Why Gallup Left the Horse Race
Gallup is a company whose name is synonymous with polling, and it’s also one with strong roots in the Princeton business community. The company was founded in 1935 at 114 Nassau Street. For many years it conducted focus groups at what is now the Off-Broadstreet Theater in Hopewell. Although it is now headquartered in Washington, D.C., with most of its operations based in Kansas, Gallup Editor-in-Chief Frank Newport is still works out of 502 Carnegie Center. The staff there writes blog posts and makes videos about what the polls mean.
Visit Gallup’s website (www.gallup.com), and you won’t see a thing on the front page about which candidate is likely to win the primary. Front and center is an article about how Trump and Clinton supporters lead in enthusiasm, alongside unemployment figures, a statistic called “Gallup Good Jobs” represented as percent of adult population (currently at 44.5 percent), and what percentage of Nebraskans feel that early childhood care is a very important investment (69 percent).
The mishmash of content in the middle of a frenzied Presidential election year reflects the organization’s move away from “horse race” polling of election campaigns and towards understanding the underlying opinions that drive political winds.
There are more sources than ever for election watchers to get their news. Although polling giant Gallup has stepped aside, private universities and small polling companies have proliferated, and a number of websites now provide sophisticated analysis of the results of those polls. For predictions, news junkies can turn to Nate Silver, Sam Wang of Princeton University (see main story), or Carl Diggler, a parody pundit created by comedy writers, whose predictions, driven not by math but by “gut, conventional wisdom, and personal experience,” have proven frighteningly accurate.
Gallup’s exit from the horse race polling business follows in the aftermath of its shortcomings in the last presidential election. In 2012 Gallup’s final polls showed Mitt Romney with a one-point lead, but when the voting was done, Obama prevailed by almost four points. Newport led a post-mortem of the firm’s polling methods and voter models that led to changes in the way Gallup does its surveys.
Newport recommended ways that the company could poll better, and perhaps regain some of the prestige it lost in 2012, but instead the company decided to exit the horse race altogether. It was a stunningly swift reversal for a company that had, until that point, enjoyed a reputation as an accurate predictor of election results going back to Franklin Roosevelt’s re-election in 1936.
Other large polling groups, such as the Pew Research Center, have also stepped back from primary polling under the same rationale that there are now more sources of horse-race polling than ever before.
But Gallup may yet put Newport’s recommended changes to the test in the general election. Once the dust has settled from the wild primary season, the Presidential contest may pose a more manageable scenario for Gallup to re-enter the race. But so far, Newport has been silent on that front, telling reporters the organization has not yet made a decision.