The WaPo Poll viewed from a height

Our friend Joshua Sharf, on his View From a Height blog, argues that, if one corrects the Washington Post poll for registration bias, the gap between President Bush and Senator Kerry closes to within the stated margin of error. Here’s Joshua’s e-mail to me on the subject:
“I already emailed Rocket about this, but I suspect you’re up and at work, where Rocket is probably sleeping off his scotch-induced bender after seeing the latest WaPo Polls.
“Here’s some comfort: if there’s one thing that I’ve learned about polls from you guys, it’s to look at the demographic breakdown. And to look for obvious outliers. (OK, that’s two, and since this email is about numbers, I should get that part right.)
“Since the Post didn’t provide these numbers, I used Excel’s Solver to tease them out of the data (one line at a time, since rounding error meant a simple matrix solver couldn’t work). Here’s the deal:
Dems: 37%
Reps: 30.6%
Inds: 32.4%
“Remember, this is from their data, so their poll numbers haven’t been adjusted to reflect the country’s actual partisan balance.
“Moreover, the poll has Kerry winning the South 55-40. Does anyone really think that Kerry is going to win the South by 15 points? Or at all? What, did they poll Castro and project the results out to all of Cuba, and then count Cuba?”
Joshua attached his spreadsheet and invited us to examine it ourselves, but I’ll leave it to the experts on our crew (the other two guys) to perform that task.

Notice: All comments are subject to moderation. Our comments are intended to be a forum for civil discourse bearing on the subject under discussion. Commenters who stray beyond the bounds of civility or employ what we deem gratuitous vulgarity in a comment — including, but not limited to, “s***,” “f***,” “a*******,” or one of their many variants — will be banned without further notice in the sole discretion of the site moderator.

Responses