Monday, August 25, 2014

Pollster Literacy

Literacy is the parent of credibility.

I don’t know if that’s  original here or  not; but ever since I passed a college course on Public Opinion at Emory University when Elmo Roper and Harris were starting out to compete with Gallup, I’ve been an addicted poll rat .

And I’ve had some bizarre wake-up calls along the way. Even today when everything is supposed to be so sophisticated and complex , there are ways to test the credibility if  not the accuracy of a public opinion survey 

For instance, within a month of the retreat of Scott Rasmussen from the poll to which he gave his name, the company began to deteriorate in credibility from the most accurate of all the American survey specialists to a clear left-leaning posture, revealing that they had returned to a sampling methodology based on the 2008 party split , rather than the 2010 / 2012 numbers . 

In these days when Reuters and Gallup show Obama’s approval at 40% or less while Rasmussen clings to a 47% approval on the same days of testing , it is bizarre .

But here is the strange thing . The syntax in the Rasmussenj website reporting is also bad. It has  frequent misspellings and slipshod editing like today’s use of “now “ for the very important verb “KNOW” !!!

I’ve noted the same thing in polls on local and state races. In Georgia today for instance, while five other surveys show David Perdue from 7 to 9 points ahead of Michelle Nunn, a WSB TV website headline gushes that Nunn “has expanded“  a lead that she has never even had , in any of the polls averaged on the Real Clear Politics unbiased and most credible site of all. The emotional one-sided awkward screed at WSB TV contributes to the embarrassment.

I’m reminded of Li’l Abner”s comic strip assertion long ago that “As any fool can plainly see ...I see !!!”

If they can’t report their findings clearly and with literate language and if they do not edit for clarity and accuracy , they probably do not do a circumspect job with the numbers.

Or  so it seems here . 

No comments: