Big D vs Twitter

[Richey]

It is obviously fair enough for people not to trust polls, as you remember when they are wrong, but you are less likely to remember when they are right.

If you take the UK for example, the polls were wrong in 2015 by overestimating Labour/underestimating the Tories, depending on how you look at it. The 2017 election went the other way a bit, and underestimated Labour. The 2019 election in the UK was pretty much polled correctly. There may have been coverage of a load of people chanting about Jeremy Corbyn, but he was always behind.

Polling companies change their weightings and their methodologies between each election in order to fix these errors. They do so because there is no money in political polling, it is just a high profile way to advertise how accurate they can be.

The US pollsters could be wrong again, but it is not likely that they would be wrong in exactly the same way (by underestimating the Republicans/overestimating the Democrats).

Specifically for the US, my understanding is that the pollsters went wrong with the undecided voters, and estimated that many would not vote and that those who would vote would split in a certain way between Trump and Clinton. As it happens, more than they estimated DID vote, and more than they estimated voted for Trump.

This time, the number of undecided voters is lower, and of those undecided voters, it is estimated that most are not so likely to vote for Trump or vote at all.

The estimations could still be wrong of course. Trump's voter base could have completely changed, or people might switch at the last minute, or people simply might not bother to vote at all. Biden is hardly an inspiring candidate in anyone's eyes really, so I would guess that the last possibility is one that Trump is hoping for.

On balance though, it isn't looking great for Trump.

 

[dantes]

So Trump isn't corrupt?

Making corrupt decisions to help out his own medium size business will help out his voters too. Policy wise, despite him being a billionaire, he's in the same boat as the people. Politicians on the other hand are in the multinational corporation boat, which is firing torpedo's at the people.

 

[Farkmaster]

By paying less tax? By not losing their health insurance? By having their jobs protected from unfair trade deals? By having their pension values recovered by a rising stock prices?

We've done this before, but the stock market doesn't actually matter a bit to the majority of Americans. About a third of americans have more than 10k in stock holdings. 50% have none.


The tax benefits were almost nothing to these same Americans, but a massive boon to the rich. That's who they were written for, obviously. We can see this because corporate tax cuts were permanent. Individual tax cuts were not. This was slight of hand so that the package would seem fiscally neutral, assuming unrealistic growth, that as we've seen, cannot be banked on. They knew that someone else would be in power then, and they'd be able to switch back into being "fiscally conservative" when it came to social programs that became unaffordable, even though they are fiscally liberal and willing to blow a budget when it comes to redistributing wealth upwards. Looking at the tax cut over the same period of analysis, it increases taxes on the middle and lower class, while providing massive benefit to the wealthy. Did this wealth translate into growth? No, it translated into cash reserves that are unprecedented.

Then, health insurance. You know, the thing that Trump tried his damndest to get rid of, for millions of americans on obamacare. He would have, too, if it were not for John McCain, who he loathes for it. He's been repeatedly asked for a plan for his healthcare, he doesn't have one, well, other than that it'll be the greatest, perfect plan, and it's coming... for the last 4 years.

Every single one of his policies exists to benefit corporations and people with large amounts of capital. Just be ok with that, if you are. Don't do his hacky salespitch for him.

 

[Farkmaster]

Making corrupt decisions to help out his own medium size business will help out his voters too. Policy wise, despite him being a billionaire, he's in the same boat as the people. Politicians on the other hand are in the multinational corporation boat, which is firing torpedo's at the people.

He puts his 400 million dollar debt to a foreign bank one leg at a time just like me!

 

[dantes]

He puts his 400 million dollar debt to a foreign bank one leg at a time just like me!

Peanuts. It's not like he took out the loan to buy a car, it's probably being used to leverage up his profits far in excess of the interest payments.

 

[Farkmaster]

Peanuts. It's not like he took out the loan to buy a car, it's probably being used to leverage up his profits far in excess of the interest payments.

Yeah, it's not like he's ever miscalculated and gone bankrupt repeatedly.

 

[Frogfish]

By paying less tax? By not losing their health insurance? By having their jobs protected from unfair trade deals? By having their pension values recovered by a rising stock prices?

Clarification :

Under Democrats : personal tax changes means ONLY those in the top 5 percentile will pay more (even then effectively those earning over $150,000).

Jobs : compare the % growth of the last three years of Obama's with Trumps first three years = they are almost identical (actually Obama's 1/2% higher in each year but who's counting eh.

Health Insurance : Trump doesn't even have a plan yet ! 4 years after he said he was going to get rid of Obamacare. All we know is that his abolition, instead of revision, would mean countless people would lose ongoing cover.

Stock Prices :Fallacy to believe governments control them.

 

[Athens]

He puts his 400 million dollar debt to a foreign bank one leg at a time just like me!

Word on the street says it's closer to 1 billion.

 

[peekay]

It is obviously fair enough for people not to trust polls, as you remember when they are wrong, but you are less likely to remember when they are right.

If you take the UK for example, the polls were wrong in 2015 by overestimating Labour/underestimating the Tories, depending on how you look at it. The 2017 election went the other way a bit, and underestimated Labour. The 2019 election in the UK was pretty much polled correctly. There may have been coverage of a load of people chanting about Jeremy Corbyn, but he was always behind.

Polling companies change their weightings and their methodologies between each election in order to fix these errors. They do so because there is no money in political polling, it is just a high profile way to advertise how accurate they can be.

The US pollsters could be wrong again, but it is not likely that they would be wrong in exactly the same way (by underestimating the Republicans/overestimating the Democrats).

Specifically for the US, my understanding is that the pollsters went wrong with the undecided voters, and estimated that many would not vote and that those who would vote would split in a certain way between Trump and Clinton. As it happens, more than they estimated DID vote, and more than they estimated voted for Trump.

This time, the number of undecided voters is lower, and of those undecided voters, it is estimated that most are not so likely to vote for Trump or vote at all.

The estimations could still be wrong of course. Trump's voter base could have completely changed, or people might switch at the last minute, or people simply might not bother to vote at all. Biden is hardly an inspiring candidate in anyone's eyes really, so I would guess that the last possibility is one that Trump is hoping for.

On balance though, it isn't looking great for Trump.

Agree with you in general. But one key aspect is that the electoral college amplifies small polling errors at regional levels to significantly different results at the national scale. If the polling and predictions are off by a small percentage which leads to incorrect predictions in MI, PA, OH - that may be the difference between 304 - 227 vs 227-304.

Second is the latent Trump voter - people who don't identify themselves as Trump voters publicly or are normally not a part of this targetted polling demographic. Pollsters say they have accounted for them but I cannot wrap my head around how they did it. There are limits to predicting human behaviors through surveys and I feel that this is one area where they have gone beyond the limits. Apparently, their surveys can get people to reveal what they are actively trying to hide? It is like trying to predict how much money businessmen have stashed in Cayman Island accounts to avoid paying taxes by surveying 100 businessmen.

 

[StevieM]

Trump isn’t a politician.

He’s also not a very good businessman - because they tend not to make money and fail.

He’s a great con man and reality tv host though.

Even then... more people tuned in to watch Biden than him the other night, apparently, so if that’s the case Trump’s truly fucked.

If Trump does lose the election I sincerely hope Biden has the chops to hold a press conference where he sits behind a desk and says “Donald Trump..... you’re fired... collect your bags and go!!!”.

I bet Dantes loves Survivor and The Bachelor.

 

[dantes]

You never slowed down to check out a car crash? Steve Scully got into a massive pile up recently taking out most of the fake news with him.

 

[redsar]

 

[The Nomad]

That shouldn't destroy his campaign.. his hair sniffing didn't destroy it. God knows how.

 

[Athens]

We've all seen that nothing destroyed trump's campaign last time around. He even mocked a disabled guy and that didn't destroy his campaign. People will vote for either of the candidates no matter what.

 

[Rosco]

Yup.

Biden joined the woke cult this week by advocating for transitioning the gender of 8 year olds.

He's still going to win

 

[Richey]

Agree with you in general. But one key aspect is that the electoral college amplifies small polling errors at regional levels to significantly different results at the national scale. If the polling and predictions are off by a small percentage which leads to incorrect predictions in MI, PA, OH - that may be the difference between 304 - 227 vs 227-304.

Second is the latent Trump voter - people who don't identify themselves as Trump voters publicly or are normally not a part of this targetted polling demographic. Pollsters say they have accounted for them but I cannot wrap my head around how they did it. There are limits to predicting human behaviors through surveys and I feel that this is one area where they have gone beyond the limits. Apparently, their surveys can get people to reveal what they are actively trying to hide? It is like trying to predict how much money businessmen have stashed in Cayman Island accounts to avoid paying taxes by surveying 100 businessmen.

first part, yes absolutely. It was at state level where things went wrong last time in terms of polls.

Second part, there is a lot of talk of shy Trumpers, just as there was in 2016. By the way, that term would mean something rather different in the UK, but also very similar, depending on your viewpoint.

Anyway, there is not a great deal of evidence for it. The idea of people who don’t wish to admit to voting for one side as they may consider it embarrassing usually means that phone polls are very different to online polls. People would tend to less likely to admit to an actual human voice that they are voting one way but are happier to admit it to an online form. As far as I know, this time the online polls are similar to the phone polls.

This also would tend to be something that happens with undecided voters. It’s one thing to not wish to declare that you are voting for Trump, but quite another to go as far as to declare support for Biden. Shy Trumpers (should they exist) would be more likely to say that they haven’t decided. As I said before, there are less undecided voters this time, and as you say, they have changed how they estimate how they believe the undecided may vote. One way is to ask a different question (various ones) and another is to estimate by demographic.

This is where most pollsters have made changes since last time, by weighting the demographics differently. The problem last time was very much that Trump had different voters to many previous Republicans. Now they can change their weighting to account for that. Actually, the result of this is more likely to overestimate Trump, which is what often happens when pollsters try to correct an error.

Obviously there are lots of potential flaws. In theory Trump could I suppose have a completely different demographic of voter compared to last time, which would skew the whole thing again. Turnout is also a huge issue this time, and one that Trump will be clinging to. People may say that they will vote, but who is to say that with Coronavirus they may actually do so? Especially if one adds the factors that they may already think Trump is doomed, so why bother, they may not feel that inspired by Biden, and who can really blame them, and they even not wish to contribute to the fairly likely chaos and unrest that may follow if Trump loses and refuses to accept it.

 

[Athens]

There's also the fact that there are lots of Republicans who can't stand Trump and who are actively trying to get rid of him, whereas the same can't be said in the democrat camp.

I still wouldn't put any money on it though.

 

[tomasjj]

Yup.

Biden joined the woke cult this week by advocating for transitioning the gender of 8 year olds.

He's still going to win

Did you sign up?

 

[redsar]

That shouldn't destroy his campaign.. his hair sniffing didn't destroy it. God knows how.

It won't, of course. But, it's ridiculous how nearly all mainstream media are biased against Trump while this guy goes under the radar. It is obvious that Biden has a long history of bullshitting, lying, even shady dealings, and yet the CNN and co. ignore it completely while they dissect everything Trump says or does...

 

[Rosco]

Did you sign up?

Nah.

The only benefit to transitioning is getting some tits. I've already got some

 

[Woland]

Can't wait for Biden to put half a dozen dems in the supreme court and overrule everything Trump did just to see all the hillbillies cry

 

[Woland]

Last night he said if he loses he might leave America. Things can actually get worse. He's going to move here isn't he? Imagine the big fat burger sauce snorting nonce moved in next door to you and started stinking the place out backing the drains up with his nappies.

 

[Athens]

I think if he loses Putin will have him sent to the gulag.

 

[Farkmaster]

Can't wait for Biden to put half a dozen dems in the supreme court and overrule everything Trump did just to see all the hillbillies cry

Yeah, the constant undoing of each other's work isn't great governance though. Plus it's not possible. The people who quit Trump's mess of an administration, as well as those scientists whose research he stopped, aren't just going to pick up where they left off. It's much much easier to break things than fix them. The politics of revenge are what got us here.

Also, I don't think the democrats have the stones to pack the court. And, if they do, it's going to devolve into still more of a farce. It's incredible to think that it functioned as an apolitical force on the back of just a sense of fair play, for decades.

 

[dantes]

It won't, of course. But, it's ridiculous how nearly all mainstream media are biased against Trump while this guy goes under the radar. It is obvious that Biden has a long history of bullshitting, lying, even shady dealings, and yet the CNN and co. ignore it completely while they dissect everything Trump says or does...

Twitter and Facebook are also deleting links to their investigation.

 

[peekay]

first part, yes absolutely. It was at state level where things went wrong last time in terms of polls.

Second part, there is a lot of talk of shy Trumpers, just as there was in 2016. By the way, that term would mean something rather different in the UK, but also very similar, depending on your viewpoint.

Anyway, there is not a great deal of evidence for it. The idea of people who don’t wish to admit to voting for one side as they may consider it embarrassing usually means that phone polls are very different to online polls. People would tend to less likely to admit to an actual human voice that they are voting one way but are happier to admit it to an online form. As far as I know, this time the online polls are similar to the phone polls.

This also would tend to be something that happens with undecided voters. It’s one thing to not wish to declare that you are voting for Trump, but quite another to go as far as to declare support for Biden. Shy Trumpers (should they exist) would be more likely to say that they haven’t decided. As I said before, there are less undecided voters this time, and as you say, they have changed how they estimate how they believe the undecided may vote. One way is to ask a different question (various ones) and another is to estimate by demographic.

This is where most pollsters have made changes since last time, by weighting the demographics differently. The problem last time was very much that Trump had different voters to many previous Republicans. Now they can change their weighting to account for that. Actually, the result of this is more likely to overestimate Trump, which is what often happens when pollsters try to correct an error.

Obviously there are lots of potential flaws. In theory Trump could I suppose have a completely different demographic of voter compared to last time, which would skew the whole thing again. Turnout is also a huge issue this time, and one that Trump will be clinging to. People may say that they will vote, but who is to say that with Coronavirus they may actually do so? Especially if one adds the factors that they may already think Trump is doomed, so why bother, they may not feel that inspired by Biden, and who can really blame them, and they even not wish to contribute to the fairly likely chaos and unrest that may follow if Trump loses and refuses to accept it.

As a person with some (not a lot) of experience with online surveys and statistical analysis, I have to disagree with that. People have become suspicious of online polls also because they are afraid the information is being stored somewhere. A shy Trump voter might not necessarily be more likely to reveal themselves in online polls also. There could be proxy questions. For example for income, we normally use questions like car ownership, the industry of employment, zipcode, etc to glean the true income. Maybe there could be similar questions for identifying shy Trump voters, but that just throws an additional layer of error into an already complicated process.

I do have an inherent skepticism of some of the statistical claims made by these consulting companies. I am not a statistical expert but have a working knowledge of most analysis and I am amazed at the confidence in which they make claims. Maybe it is just my personality.

Agree with you on the new demographic and the turnout is the key. I did read an analysis some time ago that one of the reasons for democrats making inroads in mid-terms is new voters.

 

[MomoWASright]

It’s like those people who say they care for the poor and disabled and then vote Tory.

 

[dantes]

Trump isn't a politician, so that's the end of the debate. The media labels him as thick, a nonce, a racist, a liar, a narcissist, none of it matters, they will never be able to label him as a politician. Victory is a formality from that point.

I arrive at my opinions independently, but when other people come to the same position at the same time, it's like there is some form of subliminal brainwashing happening.

 

[dantes]

Hahahaha lucky for Creepy Joe that his boss isn't Jimmy Conway.

 

[MomoWASright]

Hahahaha lucky for Creepy Joe that his boss isn't Jimmy Conway.

What he’s not letting in is that house is actually in Bootle so it’s only worth £110k.