October 30, 2016

Prediction: Trump 48, Clinton 45, Johnson 4, Stein 2, All Others 1

Heading down the home stretch, it's time to make a prediction.

I'm basing it solely off of the USC poll, which was the RAND poll in 2012. It's the only panel poll, following the same sample of individuals over time, so that any changes reflect true changes of heart and not a shift in the make-up of the respondents depending on who feels like participating. (See this earlier post for details.)

It's also an internet poll, which the Brexit results revealed were more accurate than phone polls. They even provide tablets and internet access to people without them, to get as representative of a slice of America as possible. And the sample is part of an ongoing, longer-term study independent of the election, so it won't be affected by who initially felt like joining an election-themed study during an election year.

Under its RAND branding in 2012, it was the most accurate -- more so than the IBD poll, which performed worse in 2012 than it did in '08 and '04. (IBD is not a panel but ad hoc sample, and is phone-based with about 65% being cell phones.)

The day-of estimates were Obama 49.4, Romney 46.8, for a spread of 2.6 points. Actual outcome was Obama 51.1, Romney 47.2, for a spread of 3.9. The RAND poll was off by just 1.3 points in the margin of victory, only 0.4 for the lesser candidate, and 1.7 points for the greater candidate.

I interpret the smaller error for the loser to reflect that fact that whoever admits to voting for the lesser candidate is all there is, while the greater candidate will enjoy some "everybody likes a winner" support among those who are not strongly decided.

When we look at the full series, an even more accurate picture emerges (click the earlier links for the interactive graphic):


After the summer stage was over, Obama clearly pulled away from Romney. Throughout the various ups and downs, Obama's ceiling was around 51-52, the peak being 51.2 on 10/27. That is only 0.1 point different from what he actually got. Romney's peaks throughout the series never broke 47, and that is what he ended up getting.

This indicates that the peaks are what the candidate will get -- a long-term maximum support level -- and the fluctuations below them are just people feeling wishy-washy from day to day.

Now we turn to the 2016 USC poll, which is the same methodology as before only in the present context:


After the summer conventions were over, Trump was slightly ahead or tied around Labor Day, after which he pulled clearly ahead. Trump is visibly the greater candidate, Clinton the lesser candidate.

Despite becoming tied again during the final phase of "uncertain about Trump," he's pulled away again. It's clear that his line has remained mostly above her line, and at worst tied. Unlike Obama in 2012, Trump is not the incumbent President and does not enjoy as decisive of an advantage. Other than that, its his line that obviously resembles Obama's line from last time.

So Trump will win, but by what margin?

Using the same insight of looking at peaks along the way, Trump's ceiling of support is 47-48, with the all-time peak of 47.8 on 9/18. Since Labor Day, Clinton's ceilings have been at 44-45, with a max at 45.1 on 10/23.

I'd say that makes the final outcome look like Trump 48, Clinton 45. I don't think Trump will get quite the last-minute bandwagon effect that Obama did because undecideds this time have consistently been uneasy about Trump the outsider, vs. Obama the incumbent. Even supposing he gets 49, I don't think that will come at the expense of Clinton's 45, but from the third party voters.

Speaking of whom, there are 7% left after Trump and Clinton. Stein has been polling around 2-3%, and her voters are going to be stickier because they're voting for her platform rather than merely against others or as an empty protest vote. Unlike her run in 2012, this time there was the Bernie phenomenon, many of whom are turning to her.

Johnson abandoned his campaign over a month ago, and his supporters cannot explain what he stands for, so he is just a vote-parking space for people who are uneasy about both Trump and Clinton. He polls at around 7%, but will be lucky to get half of that -- although probably close to it, given the wimpy cuckiness in the electorate about Trump the change agent. The other half just want to wash their hands of the election already and will simply abstain.

Final prediction, then: Trump 48, Clinton 45, Johnson 4, Stein 2, All Others 1. If the cucks for Johnson man up, it will shift 1 point from him to Trump (pretty iffy at this point, though).

Hope I'm wrong and it's a wider victory, but that's what the best poll suggests.

There's no evidence of a landslide, though, so please get that idea out of your head, lest you feel let down that we only won by 4 points. Remember: everyone said we were supposed to go down in flames during the primaries! Trump winning at all is a massive accomplishment for him and us the voters -- don't cry about it just because we're not going to win by 10 points!

But they say that conservatives have a lower tolerance for ambiguity, so it's either landslide or implosion in most of his supporters' minds, and we shouldn't take their claim about predicting a "landslide" literally, but simply meaning "we're going to win rather than lose".

Just one more week -- we got this!

20 comments:

  1. We can check this against Helmut Norpoth's "Primary Model," which predicts Trump's share of the two-party vote will be 52.5.

    My estimate is that he'll get 51.6 if it's 48 to 45, or 52.1 if it's 49 to 45. Either way, my projection from the USC poll is about the same as Norpoth's from his own model.

    http://primarymodel.com/

    ReplyDelete
  2. My ticket is Trump 49 - Clinton 46 - Johnson 3 - Stein 1 Others 1. The electoral vote will likely range from low 270 to something like 286, akin to George W. Bush's victory in 2004. He could reach 300 but it'll likely be close.

    ReplyDelete
  3. Don't forget the Electoral College! My personal prediction there is that Trump will win by 2 to 10 points.

    ReplyDelete
  4. What about Evan McMullin? He's supposedly at ~20% in Utah.

    ReplyDelete
  5. My prediction on the Electoral College, back in August, was 320 +/- 20. Wide range, but there was no telling. Looks like it will be on the lower side of that guess.

    http://akinokure.blogspot.com/2016/08/tempering-expectations-trump-takes.html

    McCuckin falls under All Others -- at 20% of Utah, that's a whopping 0.1% of the national vote. Only yuppie cuck Mormons are voting for him, he's a non-factor.

    ReplyDelete
  6. Random Dude on the Internet10/30/16, 7:39 PM

    I'm going to say the re-investigation is going to have a huge impact on the final results, based on the depressed Democrats I know personally who are ready to admit defeat already:

    Trump: 50
    Clinton: 45
    Johnson: 3
    Stein: <2
    McMuffin, Constitutionalists, Socialists, etc.: <1

    Trump ends up with 320-340 electoral votes, including a surprise victory that nobody saw coming, like Oregon or the whole state of Maine, not just the second district, who I think is going to Trump anyway.

    ReplyDelete
  7. Demo-wise, we've got:

    - Yuppie strivers/rich people who'll vote for Hillary (why would we expect people of actual or self-declared high status to vote for the rebel?). I'd expect transplants, and those in areas with lots of transplants, to be quite hostile to Trump.

    - Millennials (and to a lesser extent, late Gen X-ers) who'll vote 3rd party or stay home since they've never had a dog in the partisan fight and rather than prefer one candidate over the other, they just are dispirited by Boomers bickering badly for the millionth time. Obama was brilliantly marketed for a campaign against a wuss opponent; Hillary can't be marketed at all, especially when her opponent keeps pinning her down.

    - Long suffering middle aged middle class/working class whites who are energized by the lack of PC/identity politics/elitist purity rituals in Trump's campaign. High turnout, relatively speaking, though the greatest gains will be in the Rust-belt, the more down scale parts of the Northeast, and the South. In points further West, the two candidates will not be that popular since neither represents the cool affectations that Westerners prioritize. Easterners alike, Hillary wants power by any means necessary, Trump wants to create a glorious empire (enlarged by annexing the USA) to be overseen and eventually inherited by his kids. Either way, it's about concrete results, not about looking cool.

    - Blacks, reduced turnout (not THAT reduced cuz blacks tend to vote reliably, likely because the Dem machine that desperately needs blacks has conditioned them to vote).

    - Latinos. They don't vote, and I don't think they'll matter even in this election. Besides, around 1/3 will vote Trump anyway. Not partisan or numerous enough to really make much of a difference.

    Prediction? who knows. I just can't see the Pacific or the Northeast (PA and NH notwithstanding) going Trump. We've been in the culture war for so long that these culturally liberal (but behaviorally different as the East is more rooted) people to a large extent don't want to be on the same page as Middle America. I mean, since 1991 how much Hollywood garbage has portrayed Midwesterners as uptight crypto-fascists and Southerners as wild inbreds? On the more obscure message boards and blogs I read, there's a decent amount of PA people but the rest of the NE is absent not counting heartlanders who transplanted to the NE. The Scots-Irish and Germans seem to have really left a mark on PA. The rest of the NE is Yankee Anglo, "big 3" Ellis Island peoples(Jew, Italian, and potato-famine Irish), and non-whites who want nothing to do with heartland Teutonic/Scots-Irish culture right now.

    I do think we've got our best shot electorally via Mich., PA, FL, NC, Ohio, Iowa. I'm afraid that there are just too many liberal panty wastes in states like OR and Maine getting in the way to be able to wrest them from the Dems. Hell, even the Midwest BG states will be difficult enough given that each state has enough liberalism, diversity, and yuppie striving to give us trouble.

    ReplyDelete
  8. I saw a Bloomberg article claiming that per Trump insiders, not much headway was made WRT registering lower class whites who are the main Trump base. Well, at least in MN same-day registration is possible. Don't panic about early voter make-up either; early voting seems to be more common in diverse (e.g. Dem friendly) areas and partisans are more likely to vote early (duh, why wait when nothing will change your mind?). In all likelihood, plenty of once alienated non-partisan voters, some of whom may not even be registered right now, will show up to vote Trump come election day. And where are some of these early voting polls, anyway? Probably in crappy urban areas that whites avoid. Convenience wise, I can vote a quarter of a mile away on election day. Why travel further to vote earlier?
    One advantage to waiting is that I know that there's no way for anyone (the Dems included) to know my vote. I wouldn't want an early landslide for Trump, lest we encourage Dems to do even more cheating and throw more crap at Trump.
    Along those lines, I got something from some "progressive" PAC saying that they know my voting history (which was shown on the paper) and guilt-tripping me into voting (we'll know whether you did or not and it's a sacred obligation blah blah blah). They're trying to spur unmotivated younger people to vote. I wonder why.
    Oh, and the MSM and cucks have never acknowledged that the USC poll methodology and consequent results correctly indicated that Romney was a terrible candidate, while Trump is comparatively popular (and would be leading to an even greater degree had the media and the cuck brigade not repeatedly stabbed Trump in the back with purity tests, crude ad-hominem attacks up to and including Hitler comparisons, and baseless allegations).

    ReplyDelete
  9. Random Dude on the Internet10/31/16, 6:19 AM

    One other factor that seems to be important is the skyrocketing Obamacare premiums. In the past few days, millions of middle class suburbanites got letters saying their premiums will go up by double digits. You have one candidate who talks about increased competition lowering the rates and you have another candidate who spent the debates talking about how great Obamacare is. This can cost people hundreds of dollars out of pocket per month, will they take their chance with the Deal Maker in Chief or the candidate who thinks a program that just raised their rates by 30% is a positive historical accomplishment? Think about how those people, located in places like suburban Philadelphia, Columbus, Grand Rapids, etc. will take this.

    ReplyDelete
  10. Ag, you put a lot of thought into this and I like it!

    My approach to the outcome this whole time has been this: my gut tells me we're slightly behind the UK in making the pivot to populism. The transformation happens rather quickly and I've been hoping the US would be in November where the UK was in June or at least close to it.

    I feel we'll be in the (1-3%) range, just hoping we'll be on the plus side! Very possibly 4% like with Brexit because Hillary is so profoundly corrupt, but it's scary how so many people are sticking with this woman who's such bad news and an extreme sociopath. Banana Republic: shudder!

    ReplyDelete
  11. You know it's getting bad when the LA Times and USC won't release their own poll numbers.

    According to RCP's list, today's figure is Trump 47, Clinton 43. But you won't find that update on the LA Times pages (neither of the two), and the USC page refuses to load.

    Trump's final-week surge is knowledge that man was not meant to know.

    ReplyDelete
  12. The plug has officially been pulled on the Bernie people blindly stumping for Crooked Hillary.

    Nomiki Konst was a diehard Berner right up through the Convention, then instantly switched into bash Trump / shill for Hill mode. Became unreadable and unwatchable.

    Since the FBI has re-opened their investigation, she's instantly switched back to Bernie or bust mode.

    She goes on the major networks, not just the Young Turks. When someone with that level of visibility is abandoning ship, it's over.

    Anyone under 40 officially has permission to either abstain, vote Johnson or Stein, or write in Bernie's name.

    The only hardcore group among young people will be the Trump army.

    ReplyDelete
  13. Random Dude on the Internet10/31/16, 3:59 PM

    Everyone needs to keep black pilling their Democrat friends. My liberal friends have went to radio silence on this whole matter, which is surprising because for the past few weeks it was endless articles from Vox or 538 or Washington Post about how this election is all wrapped up. Even they know that it is lost. I might have to widen the victory gap Trump has from 5 points to 6 or 7. It has been an endless parade of bad news since Friday.

    LA Times is back up, I guess they got hit hard this morning. He now has a 3.4 point lead over Hillary and other than statistical noise where Hillary may gain by 0.1 points here and there, it will likely keep going up for Trump until the eighth of November.

    ReplyDelete
  14. "Chaldeans for Trump" sign at the Warren, Michigan rally.

    Make Middle Eastern Christianity Great Again.

    ReplyDelete
  15. Nah, it's going to be closer to 60-40 for Trump. Clinton may manage a handful of states.

    ReplyDelete
  16. Voters are too partisan for anything near that.

    Crooked Hillary could go an AR-15 shooting spree at a Los Angeles pre-school of color, and she'd still carry the deep blue states by double digits.

    ReplyDelete
  17. Weird things happen during political re-alignments nested within epochal transitions. We'll know soon enough.

    ReplyDelete
  18. Last time was 1896, going from Civil War / Gilded Age partisanship to the bipartisan Progressive and New Deal eras.

    McKinley won 60% of the EC votes, not a Reagan-style landslide. In 1904, it hit 70%, then 75% or more from 1920 to 1988, with a few exceptions that were divided ('60, '76).

    ReplyDelete
  19. Random Dude on the Internet11/1/16, 5:22 AM

    He keeps gaining in the LA Times poll and is up by 3.6 as of yesterday. The interesting takeaway is that he is now leading with 18-34 year olds. This has happened a few times before so it's nothing too new but it would be unprecedented that a Republican beat a Democrat with young voters overall, not just with whites.

    Interesting days ahead. Hopefully Trump can continue his trend of gaining momentum. Hillary is dredging back up the idea that Trump is a Russian puppet, something that surprisingly even the New York Times isn't buying. It made her look crazy and paranoid in the third debate when she kept hammering Trump on a fictitious Russian connection. If it's all she has, let us all enjoy the trainwreck known as the final days of the Hillary Clinton campaign.

    ReplyDelete
  20. They just interviewed the USC poll people on NPR, as the only poll that correctly called the election. I remember reading this post when it was new, thinking... we'll see whether this blogger is perceptive or off his nut. Congratulations.

    ReplyDelete

You MUST enter a nickname with the "Name/URL" option if you're not signed in. We can't follow who is saying what if everyone is "Anonymous."