Over in The Corner, there’s been discussion of “the Bradley effect.”

I was checking around to find out more about the 1982 election and came across this great article in Time magazine on the poll problems that were nationwide, not just in the California gubernatorial election. An excerpt:

Many of the national pundits relied on regional newspaper and TV-station polls, which made some notable goofs. Surveys by the New York Daily News, Long Island’s Newsday and a subsidiary of the Gannett News Service all gave Democratic Gubernatorial Nominee Mario Cuomo a final-week lead ranging from 9% to 13%; he won by 3%. The Illinois Governor’s race brought on the worst polling failures: the Chicago Tribune found Republican Thompson ahead, 53% to 34%, and the Chicago Sun-Times had the race barely closer, 55% to 40%.

(Thompson won, but of the tiniest of margins)

The most tangled polling errors came in California, where almost no one forecast Republican George Deukmejian’s 50,000-vote victory over Tom Bradley. Indeed, the Los Angeles Times ran a frontpage story on election morning about the lineup of local politicians vying to succeed Bradley as the city’s mayor. The San Francisco Chronicle’s first election extra bannered: BRADLEY WIN PROJECTED. While ABC was predicting Deukmejian’s victory, its affiliate stations in Los Angeles and San Francisco were using exit polls of their own to call the race for Bradley instead. In the Senate contest, several affiliates used local exit polls to forecast a victory for Democrat Jerry Brown, who actually lost to Republican Pete Wilson 51% to 45%.

That last sentence is what’s important. It shows a Republican surge in California, not predicted in the polls, that would have boosted Deukmejian as well as Wilson. To say that Bradley lost because he was black seems to be a re-writing of history to explain away shoddy media coverage.