Wednesday, September 04, 2002

LATEST TRIB STORY: The problem with polls

My chief worry is that the evidence in this story supports the pollsters while only a hunch supports my slant--that polls are problematic. Pollsters measure everything by numbers, and four or five distinguished stats profs I interviewed for the story say the numbers just don't support my misgivings that response rates automatically skew poll results. And I did let them make their case in the article. But I still feel uneasy about how much weight we give polls when the logistics of poll-taking have been so recently challenged by technology and social shifts. As Professor Brehm seemed to say, that's just common sense,

I got the idea for this story from this William Safire column. Which is good, but I cleared up a couple things he screwed up. I fear, though, that I did little to improve on this fine Boston Globe story (if link is busted, look for it here, except add the takes of some major names in the stats world who work in Chicago.

Recent Trib stories:
'The Note'
My Tribune archive

No comments: