What drove the tide of GOP victories in Kansas and proved many pollsters wrong? Political Science Professor Chapman Rackaway of Fort Hays State University believes the answer is simple -- Republicans just turned out in greater numbers.
In an interview with Stephen Koranda of Kansas Public Radio, Rockaway explained that the GOP accomplished that by using sophisticated data models and tools that helped them contact the right people, in the right way, and get them to vote.
That success also meant pre-election polls were thrown off because of miss-sampling of the actual voting population. The statisticians who analyzed the results of multiple polls were consequently in error as well. For example, Real Clear Politics’ aggregation of polls gave Greg Orman the lead in the Kansas Senate race, which he lost by 11 points. And Nate Silver’s FiveThirtyEight projection gave Paul Davis an 82% chance of winning over Sam Brownback.
However, Titus Bond, a Kansas City pollster who conducted surveys for a super PAC backing Pat Roberts, reported his firm showed Roberts up by six percentage points. (He won by 10 points.). In a Wall Street Journal article, Bond noted the key to getting the result right was being careful not to oversample Democrats and independents. In the same article Tom Jensen, the director of the Democratic polling firm Public Policy Polling, admits he included too many Democrats in surveys of Arkansas, Kansas and Kentucky, resulting in poll numbers that were off by between eight and 11 percentage points.
Preliminary turnout figures show about 49 percent of registered voters in Kansas marked the ballot for the U.S. Senate and governor’s race. Gov. Sam Brownback was re-elected by about 24 percent of registered voters and just under 20 percent of the voting-age population (based on 2013 population estimates). Sen. Pat Roberts was re-elected by about 26 percent of registered voters and about 21 percent of the voting-age population.
