Monday, October 30, 2006

Grad School Kool-Aid Drinkers Make Bad Pollsters

Michael Barone has a nice piece on the election coming up that notes that polls are a very imperfect art form:
Fewer people vote in off-year elections than in presidential years. In 2002, 75 million people voted. In 2004, 122 million did. My hunch is that people who identify themselves as independents are substantially less likely to vote this year than people who identify as Republicans or Democrats -- which would be good news for Republicans, since independents give Bush low job ratings. Another hunch is that the Republican turnout apparatus, with which the Democrats haven't yet caught up, will boost Republican turnout as it did in 2004, and that the resulting electorate will be more evenly divided in party identification than the electorates shown in most of the public polls.

Serious pollsters concede that there are some problems with polling. Americans have fewer landline phones than they used to, and the random digit dialing most pollsters use does not include cell-phone numbers. Larger and larger percentages of those called are declining to be interviewed.

Interviewers can inject bias in the results. The late Warren Mitofsky, who conducted the 2004 NEP exit poll, went back and found that the greatest difference between actual results in exit poll precincts and the reports phoned in to NEP came where the interviewers were female graduate students -- and almost all the discrepancies favored the Democrats.

Traditionally, the Harris Poll tilts Dem on an average of several basis points, due to methodologies designed to elicit responses favoring Dems. Ditto several other polling organizations. The reason Dems hate Rasmussen is that it's honest, and does not have ditzy grad students or other dumbed-down hires doing the polling. Hence the numbers reflect actual facts, which Dems like Dave Letterman do not recognize, but claim others [Repubs] make up.

No comments :