(PRWEB) September 23, 2004
Chicago, IL (PRWeb) September 23, 2004 Â An article entitled, ÂDewey Beats Truman Again Â Or Why We CanÂt Always Trust PollsÂ in the September 20, 2004 issue of the weekly web magazine News Informant, warns against accepting at face value that all presidential polls are as accurate as they say. Just last week, the article points out, four different reliable organizations released poll results. While three showed the two major party candidates, President George W. Bush with a lead of one percent over Senator John F. Kerry, or a statistical tie between the two, the survey from the Gallup organization showed a lead by the President of 11 percent among Âlikely voters.Â
This lack of correlation among the polls shows that they could not all be scientifically, within the 3 percent range of accuracy that they all claim, according to News Informant editor Bernard Perlstein. ÂIt would be easy to simply dismiss the one aberrant survey, Gallup, as inaccurate,Â remarks Perlstein. ÂHowever, the fact that this could happen to a respected polling organization tends to diminish the claims of scientific validity of all polls,Â he adds. Perlstein further noted that discrepancies among polls have occurred at other times during the 2004 campaign.
The article contradicts several explanations for the discrepancies in among the presidential election surveys released last week. One argument is that the Gallup survey measured the results of those identified as likely voters. However, a second Gallup survey of registered voters showed a Bush advantage of 8 percent -- still not close enough to the results of the other polls to be explained by the standard margin of error.
An additional point raised in the article is that polls tend to show large shifts in opinion over a short time. Yet these are frequently short-lived. ÂPolls do not differentiate between long-held belief and fleeting feelings,Â stated Perlstein. Generally, large shifts in opinion in short periods only occur with major events. And the best way to comprehend public opinion, the article states, is not from one poll from a single organization but from looking at the results of polls by different surveying organizations and then see how these change over time.
ÂDifferent organizations independently confirming the results of others is similar to a group of scientists replicating the experimental results of other groups of scientists. This demonstrates validity,Â Perlstein claims.
The News Informant article in its entirety, and its sources, can be found on the web at: http://www.newsinformant.com/2004_09_20/politics.php#000798. A free registration is required to view the article.
To request a free copy of the article, or to obtain reprint rights or other use of the article, please contact Media@newsinformant.com.
About News Informant Inc.: News Informant Inc. publishes News Informant, a weekly U.S.-based web news magazine focusing on news analysis of important domestic and global issues that are not sufficiently covered by the U.S. media. The magazine uses primarily respected worldwide online source material to render original information and perspectives. The periodical follows the slogan, ÂItÂs Not Just the News Â ItÂs WhatÂs Behind the News.Â
About Bernard Perlstein: Bernard Perlstein has been full-time editor of News Informant since 2003, as well as a contributor, specializing in Politics, Economics and Foreign Relations. Mr. Perlstein earned his MBA from De Paul University, and also has a B.S. in Mathematics and Computer Science. Mr. Perlstein previously majored in Government and minored in Economics at Oberlin College.
For additional information on News Informant or Mr. Perlstein, or for media inquiries, please contact:
Tonya D. Hottmann,
News Informant Inc.