Researcher: Polls Are Accurate Predictor of Election Day Results

Tuesday, 18 Sep 2012 04:58 PM

By Todd Beamon

  Comment  |
   Contact  |
  Print   |
    A   A  
  Copy Shortlink
As the presidential election draws nears, every new day brings a new poll of the nation’s electorate.

Some show how far ahead President Barack Obama is from his Republican challenger, Mitt Romney. Others show the former Massachusetts governor holding his own. Others, still, say the race for the White House is a virtual dead heat.

With Americans being bombarded with poll after poll – and their often widely divergent results – the question arises: How accurate are these polls?

Very much so, says political science professor Costas Panagopoulos of Fordham University. “It’s remarkable at how close these polls come to accurately projecting what happens on Election Day,” he said.

“What many people don’t realize is that there’s always some error associated with polls because they are samples of the population. Once you take that error into account, the range of true preferences generally tends to be right on target.”

Panagopoulos is director of Fordham’s Center for Electoral Politics. He recently published research assessing the accuracy of 23 polling organizations during the 2008 presidential election.

His study showed that Rasmussen Reports and the Pew Research Center got the results of the election dead right.

“There’s been a proliferation of polls, some better than others, but on average, polls do very well in projecting election outcomes,” he said. So far, “several hundreds, if not in the thousands” of polls have been conducted this election cycle.

While survey methodologies are generally not flawed, Panagopoulos described “error” as “bias” — and said these factors have the most impact on how a poll is conducted as well as its results.

It ranges from the number of voters polled — the “sample” — to the kinds of voters queried — “registered” versus “likely” voters — and a bevy of personal attributes, including age, sex, race, residency.

“Pollsters are generally well-trained analysts — and they pay attention to the things that could result in determining preferences that are inaccurate.

“They live and die by the accuracy of their polls, so they will do their best make sure they come up with most accurate assessment of what preferences truly are.”

The surveying technology also is a factor.

“Polling is becoming increasing difficult to execute because many of these polls are conducted by telephone – and the growing number of cellphone-only populations is making it difficult for pollsters to reach people on landlines,” he said. “They have to supplement their samples with cellphone respondents.”

And even more potential poll respondents are just saying no.

“The non-response rate has been growing over the past few years. People selected to participate in polls have refused to do so.

“People are increasingly less willing to participate in these surveys – and that is partly because of larger social phenomenon, like a general reluctance to telemarketing and other outreach that many citizens perceive to be invasive.”

And, as for party affiliation, “A scientific poll does not have a partisan dimension to it,” Panagopoulos said.

But with so many polls, the best way for the average person to assess their accuracy is to look for the margin of error. This occurs whenever a population is incompletely sampled.

“If all those estimates are within the same margin of each other, then all the polls are suggesting the same thing,” Panagopoulos said. “If you don’t know the margin of error, and you’re trying to compare across polls, you’re going to make inaccurate characterizations.”

A good margin of error is generally plus or minus three percentage points.

Further, “at the end of the day, it’s more important to track trends in preferences rather than snapshots in time,” Panagopoulos said. “Has the trend been that a growing number of people are supporting Obama or supporting Romney? What is happening over time with these polls?”

This all aside, Panagopoulos has this core rule about polls: “The closer you get to Election Day, the more accurate the polls are.

“Polls are increasingly accurate as you get closer to the election. They are reflecting what is really going on in the electorate. People are making up their minds.”


© 2014 Newsmax. All rights reserved.

  Comment  |
   Contact  |
  Print   |
  Copy Shortlink
Around the Web
Join the Newsmax Community
Please review Community Guidelines before posting a comment.
>> Register to share your comments with the community.
>> Login if you are already a member.
blog comments powered by Disqus
 
Email:
Country
Zip Code:
Privacy: We never share your email.
 
Follow Newsmax
Like us
on Facebook
Follow us
on Twitter
Add us
on Google Plus
Around the Web
Top Stories
You May Also Like

Boehner's Christmas Carol Celebrates a GOP Congress

Thursday, 18 Dec 2014 23:15 PM

Republican House Speaker John Boehner put his own political spin on "a Christmas reading" on his YouTube page Thursday,  . . .

Dershowitz: Hollywood Should Have Followed Rushdie Example

Thursday, 18 Dec 2014 22:29 PM

Hollywood should have rallied around Sony Pictures and released "The Interview" jointly online, says Harvard law profess . . .

Army Pulls ROTC Job Posting Requiring Christian Candidates

Thursday, 18 Dec 2014 22:18 PM

The Army has pulled a job posting at an Illinois religious college that required leaders of its ROTC program to be Chris . . .

Most Commented

Newsmax, Moneynews, Newsmax Health, and Independent. American. are registered trademarks of Newsmax Media, Inc. Newsmax TV, and Newsmax World are trademarks of Newsmax Media, Inc.

 
NEWSMAX.COM
America's News Page
©  Newsmax Media, Inc.
All Rights Reserved