To: Mike Dean, Common Cause Minnesota

Laura Fredrick Wang, League of Women Voter Minnesota

From: Dr. Michael McDonald, Associate Professor at George Mason University
RE: Common Cause Minnesota Memo on Redistricting Plan C1101_0

An analysis of proposed congressional redistricting plan C1101_0 by Dr. Michael McDonald at

George Mason University for Common Cause Minnesota finds that the proposed redistricting
plan is biased towards the Republican Party. Using a metric commonly used by practitioners,
academics, and the courts to assess the partisan bias of redistricting plans, Dr. McDonald finds
that the DFL would need to win 53.2% of the vote within the congressional districts to win a
majority of the congressional seats.

Dr. McDonald and his colleagues are leading the Public Mapping Project, designed to bring
more public participation and transparency to the redistricting process. More information about
the Project is available at www.publicmapping.org.

Methodology
The analysis used here is consistent with the approach used by practitioners, academics, and the
courts. The approach has been used for the past 40 years by the tie-breakers appointed to the
New Jersey Legislative Redistricting Commission by the Chief Justice of the New Jersey
Supreme Court.1 It has been used to assess partisan gerrymandering claims by Democratic and
Republican expert witnesses in the 2003 Texas re-redistricting court case and a 2002 seminal
partisan gerrymandering case in Pennsylvania.2 The methodology has been discussed extensively
in peer-reviewed academic journals.3

The idea is to measure the underlying partisan strength of districts using statewide elections. The
underlying partisan strength is one among many factors that determine election outcomes,
including the power of incumbency, national and state tides, and candidates’ campaign
resources, among many other factors. The method is as follows:

1. Calculate the vote for a statewide office(s) within the proposed districts. I describe the
methodology used to calculate these statistics in the appendix.

1
Donald E. Stokes. 1993. “Legislative Redistricting by the New Jersey Plan.” New Brunswick, New Jersey: Fund
for New Jersey.
2
Examples of expert witnesses representing Democrats, Republicans, and bipartisan commissions who have used
principal components of the method include Keith Gaddie in LULAC v. Perry, 126 S. Ct. 2594 (2006); Gary King in
Voinovich v. Quilter, 507 U.S. 146 (1993); Jonathan Katz in O'Lear v. Miller No. 222 F. Supp. 2d 850 - 2002;
Allan Lichtman in Vieth v. Commonwealth of Pennsylvania, 188 F. Supp. 2d 532 (MD Pa. 2002); and Michael
McDonald in In Re 2001 Redistricting Cases (Case No. S-10504).
33
Bernard Grofman and Gary King. 2007. “The Future of Partisan Symmetry as a Judicial Test for Partisan
Gerrymandering after LULAC v. Perry.” Election Law Journal 6(1): 10; J. Morgan Kousser. 1996. “Estimating the
Partisan Consequences of Redistricting Plans-Simply,” Legislative Studies Quarterly 21(4): 530; Janet Campagna
and Bernard Grofman 1990. “Party Control and Partisan Bias in 1980s Congressional Redistricting.” The Journal of
Politics 52(4): 1247.

1
 2. Calculate the DFL percentage of the DFL plus Republican vote, what is often referred to
as the two-party vote.

3. Apply a uniform vote swing by adding and subtracting the same percentage increment to
each district. For example, by adding and subtracting 1% to the two-party vote share in
each district.

4. Note the percentage of districts that a party constitutes a majority for a given vote share.
Of most interest is the percentage of seats a party is expected to win with 50% of the two-
party vote.

In Figure 1, I plot the seats to votes relationship for the DFL two-party share of the 2008
presidential votes tabulated within the proposed C1101_0 districts. In a fair plan, a party should
win 50% of the seats with 50% of the vote. This partisan fairness analysis indicates the DFL
needs 53.2% of the two-party vote to win a majority of the House seats.

Figure 1. DFL Seats to Votes Relationship for Plan C1101_0

Using the 2008 Presidential Vote

2
Appendix: Methodology to Calculate 2008 Presidential Vote within Proposed
Districts
These data are 2008 presidential election data obtained from The Minnesota Legislative
Coordinating Commission’s Geographic Information Services available at:
http://www.gis.leg.mn/html/download.html.

The LCC-GIS requests the following disclaimer accompany these data:

“LCC-GIS makes no representation or warranties, express or implied, with respect to the reuse of
data provided herewith, regardless of its format or the means of its transmission. There is no
guarantee or representation to the user as to the accuracy, currency, suitability, or reliability of
this data for any purpose. The user accepts the data 'as is', and assumes all risks associated with
its use. By accepting this data, the user agrees not to transmit this data or provide access to it or
any part of it to another party unless the user shall include with the data a copy of this
disclaimer.”

I perform further database work to merge these data with the 2010 census and to disaggregate
these data to the census block level, so that they may be re-aggregated into proposed redistricting
plans, following these steps: 4

1. I correspond the 2008 precincts obtained from LCC-GIS as best as possible to their
equivalents in the 2010 census geography, what are known as voting tabulation districts.
In a few instances for small-population geographies, I did not make a correspondence.

2. I apportion the precinct level election results into the smaller-sized census blocks by the
census block’s share of the precinct’s voting-age population.

3. Using what is known as a block assignment file, which lists the district assigned to each
census block, I sum the election results into the proposed districts.

The 2008 presidential vote tabulated within the proposed C1101_0 districts is presented in Table
1. I calculate the DFL percent of the two-party vote, which is the DFL vote divided by the DFL
vote plus the Republican vote. The average DFL vote share among these districts is 55.2%. I
calculate a normalized two-party vote to simulate a toss-up election, where the political parties
have equal vote shares at 50%, by subtracting 5.2 percentage points from each district’s two-
party vote share.

4
The census block level database is available at: http://www.publicmapping.org/resources/data

3
Table 1. 2008 Presidential Vote within Proposed Plan C1101_0 Districts
All
DFL Republic Others Normali
’08 an ’08 ’08 zed
Preside Preside Preside Two-
Distri ntial ntial ntial Two-Party Party
ct Vote Vote Vote Vote Vote
1 176,367 162,955 7,444 52.0% 46.8%
2 176,277 176,680 6,001 49.9% 44.8%
3 195,837 192,146 5,293 50.5% 45.3%
4 233,342 125,929 6,137 64.9% 59.8%
5 266,692 86,448 6,108 75.5% 70.3%
6 163,087 194,382 6,180 45.6% 40.4%
7 162,773 191,904 8,138 45.9% 40.7%
8 202,975 152,530 7,058 57.1% 51.9%