Data in Policy Debate: What _Isn't_ Gerrymandering?

There's an image that has been floating around social media about gerrymandering, called "how to win an election." It was posted with some pretty great discussion at the Something to Consider Forum

The image implies that the 2nd block ("BLUE WINS") is a reasonable way of breaking up the population into districts, and that the 3rd block ("RED WINS") is gerrymandering and "stealing the election." The comments on social media that we've seen suggest that this is how most people interpret the image.

But as user thomas_w points out, here's something to consider: because the population is 60% blue and 40% red, the most accurate electoral body from this population would have 3 blue candidates and 2 red candidates. The block on the right ("RED WINS") in which red "steals" the election has one more red candidate than ideal; the 2nd block ("BLUE WINS") has two more blue candidates than ideal--it's twice as wrong as the "election theft" block and completely disenfranchises the very large red population from representation in legislature.

There are three very interesting implications here.

First, it is a great example of very manipulative presentation in policy debate. Because the districts are "normal" looking rectangles that each contain the 60/40 split, it appears at first-glance to be "reasonable," even though it shuts out almost half of the population from legislative representation.

Second, it is a great example of our personal biases when reading such a graph. The use of "red" and "blue" implies "Republicans" and "Democrats," and on the right we see the "Republicans stealing the election from the Democrats." If we root for the blue team, we are probably more likely to look at this image and approve of its message, and of course we have a bias to think that a 100% blue legislation is a great answer for this population.

Third, and most importantly, it provokes us to ask: what the heck isn't gerrymandering? While some particularly heinous examples show obvious gerrymandering afoot, how can we know that other seemingly-mundane districts aren't gerrymandered?

No population is broken up so evenly as the box in the image above. In that case, the way to get proportional representation in the legislature would be to have five vertical columns or something like it, but it would create districts that are 100% noncompetitive: there is essentially no way that the blue team could win in a red district, and vice-versa.

We also know that populations aren't static: over time, populations are going to grow disproportionately and shift, so the state legislatures (in the US at least) are obligated to re-draw the lines in order to "even out" the districts. How can they do that in a way that doesn't "unfairly" disadvantage one party or another? Are they obligated to try to create competitive districts everywhere? Or are they obligated to try to create proportional representation from the state of the political distribution du jour? Or instead should they try their best to ignore the political preferences of different areas entirely and draw the lines in a more "purely geographical" way?

We'll look at the Massachusetts Congressional districting map (our home state) as an example:

Looking at this map (which was re-drawn in 2013 by the overwhelmingly-Democratic state legislature), is there "evidence" of gerrymandering? How might we tell?

Does knowing that Massachusetts consistently sends only Democrats to the House of Representatives, even though 25% of party-registered voters are Republican (and that's excluding the huge chunk of Independents), mean we have gerrymandering on our hands? 

Let us know your thoughts in comments below.

And hey! If you like what you're reading, consider supporting us on Byline. We're 100% reader-supported and independent, and your support is what will keep us going in the future.

Editor's note, 9:10 AM, 10/27/2015

Thanks to reader Ben for helping us track down the origin of this image. It was first created by Stephen Nass, a Wisconsin state senator. It was shared often on Facebook and then eventually adapted and re-configured by the Washington Post to be more nuanced. 

Editor's note, 3:25 PM, 10/27/2015

Thanks to reader Paul for pointing out a great MIT student-led class on how gerrymandering works that we think is a really interesting read for everyone interested in the topic.


Erik Fogg

We do politics, but we don't do the thinking for you.