Welcome, Guest. Please login or register.
Did you miss your activation email?
April 17, 2014, 02:22:23 pm
HomePredMockPollEVCalcAFEWIKIHelpLogin Register
News: Please delete your old personal messages.

+  Atlas Forum
|-+  General Politics
| |-+  Political Geography & Demographics (Moderator: muon2)
| | |-+  Iowa-style Redistricting II: New England Towns
« previous next »
Pages: [1] Print
Author Topic: Iowa-style Redistricting II: New England Towns  (Read 1273 times)
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« on: April 07, 2012, 10:23:30 pm »

As many have noted, counties are not the critical unit in New England. In fact counties have been dissolved as governing bodies in CT, RI, and parts of MA. Towns (and cities when so named) are the fundamental unit. For these states keeping towns intact is far more important than keeping counties (or their historical boundaries) intact.

So, the challenge here is to divide the New England states into CDs while keeping towns intact and minimizing the average population deviation. Now that I have a statistical model from the whole-county states, I'll be interested in comparing to these states with far more jurisdictions to manipulate.

As before the rule requires no district to have more than a 0.5% deviation from the ideal, and point contiguity is not allowed. Bonus points for making districts that are connected by roads internally, but we won't let that stand in the way of a great plan, such as the near-perfect split of ID.

Also, some have suggested that NJ could be on this list as well, so I'll include it (boroughs, cities townships, etc all count as towns). However, I won't extend to NY and PA since both have cities that exceed the population of one CD, and can't really fit this rule.

CT (5 CDs, 169 towns)
ME (2 CDs, 433 towns)
MA (9 CDs, 351 towns)
NH (2 CDs, 234 towns)
NJ (12 CDs, 566 towns)
RI (2 CDs, 39 towns)

Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
traininthedistance
YaBB God
*****
Posts: 2835


View Profile
« Reply #1 on: April 08, 2012, 02:56:46 pm »
Ignore

Here's a first crack at NJ, closely based on some other maps I've made:



1: 528
2: 748
3: -905
4: 372
5: 123
6: 436
7: -90, 49.3% white VAP
8: -850
9: -79
10: -475, 45.5% black VAP
11: -264
12: 454, 37.6% hispanic VAP

It is not possible to satisfy the VRA without splitting Newark and Jersey City at the very least, and my previous attempts at clean maps have generally split Elizabeth and a couple Essex County towns as well, they may be necessary too.  Since we can't do that, I elected to go for compactness rather than maximizing minority percentages.  I also tried to give lip service to respecting county boundaries as well.

This map can almost certainly be improved, I'll give it a shot sometime later.
Logged



Economic: -4.45
Social: -6.52
Yelnoc
YaBB God
*****
Posts: 6559
United States
View Profile WWW
« Reply #2 on: April 08, 2012, 02:57:06 pm »
Ignore

Here's my attempt at Maine.  I've never been there, so I'm not sure how much sense this makes.  I wanted to put the entire coastline in one district which I accomplished.  The two districts are separated by I-95 and then by Airline Road, though there is some overlap.

ME-01
VAP is 95.2% white.
60.8% voted for Obama in 2008.

ME-02
VAP is 96% white.
54.3% voted for Obama in 2008.

Logged

Goodbye
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« Reply #3 on: April 08, 2012, 03:18:56 pm »

Here's my attempt at Maine.  I've never been there, so I'm not sure how much sense this makes.  I wanted to put the entire coastline in one district which I accomplished.  The two districts are separated by I-95 and then by Airline Road, though there is some overlap.

ME-01
VAP is 95.2% white.
60.8% voted for Obama in 2008.

ME-02
VAP is 96% white.
54.3% voted for Obama in 2008.



What is the deviation for the districts? With the large number of towns, the model I have from the whole county states would predict that it should be possible to draw one district exact and the other off by one person (ME has an odd number population). The statement about exact equality should apply to NH as well.
Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
farewell
Lewis Trondheim
YaBB God
*****
Posts: 58527
India


View Profile
« Reply #4 on: April 08, 2012, 03:22:21 pm »
Ignore

Here's my attempt at Maine.  I've never been there, so I'm not sure how much sense this makes.  I wanted to put the entire coastline in one district which I accomplished.
I... I don't think that makes as much sense as, say, the current districts.
Logged

I may conceivably reconsider.

Knowing me it's more likely than not.
traininthedistance
YaBB God
*****
Posts: 2835


View Profile
« Reply #5 on: April 08, 2012, 03:50:12 pm »
Ignore

Rhode Island:



Blue district is -104, green is +103.

While I haven't rigorously proved this is optimal, I'd be shocked if a closer map was found.
Logged



Economic: -4.45
Social: -6.52
traininthedistance
YaBB God
*****
Posts: 2835


View Profile
« Reply #6 on: April 08, 2012, 03:56:09 pm »
Ignore

I suspect perfect equality is going to be hard to make in DRA for Maine, since the voting districts sometimes contain multiple towns, as well as amalgamating wide swaths of the unincorporated territory.  (Yes, Maine has unincorporated territory, which is subdivided into mostly-unpopulated townships and gores.  It's run by the state rather than by counties).

I got it down to 30 and 31, but lost that map when I tried to improve it.
Logged



Economic: -4.45
Social: -6.52
traininthedistance
YaBB God
*****
Posts: 2835


View Profile
« Reply #7 on: April 08, 2012, 04:45:03 pm »
Ignore

Perfection is possible in New Hampshire:



Except it's not perfect since the only road up to the northernmost towns in the blue district  (Errol and a couple mostly unpopulated grants) passes through the green district.  There may be a map without that problem, but I'm not the one to draw it.

Logged



Economic: -4.45
Social: -6.52
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« Reply #8 on: April 08, 2012, 05:38:58 pm »

Here's my version of exact equality for NH. It solves the road problem. With so many towns, I'm not surprised to find more than one exact solution.


Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
JohnnyLongtorso
YaBB God
*****
Posts: 6843


View Profile
« Reply #9 on: April 08, 2012, 07:04:31 pm »
Ignore

Connecticut:



CT-01 (blue) +161 (59.8 Obama)
CT-02 (green) +60 (53.3 Obama)
CT-03 (purple) 0 (63.2 Obama)
CT-04 (red) +1 (66.6 Obama)
CT-05 (yellow) -220 (60.5 Obama)
Logged
Yelnoc
YaBB God
*****
Posts: 6559
United States
View Profile WWW
« Reply #10 on: April 08, 2012, 07:18:36 pm »
Ignore

Here's my attempt at Maine.  I've never been there, so I'm not sure how much sense this makes.  I wanted to put the entire coastline in one district which I accomplished.  The two districts are separated by I-95 and then by Airline Road, though there is some overlap.

ME-01
VAP is 95.2% white.
60.8% voted for Obama in 2008.

ME-02
VAP is 96% white.
54.3% voted for Obama in 2008.



What is the deviation for the districts? With the large number of towns, the model I have from the whole county states would predict that it should be possible to draw one district exact and the other off by one person (ME has an odd number population). The statement about exact equality should apply to NH as well.
I believe +/-2000.  Not great, but I was rushed.

Here's my attempt at Maine.  I've never been there, so I'm not sure how much sense this makes.  I wanted to put the entire coastline in one district which I accomplished.
I... I don't think that makes as much sense as, say, the current districts.
Probably not from a community-of-interest standpoint (again, I know nothing about Maine) but it's more pleasing to my eye.
Logged

Goodbye
jimrtex
YaBB God
*****
Posts: 5523
Macedonia, The Former Yugoslav Republic of


View Profile
« Reply #11 on: April 08, 2012, 10:06:53 pm »
Ignore

As many have noted, counties are not the critical unit in New England. In fact counties have been dissolved as governing bodies in CT, RI, and parts of MA. Towns (and cities when so named) are the fundamental unit. For these states keeping towns intact is far more important than keeping counties (or their historical boundaries) intact.

So, the challenge here is to divide the New England states into CDs while keeping towns intact and minimizing the average population deviation. Now that I have a statistical model from the whole-county states, I'll be interested in comparing to these states with far more jurisdictions to manipulate.

As before the rule requires no district to have more than a 0.5% deviation from the ideal, and point contiguity is not allowed. Bonus points for making districts that are connected by roads internally, but we won't let that stand in the way of a great plan, such as the near-perfect split of ID.

Also, some have suggested that NJ could be on this list as well, so I'll include it (boroughs, cities townships, etc all count as towns). However, I won't extend to NY and PA since both have cities that exceed the population of one CD, and can't really fit this rule.

CT (5 CDs, 169 towns)
ME (2 CDs, 433 towns)
MA (9 CDs, 351 towns)
NH (2 CDs, 234 towns)
NJ (12 CDs, 566 towns)
RI (2 CDs, 39 towns)

Perhaps require modest recognition of counties:

Two districts may split (share parts) of at most one county.  This probably would result in better compactness, and would generally have reasonable road connectivity (you might have to dip across the district line near district boundaries, but not to great an extent.

Conceivably there could be a bonus for fewest county fragments, but there is a risk that in a state like New Hampshire, there might an extreme split that happens not to split any counties.

This might also be a better starting point for states to the west of New England which have a well-developed system of townships.

Logged
traininthedistance
YaBB God
*****
Posts: 2835


View Profile
« Reply #12 on: April 08, 2012, 11:07:02 pm »
Ignore

As many have noted, counties are not the critical unit in New England. In fact counties have been dissolved as governing bodies in CT, RI, and parts of MA. Towns (and cities when so named) are the fundamental unit. For these states keeping towns intact is far more important than keeping counties (or their historical boundaries) intact.

So, the challenge here is to divide the New England states into CDs while keeping towns intact and minimizing the average population deviation. Now that I have a statistical model from the whole-county states, I'll be interested in comparing to these states with far more jurisdictions to manipulate.

As before the rule requires no district to have more than a 0.5% deviation from the ideal, and point contiguity is not allowed. Bonus points for making districts that are connected by roads internally, but we won't let that stand in the way of a great plan, such as the near-perfect split of ID.

Also, some have suggested that NJ could be on this list as well, so I'll include it (boroughs, cities townships, etc all count as towns). However, I won't extend to NY and PA since both have cities that exceed the population of one CD, and can't really fit this rule.

CT (5 CDs, 169 towns)
ME (2 CDs, 433 towns)
MA (9 CDs, 351 towns)
NH (2 CDs, 234 towns)
NJ (12 CDs, 566 towns)
RI (2 CDs, 39 towns)

Perhaps require modest recognition of counties:

Two districts may split (share parts) of at most one county.  This probably would result in better compactness, and would generally have reasonable road connectivity (you might have to dip across the district line near district boundaries, but not to great an extent.

Conceivably there could be a bonus for fewest county fragments, but there is a risk that in a state like New Hampshire, there might an extreme split that happens not to split any counties.

This might also be a better starting point for states to the west of New England which have a well-developed system of townships.



This would be a good idea for NJ (and also for PA and NY, assuming you find some satisfactory way to deal with Philly, NYC, Hempstead, and Brookhaven); and it would work similarly well in most of the Midwest, where rural townships aren't necessarily important or powerful, but they do exist across (almost) the entire state.

This isn't really necessary in New England, my sense is that the only counties in New England which really mean anything at all to their residents are Berkshire and Aroostook.  I guess Nantucket and Dukes too, basically just the parts with a tangible geographic separation from the rest of the state.
Logged



Economic: -4.45
Social: -6.52
traininthedistance
YaBB God
*****
Posts: 2835


View Profile
« Reply #13 on: April 09, 2012, 12:53:47 am »
Ignore

This uglier NJ map drops the maximum deviation to under 500:



1: -328
2: -485
3: 170
4: 444
5: -177
6: 268
7: 490, 49.5% White VAP
8: -19
9: 452
10: -360, 42.3% Black VAP
11: -425
12: -32, 41.8% Hispanic VAP

Some more improvement is probably possible, but I suspect not much, and certainly any closer map will flaunt county lines even worse than this one.  The large towns in Newark/Jersey City area and southern Middlesex both provide significant obstacles to any further improvement.
Logged



Economic: -4.45
Social: -6.52
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« Reply #14 on: April 09, 2012, 06:24:39 am »

As many have noted, counties are not the critical unit in New England. In fact counties have been dissolved as governing bodies in CT, RI, and parts of MA. Towns (and cities when so named) are the fundamental unit. For these states keeping towns intact is far more important than keeping counties (or their historical boundaries) intact.

So, the challenge here is to divide the New England states into CDs while keeping towns intact and minimizing the average population deviation. Now that I have a statistical model from the whole-county states, I'll be interested in comparing to these states with far more jurisdictions to manipulate.

As before the rule requires no district to have more than a 0.5% deviation from the ideal, and point contiguity is not allowed. Bonus points for making districts that are connected by roads internally, but we won't let that stand in the way of a great plan, such as the near-perfect split of ID.

Also, some have suggested that NJ could be on this list as well, so I'll include it (boroughs, cities townships, etc all count as towns). However, I won't extend to NY and PA since both have cities that exceed the population of one CD, and can't really fit this rule.

CT (5 CDs, 169 towns)
ME (2 CDs, 433 towns)
MA (9 CDs, 351 towns)
NH (2 CDs, 234 towns)
NJ (12 CDs, 566 towns)
RI (2 CDs, 39 towns)

Perhaps require modest recognition of counties:

Two districts may split (share parts) of at most one county.  This probably would result in better compactness, and would generally have reasonable road connectivity (you might have to dip across the district line near district boundaries, but not to great an extent.

Conceivably there could be a bonus for fewest county fragments, but there is a risk that in a state like New Hampshire, there might an extreme split that happens not to split any counties.

This might also be a better starting point for states to the west of New England which have a well-developed system of townships.


For southern New England counties have little meaning except for history and the Census Bureau. Judicial districts have replaced that function of counties in those three states, and are not necessarily co-terminous with the old counties. Counties still have a judicial role in northern New England and maintain a governing body, but it is weak compared to town government.

As one heads out of New England, counties become stronger, so there I agree that some hybrid between county integrity and smaller unit integrity makes more sense where counties alone can't suffice. NJ sits in between. Town government is important, and when NJ does redistricting they take note of the number of municipalities, not counties, that are split. I don't have that same sense of concern about county integrity, but I would defer to current residents who could help on that point.
Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« Reply #15 on: April 09, 2012, 06:46:07 am »

Rhode Island:



Blue district is -104, green is +103.

While I haven't rigorously proved this is optimal, I'd be shocked if a closer map was found.

I prefer your map as an optimal solution, but if I remove restrictions on connectivity by road or ferry there is another. The towns are still contiguous, and the deviations are -31/+30.

Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« Reply #16 on: April 09, 2012, 11:35:30 pm »

This version for ME roughly follows the existing districts and is exact with whole towns as grouped by DRA into voting districts. CD 1 has one more person than CD 2.

Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« Reply #17 on: April 10, 2012, 11:47:40 am »

Here is my plan for MA using whole towns. The average deviation is 24.11 and the range is 82.

CD 1: -5
CD 2: +31
CD 3: +9
CD 4: -23
CD 5: -24
CD 6: -35
CD 7: +47
CD 8: -20
CD 9: +23

I suspect that it can be made better. In particular, CD 4 and 9 in SE MA together are exactly 2 districts, so a deviation of less than 23 might be possible. Also CD 1 might have an arrangement that is closer to perfect given the large number of small towns to work with in W MA.


Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
Gravis Marketing
brittain33
YaBB God
*****
Posts: 12610


View Profile
« Reply #18 on: April 10, 2012, 11:53:06 am »
Ignore

Wow. That's quite a good map without even looking at the deviations. Although the D primary in that Boston district would be ugly.
Logged
jimrtex
YaBB God
*****
Posts: 5523
Macedonia, The Former Yugoslav Republic of


View Profile
« Reply #19 on: April 10, 2012, 12:32:31 pm »
Ignore

Perhaps require modest recognition of counties:

Two districts may split (share parts) of at most one county.  This probably would result in better compactness, and would generally have reasonable road connectivity (you might have to dip across the district line near district boundaries, but not to great an extent.

Conceivably there could be a bonus for fewest county fragments, but there is a risk that in a state like New Hampshire, there might an extreme split that happens not to split any counties.

This might also be a better starting point for states to the west of New England which have a well-developed system of townships.


For southern New England counties have little meaning except for history and the Census Bureau. Judicial districts have replaced that function of counties in those three states, and are not necessarily co-terminous with the old counties. Counties still have a judicial role in northern New England and maintain a governing body, but it is weak compared to town government.

As one heads out of New England, counties become stronger, so there I agree that some hybrid between county integrity and smaller unit integrity makes more sense where counties alone can't suffice. NJ sits in between. Town government is important, and when NJ does redistricting they take note of the number of municipalities, not counties, that are split. I don't have that same sense of concern about county integrity, but I would defer to current residents who could help on that point.
I'm suggesting use of counties more as a an indication of compactness.   Counties are too few and too populous to be used in New England, while towns may be too numerous and two small to prevent gerrymandering (see the awful Massachusetts congressional map and the original Gerry-mander).

An alternative might be to measure compactness based on the number of town lines that are used as district boundaries (this is easier to measure than distance)
Logged
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« Reply #20 on: April 10, 2012, 05:20:55 pm »

Perhaps require modest recognition of counties:

Two districts may split (share parts) of at most one county.  This probably would result in better compactness, and would generally have reasonable road connectivity (you might have to dip across the district line near district boundaries, but not to great an extent.

Conceivably there could be a bonus for fewest county fragments, but there is a risk that in a state like New Hampshire, there might an extreme split that happens not to split any counties.

This might also be a better starting point for states to the west of New England which have a well-developed system of townships.


For southern New England counties have little meaning except for history and the Census Bureau. Judicial districts have replaced that function of counties in those three states, and are not necessarily co-terminous with the old counties. Counties still have a judicial role in northern New England and maintain a governing body, but it is weak compared to town government.

As one heads out of New England, counties become stronger, so there I agree that some hybrid between county integrity and smaller unit integrity makes more sense where counties alone can't suffice. NJ sits in between. Town government is important, and when NJ does redistricting they take note of the number of municipalities, not counties, that are split. I don't have that same sense of concern about county integrity, but I would defer to current residents who could help on that point.
I'm suggesting use of counties more as a an indication of compactness.   Counties are too few and too populous to be used in New England, while towns may be too numerous and two small to prevent gerrymandering (see the awful Massachusetts congressional map and the original Gerry-mander).

An alternative might be to measure compactness based on the number of town lines that are used as district boundaries (this is easier to measure than distance)

I was thinking that testing the perimeter would suffice, but your alternative may have merit, so I'll have to think how to properly count them. I'd like to keep a district like my MA CD 1 from being penalized because it splits three counties. It's clearly compact.
Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« Reply #21 on: April 11, 2012, 04:01:43 pm »

Connecticut:



CT-01 (blue) +161 (59.8 Obama)
CT-02 (green) +60 (53.3 Obama)
CT-03 (purple) 0 (63.2 Obama)
CT-04 (red) +1 (66.6 Obama)
CT-05 (yellow) -220 (60.5 Obama)

I took your two central CDs and rearranged the other three while throwing compactness to the wind. The result is a mess but the average deviation is 3.2 and the range is 13.



CD 1: +1
CD 2: +8
CD 3: 0
CD 4: -5
CD 5: -2
Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« Reply #22 on: April 12, 2012, 12:41:33 am »

Here's my better version of CT. The average deviation is 5.2 and the range is 20, so that is slightly higher than the map above, but the districts are kept much more compact.



CD 1: +2
CD 2: +3
CD 3: +9
CD 4: -1
CD 5: -11

In addition, CD 4 was 56.0-43.2 Obama, and CD 5 was 56.6-41.9 Obama so both are somewhat competitive (D+3 to D+4).
Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
muon2
Moderator
YaBB God
*****
Posts: 8079


View Profile
« Reply #23 on: April 12, 2012, 07:27:15 am »

Here's my better version of CT. The average deviation is 5.2 and the range is 20, so that is slightly higher than the map above, but the districts are kept much more compact.



CD 1: +2
CD 2: +3
CD 3: +9
CD 4: -1
CD 5: -11

In addition, CD 4 was 56.0-43.2 Obama, and CD 5 was 56.6-41.9 Obama so both are somewhat competitive (D+3 to D+4).

To follow up on jimrtex's suggested compactness measure, the number of towns touching a town in a different district in this plan is 82. That compares with 99 border towns in the plan with av dev of 3.2. One can make a model that the marginal improvement in deviation does not offset the increase in border towns.
Logged


Lunar Eclipse of April 15, 2014 with the star Spica.
Pages: [1] Print 
« previous next »
Jump to:  


Login with username, password and session length

Logout

Powered by SMF 1.1.19 | SMF © 2013, Simple Machines