L.A. Times Spins Its Poll

There has been a lot of publicity about the dueling Field and L.A. Times polls on the California recall election. Dafydd ab Hugh offers his take on the Times’ poll, which has been more favorable to the Democrats:
“[T]he real story is buried in the fourth paragraph…and handwaved away (by improper use of statistics): that even in the LA Time’s own skewed poll, Bustamante’s lead dropped from 13% in mid-August to only 5% now.
“The handwaving consisted of this line: ‘Schwarzenegger’s support ticked upward three percentage points, from 22%. Bustamante dropped five points. All of the shifts, however, are near or within the poll’s margin of sampling error.’
“But of course, it’s a single poll — so you don’t get to add the margins of error like that. And what is this about a shift being ‘near’ the MOE? Bustamante’s drop of five points is 167% of the margin, while McClintock’s rise of six points is TWICE the margin! Schwartzenegger’s rise of three points is at the outer edge of the MOE… but when you combine it with Bustamante’s drop, the Democrat’s lead dropped by 8% — which is 267% of the margin of error.
“That is very significant movement… and remember, we’re talking about the LA Times’s own poll. The Stanford University poll and SurveyUSA both show Schwarzenegger with double-digit leads over Bustamante, and with the recall leading by twenty-five points, nearly two to one.
“As usual, the real story is how the Times buries the real story.”


Books to read from Power Line