Computational Redistricting of Iowa’s Congressional Districts
Stefanie Wang, Nathaniel Merrill · 2025
Read Computational Redistricting of Iowa’s Congressional Districts by Stefanie Wang, Nathaniel Merrill (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-05848.txt.
Introduction
Stefanie Wang and Nathaniel Merrill redraw Iowa's congressional districts by computer in this 2025 paper. Iowa has four districts and strict rules that protect county lines. Fair map tests compare an enacted map against many neutral random ones.
The authors extend an algorithm made for states with few districts or strong county rules. They generate non-partisan random maps for Iowa as a yardstick. Against that yardstick, the evidence suggests the first map proposed by the state bureau was rejected over partisan bias. They also scan Iowa's 2024 election results for signs of gerrymandering. The work shows how simulation can police map fairness.
Catalog record
| CREATOR | Stefanie Wang, Nathaniel Merrill |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 23 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-05848
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Computational Redistricting of Iowa’s Congressional Districts Stefanie Wang, Nathaniel Merrill Abstract This article expands on the redistricting algorithm proposed by Chen and Rodden [ 5 ] for states with fewer than eight congressional districts, populations highly concentrated in urban areas, or state laws that require preservation of county lines. We used the updated algorithm to redistrict Iowa’s four congressional districts. Non-partisan, randomly drawn maps were used to evaluate the fairness of the enacted congressional map. Notably, the evidence suggests that the first proposed map drawn by the Iowa Legislative Bureau was rejected due to partisan bias. This article also analyzes Iowa’s 2024 election results for partisan gerrymandering. 1 Introduction In this paper, we expand on existing literature and propose methodology for evaluating whether a particular proposed legislative map is “fair,” using a case study of Congressional redistricting in the state of Iowa. With every decennial census, seats in the House of Representatives are reallocated from state to state based on updated population counts. Each state with at least two seats must prepare and approve a district map based on updated populations. In addition to the U.S. Constitution and federal law, congressional redistricting must comply with state redistricting laws and processes [ 16 ] . Attempts at enacting (and restricting) gerrymanders in the United States date back more than 200 years. There is not a specific legal standard for what a gerrymander is or isn’t, and it has generally proved difficult for plaintiffs to persuade courts that a particular plan is illegal [ 15 ] . While the general public may employ the “eye test” to mock districts with tortuous bounds, districts like that may be drawn specifically to comply with directives to connect communities of interest [ 5 ] . On the other hand, district maps with visually appealing districts may still be gerrymandered in favor of one political party through the use of packing and cracking. Packing refers to fitting as many voters from one party into as few districts as possible. The preferred candidate of the packed party is likely to win in the packed districts while the party’s total voting strength is weakened across the whole state. Cracking spreads voters from a political party into as many districts as possible in order to dilute voting strength. In this paper, we add to the algorithm in [ 5 ] to create 215 Congressional district maps for the state of Iowa based on different random initial seeds. The algorithm forces output maps to obey necessary requirements with regard to population size and contiguity. The algorithm aims to optimize compactness and is run without without regard to partisanship. We then can compare the actual enacted 2021 map to our distribution of random maps using a variety of metrics: does the actual map fall within the range of the random maps, or is it an outlier? An analysis of the vote share, efficiency gap, and compactness scores of the enacted map against the simulated maps suggest evidence of partisan bias during the redistricting process. While results from the enacted map are not outliers, it performs worse in the vote share in comparison with simulated maps. 2 Redistricting Guidelines and Procedure 2.1 Iowa Standards for Redistricting In Iowa, the redistricting process has been led by the nonpartisan Legislative Service Bureau since 1980 [ 7 ] . They are in charge of preparing an initial proposal and a potential second proposal, each of which the legislature can only approve or reject. If a third proposal is needed, then the legislature can begin to edit maps directly. 2.1.1 Equal apportionment The population in Iowa from the 2020 census was 3,190,369. The state was allocated four seats in the House of Representatives, so four districts need to be formed, each with an ideal population of 797,592.25. The Iowa Code, Section 42.4(1) requires districts to have a population deviation not to exceed 1% of the ideal population [ 8 ] . This gives a population range from 789,616 to 805,568 for each district. 2.1.2 Political boundaries A 1968 amendment to the Iowa Constitution [ 9 ] requires that individual counties not be divided into multiple congressional districts. This is relatively unique among US states: many states have at least one populous county which has more people than any single Congressional district. Even neighboring Nebraska, with three Congressional districts and less than 600,000 people in its largest county, splits two different counties between districts [ 12 ] . 2.1.3 Contiguity and compactness The redistricting amendment [ 9 ] further requires that each district be compact and contiguous. The Article requires excluding maps where a district is only connected by a pair of counties which meet only at adjoining corners. The Iowa Code specifies that “compactness” includes districts that are square, rectangular, or hexagonal in shape and excludes irregular shapes, barring any geographic or political boundaries. The code specifies two metrics, length-width compactness and perimeter compactness, to compare the relative compactness of two districts. The length-width compactness is computed by calculating the difference between the length and width, with a difference of zero signaling optimal compactness. The greater the perimeter compactness, the smaller the district perimeter [ 8 ] . These metrics for compactness are somewhat dated. We iterate the shortcomings in Iowa’s length-width compactness measure noted in [ 13 ] : Deviations that do not alter these four critical points have absolutely no effect on the district’s score. The Iowa measure is thus blind to any connivance that occurs within the four “walls” to the north, south, east, and west, although this may be very important to partisans. Polsby and Popper further remark that a district of the same shape may yield different length-width compact scores when rotated. In this paper, we will use more sophisticated metrics for evaluating the compactness of districts, introduced in Section 3.1.1 2.2 Rejection of Proposal 1 in 2021 In the 2021 redistricting process, the first map was rejected by the legislature and the second map was enacted [ 3 ] . While the initial set of maps was supported by the Democratic minority in the state legislature, the Republican majority claimed that population in the districts was unbalanced and that the shapes were insufficiently compact [ 2 ] . The legislature is not supposed to consider partisan balance or the location of incumbents when selecting maps, but news outlets also noted at the time that the first map created a likely Democratic First district. On the other hand, the second map split Linn and Johnson counties, both Democratic-leaning, between the First and Second districts, giving Republicans an advantage in both [ 3 , 6 ] . 2.3 The Redistricting Algorithm In this paper, we began by following the algorithm proposed by Chen and Rodden in [ 5 ] to generate 215 potential maps of Iowa Congressional Districts. A key component of this algorithm is to identify and work with the smallest geographical units of each congressional district. In Iowa, this is the 99 counties. The algorithm proceeds as follows: 1. Identify each county as a separate district. There are now i=99 districts. 2. Randomly select one of the i districts. Join it to its nearest neighbor by calculating the distance between