In a little more than two weeks, America is going to find out just how black Barack Obama is. And white Iowans will be the judge.
At the conclusion of the Jan. 3 Iowa caucuses, Obama will know whether all those good folks who told pollsters they were going to vote for him actually did when the question was called.
A recent Des Moines Register poll of likely Democratic caucus-goers showed Obama with a slight lead over Hillary Rodham Clinton that could easily disappear.
Obama must worry that he may be treated like black candidates in the past, who were surprised by election results that differed greatly from the polls.
Often, white voters who had said they would vote for a black candidate did not.
One of the most memorable examples of this was the 1989 Virginia gubernatorial election, in which Doug Wilder, a black man, had a double-digit lead in preelection polls. He won the race against Marshall Coleman, but by less than 1 percentage point.
More disappointing was the case of Harvey Gantt in 1990. Gantt was leading in most of the polls in his admittedly uphill race for the U.S. Senate in North Carolina, but in the end lost to incumbent Jesse Helms by 6 percentage points.
But those races were in the South, Iowans will say, suggesting Midwesterners aren't so fickle about voting for a black candidate.
Maybe, but there's also the example of Carol Moseley Braun. She won her 1992 U.S. Senate race in Illinois by a comfortable 10 percentage points, but polls had put her ahead by double that amount.
Obama, though, can take comfort in a recent analysis of the 2006 midterm elections by the Pew Research Center for the People & the Press. It suggests polling and voting are much more aligned than 10 or 20 years ago.
"Although African American candidates lost four of the five statewide races that featured black vs. white candidates, the late preelection polling tended to mirror the final outcome," said the Pew analysis written by Scott Keeter and Nilanthi Samaranayake.
The analysts noted that although two black Republican candidates for governor lost by wide margins in Ohio (Ken Blackwell by 23 points) and Pennsylvania (Lynn Swann by 20 points), polls in each state before those elections had shown similar margins. Polls also predicted Democrat Deval Patrick's landslide win in the race for Massachusetts governor.
Black Republican Michael Steele lost by 10 percentage points in his 2006 race in Maryland for a U.S. Senate seat, which was a much larger margin than had been predicted by some polls. Democrat Harold Ford's loss by 3 percentage points in Tennessee's U.S. Senate election, however, closely matched earlier polling results.
The Pew analysts said their review "suggests that fewer people are making judgments about candidates based solely, or even mostly, on race itself." They also noted that this country has come an awfully long way from 1958, when 53 percent of Americans said they would not vote for a black candidate for president. In 2003, only 8 percent of Americans expressed that prejudice.
Another thing Obama will have going for him in Iowa is the caucus process itself. It may not be as easy for a person to say one thing to pollsters and then hide his actual vote the way he could in a regular election.
Caucus participants make their choices by electing delegates to the 99 county conventions. The Republicans take a straw vote of those attending their caucus, which can be done by a show of hands or by dividing themselves into groups according to candidate.
The Democrats make it much more complex, with the voting on delegates first occurring at the precinct level (there are 1,784 precincts), then at the county level. There's even a mechanism that allows supporters of a candidate to switch sides.
The caucus results are not binding on the elected delegates, but the official Iowa Caucus Web site says "the delegates usually feel obligated to follow the wishes expressed by the caucus-goers."
That's a good thing. After all, whoever wins the caucuses would like it to be a meaningful victory - even if the meaning lasts only as long as the five days before the New Hampshire primary.
By then, we'll know whether Iowans were any more truthful with pollsters than Virginians in 1989 or North Carolinians in 1990.