- The Washington Times - Monday, March 30, 2009

WASHINGTON (AP) - Polling that ended too early and other technical shortcomings _ rather than undetected racial bias _ are the likeliest reasons so many surveys incorrectly suggested Barack Obama would defeat Hillary Rodham Clinton in the 2008 New Hampshire presidential primary, a report concluded Monday.

Clinton defeated Obama 39 percent to 36 percent in the Jan. 8, 2008, contest, even though many pre-primary polls showed Obama with solid leads. Clinton’s victory gave her a badly needed burst of momentum just five days after Obama won a surprising victory in the Iowa caucuses, the year’s first presidential contest.

Six months later, Obama captured the Democratic nomination and later the presidency.

The report, written by a panel of 11 pollsters and academics appointed by the American Association for Public Opinion Research, said polling may have ended too early to reflect last-minute shifts in voters’ attitudes. In the best-known incident, Clinton’s eyes welled up a day before the voting as she vowed to fight on no matter what happened.

“Because most of the New Hampshire polls ended data collection prior to this event, these polls would have missed any related last-minute shift to Clinton,” the study said.

The report said a failure by some pollsters to reach enough New Hampshire voters who supported Clinton could also have contributed. Clinton’s supporters were often less educated and lower income than Obama’s, people who tend to be harder for pollsters to reach.

In addition, the authors said Clinton may have been helped by being listed near the top of the New Hampshire ballot, while Obama’s name was near the bottom.

In the days after the primary, some analysts said the explanation might have been a reluctance by some white voters with negative views of black candidates to candidly tell pollsters for whom they would vote. That was discounted by the report’s authors.

“In the data we have from a wide variety of New Hampshire pre-election and exit polls, we found no evidence that white respondents over-represented their support for Obama” while talking to pollsters, said Michael Traugott, a University of Michigan political scientist and polling analyst who led the study.

The report also said a large number of new voters _ about one in five New Hampshire voters were voting for the first time _ skewed the assumptions pollsters used to calculate which types of people were likeliest to vote in the primary.

All 13 polls studied underestimated Clinton’s 39 percent showing, some by as many as 11 percentage points. Most came far closer to Obama’s 36 percent.

Traugott said he could not explain why the same pollsters, using the same techniques, accurately forecast John McCain’s 37 percent to 32 percent victory over Mitt Romney in New Hampshire’s Republican primary.

The authors, who spent more than a year gathering and analyzing data, included an unusual broadside against polling organizations that declined to provide information they requested. Just four of the 13 polls provided all the information sought.

As a result, the authors said they could not gauge if other factors contributed to the inaccurate polls, including differences in how firms compile lists of people to randomly call.


On the Net:

American Association for Public Opinion Research: https://www.aapor.org

Sign up for Daily Newsletters

Manage Newsletters

Copyright © 2020 The Washington Times, LLC. Click here for reprint permission.

Please read our comment policy before commenting.


Click to Read More and View Comments

Click to Hide