1
“Turnout is going to make the difference in this campaign and the intensity is weighing
against Gray Davis. They are more than just angry – they’re absolutely apoplectic.” –
Republican pollster Frank Luntz (
Nissenbaum 2003
).
“Right now their choices are basically: right-wing crook, right-wing boob, supposed
moderate that nobody knows what he stands for, and a cigar-smoking movie actor
killer.” – Gray Davis pollster Paul Maslin (
Nissenbaum 2003
).
More than 9.4 million California voters cast a ballot in the October 7, 2003 recall
of Governor Gray Davis and the selection of a successor, Arnold Schwarzenegger. That
total of votes was equivalent to 61.2 percent of all registered voters – that is a turnout
10.6 percentage points higher than the 2002 general election that had re-elected Davis
and 24.8 percentage points higher than the last statewide special election in 1993.
1
How did this contest generate a comparatively large turnout when at first glance it
may have many of the qualities that would lead to it being relatively ignored by voters –
including off-schedule timing, lack of federal offices on the ballot, a 77-day campaign
interrupted by legal challenges, as well as a complex recall and replacement ballot with
135 candidates?
Perhaps it was the “apoplectic” anger, as described in the quote above by Frank
Luntz, toward the governor for an energy crisis, budget deficit, and increased taxes? Or
perhaps it was the compelling circus-like uniqueness of the cast of characters on the
ballot as described by Paul Maslin? Many intriguing questions follow.
Was the electorate that turned out distinctive and different from those that had
participated in recent elections? Did the presence of Schwarzenegger on the ballot
increase the rate of participation among the 18 to 34-years-old demographic that had
1
However, the recall turnout was 9.7 percentage points lower than the state turnout in the 2000 general
election and 20.9 percentage points lower than the 1972 general election. The recall turnout was also 3.8
percentage points lower than pre-election polling had estimated (DiCamillo and Field 2003a).
Convention