County Prediction Map

[user12345]

Benchmark Politics has released a map of their 2016 prediction county by county and the lead they're predicting each candidate has there. http://benchmark.shareblue.com/counties/

[Mr. Morden]

Zero counties for 3rd party candidates?  McMullin comes up empty in Utah?

[NOVA Green]

The Oregon map looks feasible, although I think that Deschutes and Jackson (Wealthier California retirees will trend D) will be a couple more points favorable to Clinton than the model suggests, Yamhill will likely flip to Clinton and Tillamook could go Trump and doubt that Clinton will win there by 10%. Columbia will also be a little bit closer I suspect.

Also, heavily Mormon and Latino counties in Eastern Oregon will have a greater swing against the Republican nominee than the Romney 2012 numbers, so some of these counties seem a bit high on the margins.

Margins in Lane and Benton are probably a bit high for Clinton, with 3rd Party candidates over-performing among Millennials and Middle-Aged Bernie supporters.

[NOVA Green]

Some of their other county level results seem counter-intuitive glancing at a few other states....

Texas for example in Bexar County (San Anontio)  they only have at +9 Clinton, many counties in WestTex with a high proportion of Latinos that swung Democrat in 2012, still running similar margins in 2016, despite Trump likely significantly under-performing among Latinos in Tejas, but a wealthy suburban county outside of Houston Montgomery barely moving at all between '12 and '16, despite a high proportion of college-educated Anglos.

[Bakersfield Uber Alles]

Just took a quick glance at California. I feel like Orange will go Dem. I doubt that San Bernardino or Riverside will either. Trinity seems like it could go for Clinton in a normal year, but not against Trump. I'll have to make a map for CA and NV. I've done ones for Idaho and Utah that I need to post. The lack of McMullin counties in Utah disturbs me. I have McMullin winning a few in Idaho as well.

[NOVA Green]

Just took a quick glance at California. I feel like Orange will go Dem. I doubt that San Bernardino or Riverside will either. Trinity seems like it could go for Clinton in a normal year, but not against Trump. I'll have to make a map for CA and NV. I've done ones for Idaho and Utah that I need to post. The lack of McMullin counties in Utah disturbs me. I have McMullin winning a few in Idaho as well.

I have to give Benchmark and A for effort on this, and it looks like their proprietary model uses a mix of their traditional demographic based modelling from the primary season, combined with "live updates" that somehow adjust for state/national polls, which at least gives us a really cool resource to look at on a regular basis.

It does appear that their model doesn't necessarily include historical voting patterns by county so you see some weird outliers like Butte County California, which is a major college town only going +2 Clinton, which would be an astonishing reverse, but places like San Bernadino & Riverside going Trump solely because of the percentage of old Anglos living out on the Eastern regions, and high upper-income voters in the Inland Empire somehow causing these counties to swing massively towards Trump.

Additionally, it doesn't appear to incorporate data regarding religious identity, so Utah looks a bit weird as does certain counties in SE Idaho that are 90% Mormon, where we know there will be a major swings as a result of 3rd party votes.

Elections are won and lost on the margins, so although I really dig this project and what Benchmark is doing here, there are some strange county results (Still working through the states) that would completely defy any logic or conventional wisdom even within the context of this very unusual GE.

[Bakersfield Uber Alles]

Just took a quick glance at California. I feel like Orange will go Dem. I doubt that San Bernardino or Riverside will either. Trinity seems like it could go for Clinton in a normal year, but not against Trump. I'll have to make a map for CA and NV. I've done ones for Idaho and Utah that I need to post. The lack of McMullin counties in Utah disturbs me. I have McMullin winning a few in Idaho as well.

I have to give Benchmark and A for effort on this, and it looks like their proprietary model uses a mix of their traditional demographic based modelling from the primary season, combined with "live updates" that somehow adjust for state/national polls, which at least gives us a really cool resource to look at on a regular basis.

It does appear that their model doesn't necessarily include historical voting patterns by county so you see some weird outliers like Butte County California, which is a major college town only going +2 Clinton, which would be an astonishing reverse, but places like San Bernadino & Riverside going Trump solely because of the percentage of old Anglos living out on the Eastern regions, and high upper-income voters in the Inland Empire somehow causing these counties to swing massively towards Trump.

Additionally, it doesn't appear to incorporate data regarding religious identity, so Utah looks a bit weird as does certain counties in SE Idaho that are 90% Mormon, where we know there will be a major swings as a result of 3rd party votes.

Elections are won and lost on the margins, so although I really dig this project and what Benchmark is doing here, there are some strange county results (Still working through the states) that would completely defy any logic or conventional wisdom even within the context of this very unusual GE.

I have to say that I'm at least partially mistaken. San Bernardino and Riverside did go for Obama in 2012. I might've been mixing them up with the 2014 Gubernatorial race in which they both voted for Kashkari. Butts and Nevada Counties seem to go back and forth between the Dems and Reps. The problem is that those two are relatively small counties and often polling in CA seems to be vague outside of saying Bay Area, LA, other SoCal, so it can be hard to tell, which is infuriating. I still stand by my Orange County prediction though.

I was disappointed by the last two Utah polls because they don't have a breakdown by region. I'd love to see some polling in Idaho, especially to see if McMullin support spills over the border to SE Idaho (McMullin is fully on the ballot there as well).

[Bismarck]

Trump will win Vermillion county Indiana, and probably Vigo as well.

[Nyvin]

Elliot KY: +8% Hillary

hah,  I'm having a hard time believing that.

[Chunk Yogurt for President!]

Elliot KY: +8% Hillary

hah,  I'm having a hard time believing that.

Trump really isn't that popular in Kentucky.

[Devils30]

The model is too simplistic. In Pennsylvania it is probably underestimating Trump in places like Luzerne and overestimating him around Philly.

[evergreenarbor]

Washington looks about right, though I don't think Hillary will win Whitman or Cowlitz by 10 points

[heatcharger]

I'm too lazy to run a calculation but I think Arizona would flip with those margins.

[Senator-elect Spark]

HM.

[Interlocutor is just not there yet]

Just took a quick glance at California. I feel like Orange will go Dem. I doubt that San Bernardino or Riverside will either. Trinity seems like it could go for Clinton in a normal year, but not against Trump. I'll have to make a map for CA and NV. I've done ones for Idaho and Utah that I need to post. The lack of McMullin counties in Utah disturbs me. I have McMullin winning a few in Idaho as well.

She'll win the IE + OC and she'll close the gap slightly in Kern County. Just saw that registered Democrats now outnumber Republicans in Bakersfield.

Only speaking as a Riverside County resident, but I'd be stunned if Trump won here or San Bernardino. In fact, I wouldn't be surprised if Hillary wins the Inland Empire by a bigger margin than Obama '08. Voter registration has really helped the Democrats this year: http://www.pe.com/articles/county-810531-voter-democratic.html (And the gap has grown more since this was written)

[RI]

Very conservative predictions.

[BaldEagle1991]

Oklahoma counties will still be all red.

[Gass3268]

In the primaries their county maps sucked, but they were very good at getting the final statewide %. In many cases better than the other "experts" and projections.

[Dr. Arch]

No blue counties in AK seems like a longshot imo.

[HAnnA MArin County]

I hope she can win Staten Island. It just always looks so aesthetically ugly red/Atlas blue surrounded by all the blue/Atlas red.

What's up with Crawford County, Kansas being a Hillary county? I know Obama barely won it in 2008, but still random.

HaHaHa at Rowan County, Kentucky being for Hillary. Someone alert Kim Davis.

[nclib]

No blue counties in AK seems like a longshot imo.

They put every county at the same margin as the general, i.e. not giving county results for AK.