Double Dare Ya

by Ben Kraft

One of my favorite Splash classes to teach is “Gerrymandering: Theory and Practice”. By now I’ve taught it in a number of different contexts – to high school students at Splash 2014, MIT prefrosh at Firehose 2015, MIT students at Splash For Us 2015, and Spark parents at the Spark Parents’ Program 2015. I still haven’t fully refined the class to my liking, but by now I have a decent sense of how it tends to go, and how different audiences respond to it, so instead of writing a blog post version of the class, I’m writing a blog post about the class. So, while I’ll explain some things along the way where it’s necessary, I don’t expect you to learn anything about gerrymandering from this post.

Why I Teach Gerrymandering

Of all the topics in political science that are accessible to high school students (which is many of them), why gerrymandering? I really like teaching gerrymandering because by high school (and certainly by college or beyond), it’s something that a lot of people know a little bit about, but which has a lot of interesting complexity and subtlety that most people haven’t thought about. News reports often make it seem like gerrymandering, and redistricting in general, as being a simple matter of politicians being evil to a greater or lesser extent, when it’s actually much more interesting that. Similarly, it’s something that mathematicians and computer scientists often see and think is trivial – and there are actually a lot of interesting problems in gerrymandering to which math and CS can be applied that definitely aren’t trivial, which I’m always excited to share. And, of course, it has some other convenient properties – it can be made quite hands-on, and it’s something I know a decent bit about (mostly from various political science classes).

And so, while I obviously want students to come out of the class with a basic understanding of the problems of gerrymandering, I also want them to go beyond that. I want them to end up with a deeper feel for the complexity of the issue, for the philosophical questions about the nature of democracy that underlie redistricting, and for how the various goals of the process can conflict. And, of course, I hope students leave the course with some knowledge they can put to use in making sense of current events – the Shelby County and Arizona Independent Redistricting Commission cases, for example.

The Class

It’s easy to get into the weeds of complexity really quickly, so I like to start off by giving students a blank slate.

Splooshland is open season, and the students get to draw five districts however they please. Then I ask several of them to share what they drew, and, more importantly, why. This is one of the places where different classes have taken things in really different directions. MIT students, of course, were unhappy with the lack of a specified goal and constraints, and often drew fairly arbitrary districts, trying to do things like splitting population equally. Splash students, on the other hand, often didn’t think about such practical constraints, and instead thought about what they wanted the districts to be – some had some urban and some rural districts, while others had all of their districts be a cross-section of the overall population. Parents immediately started asking whether they could just have at-large elections, and why we shouldn’t just have computers draw the districts in the first place.

At this point I have the class vote on the map they like the most, and then critique it, which usually gets an interesting list of potential criteria for districts on the board. Then we move on to more constrained problems – namely, I inform them that they’re 50 years too late to have districts with widely varying populations and give them a new map, where each rectangle drawn always has 10 people in it to ease counting.

We draw 5 equipopulous districts on this in much the same way, then I show them my map.

This is clearly silly – so I show why I did it this way: I color dots based on their party affiliation.

In general, most of the class has heard of gerrymandering before, so I explain things fairly quickly, and move on to having them draw 6 gerrymandered districts on this map.

Then, with some practice under our belts, we move on to the full-size map.

Usually students manage to come up with something less ridiculous than my map, but it’s still tricky. In any case, this is all just the base problem – I want to get into the subtlety of the thing. There are lots of potential issues to consider, and not much time, so I usually just pick a couple. First of all, we can consider incumbency – while the map we drew may have looked good to start with, the three incumbents (marked with stars) aren’t going to be so happy about it.

So now students get to draw a set of districts that both advantages their party, and protects their party’s incumbents. Similarly, we can consider race – it turns out the map we drew before did a great job of splitting the minority vote (dots with green outlines), and they’re not happy about that.

And in this case it requires significant contortion to draw a majority-minority district on this map, and especially so if you want to give the blue party a majority.

This Is Real

At this point we move out of idealized Splashland and back into the real world. We start by adding to our list of things to consider when drawing districts. I hand out some real maps that illustrate some of the examples we talked about. Racial gerrymandering and the Voting Rights Act are a great example – it’s something that we worked with above, that has complex and nonobvious effects, and is in the news regularly, with Shelby County and two Alabama cases in the Supreme Court. At any given time there are some good examples from the recent news that we can talk about, and students can see how they directly fit into the maps we drew.

We talk about various methods of districting (and about the recent Supreme Court case about Arizona’s commission), and about options that avoid the districting game entirely – at-large districts, or mixed-member proportional representation – and the problems they solve and create. Sometimes I tie in the work I did for 6.S897, or real papers in political science (Chen and Rodden is a great one), to give them a sense for how one could go even deeper into the issue.