Princeton Election Consortium

A first draft of electoral history. Since 2004

New York thread

April 19th, 2016, 10:16pm by Sam Wang


Trump heading for at least 90 out of 95 delegates. If his 60% vote share holds up, I’ll guess 92. Talk among yourselves…

Wednesday 7:30am update: for now, it looks like 90 delegates. Trump got below 50% in 3 districts, therefore losing three delegates. Kasich finished first in one district, leaving 1 delegate for Trump there. Overall, the average district vote share was 60.6% for Trump, with a standard deviation of 10.0% – a bit more than my assumption of 9%.

Tags: 2016 Election · President

27 Comments so far ↓

  • martin boutte

    Hello Sam. What are your thoughts on Richard Charnin’s claims of systemic voter fraud based on exit polling discrepancies. He has argued that every one of Clinton’s victories is fraudulent. I find this incomprehensible given the margins, especially in the South. What’s your take?

  • Mark F.

    CA is very hard to predict. I live in the most Democratic district in the state (East Bay, Oakland, Berkeley). Who are the very few Republicans here voting for? I have no idea. Remember each district gives out 3 delegates, the statewide winner just gets 14.

    • Sam Wang

      Sure. Nobody knows anything. Reminds me of Peggy Noonan.

      Basically, this is the kind of thing that is said by people who are ticked off about what is going on.

    • truedson

      It does seem pretty silly. Except for Michigan, with a anomaly no one else has noticed, the polls have been very good this season. As noted the job of pundit becomes rather redundant with a site like this.

    • bks

      During the lead up to the 1992 election I asked a grad student at UC Berkeley doing a PhD in quantitative political science to explain margin of error in the polls. Since then I just assume that the MoE is 3% and feel none the worse for having completely forgotten everything he said in three hours at the blackboard. (I have much the same laissez-faire attitude towards P-values in gene association papers.)

    • Matt McIrvin

      There do seem to have been several states where the winner, whoever that was, overperformed. I think Sam’s observed that as a pattern in elections that are not very close: polls tend to understate the magnitude of blowouts.

    • Matt McIrvin

      …But there is a big difference between “these polls had a systematic error” and “all polling is useless”.

    • Xs

      The polls include undecideds, which obviously isn’t a factor in the voting. If your metric for poll accuracy is matching the final vote share, you should scale up each candidate’s number in the polls so they add up to 100%. The error doesn’t look too bad as is, even less when you take that into account.
      RedState is throwing a tantrum because they don’t like Trump.

  • Mark F.

    Cruz has to be unhappy about failing to even get 15%. I’m starting to think a 2nd ballot is unlikely.

    • Josh

      Yeah. And now Cruz is about to head into the next election day (the 26th) with the possibility of winning only a dozen or so delegates across five (!) states.

      I wonder if some party heads will start prevailing upon Cruz and Kasich to back off if (when) Trump goes five for five next week?

    • Frank

      I think the results of CA will be the one to watch to gauge how strong the likelihood of chaos at the GOP convention could be. The RCP average shows Trump +9 over Cruz, but if you use only the two polls that were conducted this month, Trump is +12.

      Trump is competing against himself now. Will he actually go over 1,300 and up to 1,400 as that internal memo by his campaign believes? The higher he goes over 1240, the less likely there will be a contested convention.

  • A New Jersey Farmer

    Polling called the races, though both winners outperformed the polls. Once Trump lowers the volume he seems to lose some luster. And interest.

    • Maddie

      Trump needs to pump up the racist volume to keep the GOP base locked down. His problem is that what locks down the (extremely white) GOP base is so toxic in its bigotry that it hurts him with (much less white) general election voters.

    • emmy

      I’m a black Trump supporter and I have to disagree. What attracts supporters isn’t quite that toxic at all. He’s broad-band.

  • 538 Refugee

    Well, my wife is relieved. She wouldn’t take my word for it that Hillary would win NY comfortably.

  • Commentor27

    Trump did well in counties neighboring Connecticut.

  • JayBoy2k

    Just heard that there a 20% minimum to gain any delegates in NY. Means that Cruz will get none and Kasich a few – possibly 3 or 4.
    CNN noted a relatively short and subdued speech by Trump. Maybe his new campaign team is having an influence.
    Hillary sounded very happy.

    • C.S.Strowbridge

      “Just heard that there a 20% minimum to gain any delegates in NY.”

      Based on the rules, Trump got 50% statewide, so he got all 14 statewide delegates. At a district level he got 50% in 23 districts, for all 3 delegates in each district. He won a plurality in 3 districts, giving 2 delegates per district. He won second place (and was above 20%) in another, giving him 1 delegate. That’s 90 in total.

      (edited for grammar and clarity)

  • Olav Grinde

    This map of the New York gives an interesting perspective on the results. Areas that Hillary won are green – while Bernie’s areas are blue.

    http://elections.huffingtonpost.com/2016/primaries/2016-04-19#NY-Dem

    • Michael Hahn

      Interesting, but not surprising! If you were to look at a county map from the last two presidential campaigns across the country, you would see a similar pattern. Sanders won in the more rural (and less populated) areas of the state, while Clinton took the most populated areas. It is the votes that count, not the counties!!

    • Commentor27

      I’d also guess that urban areas contain more minority voters compared with rural voters, and Hillary is comparably stronger with minority voters.

    • Brian Tucker-HIll

      Note she won the largest cities outside of the NYC area–Buffalo, Rochester, and Syracuse. Along with the results in Ohio, this is a good sign for how she will perform in Pittsburgh and Erie, which combined with her very likely support in and around Philadelphia should make Pennsylvania look a lot like NY.

    • Matt McIrvin

      Clinton is polling somewhat stronger in PA and MD than she was in NY.

  • Joel

    From a cursory glance, Trump is over 50% in almost every county. He’s only leading Kasich by 9 in Onondaga county, though (Syracuse).

  • bks

    Will Cruz start regressing to the polling?