Last week, I wrote about how some of my political science friends had come up with a weird thesis.
They had conducted research showing that the fate of gubernatorial candidates in general elections seemed to be related--slightly and statistically--to whether their state's flagship university had won its football game the preceding weekend. You can access that column by clicking here.
The reasoning of my political science friends (Keith E. Lee, Jr., of the University of Florida, Sydny L. Bryan of Valdosta State University, and James T. LaPlant, also of Valdosta State University) was that winning the big game on Saturday boosts the incumbent gubernatorial candidate/party by as much as 10% in the following Tuesday election. It's a matter of sense of personal well-being, kind of like your feelings about the economy or the weather.
I used that thesis to try to "predict" or at least plot theoretical scenarios for the likely outcomes of this past Tuesday's critical and highly competitive elections in Georgia, Florida, Illinois, Michigan, and Wisconsin.
Now it's grading time.
Did the model accurately explain the electoral fortunes of governors Nathan Deal (GA), Rick Scott (FL), Pat Quinn (IL), Rick Snyder (MI), and Scott Walker (WI)?
Yes and No. Actually, it was more Yes than No.
By my calculations, the "Big Saturday" and "Big Tuesday" correlations accorded with the thesis in every case except Georgia--a success rate of 80%, which probably beats most gamblers' experience with point spreads and game outcomes.
Here are results as I interpreted the football/election correlations.
• Only Gov. Nathan Deal (R) in Georgia confounded the thesis--but he's not going to complain. Deal led Jason Carter (D) by only 2.0 percent in the final pre-game polling average; and his highly-favored Georgia Bulldogs were upset by the Florida Gators. Nevertheless, Deal whipped Carter by 8.0 percent on election day. This was the one case, in my opinion, where the thesis missed the mark.
• In Florida, incumbent Gov. Rick Scott (R), who trailed Charlie Crist (D) by 1.7 percent in the final pre-game poll average, won the election by 1.1 percent after his Florida Gators defeated the Georgia Bulldogs. Put a checkmark here for the thesis.
• In Illinois, Gov. Pat Quinn (D) had a 1.0 percent lead over Bruce Rauner (R); but the Illini ran into powerful Ohio State in Columbus and took a shellacking. Then, on Tuesday, Quinn fell to Rauner by 4.9 percent. Another checkmark for the thesis.
• In Michigan, Gov. Rick Snyder (R) led Mark Schauer (D) by a 2.8 percentage margin. His Michigan Wolverines defeated Indiana; and Snyder prevailed electorally by 4.2 percent. Another checkmark for the thesis.
• Finally, in Wisconsin, Gov. Scott Walker (R) went into the election with a 2.0 percent leader over Mary Burke (D). After his Badgers defeated Rutgers , Walker cruised to a 5.7 percent victory. Another checkmark.
I contacted my three political science friends, and they put a slightly different spin on the results. Keith Lee (University of Florida) had just run the current data on the five elections. He looked at election returns in the home county of the flagship universities (which was what they had used in their original research) rather than my statewide analyses; and by his calculations, the thesis worked in three of the five cases. He found that Deal (GA) took a slight dip in Clarke County, home of the Bulldogs, who had lost on Saturday; Scott (FL) enjoyed a slight boost in Alachua County, home of the Gators, who had defeated Georgia; and Snyder saw a slight boost in Washtenaw County, home of the Wolverines, who had whipped Indiana--all predictable according to the thesis. But, contrary to that model, Quinn (IL) realized a slight boost in Champaign County, home of the Illini, who had lost to Ohio State; and Walker (WI) suffered a slight dip in Dane County, home of the Badgers, who had defeated Rutgers.
A Passing Grade.
I would have to give the football/election thesis a passing grade, based on these few cases, regardless of whether you use my statewide speculations or the more precise county analysis of my bold political science friends.
Of course, we all can think of more conventional reasons why these important gubernatorial elections turned out the way they did. One explanation is simple partisanship--all five winners were Republicans. But the fact is that this unconventional thesis probably performed as well or better than did many professional pollsters and media commentators.
What do you think? Is this thesis sound political reasoning--or pigskin frivolity?