Home

Replication Code and Data for: The Race-Blind Future of Voting Rights

Data Files, Replication Code, and Simulated Districting Plans For:
Jowei Chen and Nicholas Stephanopoulos, The Race-Blind Future of Voting Rights (Yale Law Journal, Forthcoming)

Online Appendix (PDF, 50 pages)

1. VTD-Level Shapefiles with Merged Election Data
Each state's shapefile contains the following fields:
COUNTYFP10: County FIPS Number
VTDST10: VTD FIPS Number
NAME10: VTD Name
County: County Name
Obama: Barack Obama votes (November 2012)
Romney: Mitt Romney votes (November 2012)
Source: Daily Kos

2. Block-Level Geographic Boundaries, Population, and Racial Data
Within each state's .txt data file:
geoid: 2010 Census Block geoid
Tpop: Total Population
Tvap: Voting Age Population
Bvap: Black Voting Age Population
hNBvap: Non-Black Hispanic Voting Age Population
Source: 2010 Census PL 94-171 Data
Block-level boundaries reflect Census blocks as depicted in the 2014 TIGER Data.

3. Block-Level Estimated Disaggregation of Election Data
Each state's data file contains the following fields:
block: 2010 Census Block geoid
TotPop: Total Population (2010 Census)
VAP: Voting Age Population (2010 Census)
Obama: Block-Level Estimated Barack Obama votes (November 2012)
Romney: Block-Level Estimated Mitt Romney votes (November 2012)

4. Ecological Inference (EI) Estimates, Disaggregated to 2010 Census Block Level
Each state's data file contains the following fields:
geoid: 2010 Census Block geoid
Tpop: Total Population (2010 Census)
Tvap: Voting Age Population (2010 Census)
BRvotes: EI-Based Estimate of Black Votes for Barack Obama (November 2012)
BDvotes: EI-Based Estimate of Black Votes for Mitt Romney (November 2012)
HRvotes: EI-Based Estimate of Hispanic Votes for Barack Obama (November 2012)
HDvotes: EI-Based Estimate of Hispanic Votes for Mitt Romney (November 2012)
ORvotes: EI-Based Estimate of Other Votes for Barack Obama (November 2012)
ODvotes: EI-Based Estimate of Other Votes for Mitt Romney (November 2012)

The R code used to produce the Ecological Inference (EI) estimates appears in: "R/ei_calculations.R"

5. Seed Plans for Each State

Within each state's folder:
- "blind.shp" refers to the seed plan for the race-blind MCMC algorithm.
- "race.shp" refers to the seed plan for the race-conscious MCMC algorithm.

Each seed plan is also depicted in the form of an assignment file ("blind.txt" and "race.txt") in which district assignments are listed in a single row in the same order (from left to right) as within each state's base polygon shapefile
- "blind.txt" refers to the seed plan for the race-blind MCMC Algorithm. "race.txt" refers to the seed plan for the race-conscious MCMC Algorithm.


6. Base Polygons for MCMC Algorithm for Each State

Within each state's folder:
-"base.shp" is a shapefile containing the base polygons for the MCMC algorithm.
-"touches_java.txt" is the adjacency matrix for the base polygons, used for determining contiguity.
-"lengths_java.txt" and "coordinates_java.txt" contain geographic data for the base polygons.
-"votes_java2.txt" contains geographic data, population counts, and racial breakdowns of election data (calculated from the EI estimates) for the base polygons. The relevant data fields in this file are:

cty: Polygon's County FIPS code (This is the data used for calculating each districting plan's split counties)
mcd: Polygon's Municipality FIPS code (This is the data used for calculating each districting plan's split municipalities)
Tpop: Polygon's Total Population (2010 Census)
Tvap: Polygon's Voting Age Population (2010 Census)
BRvotes: Polygon's EI-Based Estimate of Black Votes for Barack Obama (November 2012)
BDvotes: Polygon's EI-Based Estimate of Black Votes for Mitt Romney (November 2012)
HRvotes: Polygon's EI-Based Estimate of Hispanic Votes for Barack Obama (November 2012)
HDvotes: Polygon's EI-Based Estimate of Hispanic Votes for Mitt Romney (November 2012)
ORvotes: Polygon's EI-Based Estimate of Other Votes for Barack Obama (November 2012)
ODvotes: Polygon's EI-Based Estimate of Other Votes for Mitt Romney (November 2012)


The R Code used to produce these base polygon files for each state appears in "R_base/create_base".

Parameters for MCMC Algorithm Constraints: Within each state's folder, the "sim.../enact.txt" file contains the specific parameters used to contrain the MCMC algorithm for each state. Specifically, this file reports the following parameters for each state:

idealpop: Ideal population for a single district
pop: Total population of the state
sldl: Number of single-member districts in a valid districting plan
maxdev: Maximum allowable population deviation (based on the maximum population deviation exhibited by the state's enacted plan).
mcds: Total number of municipalities in the state.
spmcds: Maximum allowable number of municipalities split into two or more districts (based on the number of municipalities split into mulitple districts by the state's enacted plan).
ctys: Total number of counties in the state.
spctys: Maximum allowable number of counties split into two or more districts (based on the number of counties split into mulitple districts by the state's enacted plan).
BlackDists: Exact required number of Black opportunity districts for race-conscious MCMC algorithm (based on the number of Black opportunity districts in the state's enacted plan).
HispDists: Exact required number of Hispanic opportunity districts for race-conscious MCMC algorithm (based on the number of Hispanic opportunity districts in the state's enacted plan).
polsby: Minimum required Polsby-Popper compactness score (based on Polsby-Popper score of the state's enacted plan).


7. Computer-Drawn Redistricting Plans and MCMC Algorithm Computer Code

Redistricting plans are numbered from 1 to 1,000. Each redistricting plan is depicted in the form of both an ESRI shapefile (eg, "1.shp") and an assignment file (eg, "1.txt") in which district assignments are listed in a single row in the same order (from left to right) as within each state's base polygon shapefile.

For each state, the "Race_Blind_Redistricting_Plan_shapefiles" and "Race_Conscious_Redistricting_Plan_shapefiles" folders contain redistricting plan shapefiles. The "blind_sp" and "race_sp" folders contain assignment files (eg, "1.txt") in which district assignments are listed in a single row in the same order (from left to right) as within each state's base polygon shapefile.

"Code/MCMC_algorithm_vM20" contains Java code used for the MCMC algorithm to produce redistricting plans.

"Code/MCMC_calculations" contains R code used to analyze the MCMC algorithm-produced redistricting plans.

For each state, the "sim..." folder contains data analyzing the racial and partisan characteristics of the redistricting plans, as calculated by the R code in "MCMC_calculations".

The R code used to produce results appearing in the article and in the appendix is in: R.zip

Split Counties and Split Municipalities: Lines 55-66 of 1_Figures.R calculate split counties and split municipalities. I use the "cty" and "mcd" of each state's "votes_java2.txt" file in the base polygons folder to identify the county and the municipality, respectively, that each polygon lies within. These identifiers are used to calculate the number of split counties and split municipalities in each districting plan. A county (municipality) that is split into two or more districts is always counted as one split county (municipality). If a county (municipality) lies within a single district, then it is not considered to be split.

Home