Closed mzwu closed 2 years ago
The red line in the validation plot is the 2010 CD plan (which is probably why its population deviance is over 12%), but perhaps it would be better to show the 2020 passed plan instead as a reference?
Thanks Melissa for updating the original post with more analysis. It looks good to me; @christopherkenny or @CoryMcCartan can one of you do another final check?
See the original post that has been updated. Some additional notes:
Here's the BVAP boxplot again but with the enacted district numbers
and here's where the districts are
The new NC-09 (was 12) is the Charlotte district and NC-02 (was mostly 1, by Butterfield) are the top 2. It's good that the VRA simulation helped nudge the simulations to two districts.
NC's population grew 9.5% from 2010 and the state gained a CD. Blacks and Whites grew about equally in absolute population, with concentration in the urban areas. So I feel we don't feel we need to force another heavily BVAP district. The diversity is not ideal as is.
In 2020 NC, Cunningham won 49% of the vote, Biden 49%, Gov. Cooper 52%. So the "Performance plot" showing about 6 of the highest BVAP districts having a Dem winner sounds about right.
Two thoughts:
I can adjust the VRA constraint strength and/or target minority VAP percentage, and see where that gets us. Avoiding incumbent pairing is allowed, but not required. Should I implement that?
I did the oversampling and kept the plans with 2 MMDs. The updated validation plots and performance plot are in the PR. Thoughts? @kuriwaki @christopherkenny
Maybe I misunderstood but I thought @christopherkenny was saying to only keep plans that have Democrats winning in high BVAP districts, rather than keep plans with MMDs. In the performance plot, should we apply that filter to ordered districts 11-14? It seems all the simulted most BVAP district are "performant".
Yes, my original point was that some of the most-black districts had non-performing districts, so even though they were around 50% black, some of them would not elect Democrats, which we know is the minority preferred choice in North Carolina from recent litigation.
If the enacted plan has two performing MMDs, this is a reasonable choice, as your set up also has two performing MMDs then. The next two are probably sufficient as influence districts, based on their MVAP and party values, if there are two MMDs.
I think we are good to go here then @mzwu. Is that right @kuriwaki and @CoryMcCartan?
yes: the new edits reflected in the top post addresses the issues we had last Wednesday.
Looks good to me. The only thing is what do to now that the plan has been struck down. Generally we haven't shown plans under litigation, but I don't think there's a particular reason not to. As long as we update it with the new plan once it's decided.
Redistricting requirements
In North Carolina, under North Carolina State Constitution Article II Sections 3 & 5, districts must:
Interpretation of requirements
We enforce a maximum population deviation of 0.5%. We add a county constraint. We add a VRA constraint targeting two majority-minority districts.
Data Sources
Data for North Carolina comes from the ALARM Project's 2020 Redistricting Data Files. Data for the 2021 North Carolina ratified congressional map comes from the North Carolina State Board of Elections.
Pre-processing Notes
No manual pre-processing decisions were necessary.
Simulation Notes
We sample 6,000 districting plans for North Carolina and subset to 5,000 which contain at least two majority-minority districts. A county split constraint was used.
Validation
Checklist
TODO
lines from the template code have been removedenforce_style()
to format my coderedist_map
andredist_plans
objects, and summary statistics) have been editedAdditional Notes
Performance plot for BVAP.
Statistics for population growth and racial composition change from 2010 to 2020. In general, rural counties are losing population and urban cities are driving population growth. White and Black population have both increased, but they each comprise a lesser percentage of overall population than they did in 2010.
@kuriwaki