538 Model Megathread

[GeorgiaModerate]

Another huge difference is that Silver's model assumes that the states move together and do not act totally independent of each other. If Trump gains in North Carolina, the model assumes that he also makes some gains in Virginia, unless/until there's polling evidence to contradict that. Not the case with other models.

Yes, which is something that always bothered me. Like... does a bad IA poll impact MO, IL, MN?

Really? That's one of the model's main benefits to me. Assuming the states move independently makes no sense.

I agree to a point, although assigning the proper amount of correlation between the states will be tricky; for example, MS and AL are going to be more alike in their movements than MD and WV, to pick another pair of neighbors.  But this also has the potential to overstate the effect of outliers or other noise.

[Fmr President & Senator Polnut]

Another huge difference is that Silver's model assumes that the states move together and do not act totally independent of each other. If Trump gains in North Carolina, the model assumes that he also makes some gains in Virginia, unless/until there's polling evidence to contradict that. Not the case with other models.

Yes, which is something that always bothered me. Like... does a bad IA poll impact MO, IL, MN?

Really? That's one of the model's main benefits to me. Assuming the states move independently makes no sense.

I agree to a point, although assigning the proper amount of correlation between the states will be tricky; for example, MS and AL are going to be more alike in their movements than MD and WV, to pick another pair of neighbors.  But this also has the potential to overstate the effect of outliers or other noise.

Yes, states with similar demographics and economies/economic mixes will likely shift together. But say AZ and NM... there isn't really any pattern as to their results over the years.

[Mallow]

Another huge difference is that Silver's model assumes that the states move together and do not act totally independent of each other. If Trump gains in North Carolina, the model assumes that he also makes some gains in Virginia, unless/until there's polling evidence to contradict that. Not the case with other models.

Yes, which is something that always bothered me. Like... does a bad IA poll impact MO, IL, MN?

Really? That's one of the model's main benefits to me. Assuming the states move independently makes no sense.

I agree to a point, although assigning the proper amount of correlation between the states will be tricky; for example, MS and AL are going to be more alike in their movements than MD and WV, to pick another pair of neighbors.  But this also has the potential to overstate the effect of outliers or other noise.

Yes, states with similar demographics and economies/economic mixes will likely shift together. But say AZ and NM... there isn't really any pattern as to their results over the years.

I'm not sure what you think the 538 model does, but my understanding is that it uses the actual covariances between all the states over many election cycles in order to determine the proper correlation values. So, if NM and AZ have not historically shifted together, 538's model would not use an AZ poll to affect NM's forecast. The main way this method could fail is if certain states that have shifted together in the past do not shift together this cycle, but I think this would not be that big of an effect in the grand scheme of things.

[Fmr President & Senator Polnut]

I just think you have more faith in the model than I do.

He's too busy covering his arse and so it's far more skittish and prone to big jumps off wonky data points.

[Mallow]

I just think you have more faith in the model than I do.

He's too busy covering his arse and so it's far more skittish and prone to big jumps off wonky data points.

It's actually a very common statistical method, used in many fields (including my own) to get the most accurate "snapshot" given sparse data prone to error. It's well-studied and there's plenty of reason to have "faith" in the method.

This whole idea that the model is being somehow manipulated in real-time to suit the preconceptions of Silver et al. is silly and misinformed. The "skittishness" of the model is being overstated, too--the percentage chance for Clinton to win (n.b. NOT the change in the margin) in the "polls-plus" version of the model has changed just 13% in the past three weeks. That's the "big" change we're discussing here. And it's been driven by lots and lots of data, not just one or two data points.

[Fmr President & Senator Polnut]

Ahhhh - you're a stats person, I'm not. I'm only going by what I read and perceive.

Go with God

[dspNY]

I think his model is too volatile compared to 2012, which was not only accurate, but mirrored President Obama's internal polling that always showed him ahead

[Mallow]

I think his model is too volatile compared to 2012, which was not only accurate, but mirrored President Obama's internal polling that always showed him ahead

I'd point out that the polls-plus and the polls-only models both have Clinton ahead still, and only the latter of the two has ever shown a trump lead... and then only 50.1% to 49.9% after the RNC for one day. So basically, they have both shown Clinton ahead the whole time.

[emailking]

After he put in those Google polls, especially the DC one, I lost all faith in this model.

You are wrong. He wrote an article regarding that and he weights those tracking state polls from google and ipsos not too much.

Yep, but there are so much more of these trash polls being churned out that, even if he weighs them less, they eventually overwhelm the good pollsters, like Monmouth.

Because all of those "trash polls" tell you something. You could have polls that are all off by an average of 20 points but if you accumulate enough of them, you would know exactly where the race is.

Even if you have one or two they still tell you something. If you have 2 polls from pollsters that you believe tend to misfire by about 20ish points, but both of them show Trump ahead, then it is more likely that Trump is ahead than Clinton is ahead. Completely discounting the polls is throwing out useful and valid information. Counting the polls but weighting them down due to their suspect nature is a better course. This captures the trend they're showing while preventing the magnitude from overwhelming the less suspect polls.

If you get enough then the net result may outweigh the better polls, but that may be perfectly valid. What you have in that case is evidence that the better polls are wrong.

[Wiz in Wis]

I think his model is too volatile compared to 2012, which was not only accurate, but mirrored President Obama's internal polling that always showed him ahead

I'd point out that the polls-plus and the polls-only models both have Clinton ahead still, and only the latter of the two has ever shown a trump lead... and then only 50.1% to 49.9% after the RNC for one day. So basically, they have both shown Clinton ahead the whole time.

Again, it's also harder this year with many more methodologies. The NYT-Upshot poll also pointed out how varying enthusiasm/filters can affect LV screens. I'm not going to rip on 538 for including as much information as possible, just because I don't like the number I see.

[Likely Voter]

RNC Nate has somehow 'forgot to add in the Monmouth poll that has Clinton up 5 in Florida or the St Leo  poll. Of course Nate did remember to add in the Monmouth poll that has Trump up 8 in Iowa. Nate is looking much less credible this cycle.

they are there. The polls are sorted by weight not just age. Although the weighting is always a bit hard to figure out. There are five polls weighted higher than the Monmouth poll (which is the most recent and from a A+ rated pollster)

[Beefalow and the Consumer]

RNC Nate has somehow 'forgot to add in the Monmouth poll that has Clinton up 5 in Florida or the St Leo  poll. Of course Nate did remember to add in the Monmouth poll that has Trump up 8 in Iowa. Nate is looking much less credible this cycle.

they are there. The polls are sorted by weight not just age. Although the weighting is always a bit hard to figure out. There are five polls weighted higher than the Monmouth poll (which is the most recent and from a A+ rated pollster)

I'm guessing because at this point it's a statistical outlier.  That could change with another poll or two.

[mencken]

RNC Nate has somehow 'forgot to add in the Monmouth poll that has Clinton up 5 in Florida or the St Leo  poll. Of course Nate did remember to add in the Monmouth poll that has Trump up 8 in Iowa. Nate is looking much less credible this cycle.

they are there. The polls are sorted by weight not just age. Although the weighting is always a bit hard to figure out. There are five polls weighted higher than the Monmouth poll (which is the most recent and from a A+ rated pollster)

I'm guessing because at this point it's a statistical outlier.  That could change with another poll or two.

It is because Monmouth used a smaller sample size than the other polls (i.e. higher margin of error)

[Likely Voter]

It appears the debate about 538 model vs other models has got the notice of Nate Silver. Here is a series of tweets he sent out early this morning...

You must be logged in to read this quote.

He also retweeted this earlier they came up with comparing models....

[100% pro-life no matter what]

Trump drops to 42% in all three models with the addition of the NBC poll (adjusted to Clinton +5).  The FOX state polls are not included yet.

[KingSweden]

I imagine those state polls will nudge Clinton a little bit down, but probably not a ton since they're in states where Trump is already leading

[Gass3268]

I imagine those state polls will nudge Clinton a little bit down, but probably not a ton since they're in states where Trump is already leading

Yeah, national polls, especially major ones have a biggest effect in the system. Also it takes a few for them to cause movement. If say the next couple major polls have Clinton +5, she should spike. Problem is we probably won't get any polls again until after the debate.

[rafta_rafta]

Fox News Poll is now added. Trump moves up to 42.4%

[KingSweden]

Fox News Poll is now added. Trump moves up to 42.4%

As predicted. Adjusted towards Clinton by 1 point each for house effect

[100% pro-life no matter what]

Alabama becomes the first state to go to >99.9% Trump in any model

[Xing]

Alabama becomes the first state to go to >99.9% Trump in any model

Should've been Nevada. Trump's Casinoresortmachine™ means that Latino turnout is irrelevant (who cares which language they speak, it's never been shown that native Spanish speakers are more Democratic!) He'd lose Alabama by 30 before losing Nevada.

[KingSweden]

Clinton up to 59% after Q-Pac polls adjusted 1 point in her direction

[Ozymandias]

I thought everything about Silver's most recent post...

fivethirtyeight.com/features/clintons-leading-in-exactly-the-states-she-needs-to-win/

...was spot-on, except for this puzzling line near the end: "She has one really good Electoral College path, but it’s only one path, instead of the robust electoral map that President Obama had in 2008 and 2012."

WTH? I mean, he JUST went into tremendous detail about how Clinton's closest state, New Hampshire, is safer than Trump's five closest states. And how states don't appear to be as well-correlated this election as they have been in the past.

So saying Clinton doesn't have a robust path is just silly...

[KingSweden]

I thought everything about Silver's most recent post...

fivethirtyeight.com/features/clintons-leading-in-exactly-the-states-she-needs-to-win/

...was spot-on, except for this puzzling line near the end: "She has one really good Electoral College path, but it’s only one path, instead of the robust electoral map that President Obama had in 2008 and 2012."

WTH? I mean, he JUST went into tremendous detail about how Clinton's closest state, New Hampshire, is safer than Trump's five closest states. And how states don't appear to be as well-correlated this election as they have been in the past.

So saying Clinton doesn't have a robust path is just silly...

Hes probably talking about the 272 freiwal.

All hail and bow down before the freiwal

[Ozymandias]

Hes probably talking about the 272 freiwal.

Right, but I don't think that explains the tenor of for example this tweet:

@NateSilver538 This is NOT a robust position for Clinton though. If polling is off in one of the firewall states, she doesn't have a good backup plan.

Because if, say, she loses NH (the current 538 tipping point state), then she actually has not just one but five backup plans: NV, FL, NC, OH, & IA, all of which are more likely (according to 538) to flip to Clinton than NH is to flip to Trump.