Talk Elections

I’ve been making random Ohio maps by county for a while now and have been thinking of posting them. I briefly posted one in my signature but otherwise haven’t showed any on this site. Let me know if anyone finds them interesting or whether I should include the county names or not.

Also let me know if anyone has some ideas about other demographics that might be interesting.

I’ll start off with the 2008 Presidential Election since that’s one people know a lot about:
()
It’s mainly a battle of the northeast and the lakeshore versus everywhere else. The southeast, which is normally very swingy and often determines the outcome of state elections, swung heavily toward McCain. The northwest, which apart from Toledo is normally fairly Republican, had a very soft Republican margin. You’ll see in subsequent maps how different this was from normal.

Other places that stick out: Franklin County (Columbus) was solidly Obama while Hamilton (Cincinnati) and Montgomery (Dayton) were softer Obama Counties. Republicans win Montgomery and Hamilton more often than not. Athens County is very heavily Democratic and this will manifest itself on many maps. The Republican areas that stand out are the upper Miami Valley (the clump of 70%+ McCain counties in the northwest—this area will stand out often) and Holmes County (the Amish—this will also stand out on a lot of maps).

Compare the first map to the 2010 gubernatorial election where John Kasich (R) narrowly defeated Ted Strickland (D):
()
The biggest thing to notice is that the southeast swung back toward the Democrat this time. Athens County really jumps out. Strickland was doomed, however, because he did terrible in northeast Ohio and Columbus, failing to win either of the swingy counties (Lake and Stark) and even lost Portage County. Geauga and  Medina Counties move even further to the Republicans and Lorain, Summit, and Ashtabula are too close for a Democrat to win statewide with this type of showing.  This election is more polarized by economic issues and less by social issues as the suburban counties are much more Republican. Mercer County is still heavily Republican but the rest of the upper Miami Valley doesn’t stand out. Holmes County is heavily GOP as always. Two suburban Cincinnati Counties (Clermont and Warren) also join the 70%+ GOP ranks.

Now, for a random race no one has paid any attention to, the 2010 Ohio Supreme Court Chief Justice race between Maureen O’Connor (R) and Eric Brown (D). This was between two sitting justices for the top spot. Ohio’s Supreme Court races are always heavily slanted toward the Republicans and at one point the Democrats controlled 4 of the 5 statewide executive offices and the Republicans simultaneously held all seven Supreme Court seats. Here’s the map:
()
O’Connor won all 88 counties so it was rather boring. The southeast actually sticks out as the most heavily Democratic area.

Here’s the one I’ve already shown (and the only one with names on the counties): Percentage Catholic by county
()
Noticeable things:
Ohio’s Catholics are concentrated along the lake, in the upper Miami valley, and near Cincinnati and Youngstown. The Miami Valley has two counties (Mercer and Putnam) that are more than 50%. My home county of Erie has more than average. The southeastern and rural central parts of the state have virtually no Catholics and few live in Columbus.

Here’s Ohio’s 2004 referendum to ban gay marriage and civil unions (constitutional amendment):
()
Note: the coloring scheme is centered around 60% not 50%.
The vote passed 87/88 counties, only failing in Athens County, which sticks out a ton compared to the counties around it. The county the ban passed by the lowest margin was actually Franklin (Columbus) rather than Cuyahoga (Cleveland). The areas that stand out for the ban are the Ohio River counties, especially west of Athens, the upper Miami Valley, and Holmes County. Compared to partisan results, the ban did poorly in the Cincinnati and Dayton areas, as well as in normally Republican Geauga County in the northeast and swingy Wood County in the northwest. The ban did much better than Republicans do in the Youngstown area, carrying heavily Democratic Mahoning and Trumbull Counties with over 60%. My home county of Erie also appears more conservative than usual.

Now for some economic data, here's median household income (in thousands of dollars per year):
()
The poorest areas are in the southeast as expected. North-central Ohio is also rather poor, as are Lucas County (Toledo) and Cuyahoga County (Cleveland). Columbus, Cincinnati, and their suburbs are the wealthiest. Delaware County jumps out being the only county with a median household income over $80k a year. The exurban counties around Cleveland are fairly wealthy, Geauga and Medina in particular since those are the exurban counties without any large, poor, industrial cores.

This is a more complicated economic map, April 2011 unemployment:
()
Again the Ohio River counties fair poorly, but this time the pain is also spread across much of the northwest as well. The high unemployment in the northwest is likely an important cause of Obama’s good performance there and whether or not northwest Ohio remains happy with Obama in 2012 will be an important determining factor in the presidential election. The areas that stand out with low unemployment are the metropolitan areas of Cleveland, Columbus, and Cincinnati. Mercer County also stands out on this map, though I’m not sure why. Individual counties that stand out in a bad way are Ottawa in the northwest and Pike and Clinton in the southwest. Clinton County is home to Wilmington, which lost a huge number of jobs when DHL left and was featured in special interest stories in the national news about people coping with economic despair as a community.

This next map is the most difficult by far to read, average household size:
()
This map shows a strong urban/suburban divide with rural areas mostly in between. The Ohio River counties show lower household sizes and are older. Counties with large colleges also show small households. The only county that really jumps out on the large family side is Holmes, home of the Amish.

My last map for today is a map of shame: abortions per 10,000 people (Note: This is based on 2007 abortion statistics and 2008 population estimates.):
()
The areas that jump out immediately are Cuyahoga County and all of northeast Ohio. The other counties with many abortions are Franklin (Columbus), Hamilton (Cincinnati), and Lucas (Toledo). The counties containing poor, urban cores do much worse on this map, including my home in Erie County (Sandusky). The southeastern part of the state has the fewest abortions of any region, particularly Jefferson County (Steubenville) and Lawrence County (Ironton). The upper Miami Valley and Holmes County, the respective Catholic and Amish areas, also show noticeably lower rates. Allen County (Lima) is in the upper Miami Valley but also contains an urban core and has an intermediate rate.

This is awesome. Just wanted to let you know. :)

Ditto. TJ is one of our jewels in the crown, he really is. Just a little Indian reference there for you sbane. :)

Are the rural areas in the NW German?   Is the variation due to colonization efforts and where the colonists came from.  That is the pattern in the Texas Hill Country and also between Houston and San Antonio, where you will have some very Catholic towns and some very Lutheran ones. 

In Cleveland, Croat and Irish?  The southeast would be Scots-Irish with Columbus being an overlay of that, and you had a bit of industrialization in Steubenville and other areas just downstream from Pittsburgh.  Otherwise you could go Adams-Hardin-Carroll and everything southeast as the low percentage.

When I was drawing districts in Columbus, I was surprised at how many of the roads had German names.  It made it seem like Houston.

The Northwest is primarily German, even among non-Catholics. There are a couple counties with very high Lutheran percentages as well, such as Henry County which is majority Lutheran. The German influence stretches over that entire corner of the state, even over to Erie County. The Miami valley was settled by mostly Germans. This accounts for the strip of higher Catholic counties: Shelby, Auglaize, Mercer, Allen, and Putnam. Mercer County was built around the man-made Grand Lake St. Mary's on the Miami. Shelby and Auglaize are less Catholic becuase both have towns further from the major German Catholic settlements and Allen County has a lower percentage because it contains the city of Lima, which has a noticeable black population and more industrialization (ie. people from anywhere).

Nrotheast Ohio has larger Eastern European immigrant populations from Poland, Croatia, Serbia, Lithuania, and a handful of other places, along with some Italians, Irish, and Germans. I sort of wish I had a way of looking at the distribution within Cuyahoga County because I suspect it is strongly skewed with a much higher Catholic percentage on the west side where all these groups except the Italians settled. I may try to write something about the East-West rivalry in the Cleveland area later, but there is a considerable social difference. There's an old saying in Cleveland that the Northeast ends and the Midwest begins at the Cuyahoga River.

The Ohio river likely has higher percentages because of industrialization in general, but Jefferson County specifically is a bit different because Steubenville has Franciscan University.

I think Euclid used to be heavily Hungarian of all things.

Here is the 2010 Congressional election by county and by district:
This map has the vote in counties containing multiple districts totaled:
()
And this one just has lines drawn in at the loss of the total county vote:
()
This election was a complete blood bath where the Republicans cleaned house in pretty much every area normally competitive, with the possible exception of the southeast. Northwest Ohio was very heavily Republican, quite the turn-around from the ’08 presidential election (I haven’t made an ’08 House map yet). Large parts of the upper Miami valley were over 80% for the GOP and several counties where McCain won somewhere in the mid-50s were over 70%. Marcy Kaptur, while still winning comfortably, did terrible for a Democrat everywhere in her district other than Lucas County, only winning Erie and her Lorain County areas by small margins despite containing some heavily Democratic towns. Steve LaTourette also cleaned up all his marginal areas to win comfortably. You can clearly see the east/west division within Cuyahoga County with Kucinich only taking 53% of the vote in OH-10. Overall, the Democrats only managed low-60s in Cuyahoga County. The southeast was a little more reluctant to vote for Republican Congressmen than presidential candidates, but the GOP managed to get a slim victory in both seats.

Here is the 2008 Congressional election by county and by district:
This map has the vote in counties containing multiple districts totaled:
()
And this one just has lines drawn in at the loss of the total county vote:
()

Compare this election to the maps above. This is a bloodbath in the opposite direction as four ConservaDems are elected to seats they don't end up holding after 2010. There is a characteristic backwards-C shape of Democratic areas common to many Ohio elections: along Lake Erie, the eastern border, and along the Ohio River. The southeast is very heavily Democratic. In all of OH-6 and OH-18, Holmes and Knox are the only Republican counties. OH-16 was nearly swept by Boccieri despite containing several normally Republican counties. Kaptur and Sutton both won every county by comfortable margins. Ryan carried his part of Mahoning County with 87%.

Still, the western half of the state was heavily Republican, as was Steve LaTourette's district.

Here’s a more complicated Ohio religion demographics map. I broke it into four groups: Catholic (green), Mainline Protestant (blue), Evangelical Protestant (red), and other (black). The data is from the ARDA Database and uses the ARDA’s designations for what falls under each group. The coloring is done so that the brighter each color is the more of that group is present.
()

There are a few drawbacks to this mapping technique, many religious groups do not keep detailed membership records so the map likely underestimates evangelicals and overestimates Catholics. I also disregarded the “unaffiliated” persons because it is nearly impossible to tell what their identity is. The percentage of “unaffiliated” persons tended to be highest in rural evangelical areas so I am disinclined to call them nonreligious or other.

The Amish are considered evangelical (apparently) so Holmes County is bright, bright red.

In general, the further south you go the more evangelical an area is, while mainline Protestants are concentrated in the northern part of the state. Near the lake, virtually all of the Protestants belong to mainline denominations.

The intensity of color in Catholic areas can be slightly misleading for a number of reasons. First, large urban areas tended not only to have more Catholics but also noticeably fewer Protestants than rural areas with a similar Catholic percentage would have. Second, green seems to mask red better than blue so the green appears a bit more intense when the Protestants are Evangelical.

It’s easy to tell why Southwest Ohio is so conservative from this map: there are very few mainline Protestants. The map has a whole lot of green, red, and brown in that area.

In the Northwest, I’d like to point out Henry County. It’s colored purple because it has roughly the same number of Mainline and Evangelical Protestants, but both are mostly Lutheran (ELCA=mainline and LCMS=evangelical).

Are Pickaway and Champaign and Ashland distinct, or is it simply a case of relatively few Catholics, and balance between evangelical and mainline Protestants?

Why is Delaware so Catholic?  Would it have been so say 30 years ago when there would have been less suburban growth?  (eg is it like Putnam and Mercer with an addition of suburbanites, or is it like Franklin with fewer Blacks, who tend to be Protestant.

Totally unrelated: why are street blocks in Cleveland so long?

Pickaway, Coshocton, Harrison, and most of the other bright blue southeastern counties are mainly Methodist. Ashland is roughly evenly split between Methodists and Lutherans and has surprisingly few Evangelicals despite being the home of an Evangelical University. Champaign is a mix of lots of different mainline denominations.

As for Delaware County, I’m not completely sure which the answer is but I have a decent way of guessing. There are two main population centers in Delaware County, the Columbus suburbs in the south and the city of Delaware. Of the four Catholic Churches in the county (20,003 Catholics and only four churches), only one is in the city of Delaware. The church in the city of Delaware appears to be the largest one, however. So my guess is that before suburban sprawl Delaware County was probably more Catholic than all the counties around it but not extraordinarily Catholic, more like Wyandot or Wood Counties rather than like Putnam or Mercer. As far as the sprawl is concerned, it alone wouldn’t explain why Delaware County is much more Catholic than the other exurban Columbus counties, Licking and Fairfield, but I do think it contributed.

I would guess the reason why Cleveland blocks are so long is that most our major streets radiate east or west outward from downtown and there is little reason to build many major cross-streets between them. For example, on the east side, Carnegie, Cedar, Chester, and Euclid are all major east-west streets between downtown and University Circle. There’s really not much in between and not much reason to travel perpendicular to those streets.

Good maps, TJ.

Why are Mercer and Putnam so Catholic and still so Republican?

Mercer and Putnam were settled by Germans rather than the later waves of immigrants that were targeted more for discrimination and account for most US Catholics. This area also never experienced the extent of industrialization and unionization of many heavily Catholic areas, making many of them ancestrally Democratic.

As a whole Ohio Catholics are slightly more Republican than Ohio voters in general despite living in more Democratic areas on average. According to the 2008 Gallup exit poll, McCain actually received 51% of Ohio Catholic vote.

Basically, these areas show how rural Catholics would vote without unions, industry, or immigration as issues. On the DRA, you can actually find a whole bunch of precincts in the rural part of Mercer Countywhere McCain won ~85% of the vote. In Putnam, it isn't quite as extreme but Putnam also doesn't have a city the size of Celina with precinct that McCain only got... gasp...58% o the vote in.

I think that has more to do with the fact that there are far more blacks (usu. Protestant) than Hispanics (usu. Catholic) in Ohio. IIRC, white Catholics were still quite a bit more Democratic than white Protestants in Ohio.

You are correct. (http://www.cnn.com/ELECTION/2008/results/polls/#OHP00p1)

I tend to suspect this is mostly due to other factors beside religion considering a huge percentage of Ohio's Catholics live in pretty solidly Democratic counties.

I wonder if there is an effect on voting to live in a religiously homogenous area? That might contribute the voting patterns of Mercer and Putnam Counties (which vote very similarly to Holmes, the Amish/Mennonite stronghold).

Otherwise the best answer I can give is that Mercer and Putnam didn't experience as much industrialization and unionization. Neither county has a really large urban core, though Celina isn't negligible.

Religion being just a signifier in that particular electoral context.

Here (http://en.wikipedia.org/wiki/Maria_Stein,_Ohio)'s an example of what the rural ~85% McCain part of Mercer County is like.

Putnam is a litle more difficult to get information on but it also doesn't have any areas quite as extreme as this.

Putnam and Mercer were the only two counties in Ohio that Al Smith won, and they both trended hard toward Kennedy in 1960, though he narrowly lost them. They weren't very receptive to Kerry though, not surprisingly. Both were heavily Democratic until 1948.

Here’s a look at each party’s “high water mark”, ie. what a map would look like in a total blowout in each direction.

The first map is the Republican blow-out. In 2004, the Ohio Democrats scrambled to find a candidate after failing to draft Jerry Springer in a quixotic bid to take down the popular Republican incumbent George Voinovich. The eventually settled on Eric Fingerhut for what would be one of the most forgettable statewide elections in recent memory as Voinoich cruised to a 28-point victory, winning every county.
()
Voinovich, the three term Republican mayor of Cleveland, has an opportunity to do something no other Republican can: actually win Cuyahoga County. With that, the map looks like a map of the locations in Ohio that have an ancestral Democratic presence. The most marginal areas are along the Democratic “backwards C” from the river to Toledo. There appears to be very little difference in Republican intensity between the counties that are always very strongly Republican and normally more marginal.

---

In 2006, the nation was on fire against George Bush and Ohio was even more on fire against Governor Bob Taft, who managed to get his approval rating into the single digits. No Republican would even pretend Taft was a competent governor. Incumbent GOP Senator Mike DeWine was a sitting duck against the very liberal populist Congressman Sherrod Brown.
()
Brown cleaned up in the Southeast and the north, ensuring himself an easy victory. Unlike in the Republican blowout, this map still shows some strong pockets of GOP support, though it also wasn’t nearly as large of a blowout because the Republicans did have a decent candidate and Brown isn’t in the same league as Voinovich in terms of political clout. Holmes County, suburban Cincinnati, and the upper Miami Valley remained strongly GOP. These areas only swung about ten points while most of the state swung more than twenty. Another oddity of this map is that even though it was a Democratic blowout, DeWine still won Hamilton County, which is normally swingy.

Wouldn't it be more appropriate to use Strickland's blowout victory over Ken Blackwell as the Democratic highwater mark?  Strickland won a higher percentage of the vote and more counties than Brown did.  

Also, even though Blackwell was a terrible candidate, he was mayor of Cincinnati for a brief period of time which could partially explain the Republican overperformance in Hamilton County in 2006.

2006 Governor: Strickland (D) vs. Blackwell (R)
()

Yes, you are correct Padfoot, this is a much better map for a Democratiuc high water mark. It allows us to see which counties will vote Republican no matter what the circumstances are. Of course, Blackwell actually did decently in Cincinnati because he was the mayor. The only counties that voted strongly for Blackwell were Mercer (but without its sibling Putnam this time), Hancock, Holmes, Warren, and Clermont. Warren and Clermont are the suburban Cincinnati Counties, Mercer and Holmes we’ve talked about quite a bit already, but we haven’t pointed out Hancock yet. Hancock is primarily city of Findlay, Ohio. Findlay has a unique history as the longtime home of Marathon Oil Company. Perhaps a connection to oil is the reason why Hancock was more Republican in 2006. After all, much of the national anger was directed at George W. Bush.

Voting trend between 2000 and 2008 presidential elections (Note: the trend is adjusted for the difference in the national totals between elections)
()

The first thing to notice is that the northwest swung strongly toward Obama. This area experienced a steep decline in manufacturing and a lot of people, either out of work or experiencing pay cuts, heard then Senator Obama express in heartfelt terms how he could be the man to change the course of the nation. Many northwest Ohioans heard him and wanted to give him a chance. However, this is likely a transitory trend; Northwest Ohio swung heavily toward Republicans in 2010 and in a recent Quinnipiac survey (https://uselectionatlas.org/FORUM/index.php?topic=141422.msg3048443#msg3048443) gave the President the lowest approval rating of any part of the state.

The next point is the trend of areas with large black populations toward the Democrats, primarily in Franklin and Hamilton Counties. Cuyahoga County, despite containing far more blacks than any other county, experienced only a small Democratic trend, probably because of a Republican trend on the blue-collar white west side.

Athens County and some of the surrounding counties experienced a huge Democratic trend. Athens trended more in the Democratic direction than any other county. This is unsurprising given Athens is in many ways Ohio’s most liberal county and home to a large college town.

Delaware County, Ohio’s fastest growing county, remained strongly Republican but much less so than in 2000. The suburban growth has transformed the area into a still Republican but less lopsided county. This may be the single worst piece of information for the GOP on the map.

The collection of counties with significant Amish populations (Holmes, Wayne, Ashland, Coshocton, Tuscarawas, and Geauga) also trended toward the Democrats though not that strongly. Perhaps McCain’s candidacy caused some enthusiasm problems compared with George W. Bush?

The four main areas that trended in the Republican direction are the upper Miami Valley, the southern tip of the state, the Steubenville/St. Clairsville area near the river, and the Mahoning Valley. The southern tip of the state is made up of rural, white, ancestrally Democratic evangelicals, so a Republican trend isn’t surprising. Jefferson and Belmont Counties are blue-collar declining industrial areas full of working class whites. The Mahoning Valley (Youngstown and Warren) trending Republican may become an important headline in the future if it continues because it is a populous area that has historically voted for Democrats by lopsided margins.

The oddball county I can’t explain the Republican trend in is Clark County (Springfield). Springfield is in part a college town for a liberal arts institution and also contains a sizeable black population. Clark County is historically moderate because of some very Republican rural areas and heavily Democratic Springfield. All in all, it seems like a county that should have trended toward President Obama but it trended significantly in the other direction.

Remember that the Amish, though otherwise socially ultra-conservative (like, off the charts), are pacifists. This certainly isn't enough to get the Old Order Amish to vote Democrat (indeed, many don't vote at all), but there is a certain constituency of people who are more integrated into modern life but have an Anabaptist cultural background who are basically conservative but would sympathize with the D's on Iraq. Notice also that Lancaster PA and Elkhart IN have a very large Obama swing.

That doesn't explain why they voted so strongly for George W. Bush in 2004. If anything I would guess that a conservative pacifist would have voted for Kerry and McCain rather than the other way around.

But Kerry was for the Iraq war, unlike Obama. The people I'm thinking of are not interested in the nuances of "bringing in our allies".

I'll fully admit that I'm pretty tentative about this explanation in the case of the Ohio counties - I'm more confident about Lancaster and Elkhart, since the swing is much larger there (as well as large Obama primary victories, which again you don't see in Ohio), and they have more liberal Mennonites living there, of a sort I have some personal experience with. I don't have such a handle on the Ohio communities, which seem to be more just strict old order Amish and then people who aren't Anabaptist at all. So it's possible that the trend is caused by something completely different here. But it is a real phenomenon in the Amish/Mennonite community nationwide.

Great thread, by the way. Very interesting.

    TJ, this is a great thread.  I read somewhere that one key reason for Ohio being more GOP leaning than Michigan historically is that in Ohio the working class is spread out in many smaller cities.  In Michigan the Detroit area concentrated huge amounts of working class voters back in the auto heyday, and this led to a more cohesive pro-democrat vote, according to this theory. In Ohio, the dispersal of workers all over the state made them more likely to associate with more people from different classes etc, and less likely to vote Democratic.

I decided to try and create an ideological index for each county, separating economic and social interests. I included the 2000 and 2008 presidential election and the 2010 gubernatorial race in both scores, the vote total on Issue 2 and the current unemployment rate on the economic index, and the 2004 gay marriage referendum, the number of abortions (in 2007) per 10000 people (using 2008 population estimates), and the percentage of people in each county that are religiously unaffiliated for the social index. I hoped to catch a wide array of factors in doing this though it’s difficult to measure vague concepts like ideology. Lower scores indicate a more conservative county and higher ones indicate a liberal county. The average economic score was 50 and the average social score was 46 (the average without weighting for the population of each county).

Here’s the economic map:
()

The most conservative county was Holmes, followed by Mercer and the most liberal county was Athens, followed by Mahoning. The fiscally progressive areas form the characteristic backwards ‘C’ over the Eastern part of the state (which we see pretty consistently).

And here’s the social map:
()

The most conservative county was Mercer, followed by Putnam, with Holmes and Auglaize not far behind. The most liberal county was Athens, followed by Cuyahoga.

I was hoping the maps would show some sort of contrast: ie. counties that came out differently by each measure. Much of the state appears fairly similar on both maps, but a few areas stand out. The counties along the Ohio River are more conservative socially than fiscally. The same with the extreme northwest corner of the state. Mahoning County (Youngstown) also is much more socially moderate despite being the second most fiscally liberal county in the state. The Columbus and Cincinnati suburbs are more conservative fiscally than socially, as are Greene, Portage, and Summit Counties. I don’t think this technique is a great way of comparing Holmes County to Mercer and Putnam in some sort of running for which is more conservative since it’s really comparing apples to oranges, although it is evident that Putnam County is more fiscally moderate than Mercer and Holmes.

In honor of the Wisconsin thread, here’s a Cuyahoga County map of Ohio’s gay marriage ban vote in 2004:
()

And the race map from the DRA before someone asks:
()

Basically, the black areas voted for the ban, most of the eastern suburbs voted against it, and most of the western suburbs voted for it except Lakewood.

I love your economic/social maps by county!

'Looks like I'd fit pretty well in Lawrence county, or really anywhere in the current OH-06.

TJ, I had mentioned in a thread months ago that OH-11 was the best CD against the gay marriage ban. Are the white areas of OH-11 especially liberal or are the black areas less anti-gay than expected?

I would guess that the whites in Shaker Heights, Cleveland Heights, etc. are the most socially liberal people in Ohio, save maybe the college-related people in Athens or government workers living in downtown Columbus. Case Western is a pretty big college, too.

Most of the black areas were about 50-60% for the ban (not sure how that compares to expected). But, yes the white areas in OH-11 are very, very liberal. OH-11 includes both the really large red clump of precincts and the other liberal area near downtown. Considering how Columbus is chopped up, I'd be rather surprised if OH-11 wasn't the most liberal congressional district on the issue. However, I would bet the ban did worse the new Columbus pack district.

Agree with everything except Case Western being a pretty big college. It only has about 4,000 undergrads. Also, the Case precincts were more (not that it's saying much) in favor of the ban than most of the white parts of Cleveland Heights and Shaker Heights.

Here’s the progression of congressional results from the past decade. You can see how well that last gerrymander worked to keep a 12-6 map for most of the decade until things went awry in 2006 after Bob Ney (OH-18) ended up in prison.

2002:
()

2004:
()

2006:
()

2008:
()

2010:
()

Hamilton County gay marriage ban and racial map:
()
()

One thing I noticed was that the black areas in Hamilton County were much more liberal on gay marriage than the black areas in Cuyahoga County. There were a large number of black precincts that voted against the ban whereas in Cleveland pretty much every black precinct voted for it.

Cincinnati also has quite a few majority black precincts that vote for Obama with rather underwhelming margins. In Cleveland, pretty much every black precinct is very, very heavily Obama.

Would it be possible to do a map on the differential between the SOS race (Husted) and the gubernatorial race (Kasich)?  I am curious why Husted did so much better than the other Republicans in 2010.

Could somebody explain how the white liberal areas are in Cleveland and Cincy? What kind of people live there?

I don't know much about Cincy, but there are three main clusters of white liberals in the Cleveland area. The largest is most of the eastern suburbs, Cleveland Heights, Shaker Heights, Beachwood, Pepper Pike, Orange, and Moreland Hills. Pepper Pike, Orange, and Moreland Hills, are extremely wealthy areas with large estates. They aren't monolithically liberal, but are socially liberal in particular, although you can easily find conservatives there too. Beachwood is predominantly Jewish and Jews often are very liberal. Shaker Heights is a family-oriented bastion of progressivism and has an unusual history as the first planned suburb in the United States. It has very high taxes and very good schools. Cleveland Heights, my home ironically, is an aging hippy/hipster area and probably the most liberal area in Cleveland. The entire area has a certain level of progressive flair: I often see outlandish bumperstickers, the mascot of one of the elementary schools is "The Peacemakers", and we even have a Communist book store.

Downtown and the near west side have a lot of gentrified areas full of young professionals, the stereotypical latte-liberals, but the west side also has some degree of ancestral conservative Democrat pedigree to it. For example, you'll see a lot of religious symbols and there are conservatives who live there too even if the majority is pretty liberal. Ohio City is a bit less gentrified and there are many poor people living not too far from the luxurious new condominiums and developments. Tremont is really artistic. Edgewater and Lakewood are home to a large gay population and a lot of hipsters, but you will also find a substancial conservative minority in those areas and a lot of blue-collar Irish Catholics.

I can make a map, but Husted wasn't even the Republican who won by the largest statewide margin, State Treasurer Josh Mandel was. My guess as to why Attorney General and Governor were closer than the other two races is that they included higher profile and more controversial candidates. Strickland was a so-so governor and Kasich wasn't terribly popular even at the time of the election. DeWine was a former US Senator who made a few enemies, or at least more than Mandel or Husted, who both won comfortably as "Generic Republican" in a wave year.

Can you do the same for the 2008 Presidential election?

The reason I ask, is that the redistricting contest used the Governor, Secretary of State, and Auditor race from the 2010 election and the 2008 presidential race as their measure of partisanship.  Because 3 races were included were from 2010, about 2/3 of the vote was from 2010 (turnout in 2010 was about 2/3 of 2008).  This in itself introduces a bias.

Since the purpose was to produce "competitive" congressional races, the underlying model (at least in part) was assuming voters in a congressional district would be as likely to vote for generic Republican as they did Husted.   The only thing that I found was that Husted's opponent had been portrayed as a gun grabber because when she was a member of the Columbus city council she had favored letting Columbus have its own regulations on carrying guns, such that someone who was legal in their part of the state would commit a felony by driving into Columbus.

So I was wondering if there was some sort of rural bias towards Husted, particularly in the fairly balanced areas in SE Ohio.

I could compare Husted, but the 2010 Governor's race will definitely not give you a fair view of SE Ohio's voting patterns since Strickland is from SE Ohio and pretty popular there. Maybe I should try the State Treasurer Race since that should also be fairly non-descript.

Here's the 2008 Presidential Election broken down by county and CD:

Edit: I seemed to have made Preble County the wrong color in the original version for some reason.

()

Husted in is teal in both maps.

Here'’s Husted-Kasich. Husted did better in every single county. The county where Kasich was the closest to Husted was Cuyahoga. You can see the margins are especially large in the northwest and in Strickland’s old district, in particular the parts he represented before the 2001 redistricting. Kasich holds up much better in urban areas than rural ones.
()

And here’s Husted-Mandel. Husted does better in the northwest and Mandel does better in the northeast.  Mandel is from Cuyahoga County, so it makes sense that he would do better there. Husted was born in Detroit and leaves near Dayton, so I’m not sure why he does better in the northwest, but he does. He even manages 73% in Williams County, about 20% better than McCain.
()

I suppose there is a possibility of RTKABA vote giving Husted a few extra percentage.  Or it is possible that voters in rural areas are more discerning of individual candidates since they don't have to sort through dozens of congressional and legislative and local races.  If you don't have the money to spend on TV advertising like gubernatorial candidates, you may be willing to make yourself available to every newspaper and radio station in the state.

And even the urban areas are problematic for the contest rules.  If you use the SOS race rather than the AG race, you've added in a percentage or two boost to the 3-race average for 2010 even in the urban areas, which are particularly the areas that you don't want to overestimate the support of "generic Republican" when the contest rules are trying to make the race competitive.

My guess is that it has more to do with news because it seems to follow the lines between media markets more than a simple urban/rural divide. I'll try to see if I can find some statistics on gun ownership and concealed carry permits to really check this well. It seems like there is too much variation between Huron and Erie Counties and Ottawa, Sandusky, and Seneca Counties for at least a substantial piece of it to be media-driven. Holmes, Wayne, and Ashland also seem strange for an urban/rural split. Husted did much, much better in the Toledo area than you would otherwise expect if the split was just rural/urban. The northeast also doesn't surprise me very much because the Cleveland Plain Dealer gave Kasich a ringing endorsement and I don't think they endorsed Husted (or at least I don't remember them doing it). The southeast is skewed by Strickland, so I wouldn't consider that part of the state a useful comparison.

I think the rural/urban split may have more to do with Kasich than Husted. Kasich is unusually fiscally conservative for a state-wide Ohio politician and I think the type of voters who tend to be "moderate" in urban areas are generally more inclined to vote for fiscal conservatives whereas rural voters tend to me more socially conservative.

One other note: I only considered the two-party vote when I calculated the margins.

That matches the contest's methodology.   They went even further and simply totaled up votes.  If there was a significant 3rd party or independent vote in a particular race, it would at minimum reduce the weight of that race - but worse would likely skew the results of a race, since the other candidate would likely take more votes from one candidate than the other.

Here's a map I did not draw (or have the means of drawing) but found prepared by the City of Cleveland. It shows the condition of different blocks throughout the city taking into account the conditions of the properties in the area (ie. blighted, abandoned, foreclosed, etc).

Notice how much worse the east side is compared to the west side within the city! I'd also like to add that their idea of "regional choice" neighborhoods means "middle class" to most people.

In case anyone is wondering the black lines labeled "SPA" are the city's official neighborhood bounderies.

()

Population per Concealed Carry Permit Issued in 2011:
()

There is some pattern here between rural and suburban vs. urban and regional differences, such as much of the agricultural northwest/west-central area having few CCWs. However, this stat does not seem to match up well along ideological lines. Cuyahoga County having the fewest CCWs per capita is somewhat unsurprising but second is Coshocton (?). The third fewest is Mahoning, which isn't hugely surprising given it is pretty urban, but then fourth is Putnam, another unexpected result.

Teen Pregnancies per 1000 Females Ages 10-19 (2010):

()

This map shows higher rates in counties dominated by urban cores, but does not seem to discriminate by the size of the urban core; smaller industrial towns seem to have similar rates as larger cities like Cleveland. The lowest rates are in Deleware County (12.1), Geauga (12.2) and Holmes (14.8) while Clark has the highest (38.7).

This wouldn't be a demographics map and I don't if you are taking requests, but I'd be very interested in seeing a precinct map of Cleveland showing the decrease in turnout in the 2010 gubernatorial election compared to the 2008 presidential election.

I can try, but it will be hard to get more exact than on the municipality level because they redrew the precinct lines in most of Cuyahoga County between 2008 and 2010.

Here's a Cuyahoga County map of 2010 turnout as a percentage of 2008 turnout by municipality, except for the City of Cleveland which is broken down into area that are some combination of 2008 wards, 2010 wards, and the neighborhood stistical areas. This is about the best I can do with changing precinct and ward boundaries in most of the municipalities in the county:

()

Clearly there is a huge difference between urban and suburban areas. The area with the most consistent turnout is tiny Cuyahoga Heights which had 94% of its '08 turnout in '10 (and Kasich managed to increase upon McCain's victory margin from 2 votes to 3). For the most part the wealthier areas tended to retain voters better. The notable exception here being the high turnout in Parma, a poorer Polish and Ukrainian ethnic western suburb.

Within Cleveland itself, the poorer areas generally voted less, as well as the racially diverse neighborhoods. The diverse (45% white, 35% Hispanic, 20% black) Clark-Fulton neighborhood saw the largest fall-off in turnout. There is no obvious correllation between race and relative turnout. In fact the neighborhood with the largest relative turnout within the city itself was the far southeastern (97% black) Lee-Miles neighborhood and the second largest was the far west side's (83% white) West Park/Kamm's Corners neighborhood.

Kasich also managed to win a precinct (19-Q) in the West Park neighborhood, something John McCain failed to do.

Could some of the falloff in turnout in Cleveland be due to an ongoing drain of population, or increase in the Hispanic percentage?  Nice map by the way.

The Hispanic growth most likely played a role in the Clark-Fulton neighborhood since it (and perhaps some of the immediately surrounding neighborhoods) are the only parts of Cuyahoga County with a non-negligible Hispanic population. I would guess that is probably the reason Clark-Fulton was in the high 30s instead of mid 40s like similar neighborhoods in terms of economic class.

There is obviously some component of this from people moving, but it alone doesn't completely explain the turnout dropoff in Cleveland compared to the suburbs. Here (http://urban.csuohio.edu/nodis/maps/2010maps/chpop_cuy2000to2010_tracts.pdf) is a map of population change from 2000-2010 as a reference. I would like to point out that despite being one of fastest growing parts of the county, Downtown also experienced a large decline in turnout. Part of that is probably because Downtown's growth is largely young urban professionals who don't bother voting in non-presidential elections. Also it is widely believed that most of Cleveland's population loss occured in the first half of the last decade and has slowed since the economy crashed. Still many of the east side neighborhoods with huge turnout changes are in a state of total freefall no matter what the city's total population does, so it's probably an important contributing factor.

Thanks, awesome map!

TJ, could you do more gay marriage maps, say by precinct or state house or senate district?

While I have the data and map at hand, here's the 2010 gubernatorial race 2-party vote broken down by the same areas as before:

()

And here's the margin in each scaled by dot size:

()

Kasich did fairly well for a Republican in Cuyahoga County and probably even won the white vote. Most of Strickland's victory margin came from the heavily black neighborhoods on the east side. The other notable result is that Strickland racked up much larger margins the safer black neighborhoods than in crime ridden ones. I intentionally grouped Cleveland's two worse neighborhoods together (Kinsman and Central) and it contrasts nicely with Lee-Miles, one of the nicer neighborhoods in the city and also 97% black. Strickland carried both with similar percentages but had a 7150 vote margin in Lee-Miles and only 2674 in Central and Kinsman despite both groups having roughly the same population.

I have Lucas County:
()

To compare with race:
()

The Toledo metro area was fairly boring with the gay marriage ban passing pretty easily. The only pockets against it were around Downtown, the artsy gentrified section of the Old West End (the two darker red precincts), the University of Toledo, and the wealthy inner suburb of Ottawa Hills.

I also have Erie County done but that's even more boring. Is there anywhere in particular you're interested in? (This is easier to do precinct maps for because there weren't many changes between '04 and '08).

Here's the swing in presidential vote between 2008 and 2012.

()

Obama improved in central, south-central Ohio (the latter he did so horrendously in the first time that it would be difficult not to improve in), and the urban cores of Cleveland and the Mahoning Valley, while Romney outperformed McCain most notably in the West-Central part of the state and the eastern part of the Ohio Valley counties.

Unlike the central and western parts of the Ohio River Valley in Ohio, Democratic support for Obama hadn't collapsed to quite the same extent in 2008 in the eastern section. Romney's most improved county was Monroe, a tiny river county than Obama won by a decent margin in 2008 but lacks any defining features as to why it would remain Democratic (ie. not near Athens) while the surrounding counties become Republican.

Romney also improved upon McCain's numbers in the west-central German Catholic section of the state. Like in 2008, Mercer County was Obama's worst and this time he only managed 22% of the vote, getting fewer votes while the total cast increased.

Obama's most improved county by percentage was Licking, despite him actually losing it by a larger vote margin in 2012 than in 2008. The total number of votes increased significantly due to more suburban growth from the Columbus area.

Obama's best county was again Cuyahoga, where he outpeformed his 2008 percentage slightly but lost some of the vote margin as the total cast was about 32,000 fewer.

Epic thread. Thanks!

Again, any particular explanations for the pro-Obama swing around Columbus and directly south of that, extending into Pike etc counties? Columbus metro growth might be one, but does it explain stuff like Pike or Scioto Counties?

I'm not sure why Ross, Scioto, and Pike Counties swung toward Obama when the rest of the state did not. My guess is those counties having already swung heavily toward the GOP in 2008 from the previously normal levels and not having swung much toward the GOP otherwise (unlike the counties to the east of them that have been more Republican on other races as well). All three of them are among the poorer counties in the state and Pike has the second highest unemployment rate as of about a year ago. Pike and Scioto had fewer total votes cast for both candidates than in 2008, so there may have been a lack of Republican energy for Romney. Ross had about 200 more Obama votes than in 2008 but 1500 fewer Republican votes. Ted Strickland is from Scioto County so perhaps he's had some effect.

Perhaps somewhere in that jumble of reasons lies the answer.

Here's the 2012 map for US House by county:
()

In some places there was a pretty big disparity from the presidential numbers, for instance Wood, Hamilton, Montgomery, and Sandusky Counties were carried by the Republican candidates. Montgomery isn't all that surprising considering the popularity of Mike Turner (he was very close to hitting 60% even). The candidate effect is quite dramatic in places, like Tuscarawas County where half of it is in the contested OH-6 rematch between Charlie Wilson and Bill Johnson and the other half is in Bob Gibb's seat with a terrible Democratic opponent.

The Northeast corner of the state is much less solid for Joyce than it was for LaTourette. Joyce still won heavily Republican Geauga County by a large margin but the rest of the district was close whereas LaTourette often did about as well in Lake County as Geauga. OH-14 does have a more Republican-friendly slice of Cuyahoga County than it had before redistricting, so Joyce actually had a better total there than LaTourette normally had. However, this race had a poor Democratic candidate and far more third party and write-in votes than any other district.

Renacci had the closest race of any of the incumbents to be re-elected and it was actually the portion of Stark County that saved him as his district was given the most Republican parts of the county in the northern suburbs of Canton while Bob Gibbs had the more Democratic portion of the county.

Nice map! 

Actually, Renacci did not need Stark to pull it out. Wayne was enough to do that.  Renacci also ran pretty well in Sutton's home turf of Summit.

()

2008 Presidential Vote in Cuyahoga County broken down by the same districts as the 2010 Gubernatorial map:
()

The Democrats strength in Cuyahoga County is basically three types of areas: blacks (see the pacman shaped area of 95% precincts), blue-collar whites (the red west of black areas), and white liberals (the red east of the black areas).

Now, the 2012 Map:
()

And the swing map:
()

Romney vastly outperformed McCain in the wealthy suburban areas but underperformed McCain in the blue collar white areas in Cuyahoga County. Notice the dark blue areas on the far east side and clump of red on the near west side. In addition, a large part of this is a map of white flight, in particular the darker pink ring around the black areas. Some of the decline in Republican support on the near west side is also white flight as blacks and Hispanics are beginning to move into traditionally white west side neighborhoods. The two 1-2% R shaded neighborhoods in the center are Downtown and Ohio City/Tremont, all of which are gentrifying areas with an upperclass core that votes less lopsidedly with poor areas (including housing projects) on the periphery that vote in the usual close to unanimously Democratic fashion.

Can you make a map that has the counties which have an abortion provider? statistically, it's 13% of counties nationwide, but I wonder how that looked for the buckeye state (OH).

Thanks (FIRST POST!!!)