Skip to content
 

What’s gonna happen in November?

Nadia Hassan writes:

2016 may be strange with Trump. Do you have any thoughts on how people might go about modeling a strange election? When I asked you about predictability and updating election forecasts, you stated that models that rely on polls at different points should be designed to allow for surprises. You have touted the power of weakly informative priors. Could those be a good tool for this situation?

I received this message on 4 Apr and I’m typing this on 9 Apr but it’s 17 Aug in blog time. So you’re actually reading a response that’s 4 months old.

What is it that they say: History is journalism plus time? I guess political science is political journalism plus time.

Anyway . . . whenever people asked me about the primary elections, I’d point them to my 2011 NYT article, Why Are Primaries Hard to Predict? Here’s the key bit:

Presidential general election campaigns have several distinct features that distinguish them from most other elections:

1. Two major candidates;
2. The candidates clearly differ in their political ideologies and in their positions on economic issues;
3. The two sides have roughly equal financial and organizational resources;
4. The current election is the latest in a long series of similar contests (every four years);
5. A long campaign, giving candidates a long time to present their case and giving voters a long time to make up their minds.

OK, now to Hassan’s question. I don’t really have a good answer! I guess I’d take as a starting point the prediction from a Hibbs-like model predicting the election based on economic conditions during the past year, presidential popularity, and party balancing. Right now the economy seems to be going OK though not great, Obama is reasonably popular, and party balancing favors the Democrats because the Republicans control both houses of Congress. So I’m inclined to give the Democratic candidate (Hillary Clinton, I assume) the edge. But that’s just my guess, I haven’t run the numbers. There’s also evidence from various sources that more extreme candidates don’t do so well, so if Sanders is the nominee, I’d assume he’d get a couple percentage points less than Clinton would. Trump . . . it’s hard to say. He’s not ideologically extreme, on the other hand he is so unpopular (even more so than Clinton), it’s hard to know what to say. So I find this a difficult election to predict. And once August rolls around, it’s likely there will be some completely different factors that I haven’t even thought about! From a statistical point of view, I guess I’d just add an error term which would increase my posterior uncertainty.

It’s not so satisfying to say this, but I don’t have much to offer as an election forecast beyond what you could read in any newspaper. I’m guessing that statistical tools will be more relevant in modeling what will happen in individual states, relative to the national average. As Kari Lock and I wrote a few years ago, it can be helpful to decompose national trends and the positions of the states. So maybe by the time this post appears here, I’ll have more to say.

P.S. This seems like a natural for the sister blog but I’m afraid the Washington Post readers would get so annoyed at me for saying I can’t make a good forecast! So I’m posting it here.

12 Comments

  1. Those Washingont Post  readers would also be annoyed at you for posting something <insert blog lag time here> out of date. But as we say in the computer science biz, one person’s bug is another person’s feature.

  2. zbicyclist says:

    I have been wrong about virtually everything this election.
    For example, here I am last October on this blog, predicting that by now we would barely remember Trump:
    http://statmodeling.stat.columbia.edu/2015/10/07/gallup-gives-up-the-horse-race-as-pollsters-confront-unprecedented-obstacles-the-biggest-name-in-the-business-backs-away/

    With that caveat:

    1. Trump’s been raising lots of money in July, but not spending it. Might be the Olympics, might be saving for one last big push, perhaps with some “October surprise”.
    2. If I was living in swing states Ohio, Florida, Pennsylvania or North Carolina, I would assume there’s no point to trying to watch TV in October; time to go all-Netflix.
    3. In terms of polling, I worry that since it’s so unpopular to admit to being for Trump, that there are a lot of closet Trump voters out there.

    • thisisntimportantatall says:

      > I worry that since it’s so unpopular to admit to being for Trump, that there are a lot of closet Trump voters out there.

      Is this real? I don’t really buy the argument that people are so closeted, I find it hard to imagine. It seems like it could just as well be the other way, with there just being fewer Trump voters. How is one supposed to know for sure?

      • Andrew says:

        This:

        From on our research, it seems that when a candidate is doing less well in the polls and the news media, his supporters are less likely to respond to polls. Thus it is reasonable to believe that Trump’s support right now is higher than is reported in the polls, just as, a few weeks ago just after the Republican convention, Hillary Clinton’s support was higher than was reported in the polls at that time.

    • Anon says:

      >In terms of polling, I worry that since it’s so unpopular to admit to being for Trump, that there are a lot of closet Trump voters out there.
      Wouldn’t that (Trump supporters reluctant to reveal themselves as such) have shown up in a poll-election results discrepancy in the primaries already?

    • zbicyclist says:

      In support of my third point, I note today in 538
      http://fivethirtyeight.com/features/election-update-leave-the-la-times-poll-alone/
      it shows the “house effect” for various polls by type.
      Notably, the house effect for automated pollsters (robocalls) show Trump higher.
      The house effects for live callers generally favor Clinton.

      So, it might be true that people are a bit ashamed of telling someone they favor Trump, and a bit more likely to be willing to punch a button on their phone. In the privacy of the polling booth? We’ll find out in November.

      And yes, I realize nobody is going to read this comment, posted a week after the original post. I’m still willing to go on record that Trump will likely do a bit higher in the actual election than polls predict.

  3. Llewelyn Richards-Ward says:

    Most of us in NZ, perhaps elsewhere, don’t understand how the concept of democracy always translates into a two party race in the USA. We are unfamiliar with the idea as our democratic processes operate differently. When I ponder how this impacts onto predictive outcomes, it is almost like a Bernoulli trial for each person times the population. So with a large populace, is it unsurprising that the numbers for each candidate tend toward the probability of 0.5 — is this some variant of the law of large numbers? I also wonder how the interaction between someone being loyal to a party versus preferring a candidate impacts on prediction. One might expect a number of unpredictable core beliefs/biases would influence the ultimate voting choice of candidate or party. And then the electoral college thing confuses me completely as to how that impacts on who ‘wins’. As such, it seems to me that it would be very difficult to predict forced choice scenarios such as the US presidential election compared to scenarios in my country where political voting is spread over 4-5 parties, plus numerous small groups. I do wonder what the predictor variables are for your elections and how they are to be thought about in terms of models… we had an octopus here once that picked outcomes for a sporting event quite well — some form of octal counting basis we suspect.

    • Andrew says:

      Llewelyn,

      No, there’s nothing in the law of large numbers that would imply that election outcomes are 50/50. Back in 1984, Ronald Reagan won about 60% of the vote. Also, the electoral college is annoying and it affects campaigning, but when it comes to the actual election, the electoral college winner is almost always the same as the popular vote winner. (And, before you bring up the Bush-Gore race, in 2000 Gore won the popular vote and he also would’ve won the electoral vote had they counted all the ballots in Florida.)

      • Llewelyn Richards-Ward says:

        Thanks for the clarification Andrew

        Okay, so the argument of reverting to some mean of 50% is wrong, clearly. Therefore there probably are tangible predictors of outcome. So I read some more about this, trying to understand the party affiliation factors that might influence the election outcome. Here in NZ, preferred party equates to being in government (we use MMP) and, by definition, elects our Prime Minister. I am more puzzled. So the Democrats have 44% of the Senate, 43% of the House of Representatives, yet seem to be in front for Presidential hopefuls and are Pres. Obama’s party. The Democrats are listed as having 43.1 million members, compared to 30 million for Republicans (2012). So how does a stronger party get fewer seats in government but still elected the sitting president? I looked at time spent in office, 18 GOP Presidents for 88 years and 15 Democrat for 91 years (50/50 ish). So maybe party affiliation is not as helpful as it might seem, which is counterintuitive to me. I found some articles that suggest ethnicity, gender, State and other demographics influence outcome. Now I think I see why modelling at the State level is probably more useful, as you say above. Seems like a very difficult statistical exercise.

        • Andrew says:

          Llewelyn:

          I’m not sure what you mean by “tangible,” but in response to your three questions:

          1. Party identification and ethnicity are the most effective predictors of vote preference.

          2. Because of districting, Democrats can win approx 50% of the vote for Congress but have less than 50% of the seats.

          3. There’s a chunk of moderate voters who prefer divided government, so they are motivated to vote for a Republican president when the Democrats control congress, and vice-versa.

          • Martha (Smith) says:

            Re 2: I’d say “gerrymandering” rather than “districting” — especially when talking to someone from another country. (I may live in the most gerrymandered county in the country; https://en.wikipedia.org/wiki/Texas%27s_25th_congressional_district might give some insight. But if it doesn’t seem to make sense — well, that’s Texas politics.)

            Re 3: There are also people who vote for candidates rather than parties. For example, in one presidential election year, my mother voted Democratic for president and Republican for governor; my father did the opposite.

Leave a Reply