Predicting Who Will Be a Good President

10/29/2012 10:00 am ET | Updated Dec 29, 2012

Hiring time is coming! Sorry, it's too late to submit your resume but at least you get to be part of the hiring committee. The two finalists have gone through all of the interview rounds, kissed all the babies, shaken all the hands, memorized all their talking points and are waiting to see who the committee selects.

The committee is struggling right now, unsure of which applicant is more likely to do a good job, or, for the more cynical, which is less likely to cause damage. Committee members have been wondering what exactly should they be looking for in a candidate, what experience is most important.

Using historical data, we can identify what experience is most important in hiring a new president. How we can assess this is actually a fairly straightforward statistics exercise: we can build a predictive model of presidential quality using input variables like the educational background, business experience, political experience and military experience. Luckily we have historical data to use to measure all of the input variables and presidential quality we can use the average ratings provided by historians in recent surveys. While there are no perfectly objective ways to measure presidential quality, it is fair to say that historians' estimates will certainly succeed at separating the George Washingtons and Abraham Lincolns from the Warren Hardings and Franklin Pierces.

Here's what we found:

None of the main effects were significant. That means neither educational background, political experience, business experience nor military experience itself was a strong predictor of presidential quality. In the area of education, while highly educated FDR was considered to have been a top president, so was Harry Truman, who lacked a college degree. While Eisenhower and Washington both had substantial military experience and strong presidential terms, Civil War General Ulysses S. Grant was considered a below-average president. As for political experience, political veterans like Thomas Jefferson and Teddy Roosevelt shone while other experienced politicians like Hayes, Fillmore, Nixon and Buchanan were considered to be some of our worst presidents. Highly experienced businessmen like Washington and Andrew Jackson went on to be well-respected presidents while experienced businessmen like Hoover and George W. Bush are considered to be some of America's worst.

There was one statistically significant interaction term though it had only a limited amount of predictive power: those who had both military and business experience before becoming president tended to perform well (think of George Washington, Andrew Jackson, Abraham Lincoln and Harry Truman).

Neither of our job applicants meets this criterion but, have no fear, if whoever gets the job offer doesn't work out, the committee gets to make another hiring decision in 2016.

Statistical Note: Two types of predictive models were developed, a decision tree model and a stepwise regression model. In both models, the only variable selected was the interaction term between military experience and business experience where the R-squared of the regression model was 0.1 (meaning the model only explained about 10% of the variability in presidential performance).