To expand on what Mason explained about weighting, my understanding is that non-college educated voters were undersampled in 2016 but not adequately compensated for in weighting. This skewed things in HRC's favor because non-college voters tend to favor Trump.

This wasn't too much of an issue nationwide, as the polling predicted HRC winning the popular vote by ~4 points and she won by 2. However, the polling error in crucial states like WI was sometimes as high as 6 due to this undersampling (and poor weighting). That has since been corrected and was not an issue in 2018 polling, and presumably won't be an issue this year.

It's funny you are trying to explain biasing errors and polls while exposing your own biases. Cherry picking poll results. People predicted the bubble bursting in 2008 on housing. Some of them had unsound reasoning while making that prediction. Donks get lucky.

I joked last week about the Trump poker economy and you. It does show you and others have cases of myopia when it comes to politics.

ps Trump is everything Friedman hated. You have no moral consistency. Why is this?

I don't know why you keep harping on this like it's a major issue. We know the exact question to ask. The ballot voters will see can be reproduced. We just don't know which 500 people to ask.

Mason starting this thread to educate us all on polling, while talking about the importance of weighting by party ID in the first post, is kind of like him starting a thread to educate us all on poker and explaining that we need to move up where they respect our raises.

First, it's okay to call me Mason,

The answers to your question are no and no.

Mason

Hi d2_e4:

I have no thoughts on this except to say that if Trafalgar is right again I suspect other polling organizations will start to look seriously at their methodology which had to do with how the questions are asked.

Best wishes,
Mason

If 170 million votes are cast who does that favor?

Hi ecriture:

First, you won't really know that they're off until the election results are in. Second, other polls are showing an improvement for Trump among Afican-American voters. Also, I wonder if this 44 percent number is just for a sub-group of the black voters such as young black voters.

Best wishes,
Mason

Hi Mason,

Based on current polls, I was curious of your opinion on what odds/percentage you'd give Trump winning reelection and winning popular vote?

Trump winning by 44 percent of the black vote? No way in hell that that's accurate. Maybe 44 percent of registered black republicans.
Trafalgar also recently had Mcsally leading Kelly in Arizona by a good margin, when Kelly was ahead in every other poll for months.
FWIW, I believe the polls in general, but I don't believe the poll that just came out that puts Biden up by 17 in Michigan. No way that's accurate either

You mean Wisconsin? To give you an idea of the absurdity, that poll isn't saying Dems will gain Wisconsin. It is saying that Dems will hold Wisconsin

Explanation:
Unusually, they didn't weight for 2016 vote but did ask about it. They found that 4% fewer people than reality said they voted for Trump, and 4% more than reality said they voted for Biden. Based on this poll, Clinton won Wisconsin 50-43 - the real result is she lost it 46-47.

Either:

1) This is sampling error - you should correct (not exactly but as a first order approximation) 4 percent up for Biden, 4 percent down for Trump. Adjusted result Biden +9

2) This is a "shy Trump" factor, people not admitting they are Trump supporters. Again the best approach is to adjust +4/-4 as in 1) Biden +9

3) This is former Trump supporters in denial that they were ever Trump supporters. Polling result stands Biden +17.

Obviously I tend to be in camp 1)/2) so I ignore polling that doesn't weight by previous vote. Reality may well include a bit of camp 3) though too so my approach may underestimate big swings.

I'm more confident in my approach in an election like this where the previous election was incredibly finely balanced so to predict the result this time around we just need to know the direction of the swing relative to last time - we don't need to know the magnitude of the swing. The UK 2019 election was similar (following on from the knife edge 2017 election) and the same approach worked pretty well in that.

I mean okay. I guess we don't know Kayne won't win popular vote either. And someone might find a counter example to Fermat's Last Theorem tomorrow and disprove it. But there is no reason to credulously believe such nonsense.

Erm, nope. You probably meant a different long-standing conjecture, FLT was proved by Andrew Wiles in the early 90s. Riemann Hypothesis maybe?

Nah, I meant FLT. I was going for Kayne wins the popular vote level unlikely.

You are correct that Wiles proved FLT in the 90s. But what he didn't prove and what's almost impossible to prove in a fundamental sense, is that Peano arithmetic or ZF set theory can't simultaneously prove FLT and also produce a counter example. That's a restatement of Godel's second incompleteness theorem. We can take it to the science thread if you're intrested in more detail.

Hi ecriture:

I think what you're addressing is what I would call "response bias," and as you say it's a tough nut to fix. However, one possible way to fix it, once it's identified, is through the questions themselves. This could include rewording and / or additional questions that also produce information related to the issue at hand.

When still at Census Headquarters I was involved with a survey that was trying to get at the needs of old people. The first question was a list of 50 illnesses that an older person might suffer from. We tried to tell the sponsors that this list would bias the answers to more needs, such as ramps, in a house.

Best wishes,
Mason

First, your post has nothing to do with this thread. Second, your three points are, in my opinion, completely wrong.

The reason I found the two polling organization that appeared on Hannity interesting is that they clearly stated that their indirect approach in asking questions did give different results, and then they claimed that these results were superior to standard polling.

Now I don't know if their results will be better, but I have stated in this thread that polling problems can be created by how the questions are worded, etc. and that seems to be what these two organizations were addressing.

Mason

Just to make sure that no one takes you seriously, how the surveys were weighted was not discussed.

Mason

The answer to your question is simply how accurate do you think the current polling is. The more accurate it is the more likely Biden will win.

However, I believe, based on my knowledge of the subject, that it's getting harder and harder to produce accurate polls. So to answer your question, I don't really know.

But there are other things out there that can be somewhat indicative of what might happen. Two would be how big of a turnout does the candidate get at a campaign stop and how much campaigning he does in the battle ground states. Both of these favor Trump.

On the other hand, how much money the candidate raises and how positive/negative the news, in general, is about the candidate can also be a factor. Notice that both of these favor Biden.

Best wishes,
Mason

Not exactly. "Response bias" is certainly a type of systemic error but other types of error go into the bucket of systemic error. I thought I was using the phrase in a totally standard manner,like every statistician would. But the last stats class I took was in high school so maybe not? Either way the concept I am describing exists. Like I can make a poll with 0 systemic error for the 2016 election. We know the exact vote totals so I could simply code 1xx million Trump voters and 1xx million Hillary voters and use a random number generator to select 500 of them. This poll would have 0 systemic error, but there will still be random error. Polls for 2020 still have random error, but also have systemic, since we obviously don't know the exact vote totals yet.

I doubt many serious readers would mistake sarcasm about reasonable information coming from a Hannity show, but I appreciate any effort at clean up work to ensure clarity. Thanks, Malmuth.

Yeah, that's why pollsters **** up.

I will take Mason's guess on 44% of young black male voters and give you the other two. How much?

As much as you want. I pick the escrow.

That was the point.

This post has everything to do with this thread. You cherrypicked those polls from a biased source you watch.

1) Show your work where these polls are doing something other polls don't instead of taking them at their word.
2) I am well aware of biasing errors because of how questions are worded on polls.

And, no, I'm not wrong about anything in my post.

You gloated in the past that Trump would be good for the poker economy. Something that never came to fruition. The Trump administration's response to COVID on an economic level is something that didn't jibe well with Friedman and his apostles on an intellectual level when similar actions were taken during their time. Except maybe the tax portion of the deal. It's why you resemble Lindsey Graham or Susan Collins. You clinged on to Friedman and now embrace someone who rejects his principles thoroughly.

You wanted fun and vigorous debate. Post honestly instead of hiding your power level.