Forum Redistricting Commission

[muon2]

In any case the key issue with competitive districts is to promote a map that is responsive, ie one that allows the voters enough seats to change the partisan make-up of delegations in a way that reflects the opinions of the electorate. Since there is no proportional representation, the danger is a plan that is too effective at separating the parties into uncompetitive districts. There should be some number of competitive seats to protect against that. That's the value of measuring the polarization of a plan.

[muon2]

If this helps, here are more results by CD for my plan:

And I made two sets of PVI's; one is the standard version for Presidential races and the other was based on the five statewide races I calculated.

Very nice, and useful to see that the state results track the presidential results fairly well. BTW, did you want to vote on Items 4 and 5?

[muon2]

There is no reason for such a tight limit on range since it leads to gratuitous chops and gerrymandering.

The commission can amend an item. However, a 10% range is presumably too large, since it is the allowed state standard and the federal standard is stricter. Federal ranges of up to 1% have been upheld, and like with the VRA the commission has to determine if it wishes to push federal law by extending beyond what is clearly permissible.

Speaking of the commission, ElectionsGuy has PM'ed me to say that he declines to take a seat on the commission. That moves X into the role of commissioner. JerryArkansas now joins SLCValleyMan as an alternate.

[muon2]

I see there are no votes on Item 7, nor any questions or suggested changes. To remind the commission, under Item 3 the INEQUALITY score is used to break a tie when two plans have both the same CHOP and EROSITY scores. The higher INEQUALITY is eliminated. If one uses the range directly then a one person difference would eliminate a plan. By grouping into bands based on some statistical measure then there remains the possibility that two plans with identical CHOP and EROSITY could still both go before commission for their vote if their ranges are statistically close. The INEQUALITY table in Item 7 is one way to provide that statistical grouping. If the commission prefers they can use the range directly and not group ranges into a score.

[muon2]

One can think of each county chop as having one less whole county CD in a state, which is equivalent to having one less chop of the state as a whole. So, the table is another way of saying when two plans have statistically equivalent ranges that would be expected for that number of whole county CDs or whole county groups of CDs. For example train-B (submission 2) has 3 whole county CDs (5, 6, 9) and two whole county groups of CDs (NoVa, Richmond/Hampton Roads) for a total of 4 state chops. train didn't list the deviations for his CDs, but I calculate the range to be 2211 which is a score of 6, so it is statistically above the theoretical best.

FWIW, my range is actually 1780; from +988 (District 11) to -792 (District 5).  Knowing that these are the ranges you're choosing, I suppose just splitting Fairfax differently might get me down to a score of 5, and I'll try that at some point. I believe districts 6 and 9 together are just over 900 (they're around plus-minus 470), so pushing inequality further than that won't really work.  (Though I would argue strenuously against any efforts to allow inequality more than 1 percent; and would never myself even make a map that availed itself of half that wiggle room.)

I found my error in a thin precinct in Bedford adjacent to Lynchburg. That's why deviation is included in Item 1 for plan submissions. In any case the Item 7 table would still put the score at I=6. After the correction I did some boundary manipulation. Small shifts in Portsmouth and Norfolk gets the deviations to -405/-214/+204 for those three CDs without changing the BVAP in CD 4. Small shifts in Loudoun and Fairfax got deviations of +131/+586/+534/+298 for CDs 7,8,10,11. That would bring the range down to 1378 for a score of I=5. Those types of shifts I would expect the crowd to submit to improve plans during the submission phase.

[muon2]

There is no reason for such a tight limit on range since it leads to gratuitous chops and gerrymandering.

The commission can amend an item. However, a 10% range is presumably too large, since it is the allowed state standard and the federal standard is stricter. Federal ranges of up to 1% have been upheld, and like with the VRA the commission has to determine if it wishes to push federal law by extending beyond what is clearly permissible.

Speaking of the commission, ElectionsGuy has PM'ed me to say that he declines to take a seat on the commission. That moves X into the role of commissioner. JerryArkansas now joins SLCValleyMan as an alternate.

'Tennant' makes it clear that 'Karcher' was misread, and that "practicable" must take into account whether the deviations of a plan are necessary as a consequence of other state objectives.  There is nothing in 'Karcher' nor 'Kirkpatrick' that says 1% is an outside limit.

It happens that New Jersey, using townships could reach such a limit, and so could West Virginia.

Ideally, a Virginia commission would have established its other criteria first rather than after the population equality criteria.

I understand that, but a commission will still have a number in mind as to the outside limit of range. Even in local jurisdictions that I've worked with the body sets a range if they want to be tighter than 10% before they weigh in on other criteria. The process invariably begins by showing how large the variances have become between districts, and if the range is acceptable a local body almost always chooses to make no changes. If the post-census range is not acceptable, the body determines what is acceptable.

Practicable is still going to be tighter than substantially equal, so the federal standard is going to be stricter than the state standard. How far can SCOTUS go before it looks like the federal standard isn't really different than the state standard?

[muon2]

Ideally, a Virginia commission would have established its other criteria first rather than after the population equality criteria.

I understand that, but a commission will still have a number in mind as to the outside limit of range. Even in local jurisdictions that I've worked with the body sets a range if they want to be tighter than 10% before they weigh in on other criteria. Practicable is still going to be tighter than substantially equal, so the federal standard is going to be stricter than the state standard. How far can SCOTUS go before it looks like the federal standard isn't really different than the state standard?

Wesberry v Sanders was decided on the wrong basis.   Read Justice White's concurring opinion.  And it was a mistake to set a safe harbor of 5% for state and local redistricting.  Both lead to gerrymandering.

Consider the current Virginia litigation.  Virginia reduced the number of chops, but then claimed that some were necessary to get equal population, and therefore it was OK to have bunches for VA-3.  Had they set a higher standard for political integrity, then they couldn't have done all the chops for VA-3.

If we are going to use the fundamental principle behind 'Wesberry v Sanders', then we should be balancing CVAP.

I agree with you about how VA could have fared better in court, but that doesn't address population variance. The theoretical question of constitutional law doesn't necessarily solve the problem that would face a real commission today. In general a commission acts after a Census has released new population data. If the districts are malapportioned due to the new populations, then they must be redrawn. That forces the question as to how much variance in a CD would be construed as malapportioned after a Census, since that then sets the inequality goal for the commission. Despite the lack of clarity on that point in Karcher and other cases, experts will advise redistricting panels as to what range inequality may be a safe harbor when other neutral criteria are used in the plan. Given that SCOTUS has upheld plans of nearly 1% range, that number seems like sound advice to a commission, unless that commission specifically wants to set up a test case to demonstrate that larger ranges will survive SCOTUS scrutiny.

Adding: In Tennant WV had a range of 9.6% of the quota for its CDs after the Census and before redistricting. That would qualify as substantially equal under a state standard, but the  parties recognized that it would not be a minor variation and could not be justified for a federal plan. That prompted the legislature to redraw the districts. The adopted plan, upheld by SCOTUS, reduced 9.6% to 0.8% and that then constituted a minor variation that could be justified by the criteria used by the state. If a plan has more than a 1% range I would expect significant legal debate as to whether that can be considered a minor variation.

[muon2]

That a commission acts after the census has been released is a flaw in process.

If the Census shows the population changes to the district have induced minimal variations to equal population, there is no reason for the commission to act at all. Therefore the commission must wait for the Census results.

You must be logged in to read this quote.

I have no disagreement, other than to add that VA still had the constraint of creating a CD where the black population would be likely to elect the candidate of their choice.

You must be logged in to read this quote.

True, but in local redistricting the question is still put the body as to what maximum range they will tolerate, and many adopt a standard tighter than 10%. In light of Tennant, I would expect advisers to congressional commissions to do the same, but pose the question in terms of what range will they consider to be consistent with a minor variance: 0.79%, 1%, 2%, etc.

You must be logged in to read this quote.

The range of the AR plan was very nearly 1% (1.002% from the Statistical Abstract for 2003). Even if it was challenged and upheld it wouldn't give us much guidance beyond the 0.8% allowed in Tennant.

You must be logged in to read this quote.

And if a 2% range is ok, then someone else will propose a plan with fewer chops or less erosity and a 5% range and claim its ok by the same logic. But at some point variances cease to be minor. I don't know where that point is but it is presumably somewhere between 0.8% and 9.6% based on the facts in Tennant. As an adviser I would leave that decision to the commission.

[muon2]

The discussion/voting period for 6, 7 and 8 have passed. Item 6 has four votes in favor from Morgieb, X, Miles and fuzzy, so it is adopted. Item 7 only has one vote in favor from Morgieb, so votes are needed from other commissioners (or alternates) before it is adopted, rejected, or amended. Item 8 has no votes or discussion from the commission, so perhaps an extension of the discussion/voting period is in order.

Would it help if I move the approved items into a separate thread, so one doesn't have to scroll back through multiple pages?

[muon2]

The discussion/voting period for 6, 7 and 8 have passed. Item 6 has four votes in favor from Morgieb, X, Miles and fuzzy, so it is adopted. Item 7 only has one vote in favor from Morgieb, so votes are needed from other commissioners (or alternates) before it is adopted, rejected, or amended. Item 8 has no votes or discussion from the commission, so perhaps an extension of the discussion/voting period is in order.

Would it help if I move the approved items into a separate thread, so one doesn't have to scroll back through multiple pages?

Yeah, that'd be very helpful!  Aye on 7, btw.

I've done that, and those items now approved are in a thread titled The Muon Rules.

[muon2]

There is no area in Virginia that satisfies the first prong of the Gingles test.

I can construct a >50% BVAP CD that includes none of Richmond or Petersburg and is all whole counties/cities (Brunswick to Surry to Suffolk) plus the black areas of the independent cities of the Hampton Roads. It is compact by almost any standard geographic measure, since the erosity is confined to those relatively small areas in the Hampton Roads. The fact that it links rural and urban areas would not cause it to fail Gingles 1.

[muon2]

There is no area in Virginia that satisfies the first prong of the Gingles test.

I can construct a >50% BVAP CD that includes none of Richmond or Petersburg and is all whole counties/cities (Brunswick to Surry to Suffolk) plus the black areas of the independent cities of the Hampton Roads. It is compact by almost any standard geographic measure, since the erosity is confined to those relatively small areas in the Hampton Roads. The fact that it links rural and urban areas would not cause it to fail Gingles 1.

Brunswick is at least as far from Hampton Roads as Richmond.   I assume you crack Newport News, Hampton, Portsmouth, Norfolk, Virginia Beach and Chesapeake.   Do you exclude Isle of Wight?

You have violated traditional redistricting concepts such as integrity of political subdivisions, and unnecessary agglomeration of rural and urbam areas makes it conceptually non-compact.  Your proposed district is no better than the district turned down in Lulac v Perry

I do not exclude Isle of Wight, and as much as we might like it to be otherwise, the VRA does not care if political subdivisions are chopped in order to comply. There is nothing in Gingles to suggest that conceptual compactness has any bearing on prong 1.

LULAC is not applicable here. The question here is not whether the district meets section 2, but whether one is required by section 2. In LULAC there was already a determination that TX needed such a district and SCOTUS found that District 23 had been redrawn in such a way as to deny Latino voters as a group the opportunity to elect a candidate of their choice. District 25 was found to be a non-compact replacement, but District 23 as finally drawn after LULAC is a clear linkage of urban and rural areas with much greater separation than the VA district I suggest would meet section 2.

[muon2]

While I wait for votes on Item 8. Here's another map showing neighborhood areas for a couple of larger independent cities that might be chopped in some plans. The areas are approximated by the voting districts in DRA. Hampton districts are based on neighborhood districts recognized by the City of Hampton. Newport News districts are based on real estate market areas from the MLS.



Hampton Areas

District 1; pop 22,377; BVAP 41.8%
District 2; pop 10,487; BVAP 46.6%
District 3; pop 12,770; BVAP 44.8%
District 4; pop 13,580; BVAP 29.2%
District 5; pop 11,864; BVAP 17.0%
District 6; pop 12,846; BVAP 37.3%
District 7; pop 12,833; BVAP 59.3%
District 8; pop 17,116; BVAP 58.8%
District 9; pop 11,748; BVAP 69.6%
District 10; pop 11,815; BVAP 71.2%


Newport News Areas

Denbigh North (11); pop 52,397; BVAP 37.2%
Denbigh South (12); pop 32,093; BVAP 30.2%
Midtown West (13); pop 26,531; BVAP 14.1%
Midtown East (14); pop 36,973; BVAP 30.1%
South (15); pop 32,725; BVAP 75.2%

[muon2]

Proposed combinations of counties and independent cities.

While independent cities in Virginia are treated as county equivalents by the Census Bureau, they really don't represent separate communities of interest, and in some cases are not totally legally separate, with some sharing county offices with their surrounding counties.  The Bureau of Economic Affairs only treats independent cities as a separate entity if they have a population over 100,000, or have absorbed all of their former county.  To some extent in Virginia, "independent city" and "city" are almost synonymous.  The largest towns that are not independent cities are Leesburg and Blacksburg, both with less than 50,000 persons.

Treating independent cities as counties for redistricting purposes causes problems for erosity and chop measurements because of their typical small size.

Hampton, Newport News, and Virginia Beach qualify under both the population criteria and having absorbed their orignal county.  Chesapeake does in a certain sense, since it has 222,000 persons, and since it was created by the merger of the independent city of South Norfolk and remainder of Norfolk County, not in Norfolk or Portsmouth.   Suffolk city qualifies based on its annexation of Nansemond city, which itself was created from the remnant of Nansemond County not in Suffolk city.

The only large independent cities that have not absorbed all of their original counties are Richmond, Roanoke, and Alexandria.  This leaves Henrico, Roanoke, and Arlington counties as the only remainders of counties from which the larger cities have separated.  Roanoke County also includes the smaller independent city of Salem.   The District of Columbia originally had two counties, Washington (the portion ceded by Maryland) and Alexandria (the portion ceded by Virginia) and three municipalities, Alexandria, Georgetown, and Washington.  When the area south of the Potomac was retroceded to Virginia, it became Alexandria County, Virginia.  The portion not within Alexandria city was renamed to Arlington County in 1920.

The Bureau of Economic Affairs combines Colonial Heights with Dinwiddie County and Petersburg city.  This is problematic since Colonial Heights is barely contiguous with Petersburg, and was created from Chesterfield County.  Placing it with Dinwiddie County would give greater importance to economics than political history, and would call in to question the whole concept of using political subdivisions as representing communities of interest.  So in the following tables, I have placed Colonial Heights with Chesterfield County.

Most independent cities were created from a single county, though some have annexed areas in adjacent counties.   Galax city is the exception, having been created from almost equal population from Carroll and Grayson counties.  The Census Bureau and BEA both place Galax with Carroll County.

Independent Cities that are combined with original counties, or are treated as counties for chop and erosity analysis.

Albemarle + Charlottesville
Alexandria  (IC)
Alleghany + Covington
Augusta, Staunton + Waynesboro
Bedford + Bedford city (as of 2013, Bedford is no longer an independent city).
Campbell + Lynchburg
Carroll + Galax
Chesapeake (IC)
Chesterfield + Colonial Heights
Dinwiddie + Petersburg
Fairfax, Fairfax City + Falls Church
Frederick + Winchester
Greensville + Emporia
Hampton (IC)
Henry + Martinsville
James City county + Williamsburg
Montgomery + Radford
Newport News (IC)
Norfolk (IC)
Pittsylvania + Danville
Portsmouth (IC)
Prince George + Hopewell
Prince William, Manassas + Manassas Park
Richmond (IC)
Roanoke (IC) (Roanoke city is separate from Roanoke County because of its population)
Roanoke + Salem (but Roanoke County includes Salem (IC)).
Rockbridge, Buena Vista + Lexington
Rockingham + Harrisonburg
Southampton + Franklin
Spotsylvania + Fredericksburg
Suffolk (IC)
Virginia Beach (IC)
Washington + Bristol
Wise + Norton
York + Poquoson

Counties with no associated independent cities for measurement of chop and erosity.

Accomack
Amelia
Amherst
Appomattox
Arlington
Bath
Bland
Botetourt
Brunswick
Buchanan
Buckingham
Caroline
Charles City county
Charlotte
Clarke
Craig
Culpeper
Cumberland
Dickenson
Essex
Fauquier
Floyd
Fluvanna
Franklin
Giles
Gloucester
Goochland
Grayson
Greene
Halifax
Hanover
Henrico
Highland
Isle of Wight
King George
King William
King and Queen
Lancaster
Lee
Loudoun
Louisa
Lunenburg
Madison
Mathews
Mecklenburg
Middlesex
Nelson
New Kent
Northampton
Northumberland
Nottoway
Orange
Patrick
Powhatan
Prince Edward
Pulaski
Rappahannock
Richmond county
Russell
Scott
Shenandoah
Smyth
Stafford
Surry
Sussex
Tazewell
Warren
Westmoreland
Wythe

Since there hasn't been much discussion on this point, and only one alternate vote, this seems like a potential amendment to Item 8 before the commission. The originally proposed Item 8 is

Item 8. The primary units of redistricting in VA are the counties and independent cities. Independent cities are treated as equal to counties for redistricting, and are understood to be included when scoring describes counties. Secondary units in redistricting may include larger communities of interest made up of groups of whole counties and smaller communities of interest that wholly divide a county as adopted by the commission.

An amended version to consider is
Item 8A. The primary units of redistricting in VA are the counties. Independent cities are treated as equal to counties for redistricting when the independent cities have populations in excess of 100,000 or have absorbed all of their former county. These qualified independent cities are understood to be included when scoring describes counties. Other independent cities are considered to be part of the county that they are assigned to by the Bureau of Economic Affairs. Secondary units in redistricting may include larger communities of interest made up of groups of whole counties and smaller communities of interest that wholly divide a county as adopted by the commission.

I avoided creating a special exception for Colonial Heights. There are a number of states that have discontiguous counties and municipalities, and I'm not troubled by creating one here in the interest of uniform rules. Of course the commission is free to vote on either Item 8, 8A, or 8A with the Colonial Heights exception.

[muon2]

An alternative definition is provided in the publication Population of States and Counties of the United States, 1790 to 1990.  See Virginia section beginning on publication page 166, PDF page 177.

The BEA standard does provide an objective standard for separating larger cities.   But a state commission would be competent to make its own standard.   I regard the BEA standard for Colonial Heights to be exceptional, rather than other way around.

If creating its own standard based on population of cities, the commission might consider the following:  The largest towns in Virginia are Leesburg and Blacksburg, with populations in the mid-40,000s.   Virginia does not have "independent cities" per se, but rather cities are indepedent of their county, while municipalities which are towns are not.  The independent cities that are of comparable size, Charlottesville, Danville, and Harrisonburg are well interior to their respective counties of Albermarle, Pittsylvania, and Rockingham, such that to even reach the cities with a district boundary, the county would necessarily be split/

The next largest independent city is Lynchburg, with a population of just over 75,000.  A threshold of 50,000 might be used, since that is the threshold for classifying an area as metropolitan statistical area.  On the other hand, metropolitan statistical areas are comprised of counties, so using an independent city is somewhat anomalous, on a national scale.  Lynchburg is also problematic because of its location at the intersection of three counties.   Rather than being treated of a community of interest, it might serve as an attractor for a district boundary.

The 50K threshold is another reasonable alternative to 100K and more consistent with the use of UCCs derived from MSAs. In fact one could go a step further and define an independent city as separate from the county if it has at least 25K in an urbanized area, exactly as for the UCCs. Lynchburg shouldn't be a problem as a border community since the CoI should be addressed by the UCC.

[muon2]

Since the commission seems to be on holiday, I'll add another set of city subdivisions. In the meantime, as jimrtex notes through his suggestions, Item 8 sets the direction of the scoring for chops and erosity, since it defines the primary units that should be preserved. There are at least four variants for the commission to choose from, and the next sets of rules are on hold until this item is resolved.

Here are the subdivisions for Norfolk and adjacent cities south of Hampton Roads. The five Norfolk subdivisions follow the neighborhood service areas. The four Portsmouth subdivisions follow real estate areas. The six Chesapeake subdivisions follow the official boroughs. The seven Virginia Beach subdivisions follow the former boroughs as now used by the city public works department.



City of Norfolk
Neighborhood Area 1: pop 17,545; BVAP 66.4%
Neighborhood Area 2: pop 45,742; BVAP 68.1%
Neighborhood Area 3: pop 40,459; BVAP 54.2%
Neighborhood Area 4: pop 64,574; BVAP 25.9%
Neighborhood Area 5: pop 74,483; BVAP 23.1%

City of Portsmouth
Churchland Area (6): pop 25,739; BVAP 42.4%
Olde Towne Area (7): pop 22,172; BVAP 67.8%
Victory Area ( 8 ): pop 23,816; BVAP 69.3%
Midtown Area (9): pop 23,808; BVAP 24.1%

City of Chesapeake
Western Branch Borough (10): pop 32,624; BVAP 30.0%
Deep Creek Borough (11): pop 36,223; BVAP 35.6%
Pleasant Grove Borough (12): pop 51,806; BVAP 9.5%
South Norfolk Borough (13): pop 23,894; BVAP 53.6%
Washington Borough (14): pop 68,346; BVAP 31.6%
Butts Road Borough (15): pop 9,316; BVAP 23.1%

City of Virginia Beach
Bayside (16): pop 75,350; BVAP 22.7%
Centerville (17): pop 82,741; BVAP 23.6%
Kempsville (18): pop 56,569; BVAP 16.6%
Beach (19): pop 49,317; BVAP 13.6%
Lynnhaven (20): pop 41,511; BVAP 3.9%
Rose Hall (21): pop 66,857; BVAP 26.1%
Princess Anne (22): pop 65,649; BVAP 11.8%

[muon2]

Item 8 is still unresolved, so I want to show how it would impact future rules about scoring. One type of cluster is the urban county cluster (UCC) and is defined as those counties in the same metropolitan statistical area that have an urbanized population of 25K or more or have at least 40% urbanized area. The idea is that UCCs represent a community of interest as important as a county. Chops of a UCC in excess of the minimum required count towards the total chop count. One assumption is that UCCs must consist of more than one county, otherwise a chop of that county results in two chops instead of one, which may be excessive.

Here's the map of VA UCCs.



The pink areas don't depend on whether the independent cities (ICs) are merged into their counties or not. The numbers on those UCCs represent the minimum number of CDs to cover each of those UCCs, so an excess beyond that number creates extra chops. The orange area loses Dinwiddie county if ICs are not merged, but keeps it if Petersburg (32K) is merged into Dinwiddie.

The multiple county rule comes into play when the Item 8 question of ICs is folded in. The tan clusters meet the standard if ICs are always separate, but become single counties if merged for ICs under 25K (Montgomery/Radford) or 50 K (Albemarle/Charlottesville, Frederick/Winchester). The yellow cluster (Roanoke/Roanoke&Salem) at 100K would also become a single county cluster. So mergers could cause these clusters to disappear as UCCs.

The grey clusters would not be UCCs if ICs are separate. However, if their ICs merged they would qualify, but at that point they are just single county clusters. Augusta county has 13K (18%) urbanized and urbanized cities of 23K (Staunton) and 20K (Waynesboro). Campbell county has 17K (32%) urbanized and Lynchburg with 74K. Rockingham county has 18K (23%) urbanized and Harrisonburg with 49K. Washington county has 16K (28%) urbanized and Bristol with 18K.

[muon2]

The status of ICs under Item 8 also affect minority county clusters which is another community of interest. A black MCC consists of connected counties where the BVAP is 40% or or more. Chops of an MCC add to the chop count and discourage cracking the minority population in the MCC.

A BVAP map of VA look like this, with the following key.
yellow 25.0-33.3%
lime 33.4-39.9%
green 40.0%-49.9%
dark green > 50.0%



Contiguous counties across water without a bridge or ferry are not connected. If independent cities are all counted as counties then there are two clusters shown in medium green. If ICs are merged into their counties then the light green areas are added and there is only one large MCC.

[muon2]

Nay on item 9

Since no one else has joined in the discussion, I'd like to get your opinion X as to which elements in item 9 don't work. Is it either or both definitions or their application to the scoring?

The work on county clusters emerged from a thread in 2013 (especially the MI pull out linked on the first page) and evolved in another thread (AL delves into minority communities of interest starting on page 4), leading to the final work on UCCs stickied on this board.

The problem that was observed was that some urban areas that spanned counties were being split just to preserve county lines and it both ignored the natural community of interest of an urban center with its suburbs and tended to wash out medium sized urban vote centers by blending in lots of rural area. The problem in VRA states was to identify areas of counties dominated by a single minority interest to determine a community of interest that recognized race, but didn't make race the predominant factor in drawing a plan. The result from those threads was to identify clusters that could be objectively defined and represented a community of interest that should be given formal consideration in a redistricting plan.

Also, did you have an opinion on Item 8?

[muon2]

Fairfax county is the one county in VA that must have a macrochop. So, it will always need county subdivisions. I used the county GIS to produce this map showing the communities (CDPs) in the county as well as the red areas separate from the communities. The city and town lines on DRA match up with these areas and are the natural subunits for Fairfax. If ICs are merged into the county then Fairfax city and Falls Church would just be additional subunits.



As an example of chop counting I'll use Miles-A. Here's his plan for NoVa.



Fairfax county has 2 chops resulting in three CDs in the county.

CD 10 has the smallest chop in Fairfax with a population of 10,045. It doesn't include all of a CDP, but the remainder is in two VTDs that span CDPs so there is no chop of the county subunit.

CD 8 is the other chop with 368,013 (or 380,345 with Falls Church) and is large enough to be a macrochop. All the CDPs chopped between CD 8 and 11 are due to VTDs that span CDPs so there are no additional subunit chops.

Prince William county is also chopped, and the smaller of the two districts is CD 10 with a population of 149,996 (or 202,090 if Manassas and Manassas Park are merged in). So this is also a macrochop. CDPs don't cover Prince William as completely as they do Fairfax, so some subdivision of the undesignated area would be useful. The simplest would be to use VTDs, in which case there can be no chops of the undesignated part of the county since DRA only maps at the level of VTDs.

The total chop count in NoVa here is 3.

[muon2]

Let me continue to the southern end of Miles-A.



CD 3 has a chop of 80,107 in Henrico so its a macrochop. Within Henrico there is no chop of a CDP, so there is only the one chop.

CD 7 has a chop of 56,374 in Richmond City so it's a macrochop. If the size was kept under 36,368 it wouldn't trigger the macrochop provisions, but let's look at what that does here. By taking the larger chop, neighborhood integrity now matters as much as county integrity in rural areas. The plan chops the Southside, West End and Northside (3,790 just over the microchop limit). As drawn without the benefit of the neighborhoods, it counts as 4 towards the CHOP score. However, it would be easy to put all the West End in CD 7, Southside in CD 3 and only chop Northside, reducing the chop count by 2 but keeping CD 3 BVAP at 50.4%.

CD 7 has a chop of 36,947 in Chesterfield which is just over the threshold for a macrochop. However, there are no chopped CDPs, so there is only one chop.

Neither Dinwiddie nor Prince George are chopped if ICs are considered equal to counties under Item 8. However, if either Item 8A, 8B, or 8C is adopted, putting Hopewell in CD 4 counts as a chop of Prince George. If either Item 8A or 8B is adopted (not 8C) then Dinwiddie is chopped by putting Petersburg in CD 3. Petersburg stays separate in 8C since its population is over 25K.

CD 3 has a chop of 5,900 in James City county. The fact that CD 1 has two disconnected fragments in James City does not affect the CHOP score beyond the chop for CD 3.

CD 2 has a macrochop of 106,560 in Norfolk City. As drawn the plan chops all five neighborhood areas within Norfolk it would gain 5 extra in the CHOP score. As with Richmond, it is easy to reduce the neighborhood chops to 1, and it actually increase BVAP for CD 3.

So as drawn Miles-A scores 13 chops in the SE region as drawn. If the plan is modified to reflect neighborhoods in Richmond and Norfolk then the score drops to 7 in the SE region.

CHOP scores depend on the treatment of ICs under Item 8. If ICs under 25K are merged into their counties for redistricting, the the score goes up by one. If ICs under 50K are merged it goes up 2. This example shows the types of districts that are affected by Item 8.

[muon2]

Finally, let me look at Miles-A from the state level and look at the potential impact of Item 9.



All the county and subunit chops are in the two areas of NoVa and SE, and if I use the fully independent version of Item 8 the CHOP score is 3+13 = 16 or it reduces to 10 with the aforementioned neighborhood changes.

The urbanized area of NoVa that meets the definition of the UCC is covered by all or part of 5 CDs (1,7,8,10,11). It could be covered by as few as 4 so the score would increase to 17 if Item 9 were adopted. From a policy standpoint, this would say that it is preferable to keep Fredricksburg with the DC area, and for instance place the Williamsburg area with another CD in the SE unless it saves a county chop.

The Hampton Roads UCC area can be covered by as few three CDs, but in Miles-A is covered by 4 (1,2,3,4). Having CD 1 run from Dale City to Williamsburg could be the cause, and it is the type of district that Item 9 tries to suppress. Then again, the extra CD could also be the result of VRA compliance, and then it isn't optional. Under Item 9 the CHOP is up to 18.

The Richmond UCC can be covered by as few as 2 CDs, and in this plan there are 3 (3,4,7). The issue of Item 8 ICs does not affect Miles-A in this case. The VRA will tend to force a chop of this UCC. Like in large counties, some score increases may be unavoidable so under Item 9 the CHOP is up to 19.

None of the other potential UCCs that depend on Item 8 are affected.

The MCC originated to avoid cracking minority populations that span multiple counties and recognize them as a measurable community of interest. The goal is to discourage splits like the one here that puts Greensville and Brunswick in CD 5, unless it helps avoid chops elsewhere.

If the ICs all remain independent under Item 8 then there are two in VA. If they are merged then there is just one larger MCC. If CVAP is used instead of VAP Norfolk and Hampton are added to the eastern MCC (and the unified one). So there are four options. IC separate+VAP, IC separate+CVAP, IC merged+VAP, and IC merged+CVAP. The last one is 96.6% of the quota so it is difficult to avoid a chop.

IC separate+VAP: 3 CDs cover the western MCC and 2 CDs cover the eastern MCC for a total of 3 chops.

IC separate+CVAP: 3 CDs cover the western MCC and 3 CDs cover the eastern MCC for a total of 4 chops.

IC merged+VAP: 3 CDs cover the MCC for a total of 2 chops.

IC merged+CVAP: 4 CDs cover the MCC for a total of 3 chops.

The merged ICs avoid double counting some of the CDs in the chop count. However, the merged ICs would also increase the direct count for this plan. Without merged ICs under Item 9 the CHOP score would be 22 or 23. With merged ICs the CHOP score would range from 22 to 24.

An important part of the Commission's work is setting the basic rules, even when it gets wonky. In a real commission I would expect more time to be spent on setting the criteria than on evaluating plans. Small changes do have an effect on the balance between different factors, so I hope this example helps guide which way the Commission wants to go.

[muon2]

Aye on Item 10.

What's your preference on Item 8, since some version is needed to score Item 10? As Miles-A illustrates in the SE region, the CHOP score depends on whether the ICs are merged into their counties or not.

[muon2]

Muon2, you should put all your items in one place, preferably perhaps at the start of the thread, so it can be easily found. I can't find your list at all anymore. I am wondering if it was accidentally deleted. A map of all the state highways that measure erosity might also be helpful, along with the map of the UC's  as defined by the agreed metric.  By the way, here is a map of the border of the Hudson "urban cluster." Talk about erosity!  

I set up a separate thread to hold the approved items a while ago. I placed a link to that thread in the OP a couple of days ago since the items had moved to page 2 of the board. The unapproved items (8,9,10) aren't there yet. The difficulty seems to be that dealing with independent cities gets very technical, yet one can't construct a set of rules for VA without deciding if they are counties, subunits of counties, or either depending on the population.

As your image shows, erosity is not a constraint for the Census Bureau. That's why we look at whole political units as a cluster.

[muon2]

Muon2, you should put all your items in one place, preferably perhaps at the start of the thread, so it can be easily found. I can't find your list at all anymore. I am wondering if it was accidentally deleted. A map of all the state highways that measure erosity might also be helpful, along with the map of the UC's  as defined by the agreed metric.  By the way, here is a map of the border of the Hudson "urban cluster." Talk about erosity!  

I set up a separate thread to hold the approved items a while ago. I placed a link to that thread in the OP a couple of days ago since the items had moved to page 2 of the board. The unapproved items (8,9,10) aren't there yet. The difficulty seems to be that dealing with independent cities gets very technical, yet one can't construct a set of rules for VA without deciding if they are counties, subunits of counties, or either depending on the population.

As your image shows, erosity is not a constraint for the Census Bureau. That's why we look at whole political units as a cluster.

Perhaps you should sticky it Muon 2, along with helpful maps of urban clusters, and ideally, state highways. I have no idea where it is.

As you initially suggested I had placed it at the top of this thread by means of a link (does the link not work?). I want to limit the stickies since the board has plenty of other active threads. I plan to have another linked thread for plan submissions once the rules are finalized. I do have the sticky thread for jimrtex's UCCs which has been up since 2013, but that is up for debate in the context of the ICs in VA. If this goes well and interest is there for other states, I would provide a more permanent reference for the rules.