During the past few election cycles, we’ve regularly heard the phrase, “As Ohio Goes, So Goes the Nation,” meaning that Ohio typically votes for the candidate who wins the Presidential election. We know that this is at least somewhat accurate, but I’ve been wondering how Ohio compares to other states. Is Ohio is really the best at choosing the winners of presidential elections? Are other states just as good or better? Which states are the worst? So, I decided to analyze the voting records in order to gain a better understanding of these questions.

To do this, I obtained data from Wikipedia showing each state’s voting history back to the first presidential election in 1789. Using this data, I calculated each state’s “accuracy percentage”—the percentage of times where the state voted for the eventual winner of the election. I then visualized the data in Tableau (you can view the full visualization here).

While Ohio does well, I was surprised to see that New Mexico is actually better than Ohio, with 88.9% accuracy, compared to Ohio’s 85.2% accuracy. New Mexico has chosen the winner 24 times out of the 27 elections in which it participated (New Mexico became a state and participated in its first presidential election in 1912), while Ohio has chosen the winner in 46 of the 54 elections in which it has participated (Ohio became a state in 1803 and participated in its first election the following year).

California, Illinois, Nevada, Pennsylvania, and New York have also done quite well, all having an accuracy percentage over 79%. The worst accuracy goes to the District of Columbia which has only chosen the winner 6 times in its 14 elections, for an accuracy percentage of 42.9%. Alabama and Mississippi are also poor performers, being the only other two states with an accuracy of 50% or lower.

So, perhaps we should change the phrase to “As New Mexico Goes, So Goes the Nation.” But not so fast...Sure, New Mexico’s historical record is better than Ohio’s, but elections in the early 1900’s have relatively little to do with elections in the 2000’s. So, I wondered which state was more accurate in the past fifty years. So I created a visualization showing hits and misses for each state.

New Mexico started with a great streak, correctly choosing the President in its first 16 elections, from 1912 through 1972, but since then it has only been correct in 8 of the last 11 elections. Still quite good, but Ohio has done much better, correctly choosing the president in every one of the 14 elections since 1964. That’s a pretty amazing track record, which no other state comes close to matching. Considering the “As Ohio Goes…” phrase seems to focus on more recent elections, I think it’s fair to give Ohio credit for its ability to choose Presidents.

With Ohio being so good at this, I began to wonder about its individual counties. Which county is the best? Which is the worst? So, I obtained data from Cleveland.com showing the winner of each county for every Presidential election since 1960 and created visualizations similar to those created for each state.

The best performing county is Ottawa County, which has only missed one of the 15 elections since 1960 (93.3%), followed by Lake County (86.7%). All other counties fall below the overall accuracy of the state. The worst accuracy goes to Monroe county (46.7%), which happens to be the only county falling under the 50% mark.

So there you have it—Ohio does, indeed, deserve the title as the best state at choosing the President over the past fifty years or so. But, over the course of our short history, New Mexico takes the title.

If you’d like to view the visualization, you can see it here. And, if you find any additional insights, I’d love to hear about them.

Ken Flerlage, December 19, 2016

#### No comments:

Powered by Blogger.