Predicting the outcome of elections is big business. In the early days it was left to political professionals who would rely on their gut instincts to “feel” how the campaign was developing. This is not dissimilar to Wall Streeters who can “read the tape” to sense which way particular stocks will move. In the mid-20th century scientific polling developed, but with occasional spectacular failures. The Literary Digest poll in 1936 predicted an Alf Landon victory over FDR. Landon carried only Maine and Vermont. Everybody was wrong about the outcome of the 1948 election, epitomized by the picture of a triumphant Harry Truman holding up a copy of the Chicago Daily Tribune with its premature headline DEWEY DEFEATS TRUMAN.
In recent years, Intrade has allowed people to bet real money on the outcomes of elections, in effect measuring the gut instincts of the many. It currently has Obama’s chances at 57.3 percent and Mitt Romney at 42.3 percent.
And, of course, political science professors try as well to read the tea leaves. Larry Sabato of the University of Virginia is probably seen more often on television than other professor. He currently has the race at 237 electoral votes safe, likely, or leaning to Obama, 206 to Romney, with 95 in the tossup category.
Two professors at the University of Colorado, Kenneth Bickers and Michael Berry, have developed a prediction model based not on polling or gut instincts, but on economic factors in each of the fifty states and the District of Columbia:
According to their analysis, President Barack Obama will win 218 votes in the Electoral College, short of the 270 he needs. And though they chiefly focus on the Electoral College, the political scientists predict Romney will win 52.9 percent of the popular vote to Obama’s 47.1 percent, when considering only the two major political parties. . . .
“What is striking about our state-level economic indicator forecast is the expectation that Obama will lose almost all of the states currently considered as swing states, including North Carolina, Virginia, New Hampshire, Colorado, Wisconsin, Minnesota, Pennsylvania, Ohio and Florida,” Bickers said.
You can take this for what it’s worth, but I will point out that this model has correctly predicted the outcome for every presidential election beginning in 1980.
I am reasonably confident that Romney will beat Obama in November. If the media makes a huge issue out of Akin's comments just to avoid talking about the economy, they know that the current environment doesn't favor the President. Add to that the current fundraising gap that favors Romney and factors seem to be in place for him to win. n nHowever I'm not confident that this new model will be correct. Had the good professors isolated these factors prior to 1980, I'd take the prediction a lot more seriously. But they knew what they were looking for in advance. (Would the economic factors have predicted a Clinton victory in 1992 without the presence of Ross Perot?) That makes a lot easier to prove their conclusions.
I've predicted almost exactly that result a month ago; admittedly based on something much simpler: Obama is the most divisive, the most partisan, the most ignorant, the most obnoxious person ever to inhabit the presidency. At this point (and a month ago) he was at 47% nationally (based on RCP average) and at 47% in states with 320 EVs. There is simply no way for such a horrible person to "swing" any more votes his way. The people in his pocket today, including the ones who are holding their proverbial noses to support him, are all he is going to get. He will get no more votes because of the louse and low-life he is. So even the states he is "leading", but where he is below 50% (and almost everywhere by the insanely blue states, he is at 47% or less) he will lose in November.
“You can take this for what it’s worth, but I will point out that this model has correctly predicted the outcome for every presidential election beginning in 1980.”
Unless the model was created prior to the 1980 election, it hasn’t “predicted” any such thing. Rather, it’s correlating the economic factors with the voting patterns from the years 1980 through … 2004?, 2008? … to generate a prediction for 2012. Which is interesting if the correlation is strong, but….
If I flip a coin ten times, I can use the results to build a model that will ‘predict’ the pattern of those ten flips. But as the saying goes, “Past performance is no guarantee of future results.”
When was this "model" developed? It's easy to look backwards and Monday morning quarterback outcomes. That's not the test of a model. In fact, it's not a model at all. It's a statistical analysis of known outcomes, looking backwards. I'll also add that the sample size is very small.