Ohio redistricting thread

[Nyvin]

Hard to tell without county borders, but looks like Franklin County is split three times and Cincinnati is split.  Neither is allowed.

[lfromnj]

There's a very specific way you need to split Franklin.  Its not the worst of splits and moderately logical although the GOP gerrymander should absolutely not pair either of the splits with Delaware county for obvious reasons.

[Idaho Conservative]

Republicans should not crack Cinci IMO.  Just draw clean 12-3 map.

[dpmapper]

Republicans should not crack Cinci IMO.  Just draw clean 12-3 map.

Why not?  If it doesn't endanger the other Cincy-area district, they may as well.  You don't even have to trail it out to the east beyond Clermont County, as there are still plenty of very red areas of Hamilton.  Those plus Cincinnati city plus Clermont gives a lean R district, and then the rest of Hamilton can be added to Butler/Warren for a safe R. 

[Idaho Conservative]

Republicans should not crack Cinci IMO.  Just draw clean 12-3 map.

Why not?  If it doesn't endanger the other Cincy-area district, they may as well.  You don't even have to trail it out to the east beyond Clermont County, as there are still plenty of very red areas of Hamilton.  Those plus Cincinnati city plus Clermont gives a lean R district, and then the rest of Hamilton can be added to Butler/Warren for a safe R. 

would incentivize a ballot initiative

[Torie]

The DRA per Muon2 is using 2015 census estimate figures. The 2020 map was done manually and laboriously. It may have some small errors because I revised it a bit without going back to the drawing board, but the errors should not force a systemic map draw.

[President Punxsutawney Phil]

The DRA per Muon2 is using 2015 census estimate figures. The 2020 map was done manually and laboriously. It may have some small errors because I revised it a bit without going back to the drawing board, but the errors should not force a systemic map draw.

What is the partisanship of that 12th CD? How did it vote in 2016?

[Torie]

In 2016 in the 2020 census map the result in OH-12 for POTUS was:

Democratic   172,938   46.8%
Republican   193,790   52.4%

[Stuart98]

Tried doing an R gerrymander under the cube root rule with 24 districts (note: I apportioned with 693 total districts but included DC and PR as states; otherwise Ohio would have had 25 districts).



Conclusion: Potency of R gerrymandering would be significantly reduced w/ more districts. (And even this map required bending the rules a few times, eg splitting Hamilton and Summit thrice and Cuyahoga four times).

[Torie]

[Torie]

I might note that the 2010 Pub utterly hideous gerrymander of the Ryan seat (CD 13), by the end of the decade turned into a Dem gerrymander. Ryan is done in any reasonable map. Unless of course he would like to move to Akron and try his luck in the CD that he currently represents the entrails of.

In this chaotic era the risk of drawing dummymanders has dramatically increased. I wonder if the partisan hacks will get that memo. Color me rather pessimistic. The quantum of their obtuseness and narcissism tends to too high to facilitate their escape from purgatory, where they tend to objectify voters as their personal pawns to power.  

Just one old curmudgeon's opinion. Pay me no mind.

[palandio]

Is this a "good government" map? Because if it is, it shouldn't separate Youngstown and Warren.

Not separating Youngstown and Warren can make drawing the map a bit complicated. There seem to be several options:
a) Mahoning + Trumbull + Portage + Ashtabula + Columbiana
often combined with a Cleveland suburban/exurban district that goes from Lake to Medina and often includes some of Cleveland's D-trending eastern suburbs, forcing a triple-split of Cuyahoga and generally one of the more D-friendly designs
b) Mahoning + Trumbull + Ashtabula + Lake + parts of Geauga
goes nicely with a Summit + Portage + Geauga seat
c) Mahoning + Trumbull + Stark

All these districts have in common that they have a 2012/2016 PVI close to EVEN and that they were won by Trump in 2016 by 8-10 points and were comfortably won by Sherrod Brown in 2018. Ryan would probably have lost these districts in 2020 by about 5 points, but it would have been a fair fight.

There is of course another option:
d) Mahoning + Trumbull + Columbiana + Carroll + Jefferson + Harrison + Belmont + Guernsey + Noble + Monroe
This has a 2012/2016 PVI of about R+3 or R+4 and voted for Trump by ca. 19 points.

[Torie]

Perhaps you might like the map to the right better. It has the obvious advantage to which you referred of keeping the Youngstown metro area together. Disadvantages are that the map is more erose, makes CD-07 most probably less competitive, does not make CD-13 materially more competitive, and involves one extra mega county chop (of Summit), leaving CD-07 bereft of any whole county, although hewing to the whole county concept as a material plus factor is admittedly no big deal, particularly since Summit is a mixed bag as to its component parts, wedged in between the Cleveland and Canton metro areas.

The moral of the story is that there is no free lunch. One must choose, and in reality I suspect the map to the right in this part of the state is more likely to be adopted by the Pubs (assuming that they do not go on a gerrymandering jihad pushing the limits of the law just because they can), than the one that I prefer. So you may well get your way as it were (not that you find either map to be the one that in your mind is the closest to Aristotelian perfection, I understand ).

Both of course are worthy maps. JMO, and thank you for your comments.

[Sol]

FWIW you can do two Columbus districts fairly easy--nest one in Franklin County and the other in Delaware and Franklin, plus a cut from either Licking or Fairfield. It's the same number of county splits as on Torie's current map but the splits are more reflective of communities, since both Licking and Fairfield have Columbus suburbia but also fairly disconnected county seats.

[Torie]

FWIW you can do two Columbus districts fairly easy--nest one in Franklin County and the other in Delaware and Franklin, plus a cut from either Licking or Fairfield. It's the same number of county splits as on Torie's current map but the splits are more reflective of communities, since both Licking and Fairfield have Columbus suburbia but also fairly disconnected county seats.

The disadvantage of that is that it creates a much larger chop of the City of Columbus (and may dilute the black influence of the city CD, which I believe currently is actually represented by an AA). The CD to  the south of Columbus has no Columbus precincts within it - by design. Plus Licking and Delaware fit together well. It is true that Licking and the south Franklin CD's share of Franklin have about the same populations, so that is a plus, in that the size of the county chop would be greatly reduced. However, Licking's population is smaller by a bit, so the CD to the south would still need to cut into Franklin by about 5,000 people or so, or another CD chop into Delaware.

It would create another safe Dem seat, but the way I drew the seat, the CD is already lean Dem per the 2020 election, and is trending Dem at warp speed, so in the end it will not make much of a partisan difference. It also makes the map look more erose.


Addendum: Here is a map showing what I was saying. The BCVAP only decreases by 2.8%, so that is not really an issue. It is just the jut up from Fairfield to Licking that looks ugly. The Pubs largely control the process, so this just isn't happening - ever. What might be more likely perhaps as a not bad looking map at all really will be added in a bit. No doubt you can guess.




The perfection of Pub purloinment personified. Never put a Pub vote sink to waste, particularly if the map looks pretty. CD-12 is back in the GOP garden.

[Sol]

Joyce Beatty is Black, but she represents a seat which is majority-White and only 1/3rd Black--not exactly protected under the VRA.

Plus as you know adhering to the city of Columbus's boundaries is very difficult as they are noncontiguous and nonsensical.

Let me mock up a quick map of the Columbus area--I'll post once I finish.

[palandio]

A general question to Torie regarding his 2020 census projections:

What is the source for them? How do they diverge so drastically from DRA's 2018 extrapolations? I mean it's clear that from 2018 to 2020 demographic trends would continue for another two years, but to get from DRA's 2018 numbers to your 2020 numbers you would need an annual population decrease of 1-2% in Northeastern Ohio (or a 1-2% increase in the rest of the state). Or was there a systematic bias in the 2018 extrapolation?

[Sol]

Oh, are you using the 2020 projections? TBH I have not really been able to figure out how to use them.

[Sol]

Here's a rough pass at a Franklin-Delaware (+Licking) seat. I'm using 2018 numbers so the green district can probably shrink a little. I drew the lines in Franklin with an eye to maximize the Black %.

link

You could also go into both Licking and Fairfield, though it does require a split of a township on 2018 numbers. That said, it's probably better CoI wise. Golly, this exercise is really bringing out my inner Californian (redistricter).

[Torie]

Oh, are you using the 2020 projections? TBH I have not really been able to figure out how to use them.

I can send you my spreadsheet so that you can admire my genius if you like. PM me if so. You need a matrix chart to move county populations around, after projecting each county's population, and on the side for chops, split the population pie assuming the population change in the county is uniform. After some practice, you can move stuff around fairly quickly.

[Sol]

Oh, are you using the 2020 projections? TBH I have not really been able to figure out how to use them.

I can send you my spreadsheet so that you can admire my genius if you like. PM me if so. You need a matrix chart to move county populations around, after projecting each county's population, an on the side for chops split the pop pie assuming the population change in the county is uniform. After some practice, you can move stuff around fairly quickly.

Ah I see! I may PM you--though I may hold off since I'm a little wary of using spreadsheets given the issues with divergent growth within counties.

[Torie]

A general question to Torie regarding his 2020 census projections:

What is the source for them? How do they diverge so drastically from DRA's 2018 extrapolations? I mean it's clear that from 2018 to 2020 demographic trends would continue for another two years, but to get from DRA's 2018 numbers to your 2020 numbers you would need an annual population decrease of 1-2% in Northeastern Ohio (or a 1-2% increase in the rest of the state). Or was there a systematic bias in the 2018 extrapolation?

I used the 2010 populations from the DRA that match the census, in part because Muon2 tells me the alleged 2018 figures are from 2015. I then took the 2019 census estimates for counties, and extrapolated them forward for another nine months (to the 0.75 power). You then use the 2020 projected figures, and  for say a county chop, adjust the DRA figures according.

So, for example, if the percentage change of a county from 2010 was 10%, and you do a chop that shows 10,000 people, by locking the CD's except the one involved, and then pressing the county button with the setting to not assigned. You then on the side of your matrix chart, type in the 10,000, and it changes to 11,000 people, and then copy and paste or otherwise move the 11,000 into the column for that CD, and the row for that county. If the county is chopped once, you can in the cell in the same row for the other CD, insert an equals sign and then press the cell for the county's total projected population, a minus sign, and then the cell with the 11,000 to get the balance of the county's population to be assigned to the other CD.

[Torie]

Joyce Beatty is Black, but she represents a seat which is majority-White and only 1/3rd Black--not exactly protected under the VRA.

Plus as you know adhering to the city of Columbus's boundaries is very difficult as they are noncontiguous and nonsensical.

Let me mock up a quick map of the Columbus area--I'll post once I finish.

The black percentage issue is not a legal VRA issue here as you note, but a policy issue, that should be a factor, if not necessarily a dispositive factor,  as a legal requirement would be.

My Pub purloinment map has a black CVAP of 30% so that was not an issue with that map either. A map that would cause a black incumbent to lose, or be less likely to be replaced with another one due to bleaching in order to grab a CD for a party, that also fails to hew to the other metrics we follow, may be legal, but deserves more criticism than a map that avoids that. So I did the same thing you did and was cognizant of not only voting patterns and municipal lines, but also race.

[Torie]

Oh, are you using the 2020 projections? TBH I have not really been able to figure out how to use them.

I can send you my spreadsheet so that you can admire my genius if you like. PM me if so. You need a matrix chart to move county populations around, after projecting each county's population, an on the side for chops split the pop pie assuming the population change in the county is uniform. After some practice, you can move stuff around fairly quickly.

Ah I see! I may PM you--though I may hold off since I'm a little wary of using spreadsheets given the issues with divergent growth within counties.

In Franklin, the population change rate between Columbus and the balance of the county is very close. The difference between cities and suburbs in a country are in general far less divergent this time than they were in the run up to 2010. The cities came back.

[Sol]

Oh, are you using the 2020 projections? TBH I have not really been able to figure out how to use them.

I can send you my spreadsheet so that you can admire my genius if you like. PM me if so. You need a matrix chart to move county populations around, after projecting each county's population, an on the side for chops split the pop pie assuming the population change in the county is uniform. After some practice, you can move stuff around fairly quickly.

Ah I see! I may PM you--though I may hold off since I'm a little wary of using spreadsheets given the issues with divergent growth within counties.

In Franklin, the population change rate between Columbus and the balance of the county is very close. The difference between cities and suburbs in a country are in general far less divergent this time than they were in the run up to 2010. The cities came back.

Was thinking more wrt: Licking and Fairfield.