2020 Elections

[mahagonny]

You can't by cigarettes without reading a stark warning about the danger of smoking. It would be interesting if every prediction had to include a sentence or two about the track record of the polling organization. Assuming you could argue that being mislead by pollsters does damage. I suppose it does.

[Ruralguy]

To Clean's report, I even recall Silver on NPR several days before the 2016 election stating these predictions and saying something like " of course some battlegrounds are close enough that a late break could lead to Trump win."  For some weird reason, he still apologized for "getting it wrong." I don't think he did. Long shots win. He even said what the conditions were for such a long shot win happening.

[little bongo]

I don't like Trump, wouldn't have voted for him in 2016 and wouldn't now if I were an American, but the unwillingness of anti-Trump people to admit that there are some legitimate concerns voters have which may make them vote for him is frustrating. Fearing your home or business could be torched or vandalized is a legitimate concern for people living in those areas, and doing nothing in order to appear "woke" is not leadership.
[/quote]

As I said--he can fool some of the people all of the time. All. Of. The. Time.

It has nothing to do with being "woke." It has nothing to do with being "unwilling" to recognize legitimate concerns. It has everything to do with the campaign message of "Look! Left-wing chaos, in which Biden is a mere puppet! I'm the one who can solve this!" This leads to three critical questions:
1) Who's the freaking president?
2) Who's supposed to be solving this right now?
3) Who's absolutely not doing any solving right now?

You want to talk about being frustrated?

[lightning]

Trump galvanized the far right who were otherwise unenthusiastic about the traditional Republican platform (fiscal conservatives like Rubio and truly religious candidates like Cruz were trounced by Trump in the 2016 Republican primaries). As a reward, he brought the far right from the fringes into the mainstream, and for this their support for him is unabated. This was a voting block whose needs were not overtly acknowledged or met, even though all it took was a validation of their racist beliefs. Not even old-school Republicans wanted to go there, at least overtly (Don't get me wrong--the racist vote has always been courted, just not in a direct way). The far-right voting block only wants one thing and that's the representation and validation of who they are, and Trump gave it to them when no other modern politician wanted to. His rhetoric is part of their representation and reward. This ability to galvanize is what Biden is lacking and what Democrats lost when they abandoned Bernie, etc. in their strategy to once again fight for the votes in the middle.

So let's talk about that middle--the middle whose votes are still available to be won and have not completely made up their mind yet. The middle is the high-school educated white male in the rust belt who also has influence over their family members. This is the guy who didn't simply lose his livelihood to outsourcing and automation. He also lost his dignity. These were the votes that flipped for Trump in 2016 in states like WI and PA. These were guys who were left behind in the global tech economy and have no hope for ever catching up (too old/too far behind in education levels) to the more educated so-called "elites" who they have come to resent. Democrats have done nothing to win their vote back. Trump doesn't do anything concrete for them either, but at least he tells them it's not their fault--it's China's fault. (Hillary offered up the old strengthening of re-education programs--HEWM want good jobs and not re-education) And, Trump offers them a version of dignity, a white identity that is automatically conferred with dignity as long as the suppression of everyone else is perpetuated, and all they have to do is join him in that cause. The HEWM don't have to be Klan members. They just have to be against looting, against kneeling for the anthem, be against "socialism," and fly/wear/bear symbols that have been appropriated by the right like the flag. Easy peasy. Dignity through identity. Function is no longer important. People will give up their health care, social safety net programs, and real economic opportunity if they can live and die in dignity. The "good middle class job" is now secondary to them. That's how bad it has gotten for them, because they have given up on any politician who empathizes with their plight. The HEWM will take Trump's option. Yeah, Republicans do not do jack squat for the HEWM either, but at least Trump offers dignity. Democrats offer the HEWM re-education at best (which the HEWM does not want--they want their 1950s style job back even though they know it's not possible), and at worst Democrats make them feel like crap about themselves, just for being white. It's no wonder they flipped.

The fight is in the rust belt. Biden has got to say something--anything, that makes the HEWM feel some sense of dignity, promising an economy that brings back those old-fashioned jobs without making them feel like they are biting from the pie of "socialism."

It seems the Democrats are going for the educated professional voter who lives in the 'burbs. That may be enough to win the popular vote, but it won't be enough to win the electoral college vote.

If I sound like Michael Moore, well, he predicted the Trump victory in 2016 when every other Democrat was predicting a Clinton win.

[marshwiggle]

So let's talk about that middle--the middle whose votes are still available to be won and have not completely made up their mind yet. The middle is the high-school educated white male in the rust belt who also has influence over their family members. This is the guy who didn't simply lose his livelihood to outsourcing and automation. He also lost his dignity. These were the votes that flipped for Trump in 2016 in states like WI and PA. These were guys who were left behind in the global tech economy and have no hope for ever catching up (too old/too far behind in education levels) to the more educated so-called "elites" who they have come to resent. Democrats have done nothing to win their vote back. Trump doesn't do anything concrete for them either, but at least he tells them it's not their fault--it's China's fault. (Hillary offered up the old strengthening of re-education programs--HEWM want good jobs and not re-education)

Good analysis.

And, Trump offers them a version of dignity, a white identity that is automatically conferred with dignity as long as the suppression of everyone else is perpetuated, and all they have to do is join him in that cause. The HEWM don't have to be Klan members. They just have to be against looting, against kneeling for the anthem, be against "socialism," and fly/wear/bear symbols that have been appropriated by the right like the flag.

I'm not sure why the tone here is that being against looting somehow represents some sort of white supremacy. Honestly, is a "good Democrat" in favour of looting???

People will give up their health care, social safety net programs, and real economic opportunity if they can live and die in dignity. The "good middle class job" is now secondary to them. That's how bad it has gotten for them, because they have given up on any politician who empathizes with their plight. The HEWM will take Trump's option. Yeah, Republicans do not do jack squat for the HEWM either, but at least Trump offers dignity. Democrats offer the HEWM re-education at best (which the HEWM does not want--they want their 1950s style job back even though they know it's not possible), and at worst Democrats make them feel like crap about themselves, just for being white. It's no wonder they flipped.

More good analysis.

The fight is in the rust belt. Biden has got to say something--anything, that makes the HEWM feel some sense of dignity, promising an economy that brings back those old-fashioned jobs without making them feel like they are biting from the pie of "socialism."

It seems the Democrats are going for the educated professional voter who lives in the 'burbs.

Who are mostly going to vote for them anyway.

That may be enough to win the popular vote, but it won't be enough to win the electoral college vote.

If I sound like Michael Moore, well, he predicted the Trump victory in 2016 when every other Democrat was predicting a Clinton win.

This is the message that needs to be heard.

marshwiggle: it is clear that you either know nothing about statistical prediction(or just want to incite). Prediction is inherently inaccurate.

Nate Silver had Trump at a 27% chance in 2016. If true again, there is nearly a 20% chance of being wrong with the prediction in consecutive elections picking against Trump.

Usually, results are given witha 95% confidence interval. Which means, 19 times out of 20, they should work out. Which, given that elections are every 4 years, a wrong prediction should be followed by correct ones for about 80 years. Two wrong in a row indicates seriously flawed methodology, or ridiculously wide confidence intervals which means data samples were outrageously small.

[mahagonny]

And, Trump offers them a version of dignity, a white identity that is automatically conferred with dignity as long as the suppression of everyone else is perpetuated, and all they have to do is join him in that cause. The HEWM don't have to be Klan members. They just have to be against looting, against kneeling for the anthem, be against "socialism," and fly/wear/bear symbols that have been appropriated by the right like the flag.

Presumably the suppressed would include people who chose to live here by leaving their home country.

I'm not flying or wearing symbols of any kind. And I generally only vote negatively, that is, for the candidate I distrust the least. But I could conceivably vote for Trump if things get stupid enough on the left. For example if someone like Ibram X. Kendi (who has a National Book Award) gets the ear of the new Biden cabinet and gets a federal anti-racism department. I'm looking for just one democrat who has the sense and guts to say 'no Antiracism Department. Just...get outta here. Write your books if you must, but leave us alone. Your'e a fringe character.'

[jimbogumbo]

This is the message that needs to be heard.

marshwiggle: it is clear that you either know nothing about statistical prediction(or just want to incite). Prediction is inherently inaccurate.

Nate Silver had Trump at a 27% chance in 2016. If true again, there is nearly a 20% chance of being wrong with the prediction in consecutive elections picking against Trump.

Usually, results are given witha 95% confidence interval. Which means, 19 times out of 20, they should work out. Which, given that elections are every 4 years, a wrong prediction should be followed by correct ones for about 80 years. Two wrong in a row indicates seriously flawed methodology, or ridiculously wide confidence intervals which means data samples were outrageously small.

Wrong, and wrong. Every statistician just died a little bit. You do NOT know what a CI means. Seriously.

And no, I will not explain it to you. Look it up.

AFTR, using italics doesn't make you correct. It just makes you incorrect with italics.

[marshwiggle]

marshwiggle: it is clear that you either know nothing about statistical prediction(or just want to incite). Prediction is inherently inaccurate.

Nate Silver had Trump at a 27% chance in 2016. If true again, there is nearly a 20% chance of being wrong with the prediction in consecutive elections picking against Trump.

Usually, results are given witha 95% confidence interval. Which means, 19 times out of 20, they should work out. Which, given that elections are every 4 years, a wrong prediction should be followed by correct ones for about 80 years. Two wrong in a row indicates seriously flawed methodology, or ridiculously wide confidence intervals which means data samples were outrageously small.

Wrong, and wrong. Every statistician just died a little bit. You do NOT know what a CI means. Seriously.

And no, I will not explain it to you. Look it up.

AFTR, using italics doesn't make you correct. It just makes you incorrect with italics.

OK, so let's revisit this to make a correction.

Nate Silver had Trump at a 27% chance in 2016. If true again, there is nearly a 20% chance of being wrong with the prediction in consecutive elections picking against Trump.

If in 2016, there was a 27% chance of Trump winning, and there's a similar 27% chance in 2020, and he wins both times, the odds of that happening by chance alone is 27%*27% which equals about 7%. So it's vastly more likely that the statistics were wrong than that being wrong both times was just random.

(So indeed, they weren't 90% confidence intervals, but the odds were high enough in one direction that there should have been far more questions as to how they got it wrong. Especially since there were multiple polls.)

[Descartes]

I mean, I'm a white male in my late 30's who lives in a rustbelt/Midwest/bellweather state who has always voted Democrat from age 19 + but who voted for Trump in 2016 and is going to do so again.

In Reddit parlance, "AMA." 

[jimbogumbo]

My apologies for the thumbfingeredness. Sincerely. Yes, ,27*.27 is about .07. I have proven again (that when I go fast) I make arithmetic errors.

On the other hand, look at it this way: ,73*.73 is about .44. As a pollster you would pick the .73 every time, but you'd be wrong at least once 56% of the time. So I see pollsters being wrong a lot as a given. You and I have different expectations of them.

I'm more depressed by the follow up statement to my original reaction.

A CI is not a probability; it's is a statement of confidence in the accuracy of the estimate. The 80 years comment also resonated badly with me as it is true it is 80 years but comprised of 20 trials. Your statement that if they are wrong once they'd have to be correct the next 20 times (I edited) is just not what a CI "guarantees", if indeed CI's guarantee anything.

[marshwiggle]

My apologies for the thumbfingeredness. Sincerely. Yes, ,27*.27 is about .07. I have proven again (that when I go fast) I make arithmetic errors.

On the other hand, look at it this way: ,73*.73 is about .44. As a pollster you would pick the .73 every time, but you'd be wrong at least once 56% of the time. So I see pollsters being wrong a lot as a given. You and I have different expectations of them.

That's why getting one election wrong, even with those odds, may be excused. But to get two wrong in a row is highly unlikely unless you're doing something wrong. For an election, 73% is pretty darn high. Most are probably in the 50's or occasionally 60's. It was kind of morbidly entertaining watching the election night coverage and seeing reporters increasingly focusing on polls not yet in as the night went on waiting for the results to change to how they were "supposed" to be.

I'm more depressed by the follow up statement to my original reaction.

A CI is not a probability; it's is a statement of confidence in the accuracy of the estimate. The 80 years comment also resonated badly with me as it is true it is 80 years but comprised of 20 trials. Your statement that if they are wrong once they'd have to be correct the next 20 times (I edited) is just not what a CI "guarantees", if indeed CI's guarantee anything.

I agree that it doesn't guarantee the outcome of even the next one, but as above, it makes it highly suspect if a similar "mistake" occurs soon. It's far more probable that the poll is somehow invalid than that it was just a random sampling error.

Relating to 2016, one theory was that people didn't answer the polls truthfully. If that was indeed the case, then it means the methodology is seriously flawed if voters don't feel comfortable giving their true preferences to a poll. For instance, if it's a phone poll, and the people asking the questions somehow make people uncomfortable stating their actual preferences, then their training has been woefully inadequate.

[jimbogumbo]

Nate Silver isn't really a pollster. He is more accurately described as a meta-pollster. He averages results of many polls to create his estimate.

You are correct about the pollsters' methods likely not being accurate for the last election. That is why I said originally if the estimate was correct (but still multiplied with a thumb fingered decimal wrong).

[mythbuster]

Trump will likely lose the popular vote. I'll even go out and say he will likely lose the popular vote by more than he did last time. But even with that I think he has a very good shot of winning re-election.
   It doesn't matter if Dems get out the vote by whatever % in places like California, New York, or Massachusetts. It does matter in Florida, Wisconsin, and maybe even Arizona. What happened that Ohio is no longer the darling of the swing states?

This is why the polling doesn't really matter. It's as much about where you vote as who you vote for.

[marshwiggle]

Nate Silver isn't really a pollster. He is more accurately described as a meta-pollster. He averages results of many polls to create his estimate.

You are correct about the pollsters' methods likely not being accurate for the last election. That is why I said originally if the estimate was correct (but still multiplied with a thumb fingered decimal wrong).

This is why it's so important to address; without strong evidence to the contrary, it suggests that the method of polling employed can no longer be relied upon, even if it worked in the past. It would be extremely foolish to view that specific election as a complete one-off; it's more likely there have been cultural shifts which affect the data gathering process. (This is probably an ongoing reality, but which was highlighted by the 2016 results. For instance, if it had been expected to be a closer race, the outcome being different than the prediction wouldn't have been such a shock. It might take 3 or 4 close races in a row with "wrong" outcomes to make it apparent that there's a problem.)

[spork]

A large chunk of Trump supporters are people who are the equivalent of the kindergartener who breaks the toy when it becomes some other child's turn to play with it -- the "if I can't have what I believe I deserve, regardless of whatever terrible personal choices I've made, then burn down the whole system" mentality. They blame people who are similarly poor and structurally disadvantaged but non-white for their own circumstances and are happy to destroy the very institutions they themselves are dependent on. It's one version of crabs in a bucket syndrome: preference for blowing up the whole bucket with themselves in it to letting a few crabs escape. Trump, being a sociopath, has an inherent understanding of how to use this to his advantage.