by Andrew Kohut, President, Pew Research Center
Special to the New York Times

John McCain’s decision to suspend his campaign and to call for postponing the first debate adds yet more uncertainty to a presidential campaign that is far more difficult to predict than any of the previous six elections in which I have worked.

Figure

This is not just my opinion; the most reliable forecasters in the country — the voters themselves — agree with me. In every recent election the public has accurately picked the winner by this time in the cycle. But not this year. Two weeks ago when we asked voters to put aside their own preferences and make a prediction, 39% said McCain would win and exactly the same number chose Barack Obama. Four years ago in September, the race was close, but by a 60%-to-22% margin voters thought President Bush would be re-elected.

In 2000 at this time, voters believed Al Gore would win. But they changed their minds by late October and picked George W. Bush. In 1992 and 1996, boxcar majorities (61% and 75%, respectively) thought Clinton would win.

Why is there so little consensus on this election? For starters, voters are unsure whether John McCain, if elected, would govern differently from President Bush: 44% think he would, but an equal number think he wouldn’t. And opinion about this basic question has not changed at all since March.

Second, while the country leans Democratic because of strong discontent with President Bush and the condition of the country, only 47% of the electorate thinks that Barack Obama is well qualified to be president.

These and other campaign cross pressures have the electorate in a state of high anxiety. As many as 51% in Pew’s early September poll said the word risky applies to Obama, and almost as many — 46% — said the word applied to McCain.

Read the full commentary at nytimes.com