Skip links

The (un)Reliability of Polls


Everyone knows that if you eat after you swim, you’ll cramp-up, that we have only five senses, and that we use only ten percent of our brains. Problem is, none of these things is true.

It’s human nature to believe something that sounds at least half-way credible if it’s repeated often enough. This is as true for matters of public policy as it is for cultural anecdotes and warnings about non-existent dangers.

For example, for many years, Albert Kinsey’s flawed research was cited to proclaim that ten percent of the population was gay or lesbian. We now know this is an over-estimate by a factor of about 300 percent. Gary Gates of UCLA Law School’s Williams Institute, which “specializes in law and public policy related to sexual orientation,” in 2011 published a comprehensive “atlas” of homosexual research in which he says the total “LGBTQ” population of the United States — people who identify as being in one of these categories — is about 3.8 percent. 

The next year, a Gallup poll commissioned by the Williams Institute found even that statistic too high: In a survey of 120,000 adults, “3.4% say ‘yes’ when asked if they identify as lesbian, gay, bisexual, or transgender.”

Religious liberty and matters of conscience, essential to the very nature of the republic in which we live, also are too frequently presented in a skewed way. In September, the respected Pew Research polling and analysis firm found that 67 percent of Americans believe that “employers who have a religious objection to the use of birth control should be forced to provide it in health insurance plans.” This sounds like an overwhelming percentage until one realizes that this question embodies a false alternative. 

Read more at The Stream.

Share with Friends: