Princeton Election Consortium

A first draft of electoral history. Since 2004

David Daley and Sam Wang on gerrymandering

October 15th, 2016, 10:00am by Sam Wang


A web app for implementing Sam Wang's statistical standards for gerrymanderingLast night, David Daley and I appeared at our local bookstore, Labyrinth, to talk about gerrymandering: its effects on democracy, how technology has made it worse, and what can be done to prevent it. It’s is the subject of his new book, Ratf**ked. Thanks to Labyrinth Books and to Princeton Public Lectures, the discussion is archived for your viewing pleasure! Watch on Vimeo.

In the discussion, we talked about the role of math in diagnosing partisan gerrymanders. I have developed simple statistical tools to help provide a legal standard. These standards are available for your use at gerrymander.princeton.edu.

Tags: Princeton · Redistricting

16 Comments so far ↓

  • Brian MacDougall

    Finally got around to watching the whole presentation. Very thought provoking. I wish, however, as a Californian, that you had touched more on what happened here as a result of Prop. 11. Certainly, at the state level, before the Commission, governance was pretty stalled. We’ve moved ahead fairly steadily post-Commission, and solved a couple of persistent problems, especially with respect to the budget (by, you know, finally raising taxes!) but I’m pretty sure I wouldn’t want to be a Republican with political aspirations here, what with jungle primaries and redistricting.

    When I look at the California Commission redistricted map, the lines are fairly consistent with both the geography, political and cultural inclinations of a district; they make an intuitive sense to the extent that this native Californian understands the state. I don’t know that much about Pennsylvania, but those lines look pretty tortured. And I remember when the Commission was drawing up lines, nobody from either party was happy, which is a good indicator of impartiality. Seems to me that this kind of thing has to happen throughout the states, and not some kind of fiat from the court or the federal level. Of course, if the net effect in California was to strengthen the Democrats’ whip hand, how likely is repeating that process elsewhere?

  • Marc

    Sam,

    Per the House Generic Congressional Preference chart, it looks like nationally, the Republicans have a 7% advantage. That is Democrats have to win by 7% to get an even result. I’m sure in Pennsylvania, the margin is larger.

    Let’s assume your proposed gerrymandering test were implemented. What would be the maximum advantage a party could get, but not violate the constraint?

    I assume that some would still try to gerrymander as much as possible within your test’s standards.

  • evap

    Have you seen the announcement about the National Democratic Redistricting Committee, the new umbrella group for redistricting reform, chaired by Eric Holder? It seems like Obama has chosen gerrymandering as the (or an) issue to work on after he leaves the White House. Really good news!

  • shma

    Interesting book, although I wish Daley had spent more time on the technical method used to build the maps.

    On an unrelated topic, while looking at state-by-state differences between the major model in the NYT, I noticed there’s a PEC number for both Maine and Nebraska’s CD-2. I don’t recall seeing a CD level breakdown of evs for these states in the matlab code. Are you estimating the win probabilities for these districts in the same way as other states, or treating it differently due to the lack of polling in these districts?

  • James McDonald

    I’ve wondered if a probabilistic outcome for elections would make sense: make the odds of winning depend on the final vote percentages, so a 51% win in votes gives you a 51% (or perhaps slightly larger) chance of winning.
    This would tend to make gerrymandering a pointless effort, since small advantages in a large number of districts would (on average) just balance the corresponding large advantages in a smaller number of districts.
    It’s not even clear if it would violate the “one man, one vote” principle, although it would violate the generally assumed (but unstated) notion of what a democratic vote means.
    One other advantage would be to give occasional representation to what otherwise would be permanent minorities that never get represented.
    The actual function for the probability of winning should likely increase much faster than the vote probability, e.g. 90% odds for a 60% vote tally, just to avoid the spectacle of 1% vote winners (often with extreme ideologies) occasionally getting elected.

    • Sam Wang

      Your comment gets several things backwards. The very point of gerrymandering is to take advantage of the discrepancy between vote share and win probability. A district that is divided 60%-40% or more extremely has a >90% win probability for the dominant candidate. Therefore in a partisan gerrymander, all districts for both parties have a near-overwhelming advantage.

      In regard to “occasional representation” for minorities, this is a principle contained in the Voting Rights Act known as the establishment of majority-minority, or ability-to-elect districts. It can be achieved without the intense packing (i.e. 80%-20%) that is a hallmark of a partisan gerrymander.

  • Matthew Coons

    Did you discuss Chen and Rodden who claimed that “the Democrats’ geography problem is bigger than their gerrymandering problem”? (See NY Times article entitled “Don’t blame the maps” dated 1/24/2014.)

  • Andrew Crowder

    Creation and posting of the gerrymandering diagnostics is a distinguished public service. Thank you and congratulations, Dr. Wang.

  • Richard DuBois

    I’m hoping a new SCOTUS majority will put an end to the unequal representation.

  • Ravilyn Sanders

    When voters choose the legislators it is democracy. What is it when the legislators choose their voters? Legislatocracy?

    • J-D

      I am reminded of Bertolt Brecht’s poem ‘The Solution’.

      (It can be found on the Web if anybody wants to check it out for relevance; it’s short.)

  • DAVID DEEDS

    I was at Claremont in 1981-3 when the Rose Institute first began applying computing power to redistricting. They had a pair of IBM 360′s and as a friend told me “give me an outcome for the California Legislative and Congressional elections you’re looking for and I’ll draw you a map that will come within a couple of seats”. I remember being shocked, but later I remember being dismayed and wondering if this was the beginning of the end of a functioning democracy. 35 years it sure looks like it was.

  • Dave Kliman

    I wonder if a Supreme Court test case could be devised that would invalidate all of these renegade districts in one fell swoop, because it’s starting to look like such a case may be our only hope.

Leave a Comment