Gallup-ing into the sunset?

By Richard K. Barry

Remember how Gallup polling got it so wrong during the 2012 presidential election? The reason, as Business Insider put it, was that "its likely voter model found a drastically different electorate than the one that actually showed up." You'll recall that for much of the month leading up to Election Day, Gallup had Mitt Romney ahead of Barack Obama by a fairly large margin, which was reflected in no one else's polling numbers. However they got it wrong, they got it wrong.

At the end of it all, the folks at Fodham's political science department published a list of the most and least accurate pollsters to keep tabs on the presidential contest and Gallup came out 24th out of 27.

I can't remember who did worse, but I do recall that Public Policy Polling and Daily Kos/SEIU/PPP, two supposedly left-leaning polls, did the best.

Anyway, if you're Gallup, if you're any polling company, your accuracy is your bond, and inaccuracy has consequences.

The Washington Post reported yesterday that Gallup will no longer be doing polls for USA Today.  After 20 years of collaboration, both organizations said it was a mutual decision "based on changing media and polling landscapes."

If you really want to understand how public relations departments do their job, you can go to the Post article and read the full statements from both companies. It's all nicey-nice about the two concerns moving in different directions - kind of like what your high school girlfriend told you just before she gave you the heave-ho.

I'm sure that in order to keep the lawyers out of it USA Today decided not to say they dumped Gallup because they were bloody awful at what they do. 

It's a brutal business when everyone knows exactly how well you did your job, right down to the decimal point, but that's the business they choose to be in. They won't be around much longer if they don't up their game.