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

Happy Thanksgiving!

November 22nd, 2018, 12:42pm by Sam Wang

Dear readers, pardon the slowed rate of posting. We’re not going back into hibernation – we’re just resting. More is coming soon!

Traffic was substantially down this year compared with 2016 – by more than a factor of 10. That’s understandable for a variety of reasons, including our lighter posting regimen and the consequences of PEC’s general election calculations. But I will note that the most committed PEC readers stuck around. One measure is the comments, still solid. Another is the donation sites at left for ActBlue (D) and NRSC (R). The amount in the ActBlue is almost exactly equal* to the total in 2016, which was itself a record $362,590. In an off-year, that’s really incredible. The NRSC doesn’t have tracking – that would be interesting to see. I am thankful for your continued readership!

I am also thankful for the expanded interest in state-level action. State elections and ballot initiatives are a major part of U.S. democracy going forward. Witness the fact that we just saw redistricting reform pass in all four states where it was on the ballot (Utah just passed a few days ago), increases in the minimum wage in Missouri and Arkansas, and Medicaid expansion in half a dozen states around the country. It’s a remarkable flowering of the New Federalism.

All of this shows that we’re at some kind of pivot point in U.S. history. The current era started in the mid-1990s with Gingrich, and reached an extreme with the election of Trump. And now, the 2018 election gave a preview of what might happen next. Are we reliving the 1974 Watergate election? Are we reliving the end of the Gilded Age? The United States is going someplace new. I wonder where that is.

Happy Thanksgiving to all!

*As of November 24, the 2018 total has now exceeded the 2016 total. One important note: in 2016 there were 1,600 donors. This year’s total has come from 1,137 donors so far.

Tags: Redistricting

9 Comments so far ↓

  • 538 Refugee

    I remember saying in a post somewhere back in 2016 prior to the election the sad thing about a miss would mean that polling, i.e. data collection, was less reliable than we though. Your methods simply analyzed the available data as provided by the folks that were charged with doing that.

    I think in the end the data showed that the Republican candidate got about the same number of votes as in the past two presidential elections. I don’t know if Democrats stayed home or crossed over while some Republicans stayed home. Next cycle could be even harder to model. There are already indications that Republicans will actually go up against a sitting president in their own party so we can’t even be sure if Trump is going to get the nomination or what the backlash of that could be. I can’t even pretend to tell you what the Republican party stands for at this point. If anyone wants to take a stab at it?

    • LondonYoung

      Stab #1: Laissez-faire capitalism and legislation of 1950’s-era social standards.

      For details, see this website:
      and just take the opposite side.

    • Luis

      “Your methods simply analyzed the available data as provided by the folks that were charged with doing that.”

      Well, yes, but the analysis was poor, failing to account for the possibility that polling failures would be correlated from state to state. Nate correctly pointed that out before the election, well enough that I could explain the problem to people before the election. I like Sam, and hope he grappled with the problem, but I certainly have not read this site as religiously since it became clear that mistake was being made and not addressed.

    • Sam Wang

      Luis, I’ve already written about it. I underestimated the size of the overall polling error in my May 2016 MATLAB script. I’ve written about this in the NYT and on this site. I am not sure what you are looking for, unless you want the same thing again.

      You seem to be unaware that in early 2016, I also pointed out the commanding position of Donald Trump and Hillary Clinton for their respective nominations. I wonder why that escaped your notice.

      I would guess the problem might be that you’re mad about the election itself. In which case, it’s absolutely okay if you stop reading this site. Thanks for commenting.

    • LondonYoung

      “mistake” is a harsh word …
      All approaches are models and there is no perfect model.
      Remember – Sam makes all his code available … he shows exactly what he does – I would love to see that from Nate, but that is not how Nate operates.

      Jamming in a “correlated pollster miss” factor is just a single line of code. Coming up with a way to calibrate it? Feels like more guess than science to me.

    • Matt McIrvin

      Sam attempted to take correlated polling misses into account. He just vastly underestimated the magnitude of the effect. The big debate was whether Sam’s take on this or Nate Silver’s was closer to correct–Silver seems to have had the more realistic take, hedging with big fat tails on his electoral projections.

      Somewhat ironically, though, state-level polling was extremely accurate in the 2018 House cycle, enough so that Sam’s old approach would probably have seemed bang on. But in previous midterm cycles he’d argued that midterms were inherently harder to call…

    • LondonYoung

      I don’t know exactly how Nate handled this because he doesn’t make his code available.

      But, as Sam has said, his “hurricane model” leaves too little uncertainty on the day the hurricane hits.

  • Pechmerle

    I’m thankful that the Dems have retaken the House, with a good working majority. Now hold all four seats that are in or overlap with Orange County, CA, the place , said Ronald Reagan, “where good Republicans go to die.”

    I’m thankful that my Congresswoman, the much demonized (it.didn’!) Nancy Pelosi, will return to the Speakership.

    And in a very interesting turn, I’m cautiously thankful that C.J. Roberts finally spoke out publicly in defense of the independence of the federal judiciary. In the face of DJT’s constant attacks on the judiciary. Roberts seems to have felt a line was crossed in DJT’s most recent comments on the federal judiciary as a whole as biased. I have been hoping that Roberts would see a need to protect the Supreme Court as an institution. To do so, he will need to take on the role for so long played by Kennedy – a natural conservative who casts some liberal votes on cases before the court to maintain a degree of balance. Is he just talking the talk, or will he really do it? (He did do it in NFIB v. Sebellius, the case upholding Obamacare, in which his vote was critical and he wrote the controlling opinion.) I’m just hoping there is something real here; it would mean so much.

  • TC

    Nixon was the last Liberal president (ironically, a Republican). Then things swung to the right, and to the right again under Bill Clinton (and Gingrich). Now there is a Liberal and Left pushback against Trump et al. The next President can expand the current local, state, and national swing back. Much more could be said, but the fundamentals are obvious enough, and have been clear for two and a half years, minimum.

Leave a Comment