Redistricting, the shape of things to come

On the surface it seems like a simple idea: Every 10 years, find everyone in the United States at their place of residence, whether that’s an apartment, a nursing home, houseboat, or a tent in the woods. Count them. Everyone. Women, men, children. While you’re at it, ask them some other questions too about themselves, their work, how old they are, and a few other things. Do this as carefully and as thoroughly as you can. It won’t be easy. Then, take a total of 435 possible seats in the U. S. House of Representatives, and divide them up — reapportion them — based on the newly-counted total population in the state. A state that has lost substantial population since its residents were last counted 10 years previous might end up with one of its seats removed, and states that have become more populous will gain those. Once reapportionment of the congressional seats is complete, divide the area of the state into districts of equal population and allocate one seat to each district. When those districts are being drawn, make sure that each one is designed in a way to be fair. Depending on the state, that might mean that the district’s shape must be compact, contiguous, not cross county boundaries or other communities of interest, and not be completely different from previous years and/or influenced by the results of previous elections and/or socio-economic or demographic variables. Have the powers that be approve any changes to the districts, and voila, look forward to the next election.

Of course, in reality there is no single step of this process that is free from contention and acrimony. Before the 1960s, many congressional districts hadn’t ever been changed or updated, so with population changes over the decades they became increasingly out of balance. Now that we have greater capacity to visualize and compare our shifting population and its characteristics, the inconsistencies and injustices that undermine the important “one person, one vote” intent are no longer overlooked. The political rancor has reached the Supreme Court numerous times, currently with North Carolina and Maryland at the forefront. Missouri is entering the spotlight, and Utah is on deck as well.

If you’re reading this article, chances are you appreciate the central roles that geospatial data and digital technologies play in these processes. Likely at some point you’ve used decennial Census data in the maps you’ve made, and you’re hopeful that the upcoming 2020 results will be as comprehensive and as accurate as possible. You know how (relatively) simple it is to change the perimeter of a polygon on a map that encloses a different set of people within that space, and what the implications are for aggregating data, for zoning and the Modifiable Areal Unit Problem.

These aren’t easy matters to understand, and most people don’t. It is simple to look at a ludicrously-shaped district and highlight its violation of the principles of compactness or contiguity, but in fact, unusual shapes might be necessary to capture a heterogenous population mix, or accommodate physical features that influence how people regularly interact on the ground. It turns out that people tend to live near people like themselves. Our propensity for spatial autocorrelation in our neighborhoods makes it extremely difficult to create districts that meet an unordered list of criteria, and voter turnout doesn’t happen uniformly. Gerrymandering is an easy target to attack but fair representation is an elusive goal under the most common scenario for electing candidates: single-winner districts.

Instead of hand-wringing, chest-pounding, or head-in-sand-burying, taking baby steps towards incremental knowledge-building may be the only way to nudge the needle, as they say, towards solutions to these impasses.

With that in mind, how are these complex spatial and geospatial topics being explained to people? What tools are available to demonstrate the workflow and decision steps? Several games exist that allow drawing of district boundaries around abstract shapes or fictional landscapes. DISTRICT begins with a 2-person country and, forcing the criterion of equal population per district as the game advances, illustrates effectively how the single-winner policy results in the number of votes cast being inconsistent with the winning results. The ReDISTRICTING GAME is a comprehensive website that includes a district-drawing game with play possible at different levels and with varying objectives, along with pointers to many relevant resources. Unfortunately, like many great ideas on the web, it is no longer being maintained and users are likely to encounter play mishaps and dead links.

A more recent project is GeoCivics, led by Rebecca Theobald of University of Colorado Colorado Springs and funded to date with a seed grant from the National Center for Research in Geography Education. Theobald believes strongly that local and regional geographic knowledge increases the sensitivity that citizens will have towards divisive and poorly designed districts. Her deployment of National Geographic’s super-sized maps is one strategy to allow both youth and adults the chance to visualize the representation of their state and its landscapes where district boundaries cross. When adults find themselves now living in new or unfamiliar places, or they have long forgotten their fourth-grade lessons on state-wide geography and history, they lack an understanding of how their state’s population patterns have shifted over time. Notes Theobald, “Looking at information like population numbers in isolation gives no sense of the people and places on the ground.”

Another component of the GeoCivics project are individual, state-wide ArcGIS Online Redistricting Exercises with instructions for building population-equal districts comprised of contiguous counties, created by Anita Palmer of GISetc. Out-of-the-box AGO is slightly clunky and limited for this, but within a few minutes I had followed the provided instructions and created two quasi-equal-population districts for my state of New York. Manhattan and six of its neighboring counties became District A and the remaining 55 counties of the state were grouped as District B. The story of New York politics in one map!

How do the real legislators and their professional staff accomplish these tasks? They use GIS. Since the 1990s the digital tools have replaced the use of paper maps and colored pencils in this realm; even shapefiles are defined and described in “Into the Thicket: A Redistricting Starter Kit for Legislative Staff.” Software companies have designed customized functionality to support the iterative process, such as Caliper’s Maptitude for Redistricting and Esri’s Redistricting tool. Smaller companies such as ZillionInfo has produced iRedistrict and is targeting other boundary-needing audiences such as schools and police departments as well. The open-source Auto-Redistrict and DistrictBuilder further diversify the list of available products.

The software platforms aim to streamline the workflows by making the data and its visualization easier to access and manipulate, but oversimplifying is equally unhelpful in the end. These dynamic situations require prowess with mathematics, awareness of physical and socio-economic geographies, cognizance of Census data’s quirkiness, sensitivity to history, and savviness with the political machine. Add a bunch of lawyers into the mix and, well, here we are.

Additional Resources