Using Urban County Clusters To Guide Redistricting

[jimrtex]

Is the above definition different from the urban cluster definition provided by the census?  I thought the lighter shade counties were as defined by the census for non core counties or whatever the terminology was.

The areas in green show the urban areas in the Detroit area.  Detroit, Toledo, Flint, Ann Arbor, South Lyon-Howell, and Monroe are urbanized areas, that is they have a population greater than 50,000.

Adrian, Lapeer, Milan, Richmond, Goodrich, Chelsea, Dundee, Blissfield, and Almont are urban clusters, that is, an urban area with a population of less than 50,000.

Urban areas are based on density, mostly without regard to political boundaries, using census blocks and census tracts, to measure density.  Urban areas may also have "jumps" and "hops" to connect areas of urban density.  As you drive down a highway between alternating areas of cows and subdivisions, you might remain within an urban area,

If you were driving between the Detroit Urbanized Area and the Ann Arbor Urbanized Area, you could not tell where one ends and the other begins.  The separation is artificial.

Core Based Statistical Areas (CBSA) are based around a core consisting of an urban area.  For each county, the urban area with the largest population in the county.   That is:

St. Clair:  Port Huron UA
Lapeer: Lapeer UC
Genesee: Flint UA
Macomb: Detroit UA
Oakland: Detroit UA
Livingston: South Lyon-Howell UA
Wayne: Detroit UA
Washtenaw: Ann Arbor UA
Monroe: Monroe UA
Lenawee: Adrian UC

These are central counties of a CBSA.   Or more accurately, the central counties of potential CBSA.  Because Macomb, Oakland, and Wayne share the same core urban, they are treated as a unit, for the next step in delineating CBSA.

Commuting flows are used to determine whether counties are outlying counties of a CBSA.   Lapeer, St.Clair, and Livingston becoming outlying counties of the Detroit Metropolitan Statistical Area, because 25% or more of their employed residents work in the 3 big counties.   They lose their status as central counties (or potential central counties).  Genesee, Washtenaw, Lenawee, and Monroe retain their separate status because of lesser relative commuting flows.  Outlying counties need not have a core urban area.  Sanilac County could have qualified as an outlying county.   However, its primary flow is to St.Clair, and a county can't become an outlying county, based on commuting to another outlying county.  If St.Clair had retained its independent status as Port Huron Metropolitan Statistical Area, Sanilac potentially could have been included.

CBSA are classified as Metropolitan Statistical Area or Micropolitan Statistical Area based on whether their core urban area is an urbanized area or an urban cluster.  Adrian Micropolitan Statistical Area is comprised of Lenawee County.   Detroit, Flint, Ann Arbor, and Monroe are metropolitan statistical areas.

[Torie]

And we ignore outlying counties (the lighter shade of pale counties) in an urban cluster, right, for the proposed rule set under discussion?

[jimrtex]

OK thanks then Jimtex. No change. We are still talking about, for redistricting purposes, multi county "urban areas" as defined by the Census Bureau, and nothing more. This is a Goldilocks issue - too many multi county urban areas, and it could spit out a map mess, too few and too many urban chops, and too much gerrymandering potential. But you have come up with something where coincidentally or not, it sails between the Scylla and Charybdis, and comes out just about right - statistics and intuition for at least one time harmoniously coincide. Now if we could just get to the same place with erosity, life would be beautiful.

My definition of Urban County Cluster is not the same as that employed by the Census Bureau for Core Based Statistical Area, Consolidated Statistical Area, or central county cluster (the Census Bureau does not really define central county cluster, but does refer to a group of central counties that share an urbanized area as the core, as a "cluster".  That is likely where I came up with the term - after reading the OMB definition of a CBSA.

Unfortunately, the Census Bureau refers to blobs of high density as "Urban Areas" and smaller urban areas as "Urban Clusters".  Neither are based on counties for the most part (in dividing larger urbanized areas, the division line is placed near the county line).

So we have three definitions from the Census Bureau for "groups of counties".

Consolidated Statistical Area
Metropolitan Statistical Area
Central Counties of Metropolitan Statistical Area.

None of these embody the particular concepts that I am trying to capture.

Consolidated Statistical Area - Too big.

Metropolitan Statistical Area - May include too much rural territory.   Also the definition based on commuting, may exclude nearby towns, while including more remote rural territory.   If the core area is below the population needed for a congressional district, then the surrounding areas may end up being included anyway; and if the extra area forces another district, then both districts may extend far outside the metropolitan area.

Central Counties - the basic problem is that there is not a 1:1 relationship between urbanized areas and counties.  The Mission Viejo--Lake Forest--San Clemente, CA Urbanized Area has 600,000 persons, but is ignored when defining metropolitan areas, because Los Angeles-Long Beach-Anaheim, CA urbanized area has 2.4 million persons in the county.

On the other hand, the 600,000 persons in Los Angeles--Long Beach--Anaheim, CA Urbanized Area who live in San Bernardino County are ignored because even more live in Riverside--San Bernardino, CA Urbanized Area.  If they were reversed, then San Bernardino would be a central county of the Los Angeles metro area.

So what I am proposing is using Metropolitan Statistical Areas as a starting point, and removing the less urban counties.   These are necessarily outlying counties, but not all the outlying counties.

Other outlying counties are retained based on the presence of urbanized areas.  So the areas are those with major concentrations of population that are linked by significant commuting.

[jimrtex]

And we ignore outlying counties (the lighter shade of pale counties) in an urban cluster, right, for the proposed rule set under discussion?

The counties in the whiter shade of pale are outlying counties, but they are not all of the outlying counties.  They might be considered the most outlying counties, those that are connected only by commuting, and without significant urbanization.

[jimrtex]

Continuing with the northeast.




Connecticut

Hartford-West Hartford-East Hartford, CT 1212  1212  (3): Hartford County 894 UA; Middlesex County 166 UA; and Tolland County 153 UA.

Bridgeport-Stamford-Norwalk, CT 917  917  (1): Fairfield County 917 UA.

New Haven-Milford, CT 862  862  (1): New Haven County 862 UA.

Norwich-New London, CT 274  274  (1): New London County 274 UA.

Worcester, MA-CT 118  118  (1): Windham County 118 UA.


Delaware

Philadelphia-Camden-Wilmington, PA-NJ-DE-MD 538  538  (1): New Castle County 538 UA.

Salisbury, MD-DE 197  197  (1): Sussex County 197 UA.

Dover, DE 162  162  (1): Kent County 162 UA.


District of Columbia

Washington-Arlington-Alexandria, DC-VA-MD-WV 602  602  (1): District of Columbia 602 UA.


Illinois

Chicago-Naperville-Elgin, IL-IN-WI 8587  8587  (9): Cook County 5195 UA; DuPage County 917 UA; Lake County 703 UA; Will County 678 UA; Kane County 515 UA; McHenry County 309 UA; Kendall County 115 UA; DeKalb County 105 UA; and Grundy County 50 UC.

St. Louis, MO-IL 704  572  (8): St. Clair County 270 UA; Madison County 269 UA; Macoupin County 48 NQUA; Clinton County 38 NQUA; Monroe County 33 UA; Jersey County 23 NQUA; Bond County 18 NQUA; and Calhoun County 5 NUA.

Peoria, IL 379  322  (5): Peoria County 186 UA; Tazewell County 135 UA; Woodford County 39 NQUA; Marshall County 13 NUA; and Stark County 6 NUA.

Rockford, IL 349  349  (2): Winnebago County 295 UA; and Boone County 54 UA.

Champaign-Urbana, IL 232  201  (3): Champaign County 201 UA; Piatt County 17 NQUA; and Ford County 14 NQUA.

Davenport-Moline-Rock Island, IA-IL 214  198  (3): Rock Island County 148 UA; Henry County 50 UC; and Mercer County 16 NQUA.

Springfield, IL 210  197  (2): Sangamon County 197 UA; and Menard County 13 NQUA.

Bloomington, IL 186  170  (2): McLean County 170 UA; and De Witt County 17 NQUA.

Carbondale-Marion, IL 127  127  (2): Williamson County 66 UA; and Jackson County 60 UA.

Kankakee, IL 113  113  (1): Kankakee County 113 UA.

Decatur, IL 111  111  (1): Macon County 111 UA.

Danville, IL 82  82  (1): Vermilion County 82 UA.

Cape Girardeau, MO-IL 8  0  (1): Alexander County 8 NQUA.


Indiana

Indianapolis-Carmel-Anderson, IN 1888  1873  (11): Marion County 903 UA; Hamilton County 275 UA; Hendricks County 145 UA; Johnson County 140 UA; Madison County 132 UA; Hancock County 70 UA; Morgan County 69 UA; Boone County 57 UA; Shelby County 44 UC; Putnam County 38 UC; and Brown County 15 NUA.

Chicago-Naperville-Elgin, IL-IN-WI 708  660  (4): Lake County 496 UA; Porter County 164 UA; Jasper County 33 NQUA; and Newton County 14 NUA.

Fort Wayne, IN 416  355  (3): Allen County 355 UA; Whitley County 33 NQUA; and Wells County 28 NQUA.

Louisville/Jefferson County, KY-IN 277  209  (5): Clark County 110 UA; Floyd County 75 UA; Harrison County 39 NQUA; Washington County 28 NQUA; and Scott County 24 UC.

South Bend-Mishawaka, IN-MI 267  267  (1): St. Joseph County 267 UA.

Evansville, IN-KY 265  239  (3): Vanderburgh County 180 UA; Warrick County 60 UA; and Posey County 26 NQUA.

Lafayette-West Lafayette, IN 202  173  (3): Tippecanoe County 173 UA; Carroll County 20 NQUA; and Benton County 9 NUA.

Elkhart-Goshen, IN 198  198  (1): Elkhart County 198 UA.

Terre Haute, IN 172  135  (4): Vigo County 108 UA; Clay County 27 UA; Sullivan County 21 NQUA; and Vermillion County 16 NQUA.

Bloomington, IN 160  138  (2): Monroe County 138 UA; and Owen County 22 NUA.

Muncie, IN 118  118  (1): Delaware County 118 UA.

Michigan City-La Porte, IN 111  111  (1): LaPorte County 111 UA.

Kokomo, IN 83  83  (1): Howard County 83 UA.

Columbus, IN 77  77  (1): Bartholomew County 77 UA.

Cincinnati, OH-KY-IN 64  50  (3): Dearborn County 50 UC; Union County 8 NUA; and Ohio County 6 NUA.


Iowa

Des Moines-West Des Moines, IA 570  543  (5): Polk County 431 UA; Dallas County 66 UA; Warren County 46 UC; Madison County 16 NQUA; and Guthrie County 11 NUA.

Cedar Rapids, IA 258  211  (3): Linn County 211 UA; Benton County 26 NQUA; and Jones County 21 NQUA.

Waterloo-Cedar Falls, IA 168  131  (3): Black Hawk County 131 UA; Bremer County 24 NQUA; and Grundy County 12 NUA.

Davenport-Moline-Rock Island, IA-IL 165  165  (1): Scott County 165 UA.

Iowa City, IA 153  131  (2): Johnson County 131 UA; and Washington County 22 NQUA.

Sioux City, IA-NE-SD 127  102  (2): Woodbury County 102 UA; and Plymouth County 25 NQUA.

Omaha-Council Bluffs, NE-IA 123  93  (3): Pottawattamie County 93 UA; Mills County 15 NQUA; and Harrison County 15 NQUA.

Dubuque, IA 94  94  (1): Dubuque County 94 UA.

Ames, IA 90  90  (1): Story County 90 UA.


Kansas

Kansas City, MO-KS 820  778  (5): Johnson County 544 UA; Wyandotte County 158 UA; Leavenworth County 76 UC; Miami County 33 NQUA; and Linn County 10 NUA.

Wichita, KS 631  599  (5): Sedgwick County 498 UA; Butler County 66 UC; Harvey County 35 UC; Sumner County 24 NQUA; and Kingman County 8 NQUA.

Topeka, KS 234  178  (5): Shawnee County 178 UA; Jefferson County 19 NQC; Osage County 16 NQUA; Jackson County 13 NQUA; and Wabaunsee County 7 NUA.

Lawrence, KS 111  111  (1): Douglas County 111 UA.

Manhattan, KS 93  71  (2): Riley County 71 UA; and Pottawatomie County 22 NQUA.

St. Joseph, MO-KS 8  0  (1): Doniphan County 8 NQC.


Kentucky

Louisville/Jefferson County, KY-IN 959  918  (7): Jefferson County 741 UA; Bullitt County 74 UA; Oldham County 60 UA; Shelby County 42 UC; Spencer County 17 NUA; Henry County 15 NUA; and Trimble County 9 NQC.

Lexington-Fayette, KY 472  472  (6): Fayette County 296 UA; Jessamine County 49 UC; Scott County 47 UC; Clark County 36 UC; Woodford County 25 UC; and Bourbon County 20 UC.

Cincinnati, OH-KY-IN 425  369  (7): Kenton County 160 UA; Boone County 119 UA; Campbell County 90 UA; Grant County 25 NQUA; Pendleton County 15 NUA; Gallatin County 9 NUA; and Bracken County 8 NUA.

Bowling Green, KY 159  114  (4): Warren County 114 UA; Allen County 20 NQUA; Butler County 13 NUA; and Edmonson County 12 NUA.

Elizabethtown-Fort Knox, KY 148  106  (3): Hardin County 106 UA; Meade County 29 NQC; and Larue County 14 NQUA.

Owensboro, KY 115  97  (3): Daviess County 97 UA; McLean County 10 NUA; and Hancock County 9 NQUA.

Clarksville, TN-KY 88  74  (2): Christian County 74 UC; and Trigg County 14 NQUA.

Huntington-Ashland, WV-KY-OH 86  86  (2): Boyd County 50 UA; and Greenup County 37 UA.

Evansville, IN-KY 46  46  (1): Henderson County 46 UA.

[jimrtex]

Maine

Portland-South Portland, ME 514  514  (3): Cumberland County 282 UA; York County 197 UA; and Sagadahoc County 35 UC.

Bangor, ME 154  154  (1): Penobscot County 154 UA.

Lewiston-Auburn, ME 108  108  (1): Androscoggin County 108 UA.


Maryland

Baltimore-Columbia-Towson, MD 2710  2710  (7): Baltimore County 805 UA; Baltimore city 621 UA; Anne Arundel County 538 UA; Howard County 287 UA; Harford County 245 UA; Carroll County 167 UA; and Queen Anne's County 48 UA.

Washington-Arlington-Alexandria, DC-VA-MD-WV 2304  2304  (5): Montgomery County 972 UA; Prince George's County 863 UA; Frederick County 233 UA; Charles County 147 UA; and Calvert County 89 UC.

Salisbury, MD-DE 177  177  (3): Wicomico County 99 UA; Worcester County 51 UC; and Somerset County 26 UC.

Hagerstown-Martinsburg, MD-WV 147  147  (1): Washington County 147 UA.

California-Lexington Park, MD 105  105  (1): St. Mary's County 105 UA.

Philadelphia-Camden-Wilmington, PA-NJ-DE-MD 101  101  (1): Cecil County 101 UA.

Cumberland, MD-WV 75  75  (1): Allegany County 75 UA.


Massachusetts

Boston-Cambridge-Newton, MA-NH 4134  4134  (5): Middlesex County 1503 UA; Essex County 743 UA; Suffolk County 722 UA; Norfolk County 671 UA; and Plymouth County 495 UA.

Worcester, MA-CT 799  799  (1): Worcester County 799 UA.

Springfield, MA 622  622  (2): Hampden County 463 UA; and Hampshire County 158 UA.

Providence-Warwick, RI-MA 548  548  (1): Bristol County 548 UA.

Barnstable Town, MA 216  216  (1): Barnstable County 216 UA.

Pittsfield, MA 131  131  (1): Berkshire County 131 UA.


Michigan

Detroit-Warren-Dearborn, MI 4296  4296  (6): Wayne County 1821 UA; Oakland County 1202 UA; Macomb County 841 UA; Livingston County 181 UA; St. Clair County 163 UA; and Lapeer County 88 UC.

Grand Rapids-Wyoming, MI 989  866  (4): Kent County 603 UA; Ottawa County 264 UA; Montcalm County 63 NQUA; and Barry County 59 NQUA.

Lansing-East Lansing, MI 464  464  (3): Ingham County 281 UA; Eaton County 108 UA; and Clinton County 75 UA.

Flint, MI 426  426  (1): Genesee County 426 UA.

Ann Arbor, MI 345  345  (1): Washtenaw County 345 UA.

Kalamazoo-Portage, MI 327  250  (2): Kalamazoo County 250 UA; and Van Buren County 76 NQUA.

Saginaw, MI 200  200  (1): Saginaw County 200 UA.

Muskegon, MI 172  172  (1): Muskegon County 172 UA.

Jackson, MI 160  160  (1): Jackson County 160 UA.

Niles-Benton Harbor, MI 157  157  (1): Berrien County 157 UA.

Monroe, MI 152  152  (1): Monroe County 152 UA.

Battle Creek, MI 136  136  (1): Calhoun County 136 UA.

Bay City, MI 108  108  (1): Bay County 108 UA.

Midland, MI 84  84  (1): Midland County 84 UA.

South Bend-Mishawaka, IN-MI 52  52  (1): Cass County 52 UA.


Minnesota

Minneapolis-St. Paul-Bloomington, MN-WI 3223  3101  (14): Hennepin County 1152 UA; Ramsey County 509 UA; Dakota County 399 UA; Anoka County 331 UA; Washington County 238 UA; Scott County 130 UA; Wright County 125 UA; Carver County 91 UA; Sherburne County 88 UA; Chisago County 54 NQUA; Isanti County 38 UC; Le Sueur County 28 NQUA; Mille Lacs County 26 NQUA; and Sibley County 15 NUA.

Duluth, MN-WI 236  236  (2): St. Louis County 200 UA; and Carlton County 35 UC.

Rochester, MN 207  144  (4): Olmsted County 144 UA; Wabasha County 22 NQUA; Fillmore County 21 NQUA; and Dodge County 20 NQUA.

St. Cloud, MN 189  189  (2): Stearns County 151 UA; and Benton County 38 UA.

Mankato-North Mankato, MN 97  97  (2): Blue Earth County 64 UA; and Nicollet County 33 UA.

Fargo, ND-MN 59  59  (1): Clay County 59 UA.

Grand Forks, ND-MN 32  32  (1): Polk County 32 UA.

La Crosse-Onalaska, WI-MN 19  19  (1): Houston County 19 UA.


Missouri

St. Louis, MO-IL 2084  2052  (7): St. Louis County 999 UA; St. Charles County 360 UA; St. Louis city 319 UA; Jefferson County 219 UA; Franklin County 101 UC; Lincoln County 53 UC; and Warren County 33 NQUA.

Kansas City, MO-KS 1189  1085  (9): Jackson County 674 UA; Clay County 222 UA; Cass County 99 UA; Platte County 89 UA; Lafayette County 33 NQUA; Ray County 23 NQUA; Clinton County 21 NQUA; Bates County 17 NQUA; and Caldwell County 9 NUA.

Springfield, MO 437  353  (5): Greene County 275 UA; Christian County 77 UA; Webster County 36 NQUA; Polk County 31 NQUA; and Dallas County 17 NQUA.

Joplin, MO 176  176  (2): Jasper County 117 UA; and Newton County 58 UC.

Columbia, MO 163  163  (1): Boone County 163 UA.

Jefferson City, MO 150  120  (4): Cole County 76 UA; Callaway County 44 UC; Moniteau County 16 NQUA; and Osage County 14 NUA.

St. Joseph, MO-KS 119  89  (3): Buchanan County 89 UA; Andrew County 17 NQUA; and DeKalb County 13 NQUA.

Cape Girardeau, MO-IL 88  76  (2): Cape Girardeau County 76 UA; and Bollinger County 12 NUA.

Fayetteville-Springdale-Rogers, AR-MO 23  0  (1): McDonald County 23 NQC.


Nebraska

Omaha-Council Bluffs, NE-IA 742  676  (5): Douglas County 517 UA; Sarpy County 159 UA; Cass County 25 NQUA; Saunders County 21 NQUA; and Washington County 20 NQUA.

Lincoln, NE 302  285  (2): Lancaster County 285 UA; and Seward County 17 NQUA.

Grand Island, NE 82  59  (4): Hall County 59 UA; Hamilton County 9 NQUA; Merrick County 8 NQUA; and Howard County 6 NUA.

Sioux City, IA-NE-SD 27  21  (2): Dakota County 21 UA; and Dixon County 6 NUA.


New Hampshire

Boston-Cambridge-Newton, MA-NH 418  418  (2): Rockingham County 295 UA; and Strafford County 123 UA.

Manchester-Nashua, NH 401  401  (1): Hillsborough County 401 UA.


New Jersey

New York-Newark-Jersey City, NY-NJ-PA 6471  6471  (12): Bergen County 905 UA; Middlesex County 810 UA; Essex County 784 UA; Hudson County 634 UA; Monmouth County 630 UA; Ocean County 577 UA; Union County 536 UA; Passaic County 501 UA; Morris County 492 UA; Somerset County 323 UA; Sussex County 149 UA; and Hunterdon County 128 UA.

Philadelphia-Camden-Wilmington, PA-NJ-DE-MD 1317  1317  (4): Camden County 514 UA; Burlington County 449 UA; Gloucester County 288 UA; and Salem County 66 UA.

Trenton, NJ 367  367  (1): Mercer County 367 UA.

Atlantic City-Hammonton, NJ 275  275  (1): Atlantic County 275 UA.

Vineland-Bridgeton, NJ 157  157  (1): Cumberland County 157 UA.

Allentown-Bethlehem-Easton, PA-NJ 109  109  (1): Warren County 109 UA.

Ocean City, NJ 97  97  (1): Cape May County 97 UA.


New York

New York-Newark-Jersey City, NY-NJ-PA 13039  13039  (12): Kings County 2505 UA; Queens County 2231 UA; New York County 1586 UA; Suffolk County 1493 UA; Bronx County 1385 UA; Nassau County 1340 UA; Westchester County 949 UA; Richmond County 469 UA; Orange County 373 UA; Rockland County 312 UA; Dutchess County 297 UA; and Putnam County 100 UA.

Buffalo-Cheektowaga-Niagara Falls, NY 1136  1136  (2): Erie County 919 UA; and Niagara County 216 UA.

Rochester, NY 1080  946  (6): Monroe County 744 UA; Ontario County 108 UA; Wayne County 94 UC; Livingston County 65 NQUA; Orleans County 43 NQUA; and Yates County 25 NQUA.

Albany-Schenectady-Troy, NY 871  838  (5): Albany County 304 UA; Saratoga County 220 UA; Rensselaer County 159 UA; Schenectady County 155 UA; and Schoharie County 33 NQUA.

Syracuse, NY 663  663  (3): Onondaga County 467 UA; Oswego County 122 UC; and Madison County 73 UC.

Utica-Rome, NY 299  299  (2): Oneida County 235 UA; and Herkimer County 65 UC.

Binghamton, NY 252  252  (2): Broome County 201 UA; and Tioga County 51 UA.

Kingston, NY 182  182  (1): Ulster County 182 UA.

Glens Falls, NY 129  129  (2): Warren County 66 UA; and Washington County 63 UA.

Watertown-Fort Drum, NY 116  116  (1): Jefferson County 116 UA.

Ithaca, NY 102  102  (1): Tompkins County 102 UA.

Elmira, NY 89  89  (1): Chemung County 89 UA.

[jimrtex]

North Dakota

Fargo, ND-MN 150  150  (1): Cass County 150 UA.

Bismarck, ND 115  109  (4): Burleigh County 81 UA; Morton County 27 UA; Sioux County 4 NUA; and Oliver County 2 NUA.

Grand Forks, ND-MN 67  67  (1): Grand Forks County 67 UA.


Ohio

Cleveland-Elyria, OH 2077  2077  (5): Cuyahoga County 1280 UA; Lorain County 301 UA; Lake County 230 UA; Medina County 172 UA; and Geauga County 93 UA.

Columbus, OH 1902  1802  (10): Franklin County 1163 UA; Delaware County 174 UA; Licking County 166 UA; Fairfield County 146 UA; Pickaway County 56 UC; Union County 52 UC; Madison County 43 UC; Perry County 36 NQUA; Morrow County 35 NQUA; and Hocking County 29 NQUA.

Cincinnati, OH-KY-IN 1625  1581  (5): Hamilton County 802 UA; Butler County 368 UA; Warren County 213 UA; Clermont County 197 UA; and Brown County 45 NQUA.

Dayton, OH 799  799  (3): Montgomery County 535 UA; Greene County 162 UA; and Miami County 103 UA.

Akron, OH 703  703  (2): Summit County 542 UA; and Portage County 161 UA.

Toledo, OH 610  567  (3): Lucas County 442 UA; Wood County 125 UA; and Fulton County 43 NQUA.

Youngstown-Warren-Boardman, OH-PA 449  449  (2): Mahoning County 239 UA; and Trumbull County 210 UA.

Canton-Massillon, OH 404  376  (2): Stark County 376 UA; and Carroll County 29 NQUA.

Springfield, OH 138  138  (1): Clark County 138 UA.

Mansfield, OH 124  124  (1): Richland County 124 UA.

Lima, OH 106  106  (1): Allen County 106 UA.

Wheeling, WV-OH 70  70  (1): Belmont County 70 UA.

Weirton-Steubenville, WV-OH 70  70  (1): Jefferson County 70 UA.

Huntington-Ashland, WV-KY-OH 62  62  (1): Lawrence County 62 UA.


Pennsylvania

Philadelphia-Camden-Wilmington, PA-NJ-DE-MD 4009  4009  (5): Philadelphia County 1526 UA; Montgomery County 800 UA; Bucks County 625 UA; Delaware County 559 UA; and Chester County 499 UA.

Pittsburgh, PA 2356  2356  (7): Allegheny County 1223 UA; Westmoreland County 365 UA; Washington County 208 UA; Butler County 184 UA; Beaver County 171 UA; Fayette County 137 UA; and Armstrong County 69 UC.

Allentown-Bethlehem-Easton, PA-NJ 712  712  (3): Lehigh County 349 UA; Northampton County 298 UA; and Carbon County 65 UA.

Scranton--Wilkes-Barre--Hazleton, PA 564  535  (3): Luzerne County 321 UA; Lackawanna County 214 UA; and Wyoming County 28 NQUA.

Harrisburg-Carlisle, PA 549  504  (3): Dauphin County 268 UA; Cumberland County 235 UA; and Perry County 46 NQUA.

Lancaster, PA 519  519  (1): Lancaster County 519 UA.

York-Hanover, PA 435  435  (1): York County 435 UA.

Reading, PA 411  411  (1): Berks County 411 UA.

Erie, PA 281  281  (1): Erie County 281 UA.

East Stroudsburg, PA 170  170  (1): Monroe County 170 UA.

State College, PA 154  154  (1): Centre County 154 UA.

Chambersburg-Waynesboro, PA 150  150  (1): Franklin County 150 UA.

Johnstown, PA 144  144  (1): Cambria County 144 UA.

Lebanon, PA 134  134  (1): Lebanon County 134 UA.

Altoona, PA 127  127  (1): Blair County 127 UA.

Youngstown-Warren-Boardman, OH-PA 117  117  (1): Mercer County 117 UA.

Williamsport, PA 116  116  (1): Lycoming County 116 UA.

Gettysburg, PA 101  101  (1): Adams County 101 UA.

Bloomsburg-Berwick, PA 86  86  (2): Columbia County 67 UA; and Montour County 18 UA.

New York-Newark-Jersey City, NY-NJ-PA 57  0  (1): Pike County 57 NQUA.


Rhode Island

Providence-Warwick, RI-MA 1053  1053  (5): Providence County 627 UA; Kent County 166 UA; Washington County 127 UA; Newport County 83 UA; and Bristol County 50 UA.


South Dakota

Sioux Falls, SD 228  214  (4): Minnehaha County 169 UA; Lincoln County 45 UA; Turner County 8 NUA; and McCook County 6 NUA.

Rapid City, SD 135  101  (3): Pennington County 101 UA; Meade County 25 NQUA; and Custer County 8 NUA.

Sioux City, IA-NE-SD 14  14  (1): Union County 14 UA.


Vermont

Burlington-South Burlington, VT 211  157  (3): Chittenden County 157 UA; Franklin County 48 NQUA; and Grand Isle County 7 NUA.


Virginia

Note: Independent Cities were treated as part of their original county.  When a independent city has absorbed the entire county (eg Hampton, Newport News, Virginia Beach), the independent city is shown rather the former county name.  Norfolk and Portsmouth were created from Norfolk County, and then the remnant of the county became Chesapeake.  They are all listed under Norfolk County.

Washington-Arlington-Alexandria, DC-VA-MD-WV 2677  2656  (11): Fairfax County (Fairfax city, Falls Church city) 1117 UA; Prince William County (Manassas city, Manassas Park city) 454 UA; Arlington County (Alexandria city) 348 UA; Loudoun County 312 UA; Spotsylvania County (Fredericksburg city) 147 UA; Stafford County 129 UA; Fauquier County 65 UA; Culpeper County 47 UC; Warren County 38 UC; Clarke County 14 NQUA; and Rappahannock County 7 NUA.

Virginia Beach-Norfolk-Newport News, VA-NC 1641  1632  (10): Norfolk county (Chesapeake city, Norfolk city, Portsmouth city) 561 UA; Virginia Beach city 438 UA; Newport News city 181 UA; Hampton city 137 UA; Suffolk city 85 UA; James City County (Williamsburg city) 81 UA; York County (Poquoson city) 78 UA; Gloucester County 37 UA; Isle of Wight County 35 UC; and Mathews County 9 NUA.

Richmond, VA 1208  1063  (13): Henrico County (Richmond city) 511 UA; Chesterfield County (Colonial Heights city) 334 UA; Hanover County 100 UA; Dinwiddie County (Petersburg city) 60 UA; Prince George County (Hopewell city) 58 UA; Caroline County 29 NQUA; Powhatan County 28 NQC; Goochland County 22 NQC; New Kent County 18 NUA; King William County 16 NQUA; Amelia County 13 NUA; Sussex County 12 NUA; and Charles City County 7 NUA.

Roanoke, VA 309  247  (4): Roanoke County (Roanoke city, Salem city) 214 UA; Franklin County 56 NQUA; Botetourt County 33 UA; and Craig County 5 NUA.

Lynchburg, VA 253  238  (4): Campbell County (Lynchburg city) 130 UA; Bedford County (Bedford city) 75 UA; Amherst County 32 UA; and Appomattox County 15 NUA.

Charlottesville, VA 219  142  (5): Albemarle County (Charlottesville city) 142 UA; Fluvanna County 26 NQUA; Greene County 18 NQUA; Buckingham County 17 NUA; and Nelson County 15 NUA.

Blacksburg-Christiansburg-Radford, VA 178  146  (4): Montgomery County (Radford city) 111 UA; Pulaski County 35 UC; Giles County 17 NQUA; and Floyd County 15 NUA.

Harrisonburg, VA 125  125  (1): Rockingham County (Harrisonburg city) 125 UA.

Staunton-Waynesboro, VA 119  119  (1): Augusta County (Staunton city, Waynesboro city) 119 UA.

Winchester, VA-WV 105  105  (1): Frederick County (Winchester city) 105 UA.

Kingsport-Bristol-Bristol, TN-VA 96  73  (2): Washington County (Bristol city) 73 UA; and Scott County 23 NQC.


West Virginia

Charleston, WV 227  193  (3): Kanawha County 193 UA; Boone County 25 NQUA; and Clay County 9 NUA.

Huntington-Ashland, WV-KY-OH 216  194  (4): Cabell County 96 UA; Putnam County 55 UA; Wayne County 42 UA; and Lincoln County 22 NUA.

Morgantown, WV 130  96  (2): Monongalia County 96 UA; and Preston County 34 NQUA.

Beckley, WV 125  125  (2): Raleigh County 79 UA; and Fayette County 46 UA.

Hagerstown-Martinsburg, MD-WV 104  104  (1): Berkeley County 104 UA.

Parkersburg-Vienna, WV 93  87  (2): Wood County 87 UA; and Wirt County 6 NUA.

Wheeling, WV-OH 78  78  (2): Ohio County 44 UA; and Marshall County 33 UA.

Weirton-Steubenville, WV-OH 55  55  (2): Hancock County 31 UA; and Brooke County 24 UA.

Washington-Arlington-Alexandria, DC-VA-MD-WV 53  53  (1): Jefferson County 53 UC.

Cumberland, MD-WV 28  0  (1): Mineral County 28 NQUA.

Winchester, VA-WV 24  0  (1): Hampshire County 24 NUA.


Wisconsin

Milwaukee-Waukesha-West Allis, WI 1556  1556  (4): Milwaukee County 948 UA; Waukesha County 390 UA; Washington County 132 UA; and Ozaukee County 86 UA.

Madison, WI 605  582  (4): Dane County 488 UA; Columbia County 57 UC; Green County 37 UC; and Iowa County 24 NQUA.

Green Bay, WI 306  248  (3): Brown County 248 UA; Oconto County 38 NQUA; and Kewaunee County 21 NQUA.

Appleton, WI 226  226  (2): Outagamie County 177 UA; and Calumet County 49 UA.

Racine, WI 195  195  (1): Racine County 195 UA.

Oshkosh-Neenah, WI 167  167  (1): Winnebago County 167 UA.

Chicago-Naperville-Elgin, IL-IN-WI 166  166  (1): Kenosha County 166 UA.

Eau Claire, WI 161  161  (2): Eau Claire County 99 UA; and Chippewa County 62 UA.

Janesville-Beloit, WI 160  160  (1): Rock County 160 UA.

Wausau, WI 134  134  (1): Marathon County 134 UA.

Minneapolis-St. Paul-Bloomington, MN-WI 125  125  (2): St. Croix County 84 UC; and Pierce County 41 UC.

Sheboygan, WI 116  116  (1): Sheboygan County 116 UA.

La Crosse-Onalaska, WI-MN 115  115  (1): La Crosse County 115 UA.

Fond du Lac, WI 102  102  (1): Fond du Lac County 102 UA.

Duluth, MN-WI 44  44  (1): Douglas County 44 UA.

[Torie]

"By restricting transfers of population between regions they restrict certain forms of gerrymandering along the same philosophy as the MI rules."

I don't think that you have made the case above. Heck once you get down to the minimum number of chops you think will work while containing erosity, you can reverse engineer the regions. Please give me an example where when limiting erosity and chops (with micro-chops favored as compared to Micro chops, either having no penalty or half a penalty or whatever),  one can still play gerrymandering games, while by doing the regional strait jacket right out of the box, those games are shut down.

In Michigan, erosity was not a negative, and there was no micro-chop factor (the former being far most important, particularly intra-county, where one cheery picked localities, particularly in Oakland County, and to a lessor extent in Macomb, Ingham, Saginaw, and Washtenaw). Erosity measures shut down intra county chop games because even if not that visible on a statewide map, it is always possible to improve the erosity score along the edges of a CD by doing the chop based on geography, not politics.

And this all assumes of course, that there is no mechanism for either party to veto some of the high scoring maps (which surely would kill off most of the gerrymandering games that favor one party right out of the box in any event). Moreover, the computer will spit out the high scoring maps one presumes, just as one presumes that computers are non-partisan.

[muon2]

"By restricting transfers of population between regions they restrict certain forms of gerrymandering along the same philosophy as the MI rules."

I don't think that you have made the case above. Heck once you get down to the minimum number of chops you think will work while containing erosity, you can reverse engineer the regions. Please give me an example where when limiting erosity and chops (with micro-chops favored as compared to Micro chops, either having no penalty or half a penalty or whatever),  one can still play gerrymandering games, while by doing the regional strait jacket right out of the box, those games are shut down.

In Michigan, erosity was not a negative, and there was no micro-chop factor (the former being far most important, particularly intra-county, where one cheery picked localities, particularly in Oakland County, and to a lessor extent in Macomb, Ingham, Saginaw, and Washtenaw). Erosity measures shut down intra county chop games because even if not that visible on a statewide map, it is always possible to improve the erosity score along the edges of a CD by doing the chop based on geography, not politics.

And this all assumes of course, that there is no mechanism for either party to veto some of the high scoring maps (which surely would kill off most of the gerrymandering games that favor one party right out of the box in any event). Moreover, the computer will spit out the high scoring maps one presumes, just as one presumes that computers are non-partisan.

You are correct that I could reverse engineer a map, but as you may recollect from our OH discussion I reverse engineered your design to regions. But then I started from those regions and made a stronger map. Presumably in a contest situation your plan would suffer compared to those that started from regions just as I could improve on your one example. I know it helped me greatly in the OH competitions. A successful strategy is worth considering as part of the process. Remaining erosity concerns can be handled by rules limiting population shifts which was my reference to the MI rules.

Your last sentence is actually the catch and for me makes a unique case for regions. Given lots of data and some constraints, computers aren't very good at drawing maps. Sure you can put in things like the splitline algorithm, but that gives results that are fairly unacceptable to the public. I've had a chance to look at a number of redistricting algorithms and all have some similar problems. The computational task is in a category called NP-hard and that means the best result will probably be an approximation to the ideal.

One of the most successful tools to make a good approximation to an NP-hard problem is to use an iterative algorithm based on moving from coarse grain data to fine grain data. The existence of counties and county subdivisions provide a natural set of data granularity tiers. However, when I looked at counties alone as a coarse data set the result tended toward erose results dominated by county preservation, or it generated plans with unusual chops like my carve up of Lansing. Forcing compact regions overcomes this problem, so the question becomes one of identifying the standards for regions. This convergence on urban county clusters seems to be the right track for a standard.

I wish that DRA had a whole county coloring option like most redistricting software. If it did I posit that you would find regions to actually be quite simple to operate. Unfortunately, there's nothing in DRA to guide a county or regional strategy so I'm forced to build my own spreadsheet as a data source to find acceptable regions.

[Torie]

You have explained well Mike how regions may be a useful tool for finding the best map, but I still find nothing as to why it needs to be some kind of rule once regions are defined by someone. That is the point.

[muon2]

You have explained well Mike how regions may be a useful tool for finding the best map, but I still find nothing as to why it needs to be some kind of rule once regions are defined by someone. That is the point.

There are two areas where I see that they might move beyond mere tool. One is the subject of population shifts in the form of chops. You would like to create chops to reduce erosity, I would like to restrict that to address the most erose districts. Limiting the population shifts between regions provides a clear demarcation of the number of permitted shifts. Without regions there's really little to restrict shifts. jimrtex's population flow maps are good examples of what I mean.

The second area I just introduced is for VRA districts. Federal law for VRA districts will trump any other redistricting rule. However, by constraining a VRA district to a region with applicable rules to the extent practicable can reduce the ability of a mapmaker to say that a given district was required by the VRA, when in fact it was a political choice that violates a host of rules. Rules for VRA districts can be done without regions, but testing them to insure that the VRA wasn't merely an excuse is harder.

One final area is a procedural point that is favored by jimrtex. If I understand the type of method he'd like to see, then once there are regions defined by the first step in the process, public input and focus can be directed at the separate regions on a subsequent step. For a commission this seems akin to the "grid" step used in AZ prior to the actual districting.

[Torie]

Still not persuasive, and you did not address the reverse engineering bit. You have not made the case that these further restraints are necessary to avoid Gerrymandering games either. And it's too complicated. I might note that the AZ commission, after approving the grid map, essentially threw it in the dumpster. I would never agree to any regional map myself unless and until I saw what actual maps it allowed. It would be nutter to do otherwise.

Given that the overall conclusion you make is so amorphous, the only way to persuade is with real world examples, as to how using regions as a rule rather than a tool avoids mischief not otherwise constrained by chop and erosity scores. I suspect the odds are better than even, that you can't produce such an example. But I have an open mind to reconsider if such examples are out there. Your Michigan example just didn't cut it - that was apples to oranges. Think about this all as writing a legal brief, trying to persuade a judge, and making it comprehensible to a judge. That might prove more efficacious.

Just a suggestion of course. Another option is to put me on "ignore," as I kvetch from the bleachers.

[muon2]

Still not persuasive, and you did not address the reverse engineering bit. You have not made the case that these further restraints are necessary to avoid Gerrymandering games either. And it's too complicated. I might note that the AZ commission, after approving the grid map, essentially threw it in the dumpster. I would never agree to any regional map myself unless and until I saw what actual maps it allowed. It would be nutter to do otherwise.

Given that the overall conclusion you make is so amorphous, the only way to persuade is with real world examples, as to how using regions as a rule rather than a tool avoids mischief not otherwise constrained by chop and erosity scores. I suspect the odds are better than even, that you can't produce such an example. But I have an open mind to reconsider if such examples are out there. Your Michigan example just didn't cut it - that was apples to oranges. Think about this all as writing a legal brief, trying to persuade a judge, and making it comprehensible to a judge. That might prove more efficacious.

Just a suggestion of course. Another option is to put me on "ignore," as I kvetch from the bleachers.

I think we view "rule" in different ways. To me this is like a student showing their work on a problem, it's not enough to just write down the answer. Of course I am by profession suspicious of an answer that can't be supported by steps, and in redistricting I am suspicious of the mapper's motive. And I claim its not just being a stickler for the formality, it seems to me that I've been able to improve on your offerings by reversing them back to regions then going forward to districts, and I would offer those as real world examples. That alone suggests that those who don't proceed from regions will lose out to those who do. Why not give everyone that advantage up front and require that they show the step?

That said, I'm still nervous about VRA districts that could or should creep outside of a single urban county cluster. I picked on LA precisely because it would have that feature. I think NC would be interesting in that regard as well. What constraint(s) would you impose to prevent gerrymandering VRA districts or their neighbors? I submit that erosity by itself probably won't do given some of the shapes that have survived legal challenge.

[jimrtex]

Still stuck on regions as some sort of rule rather than a convenience tool. I still don't understand that, and don't agree with that. Absent the regional rule (whatever the regions are for LA), would that make Muon2's map any less erose? Or by "regions" here do we just mean the multi county urban areas?

The black CD in the first map is a mouthful, and I think probably a map killer inasmuch as the second map suggests it is possible to draw something far less offensive and meet the VRA requirements. And do the VRA requirements really dictate that a NO CD suck up the black precincts of Baton Rouge?  Are NO and BR one community of interest?  I tend to doubt it. Heck the each have their own urban cluster.  I tend to doubt the first map's black CD is required by the VRA either as one community of interest.

In a state like Iowa, there are 4 regions, each with the population of one congressional district.

In another state there might be one region per congressional district, but they may have greater deviation.   I see the process as sequential, with first regions defined, and then the regions refined, either by chops or division of multi-district regions.

Regions made up of whole counties can be proposed by ordinary citizens, they are straightforward to evaluate for equality and erosity, and they may be evaluated and rated by ordinary citizens.

While counties are typically the best choice for composing districts, this is not always possible.   Some counties are too large, or it may be impossible to combine them in a manner that permits whole-county districts.  

In other cases, the only combinations possible require a district extending outward from the metropolitan area.  It is quite easy in Minnesota to have a district running from Anoka to Superior, Wright to Grand Forks, Carver to Sioux Falls, and Dakota to Lacrosse, with the only impediment the state lines.  

Requiring districts to conform to urban county clusters retains the overall simplicity of districts based on counties, but permits districts in urban areas to have a relaxed basis.  There would still be restrictions based on minimum number of county cuts, and double spanning.

[Torie]

I really don't know why this is so difficult. You again make a good case that regions are a useful tool in coming up with the best maps, but not that it needs to be a rule. And you can be suspicious of a student's motives all you want, but if it is clear that there is no cheating, and he gets the highest score on the test, what is the problem?  In the end, you evaluate the work product, not how it was fabricated, for scoring purposes.

Anyway it is clear to me that maybe slapping county populations on an excel spreadsheet, organized by contiguity, and moving those numbers from column to column to find where the micro-chops might be, then draw your CD's, and then reverse engineer what the regions are that you hewed to, is the only way to make you happy while basically ignoring your regional rule at the same time. I guess that is a deal - you have your rule, and I have a way to finesse the rule, and make it but a tool, so that we are both happy.

I am confused about the contretemps about VRA districts. If you think a community of interest is there that requires a minority majority CD per judicial precedent, you draw the least erose with the fewest chops generating CD that is possible to contain the "damage" as it were. I think what you might be fussing about is that going about that exercise might make a hash of your ex ante regions. Is that it?

[jimrtex]

Edit: Show all of Logan, UT-ID metropolitan area.




Arizona

Phoenix-Mesa-Scottsdale, AZ 4193  4193  (2): Maricopa County 3817 UA; and Pinal County 376 UA.

Tucson, AZ 980  980  (1): Pima County 980 UA.

Prescott, AZ 211  211  (1): Yavapai County 211 UA.

Lake Havasu City-Kingman, AZ 200  200  (1): Mohave County 200 UA.

Yuma, AZ 196  196  (1): Yuma County 196 UA.

Flagstaff, AZ 134  134  (1): Coconino County 134 UA.

Sierra Vista-Douglas, AZ 131  131  (1): Cochise County 131 UA.


California

Los Angeles-Long Beach-Anaheim, CA 12829  12829  (2): Los Angeles County 9819 UA; and Orange County 3010 UA.

San Francisco-Oakland-Hayward, CA 4335  4335  (5): Alameda County 1510 UA; Contra Costa County 1049 UA; San Francisco County 805 UA; San Mateo County 718 UA; and Marin County 252 UA.

Riverside-San Bernardino-Ontario, CA 4225  4225  (2): Riverside County 2190 UA; and San Bernardino County 2035 UA.

San Diego-Carlsbad, CA 3095  3095  (1): San Diego County 3095 UA.

Sacramento--Roseville--Arden-Arcade, CA 2149  2149  (4): Sacramento County 1419 UA; Placer County 348 UA; Yolo County 201 UA; and El Dorado County 181 UA.

San Jose-Sunnyvale-Santa Clara, CA 1837  1837  (2): Santa Clara County 1782 UA; and San Benito County 55 UC.

Fresno, CA 930  930  (1): Fresno County 930 UA.

Bakersfield, CA 840  840  (1): Kern County 840 UA.

Oxnard-Thousand Oaks-Ventura, CA 823  823  (1): Ventura County 823 UA.

Stockton-Lodi, CA 685  685  (1): San Joaquin County 685 UA.

Modesto, CA 514  514  (1): Stanislaus County 514 UA.

Santa Rosa, CA 484  484  (1): Sonoma County 484 UA.

Visalia-Porterville, CA 442  442  (1): Tulare County 442 UA.

Santa Maria-Santa Barbara, CA 424  424  (1): Santa Barbara County 424 UA.

Salinas, CA 415  415  (1): Monterey County 415 UA.

Vallejo-Fairfield, CA 413  413  (1): Solano County 413 UA.

San Luis Obispo-Paso Robles-Arroyo Grande, CA 270  270  (1): San Luis Obispo County 270 UA.

Santa Cruz-Watsonville, CA 262  262  (1): Santa Cruz County 262 UA.

Merced, CA 256  256  (1): Merced County 256 UA.

Chico, CA 220  220  (1): Butte County 220 UA.

Redding, CA 177  177  (1): Shasta County 177 UA.

El Centro, CA 175  175  (1): Imperial County 175 UA.

Yuba City, CA 167  167  (2): Sutter County 95 UA; and Yuba County 72 UA.

Hanford-Corcoran, CA 153  153  (1): Kings County 153 UA.

Madera, CA 151  151  (1): Madera County 151 UA.

Napa, CA 136  136  (1): Napa County 136 UA.


Colorado

Denver-Aurora-Lakewood, CO 2543  2490  (10): Denver County 600 UA; Arapahoe County 572 UA; Jefferson County 535 UA; Adams County 442 UA; Douglas County 285 UA; Broomfield County 56 UA; Elbert County 23 NUA; Park County 16 NUA; Clear Creek County 9 NUA; and Gilpin County 5 NUA.

Colorado Springs, CO 646  646  (2): El Paso County 622 UA; and Teller County 23 UC.

Fort Collins, CO 300  300  (1): Larimer County 300 UA.

Boulder, CO 295  295  (1): Boulder County 295 UA.

Greeley, CO 253  253  (1): Weld County 253 UA.

Pueblo, CO 159  159  (1): Pueblo County 159 UA.

Grand Junction, CO 147  147  (1): Mesa County 147 UA.


Hawaii

Not shown on map.

Urban Honolulu, HI 953  953  (1): Honolulu County 953 UA.

Kahului-Wailuku-Lahaina, HI 155  155  (2): Maui County 155 UA; and Kalawao County  NUA.


Nevada

Las Vegas-Henderson-Paradise, NV 1951  1951  (1): Clark County 1951 UA.

Reno, NV 425  421  (2): Washoe County 421 UA; and Storey County 4 NQC.

Carson City, NV 55  55  (1): Carson City 55 UA.


New Mexico

Albuquerque, NM 887  871  (4): Bernalillo County 663 UA; Sandoval County 132 UA; Valencia County 77 UA; and Torrance County 16 NQUA.

Las Cruces, NM 209  209  (1): Doña Ana County 209 UA.

Santa Fe, NM 144  144  (1): Santa Fe County 144 UA.


Utah

Salt Lake City, UT 1088  1088  (2): Salt Lake County 1030 UA; and Tooele County 58 UC.

Ogden-Clearfield, UT 597  588  (4): Davis County 306 UA; Weber County 231 UA; Box Elder County 50 UA; and Morgan County 9 NQUA.

Provo-Orem, UT 527  517  (2): Utah County 517 UA; and Juab County 10 NQUA.

St. George, UT 138  138  (1): Washington County 138 UA.

Logan, UT-ID 113  113  (1): Cache County 113 UA.

Farmington, NM 130  130  (1): San Juan County 130 UA.

[jimrtex]

Alaska

Not shown on map.

Anchorage, AK 381  381  (2): Anchorage Municipality 292 UA; and Matanuska-Susitna Borough 89 UC.

Fairbanks, AK 98  98  (1): Fairbanks North Star Borough 98 UA.


Idaho

Boise City, ID 617  581  (5): Ada County 392 UA; Canyon County 189 UA; Gem County 17 NQUA; Owyhee County 12 NQUA; and Boise County 7 NUA.

Coeur d'Alene, ID 138  138  (1): Kootenai County 138 UA.

Idaho Falls, ID 133  104  (3): Bonneville County 104 UA; Jefferson County 26 NQUA; and Butte County 3 NUA.

Pocatello, ID 83  83  (1): Bannock County 83 UA.

Lewiston, ID-WA 39  39  (1): Nez Perce County 39 UA.

Logan, UT-ID 13  0  (1): Franklin County 13 NQUA.


Montana

Billings, MT 159  148  (3): Yellowstone County 148 UA; Carbon County 10 NUA; and Golden Valley County 1 NUA.

Missoula, MT 109  109  (1): Missoula County 109 UA.

Great Falls, MT 81  81  (1): Cascade County 81 UA.


Oregon

Portland-Vancouver-Hillsboro, OR-WA 1790  1790  (5): Multnomah County 735 UA; Washington County 530 UA; Clackamas County 376 UA; Yamhill County 99 UC; and Columbia County 49 UC.

Salem, OR 391  391  (2): Marion County 315 UA; and Polk County 75 UA.

Eugene, OR 352  352  (1): Lane County 352 UA.

Medford, OR 203  203  (1): Jackson County 203 UA.

Bend-Redmond, OR 158  158  (1): Deschutes County 158 UA.

Albany, OR 117  117  (1): Linn County 117 UA.

Corvallis, OR 86  86  (1): Benton County 86 UA.

Grants Pass, OR 83  83  (1): Josephine County 83 UA.


Washington

Seattle-Tacoma-Bellevue, WA 3440  3440  (3): King County 1931 UA; Pierce County 795 UA; and Snohomish County 713 UA.

Spokane-Spokane Valley, WA 528  471  (3): Spokane County 471 UA; Stevens County 44 NQUA; and Pend Oreille County 13 NQUA.

Portland-Vancouver-Hillsboro, OR-WA 436  425  (2): Clark County 425 UA; and Skamania County 11 NUA.

Kennewick-Richland, WA 253  253  (2): Benton County 175 UA; and Franklin County 78 UA.

Olympia-Tumwater, WA 252  252  (1): Thurston County 252 UA.

Bremerton-Silverdale, WA 251  251  (1): Kitsap County 251 UA.

Yakima, WA 243  243  (1): Yakima County 243 UA.

Bellingham, WA 201  201  (1): Whatcom County 201 UA.

Mount Vernon-Anacortes, WA 117  117  (1): Skagit County 117 UA.

Wenatchee, WA 111  111  (2): Chelan County 72 UA; and Douglas County 38 UA.

Longview, WA 102  102  (1): Cowlitz County 102 UA.

Walla Walla, WA 63  59  (2): Walla Walla County 59 UA; and Columbia County 4 NQUA.

Lewiston, ID-WA 22  22  (1): Asotin County 22 UA.


Wyoming

Cheyenne, WY 92  92  (1): Laramie County 92 UA.

Casper, WY 75  75  (1): Natrona County 75 UA.

[jimrtex]

I really don't know why this is so difficult. You again make a good case that regions are a useful tool in coming up with the best maps, but not that it needs to be a rule. And you can be suspicious of a student's motives all you want, but if it is clear that there is no cheating, and he gets the highest score on the test, what is the problem?  In the end, you evaluate the work product, not how it was fabricated, for scoring purposes.

Anyway it is clear to me that maybe slapping county populations on an excel spreadsheet, organized by contiguity, and moving those numbers from column to column to find where the micro-chops might be, then draw your CD's, and then reverse engineer what the regions are that you hewed to, is the only way to make you happy while basically ignoring your regional rule at the same time. I guess that is a deal - you have your rule, and I have a way to finesse the rule, and make it but a tool, so that we are both happy.

I am confused about the contretemps about VRA districts. If you think a community of interest is there that requires a minority majority CD per judicial precedent, you draw the least erose with the fewest chop generating CD that is possible to contain the "damage" as it were. I think what you might be fussing about is that going about that exercise might make a hash of your ex ante regions. Is that it?

You're assuming that the final map is the only product that will be evaluated.  There would be no harm if you were to devise a final map, and then derive the regions, and present that.  Others might present a different regional map, or improve yours, and the final map might be different, even if your regional map was chosen.

[muon2]

I think you are adjusting your regions to fit your VRA districts, and then you are cheating on the rules to draw the VRA district, by placing the whole counties but connecting squiggly lines in the 3 larger counties.

It first seems reasonable to identify any sub-county regions that are used for districts (in advance).   In Minnesota, counties and townships can be used, with perhaps an adjustment to place all of city in a single county (eg St. Anthony would be treated for redistricting purposes as if it were entirely in Hennepin County).   Larger cities should also be divided into neighborhoods.  That we delineate neighborhoods in Minneapolis does not imply that Minneapolis will be divided, any more than that Cottonwood County has townships.  The neighborhoods in Minneapolis could also be used for legislative districts.   It is not desirable that these neighborhoods have equal population.  Communities of interest do not come in quantum units.  And if they are all the same size, it is hard to combine to get a particular population.  So city council districts should not be used, which also avoids the possibility that they were drawn for political purposes.

In Michigan much the same rules can be used, though in that case, Detroit must be divided.  

Ohio probably needs more work, since cities extensively cross county boundaries, some cities incorporate townships, and others do partially.

Louisiana doesn't have defined townships. but the Public Land Survey System exists in most of the state, and the prime meridians and townships are defined, so that townships could be constructed.

These could then be refined.  The first step would to conform them to county boundaries.  Then  partial townships (eg less than 10 sections), or with few residents (eg less than 200) could be merged.  Boundaries could be adjusted to follow rivers or ridge lines.  Road connectivity could be considered (eg if education were being provided on a township basis, would an extreme bus route be needed.   Towns above a certain size could be elevated to township status, or perhaps absorb the surrounding area.

A VRA district could be constructed from these smaller redistricting units.  They need not be contiguous.   Then simply remove that territory and population from the state, as if the areas were depopulated, and then redistrict the remainder of the state.

If VRA districts are required and the is a region-based approach then it must comport with the federal law. I looked at the statewide VAP and saw that it was almost a third of the population and with a history of racial bloc voting I sought a compact area of 50% BVAP for a district. Then I looked at the BVAP in each cluster and saw that none of them alone would create a region that had a 50% BVAP district with our rule that a cluster be divided among the fewest number of districts. I tried two combinations of clusters that did reach 50% BVAP and as I drew the districts I sought to reduce chops and erosity.

If I had a tool that had clear county subdivisions like townships I would have used them to the extent possible within the constraints of the VRA. Since redistricting is a process involving electoral districts there is a bias in any redistricting software to provide subunits related to elections. In the half of the states with townships as subdivisions, those entities are used to form voter tabulation districts. It is easy to agree on those units as building blocks for redistricting because they are electoral in nature. The other half of the states may use cities in some cases, but the unincorporated land in the counties has no grouping between the county and the VTD. It will be hard in those states to identify an agreed subunit. The census does provide a subunit, but it is not generally used in the states.

[traininthedistance]

The LA example had me curious to see what could be done in other Deep South states vis-a-vis the VRA.

Alabama is actually a good deal easier to accommodate than I had feared, for the simple reason that the Birmingham cluster is about one-and-a-half districts large, just like New Orleans.  So you cut out inner Birmingham and put it with the Black Belt, draw AL-6 to encompass the rest of the cluster, and go from there.  It's not perfection: having 6 and 7 as immovable blocks around which everything else must be drawn (well, there is wiggle room in the composition of 7) leads to some elongated shapes, and the best I could get on deviation with otherwise whole counties is over the 0.5% threshold (1 is +4634, 5 is -4058).  But another one or two split counties should resolve these issues, and what is here can serve as a rough, improvable proof of concept.



Mississippi is trivial, since all of the urban clusters are small enough that they must take in rural areas.  Just append the Jackson area to as much of the Black Belt as will fit, and voila.  District 2 is 52.6 Black VAP, and 55.8% Obama.

Of course, one can only imagine the howls from Rankin and Madison counties...

[traininthedistance]

Just for funsies:

It is technically possible to draw a black-majority district in Mississippi that is a) whole counties, and b) does not include the Jackson urban cluster.



Obviously we have left quaint notions like "compactness" entirely behind here.  But the point is, there are options in MS.  The black population is quite high.

[Torie]

Please remind me again, just why it is preferable to ban one extra chop of an urban cluster, while on the other hand, one extra county chop not involving an urban cluster, is just one more chop, a demerit yes, but to be balanced against erosity factors? Thanks.

[traininthedistance]

On the issue of minority geographical compactness again, this Brief filed in a NY case attacking a minority district in Nassau because it was not geographically compact, taking in disparate places, and a re-reading of Abrams v Johnson (with the court commenting on the combining of disparate black population nodes syndrome), are elucidative in suggesting that (i) drawing a CD that picks up two disparate minority nodes is not required under Section 2, and (ii) if by picking up those disparate nodes, other appropriate redistricting principles are abandoned to that end, that CD may itself be an illegal racial gerrymander.

The more and more I explore this, the more and more I am persuaded that the scope of Section 2 in forcing the drawing minority CD's that violate other redistricting principles, and would never be drawn absent race as a factor, is very narrow indeed. If a reasonable minority CD can be drawn, that does not violate other principles in play, and is just one reasonable choice among others, than draw it. Otherwise, it's not required, and if too gross, is itself illegal. Indeed after having done this little excursion, I will be more circumspect in drawing minority CD's than I have been heretofore. I believe that while my former interpretation of the scope of the VRA here was narrower than Muon2's, it was not narrow enough. I just won't be breaking many other redistricting rules to draw them, particularly if the only reason to do so, was to pick up geographically disparate minority population nodes, as opposed to a contiguous one, that just happens to spill inconveniently over a number of local jurisdictional lines - particularly, in our chop obsessed minds here, county ones.

So Judge Torie hereby rules that the NO-BR black barbell populated CD is not required under Section 2 of the VRA, but neither is it illegal, because it does reasonably hew to other redistricting principles in play, or at least not totally shred them.

What would Judge Torie say about the black-majority districts I drew for MS and AL above?  I assume that both would be non-illegal, but I can't tell whether you would consider them required or not.  (I would hope that at the very least the Mississippi black district would be required).

And what would Judge Torie say about NC-1?

Please remind me again, just why it is preferable to ban one extra chop of an urban cluster, while on the other hand, one extra county chop not involving an urban cluster, is just one more chop, a demerit yes, but to be balanced against erosity factors? Thanks.

I do not believe in a blanket ban of urban cluster chops (in large part because a) there is always the question of how strictly you define the urban cluster, and b) because I believe that all sorts of chops can have mitigating factors), but would rather see them considered as simply a demerit.  How much of a demerit, I am still wrestling with: right now my gut says that chopping into the urban cluster more than necessary probably deserves to be marked off at the rate of one-and-a-half county cuts, possibly less if it's a marginal county and possibly more if it's a central county.  And also that the scoring system should allow some leeway, but not much.

[traininthedistance]

It also has just occurred to me now, because I'm an idiot, that:

The concept of Urban County Clusters is to handle big blobs of population which are difficult if not impossible to divide into equal population districts based on county boundaries, or which may result in perverse outcomes when they do - such as a large suburban county with the population for half a district stretching across the state to pick up the rest of the district.  In addition in urban areas, county boundaries are a weaker source of identification, with residents, workers, and shoppers crossing the county boundary oblivious to its existence.

An Urban County Cluster means "county-based districts don't work as well here, identify a compact area that contains the blob and can be represented by one or more districts, and treat them separately.

is actually saying that county chops should not really be minded so much within the cluster (perhaps intra-cluster chops should be counted at a discount rate, if at all, in any scoring system?), and that perhaps within the state as a whole the UCC is to be treated as a super-county for the purposes of chop identification and minimization.

Which is a stance I wholeheartedly endorse.

[Torie]

It also has just occurred to me now, because I'm an idiot, that:

The concept of Urban County Clusters is to handle big blobs of population which are difficult if not impossible to divide into equal population districts based on county boundaries, or which may result in perverse outcomes when they do - such as a large suburban county with the population for half a district stretching across the state to pick up the rest of the district.  In addition in urban areas, county boundaries are a weaker source of identification, with residents, workers, and shoppers crossing the county boundary oblivious to its existence.

An Urban County Cluster means "county-based districts don't work as well here, identify a compact area that contains the blob and can be represented by one or more districts, and treat them separately.

is actually saying that county chops should not really be minded so much within the cluster (perhaps intra-cluster chops should be counted at a discount rate, if at all, in any scoring system?), and that perhaps within the state as a whole the UCC is to be treated as a super-county for the purposes of chop identification and minimization.

Which is a stance I wholeheartedly endorse.

No, I think the idea is that a county should not be divided without penalty from its urban cluster to round out another CD, except necessarily, part of one county if the urban cluster's population is a bit short of an some integer of CD's, so a chop has to occur anyway. Allowing  chops within an urban cluster than can be avoided, is the fertilizer for gerrymandering, except to the extent justified by erosity issues.