A Tribune poll of 1,099 registered voters published Oct. 11 (Page 1) showed Peter Fitzgerald leading Carol Moseley-Braun by 13 percentage points (49 percent to 36 percent), with a claim of accuracy within “plus or minus 3 percentage points from results obtained if all likely voters in Illinois were questioned.”
But when all voters were questioned on Election Day, there was only a 4 percentage point difference (51 percent to 47 percent). Why were you so wrong? There have been references to a surge in minority voters, last-minute advertising, low voter turnout (perhaps a result of believing your polls?), etc., but no hint that the polling methods might be seriously flawed in claiming that 1,099 people can accurately reflect the opinion of more than 3 million voters.
The many election surprises nationally should encourage the media to re-examine their polling practices.
