Manchester, NH -- The conventional wisdom today about the results of the New Hampshire Democratic primary does not take into account the complexity of the multiple forces at play.
There were at least five factors influencing the outcome in Hillary Clinton's unexpected victory here on Tuesday: the so-called "Bradley effect;" the different ways in which primaries and caucuses filter voters; the geographical distribution of the turnout increase in New Hampshire between 2004 and 2008; a shifting gender gap; and the greater number of young voters and college students in Iowa than in NH.
The outcome of the Democratic contest here suggests the emergence of the so-called "Bradley/Wilder/Dinkins effect" -- a discrepancy between election results and poll data in races in which African American candidates win 5 to 10 percentage points fewer votes than predicted. (The effect was first noted in Tom Bradley's 1982 race for California governor, when polls pointed to a win but the result was defeat by a white opponent.)
"Anytime you've got white undecided voters pulling the lever choosing between a white and a black candidate, that is when the race issue is most important," notes Drew Westen of Emory University. "Both campaigns' internal polls showed a 10 to 12 point Obama lead; to see that evaporate into a three-point loss, when he didn't have any gaffes, that has a ring to it."
According to Andrew Kohut of the Pew Research Center, "The failure [of polling] on the Democratic side has to do with the fact that Clinton ran best among groups of voters who most often refuse polls -- poorer, less well-educated people. These are also the very people who are reluctant to vote for a black candidate."
A separate finding - that Clinton performed best in New Hampshire yesterday in communities with the highest turnout increases between 2004 and 2008 - lends indirect support to the Bradley effect thesis. MIT political scientist Charles Stewart found that "Clinton did better, on average, in the towns that had the biggest [positive] change in turnout across the four years. Conversely, Obama did worse." Stewart's analysis of NH voting data showed that a "strong predictor of how Obama did, town by town, are the results from '04. Dean country is now Obama country."
Stewart also found that, conversely, John Kerry country has become Clinton country: "In the towns where Kerry beat Dean in 2004 [Clinton country now], turnout increased by a total of 31.4% between 2004 and 2008; where Dean beat Kerry, turnout increased by a total of 27.5%. That's not a huge number, but in a close race, you need everything you can get."
Clinton's machine may well have proved more effective than Obama's in pulling out new voters. Another way to interpret the finding is that a small but significant fraction of white voters turned out to actually cast ballots against a black candidate. The two interpretations are not mutually exclusive.
A number of political analysts question a Bradley effect in New Hampshire, including Ben Smith of Politico and Matthew Yglesias of The Atlantic. The difference of opinion will likely persist among political scientists and pollsters.
Yet if race has become a factor, it will challenge a central premise of this year's primary election: that the historic movement behind Obama reflects an electorate that has become sufficiently color blind - 54 years after Brown v. Board of Education -- to elect a presidential candidate who happens to be black.
Results of the Iowa caucus last week raised expectations less sustainable in primary states: a Democratic caucus acts as a filter, its cumbersome rules limiting participation to the relatively more committed social activists.
At the same time that race apparently worked against Obama at the margins in the Granite State, gender favored Clinton. Clinton lost among men by almost identical percentages, 12 points in Iowa (23-35) and 11 in New Hampshire (29-40). Among women, however, although Clinton lost by 5 points in Iowa (30-35), she won by solid 10 points in New Hampshire (46-36), an overall 15 percentage point spread.
Another point: there are substantially fewer young and student voters in New Hampshire than in Iowa -- 22 percent of Democratic Iowa caucus goers were under the age of 30, compared to 18 percent of New Hampshire primary voters. Stanford political scientist David Brady points out that the student bodies at the University of Iowa, Iowa State and the University of Northern Iowa far outnumber their New Hampshire counterparts.
Obama has done exponentially better than the three African Americans who sought the presidential nomination before him: Shirley Chisholm, Jesse Jackson and Al Sharpton. Nonetheless, at the moment race remains the most salient of factors.
Few Democratic politicians, especially Clinton, would publicly cite Obama's race as a liability. But in brutal political terms, this unstated vulnerability may be used to persuade leaders of the party establishment and fundraising network to stick with Clinton through what is now sure to be a bruising fight until at least February 5, and perhaps all the way to the convention in August.
Obama, in the coming months, will have to show that he is, in the words of U.C. Berkeley political scientist Jerry Lubenow, "more than a quicksilver candidacy, shiny and appealing but elusive and difficult to grasp, long on poetry but short on prose.... Hope is a difficult target to attack, but at some point he will have to define a more precise destination and map a path to it. As he addresses that reality he becomes more vulnerable."
At the core of the Obama campaign is a message of possibility -- what he calls hope -- that transcends racial politics. The electorate is on a steep learning curve, and Obama may yet prove to be the candidate who moves the country closer to its full potential in terms of equality.
Clinton, in turn, cannot afford to defeat Obama in a way that offends blacks or racially liberal whites -- both crucial constituencies for Democratic victory in November.
Clinton is a polarizing figure and, consequently, if she is the 2008 nominee, she will need a high turnout among core Democratic voters in order to have a chance at winning the White House. Any voters she drives away during the primaries may well stay at home or vote against her on November 4, a development she has to avoid at all costs.