2012 White Obama Lovers Map by County Project: COMPLETE!

[Bandit3 the Worker]

I notice that even among whites, Obama did better in the urban areas, but you can still see traces of the pre-2000 voting patterns on this map (from before the days of the urban/rural split). For instance, despite Cincinnati being a huge urban area, Obama generally does better the further you get from Cincinnati, until you pass the next closest major cities. On this map, Obama actually did better in some very rural areas than in some very urban areas. It's like it's 1988 all over again.

Cincinnati used to be much more conservative than the nearest other metropolitan areas - Dayton, Lexington, Louisville, Indianapolis, Columbus, Huntington - even though Cincinnati at the time was bigger. This map has traces of that old pattern.

[I spent the winter writing songs about getting better]

So Obama did better than whites in Indiana than Virginia and almost as well as Florida. Kind of puts things into perspective about the demographics of those states considering how they all vote.

Kind of reminds me of a point made on DailyKos once, Obama did better amongst whites in Michele Bachmann's district than he did in the country at large.

[Bandit3 the Worker]

Kind of reminds me of a point made on DailyKos once, Obama did better amongst whites in Michele Bachmann's district than he did in the country at large.

This is part of the old pre-2000 pattern. Minnesota used to be one the very most Democratic states, but now it's been surpassed by more urban states.

[ElectionsGuy]

I have a question. If Philadelphia voted 85% Obama, with whites voting 81% Obama and blacks voting around 95% Obama (I think that's a safe assumption), how does that add up to 85%? Do Philadelhia blacks turn out badly?

[traininthedistance]

Pennsylvania has been added.

PA 2012 statewide share of the white vote for Obama: 44.0%


I invite those familiar with PA to give it an extra-close inspection, as the general variance in color throughout many areas makes me wonder if I could have made any mistakes. This may just have to do with PA having relatively few, large counties.

The large populated areas all pass the smell test– in particular I am not surprised to see so much of a difference between Lackawanna and Luzerne, though you might have been.  

What is an issue– and it is an issue in the majority of states, not just PA, is rural prisons skewing the numbers in certain low-population counties, by basically air-dropping nonvoting minorities into heavily white areas.  I was wondering about Forest County, and checked out its townships in Wikipedia, and lo and behold it has a 274-bed juvenile detention center:

http://en.wikipedia.org/wiki/Howe_Township,_Forest_County,_Pennsylvania

Not a lot of beds... but given how small Forest is, that moves the needle quite a bit.  I'm sure there are several other examples in PA, which I could try and dredge up at some point for you.  And, also, examples in many other states.

Correcting for this phenomenon will be a pain, I'm sure.  Sorry.

I actually have been doing this, at least for the South, because...well, it's obvious why. MO, WV, OH, PA, MD & DE were not assessed for prison populations, in part because I was expecting the racial discrepancy between the prison and the county it is in to be less than in southern states. For these states, I do need to probably go back and compile a list of the counties where institutions are and assess the situation, but I was hoping to get by except in cases where a county appeared very out of place with its neighboring counties.

You should be safe with WV (since it is so white overall) and DE (because there are only three counties and they're all large).  The others are probably worth checking out: Forest is probably the largest outlier but I wouldn't be surprised to see a couple others here and there.

[YaBoyNY]

I have a question. If Philadelphia voted 85% Obama, with whites voting 81% Obama and blacks voting around 95% Obama (I think that's a safe assumption), how does that add up to 85%? Do Philadelhia blacks turn out badly?

Whites generally turn out more than blacks in Philadelphia, even with the Obama campaign's excellent attempts at increasing minority turnout.

[Adam Griffin]

I have a question. If Philadelphia voted 85% Obama, with whites voting 81% Obama and blacks voting around 95% Obama (I think that's a safe assumption), how does that add up to 85%? Do Philadelhia blacks turn out badly?

Whites generally turn out more than blacks in Philadelphia, even with the Obama campaign's excellent attempts at increasing minority turnout.

WARNING: LONG BUT VERY INTERESTING POST AHEAD

Yes, there are multiple interesting observations I have made over the years of doing this sort of projection, and oddly enough they almost all involve black voting populations (either being an influence on other voting blocs or the subject of the intrigue directly). Hopefully I'm not using an example here that would seem offensive, but it reminds me of a study I once read about where when a population of deer on an island began to exceed its sustainable population, all sorts of interesting things started organically happening to the deer (one notable observation was that rates of homosexuality among the deer skyrocketed). It's as if in some cases there are magical, subconscious effects being felt throughout communities and influencing voting trends.

The first observation relates to this quote: the larger the black population in a county, the more likely they are to be a lower share of the vote (but only after a certain point). In counties with very small black populations, their share of the electorate by my observations tends to be larger - but not dramatically larger - than their share of the electorate. Once you begin to encounter counties that are 20, 30 maybe even 40% black, there's more range to it: black share of the population usually isn't under-represented, but it will either be on par with the population or larger; in the latter case, larger (proportionally) than in the counties with less than 20% black populations. Once you get close to being a majority-black area, however, the discrepancy begins heading in the opposite direction and blacks become a (in some cases, substantially) smaller share of the electorate. This seems to be the case in smaller and larger counties, and even in cases where the share of the black population is small as a percentage but the county/number of blacks overall is larger, it holds true (the point in mentioning this is that in larger counties where blacks may be 10% of the population, there is enough socioeconomic variance for them to be just as representative of a much larger community all the same). I've been able to explicitly confirm this in Georgia, where voter registration and turnout by race is recorded by SoS.

Basically, something like this:



The second observation I posted about on the first page of this thread. Essentially, there are areas - often Southern Black Belt counties - that are solidly-majority black (you start seeing the trend at 55% black or so, but really takes off once you're in the 60s) where whites suddenly become more Democratic than their counterparts in racially-divided or majority-white counties with substantial black populations. These are counties where, no matter how you tinker with black share of the electorate or even literally maximize black vote to 100% D, that you can reasonably come to the conclusion that whites are not voting substantially more Democratic than in neighboring counties. Most states do not record voter registration by race, but Georgia records both voter registration and turnout by race, which allowed me to see that this was in fact true. Here is an example:

2012 electorate, Hancock:

77.34% black
21.21% white
1.45% other

The others/unknowns don't even play a significant role here. Using the formulas, I get this:

(77.34*0.95) + (1.45*0.65) = 74.41%
80.92 (Obama vote %) - 74.41 = 6.51
6.51/21.21 = 30.69% of white vote for Obama

Even if you count all of the unknowns as whites: 32.87%
And count all of the unknowns as blacks: 28.62%

Still using Hancock as a example (since I've been too lazy to go back and do the others yet), let's assume 100% black support and 100% of all other/unknown voters are black as well (which means Obama got 100% of their votes, also):

(77.34*1.0) + (1.45*1.0) = 78.79%
80.92 (Obama vote %) - 78.79 = 2.13
2.13/21.21 = 10.04% of white vote for Obama

That'd match the surrounding counties pretty well, but is a huge stretch to assume that every single voter who wasn't white voted for Obama, and that every unknown voter was also non-white.


Let's consider another scenario, where Obama gets 97% of the black vote and 80% of the other vote (in line with your theory that other/unknowns match that of the remainder of the electorate):

(77.34*0.97) + (1.45*0.8 ) = 76.18%
80.92 (Obama vote %) - 76.18 = 4.74
4.74/21.21 = 22.35% of white vote for Obama

This would still be quite a bit out of whack with a lot of the other surrounding counties.

Pictured here is Hancock County (31%), the fifth-most white Democratic county in the state (only behind core ATL counties and Athens-Clarke) and blackest county in the state:



You see this trend also in Greene County, AL (blackest county in the state), as well as to a lesser degree yet statistically significant all throughout the black belt. Again, at least in Georgia's case (all of the other states but 1 require refined projections to be made), I was able to explicitly confirm the existence of this trend via turnout by race, and there are very few (often less than 2%) non-white, non-black voters here to make the math difficult.

So basically, this:

[smoltchanov]

Very interesting, thanks! I thought about such numbers too and came to very simple (may be - even primitive) hypothesis, which could, naturally, be very wrong, but still: whites, who continue to live in black-majority counties, learned (and learned very good) how to live in counties governed by blacks. They adapted well, Those, who cound't - moved. Mostly - to areas with relatively few blacks. And "border counties" (with 40-45% blacks) are exactly those where power still mostly "belongs" to whites, but situation may change very soon. Here - the obstinance and desire to hold onto that power as long as possible and delay impeding transition for that time.

In some aspect the same hypothesis explains why in 50th or 60th it was different - then it were whites from black-majority counties (like WCC founder from Indianola, Mississippi) who were most intransigent in their support of segregation. They knew exactly well that they will lose most of their political power almost immediately after blacks begin to vote "en masse". Not so in, say, hilly NE Mississippi, with most counties being 20-25% black - of course whites there were mostly for segregation too, but integration was less threat for their political power..

[Adam Griffin]

Very interesting, thanks! I thought about such numbers too and came to very simple (may be - even primitive) hypothesis, which could, naturally, be very wrong, but still: whites, who continue to live in black-majority counties, learned (and learned very good) how to live in counties governed by blacks. They adapted well, Those, who cound't - moved. Mostly - to areas with relatively few blacks. And "border counties" (with 40-45% blacks) are exactly those where power still mostly "belongs" to whites, but situation may change very soon. Here - the obstinance and desire to hold onto that power as long as possible and delay impeding transition for that time.

In some aspect the same hypothesis explains why in 50th or 60th it was different - then it were whites from black-majority counties (like WCC founder from Indianola, Mississippi) who were most intransigent in their support of segregation. They knew exactly well that they will lose most of their political power almost immediately after blacks begin to vote "en masse". Not so in, say, hilly NE Mississippi, with most counties being 20-25% black - of course whites there were mostly for segregation too, but integration was less threat for their political power..

I don't think I've ever mentioned it here (maybe in chat), but your first paragraph is exactly what I have suspected, too. They no longer have some irrational "fear of the unknown" and realize that the world won't end.


Illinois has been added.

IL 2012 statewide share of the white vote for Obama: 45.4%

[Adam Griffin]

And just for quick reference thus far, a map showing who won the white vote by county. Because of there being a greater likelihood of a county's white share of the vote for Obama being under-represented than over-represented, it displays all counties where Obama won 48% of the white vote or more.

[Mr. Illini]

Excellent work!

[Skill and Chance]

And just for quick reference thus far, a map showing who won the white vote by county. Because of there being a greater likelihood of a county's white share of the vote for Obama being under-represented than over-represented, it displays all counties where Obama won 48% of the white vote or more.

Interesting, so Allegheny and Fairfax were <50% wins for Obama?

[TDAS04]

Southern Illinois looks a lot like Indiana, no surprise.

[Adam Griffin]

Interesting, so Allegheny and Fairfax were <50% wins for Obama?

49% each, according to my calculations.


Since this graphic is now getting so large, I am making preparations to convert the main graphic over to a Google Fusion Tables Map, which can be viewed here. It does not currently have all of the states I have completed (some states' files are on another computer), and there are currently a couple of small errors I will need to fix (Augusta is missing in VA, two MS counties are missing and SD appears randomly on the map). This will be the only way in which to display all of the data in one map that everyone can access, and despite my therapeutic love for coloring in maps by hand, this will be more efficient and probably more aesthetically pleasing in the end as well.

Additional note: if you see counties on the new FT map that do not match the shades of the one on my original map, this is because I was lazy and did not manually adjust the original spreadsheets after making the corrections. I'll be fixing these shortly as well.

[ElectionsGuy]

Is there any specific reason why Kosciusko, Indiana is so Republican?

[Adam Griffin]

Is there any specific reason why Kosciusko, Indiana is so Republican?

The only thing that really stuck out to me demographically was this from the 2000 Census:

You must be logged in to read this quote.

[Skill and Chance]

Is there any specific reason why Kosciusko, Indiana is so Republican?

The only thing that really stuck out to me demographically was this from the 2000 Census:

You must be logged in to read this quote.

So that means lots of stay-at-home moms?  Or are the medians calculated only among those with incomes?

[Adam Griffin]

Is there any specific reason why Kosciusko, Indiana is so Republican?

The only thing that really stuck out to me demographically was this from the 2000 Census:

You must be logged in to read this quote.

So that means lots of stay-at-home moms?  Or are the medians calculated only among those with incomes?

I'm pretty sure it's just a per capita measurement of individuals over the age of 15 - could be wrong.


While I may continue with the PNG-based map in the future, I have discontinued it for the time being. The interactive map is now fully operational and deployed, and I will be updating it instead of the PNG for the foreseeable future. If for some reason you cannot see the map from its hosted page, please let me know - there might be some existing permissions issues to work out.

Here is a screenshot of and link to it, which can be fully accessed from the original post. You can click on each county for its name and the share of the vote Obama received.

New Interactive Map

[I spent the winter writing songs about getting better]

That type of gender gap is usually indicative that the economy is based primarily off jobs that employ mostly men, like construction or factories. You see a similar effect in the coal counties of West Virginia too.

[Adam Griffin]

NJ has been added.

2012 statewide share of the white vote for Obama: 44.7%

[ElectionsGuy]

Adam, you may want to check out Florida. There's some counties that are probably mixed up (I noticed it when I clicked on Seminole County and it said Santa Rosa (where Obama got 18%!)

[dpmapper]

Richmond city and Richmond county in Virginia have their percentages reversed, it seems. 

[Adam Griffin]

Adam, you may want to check out Florida. There's some counties that are probably mixed up (I noticed it when I clicked on Seminole County and it said Santa Rosa (where Obama got 18%!)

Thanks for the heads-up. It looks as if there were five counties in Florida that ended up transposed (Seminole, Saint Johns, Saint Lucie, Sarasota & Santa Rosa) due to the "Saint"/"St" difference/ordering between my geography column and my data column. I did my best for each state to check for that but in the case of the "Saints", I may have simply looked at the beginning of each county without double-checking that the right Saints were in the right order.

It should be showing up fine now (it may take a hard refresh, depending). Let me know if any more errors appear.  

Richmond city and Richmond county in Virginia have their percentages reversed, it seems.  

Well, crap - and just as I was making the post above. By default, Baltimore's was like that as well when I added to GFT, but I caught it. Should be fixed now (as I said above, a hard refresh may be needed before it shows).

[YaBoyNY]

Interesting to see the well-explained difference between Northern and Southern metros illustrated. Democrats win the Northern white vote not only in the cities themselves, but in many of their suburbs and satellite cities as well, and keep it close in the areas where they don't win.

[smoltchanov]

Interesting to see the well-explained difference between Northern and Southern metros illustrated. Democrats win the Northern white vote not only in the cities themselves, but in many of their suburbs and satellite cities as well, and keep it close in the areas where they don't win.

Exactly. And in the South suburbs (even more - exurbs) are, usually, heavily republican, and frequently not only cancel Democratic majorities from big cities, but outweigh them. To some extent that phenomenon exist in some Midwest (Milwaukee, for example) metropolitan areas too..