Fair Redistricting (PA aftermath)

[cvparty]

No, the remaining issue was chopping a voting district that contains multiple towns to get better population equality by cherry picking only some of the towns in the VD. But you said no. It is probably true that no voting district chops a unorganized town, so that distinction does not matter. If a VD did, than it would matter as to whether or not an unorganized town is treated the same as an organized town.

Okay, so yes your question "So one uses the voting districts, they cannot be chopped, and for whether or not you have chopped a municipal unit, just compare the lines of the voting districts, to the 'town' lines on the Leaps map, and whatever that indicates, is what matters, and nothing else" is correct. I don't know about the other panelists, but I personally am not going to be fanatical about county and town lines, just as long as the district is reasonably compact and representative. In my opinion I think the other criteria are better and more important measures of a map's fairness. I really would not bend over backwards and split up metro areas and make oddly shaped districts just to have population equality with completely whole counties.

[jimrtex]

It might carry weight which map is more exact if the maps otherwise tie, but whatever. And you might be right, that there is no voting district that takes in parts (slices, dices, bits, pieces, fragments) of two towns (are you sure?). But the participants will have to figure out digging out the town maps, which voting districts chop a town, and which do not (see map below where one town (South Oxford) is split into two voting districts, with one voting district therein also taking in another town (Stoneham) to confuse matters further). Good luck with that. The voting district numbers give no clue as to whether a split is involved or not.

Maine SOS web site has election results by town, and the Census Bureau has shapefiles and population totals by town. It should trivial to compensate for deficiencies in DRA.

[jimrtex]

Regarding Maine, I am not sure using towns is a very good idea. Many of the towns are defunct, although where the lines are drawn between CD's, in all probability the towns involved will have some governmental life to them (the defunct towns are mostly in the very low population density north woods). But even using block groups on the DRA rather than election districts (which are larger), the block groups do not necessarily match the town lines. For example, looking at the maps below, if I added the town of Sweden in Oxford county to the blue CD, I would get very close to exact population equality (Sweden has a population of 391). But I can't do that. There are two block groups that together cover Sweden and Lovell, but the line between the two block groups does not match the town line between Sweden and Lovell. Other block groups combine more than one town, as one can see with the block group that takes in Stow, South Oxford and Stoneham.

Sure, one could invite participants to search out the town maps on the internet, figure out which towns still have meaning rather than having been consigned to the ash heap of history, look up the populations, and then photoshop the map to draw a line chopping a block group, and adjust the population accordingly. I doubt you guys want to have that level of complexity here, particularly with tight time frames, which might discourage participants who want to play, but don't want this activity to be their chief hobby in life.

Iowa by the way has towns (townships). They are all defunct so far as I know (and I do know they are in Madison County).

That is not true for Iowa. Townships are not independent governments (municipalities), but do have elected trustees.

Sweden: DJT 116, HRC 115
Stoneham: DJT 89, HRC 77

[Torie]

It might carry weight which map is more exact if the maps otherwise tie, but whatever. And you might be right, that there is no voting district that takes in parts (slices, dices, bits, pieces, fragments) of two towns (are you sure?). But the participants will have to figure out digging out the town maps, which voting districts chop a town, and which do not (see map below where one town (South Oxford) is split into two voting districts, with one voting district therein also taking in another town (Stoneham) to confuse matters further). Good luck with that. The voting district numbers give no clue as to whether a split is involved or not.

Ah well. I had hope that one grouping might be better than the other.

So this would be my initial submission. I checked the town maps and there is only one county chop and no town chops and does not chop a UCC. CD 1 (blue) has a deviation of +75. I think this is as low as I can get erosity using counties as the primary unit and towns as the subunit.

Your map is using block groups, rather than VD's, right? So there is no partisan data, although obviously it could be closely estimated by redrawing a close approximation of the map using VD's. Everybody should draw using the same map I would think.

[muon2]

It might carry weight which map is more exact if the maps otherwise tie, but whatever. And you might be right, that there is no voting district that takes in parts (slices, dices, bits, pieces, fragments) of two towns (are you sure?). But the participants will have to figure out digging out the town maps, which voting districts chop a town, and which do not (see map below where one town (South Oxford) is split into two voting districts, with one voting district therein also taking in another town (Stoneham) to confuse matters further). Good luck with that. The voting district numbers give no clue as to whether a split is involved or not.

Ah well. I had hope that one grouping might be better than the other.

So this would be my initial submission. I checked the town maps and there is only one county chop and no town chops and does not chop a UCC. CD 1 (blue) has a deviation of +75. I think this is as low as I can get erosity using counties as the primary unit and towns as the subunit.

Your map is using block groups, rather than VD's, right? So there is no partisan data, although obviously it could be closely estimated by redrawing a close approximation of the map using VD's. Everybody should draw using the same map I would think.

The original map was drawn in voting district mode and because of combined towns I was left with a deviation of 521. To get the chop I wanted I switched to block group mode (swapping Chesterville and New Vineyard). I found that also had PVI info that matched the values that appeared in voting district mode. Since I had partisan data from the block groups and my preferred chop, I went with that.

[jimrtex]

Regarding Maine, I am not sure using towns is a very good idea. Many of the towns are defunct, although where the lines are drawn between CD's, in all probability the towns involved will have some governmental life to them. But even using block groups on the DRA rather than election districts (which are larger), the block groups do not necessarily match the town lines. For example, looking at the maps below, if I added the town of Sweden in Oxford county to the blue CD, I would get very close to exact population equality (Sweden has a population of 391). But I can't do that. There are two block groups that together cover Sweden and Lovell, but the line between the two block groups does not match the town line between Sweden and Lovell. Other block groups combine more than one town, as one can see with the block group that takes in Stow, South Oxford and Stoneham.

I use the 2010 voting districts option, not block groups. ME consolidates its voting districts across towns for polling efficiency. I believe that when towns are consolidated into a voting district the voting district consists only of whole towns so no chops are needed. In any case the panel has expressed less interest in exact equality as long is the maximum deviation is under 0.5%. So going from under 400 deviation to under 100 deviation may not carry much weight.

Maine conducts its elections by towns, except the unorganized towns, which are generally unorganized because there are no people to organize them.

The VTD's are what Maine submitted to the Census Bureau, and appear to be Senate-House intersections (they are relatively small because Maine has so many House districts). The numbers are of the form: sshhhi, where ss is the senate district, hhh is the house district, and i is an index.

[Torie]

It might carry weight which map is more exact if the maps otherwise tie, but whatever. And you might be right, that there is no voting district that takes in parts (slices, dices, bits, pieces, fragments) of two towns (are you sure?). But the participants will have to figure out digging out the town maps, which voting districts chop a town, and which do not (see map below where one town (South Oxford) is split into two voting districts, with one voting district therein also taking in another town (Stoneham) to confuse matters further). Good luck with that. The voting district numbers give no clue as to whether a split is involved or not.

Ah well. I had hope that one grouping might be better than the other.

So this would be my initial submission. I checked the town maps and there is only one county chop and no town chops and does not chop a UCC. CD 1 (blue) has a deviation of +75. I think this is as low as I can get erosity using counties as the primary unit and towns as the subunit.

Your map is using block groups, rather than VD's, right? So there is no partisan data, although obviously it could be closely estimated by redrawing a close approximation of the map using VD's. Everybody should draw using the same map I would think.

The original map was drawn in voting district mode and because of combined towns I was left with a deviation of 521. To get the chop I wanted I switched to block group mode (swapping Chesterville and New Vineyard). I found that also had PVI info that matched the values that appeared in voting district mode. Since I had partisan data from the block groups and my preferred chop, I went with that.

What version of the DRA had PVI data for block groups? 2.2 did not.

[Kevinstat]

Regarding Maine, I am not sure using towns is a very good idea. Many of the towns are defunct, although where the lines are drawn between CD's, in all probability the towns involved will have some governmental life to them. But even using block groups on the DRA rather than election districts (which are larger), the block groups do not necessarily match the town lines. For example, looking at the maps below, if I added the town of Sweden in Oxford county to the blue CD, I would get very close to exact population equality (Sweden has a population of 391). But I can't do that. There are two block groups that together cover Sweden and Lovell, but the line between the two block groups does not match the town line between Sweden and Lovell. Other block groups combine more than one town, as one can see with the block group that takes in Stow, South Oxford and Stoneham.

I use the 2010 voting districts option, not block groups. ME consolidates its voting districts across towns for polling efficiency. I believe that when towns are consolidated into a voting district the voting district consists only of whole towns so no chops are needed. In any case the panel has expressed less interest in exact equality as long is the maximum deviation is under 0.5%. So going from under 400 deviation to under 100 deviation may not carry much weight.

Maine conducts its elections by towns, except the unorganized towns, which are generally unorganized because there are no people to organize them.

The VTD's are what Maine submitted to the Census Bureau, and appear to be Senate-House intersections (they are relatively small because Maine has so many House districts). The numbers are of the form: sshhhi, where ss is the senate district, hhh is the house district, and i is an index.

The last digit is the County Commissioner district.  Also, these are the pre-2013 State House, State Senate and County Commissioner districts.  I find the block groups much better overall for Maine, but they have issues too, as has been shown.

I'm working with Amanda Rector, who's Maine's State Economist but also the "Governor's liaison" to the Census Bureau's Redistricting Data Program, to get better voting districts in the future (although most of our conversations so far have concerned census block lines as the Redistricting Data Program was in its Block Boundary Suggestion Project phase until last year).  Like each town with its own polling place being its own voting district, and also each ward in a city being its own voting district, like New Hampshire does now (their 2000 voting districts were as bad as or worse than Maine's, but they really improved in 2010).  Or perhaps the voting districts could be the intersections of the towns/city wards as well as the State House districts, State Senate districts and County Commissioner districts.  The Congressional district lines in Kennebec County haven't been voting district lines there (unless they're also one of the other three district lines), but if you had the borders of organized towns as voting district boundaries that would take care of the Congressional district lines anyway, with all the Congressional districts Maine has had in recent memory at least.

[jimrtex]

Maine conducts its elections by towns, except the unorganized towns, which are generally unorganized because there are no people to organize them.

The VTD's are what Maine submitted to the Census Bureau, and appear to be Senate-House intersections (they are relatively small because Maine has so many House districts). The numbers are of the form: sshhhi, where ss is the senate district, hhh is the house district, and i is an index.

The last digit is the County Commissioner district.  Also, these are the pre-2013 State House, State Senate and County Commissioner districts.  I find the block groups much better overall for Maine, but they have issues too, as has been shown.

I'm working with Amanda Rector, who's Maine's State Economist but also the "Governor's liaison" to the Census Bureau's Redistricting Data Program, to get better voting districts in the future (although most of our conversations so far have concerned census block lines as the Redistricting Data Program was in its Block Boundary Suggestion Project phase until last year).  Like each town with its own polling place being its own voting district, and also each ward in a city being its own voting district, like New Hampshire does now (their 2000 voting districts were as bad as or worse than Maine's, but they really improved in 2010).  Or perhaps the voting districts could be the intersections of the towns/city wards as well as the State House districts, State Senate districts and County Commissioner districts.  The Congressional district lines in Kennebec County haven't been voting district lines there (unless they're also one of the other three district lines), but if you had the borders of organized towns as voting district boundaries that would take care of the Congressional district lines anyway, with all the Congressional districts Maine has had in recent memory at least.

In a way it would be better if the Census Bureau got rid of VTD's, since they often change after the census, as state and local election boundaries change. Alternatively, they should retabulate the census results like they do for legislative and congressional districts.

[muon2]

Maine conducts its elections by towns, except the unorganized towns, which are generally unorganized because there are no people to organize them.

The VTD's are what Maine submitted to the Census Bureau, and appear to be Senate-House intersections (they are relatively small because Maine has so many House districts). The numbers are of the form: sshhhi, where ss is the senate district, hhh is the house district, and i is an index.

The last digit is the County Commissioner district.  Also, these are the pre-2013 State House, State Senate and County Commissioner districts.  I find the block groups much better overall for Maine, but they have issues too, as has been shown.

I'm working with Amanda Rector, who's Maine's State Economist but also the "Governor's liaison" to the Census Bureau's Redistricting Data Program, to get better voting districts in the future (although most of our conversations so far have concerned census block lines as the Redistricting Data Program was in its Block Boundary Suggestion Project phase until last year).  Like each town with its own polling place being its own voting district, and also each ward in a city being its own voting district, like New Hampshire does now (their 2000 voting districts were as bad as or worse than Maine's, but they really improved in 2010).  Or perhaps the voting districts could be the intersections of the towns/city wards as well as the State House districts, State Senate districts and County Commissioner districts.  The Congressional district lines in Kennebec County haven't been voting district lines there (unless they're also one of the other three district lines), but if you had the borders of organized towns as voting district boundaries that would take care of the Congressional district lines anyway, with all the Congressional districts Maine has had in recent memory at least.

In a way it would be better if the Census Bureau got rid of VTD's, since they often change after the census, as state and local election boundaries change. Alternatively, they should retabulate the census results like they do for legislative and congressional districts.

Since any state realign their VTDs after the Census, it seems that the best solution would be for the Census to retabulate VTDs in 2023 based on 2020 data (looking ahead to the next cycle).

[jimrtex]

Perhaps you could provide a link to where one finds all these town maps.

Census Bureau.

Geography > Maps&Data > Reference Maps > Census Reference Maps > County Block Maps.

For a given county, the 000 map is the county index map to the other maps, but is probably sufficient for your purpose.

Geography > Maps&Data > TIGER products > KML - Cartographic Boundary Files

Select the county subdivision file, and drag into Google Earth.

Geography > Maps&Data > TIGER products > TIGER/Line Shapefiles

Geography > Maps&Data > TIGER products > Cartographic Shapefiles

Either can be used with a GIS program such as QGIS.

You can also use American Fact Finder and draw a map.

[muon2]

Your map is using block groups, rather than VD's, right? So there is no partisan data, although obviously it could be closely estimated by redrawing a close approximation of the map using VD's. Everybody should draw using the same map I would think.

The original map was drawn in voting district mode and because of combined towns I was left with a deviation of 521. To get the chop I wanted I switched to block group mode (swapping Chesterville and New Vineyard). I found that also had PVI info that matched the values that appeared in voting district mode. Since I had partisan data from the block groups and my preferred chop, I went with that.

What version of the DRA had PVI data for block groups? 2.2 did not.

I just now started DRA from scratch for ME using block groups and it loads the extra data files and provides a PVI based on the 2012/2016 presidential races.

[jimrtex]

Perhaps you could provide a link to where one finds all these town maps.

I search for them separately, eg franklin county maine town map. I can usually find one in the images. There are too many in ME overall to see them clearly on a state map.

You of course love all this navigating of the maze I understand. My map beats yours on equality, if I get to chop a voting district, to tease out the glorious town of Sweden.  Yours no doubt wins on erosity using your metric, which has not been adopted by the panel. Rather their metric is more alone the lines of the pornography test, that you know it when you see it. Some might think something that is elongated, traversing the state from north to south in an emaciated way, but scores will using your metric, because it hugs a state line, is more erotic than your metric scoring. But that is part of the fun I guess. 

Anyway, just out of curiosity, can a poster chop a voting district in the way I did, or not?  What is the rule?

Yes.

Voting Districts in Maine are fairly arbitrary and have little to do with actual voting practices. A state can tell the Census Bureau that we want data tabulated for these areas.

In Hudson, NY somebody probably gave the census bureau a paper map of precinct boundaries. The Census Bureau probably said we don't split block boundaries, so we'll just combine the submitted 7 precincts into 3 VTDs and adjust some boundary lines to follow streets - including the wrong streets in some places (remember that Ward 5 had 2 precincts, and until fairly recently Ward 3 also had two).

The census bureau tabulates the census data based on VTD's, and then the folks in Hudson treat it as gospel for wards 3 and 5, and realize that wards 1, 2, and 4 have to be estimated from "VTD 1-1 2-1 4-1" and botch that. Since 2010, the Census Bureau has permitted VTD's to follow non-visible boundaries. Right now, VTD definitions are being updated. The mayor or county board of elections might have some leverage with regard to the census liaisons in Albany, who otherwise will mostly be interested in their bon bon supply.

Back to Maine. Maine doesn't have particularly useful VTD's. They are the intersection of house, senate, and county commissioner's districts. At best they could be described as areas with the same ballot style. Maine does not administer elections at the county level. They are administered by towns. Someone preparing the DRA data had to take the town results and aggregate them into VTD's. DRA should probably use towns and wards in larger cities for its basic units. But if you were redistricting Maine, you would use towns.

[jimrtex]

and you can't split south oxford in DRA anyway...

Actually I showed above that you can. But whether the Leips map is accurate or not, pretending that it is, makes total sense to me. Looking for other maps on the internet does not. That still leaves the question, as to whether a participant can chop a voting district, in order to avoid a town chop, or to take in just one of several towns that are in the voting district.

It is odd that Leaps calls Albany an unorganized township the Leips map rules, and you say assume any designation of UT thereon, as being no different from where it says "Town," than that puts that matter to bed.

Actually, Albany is an unorganized town with no functions these days ("defrocked" some time ago as a real town). But if, along with the other two bits that you mentioned as constituting the "unorganized territory" of South Oxford (which other two bits might not even have a UT status (I can't find that out on the internet), that does not matter, we can cease to worry about that.

The Census Bureau divides counties into county subdivisions. In the northeastern quarter of the country, they correspond to minor civil divisions, towns and cities. In the rest of the country they are pretty arbitrary.

In Maine, not all territory has an organized government. Some areas were never settled, and never had a formal local government. In other areas, the local government has been deorganized. That happened in 1933 to Albany town. It reverted to Albany township. Townships represent the territory, towns represent the government.

Through 1950, the Census Bureau reported the population for these areas without local government. In 1960, they reported the population in each county that did not have a local government.

Since then, the Census Bureau has reported areas without local government as county subdivisions, designated by the Census Bureau as Unorganized Territory. In Oxford County, there are three such areas, South Oxford UT, North Oxford UT, and Milton UT. Remember, these are designations by the Census Bureau.

In Maine, "Voting Districts" are NOT districts used for voting. The Census Bureau permits states to designate areas as VTD's which it tabulates census data for. For example, in Hudson, there are 3 VTD's. One corresponds to roughly precincts 1-1, 2-1, and 4-1, one corresponds to precincts 3-1 and 3-2, and one corresponds to precincts 5-1 and 5-2. These are not identical to either the old ward boundaries, nor to the areas actually used for elections. Whoever prepared the election data for DRA for Hudson, presumably totaled the ward results for 1, 2, and 4 and assigned that to "VTD 1-1 2-1 4-1", for 2016, they would have used the new wards, even though they don't correspond to the map.

DRA has made a decision to use "Voting Districts". In some states these do correspond to election precincts. In others the election results have to allocated. In Maine and Hudson, this means that election data reported by the state has to be adjusted to match the chosen census geography.

In Maine, elections are conducted by towns. There are no county boards of elections. This is a problem if there is an area with voters, but no government. State law provides that the county provide for polling places for these areas. In some cases, the county contracts with a neighboring town. Voters from the townships are marked with a "T" so that they are not permitted to vote in town elections.

Albany Township is unusual in that it has a substantial population, greater than some organized towns in Oxford County. It has 800 persons in 1900, and had declined to around 240 by 1950, probably due to a decline in logging jobs, and perhaps to establishment of the White Mountain National Forest, if private land was being purchased. It has since rebounded to around 500 persons (it appears that 80% or so of the South Oxford UT population is in Albany). You don't need to be a logger to live in the forest. You can have a second home, that becomes a first home. You can drive into Norway or Rumford to work or for groceries. Albany is only about half in the national forest, so someone might have a home in the woods, and rent a bed and breakfast, either extra rooms or an adjacent cabin.

So it appears that Albany Township votes in Albany and the county has appointed a municipal clerk and register. Mason Township and Batchelder Grant, the portions of the South Oxford UT to the west are entirely in the national forest and much less populous. I have not found out where they vote. It would make sense to vote in Bethel, since that is the same senate, house, and commissioner's district. You would only have to keep the township voters from voting in town races.

[America Needs a 13-6 Progressive SCOTUS]

I am going to just put New Hampshire in now:

https://uselectionatlas.org/FORUM/index.php?action=gallery;sa=view&id=19719

[America Needs a 13-6 Progressive SCOTUS]

I have 2 maps for Connecticut drawn already.

[cvparty]

solid ya gotta wait till we get there dw it’ll be soon

[Mike Thick]

Sorry for being a bit inactive — auditions and stuff. I support Torie’s Maine plan.

I have maps for CT, RI, and NH that I will post soon.

[America Needs a 13-6 Progressive SCOTUS]

I am going to just put New Hampshire in now:

https://uselectionatlas.org/FORUM/index.php?action=gallery;sa=view&id=19719

Now that this has started, I stand by this.

[Kevinstat]

Perhaps you could provide a link to where one finds all these town maps.

I'm sorry I didn't reply to this earlier.  You already seem to know about this directory of 2000 census-era maps of county subdivisions (which New England towns are often treated as by the Census Bureau) and places, although maybe the map of Oxford County towns I thought you posted was copied from someone else (I saw it in a post of yours and not in a quote box).  To get the 2010 census version, click on "Parent Directory" and then click on "GARM2010/", which takes you here.  You might not be able to be sure in all cases whether a line between two block groups or two voting districts follows a town line, but in most cases you'll get a pretty good idea if it lines up or not.

It just occurred to be as I was writing this that you might have been talking about, say, maps with the boundary between Albany Twp. and Mason Twp. in South Oxford UT.  For those boundaries, the most user thing I've found for Maine (although I haven't looked that hard) is the MaineDOT Public Map Viewer.  I sometimes think of Maine having two sets of sub-county municipal boundaries, the Census County Subdivisions and the DOT/DeLorme set (DeLorme being the former publisher of The Maine Atlas and Gazetteer).  They're mostly identical in actual towns, with the DOT/DeLorme "Minor Civil Divisions" being subsets of the Census County Subdivisions, but some of the County Subdivision boundaries cut across townships, and the Penobscot Indian Island Reservation, which is its own Census County Subdivision, has it's territory parceled out between surrounding towns in the DOT/DeLorme MCDs, and the Passamaquoddy Pleasant Point Reservation is treated as part of the town of Perry.  The Penobscot Reservation, which stretches as far as Medway where the Penobscot River branches (it's pretty much the territory in the river and the islands within), is divided in the DOT/DeLorme MCDs roughly along the main channel and along extensions of the town lines from off shore.  My current Maine Atlas (I've gone through a bunch, some of which I've drawn county lines (before they highlighted those), school district lines, congressional or legislative district lines, etc.), has the land territory in the reservations in pink, but it does the same for the non-reservation "Trust Territories" (I think that's what I've heard them called) that I think were purchased by the tribes as part of the Maine Indian Claims Act in the late 1970s.  The Passamaquoddy have some Trust Territory along the Quebec border just northeast of where U.S. Route 201 crosses it and becomes (Provincial?) Route 173, which I'm pretty sure is far west of historical Passamaquoddy land.

Anyway, I think you may find those links helpful.  You could also see some changes in the Census County Subdivisions between the 2000 and 2010 censuses, like the former town of Madrid becoming part of East Central Franklin UT (interestingly, it's in a different census tract from the remainder of East Central Franklin but the same one as West Central Franklin UT which the former town also shared a border with, albeit a smaller one).

Since at least the 1990s, and I think since the 1980s, legal descriptions of election district lines, in statutes, court rulings, what have you, have referenced the Census-defined units.

[Mike Thick]

Resigning from the panel. Sorry for holding up the panel, at this point it's pretty clear that I don't have the available time to really commit like most of you clearly are

[America Needs a 13-6 Progressive SCOTUS]

Resigning from the panel. Sorry for holding up the panel, at this point it's pretty clear that I don't have the available time to really commit like most of you clearly are

I would be willing to take your place.

[President Punxsutawney Phil]

I am a bit confused about what how we are proceeding now in regards to procedures. Could someone in the know give me a rundown?
The impression I get is that we've split the US into 4 bits and we're running those concurrently?

[Strudelcutie4427]

I am a bit confused about what how we are proceeding now in regards to procedures. Could someone in the know give me a rundown?
The impression I get is that we've split the US into 4 bits and we're running those concurrently?

We’re running several at the same time I think just to save time. We just need your vote on the NH one in the north east thread and I think we’re allowed to start on the first States in the other regions

[President Punxsutawney Phil]

I am a bit confused about what how we are proceeding now in regards to procedures. Could someone in the know give me a rundown?
The impression I get is that we've split the US into 4 bits and we're running those concurrently?

We’re running several at the same time I think just to save time. We just need your vote on the NH one in the north east thread and I think we’re allowed to start on the first States in the other regions

Where are all the NH maps posted in a row, a single post, to make it more obvious what we're voting for?