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

Top 10 GOP candidates

August 4th, 2015, 2:58pm by Sam Wang

…with appropriate error bars! From Danielle Kurtzleben at NPR.

Tags: 2016 Election

50 Comments so far ↓

  • 538 Refugee

    The unanswerable question is if the ‘kid’s table’ tier would have really ever garnered enough serious support to make a run. It seems all the money is already being gobbled up and that is the real test of a politician these days. Can you raise campaign funds? It might be interesting to see where the eventual nominee is on the current list though.

  • Amitabh Lath

    I had to do this sort of choosing and culling once when I was on the organizing committee of one of the large physics conferences. You either pick a few solid ones from the submitted abstracts, or you end up with a bunch of 3-minute talks.

    Of course there was pushback from the cut ones. Same as Fox will get whoever they choose.

  • A New Jersey Farmer

    In the end, the candidates will most likely get about 8 minutes of airtime, so the ones left out won’t lose all that much. Then there will be a reshuffling once Trump flames out.

  • bks

    I give up. Bring on the GOP Christian police state. –bks

    • 538 Refugee

      Google’s news feed showing several sites with stories saying Trump’s attack on Megan Kelly will be his demise. We are about to find out if these were off the cuff gaffs or strategically planned pandering to the lowest common denominator for political gain.

  • bks

    I think it will take at least a full week for the aftermath of the “debate” to show up in the polls. I’ll be most interested in those conducted between the 13th and the 20th of August. –bks

  • 538 Refugee

    I won’t touch the poll name with a 10 foot pole.

    New NBC News/Survey Monkey Poll: Donald Trump Still in the Lead After Debates

  • Olav Grinde

    In my opinion, “Who are the top 10?” is the wrong question. Far better to ask: “Which GOP Presidential candidate currently enjoy significant voter backing?”

    We see that only Donald Trump, Jeb Bush and Scott Walker enjoy double-digit scores in the polls. The others lag far behind.

    So, honestly, Fox News should have settled for a three-way debate. Perhaps augmented by a 14-part infomercial from the other contenders.


  • bks

    Trump still leading, Bush biggest debate loser, Carson biggest winner. No bump for Rubio, despite the softball questioning:

  • Amitabh Lath

    CNN has a poll out, 1k RV, similar numbers as the others (Trump 24%, Bush 13%).

    Interestingly, favorable/unfavorable for both look similar (38/58 for Trump, 35/57 for Bush).

    I confess watching the political pundit class evokes a little schadenfreude. Feels like classical physicists at the beginning of the 20th century watching quantum and relativity results pour in; realizing the rules of nature they had assumed universal were only valid in a specific environment.

  • bks

    What’s stunning, though, is how this compares to Jeb Bush. Donald Trump’s net favorability among all voters is higher than Jeb Bush’s. Margins of error, etc., but for Pete’s sake. Tell someone in March that Donald Trump would be leading the Republican field by double digits and they’d roll their eyes. Tell them that Trump would be viewed as or more favorably than Bush and they’d have had you committed.

    • Sam Wang

      On the other hand, when it comes to favorable/unfavorable numbers, Jeb is not a high bar to clear. Trump’s still in the basement with a lot of GOP voters. Now he’s about where Gingrich, Cain, Santorum, and Bachmann where at the moment they crashed and burned in 2011.

    • bks

      I believe that in August/September of 2011 it was Rick Perry who was leading all the polls. Then his terrible performance in the September debate (not the Oops debate, that was in November) ended his hopes. Bachmann and Cain and Gingrich then had their moment in the sun, but like Icarus, their wings could not endure the burning light. I still find it hard to believe that Trump could capture the nomination, but I don’t think that his downfall is going to be due to his public performances. –bks

  • Amitabh Lath

    Has anyone sensible commented on the systematic uncertainties of the Trump numbers? I am still wondering if they are inflated by his celebrity.

    The analogy is with HIV testing. Let’s say a test is 99% efficient in detecting HIV when present, but gives 5% false positives. And 0.1% of the population actually has HIV. Then the group who test positive would be dominated by people who don’t have HIV but tested positive falsely, rather than people who actually have HIV and tested correctly.

    Now imagine Likely Voter screens. In the past type 2 errors didn’t matter because let’s face it, if you had heard of Herman Cain, you were pretty likely to vote in a primary anyway.

    Now there is a huge population that has heard of Trump, but isn’t the usual primary voter. Even if the rate of them leaking into the LV pool is low, N is so large that they could easily skew his numbers up.

    Has there ever been a candidate with this level of celebrity? Are there any estimates of how much this effects the polling?

    Any thoughts? If anyone knows how LV screens work maybe explain why this wouldn’t matter?

    • Amitabh Lath

      One way to check if Trump’s bump is real: check if some very exclusive demographic that makes up a large part of the Trump fan club, but never ever shows up to vote in the Republican primaries, is passing the LV filters.

      Say, 18-25-yr-old-white-males-no-college-drives-Camero. Assuming they don’t historically vote in Republican primaries but are suddenly a big part of your LV pool, then you know there is an issue.

      Of course as 538 Ref says, this may be real, these Trump fans may actually go out and pull the lever on primary day.

  • 538 Refugee

    My fear is that this may be more real than we want to believe.

    Rosalynn Carter says about Ronald Reagan, “I think this president makes us comfortable with our prejudices.”

    It is convenient, but probably superfluous, to be able to add “illegal” in front of “immigrant”. Now even ‘natural born’ isn’t good enough for Trump? Is he circling the wagons for an increasingly threatened segment of the population that is en-route to lose it’s numerical advantage?

  • pechmerle

    There have been at least three candidates for president with this level of celebrity (much higher, actually):
    G. Washington, U.S. Grant, and General Eisenhower.
    Not that that helps with your statistical question.

  • bks

    Irony note: The CNN GOP debate (which apparently will exclude Fiorina) is on 16 September, which is Grito de Dolores, Mexico’s Independence Day. –bks

    • Paul

      Unless I have missed a CNN update, the decision on who is invited into the top 10 part of the debate cannot be made until after factoring in any polls completed on September 10.

    • bks

      The problem for Fiorina is that most of the polling that CNN will use has already been published:

    • Paul

      Thanks to reply from “bks.” I did not realize that the CNN selection window goes all the way back to July 16. In addition, they must not be counting the August 10 Rasmussen poll where Fiorina received 9% of the responses. What seems unfair is that a candidate, in this case, Fiorina, who was not at the main table at the 1st debate (Fox), appears unlikely to make it to the main table at the Sept. 16 CNN debate due to the rules setup by CNN. I would prefer a shorter window for the CNN debate which would have biased the results to later polls rather than earlier polls. Whatever the value of these debates, CNN seems to have setup rules which all but guarantee the same set of candidates at the main table as in the Fox debate.

  • Matt McIrvin

    Something very odd is going on with head-to-head Democrat vs. Republican polling right now. Hillary Clinton has reasonable leads over most of the Republican candidates in state polls, but Quinnipiac (not an outfit with a lot of Republican bias in the past) keeps releasing swing-state polls showing her losing badly to nearly all of them. They’ve claimed this now about Colorado, Iowa, Florida, Ohio and Pennsylvania.

    By most historical analyses of this sort of thing, one would think these results are simply starkly inconsistent. She can’t be losing all these states if she’s leading nationally. Her lead in non-swing states would have to be so large as to be implausible.

    But I suppose it’s possible that Hillary Clinton’s lead is somehow concentrated in blue states to an unprecedented degree, or maybe she has some unforeseen appeal to a minority in deep-red states.

    It seems more likely that something is up with these Quinnipiac polls. But not many outfits are doing state head-to-head polling right now.

    • Matt McIrvin

      “Hillary Clinton has reasonable leads over most of the Republican candidates in state polls”

      Read “national polls” there, of course.

    • Matt McIrvin

      …also I guess Quinnipiac only has her losing to Rubio in Ohio, not all the others.

    • Amitabh Lath

      Quinnipiac poll seems to have oversampled Republicans. In FL they have N=1093, with 477 R and 345 D (rest Independent I guess).
      I presume they corrected with some demographic based algorithm but when your raw data has a large bias then your correction coefficients are big and it’s all a big mess.

  • Matt McIrvin

    The update at this link says their samples are unreasonably white and old compared to the 2012 electorate, which would neatly explain all of this:

  • 538 Refugee

    Where have we heard this before?

    2. But there’s still a caveat

    Trump’s success, though, has always carried the same asterisk: A lot of people still say they wouldn’t consider voting for him.

  • Amitabh Lath

    I have a question about the favorable/unfavorable numbers that a lot of people (including Sam) have been pointing to as indicators of Trump’s inability to win the Republican primary.

    I do not understand the mechanism by which unfavorables turn into defeat. At the very least you would need to know the correlation matrix.

    For instance, if Trump unfavorables correlate highly with Bush unfavorables, then those people may well stay at home, and Trump wins.

    (of course instead of staying home they could write in a protest candidate or vote Pataki etc. if the symbolism of casting a ballot is important to them)

    • Sam Wang

      The empirical relationship is here. The mechanism? Basically that you wouldn’t vote for someone you felt unfavorable about. So it’s a necessary condition to have positive net favorables…but not sufficient.

    • Amitabh Lath

      Yes, Christie being 4sigma more unfavorable than the mean probably indicates he is buttered toast.

      But this highlights my point, we should look at relative favorability, people who dislike the whole set: Trump, Bush, Walker… are probably not going to vote. And Trump’s relative unfav is similar to the others.

  • bks

    Ed Rollins was the campaign manager for Reagan and for H.Ross Perot:

    Yes, Donald Trump can win, but here’s what it will take
    By Edward J. Rollins, August 17, 2015


  • Amitabh Lath

    Talk of a Trump “ceiling” seems a lame attempt to render in quantitative language what is basically a pundit gut feeling: Republican primary voters wouldn’t be crazy enough to nominate him, would they?

    It appears they would, and will.

    • 538 Refugee

      I think the “ceiling’ is real but a moving target. Trump has grabbed the lead by taking a hard anti (illegal) immigrant stance. Being unapologetically insensitive played well to a large segment and that’s all his current supporters need to continue to support him as he drifts to the center (eventually) to give cover to people on the periphery. That drift continues as long as he can make it work, gradually growing his base of support. At the end of the day you count on the rest to vote against your opponent. That’s how you attack the ‘unfavorables’ rating. Drifting to the center after securing the ‘fringe’ seems to be pretty standard political fare.

      That said, it is still a delegate numbers game that I don’t care to delve into. Who drops out, when, where that support goes, etc…

    • Amitabh Lath

      Yes, Trump has managed to carve out the perfect message unhindered by obligations to Wall Street tycoons or Chamber of Commerce or Grover Norquist.

      His stance on Social Security and Medicare is “don’t touch it at all” which resonates with the elderly a lot better than the usual Republican “must privatize but will try to spare older folks”.
      This is not a stance that candidates dependent on Wall Street money can afford to take.

      Couple that with the know-nothing levels of nativism that he calculated would connect with the bubbling xenophobia, and you have got the white majority that Sean Trende wrote about when he crunched the numbers post-2012 and declared that the Republicans didn’t really need to go after the hispanic vote if they could boost the white vote.

      Trump is boosting away.

    • Matt McIrvin

      The thing is, if you just try to improve on Romney’s or Reagan’s share of the white vote by appealing to naked racism, you’ll probably turn out the non-white vote as well.

      Consider also that a Trump v. Clinton race in particular would probably have the most titanic gender gap of all time, which I guess cuts both ways (but there are usually slightly more women voting than men).

  • bks

    A majority of Republican voters believe that Donald Trump will be their party’s presidential nominee, according to a Rasmussen poll released Friday.

    The poll found that 57 percent of Republican voters think Trump is likely to win the nomination, up from 27 percent two months ago, when the billionaire businessman officially launched his campaign.

    • Amitabh Lath

      I wonder who President Trump will install as Secretary of Energy. I guess a physicist like Chu or Moniz is too much to hope for.

      I for one would like to welcome Secretary Kardashian and look forward to discussing the LHC upgrade with her.

  • Amitabh Lath

    Interesting if disturbing read in the latest New Yorker about people who are really fired up about Trump.

    • 538 Refugee

      Perhaps a more in depth and well documented version of what has been said here. We have to be mindful of Sam’s moderation responsiblities though. ;)

      Racism is the ultimate form of lazy. You’re ‘better’ because you were born ‘better’. No effort required.

      I got a good chuckle out of the landscaper whose parents told him to get a ‘normal’ job if he wants to be happy. Landscaping is not a year-round job in Cincinnati yet he expects the same standard of living as his two working (teacher) parents enjoy? Wouldn’t that require something like a welfare subsidy? Maybe he is in landscaping to keep all of those illegal immigrants from getting those jobs?

Leave a Comment