From Manjoo’s Salon article: “For instance, in the Bush strongholds — where the average completion rate (of exit poll surveys) was 56 percent — it’s possible that only 53 percent of those who voted for Bush were willing to be polled, while people who voted for Kerry participated at a higher 59 percent rate. Meanwhile, in the Kerry strongholds, where Mitofsky found a 53 percent average completion rate, it’s possible that Bush voters participated 50 percent of the time, while Kerry voters were willing to be interviewed 56 percent of the time. In this scenario, the averages work out to the same ones Kennedy cited: a 56 percent average response rate in Bush strongholds, and a 53 percent average response rate in Kerry strongholds. But in both Bush strongholds and Kerry strongholds, Kerry voters would have been responding at a higher rate, skewing the poll toward Kerry.”
From Michael Collins’s Scoop critique of Manjoo: “This critical paragraph consists of simple verbal calculations, plus or minus three. There is nothing else there except Manjoo’s words surrounding numbers which conveniently counter the statistical and mathematical analyses Kennedy cites. This is simply amazing. Meaningless words and numbers are produced to refute Kennedy’s sources without any basis whatsoever. None. The run on sentences above are based entirely on the phrase “it’s possible.”
Jonathan Simon comments:
Of course the illogic of the Manjoo math above is even worse than Collins’s description of it. The averaging that Manjoo indulges in (59K + 53B / 2 = 56 in Bush strongholds and 56K + 50B / 2 = 53 in Kerry strongholds), would be possible only if there were an equal number of Kerry and Bush voters–that is, if K = B–in each set of strongholds. By definition, the strongholds represented unequal (in fact >80/20) divisions of Kerry and Bush voters (hence “strongholds”), so Manjoo’s facile averaging is false, and is in fact mathematical nonsense, as would be obvious to a pre-algebra middle school student. It was in fact extremely difficult, if not impossible, to make the data given by Edison/Mitofsky fit the “reluctant responder” hypothesis, though the debate on whether it was technically possible or not rested on genuine mathematical analysis of a fairly high order, rather than the garbage thrown into the pot by Manjoo.
What is particularly aggravating about this is that legitimate calculations of actual rigor have been being carried out by many of us since literally seconds after the release of the relevant data in January 2005, grappling with the fairly complex mathematical challenges posed by an incomplete data set. All this was public and available to Manjoo, had he cared to look it up. For him to attempt to disparage Kennedy’s analysis (and that of the rest of the analysts who have worked long and hard) by using simplistic, false and inane arithmetic (for pete’s sake), reveals either an unconscionable shoddiness of analysis and argument or a driving motive to disparage without letting either facts or basic math get in the way (and unfortunately this shoot-from-the-hip-while-talking-on-a-cell-phone utter absence of rigor characterizes Manjoo’s attack at many other points as well).
In either case, Manjoo owes not only Kennedy, but the rest of us who have devoted our energies to this task, a retraction and an apology.