Largest city (or municipality) in each state to vote for Trump

[Adam Griffin]

Alabama - possibly Huntsville (4)

31% black and has a sizeable number of white HRC voters by Ala. standards.

Huntsville would be really close. Obama carried it by 1 in 2008 with the county as a whole going to McCain by 15. In 2016, Trump won the county by 16. Based on the broader trends in last year's election, it wouldn't surprise me if unincorporated Madison County swung to Clinton while Huntsville proper swung to Trump (compared to 2008). The real question is by how much?

Unfortunately, there is no real way to know given the fact that Madison County has since consolidated precincts with no regard for municipal boundaries.

[Adam Griffin]

Trump definitely lost Nashville and Memphis and definitely won Franklin (#7).  He likely won Clarksville and Murfreesboro (#'s 5 and 6) too, but I am not quite sure on Knoxville (#3) and Chattanooga (#4).  I might be able to run those numbers when I have some time!

Yeah, looks like I'm late to the thread, but I could have saved you guys some time by saying that there was no way Chattanooga voted for Trump. The city proper tends to fall between 55-60% Democratic in most straightforward elections. I believe Obama got around 58% in 2008.

The county results as a whole can be misleading; Chattanooga is half the county but probably slightly less in actual votes, but the non-Chatt portions of Hamilton County basically vote like any Deep South area (70% GOP or more). In Knox County there's a bit more homogeneity between Knoxville and the rest of Knox than in Chattanooga/rest of Hamilton.

Of the four biggest cities, Knoxville is the most Republican; if you start with it, you can usually figure out whether you need to drill down more. If Knoxville is won by a Democrat, then you can almost certainly bet Chattanooga was as well - and by even more.

[Adam Griffin]

Alabama Update:

I'm somewhat convinced Huntsville (#4) has gone to Clinton.

I'm not entirely convinced that this is true, for reasons I outlined on the first page:

Huntsville would be really close. Obama carried it by 1 in 2008 with the county as a whole going to McCain by 15. In 2016, Trump won the county by 16. Based on the broader trends in last year's election, it wouldn't surprise me if unincorporated Madison County swung to Clinton while Huntsville proper swung to Trump (compared to 2008). The real question is by how much?

Unfortunately, there is no real way to know given the fact that Madison County has since consolidated precincts with no regard for municipal boundaries.

The more I've thought about it, the more I tend to believe that the broader discrepancy in 2016 when compared to 08/12 (that in bigger counties that swung slightly to Clinton and Trump, the more urban areas saw Democratic turnout drop and the more suburban areas swung to Clinton) would apply here. Huntsville's Democratic base is heavily reliant on black voters: exactly who would and did drop off in big numbers throughout the South and beyond in 2016.

When I think about the '08 numbers I posted above along with that, I imagine the gap (16 points in 2008) would have narrowed somewhat. Trump won the county by 1 point more than McCain, but it very well may be that instead of the Dem winning Huntsville by 1 and losing the rest of Madison by 30, it's a case of the Dem losing Huntsville by 3 and losing the rest of Madison by 25 or so.

How did you parse out the results? When I looked at Huntsville versus Madison's current precincts, it looked like too much of a hot mess to really accurately gather results by merely adding up precincts. Are you using Census data to approximate population portions that are in the city boundaries of each split precinct or comparing side-by-side elections on the same day in both Madison and Huntsville to get an idea based on proportionate turnout/votes cast?

[Adam Griffin]

I used census data to help approximate what percent of a precinct's population is not within municipal borders. It's not as bad as one might think in some areas, because much of the annexed territory has next to no people living there in some parts of these salients.

Also, your intuition about Clinton losing ground due to lower black turnout is correct and gaining ground in suburban areas is correct, but the conclusion drawn from that is wrong.

If you look at the stretch of development from 4 Mile Post Road SE to the Tennessee river, all of that Suburbia (territory one would expect to swing to Clinton, and indeed did so by a strong amount) is within Huntsville proper. There are also White areas closer to the urban core that swung strongly to Clinton.

Rural and exurban areas still swung to Trump as well.

All of the precincts that Clinton carried by a strong margin are completely inside Huntsville (in fact, every precinct Clinton carried in Madison County, with the exception of Triana City Hall precinct, is entirely within Huntsville proper). The question is simply, "are there enough Trump areas left to overcome Clinton's margins in Northern Huntsville". Even making favorable assumptions regarding the division of complicated precincts, I don't think one can justify Trump doing well enough in remaining Huntsville to win the city.

After cleaning up the formatting, I'll post my data here sometime later today, but I stand behind it.

Sounds like you've done the homework, so I'll defer to your data. I guess it does make sense given that Huntsville has a decent amount of sprawl and would take in a lot of areas where Democratic improvement over 2012 would be naturally present - that combined with the broader white demographic makeup of the city could definitely be enough to counteract black drop-off relative to 2008's performance.

[Adam Griffin]

Here's a spreadsheet for Arlington:

Arlington 2016 Spreadsheet

I haven't taken a close look at the data to see where partial precincts might be an issue (I did notice a few where there were 1000+ presidential votes but practically 0 Prop 1 votes), but from a quick look at the old DRA data, precinct boundaries conformed pretty well to municipal ones in previous cycles. I noticed that the presidential race had about 17k more votes from these combined precincts than the Proposition 1 total, so it's likely there are a few precincts that need to be removed from the presidential column due to there being almost no voters in municipal boundaries in those precincts.

With that being said, I think even after the numbers are polished, Clinton carried it unless those partial precincts are overwhelmingly Democratic:

Candidate	# Votes	Percentage	Margin
Clinton	65,404	49.13%	+3.37
Trump	60,910	45.76%
Other	6,808	4.71%

[Adam Griffin]

^^^ Just did a quick comparison of vote totals between Prop 1 and 2016-Pres for each Arlington precinct where there was a discrepancy. After all was said and done, the margin shifted in favor of Clinton by about 300 votes (+/-100); virtually no change. I'm very confident Clinton won Arlington by about 3.5 points and just shy of 50%.

EDIT: corrected # typo in previous post, but did not change prelim count to reflect further precinct analysis mentioned here

[Adam Griffin]

Maryland:

Glen Burie (#7) confirmed Trump by precinct summation:

Trump - 9,764 (48.28%)
Clinton - 9,049 (44.75%)

Arizona

As for Arizona, I don't think Mesa (#3) needs to be checked (it doesn't look like a Trump blowout, but I don't see how Clinton carried the place), but I guess someone could do it if they really wish to check it.

I'm going to take a look at Mesa just out of curiosity; I'm transposing the data from the Maricopa County's crappy PDFs and am about halfway done (going in numerical order; currently 52-38). Gotta step out for a bit but will finish it when I return.

[Adam Griffin]

Mesa, AZ (excluding precinct splits):

Trump 94793 (53.13%) +15.77
Clinton 66655 (37.36%)
Johnson 9076 (5.09%)
Other 5396 (3.02%)
Stein 2512 (1.41%)