This is what passes for polling analysis in the mainstream media -- from USA Today:
Bush's supporters say his determination helped him win a second term. A Los Angeles Times poll taken a year ago found that 56% of voters said Bush was "too ideological and stubborn." But on Election Day, surveys of voters found that of the 17% who said they voted for the candidate they thought was a strong leader, 87% voted for Bush.
Did it occur to Keen and Kiely that these findings aren't inconsistent? We've known for decades that people's survey responses about candidate characteristics are closely related to which candidate they support (in the jargon, they're "endogenous"). That's why almost everyone in the 2004 exit poll who chose "strong leader" as the most important quality of the candidate they chose voted for President Bush -- it was one of the central rationales of his candidacy. But that doesn't necessarily contradict the finding that 56% of Americans think Bush is "too ideological and stubborn"; the two groups are likely to be mutually exclusive, or very nearly so.
Comments