figure out why.
Cricket, cricket …
