The Wilder Effect

The Wall Street Journal has an interesting article about minority candidates in close races. Some call it the 15% lie. Others call it the Wilder Effect. The idea is basically that when a pollster calls people on the phone to ask whom they will support in the election, some people will tell the pollster that they will support the minority candidate (even though they will really support the white candidate) just because they do not want the pollster to believe they are racist. For example, consider the Tennessee Senate race where democrat Harold Ford, an African American, is running against republican Bob Corker, who is white. If a democrat who is white receives a phone call from a pollster, he may feel the pollster is judging him as racist if he claims to be a democrat who is also supporting Corker. As a result, minority candidates typically underperform their poll numbers in the general election. In a famous race in November 1989, democrat Douglas Wilder, an African American then serving as Virginia lieutenant governor, ran for governor against Republican attorney general Marshall Coleman. Exit polls (face to face interviews after voters leave the voting area) during the day of the election showed Wilder ahead by 10%. In the end, Wilder won the race by only 0.38%. Why were the exit polls so wrong? Most assume it was the "Wilder Effect" at work -- which may be even stronger in face to face interviews (versus telephone surveys). This year, the "Wilder Effect" may be witnessed by democrat Harold Ford in the Tennessee's senate race, republican Michael Steele in the Maryland senate race, and Kenneth Blackwell in the Ohio governor's race.

Just to be clear - the "Wilder Effect" does not occur because voters are racist. It occurs simply because they do not want to appear as racists to the interviewer. I'm guessing this effect may not be a factor in uncompetitive races - such as Alan Keyes disastrous Illinois senate campaign in 2004.