Using Credit Scores, Betting Odds and Polls to Predict the 2016 Election

I think it is hard to deny that the 2016 United States Presidential election continues to be an unpredictable spectacle that is making history. I have been trying to follow it in terms of statistics, which leads me to regularly be directed to two interesting websites.

The first website is where people can bet (albeit, technically off-limits to Americans) on the outcome of the election. A betting site like this is effectively a futures market, and I visit the site twice weekly to see what the odds are for people putting real money down on their bets. Since about January, the odds have consistently suggested a Clinton-Trump matchup, with Clinton carrying a respectable lead, but not an enormous one. This isn’t my opinion or spin on the facts; this is what real money is betting on. Check out the archived results, and you will see some interesting trends throughout the nomination season.

The other site I like to visit is which is a relatively new blog established by Nate Silver. Silver is respected for creating a way of coalescing an analysis of several polls to provide a strong consensus on election outcomes. He accurately predicted how 49 of 50 states voted in 2008, and how 50 of 50 states voted in 2012. Silver’s blog isn’t predicting the next president just yet, but it is closely following primary polls. They have great graphs showing how each candidate’s support has tracked over time, and shows how “on target” each candidate’s strategy is that would best secure their nomination. The most fascinating recent developments show Sanders is not too terribly off track to still have a chance at winning the nomination, and that Trump’s support may be waning enough to result in a brokered convention. Again, this isn’t my opinion, these are what the numbers and polls are actually saying.

I think the most interesting article I’ve read this week happened to come from CNBC: . It appears CNBC tried to track how people will vote based on their credit scores. People with the worst credit scores, defined as 620 and below, were most likely to vote for Clinton, followed by Sanders, and then Trump. And people with the best credit scores, defined as 720 or greater, were most likely to vote for Kasich, followed by Rubio and then Cruz. The poll said it was conducted online and had 765 participants. Without more concrete information on how it was conducted and given the very small sample size, I don’t think this poll really tells us much of anything!

It is always important to remember that this data is simply a snapshot in time, and what truly matters more than anything is that every able, eligible voter makes it to the ballot box to support their desired candidate. The United States has some of the worst voter turnout of developed countries. In the 2012 election, only 53.6% of eligible people actually voted in the US, but 84.3% of people registered to vote did vote. It appears if we can convince people to register to vote, there is a good chance they will participate!