« Vietnam Warnings Since 1975: Wrong Every Time? | Main | Weblog Change-- Notifier for Comments »

September 23, 2004

Candidates' Nightly Polling

Via Drudge, the Washington Times has an interview with Karl Rove that tells us about the extent of polling in presidential campaigns: ...


... Mr. Rove said he thinks the president is five to six percentage points ahead of Mr. Kerry nationally, although the Bush campaign is not conducting national polls. However, Mr. Rove said, "We have an army of pollsters" doing extensive sampling in battleground states.

"We've taken all the battleground states and molded them together so that we're doing 600 sample a night or 800 sample a night in every battleground state and then aggregating all of those," he said.

"So we're talking about literally interviews in the thousands every night," he said. "And you run three nights of those, and you're talking tens of thousands of interviews."

It makes sense not to do national polls, since they wouldn't affect a candidate's decisions. What is striking is that Bush is doing a *nightly* scientific poll in each battleground state. The public pollsters take three nights per poll of that size.

Two ideas:

(1) The candidates' actions might tell us more than the public polls we see, since the candidates see better and more frequent polls.


(2) Someone should press the campaigns to reveal their data after the election, just for scholarly and historical purposes. It sounds like great data.

Posted by erasmuse at September 23, 2004 09:42 AM

Trackback Pings

TrackBack URL for this entry:
http://www.rasmusen.org/mt-new/mt-tb.cgi/223

Comments

Post a comment

¡Comment registration is required but no TypeKey token has been given in weblog configuration!