DRA is the most useful free online “expert” to help us evaluate and analyze the maps. DRA 2020 not only provides excellent tools for drawing districting maps, it also provides an array of analytics that are second to none in the realm of map-drawing programs easily accessible to the general public.

Whether one agrees with the Governor’s Assembly map or not, one should be sympathetic with the predicament he faced. Black Leaders Organizing for Communities (BLOC), Voces de la Frontera, the League of Women Voters of Wisconsin and others pushed for a map containing seven Black Voting Age Population (BVAP) majority districts in the Milwaukee area. The Legislature wanted a map that changed the current six BVAP map the least. The Wisconsin Supreme Court (SCOW) ruled for a least-change map that did not violate the Voting Rights Act (VRA) and left it up to the parties involved to solve its problem. The challenge for Governor Evers was to draw a map with superior core retention yet be reasonably bullet-proof against any VRA complaint. To fully understand the situation, it is helpful to review the chronology of events leading up to the presentation of the maps to SCOW on December 15, 2021.

BLOC v Spindell, filed on Sept. 7, argued that the current six Black Voting Age Population (BVAP) majority district Assembly map violated section 2 of the Voting Rights Act of 1965. In promoting a seventh BVAP majority district BLOC’s petition revealed to the public Ken Mayer’s seven BVAP majority district map as an illustration. Indeed, BLOC’s December 15 brief to the court presented an Assembly map having seven BVAP majority districts in the narrow 50.2% to 52.3% range. Its core retention according to DRA is 84.3%.

On October 28 Speaker Vos presented the Legislature’s Assembly map claiming, “Our goal from start to finish was to produce a ‘least-change’ map that prioritized core retention while adjusting for population change.” Feeling confident that its map would win the SCOW contest, the Legislature chose not to make any changes to this map in its December brief. DRA computes a core retention for this map at 84.5%, slightly better than BLOC’s percentage, but beatable. The VRA seemed to be secondary in the Legislature’s priority list as its map contained only five BVAP majority districts, District 10 having only a 47.2% BVAP. BVAP percentages for the other five districts according to DRA were: 11 (73.3), 12 (57.0), 16 (54.1), 17 (61.8), 18 (52.6).

There is some debate on what percentage constitutes packing which is disallowed by the VRA. BLOC posits that anything over 55% constitutes packing, others would say that up to 60% is fine. The BVAP of District 11 seems inordinately high, especially since it is neighbor to sub-majority District 10. The Legislature stated that its map was not race-motivated and cites the BVAP percentages as proof.

On November 11 the Legislature’s maps were officially adopted by the Republican-dominated Assembly and Senate. On the same day the People’s Maps Commission (PMC) maps were soundly defeated my members from both parties. The Democratic opposition was led by Milwaukee Senator Lena Taylor and Representative Sylvia Ortiz-Velez stating they did not believe that the maps complied with the VRA. The BVAP percentages for the PMC maps were: 10 (56.3), 11 (47.4), 12 (58.0), 16 (45.2),

17 (44.9), and 18 (45.7). The PMC said very little in its report to establish that its maps complied with the VRA, a critical mistake given that four of their districts were sub-majority BVAP. The Governor most assuredly took note. Since the PMC map was drawn before the “least-change” criterion was implemented the Governor had an easy way out to distance this troubled map from his.

After soliciting feedback from the various parties involved, SCOW in its November 30 order decided that “least-change” would be the key criterion determining the winner of the map contest. The Court also reserved the right to modify any map to comply with federal and state laws. Any discussion of partisanship bearing on the map-drawing was forbidden. The Governor’s roadmap to success was determined. Draw a map that advantaged the Democrats a bit more than the current map and considerably more than the Legislature’s map which tilted in favor of the Republicans severely. Make sure to have seven BVAP majority districts to protect against a VRA challenge. Since considerable changes in the Milwaukee would hurt core retention, be extra diligent in the rest of the state to minimize change. Like the PMC, the Governor skimped on VRA analysis, relying too much on BLOC to carry the water for a seven district BVAP majority map. The BVAP percentages for the Governor’s map ranged from 50.1% to 51.4%. Like the BLOC map, District 14 was considerably revamped to be the seventh district. Core retention came in at 86.3% easily outpacing the Legislature’s map and winning the contest and SCOW’s endorsement for the moment.

Wisconsin is a purple state. It went for Donald Trump in 2016 and Joe Biden in 2020, each win by less than a 1% difference in votes. It has a Democratic senator, Tammy Baldwin and a Republican one, Ron Johnson. However, in state elections for the past decade our state has been decidedly Red. Why the difference? Democrats say, “Give us fair maps and we will win our fair share of elections.” Republicans counter with, “You’re losing because our local candidates are better.” How much of a factor does partisanship play in the outcome of a regional election in Wisconsin? An analysis of the 2020 Assembly elections provides a clue.

Dave’s Redistricting App (DRA) uses a composite of six recent statewide elections (2016 Pres, 2020 Pres, 2016 Sen, 2018 Sen, 2018 Gov, 2018 AG) to determine the partisan leanings of each Assembly district. Note that no election data from actual Assembly races is used. Below on the left is a stylized Assembly map of Wisconsin (99 districts total) with the nine principal districts of Milwaukee attached on the east. The map is colored with respect to how DRA classified each district’s partisan lean. The map on the right shows the actual 2020 election outcomes.

Of the 95 districts that leaned toward one party or the other, only one district (51) had a 2020 election result that was NOT predicted by the party lean. All four of the extremely competitive districts (13, 14, 23, and 94) went to the Democrats in 2020. The remarkably predictive capacity of the partisan lean map for the 2020 elections seems to indicate that the Democrats have a point. The maps matter.

What changes might we see with the new 2022 Assembly map? Checkout the side-by-side comparison of the 2020 DRA Partisan Lean with that of the 2022 map.

Here are a few observations.
1) Two competitive districts in 2020 (13 and 14) are split between the parties in 2022, District 13 going Republican and District 14 solidly Democratic.
2) Six Republican districts in 2020 (21, 42, 55, 85, 88, 96) join 23 and 94 as competitive in 2022.
3) District 37, Republican leaning in 2020, swings to a deeply Democratic district in 2022.

In terms of partisan lean the 2022 Assembly map, though still heavily tilted Republican, is a significant improvement for the Democrats.

Four western Wisconsin Assembly districts played a huge role in the Wisconsin Supreme Court (SCOW) decision to select Governor Evers’ map (GUV) over the one proffered by the Legislature (SB 621). Assembly District 30 and its three surrounding districts, 28 to the north, 29 to the east and 93 to the south, contributed over half of the core retention differential that provided the decisive edge in the SCOW decision. Core retention is calculated as that fraction of a district in the enacted map that is contained in the same numbered district of the proposed map. The higher the proportion, the more people stay in the same district. (See table below generated from statistics generated by Dave’s Redistricting App.) The accumulated fractional difference of 0.9141 multiplied by the average district population of 59,533 amounts to over 54,400 more people remaining in their current district under GUV compared to SB 621.

Justice Brian Hagedorn writes in the majority SCOW opinion that the primary reason of selecting Gov. Evers’ Assembly map is because it retains 96,178 more people in their districts than SB 621. Thus, the four western Wisconsin districts while comprising just 4% of the Wisconsin population contribute to over 56% of the core retention differential.

The table summarizes some of the metrics provided by Dave’s Redistricting App for the six maps contesting for the prize of best Assembly map that changes Act 43 the least. They are sorted by the last column which gives the average percentage of population retained by Act 43 districts in the same district in the new map. By this metric, GUV changes Act 43 the least and MathSci changes it the most. On the other hand, looking at the yellow columns which hold compactness ratings, one could say that Hunter improves on compactness the most and SB 621 the least. All the maps felt obligated to diminish the number of counties that are split (first tan column), but MathSci would claim to do the best job while Bewley the worst. MathSci also does the best in minimizing the number of times counties are split (second tan column) and GUV would be the worst at that objective. Finally, GUV and BLOC manage to increase the number of majority minority VAP districts to 9 (last green column) and MathSci does the poorest.

The Court faces the complex problem of choosing a map that changes Act 43 “the least” and yet being mindful of traditional principles of redistricting, changes some aspects of Act 43 the most. A map that takes on the challenge of increasing Act 43’s compactness should be welcome, but how much different from Act 43 does that make the map? County and municipal splits should be kept at a minimum, but at what cost to core retention? It would be hard to justify selecting a map like SB 621 that has a lower number of majority minority districts over Bewley just because it has a higher core retention percentage. The selection problem is a complex one and depends heavily on which principles the Court wishes to emphasize and which principles it does not. There is no road map for deciding.

What role does the first column of numbers play in the decision? Is the lowest total deviation the best? All things being equal, the answer would be yes, but as we have seen, all things are not equal. How much does the pursuit of total deviations less than 1% restrict the flexibility needed to achieve other desirable map characteristics? SB 621’s 0.76 puts it at 1.78% fewer people (1060) per district retained than Governor Ever’s map. That is over 100,000 more people moved to a different district in the state. MathSci’s 0.74 seems to hurt its ability to lower municipal splits and achieve more majority minority districts. The quest to minimize total deviation can interfere with the map-makers ability to increase compactness, reduce splits, and even construct districts that conform to the Voting Rights Act of 1965. Only a handful of states in 2010 had State House districts with total deviations less than 1%. Perhaps the other states were wise not to succumb to the fallacy of delusive exactness.1

1. https://law.justia.com/cases/federal/district-courts/FSupp/793/859/1369102/

Concerned citizens of WI have filed a Friends of the Courts brief with the WI Supreme Court on January 4. Come hear the Why and the What of that brief from Don Leake, WIMAP team leader, and one of the attorneys who filed the brief pro bono.

Then zoom out with our guest speaker, Mel Barnes, from Law Forward who will give us a big picture update from the WI Supreme Court case and help us understand what to expect from the Federal Courts. The deadline for new maps is March 1, 2022 so new legal developments will happen fast!

Also, we want to hear from you! 2022 will require a brand new approach to Fair Maps and Strengthening Our Democracy. Join us to let us know what is important for you to work on in 2022. What ideas do you have to create change?

We will hear from NSFM’s own Cheryl Maranto, a member of the Wisconsin Map Assessment Project, and one of the 36 Concerned Voters of Wisconsin who filed a “friend of the court” brief in the Wisconsin Supreme Court’s redistricting case on January 4. Cheryl was co-author of the written material that formed the basis of the brief.

Daniel Hessel from the Harvard Election Law Clinic will recap the WI Supreme Court case for us, delving into the most recent pleadings, and telling us what might happen next.

And returning to our meeting is Walter Olsen, the Republican co-chair of Maryland’s Citizen Redistricting Commission. He will tell us how things are going in Maryland, a state gerrymandered by Democrats. No matter who gerrymanders, it’s wrong.

As population moves eastward from Dane County in rebalancing Milwaukee County it is inevitable that some districts change greatly.

Such is the case for SB 621’s AD 31, 33, and 43 which have Act 43 core retentions of 60.7%, 56.4%, and 46.7%. Although the districts might change dramatically, the mapper whose eye is on least change would try to keep municipalities, especially smaller ones, intact as they are communities of interest.

Whitewater (pop. 14,889) is a small city that houses a campus of the UW system, lies in two counties, and is split by SB 621 between AD 31 (white) and AD 33 (blue). Whitewater is not divided by Act 43, completely embedded in AD 43 (pink).

Sometimes it is necessary to split municipalities when rebalancing, but not in this case. Below is a small revision of SB 621’s map that allows Whitewater to stay united as a community in District 31.

Note that this map to the right transfers the Jefferson County portion of Whitewater (pop. 3,168) to District 31 while compensating by reassigning two townships, Johnston (pop. 766) east of Janesville and LaGrange (pop. 2,472) southwest of Whitewater to District 33.

Since Lagrange Township was originally in Act 43’s AD 33 and Johnson Township was originally in Act 43’s AD 31 this results in 2,472 – 766 = 1706 more people being retained within their Act 43 district.

It’s a win-win situation, community of interest preserved and change lessened.

Changes made to Act 43’s ADs 23 and 24 by SB 621 are stunning. Instead of reaching down into Milwaukee County, SB 621’s AD 24 extends a large arm up to include Grafton Village and Township displacing over 20,000 people. SB 621’s AD 23 absorbs most of the people from Milwaukee County that exit AD 24.

Thomas Bryan, the Legislature’s expert quantifies this extreme change on page 57 of his report. Only 54.2% of Act 43’s AD 24 population remains in SB 621’s AD 24. This is the fifth worst retention rate in SB 621. The corresponding changes made to AD 23 are not as remarkable, but poor, nonetheless. Bryan states that 66.2% of Act 43’s AD 23 population stays within SB 621’s AD 23. The surplus populations of Act 43’s AD 23 and 24 are 1,228 and 1,204 respectively, so one could expect a maximum core retention of about 98%.

The third image to the right is an excerpt from a least change map (1.95% total deviation) which shows what a least change revision of ADs 23 and 24 could look like.

View this least change map in DRA.

“Our goal from start to finish was to produce a ‘least-changes’ map that prioritized core retention while adjusting for population change.”
– Speaker of the Wisconsin State Assembly Robin Vos, from testimony on SB 621/AB 624 before the Senate Committee on Government Operations, Legal Review, and Government Operations and the Assembly Committee on State Affairs on October 28, 2021.

If Speaker Vos’s goal was to construct a map that moved as few people as possible, then the Legislative Reference Bureau did not do a great job. Or perhaps the LRB were following additional instructions. How else can it be explained that the number of competitive Assembly districts (winner having at least a 10% voting advantage) decreased from 23 in an unbalanced Act 43 map to just 17 in SB 621?

For example, the district in which I live, 30, has a surplus of 3,202 people (an excess of 5.38%) according to the new Census. If one were trying to minimize the difference between Act 43 while balancing the population, the natural thing to do would be to subtract some smaller precincts from the district. In other words, the new District 30 should be contained entirely within the old District 30 and the core retention should be 100%. SB 621 retains only 80.54%.

Why does this happen? Perhaps the real priority was reducing competitiveness in the district. Using a composite of voting records from six recent elections, the web app DRA 2020 computes the unbalanced Act 43 District 30 Republican winner would get 50.58% of the vote compared to a 46.03% share for a Democrat, truly competitive. Under SB 621 the split is 54.37/42.48, a definite Republican lean.

SB 621 touts its core retention of 84.17% as proof that it is a “least-change” map, yet there are maps with higher core retentions that do not decrease competitiveness so drastically. WIMAP has constructed a map that has a core retention of over 86% and manages to actually increase the number of competitive districts to 26. How was this map constructed? It was simply a matter of modifying SB 621 so that it looked more like Act 43. The following ten districts are competitive in this map, but are not in SB 621: 13, 14, 15, 23, 24, 30, 42, 68, 82, 92.

Our conclusion: as a “least-change” map, SB 621 is most definitely a wolf in sheep’s clothing.

Arguably the most significant change to the current Assembly map (Act 43) in the Legislature’s proposed map (SB 621) is within the 5th Senate district comprised of Districts 13, 14, and 15. SB 621 transforms the infamous triplet of horizontally oriented Milwaukee County “cracking” finger districts into two squarish districts (13 and 15) linked by a sinuous District 14. In SB 621 District 13 is pretty much the city of Brookfield, its surrounding township, and the village of Elm Grove. District 15 is the square city of New Berlin with a bit of West Allis and a part of Waukesha township, carefully avoiding the city of Waukesha. District 14 winds its way through Wauwatosa and West Allis in Milwaukee County to connect the other two districts.

The political impact of this transformation is stunning. In the Act 43 map, Districts 13, 14, and 15 were all very competitive. SB 621 eliminates the competition. DRA 2020 gives the percentages for Democrat and Republican voting over seven recent statewide elections from 2016-2020. See Table 1 on the right.

Note that Senate District 5 goes from an average Republican lean of 49.95% to 53.40%. In the 2018 Senate election Republican Dale Kooyenga defeated Democrat Julie Henszey by a margin of 51.1% to 48.7% proving the competitiveness of the district. The Republican advantage will most certainly increase with SB 621.

In the 2020 Assembly elections: Democrat Sara Rodriguez defeated Republican Rob Hutton 50.9% to 49.0% in District 13, Democrat Robyn Vining defeated Republican Bonnie Lee 54.0% to 45.9% in District 14, and Republican Joe Sanfelippo defeated Democrat Jessica Katzenmeyer 54.7% to 45.1% in District 15. It appears that the proposed map is designed to reverse the current two to one Democratic advantage in favor of the Republicans.

WIMAP identified 13 maps submitted to the Legislature’s Draw Your District (DYD) portal with total population deviations of less than 2%. We imported them, along with SB 621’s Assembly map, into the web app Dave’s Redistricting App 2020 (DRA 2020) to see how the maps are rated on widely accepted measures of: 

Proportionality, 
Competitiveness, 
Minority Opportunity,
Compactness, and 
County Splitting. 

All ratings are from 0 to a top score of 100. The results are sorted below.

It is easy to see that SB 621 RATED THE WORST THREE OF THE FIVE CATEGORIES and was well below average in County Splitting. Compactness of districts is a Wisconsin Constitutional requirement. Maximizing compactness and minimizing county splitting are traditional criteria cited in the Legislative Reference Bureau’s (LRB) 2020 Guide for Redistricting. They are also two of six “plan values” listed in the DYD portal.

Even though it is MADE TO APPEAR in the DYD portal that maximizing Core Retention and minimizing Senate Disenfranchisement are traditional Wisconsin redistricting principles, these two “values” are not listed as traditional principles by the LRB. They seek justification in the NONBINDING RESOLUTION, JR 63, passed by a partisan legislature on September 28, 2021.

DRA 2020 organizes its metrics in five general categories: 

Proportionality reflects how well the results of the Assembly elections reflect those of statewide elections from President 2016, President 2020, U.S. Senate 2016, U.S. Senate 2018, Governor 2018, and Attorney General 2018.

Competitiveness attempts to measure the degree that the map provides competitive districts (typically with the winner getting less than 55% of the vote, but it is more complicated than that).

Minority Opportunity is concerned with the maps’ ability to give minorities an opportunity to elect candidates of their choice (see the Voting Rights Act of 1965).

Compactness (required by the Wisconsin Constitution) measures how square the districts are.

County Splitting measures how well counties are kept intact, not split up.

DRA 2020 is a free web application that can be used to create, view, analyze and share redistricting maps for any state. DRA includes demographic data from the 2020 Census and election data from 2016 to 2020. Official congressional and legislative district maps are included and can be used as a starting point for a new map, or one can be created from scratch. A rich feature set makes it easy to create and modify maps while keeping them within the accepted parameters. DRA 2020 also includes a rich set of analytics, including measures of competitiveness, proportionality, minority opportunities, compactness, and splitting, as well as bias and responsiveness.

Dave Bradlee created the original DRA ten years ago. He has over 40 years of software engineering experience at Data General, Microsoft, and LabKey Software. He helped build compilers for Fortran, Pascal, C, and Visual Basic; worked on Microsoft Visio; was development manager for Microsoft’s Natural Language Group; and developed software for medical researchers. Dave has a BS in Mathematics and Computer Science from Union College (NY) and a PhD in Computer Science from the University of Washington.

Thank you for your interest in WIMAP. We’d love to hear from you.