Content uploaded by Margo Anderson
Author content
All content in this area was uploaded by Margo Anderson on Jun 03, 2021
Content may be subject to copyright.
1
May 31, 2021
Redistricting Criteria and Testing the Adequacy of the the 2020 Census
Disclosure Avoidance System
Margo Anderson
Distinguished Professor, Emerita, History & Urban Studies
University of Wisconsin – Milwaukee
Milwaukee, WI 53201
margo@uwm.edu
This is an explication of the steps required to align redistricting criteria for the PL
94-171 files with the Disclosure Avoidance System [DAS] planned for the 2020
Census, as of May 2021. It is also, by implication, an explication of the steps that
should have been undertaken between 2017 and 2021 with the redistricting
community, particularly the committees organized by the National Council of State
Legislatures [NCSL], as the Census Bureau built the new DAS.
1
It draws from recent
reports submitted to the Census Bureau on the utility of the April 28, 2021
Demonstration Product [select citations included below]. As is clear from these
reports, the steps were not taken, resulting in a flawed Disclosure Avoidance
System, and hence flawed PL 94-171 files.
Criteria for drawing districts and their impact on the April 28, 2021 DAS’d PL
94-171 file
Initial Criteria:
1. Equal Population Size
2. Compactness
3. Contiguity
4. Communities of Interest
5. Respect for Political Boundaries
Voting Rights Criteria:
1. Can a majority/minority district be drawn that meets criteria 1-5 above?
2. Does the minority vote pattern reveal a preference for one candidate?
3. Does the majority vote systematically in such a way that the preferred minority
candidate is not elected?
1
For further analysis of this point see, “Margo Anderson’s Motion for Leave to File
Amicus Curiae Brief in Support of Plaintiffs,” Alabama vs. Department of Commerce,
Civil Action No. 3:21-CV-211-RAH-KFP, available at
https://www.brennancenter.org/sites/default/files/2021-
04/Amicus%20Brief_margoanderson_2021-04-16.pdf. Accessed May 31, 2021.
2
[Criteria 2 and 3 are used to test “racially polarized voting,” or RPV].
Evaluation:
*The current DAS’d PL 94-171 file leaves census geography, i.e., the “spine,” and “off
spine” shapes, in tact. Line drawers can align their line drawing to the tabulations in
the DAS’d PL 94-171 file. Main criteria 2, 3, and 5 [in black above] are met.
* The Census Bureau claims that their metrics demonstrate that initial criteria 1 and
4 [in red] are also met. Analyses of the April 28, 2021 DAS’d PL 94-171 tabulations,
however, indicate that the tabulated numbers do not meet initial redistricting
criteria 1 and 4. Districts deviate from the allowable equal population size deviation
standard. Characteristics of block populations produce anomalous results on
characteristics of the block populations: children in blocks with no adults; minority
population in blocks with no minority population in the official 2010 tabulations; a
questionable number of blocks with households with 15 or more people. These
anomalies also affect the first voting rights criterion, since the noise injection moves
minority majority districts above and below the 50% districting threshold used to
draw such districts.
*Voting Rights Criteria 2 and 3 [in blue] require the redistricter to examine
additional data files of small area voting patterns to evaluate racially polarized
voting. To date, the Census Bureau has not provided RPV evaluations of the DAS’s
PL 94-171 tabulations at the voting precinct or ward level in existing majority
minority districts.
Select Listing of Technical Reports with Critiques of the DAS Relevant for
Redistricting:
Andrew Beveridge, “As Census Prepares for First 2020 Data Release, Questions
Arise About Accuracy of Next Data Release,” April 23, 2021,
https://www.socialexplorer.com/blog/post/as-census-prepares-for-first-2020-
data-release-questions-arise-about-accuracy-of-next-data-release-11427
Andrew Beveridge, Does the Quality of the Census April 28, 2021, Census
Demonstration Product (with an Epsilon of 12.3) Mean that Such a Product Would
Be “Fit for Use” for Redistricting? An Assessment Based Upon “Traditional
Redistricting Principles,” May 28, 2021, in possession of author.
Thomas M. Bryan, “Declaration,” filed May 17, 2021,
https://vhdshf2oms2wcnsvk7sdv3so.blob.core.windows.net/thearp-
media/documents/Decl._of_Thomas_M._Bryan_5.17.21.pdf, and prior filings at the
American Redistricting Project site.
3
Christopher Kenny, et al., “The Impact of the U.S. Census Disclosure Avoidance
System on Redistricting and Voting Rights Analysis,” May 28, 2021, https://alarm-
redist.github.io/posts/2021-05-28-census-das/Harvard-DAS-Evaluation.pdf and
https://alarm-redist.github.io/posts/2021-05-28-census-das/.
William P. O’Hare, “Analysis of Census Bureau’s April 2021 Differential Privacy
Demonstration Product: Implications for Data on Children,” May 2021,
https://secureservercdn.net/198.71.189.253/r75.9ab.myftpupload.com/wp-
content/uploads/2021/05/Implications-of-Differential-Privacy-for-kids-Final-5.21-
1.pdf and https://countallkids.org/report-analysis-of-census-bureaus-april-2021-
differential-privacy-demonstration-product-implications-for-data-on-children/
David Swanson, et al., “The Census: Protecting Privacy versus Creating Useless
Data,” May 18, 2021 https://countallkids.org/report-analysis-of-census-bureaus-
april-2021-differential-privacy-demonstration-product-implications-for-data-on-
children/
David Van Riper, Jonathan Schroeder, and Steven Ruggles, IPUMS-NHGIS
Institute for Social Research and Data Innovation, University of Minnesota,
Feedback on the April 2021 Census Demonstration Files,” in possession of author.
Additional Available Source Compilations Critiquing the DAS:
Southern Demographic Association, “Articles on Disclosure Avoidance,” continually
updated http://sda-demography.org/news/10524019
The American Redistricting Project, Alabama v. United States Department of
Commerce (2021), continually updated, https://thearp.org/litigation/alabama-v-
united-states-department-commerce-2021/