Princeton Election Consortium

Innovations in democracy since 2004

Outcome: Biden 306 EV (D+1.2% from toss-up), Senate 50 D (D+1.0%)
Nov 3 polls: Biden 342 EV (D+5.3%), Senate 50-55 D (D+3.9%), House control D+4.6%
Moneyball states: President AZ NE-2 NV, Senate MT ME AK, Legislatures KS TX NC

The Princeton Gerrymandering Project is live!

August 4th, 2017, 7:34am by Sam Wang

I am happy to announce our revamped site at This is part of the Princeton Gerrymandering Project’s tooling up for the coming several years of work by courts and reformers.

The site now has an interactive map showing the results of three simple gerrymandering tests, applied in all the states. It allows you to upload data more conveniently than before. Finally, it has a tutorial, as well as links to background reading and current court cases. Check it out!

All of this was done by the team: Rob Whitaker, Brian Remlinger, Aimee Otsu, Sung Chang, and Naomi Lake. Hats off to their great effort.

Tags: Redistricting · Site News

11 Comments so far ↓

  • LondonYoung

    Virginia flips from “definitely Gerrymandered” in 2012 to “unlikely to be Gerrymandered” in 2014. Was there a court order redistricting in between?

    • Sam Wang

      A gerrymander can disappear if the offending political party has a good year, at which point they legitimately earn what they took in the previous election. For this reason, GOP gerrymanders look smaller in 2014 by our tests.

      For 2016, one district was redrawn by court order:

    • LondonYoung

      I am wary of the incumbency effect.
      If one party manipulates the borders to win elections in year N, and then enjoys an incumbency effect that defeats the tests in year N+2, I am tempted to call it a gerrymander all the same. Incorporating incumbency seems hard.

      Of course, there is also the fact the republicans seem more willing to vote when the presidency is not at stake than democrats are.
      Maybe, like Cook PVI, tests should be based on two or more consecutive elections.

  • Amitabh Lath

    Gerrymandering seems to be exclusively an Republican effort. As much as I would like to believe that Democrats eschew gerrymandering on moral grounds I suspect this is not so. Why the asymmetry?

    • Sam Wang

      Two reasons, as far as I can tell, based on opportunity and motive.

      (1) Since 2000, more states have under single-party Republican control of redistricting. They used that opportunity.

      (2) Before 2000, Democrats were focused on single-incumbent protection. Building one safe district can help a party overall (usually if the win is 70%), or neither (in between). And I think they assumed they’d control the House forever, which reduces the motive.

      Note that there are Democratic gerrymanders, most notably in Maryland.

    • LondonYoung

      At present the dems control the government in eight states, but in CA an outside commission draws the boundaries. In the remaining seven states, the dem have achieved 30 out of 32 seats. The last two GOP seats are in rural inland areas that are very hard to gerrymander away.

    • Matthew McIrvin

      Also, modern software for optimizing gerrymanders didn’t exist in the days when Democrats had more opportunity to lock in party dominance through gerrymandering. Those products have probably made gerrymandering far more effective than it was previously.

    • LondonYoung

      Don’t forget that gerrymandering only works when you can predict how people will vote. It is only recently that identity politics have become so strong that this has become easy.

      Today, if a congressperson is from a district that the opposite party took in the presidential election they are a target who won’t last long. It has not always been this way.

  • Joseph Elfelt

    Everyone is welcome to use the Google + GIS maps I have produced that show congressional districts and state legislature districts. The purpose of these maps is to help shed light on gerrymandering.

    And if you have seen these maps before, then please note that the way the congressional district data is displayed has just been improved as follows.

    1. Congressional districts now have semi-transparent colors (instead of solid colors) representing party affiliation of the representative.
    2. When you click a congressional district you see a display that includes social media links for the representative and both senators.
    3. The congressional maps will often display faster since my server (VPS – Virtual Private Server) is now hosting most of the data.

    To open these maps:
    2. Click or tap the basemap button (next to the “Menu” button)
    3. Look under the “Overlay” heading. Mobile users need to scroll down.
    4. Select “State_legislature_districts” or “Congress_districts”
    5. Select a state

    For more information please click “Map Tips” in the upper left corner. Some of the information on the “Tips” page is different depending on whether you are looking at the state legislature maps or congress maps.

  • Alexander Dubbs

    I am a researcher at NYU interested in redistricting and gerrymandering. I have an algorithm which I think can create fair districts for any state that pass the mean-median objective, and possibly others. The problem is that I don’t have real data, I would need the latitude, longitude, and party of every voter in a state in order to create districts and prove they are fair. Do you have this data? See my linkedin profile for my publication history.

Leave a Comment