Princeton Election Consortium

A first draft of electoral history. Since 2004

Thoughts on Faulty Priors

January 22nd, 2016, 9:15am by Sam Wang

Update: Reader Sourav takes a different view: “In this age it is hard for a Black Swan event to happen without any indication….People touting the ‘But Herman Cain was ahead this time 4 years ago’ theory were misreading the data. No one in 2012 had a consistent lead for so long. Trump has been the most retweeted candidate, highest Google search volumes. If the data pundits ignored all this, they did it at their own peril.” Sourav also has some analytics of Trump’s unusually large announcement bump here.

Like I said two weeks ago, Donald Trump is the strongest GOP Presidential front-runner since George W. Bush. Nothing’s happened to change that.

In this week’s news, Sarah Palin endorsed Trump. Then Bob Dole said Trump would be better than Ted Cruz. There are exceptions, but that’s quite a range. Belatedly, The Upshot and Nate Silver are coming around, which is quite a change.

This leads me to a question I have been pondering: what can we learn when quantitative punditry goes off track?

At its best, “data punditry” can help us see past the noise of individual perspectives. Good analysis can help us separate concrete evidence from our biases. But the problem is that we still have to have some prior framework for interpreting data. In this case, Silver and The Upshot weighed in at a time when polls were not very predictive. They did the natural thing: examine other indicators. Now Silver is finally catching up to the polling data, and The Upshot is starting to reflect Trump’s dominance.

What if our biases are so strong that we are unable to recognize a major phenomenon even as it stares us in the face? I am familiar with this experience, having committed a significant error myself in summer 2014.

This year, the major phenomenon is the durability of Trump in a highly divided field. Last year, analysts faced a conflict:  (1) polls were starting to point strongly toward Trump as more than a flash in the pan, yet (2) traditional markers of a strong candidate (“The Party Decides”, fund-raising) pointed toward someone else. And mentally committing to “The Party Decides” made it hard to accept concrete evidence of public opinion. Which is how some data pundits ended up lagging behind more traditional journalists. Will Trump get the GOP nomination? I can’t say with certainty, but it seems probable. Now, essays like The Six Stages of Trump Doom sound a lot like “non-data” punditry. Very few analyses from summer are wearing well.

As I said, I am no stranger to this kind of mistake. In August 2014, I didn’t see that year’s wave election coming. About a month later, this reality was forced upon me, and I gradually changed my tune. So yes, predictions can fail. The question is: what can one learn from failures when they happen?

Let me step back and just offer a few thoughts dating back to 2004, when I first got into this game. At that time, a state-poll-based snapshot of Kerry v. Bush gave a helpful picture of the campaign’s ups and downs. Electorally speaking, things went right down to the wire, a tough situation for making predictions. There was some media attention that year, but not that much. The big boom in election data analysis, 2008, was a fairly easy year, since Obama was pretty obviously headed for victory. In such a safe environment for forecasting, there was room to drill into the details of pollster reliability, that kind of thing. For some people, it was nerd heaven. The same was true for 2012.

In 2016, things may not be so favorable for the data nerds.

Most predictions are predicated on the idea that past history can be extrapolated to predict the future. Ray Fair’s Bread and Peace model, “The Party Decides” – these decades-long trends are used to extrapolate into the near future. But these models are only as good as the likelihood that the trends will continue. They are only useful under “normal” conditions – and can’t capture “black swan” events that upend the whole playing field. And once in a while, one of the major political parties undergoes a major realignment, as XKCD’s Randall Munroe has graphically illustrated.

Since the 1980′s, the normal condition has been that the Democratic and Republican parties have been stable, even as they drifted apart from one another. 1994 was a big turning point because it ushered in Gingrich-style polarization. That led to rightward movement, which reached crisis proportions during the Obama administration. Limbaugh begat Gingrich, Gingrich begat Palin, Palin begat Trump. Now here we are, with the most authoritarian one-eighth of the population calling the shots for one of the two major political parties. At the national level, the Republican Party seems to be at a breaking point (though it remains strong at lower levels).

Data punditry isn’t equal to the task of capturing this year’s weird events. For this reason I have been found prognostication by FiveThirtyEight and The Upshot to be an incomplete description. U.S. politics is in incredible ferment, and data punditry isn’t equal to the task of capturing it. We seem to be entering a phase where like other commentators, data pundits are running alongside the crowd, trying to make sense of events. I think the two types of reporting are going to need one another. At least until the general election campaign, when things should get predictable again.

Or not.

Tags: 2016 Election

55 Comments so far ↓

  • bks

    What could go wrong?

    One more big change for both parties? Microsoft has designed a cloud-based app for each party to make the caucus process easier. Iowa has an electoral process like nothing else in the country, and now it will be powered by similarly unprecedented technology.

  • Petey

    Back to the crucial matter of #DelegateAllocation:

    Via Politico: Said a top official of a rival GOP campaign: “If Donald Trump wins Iowa, I think he has won—period. Ted Cruz is supposed to win Iowa. If Trump wins, he’ll be on a trajectory to come out of the SEC primaries [March 1] with close to triple the delegates of anyone else.”

    Triple the delegates! And the then we head from the not-really-proportional states into WTA and ‘hybrid’ pseudo-WTA states. Still think if we stay on the current trajectory, this thing is over by 3/16 or 4/1.

    Due to different #DelegateAllocation methods, if Sanders can produce real strength after the first two states, the D process would last longer.

    (And worth noting that Nate Cohn directly asserted to me on Twitter that the NYT had a multi-part intensively reported series on #DelegateAllocation implications ready to run in January. Well, he’s got 3 days left, but I’m thinking the real takeaway is to never trust a man named “Nate”.)

  • Kevin

    538 published an “It’s Rubio or Bust” article:

    The line “Yet the notion that Rubio is the ‘Republican Obama’ also makes some GOP voters hesitant to support him” is a cute way of saying that the racism of some GOP voters may be self-defeating.

    • Amitabh Lath

      Thanks for the link but this article is emblematic of the basic problem of punditry.

      I was reading along happily until I started encountering statements like this:

      “It’s hard to imagine Clinton matching the share of Latino voters that Obama won in 2012…”

      and then:

      “It’s also hard to imagine Clinton matching Obama’s 60 percent among 18-to-29-year-olds against a candidate two decades younger…”

      Seriously? You know what’s hard to imagine? America voting twice for an African American president. You know what else is hard to imagine? Poor, working class whites supporting a flamboyant New York billionaire.

      Stop trying to imagine things and stick to numbers. That’s supposed to be your wheelhouse.

    • Petey

      “Thanks for the link but this article is emblematic of the basic problem of punditry … Stop trying to imagine things and stick to numbers.”

      - You do understand that elections, this far out from the voting, are as much social science as physical science, no? And just as with most social science, sole reliance on numbers will only get you so far. That applies equally to the invisible primary, and the general election 9 months out. There are limited strongly predictive numbers to rely on in such situations.

      - Wasserman is my absolute favorite numbers guy,and has been for many years. He really is basing his commentary on numbers, and making (mostly) sound observations deriving from them.

      - His core thesis on Rubio is shared by pretty much everyone. It may be wrong, but it’s shared across the board. The D’s universally fear Rubio above all, as you can see by very many of their actions and words. For the R’s, did you see last night’s Luntz pseudo-”focus group” / propaganda session? It was a pure Rubio commercial. And Luntz wasn’t doing that for his own amusement.

      Now, maybe everybody is wrong, and it’s all imagination. And I don’t agree with every single word of Wasserman’s piece. But still, it’s head and shoulders above almost all 538 pieces we’ve gotten this cycle.

      (Worth noting that parties very often don’t nominate their “best” general election candidate. Which is why I found the Nate Silver non-mea-culpa “We weren’t wrong, the GOP is wrong” argument so damn weird.)

  • bks

    If you have a strong stomach, read the transcript of Rush Limbaugh’s show today:
    It’s a paean to Trump.

  • JayBoy2k

    I saw this Poll and thought it would be worth sharing. Iowa Monmouth poll:

    ” Trump does better among those who do not have a history of taking part in party elections. This includes 44% support among registered Republicans who are general election
    voters and 50% support among registered independents who say they will attend their local Republican caucus on Monday.”
    Likely the pundits and prognosticators failed to take account of these non traditional voters in a Republican caucus primary. If there is a larger volume of voters (>>>122k) that carries Trump , that will be the story out of Iowa.

    • bks

      In that Monmouth poll (the real Monmouth, not the fake Monmouth) it’s Trump 30 Cruz 23. In the same poll six weeks ago it was Trump 19 Cruz 24. Yow!

  • Amitabh Lath

    Nate Silver’s mea culpa chastises the Republican “establishment” for not doing more to stop Trump, and sooner (maybe they were listening to experts like Silver who assured them that Trump would soon implode :)

    I think Silver and others who expected some fancy chess moves from “establishment” types suffer from assumptions of competency. We’ve all had experiences working with people who look great on paper (excellent GPA, good letters) who work out great usually, but cannot function in novel situations thinking on their feet.

    I suspect when Trump started winning, it was easier for everyone to assume a) he will fold soon, and b) someone else has got this.

  • adrian crutch

    …I always add the numbers in these “polls”…and they never add up to 100%…a quick look at four ranged from 96 to 92%…so are the “leftovers” the undecided and the quack candidates?…4 to 8% is a lot…

  • ThePunditsAreStupid

    You know, what is interesting is that a lot of pollsters will ask a question along the lines of: “Regardless of who you support, who do you think will be the eventual nominee”

    Trump has been clearly leading the question in those polls since at least October, and I think maybe even September. The likes of Nate Silver would have probably laughed at that at the time, but they are starting to look right. It would be interesting to see how accurate of a predictor this question has been in previous cycle.

  • Petey

    Pretty odd semi-mea culpa from Nate Silver. Seems to rely on two points:

    1) We read Our Particular Bible Wrong.

    2) We weren’t wrong. The Republican Party is Wrong.

    The piece is not without a number of decent insights, but Nate still doesn’t seem to be Grasping Reality with All Tentacles, as Brad DeLong’s current motto would have it. He’s still not coming to terms with his real Faulty Priors that have shaped his poor performance this cycle.

    I also found it interesting in that it linked to an Enten piece from mid-December I’d missed that I found revealing of the 538 mindset in three ways.

    - Enten was convinced Giuliani would be the ’08 nominee.

    - His lesson learned was the (correct) one in that polling during the ‘invisible primary’ has limited informational value.

    - However, when he tried to transfer that (correct) lesson to 2016, he still didn’t see the forest for the trees. He didn’t have a clue as to why Trump’s lead might have dramatically different meaning than Giuliani’s lead.

  • Amitabh Lath

    Thinking long term, it seems that the pollster model of Likely Voter will need to be more sophisticated. From what I understand, LVs are determined from some screening questions.

    I would posit that likelihood to vote be considered a function of candidates in play. In other words, a given voter (or voter demographic) would have different LV weights associated depending on who was running. So there may be a group that is extremely likely to vote if say Trump or Obama is on the ticket but not so much otherwise.

    I believe some polls do ask about how excited you are for candidate X, but the LV filter is still a constant.

    • 538 Refugee

      I think the telephone polling technique is probably dying. Illegal telemarketing is taking its toll on people willing to pickup the phone for a number they simply don’t recognize. Social media may be a possibility. I think it is going to take a fundamental change more than just tweaking the likely voter model.

  • bks

    Brad Delong is also thinking about priors

    …However, the herds and hordes of journalists and political scientists are not coming to grips with this. Rather than come to grips with this, they work hard to “save the phenomena” and save their models–analyzing the rise and durability of Donald Trump by making the smallest possible tweaks to what they thought last year. They are not stepping back and absorbing the lesson. They do not want to recognize that the rise and durability of Trump teaches them that what they thought last year was wrong. They do not want to face the reality that they need to pretty much throw everything away and start over….

    • RefreshStream

      But is it really a zero probability event? My explanation of the Trump rise is that it has been powered by his stance on immigration than anything else (remember this was the first position he put out). His celebrity, demeanor, and aggressiveness help, but what really fundamentally made him stand out was immigration. I saw the reaction in real time on alt-right blogs online.

      I think if you look at polling data on issues and try to find a large disconnect between the establishment consensus and something a large (but not majority) of people feel very passionate about, immigration is up there, and that points to an opening for a someone to exploit. So maybe the rise of someone like Trump wasn’t entirely a zero probability event?

    • bks

      DeLong is saying that it was a zero probability event, not that it is a zero probability event.

    • Amitabh Lath

      For a zero probability event to happen means that your visualization of the Cosmic All is simply wrong–or it would not strike you as a zero probability event.

    • RefreshStream

      bks – I should have written was, you’re correct, but as the rest of my comment was directed to Trump’s rise, I think the main thrust of my argument (the rise of Trump or someone like Trump in light of strong dissatisfaction with immigration among a portion of the population) still stands. Or maybe I’m engaging in post-hoc reasoning and drawing targets around where the arrows hit. But as there were commentators who saw the rise of Trump very early, I think it’s more likely that the “this was a zero probability event” crowd had faulty priors, even if they’re correct on criticizing their peers who haven’t adjusted their worldview.

    • bks

      Amitabh, that is, I think, DeLong’s point.

    • Amitabh Lath

      It’s a quote taken directly from DeLong.

    • 538 Refugee

      Reminds me of a customer.
      Customer, “Why can’t it just do this?”
      Me, “That would violate the laws of physics.”
      Customer, “Gxx Dxxxxx Government!”

      Yes I could make this stuff up, but sadly, I don’t have too.

  • Daz

    I reckon the problem is that 538 weighed in early. The temptation to say _something_ about Trump must have been overwhelming given that he is all the media is talking about, but basically a more prudent approach would be to say “The evidence on all of this is going to be weakly predictive until March.” There is no shame in saying “I don’t know.”

  • fladem

    Front runners at this point since 1980 (relative to Iowa)

    All who were over 40 nationally won.
    Only 1 exception:
    Clinton in ’08

    Front runners under 35 at this point
    Gart Hart in ’88 (lead nationally until Iowa)
    In ’12 Romney and Gingrich

    You are really wrong in one respect: Hillary’s national number was higher than Trumps by about 8 points, and her lead was bigger.

  • Amitabh Lath

    I suspect some retrospective analysis/post-mortem is inevitable. If Trump wins the early states expect most pundits to pen such pieces (if he loses, it will be pollsters tearing their hides).

    A major issue I hope gets examined is credibility. I suspect Trump built up quite a lot of credibility among his crowd with an unrelenting pursuit of Obama’s birthplace. The White House releasing the birth certificate was a major win because it showed they were being heard (seriously enough to be lied to, in their opinion).

    Concurrently, the very people who feature in works like “The Party Decides” were losing credibility. When it becomes a common trope that a candidate is going to “veer right” in the primaries and then “correct to the center” in the general (recall the “etch a sketch” comment from Romney’s camp), disillusionment is to be expected. One could say the Trump brigades are behaving rationally.

    • Petey

      “A major issue I hope gets examined is credibility. I suspect Trump built up quite a lot of credibility among his crowd with an unrelenting pursuit of Obama’s birthplace.”

      I think this is an insightful and valid point, Amitabh. But beware! You’re sliding down a slippery slope into … punditry!

      Which, of course, hits directly on my hobbyhorse of the thread: due to the low informational value of polling numbers during the “special case” of the invisible primary, you need some punditry to properly interpret the meaning of the polling numbers.

      Otherwise, one wouldn’t have been able to differentiate the meaning of Trump’s polling numbers from Giuliani’s or Dean’s or many similar cases. I think the Sin Of The Two Nates was simple atrocious punditry. They made poor decisions about how to interpret the meaning of Trump’s numbers.

      Your point here is one good corrective. Elsewhere in the thread, I laid out my real-time punditry on why I took Trump’s numbers seriously early on.

  • Sourav

    I have a slightly different view. I don’t think Trump’s strength was so unpredictable. Yes, initially we all thought this was a temporary bump, like Cain/Bachmann received in 2012 (in fact, I am rather ashamed to admit, I had written that this bump is normal for a candidate after announcing, .
    However, after a while, it was clear that Trump’s lead was consistent. People touting the “But Hermann Cain was ahead this time 4 years ago” theory were misreading the data. No one in 2012 had a consistent lead in the polls for so long, as Trump had had. If the data pundits ignored that consistent lead, they did it at their own peril.

    Yes, black swan events might upset everyone’s calculations, but in this social media age, it is hard for a Black Swan event to happen without any indication.

    Donald Trump has shown all indications of leading the race. He has been consistently up in the polls, he has been the most retweeted candidate, highest google search volumes. To disregard all this, is to disregard the data.

    Now, whether he will win only time will say. But those who underestimated him, in my view were ignoring many signs in the data.

    • RefreshStream

      I agree. 538 and Nate Silver irk me much more than any other website or commentator who consistently underestimated Trump because their whole raison d’etre was quantitative, data based prediction. Instead, they overweighted data that was clearly meaningless (endorsements), underweighted polling data, and disregarded other data that would have showed Trump was the leading candidate for a long time.

      My guess is they will try to spin this as “such an unpredictable event that data was useless” when, as you rightly point out, it was quite the opposite. If there is any justice in the world, Silver’s reputation will be far diminished.

  • Roger King

    Just want to make sure we are all aware no one has voted yet. Can you really claim Nate Silver or anyone else is way off on their analysis before we know how the voting plays out?

    • bks

      I don’t mean to speak for Sam, but I think he is saying above that Nate Silver is changing his analysis in a data-driven direction, not that he’s right or wrong.

  • 538 Refugee

    One of the few things I remember from Psych 101. We seem to have a ‘need to know’ and many don’t seem adverse to filling in the gaps with whatever is at hand when they don’t.

    If you don’t have enough data then you need to extend it.

    Easy Punditry Helper Recipe:

    1. Mix in staled hand picked historical data with your fresh data until you have the flavor and volume needed.
    2. Season to taste with your expectations.
    3. Strain off the error bars. (IMPORTANT! Do NOT omit this step!)
    3. Serve with a sense of authority.
    4. Have Reality Check Insurance just in case. (This step is optional depending on how susceptible your guests are to reality checks.)

    Seriously though, I think we all like to think we know and end up downplaying the error factor and the possibility there is stuff we just don’t know yet.

    • Kevin

      To be fair, there is a whole literature within political science of looking at factors that explain election outcomes without reference to polling. I think Drew Linzer manages to incorporate some of this information in a sensible manner–and it doesn’t hurt to look at polls-only approaches in the context of other indicators and information. That said, I sympathize with Amitabh’s point of view.

    • Amitabh Lath

      Drew Linzer is a great example of someone who is transparent about polls and their shortcomings and his approach is aboveboard and strictly Bayesian.

      My problem is people who look at the data and see that it is not giving them the answer they believe it should. They then pull in other variables (say, “unfavorables”, or “more people believe the moon landings were faked than like Trump”).

      This is actually quite common in undergrad physics labs (where the correct answer is well known, but hard to get).

    • 538 Refugee

      Kevin. I was careful to say “many” because there are some earnest efforts to try and stay neutral using data. Science dictates when your hypothesis doesn’t match the data you need to rethink your hypothesis. Politics is opinion though so if the opinion doesn’t match the data you can still change the opinion. That is the reason for the last sentence.

      I’m sure Trump had the high negatives the early polling suggested but that was a variable in the political equation, not a constant. That’s where many went wrong. We all knew that number wasn’t written in stone but we wanted it to be true and treated it as such. Look for that number to rise further in fact. After pandering shamelessly to the far right he is now doing the expected moderation and taking flack from the right because “deal making” is compromise which equals defeat to them but when they are left with the Trump/Sanders ;) choice which way will they go?

      Bottom line. Politics is opinion driven so it invites us to offer them up either consciously or subconsciously.

    • Petey

      “My problem is people who look at the data and see that it is not giving them the answer they believe it should. They then pull in other variables…”

      Sure. Randomly pulling in other variables to make your result “correct” is very bad practice and should be condemned.

      But my whole point is that the “special case” of the invisible primary requires pulling in subjective information about alignment of forces in a party to supplement pure data. In other words, some punditry, (which I well know is a dirty word to you), must be incorporated for the invisible primary.

      Otherwise, to take my above examples, how could you tell from data that Dean and Giuliani weren’t favorites, or that HRC in ’08 wasn’t the overwhelming favorite the data showed?

      (To show my work in the current race: I became an early believer in Trump’s polling data due to three entirely subjective factors: 1. The logjam in the ‘establishment’ lane. 2. The size/enthusiasm of Trump’s rallies in the South. 3. His ability to overwhelmingly win a battle against against Fox. That last item was probably most important for me.)

  • Matt McIrvin

    The thing is, 538 hasn’t really been doing data punditry when it comes to the Republican primary race. They look at their data, it tells them something they can’t believe and they cast about for something else to say. If Trump gets the nomination, it won’t be a strike against data-driven prediction, quite the contrary.

    • bks

      We have no alternative but to trust the data. Great discoveries start with the scientist saying, “Hmm, that’s peculiar.” not “Oh, that’s an outlier.”Unskewing doth make morons of us all.

    • Amitabh Lath

      Not to reignite this old argument, but 538 has always been in favor of adding “fundamentals” to the poll numbers.

      One can point out that this only makes uncertainties impossible to determine until blue in the face, but pundits gotta pundit.

    • Matt McIrvin

      Saying that results are outliers is fine, as long as you have statistical reason to believe that they really are outliers, and your observations are consistent. What we often get in election years is people touting the results of the latest SHOCK POLL when there are a dozen polls, which is usually a sign of cherry-picking.

  • Amitabh Lath

    The moral of this cycle: if you want to know how people will vote, ask them (polls). Yes, polls can be wrong (UK, Israel) but better than all other metrics, and at least one can start some sort of error analysis (likely voter filters, unsampled demographics).

    When people start adding variables because they do not like/believe what the polls are saying, beware.

    Remember all the analysts that said “Look at Trump’s unfavorables, those numbers are impossible to move…” Except they did.

    • Matt McIrvin

      Remember, Sam himself added a variable in 2004 that led him to predict a Kerry win, when his aggregate of untweaked election-eve polls was calling the election exactly as it went down.

      And then he looked back at what he had done and didn’t do that again, which is one of the reasons I keep coming back here.

    • Petey

      “The moral of this cycle: if you want to know how people will vote, ask them (polls).”

      Actually, while polls certainly shouldn’t be ignored, they have somewhat limited informational utility during the ‘invisible primary’. To pick three examples out of many:

      - Think of Rudy Giuliani in fall 2007.
      - Think of Hillary Clinton in fall 2007.
      - Think of Howard Dean in fall 2003.

      Now, I bet against all those plays in the prediction markets, because I thought they were wildly overpriced due to the polls. (Clinton was the only one I thought to be a genuine favorite, but even she was wildly overpriced.)

      This year, sure, Trump was underpriced despite the polls. But polls alone weren’t why I’ve thought Trump was the favorite for months. The invisible primary is odd stuff…

    • Amitabh Lath

      Yeah, I’m not saying polls are perfect (I pointed out UK and Israel as major poll fails, Petey points out other instances).

      What I am saying is everything else is crap.

    • Petey

      “Yeah, I’m not saying polls are perfect (I pointed out UK and Israel as major poll fails…”

      I honestly think it goes further than that Amitabh.

      Polling in the ‘invisible primary’ isn’t necessarily wrong, as in the UK and Israel general election examples. Polling in the ‘invisible primary’ is just of limited informational utility.

      The polls showing Dean, Clinton, and Giuliani with massive leads weren’t off by double-digit margins. They were snapshots of something quite real with limited informational utility.

    • Amitabh Lath

      Yes, some people don’t make up their minds until the last minute, and some actually change their minds. Sometimes for no apparent reason! (Something happened for sure in the 3rd week of September 2014 to shift the midterm elections but damn if anyone knows what it was).

      People are just unreliable.

      Better to stick with subatomic particles when making measurements. Way simpler.

    • Petey

      I really don’t want to get in a noise over signal back and forth with you, Amitabh. I tend to generally agree with you.

      All I’m saying is that the invisible primary is a special case, where polls have far less informational value than during general elections, or even during actual primary season.

      You can see this in cycle after cycle after cycle. There are shifts in election season, but not the kind of multiple-double-digit shifts we regularly see in the invisible primary.

      Let’s say election season is Newtonian physics. Polls work pretty damn well as long as your data is good. The invisible primary is Einsteinian physics. Different rules apply. Massive poll leads may not mean much.

      (And FWIW, I’ve thought for many months the poll leader in the current GOP race was indeed the favorite. Polls just haven’t been the only thing I’ve been factoring in, because of the special case of the invisible primary.)

  • Petey

    One more example on ‘alignment of forces’ during the invisible primary. I’m old enough to remember the ’92 Dem nomination race. And I knew Clinton was a dead-lock for the nomination in Fall ’91 even before “consensus favorite” Cuomo decided not to run. Why?

    1) No Jesse Jackson or White Southerner was running, thus leaving the South open to a total Clinton sweep.

    2) Harold Ickes, a highly influential D power broker who’d been with Jesse in ’88 signed up with Clinton. That gave Clinton real power on the left, to go along with the center/right power he brought himself. Full spectrum.

    After Ickes, game over. I always think Cuomo stayed out because he recognized the alignment of forces.

  • JayBoy2k

    I think that this morning’s National Review article attacking Trump directly is an indication of how far we are “off” the Party_Controlled and orchestrated election cycle. I especially liked this quote ” Trump nevertheless offers a valuable warning for the Republican party. If responsible men irresponsibly ignore an issue as important as immigration, it will be taken up by the reckless. If they cannot explain their Beltway maneuvers — worse, if their maneuvering is indefensible — they will be rejected by their own voters. If they cannot advance a compelling working-class agenda, the legitimate anxieties and discontents of blue-collar voters will be exploited by demagogues.
    Read more at:
    Very few voters are in lock step with everything the National Parties endorse. I am certainly not a democrat or progressive, but I have some progressive views and do not agree with the policy or procedures of the establishment GOP leadership. When a significant chunk of your Party’s base rebels against business as usual ( certainly Trump/Cruz/Carson voters but maybe Sanders voters also), many of the predictive factors may not apply. Get your popcorn, It is going to be a wild ride.

  • Petey

    “Data punditry isn’t equal to the task of capturing this year’s weird events. For this reason I have been found prognostication by FiveThirtyEight and The Upshot to be a fairly incomplete description. U.S. politics is in incredible ferment, and data punditry isn’t equal to the task of capturing it.”

    A few points:

    - I think your headline captures the essence incredibly well, Sam. “The Party” decided on Goldwater in ’64, but, ummm, well…

    - This has indeed been a miserable performance from the two Nates. Just repeatedly ignoring obvious evidence for months and months on end.

    - Pure #DataJournalism is a relatively weak tool among the many others during the ‘invisible primary’. And not just during ‘weird’ years. That period is best analyzed by not only by parsing the polls correctly, but more crucially, by looking at how the pertinent forces are aligned. For example, I thought the turning moment in the GOP race was during the first debate, when Fox threw all their big artillery at Trump, and Trump walked away victorious and unscathed. That should’ve been the signal that the GOP countervailing forces were paper tigers.

    - In accord with the last point, I’ve found the consistently best analysis of the race coming from reporter/pundit Dave Weigel. He’s been close enough to the ground that he’s been ‘getting’ the alignment of forces.

    - Looking at the implications of delegate allocation is one place where #DataJournalism can play a highly useful part during this phase. And the two Nates have just ignored the topic. Kudos to you, Sam, for being early to the topic, even before you started doing actual models.

    (The original title of the post was The Party Decides – Or Falls In Line. – Sam)

    • pechmerle

      The Party did not decide on Goldwater in 1964. From Wikipedia:

      “He alarmed even some of his fellow partisans with his brand of staunch fiscal conservatism and militant anti-communism. He was viewed by many traditional Republicans as being too far on the right wing of the political spectrum to appeal to the mainstream majority necessary to win a national election. As a result, moderate Republicans recruited a series of opponents” most notably including Nelson Rockefeller to run against him in primaries.

      The conservative insurgents carried Goldwater to victory in the primaries and the convention, though Goldwater didn’t lock up the nomination until California’s winner take all primary in June.

      As I said in the prior thread, despite that thrashing the party still moved toward the right, just happening to get lucky with Reagan as the conservative preference in 1980 — a unique bundle of campaigner attributes.

      This year — Trump has a very different, but still unique, bundle of campaigner attributes. In some ways, he’s more like Goldwater: “Extremism in defense of liberty is no vice.” In May 1964 Goldwater suggested the use of “low-yield atomic weapons” against North Vietnam.

      Times change; I don’t claim to know what will happen in the rest of this year any more than anybody else can. But a view from the history suggests a general election outcome more like Goldwater’s than Reagan’s, just because of the huuuuge difference in personal attractiveness between these candidates to the general electorate.

    • Mark F.

      While a more moderate candidate would have likely done a bit better than Goldwater, no candidate was going to beat the Democrats that year. Only an insane person could have imagined Nelson Rockefeller winning a year after Kennedy’s assassination and Johnson’s popular welfare and civil rights legislation.

  • bks

    Richard Hofstadter (1916-1970) wrote:

    It is ironic that the United States should have been founded by intellectuals, for throughout most of our political history, the intellectual has been for the most part either an outsider, a servant or a scapegoat.