Princeton Election Consortium

Open thread, post-Kennedy withdrawal

—

by

Sam Wang

Good morning! I’m fairly occupied with the new individual voter-empowerment tool Vote Maximizer.org, the start of fall term, as well as an imminent TED talk release on September 10. Lots of things happening here.

In the meantime, some random notes:

  • The withdrawal of Robert F. Kennedy Jr. from the Presidential race is of little consequence. As most of you know, minor candidates usually fade in the home stretch. This process was already under way even before he withdrew. At PEC we were insulated from this, since the poll scraper takes two-way Harris-vs-Trump state polls from FiveThirtyEight.com when available, and uses surveys with other candidates if that’s all the polling organization did.
  • PEC daily updates currently take place at 10:00am, 2:00pm, and 10:00pm. Just check then – and use your energy to do something for your side (D or R)!
  • The Senate tracker has a problem. It should be repaired soon, and also incorporate the Nebraska race, where something interesting is brewing. See this article about independent candidate Dan Osborn.

Comment away! (P.S. I was replying in comments, but will be slowing down on that for a while.)

Comments

19 responses to “Open thread, post-Kennedy withdrawal”

  1. lgd919219 Avatar
    lgd919219

    Curious—what caused the 0.5 movement towards Ds in the last few hours? Only polls released were (I think) the Fox polls. All within 1-2 pts of each other and within MoE.

    Reply
  2. Sam Wang Avatar
    Sam Wang

    Probably North Carolina? https://projects.fivethirtyeight.com/polls/president-general/2024/

    Also, the point of the meta-analysis is that it denoises polls by combining them. The fact that an individual poll is within its own margin of error is not particularly relevant, since it all goes into a larger mix that reduces the effective uncertainty.

    Reply
  3. lgd919219 Avatar
    lgd919219

    Got it. Thanks!

    Reply
    1. Sam Wang Avatar
      Sam Wang

      Actually, it probably is the Fox polls you cite. The link I gave is the entire FiveThirtyEight presidential feed. If you are following minutely, you can sort that feed by date added to see what came up last.

      The current fluctuations are noise until there is visible movement out of the zone visible over the last week or so. Also, to reiterate, there is nothing magical about crossing the zero line. If it gets farther than 1.0%, get interested; farther than 2.0%, it’s convincing.

      Reply
  4. MarkS Avatar
    MarkS

    The votemaximizer.org link in the opening paragraph is broken.

    Reply
    1. Sam Wang Avatar
      Sam Wang

      Thank you! Fixed now.

      Reply
  5. ArcticStones Avatar
    ArcticStones

    How many House seats does a 3.0% meta-lead (or the current 2.6% margin) translate into?

    Your grey zone is now completely above the line indicating D control. Do I understand correctly that there is now > 90% probability that Democrats will flip the House?

    Technical issues: Is it necessary with the time-consuming reCAPTCHA confirmation for known PEC commenters? Also, the Replies are not placed below the relevant comments.

    Reply
  6. Curious John Avatar
    Curious John

    Hey there–
    What happened to the Senate charting ? All of the data points–that is, the entirety of the chart, past and current–have shifted one-seat in the R direction in the last 24 hours. You mentioned in the post that “the Senate tracker has a problem” but didn’t specify what happened.
    Thanks !

    Reply
    1. Sam Wang Avatar
      Sam Wang

      Indexing error, kind of embarrassing, classified 49 D/I seats as 50, etc. Frankly it was bothering me for some time because it was apparently wrong, but Vote Maximizer was more urgent. If it affected your ability to be an activist then I apologize. I will say more later.

      Reply
  7. Sam Wang Avatar
    Sam Wang

    @ArcticStones, let me see what is possible. I do not know why threading is not working for you…

    Reply
  8. Curious John Avatar
    Curious John

    Thanks for the quick reply. So, is it correct now, or is the one-down step the error ?

    Reply
    1. Sam Wang Avatar
      Sam Wang

      It is correct now. Democrats/Independents are favored for 49 seats as the modal outcome. Lots of possibilities for that to change, with changes in the D-ward direction apparently easier. See the President/House/Senate graph for a sense of which way the wind is blowing for now.

      Reply
  9. Curious John Avatar
    Curious John

    I see. You added some additional discussion to the reply you made, which answers my second question. Thanks !

    Reply
  10. MarkS Avatar
    MarkS

    Why is the Senate control line drawn between 49 and 50 D seats? That’s only the case Harris/Walz win the POTUS electiion. I think you should have a gray zone surrounding 50.

    Reply
  11. Marc Avatar
    Marc

    Why does Montana have the most vote power for Senate control? It’s nowhere near the closest race. That ought to go to Texas?

    Reply
    1. Sam Wang Avatar
      Sam Wang

      https://samwang.substack.com/p/democracy-moneyball

      Reply
  12. marco Avatar
    marco

    For horserace purposes, to what extent do you think it’s appropriate to make adjustments for states where abortion is on the ballot?

    Do we expect a larger than normal turnout for women? For democrats?

    Reply
  13. Jeff Avatar
    Jeff

    Hi Sam, minor bug on the Congress page. The 538 poll links for senate are now “https://projects.fivethirtyeight.com/polls/senate/2024/[state]”. The current links only show 2020 and 2022 polls.

    Thanks for your excellent analysis!

    Reply
  14. nhm Avatar
    nhm

    Is the house estimate of “meta lead” broken (there has been zero movement for a couple weeks in the estimate).

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *