Is Heitkamp Really DOA In The Election This Time? 2018 GE Edition

[Co-Chair Bagel23]

The mood over Heitkamp over the past few days even on a platform such as Atlas which tends to be very optimistic for democrats has been getting pretty somber and bleak. I think she probably loses by mid single digits, but I don't think she is DOA in this race quite yet. I'd still give her a 15-20% of holding out. That's not saying much, but I still think that she still has a political pulse, just very faint, but the chance for an upset is too big for me to say DOA. Do you think she is DOA this time around?

[Xing]

I wouldn't say she's DOA just yet, but it certainly doesn't look good for her.

[BudgieForce]

Ask me in 2 weeks.

[SCNCmod]

It looks likes she closed pretty strong in 2010... here are the RCP Polls from 2010:
https://twitter.com/DC_para/status/1048063345847427072

[Antonio the Sixth]

I wouldn't say she's DOA just yet, but it certainly doesn't look good for her.

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

She is not DOA. Her position seems to be is a bit worse than it was thought to be a week or two ago, but not hugely so. The recent polls are bad (and indeed, should be taken into account, which is the only reason why her position is a bit worse than before), but as has been said many times, ND polling is uniquely difficult and unreliable for obvious and well-trodden reasons.

So, insofar as people's opinions of her re-election prospects have deteriorated, that is primarily because those opinions were unjustifiably rosy to begin with - see for example the 538 faux-"fundamentals."

She is running in North-Fricken-Dakota, and that fact is not something that was just discovered a few days ago. Let me say it again. North-Fricken-Dakota.

That does not mean she is DOA. She won previously, after all, in a year that is likely to be less good (but not bad) for Dems overall. And that fact - that she was elected previously in a year that is likely to be less good (but not bad) for Dems overall was previously, and still is today, the best reason to not write her off or think that she is DOA.

She is in no way assured of being re-elected, and is not in any way a clear favorite, but that ought not to be breaking news to anyone who knows anything about politics, or anything about North Dakota.

This is one of those races where there is high variance in what you should expect. It is quite possible she could win fairly narrowly, and also is quite possible she could lose by 10 (or more).

[TML]

As I said earlier, it is still too early to count her out on the basis of limited polling.

At this stage six years ago, one could have also made the case that she was on track to lose based on available polling back then. In fact, during that year, she was considered an underdog by several major political prediction outlets even during the final days leading up to Election Day. Plus, 538's fundamentals favor her, and 538 notes that over the past 30 years, the polls have more often than not moved in the direction of the fundamentals during the final stretch. Back in 2012, that's exactly what happened, as her deficit had narrowed considerably by late October.

I will reserve judgment for now, but I will say this: if there are high-quality polls conducted in late October/early November which show her with a deficit as big as now (if not bigger), then I can probably say that she is on track to lose. On the other hand, if the polls at that time show a substantially narrower margin than now (or even a lead for her), then the race is, at a minimum, a jump-ball.

[Pericles]

Can someone explain why 538's fundamentals are 'faux fundamentals' rather than just stating it as fact? Something conflicting with your preconceived notions is not fact, and they clearly show where they're coming from with these numbers so it's not some magic number someone pulled out of their ass(unlike so many numbers on Atlas).

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Can someone explain why 538's fundamentals are 'faux fundamentals' rather than just stating it as fact? Something conflicting with your preconceived notions is not fact, and they clearly show where they're coming from with these numbers so it's not some magic number someone pulled out of their ass(unlike so many numbers on Atlas).

There is a breakdown of what goes into their "fundamentals"

https://projects.fivethirtyeight.com/2018-midterm-election-forecast/senate/north-dakota/

Heitkamp gets +13.6 for incumbency.

In particular, this advantage that the model is giving Heitkamp is so large because North Dakota is a small state.

This is something that Nate Silver got from regressing past election results, and found that it helped to explain those past election results (at least given the other factors in the model).

But making a good model is more than just throwing things into a model and finding that they have good at explaining the past. Making a good model is also dependent on having good domain knowledge, not just on throwing in any old thing that improves the fit of a model. You can overfit data, and fit it in a way that explains the past better, but makes it less likely to explain the future better.

Nate Silver's domain knowledge is certainly better here than most pundits, but his domain knowledge for House/Senate midterm elections is not as good as his domain knowledge of Presidential elections, or of baseball.

If his domain knowledge were greater, he would likely be more skeptical of incumbency advantage, and in particular would be aware that there is a general trend towards it becoming less important over time, which means that it is likely to have less explanatory power for predicting future results than for explaining past results on which the model is fitted.

[Pericles]

Silver seems to have considered incumbency advantage decreasing, he gave Heitkamp a large one as it is a small, demographically distinctive state which he found have stronger incumbency advantages. Nelson's advantage is only 5.9% because Florida is a large state, and Cruz's is 5.4% similarly. The advantage in the House is also smaller than in the Senate. Furthermore, you could argue that in 2018 incumbency advantage is much smaller, that might be true, but if there's insufficient evidence then inserting your assumptions into the model would make it less reliable. Furthermore, this isn't that surprising, Democratic incumbents have been doing pretty well this cycle overall(in comparison to their state's partisanship), and even in recent elections incumbents have done well. Naturally these models aren't perfect, perhaps your assumptions are right but making a model driven by assumptions is a bad idea. Just because a model's findings are surprising doesn't mean it's invalid, I'll refer you to what Nate said on this;

You must be logged in to read this quote.

https://fivethirtyeight.com/features/election-update-why-our-model-thinks-beto-orourke-really-has-a-chance-in-texas/

[Former Dean Phillips Supporters for Haley (I guess???!?) 👁️]

Silver seems to have considered incumbency advantage decreasing, he gave Heitkamp a large one as it is a small, demographically distinctive state which he found have stronger incumbency advantages. Nelson's advantage is only 5.9% because Florida is a large state, and Cruz's is 5.4% similarly. The advantage in the House is also smaller than in the Senate. Furthermore, you could argue that in 2018 incumbency advantage is much smaller, that might be true, but if there's insufficient evidence then inserting your assumptions into the model would make it less reliable. Furthermore, this isn't that surprising, Democratic incumbents have been doing pretty well this cycle overall(in comparison to their state's partisanship), and even in recent elections incumbents have done well.

Let me restate and try to be more specific, also. The problem with the way incumbency is treated is not just that its effect is decreasing over time (though this is a problem, and is really mostly a function of increasing partisanship). That is just one aspect/example of the problem. More broadly, or additionally, the problem with incumbency is also is that it is standing in for omitted variables in the model. In this sense, the problems with how incumbency is treated in the 538 model are actually similar to the conceptual problems with the idea of "candidate quality" - it is to a significant degree simply a residual variable - whatever cannot be explained by other actual factors tends to get picked up and thrown into the bucket of "incumbency" (which means that the variable called "incumbency" in the model is not actually measuring true incumbency).

One of the major and important differences between Congressional/Senate elections and Presidential elections (and here is where domain knowledge in recognizing and understanding this difference comes in to play) is that most elections in Senate and House races are not seriously contested. In Presidential elections, it is true that some states are not contested (in the sense that the candidates run no deliberate campaign in those states), but nevertheless the voters have much more information than in Senate/House races that are not contested and the races are higher profile, and they still get a lot of news about the race. This is one of the main reasons why Presidential results are generally easier to predict and more systematic, because Presidential races are always contested seriously, to pretty much the maximum extent possible, whereas in Senate/House elections, this varies a great deal.

And in elections that are not seriously contested, incumbents do very well. This is not per se because they are "incumbents" and is not really "incumbency effect" per se - rather it is simply that the election was not seriously contested - but nonetheless if you run a regression of election results data that includes, for example, the 57th time Kent Conrad, Pat Leahy, or Chuck Grassley ran for a (basically uncontested) re-election against a non-serious challenger, then you will end up finding an artificially inflated effect of "incumbency." This is because your specification *should* include, essentially, something like a dummy variable for "this election was actually seriously contested." In the 538 model, fundraising does actually partially serve that purpose, but it won't ideally capture the true thing that is making a difference here. Name ID - if there were data for this and it were included, could similarly serve as an imperfect proxy variable for this.

You must be logged in to read this quote.

OK, here I just have to say, all modeling is necessarily making assumptions. Models are inherently foundationalist constructs used to explain data. The process of specifying a model consists entirely of making choices about what variables to include and what variables to not include (and also what functional forms and other methodologies to use). These choices are inherently arbitrary and subjective, but yet some choices are definitely better than others (and that is where domain knowledge and also, indeed, intuition comes in). So any such idea that there is some way to make a model that relies more or less on making assumptions is just wrong.

Indeed, a model is not "invalid" just because its findings are surprising (but in truth it is a category error to apply the concept of "validity" to models - testing "validity" is just not something that mathematical models do or can do). But if a model's findings are too surprising, that is often a good hint that either there is some problem or limitation with your domain knowledge, or else that you are missing important omitted variables, or maybe that you are using a flawed/inappropriate functional form somewhere, etc. Surprises and things that simply don't make sense are hints that your model can be improved. Experience and time spent looking at these anomolies and finding order/patterns within the swirl of anomolies hones your intuition and helps you to then go back and make better models that will not omit the wrong variables, that will be specified in an alternate way, and also that will not overfit. Model making is really, above all, a matter of art and judgement.

[Sir Mohamed]

She's not DOA, but very vulnerable. If Dems lose a seat, it's ND. That's what I said early on. Hope it will be the only seat we lose.

[Adam Griffin]

Until we see more polling and/or a real trend developing, I'll continue to believe that MO is our most vulnerable seat.

[IceSpear]

The double standards here are pretty amusing. You'd all laugh at anyone who thought a Republican consistently behind by double digits in a deep blue state could still win, and for good reason.

[jamestroll]

If Kavanaugh gets confirmed without her vote I suspect the chaos will lighten and the race could tighten significantly.

[SnowLabrador]

She is now that she's voting against Kavanaugh.

[Figueira]

Not enough info to call her "DOA" but I'd be surprised if she pulls this off.

[Politician]

Not enough info to call her "DOA" but I'd be surprised if she pulls this off.

[GM Team Member and Senator WB]

Not enough info to call her "DOA" but I'd be surprised if she pulls this off.

this but also this:

If Kavanaugh gets confirmed without her vote I suspect the chaos will lighten and the race could tighten significantly.

[Storebought]

Yes, obviously. If this were a Republican running with these poll averages, for so long, everyone commenting on this thread would agree.

Just because ND is "cheap" to advertise in doesn't mean that Democrats should spend money advertising there. Just because something is cheap doesn't mean its worth buying.

[Cold War Liberal]

Not enough info to call her "DOA" but I'd be surprised if she pulls this off.

this but also this:

If Kavanaugh gets confirmed without her vote I suspect the chaos will lighten and the race could tighten significantly.

[Brittain33]

Can someone explain why 538's fundamentals are 'faux fundamentals' rather than just stating it as fact? Something conflicting with your preconceived notions is not fact, and they clearly show where they're coming from with these numbers so it's not some magic number someone pulled out of their ass(unlike so many numbers on Atlas).

There is a breakdown of what goes into their "fundamentals"

https://projects.fivethirtyeight.com/2018-midterm-election-forecast/senate/north-dakota/

Heitkamp gets +13.6 for incumbency.

In particular, this advantage that the model is giving Heitkamp is so large because North Dakota is a small state.

He had to do that after Montana and North Dakota were the only two states he called wrong in 2012, and called badly wrong.

[KingSweden]

In wouldn’t say DOA, but she’s definitely not favored as of today

[Pyro]

If Kavanaugh gets confirmed without her vote I suspect the chaos will lighten and the race could tighten significantly.

Exactly.

[Pollster]

No reason to put too much stock in North Dakota polling, especially public/internal-made-public North Dakota polling.

Looking at the full picture, things don't look good for her, but she is sitting on a massive heap of money (by North Dakota standards), no doubt raised a massive amount yesterday, and expectations are lowering. If anybody is in a position to close strong with a last-minute push, it's her.