What makes a Community of Interest?

[Born to Slay. Forced to Work.]

This may sound kind of obvious, but I have seen a lot of discussion of what makes a Community of Interest. People have pointed to media markets, commuter patterns, racial demographics, and other markers as to define communities, but people almost always draw a different conclusion. Its also kind of awkward if you're not from a state trying to not divide communities and often that happens unintentionally. I thought it would be helpful to make a thread where people could discuss their COI in their states.

I'm working on my idea of South Carolina, so I'll post that as soon as I can.

[Born to Slay. Forced to Work.]

So for South Carolina theres a bunch of facets at play, I'm mostly going to focus on counties, but that leaves a lot on the cutting room floor. I'm from the northeastern part of the state and live in the Midlands now so most of my view is coming from that. I have very little experience with the Savanah river valley counties, so if you are someone who does know more about that part of the state please correct me.



First of for my state there is a main Upstate, Midlands, and Lowcountry divide. Shown here in with "true" upstate in dark red. Midlands are in blue, and the low country in green. However there's a lot of counties that sorta can feel like both depending on where you are and how you define the midlands and low country border.

The biggest of these is the PeeDee, which is constantly the "where do I put this" when making maps. Historically is been paired with the Grand Strand. This brings me to my second map


Where I try to lay out regions. Colors are functioning the same, more red/blue/green the more it fits into those regions. Many of these don't have official names (there's some that the tourist board assigns that would fit, but I've never seen them used outside of that). This one is the one I'm most unsure about, as there is a lot of salience between regions and not clear cut borders. For instance, Walterboro, St. George, and the Lake Moultrie region would be better aligned with Orangeburg then with Summerville. But Daniel Island and Summerville and 100% connected to Charleston. The only thing these two counties have in common is their suburbanization of the city, which brings me to my last map.



one of county pairings. There are several county pairs in the state that really in any fair map should go together if possible. Some of these (Spartanburg and Greeneville) are possibly less related, and too populous to have to be drawn together, and thus it's fair to break them up. But most of these are low population and can easily be stitched together in a fair map

[muon2]

Most of the time CoI in redistricting refers to geographic units smaller than a county.  One problem with entirely tiling a state into multi-county CoIs is that population equality is going to force chops in some number of them. I did a regression analysis that shows the likely inequality based on the mean number of geographic units (like counties) per region (see item 6 in the muon rules thread.)

For regional CoI I recommend the work jimrtex did in putting together the Urban County Clusters (UCCs) based on collaborative threads here years ago, with the final version stickied on the board. The UCC avoids the population problems in mentioned above by recognizing that the CoIs tend to be stronger in metro areas, so it leaves the rural areas to float between districts to aid population equalization. The muon rules provide a means to reward plans that keep those UCCs as whole as possible without using them to dilute rural areas unnecessarily.

[Sol]

If it's alright I might chime in with some North Carolina CoIs!

The way I try to think about CoIs is a bit like isoglosses in linguistics--every state has multiple Communities of Interest, often overlapping and cross-cutting.

Some major CoI divides in North Carolina:

The three way topographical divide between the Coastal Plain, Piedmont, and the Mountains:
 
(image from NC Sec. of State Website)

https://www.ncregions.org/wp-content/uploads/2016/01/NC-map.jpg

NC Councils of Government (image from Council of Governments website)

And of course metro areas.

[jimrtex]

Most of the time CoI in redistricting refers to geographic units smaller than a county.  One problem with entirely tiling a state into multi-county CoIs is that population equality is going to force chops in some number of them. I did a regression analysis that shows the likely inequality based on the mean number of geographic units (like counties) per region (see item 6 in the muon rules thread.)

For regional CoI I recommend the work jimrtex did in putting together the Urban County Clusters (UCCs) based on collaborative threads here years ago, with the final version stickied on the board. The UCC avoids the population problems in mentioned above by recognizing that the CoIs tend to be stronger in metro areas, so it leaves the rural areas to float between districts to aid population equalization. The muon rules provide a means to reward plans that keep those UCCs as whole as possible without using them to dilute rural areas unnecessarily.

In thinking about Minnesota, I was wondering whether municipalities should have a special place in the statute but rather could be recognized as communities of interest by a simple vote of the city council. This could also apply to groups of towns/cities. Minnesota apparently makes any city independent of its original township, at least as far as defining MCD. But the township and city by joint resolution could be recognized as a COI. There could be some sort of limit to the population of these aggregate COI, since the intent would be that they be contained in a single legislative district (e.g. maximum population of 10,000 unless contained within a single municipality).

This would permit municipalities that cross county boundaries to be treated as a COI. This could also apply to Indian Reservations.

Counties for redistricting purposes could be aggregation of municipal COI. This might permit the eastern portions of St.Cloud as well as Sauk Rapids to be included with Stearns. The same might be done for North Mankato (both Manakato and North Mankato are in both Blue Earth and Nicollet counties).

Minnesota has Regional Development Councils (since 1969) in most of the state. They appear to be more formal than the RCOGs in states like Illinois, but less formal than in Michigan or Texas. There are also Regional Transportation Coordinating Councils that cover the state except for the metro area. These are mandated by state statute, but are quite new. The 7-county Metro Area has a formal government.

*A curiosity is that Hennepin County no longer has MCD recognized by the Census Bureau estimates program since all of the county other than the non-populated Fort Snelling UT are incorporated cities.

In other counties, cities are treated as MCD. The Census Bureau does estimate populations for incorporated cities. The cities in Hennepin are treated as county subdivisions, but these are only considered to be statistical entities, and no estimates are produced.

[jimrtex]

If it's alright I might chime in with some North Carolina CoIs!

The way I try to think about CoIs is a bit like isoglosses in linguistics--every state has multiple Communities of Interest, often overlapping and cross-cutting.

Some major CoI divides in North Carolina:

The three way topographical divide between the Coastal Plain, Piedmont, and the Mountains:
 
(image from NC Sec. of State Website)

https://www.ncregions.org/wp-content/uploads/2016/01/NC-map.jpg

NC Councils of Government (image from Council of Governments website)

And of course metro areas.

These have been used in development of congressional maps.

[beesley]

-skip-

Surprised that Brunswick and New Hanover counties wouldn't be grouped together.

[jimrtex]

-skip-

Surprised that Brunswick and New Hanover counties wouldn't be grouped together.

This was my reasoning at the time.

https://talkelections.org/FORUM/index.php?topic=285238.msg6077716#msg6077716

My initial analysis found that the Lumber River and Cape Fear RCOG had a population equivalent to 0.981 districts. If I hadn't had problems elsewhere I would have considered that an ideal district.

I appear to have concentrated on keeping as many districts outside the Charlotte-Greensboro-Raleigh triangle as possible, initially aiming for a 2:7:4 division. The two western districts closely follow the RCOG's, and the northeastern district follows the RCOG's along Albermarle Sound and points west. I was likely having DRA concerns there.

I also was aiming for whole county districts. My second to last map only split Wake, Mecklenburg, and Guilford.

Technically, Brunswick is part of the Myrtle Beach (SC) UCC, because slightly more of the county is in the Myrtle Beach UA vs. the New Hanover UA (a very tiny difference). If you were delineating COI from a North Carolina perspective, you would have placed Brunswick and New Hanover together.

You might be able to rotate the eastern districts slightly clockwise.

I may attempt a new effort at a 14-district NC and current population. Though the new district will quite likely be in the central part of the state (splitting my I-95 district?), it might make the creation of 4 districts in the east easier.

[Nyvin]

Most of the time CoI in redistricting refers to geographic units smaller than a county.  One problem with entirely tiling a state into multi-county CoIs is that population equality is going to force chops in some number of them. I did a regression analysis that shows the likely inequality based on the mean number of geographic units (like counties) per region (see item 6 in the muon rules thread.)

For regional CoI I recommend the work jimrtex did in putting together the Urban County Clusters (UCCs) based on collaborative threads here years ago, with the final version stickied on the board. The UCC avoids the population problems in mentioned above by recognizing that the CoIs tend to be stronger in metro areas, so it leaves the rural areas to float between districts to aid population equalization. The muon rules provide a means to reward plans that keep those UCCs as whole as possible without using them to dilute rural areas unnecessarily.

This kinda sounds like it'd work against Democrats more often than not though.   Urban areas would be bound together in a small number of districts and tightly pack Democratic voters, while rural areas could be slide in and out of districts to the benefit of Republicans.   

There should be some mechanism to keep the most sparsely populated areas of a state together in the same districts as there is to be keep densely populated areas of the state together, otherwise with the current political geography of the country the metrics would have a Republican favoring bias.

[muon2]

Most of the time CoI in redistricting refers to geographic units smaller than a county.  One problem with entirely tiling a state into multi-county CoIs is that population equality is going to force chops in some number of them. I did a regression analysis that shows the likely inequality based on the mean number of geographic units (like counties) per region (see item 6 in the muon rules thread.)

For regional CoI I recommend the work jimrtex did in putting together the Urban County Clusters (UCCs) based on collaborative threads here years ago, with the final version stickied on the board. The UCC avoids the population problems in mentioned above by recognizing that the CoIs tend to be stronger in metro areas, so it leaves the rural areas to float between districts to aid population equalization. The muon rules provide a means to reward plans that keep those UCCs as whole as possible without using them to dilute rural areas unnecessarily.

This kinda sounds like it'd work against Democrats more often than not though.   Urban areas would be bound together in a small number of districts and tightly pack Democratic voters, while rural areas could be slide in and out of districts to the benefit of Republicans.   

There should be some mechanism to keep the most sparsely populated areas of a state together in the same districts as there is to be keep densely populated areas of the state together, otherwise with the current political geography of the country the metrics would have a Republican favoring bias.

Our model state when we started was MI. The problem was that it was too easy to split the Dems in the 3-county Lansing metro under the guise of whole counties. We found that metro areas as defined by the Census were too big and grabbed a lot of unnecessary rural areas. That led to the specific metrics for selecting certain counties in a metro area for inclusion in the UCC. We tested the model with a variety of plans and found it kept the political balance reasonably well. We followed with plans for many states, and never found one where it did worse politically than metrics that ignored the UCCs.

Ideally metrics should not be shifted just because there is political realignment. That said, it's worth testing the metrics with new election data to see if there's an unanticipated bias. That requires plan makers to follow a set of metrics, an idea that seems to have fallen a bit out of favor compared to 10 years ago.

[Badger]

I can tell you that a community of interest is frequently not easily parceled by municipal or even County boundary. There are many Central Ohio suburbs of Columbus like West Jefferson, Pickerington, Dublin, Westerville, which are largely or entirely outside Franklin County, but whose identity is far more with Columbus then their own county seats. In Pickerington for example, I know that people there get the Columbus Dispatch not the Lancaster Eagle Gazette. If they travel outside of Pickerington or Canal Winchester area for shopping, entertainment,  employment, Etc, it is almost always towards Columbus, except if one has to go deal with County government agency in Lancaster.

[Sol]

Here's a relevant topic I started a little bit ago, on sub-county communities of interest.

[Secretary of State Liberal Hack]

An expo-facto argument in favour of boundaries that favour my party.

[muon2]

An expo-facto argument in favour of boundaries that favour my party.

I've seen that, for sure. In 2011 the IL Dems wrote a extensive resolution to accompany their new legislative plan. The communities of interest were written in detail and had little to do with the ones presented in 19 public hearings. However, their described CoIs exactly matched their plan and they admitted in floor debate over the map that it was drawn explicitly to favor their party.

[Nyvin]

'The primary factor should be if the voters of the district can easily identify what district they're in. 

[jimrtex]

An expo-facto argument in favour of boundaries that favour my party.

I've seen that, for sure. In 2011 the IL Dems wrote a extensive resolution to accompany their new legislative plan. The communities of interest were written in detail and had little to do with the ones presented in 19 public hearings. However, their described CoIs exactly matched their plan and they admitted in floor debate over the map that it was drawn explicitly to favor their party.

The draft Minnesota law in the complaint says that the sponsor of a plan must only provide definitions of communities of interest if he is asserting that his plan is preserving them.

"If the sponsor of a plan asserts that it preserves a community of interest, maps of the plan must include a layer identifying the census blocks within the community of interest. The plan must be accompanied by a description of the research process used to identify the community of interest and a communities of interest report listing any district or districts to which the community of interest has been assigned. The report must also show the number of communities of interest that are split and the number of times a community of interest is split"

It is not clear what a sponsor is. It's only used in this context.

If a community will exist in 2021 and persist throughout the decade, then surely it existed in 2019.

[muon2]

'The primary factor should be if the voters of the district can easily identify what district they're in.  

I agree. That's the advantage of dividing the state into identifiable geographical subunits and then putting the subunits together to make districts. Larger districts like congressional work better if one starts with larger subunits. Smaller districts need smaller subunits to start. I've found that subunits that are no larger than 10% of the ideal district size work best.

This is from the analysis I did on whole county (and whole New England town) maps that were drawn for congressional districts on threads here last cycle. A range of 3.5 (average of 10 geographic units per district) is just under 0.5% of a CD. A range of 2.5 (average of 20 geographic units per district) is just under 0.05% of a CD.



The analysis would generally apply to any collection of subdivisions, not just counties. For example one could use school districts in populous suburban areas as better indicators of communities of interest that are easily identified by voters.

[Born to Slay. Forced to Work.]

'The primary factor should be if the voters of the district can easily identify what district they're in.  

I agree. That's the advantage of dividing the state into identifiable geographical subunits and then putting the subunits together to make districts. Larger districts like congressional work better if one starts with larger subunits. Smaller districts need smaller subunits to start. I've found that subunits that are no larger than 10% of the ideal district size work best.

This is from the analysis I did on whole county (and whole New England town) maps that were drawn for congressional districts on threads here last cycle. A range of 3.5 (average of 10 geographic units per district) is just under 0.5% of a CD. A range of 2.5 (average of 20 geographic units per district) is just under 0.05% of a CD.



The analysis would generally apply to any collection of subdivisions, not just counties. For example one could use school districts in populous suburban areas as better indicators of communities of interest that are easily identified by voters.

I’m curious how counties play a role in states like Georgia, North Carolina, Texas, New York, and California when trying to run broad analysis like this. These states run the gambit of county size with some having counties that could have a dozen districts, or districts that could have a dozen counties.

I think a lot of this has to do the fact I come from a fuzzy area between major COIs so I’m especially interested as what to do with the buffer regions. I’d be interested to see a map of how you’d draw subsets in a given state so I could get a better idea of the concept if you happen to have a map handy

[muon2]

'The primary factor should be if the voters of the district can easily identify what district they're in.  

I agree. That's the advantage of dividing the state into identifiable geographical subunits and then putting the subunits together to make districts. Larger districts like congressional work better if one starts with larger subunits. Smaller districts need smaller subunits to start. I've found that subunits that are no larger than 10% of the ideal district size work best.

This is from the analysis I did on whole county (and whole New England town) maps that were drawn for congressional districts on threads here last cycle. A range of 3.5 (average of 10 geographic units per district) is just under 0.5% of a CD. A range of 2.5 (average of 20 geographic units per district) is just under 0.05% of a CD.



The analysis would generally apply to any collection of subdivisions, not just counties. For example one could use school districts in populous suburban areas as better indicators of communities of interest that are easily identified by voters.

I’m curious how counties play a role in states like Georgia, North Carolina, Texas, New York, and California when trying to run broad analysis like this. These states run the gambit of county size with some having counties that could have a dozen districts, or districts that could have a dozen counties.

I think a lot of this has to do the fact I come from a fuzzy area between major COIs so I’m especially interested as what to do with the buffer regions. I’d be interested to see a map of how you’d draw subsets in a given state so I could get a better idea of the concept if you happen to have a map handy

For the study used to make the graph I used plans that nested as many whole districts in a county as possible and attached the remainder to other counties. That's why the graph refers to regions, since some reflect multiple districts due to large counties. It doesn't imply that nesting districts is best, but nesting them made a uniform approach to handling large counties across the states.

There are lots of threads from 2012-2015 that looked at rational ways to subdivide large counties in different states. Most are available through the old 2010 redistricting directory thread. IIRC MI, WA, and VA got a lot of attention on big county subdivisions.

One important consideration we found based on the plans drawn and the observations of real legislative gerrymanders is that any geographic subdivisions have to be specified in advance and agreed to by all sides. It was too easy to construct gerrymanders (sometimes subtle) and then define the CoIs after the fact. Agreeing on subdivisions in advance block that sort of gerrymander.

[Skill and Chance]

Whatever a local politician wants it to be, which is why this criterion should not be used!

[jimrtex]

I’m curious how counties play a role in states like Georgia, North Carolina, Texas, New York, and California when trying to run broad analysis like this. These states run the gambit of county size with some having counties that could have a dozen districts, or districts that could have a dozen counties.

Under the Texas Constitution, House districts are required to be made up of whole counties.

(1) Single counties electing multiple representatives.
(2) Multiple counties electing a single representative.
(3) Multiple whole counties or surpluses of whole counties electing a single representative.

The 3rd type are called floterial districts. Imagine you had a county entitled to 1.2 representatives. It is entitled to one district that would be the whole county. It also has a surplus of 0.2 districts, and the whole county would be treated as a county with 0.2 districts in forming a single-member multi-county district.

Historically, Texas had lots of counties entitled to 1 or 2 districts. Very few counties were entitled to more. Counties on the frontier might be entitled to less than a whole district.

So a county entitled to 1.2 districts might be given 1 district, and one entitled to 1.8 might be give two districts. Two counties entitled to 1.5 districts each might be entitled to one district, and together form another district.

Or counties entitled to 1.6 districts and 1.4 districts might be paired in a similar fashion. Note that in the this case, the larger county would have 53% of the electorate.

But if you require greater equality, you might find a county entitled to 1.6 representatives combined with two smaller counties entitled to 0.2 representatives. The larger county would form 80% of the electorate in the floterial district.

For this reason, floterial elections usually violate equal protection and are unconstitutional.  They might be constitutional in certain forms. Imagine a legislature of 150 members. The state could be divided into 100 single-member districts, and then districts paired to elect another 50 members. No voters are injured, they participate in the election of two representatives, in both cases in full parity with other voters.

After Reynolds v Sims, Texas legislators (read Democrats) decided they could ignore the Texas Constitution and carve up districts without regard to counties.

But the Texas Supreme Court determined that the Texas Constitution could be harmonized with equal protection, by treating a sub-area of a county that contains a population equal to the surplus as a "whole" county in forming districts of type 3A.

So if there were two adjacent counties entitled to 1.6 and 1.4 districts, there would be one district in a portion of County A with a population equal to 1.0 districts; another in County B with a population equal to 1.0 districts, and a third district comprising the remainder of the two counties.

Since these types of quasi-floterial districts are technically in violation of the constitution, whole county apportionment is favored.

At-large elections are generally unconstitutional when minorities are involved, since the majority may elect all the members when there is racially polarized voting. When they applied this to at-large elections to the House in Texas, this was found to be true in every county but in Hidalgo. So it has become the practice in Texas to draw single member districts throughout the state. The Texas Supreme Court has ruled that this is a manner regulation. That is if a county comprises a 14-member district, the individual representatives may be elected by subdistrict.

Reapportionment in Texas is then a two-step process. (1) Find single counties entitled to a whole number of representatives (within 5% deviation); Find counties entitled to more than a district, and combine their surpluses with surpluses from other similar counties or smaller counties to form a single-member district. Draw multi-county single member districts.

(2) Divide larger counties into single-member districts.

This is where gerrymandering occurs. Cities in Texas are often not compact.

[Born to Slay. Forced to Work.]

I’m curious how counties play a role in states like Georgia, North Carolina, Texas, New York, and California when trying to run broad analysis like this. These states run the gambit of county size with some having counties that could have a dozen districts, or districts that could have a dozen counties.

snip

This is incredibly interesting and unique in how it’s done. I’m interested to see how it holds up in an increased polarized environment. Looking at the TX House of Representatives map the districts aren’t too appalling gerrymandered outside of the major cities, but even then I’ve seen worse. Harris County is probably the worst of all of them. I really like this method. I imagine it makes election running and elections coverage much easier as you don’t have to deal with much cross county issues like you have in South Carolina where we have districts that take parts of 4 counties while having none of the whole so you get small counties having to create dozens of ballots. Do they apply this to congressional seats?

[President Punxsutawney Phil]

I’m curious how counties play a role in states like Georgia, North Carolina, Texas, New York, and California when trying to run broad analysis like this. These states run the gambit of county size with some having counties that could have a dozen districts, or districts that could have a dozen counties.

snip

This is incredibly interesting and unique in how it’s done. I’m interested to see how it holds up in an increased polarized environment. Looking at the TX House of Representatives map the districts aren’t too appalling gerrymandered outside of the major cities, but even then I’ve seen worse. Harris County is probably the worst of all of them. I really like this method. I imagine it makes election running and elections coverage much easier as you don’t have to deal with much cross county issues like you have in South Carolina where we have districts that take parts of 4 counties while having none of the whole so you get small counties having to create dozens of ballots. Do they apply this to congressional seats?

Texas does not do this for congressional districts, but in a world where we had more deviation allowed and more incentive to follow adherence to county lines, I'd want it to.

[jimrtex]

I’m curious how counties play a role in states like Georgia, North Carolina, Texas, New York, and California when trying to run broad analysis like this. These states run the gambit of county size with some having counties that could have a dozen districts, or districts that could have a dozen counties.

I think a lot of this has to do the fact I come from a fuzzy area between major COIs so I’m especially interested as what to do with the buffer regions. I’d be interested to see a map of how you’d draw subsets in a given state so I could get a better idea of the concept if you happen to have a map handy

North Carolina requires House districts made up of one or more whole counties. You might not be able to place some counties together and have the population equivalent to one district so you end of with larger groupings electing 2 or 3 representatives at large.

In addition, North Carolina had VRA districts, where they would piece together contorted geographies to get enough population that was 50% black. These sometimes included point-connected districts that would cross over. These districts ignored county lines. The legislature would then group the remnants together in whole-county districts.

These districts were found to violate the VRA and not satisfy the Gingles test. There was not a black CVAP majority in a compact area. Just because the boundary around a district is continuous it does not make it compact. In addition, voting was not so polarized as to require creation of black majority CVAP districts.

The NC Supreme Court also found multi-member districts to be unconstitutional. So what happens now is that groups of counties that are equivalent to a whole number of representatives are gathered together then divided. Ideally there would be no divisions except wholly within a county. Larger counties such as Mecklenburg, Wake, Guilford likely can have a whole number of districts. But in other areas division of small counties is unavoidable. The goal is to minimize these divisions that violate the state constitution.


New York has some sort of requirement to respect county lines for Assembly districts. In some cases it works. There are a whole number of districts in Nassau and Suffolk counties, and exactly one district crosses the county line.

In additional all NYC assembly districts are confined to a single borough. This is not something special for NYC, but the fact that larger counties are more likely to be able have a whole number of districts with a reasonable deviation, and Richmond happens to match up.

Upstate it is hard to see how the rules are applied. They do on occasion match county lines. But Assembly District 101 is actually an application of the rules - I think it is leftovers from other districts.