US House Redistricting: Arkansas

[JohnnyLongtorso]

Not much is probably going to happen in Arkansas, but I thought I'd throw together a 2-2 map of the state:



AR-01 (blue) becomes a completely safe Republican district; it went 65-32 for McCain.
AR-02 (green) adds Jefferson County in order to become more Democratic; it's now 51-48 McCain.
AR-03 (purple) shrinks a bit. Safe Republican, naturally, it's 64-34 McCain.
AR-04 (red) reaches up across the eastern half of the state, but in order to keep Mike Ross satisfied, doesn't turn the district into a Democratic stronghold. It's 56-41 McCain.

Since the state requires no county splits, the population numbers are a little high for AR-03 and a little low in AR-02 (by about 5,000 voters). AR-01 and AR-04 are pretty close to the target.

[minionofmidas]

Trying to keep the basic setup leads to some pretty strange transfers, IIRC, drawing all the districts northwestwards. At least, the way I did it once with estimates data did. Doing something like your map would arguably make more sense... at least that still has three districts that represent a recognizable region; though the blue district gets it up the shaft in that respect.

[muon2]

Not much is probably going to happen in Arkansas, but I thought I'd throw together a 2-2 map of the state:




Since the state requires no county splits, the population numbers are a little high for AR-03 and a little low in AR-02 (by about 5,000 voters). AR-01 and AR-04 are pretty close to the target.

It looks like Logan has a small piece of AR-3.

[JohnnyLongtorso]

Not much is probably going to happen in Arkansas, but I thought I'd throw together a 2-2 map of the state:




Since the state requires no county splits, the population numbers are a little high for AR-03 and a little low in AR-02 (by about 5,000 voters). AR-01 and AR-04 are pretty close to the target.

It looks like Logan has a small piece of AR-3.

Whoops, you're right; thankfully fixing that evens out the population a bit.

[jimrtex]

Since the state requires no county splits, the population numbers are a little high for AR-03 and a little low in AR-02 (by about 5,000 voters). AR-01 and AR-04 are pretty close to the target.

I'm not sure that there is actually a requirement of no county splits.  I couldn't find anything in the Constitution.

In 2001, the legislature enacted the current county-based boundaries, and then added a section that defined what would happen if those were overturned.

AR-1 has a perfect population (a deviation of 0 from the ideal of 668,360.

AR-2 was short 2,302 and would have taken 4 precincts from Pope county.

AR-4 was short 2.094 and would have taken 5 precincts from Franklin county.

AR-3 has an excess of 4,396 and would have lost the above areas.

So they actually had a 0.97% total deviation,  and AR-2 was 0.65% over.

[freepcrusher]

arkansas is one of the few southern states left where democrats control just about everything in the statehouse. I expect them to draw as favorable a map to democrats as possible.

[jimrtex]

arkansas is one of the few southern states left where democrats control just about everything in the statehouse. I expect them to draw as favorable a map to democrats as possible.

They don't have that big of a majority.

If you are Democratic representative from a weird appendage, do you vote for the bill?

They have been so dominant for so long, they probably aren't easily whipped.

[muon2]

Here's my first version of a remap based on whole counties. I only moved 13 counties and substantially improved on the population equality from the 2000 map. This version has an average deviation of 761 (0.10%) and a range of 2151 (0.30%).



CD 1 (cyan): 727,456
CD 2 (yellow): 729,395
CD 3 (forest): 729,607
CD 4 (pink): 729,460

Edit: The map was incorrectly painted in the OP, and is now correct. The variances were correct.

[jimrtex]

Here's my first version of a remap based on whole counties. I only moved 14 counties and substantially improved on the population equality from the 2000 map. This version has an average deviation of 761 (0.10%) and a range of 2151 (0.30%).



CD 1 (cyan): 727,456
CD 2 (yellow): 729,395
CD 3 (forest): 729,607
CD 4 (pink): 729,460

In 2001, they actually amended the existing redistricting statute, as opposed to simply striking everything and replacing (see Act 1840, 2001 session).  Only 4 counties were shifted:

Baxter from 3 to 1; and  Logan, Polk, and Scott from 3 to 4.

So they might opt for less equality, it there are fewer adjustments.

[Frodo]

arkansas is one of the few southern states left where democrats control just about everything in the statehouse.

One of two -if you count West Virginia as a southern state.

[tpfkaw]

Democrats still control the Mississippi statehouse, though they'll lose it this year.

[minionofmidas]

Maryland?

[muon2]

In 2001, they actually amended the existing redistricting statute, as opposed to simply striking everything and replacing (see Act 1840, 2001 session).  Only 4 counties were shifted:

Baxter from 3 to 1; and  Logan, Polk, and Scott from 3 to 4.

So they might opt for less equality, it there are fewer adjustments.

It is possible to move only 4 counties and have the population be within the same range as the current map. Move Pope, Johnson, and Newton from CD 3 to CD 4, and move Van Buren from CD 2 to CD 1. The resulting map has an average deviation of 0.35%, a range of 0.98%, and CD 2 has the maximum deviation of 0.70% over the ideal.

However, if someone doesn't like the map that moves a minimum of counties, they could sue. They would claim that it was practicible to move the 13 counties in my example and have a smaller deviation and range, yet still keep counties whole.

Edit: Here's the map and populations for this version.


CD 1 (cyan): 727,342
CD 2 (yellow): 734,082
CD 3 (forest): 726,940
CD 4 (pink): 727,554

[Frodo]

Democrats still control the Mississippi statehouse, though they'll lose it this year.

I think he was talking about Democratic-controlled state legislatures in the South -not individual chambers.  

Maryland?

When did the Confederacy ever officially claim Maryland?  

[minionofmidas]

Democrats still control the Mississippi statehouse, though they'll lose it this year.

I think he was talking about Democratic-controlled state legislatures in the South -not individual chambers.  

Maryland?

When did the Confederacy ever officially claim Maryland?  

When did the Confederacy ever include all of the South?

[nclib]

Maryland isn't part of the South demographically or culturally, and barely is latitudinally.

As for Arkansas's map, any chance the legislature will make a map that will strengthen the Dems' chance at either AR-1 or AR-2?

[muon2]

As for Arkansas's map, any chance the legislature will make a map that will strengthen the Dems' chance at either AR-1 or AR-2?

The map I describe with the minimum county movement is probably better for the Dems. However, if it is the GOP has a strong challenge along the lines I suggest.

However, its hard to know what data the Dems should use to improve their chances. The federal and state races are strongly diverging, and the federal votes in the last decade don't look good. Their best play would be to redivide CD 1 and 2 and try to link Little Rock to the Mississippi.

In 2001, they actually amended the existing redistricting statute, as opposed to simply striking everything and replacing (see Act 1840, 2001 session).  Only 4 counties were shifted:

Baxter from 3 to 1; and  Logan, Polk, and Scott from 3 to 4.

So they might opt for less equality, it there are fewer adjustments.

It is possible to move only 4 counties and have the population be within the same range as the current map. Move Pope, Johnson, and Newton from CD 3 to CD 4, and move Van Buren from CD 2 to CD 1. The resulting map has an average deviation of 0.35%, a range of 0.98%, and CD 2 has the maximum deviation of 0.70% over the ideal.

However, if someone doesn't like the map that moves a minimum of counties, they could sue. They would claim that it was practicible to move the 13 counties in my example and have a smaller deviation and range, yet still keep counties whole.

[jimrtex]

Here's my first version of a remap based on whole counties. I only moved 14 counties and substantially improved on the population equality from the 2000 map. This version has an average deviation of 761 (0.10%) and a range of 2151 (0.30%).



CD 1 (cyan): 727,456
CD 2 (yellow): 729,395
CD 3 (forest): 729,607
CD 4 (pink): 729,460

In 2001, they actually amended the existing redistricting statute, as opposed to simply striking everything and replacing (see Act 1840, 2001 session).  Only 4 counties were shifted:

Baxter from 3 to 1; and  Logan, Polk, and Scott from 3 to 4.

So they might opt for less equality, it there are fewer adjustments.

It is possible to move only 4 counties and have the population be within the same range as the current map. Move Pope, Johnson, and Newton from CD 3 to CD 4, and move Van Buren from CD 2 to CD 1. The resulting map has an average deviation of 0.35%, a range of 0.98%, and CD 2 has the maximum deviation of 0.70% over the ideal.

However, if someone doesn't like the map that moves a minimum of counties, they could sue. They would claim that it was practicible to move the 14 counties in my example and have a smaller deviation and range, yet still keep counties whole.

Which is almost identical to the situation that they had in 2001, when the statute included additional language that would have split two counties if anyone had challenged the plan in court.  The only difference is that AR-1 had the precisely ideal population in 2000.

I don't think that Arkansas has a legal requirement to not split counties, nor to ignore existing districts like in Iowa.

So Arkansas will argue they have a rational state purpose: to not split counties and to maintain existing representative-constituent relationships, while getting under 1% total deviation.  And the legislature has already expressed a policy preference to split counties rather than do large reshuffling merely to make districts a couple 1000 closer.  They also have a historical record of similar plans.

[jimrtex]

Maryland isn't part of the South demographically or culturally, and barely is longitudinally.

Historically, Maryland was divided North/South, with the Chesapeake serving as a transportation link rather than a barrier.   There weren't any Washington suburbs.

[JohnnyLongtorso]

The Eastern Shore is very much like the Deep South, culturally speaking.

[muon2]

As for Arkansas's map, any chance the legislature will make a map that will strengthen the Dems' chance at either AR-1 or AR-2?

The map I describe with the minimum county movement is probably better for the Dems. However, if it is the GOP has a strong challenge along the lines I suggest.

However, its hard to know what data the Dems should use to improve their chances. The federal and state races are strongly diverging, and the federal votes in the last decade don't look good. Their best play would be to redivide CD 1 and 2 and try to link Little Rock to the Mississippi.

Here's a map that does what I suggested. CD 2 (yellow) would have voted for Obama by about 3% over McCain. Needless to say, CD 1 becomes hard R. It also does increase the GOP lean of CD 4 by a couple of points, however.

As with other plans for AR, counties are kept intact. The average deviation is 1956 (0.27%) and the range is 4866 (0.67%).



CD 1 (cyan): 731,806
CD 2 (yellow): 730,066
CD 3 (forest): 726,940
CD 4 (pink): 727,106

[Platypus]

D1 (blue): 708,934 (-602)

D2 (orange): 710,676 (+1,140)

D3 (green): 710,264 (+728)

D4 (red): 708,269 (-1,267)

[Platypus]

Why do I have a different state population total to you, Muon? I'm using new estimates on 2000 block groups.

[Smash255]

Why do I have a different state population total to you, Muon? I'm using new estimates on 2000 block groups.

Muon is using the updated 2010 Census Data, you can find the population by county for some states including Arkansas on the Census Bureau website.  The 2010 Census Data update is not yet available on Dave's site (though some other states are).

[muon2]

Why do I have a different state population total to you, Muon? I'm using new estimates on 2000 block groups.

Muon is using the updated 2010 Census Data, you can find the population by county for some states including Arkansas on the Census Bureau website.  The 2010 Census Data update is not yet available on Dave's site (though some other states are).

That's right. I loaded the county data from Census 2010 into spreadsheet and used that to build districts. My plans are equivalent to real plans that AR may be considering.