Polsters' Dirty Little Secret

Like many political enthusiasts, I've been following the current political process like a rabid sports fan playing fantasy football. The polls are my scoreboard and any uptick by one side or the other elicits either standing cheers or woeful groans.


Unfortunately, the current polls have a big problem in their scorekeeping method that falsely skews the numbers in John McCain's favor. For those who have a financial stake in keeping the race close (news media, political fundraisers, etc.) this little handicapping method is conveniently overlooked. However, a careful observation of the polling process shows the errors.

First, polsters do not call cell-phones. That means that households without landlines will never be polled. How big of an issue is this? According to a report published last December by Stephen Bloomberg, a senior scientist at the federal Centers for Disease Control and Prevention*, one in eight American households have cell phones but lack traditional land lines. That's 12.5% of the total population.

Polsters argue that they mathematically adjust their numbers to account for this missing population. However, these polsters assume that this population votes the same way as similar demographic groups. Here's the problem with that. According to Bloomberg, this population is more likely to be "young, poor, male, and Hispanic." In addition, "11 percent of whites and 14 percent of black adults had only cell phone service." A third of adults under 30 had only cell phones. This group describes a demographic that supports Barack Obama by more than 60 percent.

That un-polled support may only affect the numbers by a few overall percentage points, but in a race that has been tied for much of the past few months, those few points are significant. So be careful when you see poll numbers published by television news, especially cable news.



*As published on cnn.com (December, 2007).