Monday, December 6, 2010

Can we believe polls?

Rasmussen, Gallop, Pew, and Real Clear Politics.  These are all polls that some of us live by from day to day.  Marketers rely on polls.  Politicians live and die by polls.   Sadly, some vote by polls.

But now comes an article in Scientific American by Charles Seife that sheds new light on polls and those who report on polls.  The poll in question was released in September and concerned a “religion test” of 32 questions.  It pitted atheists/agnostics against several groups of believers.

The poll seemed to show that non-believers know more about religion than believers.  Time magazine headlined “Atheists know more about religion than believers”.

The article’s author pointed out that a Pew’s research was flawed because the sample for non-believer was small and the statistical margin of error revealed that there was no difference between the two groups polled.

Nevertheless, the headlines screamed and once again the public shrugged it shoulders and some once again may have accepted these polls on face value.

But should we accept the results of polls?  Back in the 1970’s a scandal rose over the so-called push-polls.  Candidates conducted bogus polls by polling only those groups who would ordinarily voted for said candidate.  The results was released to the media and touted in political ads.  Sure enough, the results in the other polls followed the bogus push-poll.

“The press leaped on the atheists versus believers headlines without critically examining the numbers.  The Pew study revealed less about our faith in God than it did about our faith in polls – which, far too often, is blind.”  Seife said.

The article is in the December 2010 edition and is  entitled, “The Science of ‘Disestimation’; Why we shouldn’t put our faith in opinion polls”.