Fooled by exit polls.

Byline: The Register-Guard

According to the Election Day exit polls, President John Kerry would lead a nation whose voters care above all about moral values.

The exit polls' prediction of last Tuesday's result has been deflated by the fact of President Bush's victory. Yet those same polls' reading of voters' motivations has attained the status of conventional wisdom. Karl Rove, the president's political adviser, placed moral values third on Bush supporters' list of concerns - and he has a better record than the exit pollsters.

The exit polls were so far off that they came close to recreating some "Dewey Defeats Truman" moments. By mid-afternoon Tuesday, when the data pointed to a win for Kerry, the Massachusetts senator was advised to begin crafting a victory speech. Drug and defense stocks fell on Wall Street as news of the exit polls' findings circulated on the Internet. Pro-Bush radio talk show hosts, scanning leaked poll results on the Drudge Report and other sites, took on a funereal air. British Prime Minister Tony Blair went to bed believing Kerry had won.

All this happened before the only polls that really matter - the polls where voters cast their ballots - had closed.

The first real results made it clear that there was something wrong with the exit polls. Rove spotted the problem right away: The exit polls, he said in a Fox News interview, had Bush losing by 19 percentage points in Pennsylvania and by 17 points in New Hampshire. Though Kerry ultimately won both states, the actual results were much closer and Bush was the clear winner nationwide.

Exit polls are conducted by interviewers who approach voters as they leave their polling places. A growing problem is inherent in the technique: Increasing numbers of Americans don't cast their ballots at polling places. Exit polls are impossible in Oregon, where all ballots are cast by mail. They're only slightly less problematic in many other states, where the rate of absentee voting approaches 50 percent.

What's more, the first exit poll interviews are unavoidably conducted early in the day, creating a skewed sample. Certain types of voters are less willing to be interviewed than others. The result is a widening gap between exit polls and reality.

In contrast, the accuracy of pre-election polls was widely demonstrated last week. Most showed Bush with a lead of a few percentage points, but pronounced the race a statistical tie because the difference was within the polls' margin of error. Polls for The Associated Press, the television networks and national newspapers came close to reflecting the actual result. The final poll conducted by the Pew Research Center for the People & the Press hit the nail on the head, with Bush at 51 percent, Kerry at 48.

Despite the drubbing they received, the exit polls have shaped the national discussion of the election results. In particular, the finding that more than one in five voters supported Bush because of "moral values" has launched a thousand post-election analyses.

Rove interprets the results differently. He put "moral values" in third place, behind the war in Iraq and economic issues such as jobs and taxes.

The phrase "moral values" is generally understood to be shorthand for such issues as gay rights and abortion, but could be stretched to encompass everything from Michael Moore's films to stem-cell research. The category's breadth limits its usefulness in explaining voter behavior. If Rove is right, voters had more specific, concrete issues in mind last Tuesday.

Recent experience has shown that when it comes to understanding voters, it's not a good idea to bet against Rove.