COPIED AS READ.
"In 2000 Zogby drew maximum attention. During much of October, George W. Bush had small leads over Al Gore in most polls. But after the news on the Thursday before the election of Bush’s 1976 DUI, Zogby’s tracking polls showed movement to Gore and ended with him in the lead. Gore did indeed win the popular vote, and in some quarters Zogby was hailed as a prophet.
But not among many of his fellow pollsters. Zogby admits that some of his methods are unusual. Most pollsters use random digit dialing so they won’t miss those with unlisted numbers. Zogby says there is no political difference between people with listed and unlisted phones, and his interviewers call only listed numbers. “It reduces caller fatigue and unproductive interviews,” he says. Most pollsters place calls only in the evening, when most people are home; Zogby’s interviewers call all day long and, he says, reach people who aren’t reachable from 6 to 9 p.m.
“Zogby is not a reputable pollster,” opines Warren Mitofsky, head of Mitofsky International, one of the two firms conducting 2004 exit polls for the broadcast and cable news networks. “He is more a salesman and a self-promoter than a pollster.”
Most controversial is Zogby’s weighting of poll results not only by the customary demographic factors but also by party identification, to get what he considers the right percentages of Republicans and Democrats. Zogby says he also adjusts results according to his feel for what is happening on the ground. Preposterous, say most pollsters. Justified, say Zogby fans. “There’s artwork involved,” says Zogby. “I’m not a statistician, I’m a historian. I’m used to using soft methods.” On the day before the New Hampshire primary, Zogby reported that John Kerry was leading Howard Dean by just 31 to 28 percent–much closer than in other tracking polls. To get that result he included for the first time those leaning toward a candidate. Then on primary day, he announced that Kerry’s lead in the two-day tracking had risen to 37 to 24 percent–a big shift, reflecting undecideds who went for Kerry.
But that one day’s tracking was based on a small sampling of only about 350 respondents. Sneered Slate.com’s Mickey Kaus: “Here is that final Zogby New Hampshire poll describing a huge Kerry miracle surge, undetected by others, that just happens to bring Zogby into line with other polls.”
Zogby admits mistakes. “I blew some states,” especially the 2002 Illinois gubernatorial and Colorado Senate races. But, he brags, “There were a number of states I got right, and no one else did.” He knows that other pollsters look askance at his methods. But it’s also true that everyone in the political world and many in business are eager to see his latest results.
Also read the following:
-
https://fivethirtyeight.com/features/worst-pollster-in-world-strikes-again/ -
http://www.politico.com/story/2008/11/zogby-wont-duplicate-obama-poll-015829