The prestigious Gallup polling organization is losing some of its cachet, as clients and the public wonder why their predictions for the 2012 presidential election were so far off for so long.
And now, USA Today has announced they will not renew their polling contract with Gallup, after 20 years of collaboration.
The two organizations say the decision is mutual, chalking it up to differences of opinion and the "changing and evolving" nature of their relationship.
Both companies are equally vague, so it's hard to know who initiated, but USA Today says it is already finalizing talks with another polling firm.
They have plenty to choose from. Polling outfits have sprung up as the public grows hungry for data.
And USA Today can do far better than Gallup, which was one of the least accurate firms during last year's election.
At times, Gallup showed Republican challenger Mitt Romney with as much as a 7-point lead over President Obama, heights never seen in any other polls.
Gallup's use of "rolling" averages aggregates data over a week, smoothing out spikes, but also amplifying errors in methodology. It seems likely that Gallup's proprietary model of likely voters is somehow skewed, though after the election, they may have been tweaking it.
At first glance, Gallup seems to have a good model. They use live-callers and poll cellphones, which prevents a fair amount of conservative bias. But they may have bought into pundits' proclamations that turnout in 2012 wouldn't equal turnout in 2008, an assumption that turned out to be wrong.
Lower turnout generally results from a dearth of Democratic voters, so many of Gallup's numbers showed Romney doing better among almost all demographics than he actually did. Gallup showed Romney and Obama tied among women, for example.
Of course, USA Today could pick firms with equally bad track records, like Rasmussen, which consistently skews conservative. Rasmussen and Gallup were the only two polling outfits still predicting a Romney victory even on the eve of the election, and Rasmussen was the only organization that predicted a Romney win in Ohio.