DEBUNKING THE ‘BRADLEY EFFECT’
By Ken Khachigian
Washington Post
article/2008/10/30/AR2008103002396.html?hpid%3Dopi nionsbox1&sub=new
Oct 30 2008
They call it "the Bradley effect."
Pundits and politicians speak of it in ominous tones. It surfaced in
New Hampshire in January, when Barack Obama’s eight-point lead on the
eve of that state’s primary dissolved into a shocking come-from-behind
victory for Hillary Rodham Clinton. Could it have been the Bradley
effect? Chris Matthews of "Hardball" and a host of other talking
heads thought so.
As Obama continues to hold a lead in the presidential polls against
John McCain, the specter of the Bradley effect still haunts the
race. It’s a reference to the 1982 California governor’s race, which
Los Angeles Mayor Tom Bradley, an African American, lost to state
Attorney General George Deukmejian even though a popular election-eve
poll showed Bradley ahead by a solid seven points. If Obama should
lose on Nov. 4, there are those who’ll maintain it was the Bradley
effect at work. Even in faraway Kenya, the Los Angeles Times found
a Nairobi choreographer to quote: "There’s this thing called ‘the
Bradley effect’ that we are all very afraid of."
Enough. This urban legend that holds that white voters may be
telling pollsters they’re voting for Obama while they’re secretly
harboring racial reservations about him deserves to be banished from
our political conversation. As a senior strategist and day-to-day
tactician in Deukmejian’s 1982 campaign, I’m happy to send it packing
once and for all.
There were several reasons why Bradley lost the governor’s race
in 1982 — and none of them had to do with race. In the last two
weeks of that campaign, Bradley was cruising through California on a
languid victory tour. Conventional wisdom and early polling had made
him smug and complacent. Nine days before the voting, a United Press
International story observed: "Los Angeles Mayor Tom Bradley hasn’t
even been elected governor yet, but Democrats already are talking
him up as a potential candidate for vice president in 1984." You
could hardly blame them. Deukmejian’s campaign manager had resigned
three weeks before Election Day, and the political obituaries for
the Republican candidate had become routine.
With our backs to the wall, the "Duke’s" campaign regrouped. We
got a large infusion of late cash from loyal supporters and shed
our defensive posture in favor of hard-hitting messages homing in
on Bradley’s two principal vulnerabilities: non-Angeleno antipathy
toward Los Angeles and the mayor’s "soft-on-crime" liberalism.
With a little more than a week left, I drafted copy for two new
television commercials. The first built on Bradley’s opposition to the
death penalty and California’s Victims’ Bill of Rights, both of which
had been overwhelmingly approved by state voters. Four former chiefs
of Bradley’s own police department had endorsed Deukmejian, the author
of California’s death-penalty statute and other tough-on-crime laws.
A second commercial sharply exploited the wariness that other major
California cities felt toward Los Angeles, something that surveys
by our pollster, Lance Tarrance, showed to be a sure vote-getter in
San Diego, the San Francisco Bay Area and growing suburbs across
the state. Hence the tag line: "We deserve a governor for all of
California’s cities, not just one." It’s noteworthy that no Los
Angeles mayor has been elected governor in modern California history.
But rural California wasn’t in line merely to reject Los Angeles. There
were two other central concerns. First, guns. Gun control advocates
had put an initiative to freeze handgun sales — Proposition 15 —
on the ballot. The NRA and the firearms industry raised millions of
dollars to run the "No on 15" campaign and, through California gun
stores, registered 300,000 new voters, few of whom were likely to
vote for gun-control advocate Bradley.
Add that to Bradley’s unpopularity among Central Valley farmers —
he’d supported the United Farm Workers’ grape boycott and couldn’t
escape being identified with vastly unpopular Gov. Jerry Brown —
and it’s easy to see why rural California flocked to the polls to
voice its opposition to his candidacy.
Finally, exit polls showing Bradley winning were skewed by
the unprecedented wave of absentee voters. In early September,
the state GOP apparatus had set in motion a campaign to promote
absentee-ballot voting — something quite common today but more unusual
a quarter-century ago. The party’s push contributed to more than a
half-million absentee voters, 50 percent more than in the previous
gubernatorial election. As so many other observers, Democrat as well as
Republican, have noted, Bradley may well have won with actual precinct
voters. But he was swamped by overwhelmingly Republican absentee
ballots counted late into the night and the early morning hours.
Analysts shouldn’t overlook an element of flawed polling that
contributed to the Election Day surprise. Tarrance continued his
tracking polls for the Deukmejian campaign right up through the eve of
Election Day. His final tracking poll was taken on Sunday and Monday
nights and showed Deukmejian within one point of Bradley, confirming
the steady gains we’d been making since taking the offensive in the
last two weeks. Mervin Field, whose firm was then the state’s gold
standard of polling, took his final poll over the weekend, including
both Friday and Saturday, the two days when it’s most difficult to
reach the most valid samples of voters. Field not only may have had
sampling errors, but his timing was also massively flawed and failed
to capture Deukmejian’s surging momentum.
Lost in all the shallow analysis of that gubernatorial campaign is
what might really be called "the Deukmejian effect." Less than three
weeks before election day, Field had released a poll analysis showing
that about 5 percent of voters were disinclined to vote for an African
American candidate. But the poll also found something else: "On the
other hand 12 percent of the voters say they are disinclined to vote
for a candidate of Armenian descent (which describes Deukmejian). In
the face of these findings, it would appear that Deukmejian’s Armenian
background could potentially act as a greater drag to his chances of
becoming governor than being Black works against Bradley."
Indeed, when poll callers recorded verbatim responses to the question
"What is the main reason you have for voting for/against George
Deukmejian?", an alarming number included angry attacks on the
candidate’s Armenian ancestry. Moreover, his was not exactly the most
"average"-sounding of names. Deukmejian’s six immediate predecessors
had been named Brown, Reagan, Brown, Knight, Warren and Olson. Race
aside, "Bradley" was a familiar name that fit the mold.
In the end, voters managed to sort everything out — and Deukmejian’s
ancestry and Bradley’s race may have canceled each other out. Polling
errors, absentee votes, gun rights activists, anti-L.A. sentiment
and Tom Bradley’s liberal positions all added to his narrow defeat —
by just a little more than 1 percent of the vote — by a better-run
campaign that created driving momentum in its final days. Any notion of
race as an issue was put to bed when Bradley sought a rematch in 1986,
and Deukmejian trounced him by 23 percent — the biggest gubernatorial
landside in California in the last half of the 20th century.
As Tarrance recently wrote, the "Bradley effect" is a "theory in
search of data." If we want honest debate about the role of race in
elections, it’s time to put a stake through its heart.
Ken Khachigian, a California lawyer, was a senior aide to President
Ronald Reagan.