Vote totals are officially calculated at the state, county, and precinct levels, but not at the city level. There’s been recent discussion on how major cities voted in the 2016 presidential election, and questions about what was the biggest city that Trump won. City-level vote choice can be calculated using tools or data other than the presidential voting data, but all of those make use of precinct level data to build up to city level totals. In the past, I’ve calculated city voting results by looking at city district maps made up of voting precinct names. This gets tedious, and there are better ways to make these calculations. Often times, precinct presidential results come with results for other ballot measures or races. If there are some ballot questions that are specific to a city, one can take those precincts and only look at presidential vote there to get vote choice within only that city.
That’s what I tried to do in calculating the vote choice in Oklahoma City–a potential “big city” victory for Trump. I obtained precinct data for the entire state from the Oklahoma State Election Board. There were 166 different questions on the ballot in Oklahoma during the 2016 November election, and I originally thought a proposition measure–which was called “PROPOSITION NO. 1 (SCHOOL BUILDING MAINTENANCE, SAFETY AND GENERAL EQUIPMENT) OKLAHOMA CITY PUBLIC”–was voted on by the entire city of OKC and only residents in that city. I took the codes of precincts that recorded votes for this measure, calculated vote choice for these precincts, and got a 49.9-43.5 win for Clinton in the city. However, it turns out only portions of OKC–a specific school district–voted on the proposition, and not the entire city (thanks to two reporters from Oklahoma, Laura Eastes and Sarah Stewart, for informing me of this).
Thus, I had to find a new way to narrow down OKC-only precinct names/codes. I called the Oklahoma County and Oklahoma State Election Boards (big thanks to them for the help), and the latter pointed me to results for a 2014 mayoral race in OKC. I was able to get codes for precincts that voted in this election, and thus I was able to get a set of 235 precinct codes for just OKC–only OKC residents voted for the OKC mayoral race. (Side note: Initially, I also assumed that the entirety of OKC was in Oklahoma County–that’s not true, as some OKC precincts are also in Canadian, Cleveland, and Pottawatomie Counties, but this wasn’t a problem with my new approach.) With this new set of data for OKC only, I was finally able to conclude that Donald Trump in fact beat Hillary Clinton in OKC, and not the other way around as I initially calculated. I apologize for this mistake. Thanks for all those on Twitter (especially John Kenney) and Oklahoman reporters and election board workers who helped me correct it. Given Nate Cohn’s collection of vote totals in other cities on Twitter, it seems very likely that Oklahoma City was the most populous (the largest “big city”) that Trump won in 2016, which only comes in as the 27th most populated in the US. Mesa, AZ is likely the next biggest city that Trump won, which ranks as the 38th most populous. Here’s the final data for OKC:
Total votes: 278,327