- The Washington Times - Wednesday, November 5, 2008

Much of the political discourse in the nation’s longest presidential campaign was shaped by a daily deluge of opinion surveys that gauged American sentiment du jour. News coverage often hinged on the few digits that separated Sens. Barack Obama and John McCain.

In the end, were the polls right?

In many ways, yes. Stringent polling techniques and new disciplines imposed on pollsters themselves contributed to accuracy. The public, too, has a role. They have been conditioned to heed polls, thanks to the ever hungry mainstream news media, which needed numbers to fuel their prime-time dramas and burgeoning headlines.

The statistical separation between Mr. Obama and Mr. McCain ranged from one percentage point to about 10 in recent weeks, according to a spate of polls from news organizations and research groups - often deeming Mr. Obama’s lead “comfortable.”

That pattern persisted far into election night.

The earliest speculation based on real voting numbers found Mr. Obama with a three-point lead nationwide with the popular vote, a six-point lead in Florida, a three-point lead in North Carolina. Mr. McCain had a three-point advantage in Indiana and a 12-point lead in Kentucky. The hues of the states from red to blue were charted with county-by-county precision in some battleground regions.

“It’s almost been a perfect storm, brewing in favor of Senator Obama,” noted CNN analyst William Bennett.

Exit polls and their fallout have worried Mr. McCain’s campaign for months.

Strategist Bill McInturff warned Tuesday that exit polls “could overstate” voter support for Mr. Obama - citing a previous Fox News survey that revealed 46 percent of Obama supporters said they were very likely to participate in the exit polls, compared to 35 percent of McCain supporters.

A contrite press has vowed to be cautious about the siren call of exit polls for years. No one wants a repeat of the 2000 race, when the press pronounced early winners based on faulty exit polls. And memories of the midterm two years later haven’t faded either; revamped exit polling from the now defunct Voter News Service also failed.

Since 2004, news organizations have adhered to fact-based methodology established by the National Election Pool - a consortium of broadcasters and the AP, using standards meant to discourage eager journalists from declaring early winners to garner big audiences.

“Individual polls can swing wildly. If you report available polls, you give people the idea they are fluctuating wildly from one hour to the next. No wonder voters get whiplash,” said Sam Feist, polling director for CNN.

In recent days, the network used a “Poll of Polls” - which averaged numbers from six major daily polls - as a centerpiece of its political coverage.

“That way, you don’t see wild swings. You see trends,” he said.

Exit polls of voters leaving their polling places are a different snapshot all together.

“Unlike the last election, exit poll data was quarantined until late afternoon to prevent leaks of information that may not be complete or valid, and is very much open to misinterpretation in news reports,” Mr. Feist said.

The General Election Exit Poll conducted by Edison Media Research and Mitofsky International included more than 100,000 interviews at 1,000 polling locations.

The AP - which used 5,000 field correspondents to glean information from voters - made its assessments based on voter turnout, previous voting patterns, exit polling, telephone polls of absentee or early voters, and the “experience of AP journalists who have covered the campaigns,” the wire service said.

“When our analysis indicates a candidate has that significant winning margin, we will call the race at the scheduled poll close hour even if voting has been extended briefly in selected precincts. As always, we do not call races until we are confident of that winning margin, whether at poll close or many hours or days later.”

Copyright © 2016 The Washington Times, LLC. Click here for reprint permission.

blog comments powered by Disqus


Click to Read More

Click to Hide