**** Elections thread *****

[wufwugy]

I would note that the incorrect beliefs about Clinton reflect the Democratic establishment wisdom. The Republican establishment had the same issues. They believed Jeb was the best candidate and believed that Sanders was the dream candidate to run against. No and no. Jeb inspired nobody and at best could have squeaked by a bad opponent, while Sanders was actually formidable. Sanders was the only hope for the Democrats to keep the White House.

80% of political consultants need to get their asses canned. I would say they have no idea what they're doing, but that may not be true. They're all lawyers. It's like the MPAA/RIAA stuff. The organizations gain nothing, while the lawyers gain everything by convincing the organizations their services are valuable. No body of professionals jump from failure to failure like political consultants, all the while making bank.

[Poopadoop]

80% of political consultants need to get their asses canned. I would say they have no idea what they're doing, but that may not be true. They're all lawyers. It's like the MPAA/RIAA stuff. The organizations gain nothing, while the lawyers gain everything by convincing the organizations their services are valuable. No body of professionals jump from failure to failure like political consultants, all the while making bank.

It seems like a job where if your party wins you're a genius with great instincts and if it loses you're the opposite. Their whole process strikes me as based on snake oil and tarot cards and if there's any justification for what they do it's that half the time they're right.

[wufwugy]

That sums it up.

[wufwugy]

There is a good deal of correlation with the Democrat zeitgeist and the Remain camp in Brexit. Just like the Remain campaign, Democrats believe that history is on their side, that they're inevitable, that they don't have to do much, that they can just scare people with fakery into agreeing with them, and that the parameters that have failed them previously (yet worked a long time ago!) will work for them this time.

Political power makes people lazy.

[Savy]

Brb going to look up why trump was a given to lose the nomination.

[Poopadoop]

Geez, they woke her up from a nap to answer questions about some bombing? Fucking savages.

[OngBonga]

To be fair, that's what I felt like on Saturday morning when I was making crumpets and my mate was asking me when I'd have the rent.

[wufwugy]

I'd like to see why people say she's drunk there.

[Poopadoop]

Because it sound worse than 'drowsy'?

She doesn't look drunk to me, she's not slurring her words or anything. She looks more like she just woke up and/or her meds haven't kicked in yet.

[OngBonga]

I'd like to see why people say she's drunk there.

Isn't she on a plane? If she's just woke up from a nap, and she's jetlagged, this is reasonable, especially considering she's not been well.

I'll give her the benefit of the doubt here.

[Poopadoop]

Interesting contrast between her groggy response and Trump's 'ZOMFG the shit's hit the fan again, better get tough folks!' reaction (before he even really knew what had happened he assumed it was a bombing). Neither one of them really strikes me as someone I'd want in charge.

[Poopadoop]

Isn't she on a plane? If she's just woke up from a nap, and she's jetlagged, this is reasonable, especially considering she's not been well.

Ya, but you'd think that's the kind of news that would perk her up a bit, jetlagged, sick or whatever.

[wufwugy]

I heard the other day accusations about her being an alcoholic but didn't find much. She could be high functioning. I'm mostly joking when I say this: the best piece of evidence that she is high functioning alcoholic is that she spoke to the press in the first place. Need that liquid courage.

[wufwugy]

This was Trump's response:

https://twitter.com/realDonaldTrump/...94113834917888

I would like to express my warmest regards, best wishes and condolences to all of the families and victims of the horrible bombing in NYC.

Are you talking about something else?

[Poopadoop]

Watch his initial reaction as he's about to leave his rally. That's what I'm talking about.

[wufwugy]

I get the sentiment. I don't like that it's good campaigning, but it is good campaigning. Fortunately, the signs are that he uses rationality much more than his public persona appears to. The birther thing I posted is a good example of it. Many things people see from Trump they don't like are the "pacing" portion of the "pacing and leading" tactic that Trump uses all the time to great effect. He has gotten many people to agree with him by pacing them, and then he has flipped it and gotten them to change their beliefs by leading them.

Trump winning the nomination is one of the best things to happen to me because it was the trigger that got me to study persuasion. This stuff is truly amazing.

[Poopadoop]

I guess the question is how much of his rhetoric is just trying to appeal to people and win the election and how much he really is the type of guy who would act rashly. Hopefully more weighted towards the former.

[wufwugy]

I definitely think it's the former. He seems among the least rash of those who ran for President this cycle. Example, just the other day he said he would reveal his health records on Dr. Oz, then his manager said he wouldn't, then the media blew up calling him a liar, then he went on Dr. Oz and surprise revealed his health records. This was orchestrated. This particular instance lasted a short enough period of time that we can more easily evaluate it, but most of his orchestrations have been much longer form, and they're where the accusations of rash came from.

One of the better examples of long form orchestration is something Cialdini discussed in Influence: the contrast principle. Trump started out trying to look crazy so that the media would call him crazy. This put Trump in a position of control, where he could turn presidential, making the media look stupid and showing low-information voters (the vast majority of them) who tune in late in the cycle that he's totally normal. This was something Scott Adams was saying he would do many months ago, and his prediction was correct. Trump has turned the tables merely by holding his tongue, and it works so effectively because people expect him to be a lunatic. It makes him look even more presidential than Clinton and more than he could have otherwise.

He is the best persuader alive.

[wufwugy]

An example commonly used to denigrate Trump that I think is intentional persuasion on multiple tiers: his goofy physical appearance and adjusting that for contrast. Why is his hair stupid and face orange? Given what we know about Trump it's probably a way to get attention. Goofy, unique, outlandish appearances garner attention. Look at Don King. Look at Dennis Rodman. These people would not be as popular if they looked normal. This can explain why Trump created his own unique and eye-catching look. But during the election he's gone further than just getting attention; instead he's using that as a tool to subvert expectations. He's gradually lightening his tan and normalizing his hair. At the debates, many people will tune in to see a clown yet end up seeing a relatively normal person. By the contrast principle, this will influence them to want to vote for Trump more than most other things.

[Poopadoop]

Lol @ this guy. Objective reporting at it's finest.

[JKDS]

When I hear "major meltdown", I expect to see something akin to "FUCK IT, WE'LL DO IT LIVE. FUCK IT. WE'LL JUST DO IT LIVE". But maybe I'm asking for too much.

[Poopadoop]

When I hear "major meltdown", I expect to see something akin to "FUCK IT, WE'LL DO IT LIVE. FUCK IT. WE'LL JUST DO IT LIVE". But maybe I'm asking for too much.

Objective video titling at it's finest. Kind of like the 'Hillary is drunk' title when she obv. isn't.

That said, for a guy reporting on a major news network, he's definitely triple jumped over the line of dispassionate reporting.

[wufwugy]

Big LOLs. Big league, bigly.

https://twitter.com/regated/status/777638188580036608

[Savy]

Trump thinking of the long game in 2011 lol I've never heard such bollocks. This was all a publicity stunt. At the very best he's looking at past incidents that happened and trying to somewhat optimise them. The random outcomes of decisions are impossible to predict for a week years is just stupid to think.

[wufwugy]

Trump thinking of the long game in 2011 lol I've never heard such bollocks. This was all a publicity stunt. At the very best he's looking at past incidents that happened and trying to somewhat optimise them. The random outcomes of decisions are impossible to predict for a week years is just stupid to think.

It was a publicity stunt too. Regardless, it makes more sense as a setup. When pacing and leading is easy after you know what pacing and leading is, when so much of what you do looks like pacing and leading, when the results you get from what looks like pacing and leading are remarkable, it's probably not coincidence and is instead pacing and leading.

[OngBonga]

I don't think it's a question of predicting outcomes of decisions, it's a matter of understanding how the public perceive certain aspects of a person, and predicting how those perceptions change over time as a result of certain actions and events. It's basically psychology, there's no Nostradamus shit going on.

[OngBonga]

In cases where it appears he has predicted an event of some sort, then sure, it's opportunism for publicity. But in the case of things like appearance, I think wuf makes some interesting points. He might or might not be right, but the idea that Trump is an expert when it comes to manipulating the way people think, as he calls "persuasion", that's not unreasonable.

[BankItDrew]

Wuf, that link for polls you shared from USC has been added to my bookmarks and I check it daily. I find it so confusing still as to how Trump is seemingly a huge favorite to win the popular vote according to them while all betting sites are still headed in the opposite direction. I assume this is just because sports bettors are not yet tapping into the true nature of the nations current feelings.

I'm also not convinced in the analysis of Scott Adams regarding Trump as a master of persuasion. I'm on ImSavy's side of the spectrum in thinking Trump doesn't plan as much as you think but rather reacting and capitalizing on his past mistakes.

A Trump presidency scares me. I think he's totally clueless and narcisistic. He's not doing it for the good of the people but rather the good of his branding.

[Poopadoop]

Wuf, that link for polls you shared from USC has been added to my bookmarks and I check it daily. I find it so confusing still as to how Trump is seemingly a huge favorite to win the popular vote according to them while all betting sites are still headed in the opposite direction. I assume this is just because sports bettors are not yet tapping into the true nature of the nations current feelings.

I don't know what this link is you're referring to, but most polls have Clinton ahead afaik.

Regardless of the polls, betting odds change depending on how many people are betting and on whom (also afaik), which collectively may be different from what an objective person would do. So if the polls go in favour of the odds against Clinton being longer and longer it's because more people are betting on Trump.

A Trump presidency scares me. I think he's totally clueless and narcisistic. He's not doing it for the good of the people but rather the good of his branding.

...or for this ego. I don't think Clinton would make a good pres either, I just think she's significantly less likely to start World War III after a car backfires somewhere in the US.

[Poopadoop]

When her meds are just right and she does actually come out and speak, she seems to me eminently more even-handed than Trump.





On another topic, does anyone else think this scene looks staged, like she's not actually standing in front of her plane here?

[OngBonga]

Don't be fooled by the bookies when it comes to politics. Brexit was about 5:1 at the start of voting.

[OngBonga]

What you have to understand is that when there is a huge public debate raging, many people just quietly keep their thoughts to themselves and only express their opinion in the form of voting. I think such behaviour has a tendancy to skew the polls in favour of the candidate or option that the media favours, because those who oppose the media's idea of the "positive" option are less likely to engage in public discussion or polls.

This is what I think happened during our referendum. Those who wanted out were sick of being accused of racism, so just kept their thoughts to themself until the vote, making it seem that the "remain" camp were ahead. I see parallels when it comes to Trump.

[Poopadoop]

What you have to understand is that when there is a huge public debate raging, many people just quietly keep their thoughts to themselves and only express their opinion in the form of voting. I think such behaviour has a tendancy to skew the polls in favour of the candidate or option that the media favours, because those who oppose the media's idea of the "positive" option are less likely to engage in public discussion or polls.

I don't have any detailed information about the whole polling issue, but I assume people are being polled with anonymity. It doesn't guarantee they're going to give an honest answer, but it seems to remove (or at least reduce) any possible embarrassment about who or what a person says they're voting for.

On an aside, I think an issue was raised with the accuracy of polls in general after one of the polls last year in some election. The poll itself ended up being way off. Well, no statistician is going to argue a poll is 100% accurate, and there's always a chance any given poll will be way off because it's drawing a sample and is not an exhaustive measure of the entire population.

There may be issues with how samples are drawn in some cases and whether they're representative. One example I recall hearing about was a telephone poll conducted in some distant year when home telephones were still only owned by relatively well-off people. Not surprisingly the sample wasn't representative because it only included people who owned telephones, and the poll was miles off the mark. But I think sampling issues are much more widely controlled for these days and so without knowing too much about how these election polls are being done I'd hazard a guess that most of them are as accurate or nearly as accurate as they claim to be (i.e., fairly accurate but not precise).

This of course assumes the pollsters aren't setting them up to give biased results, either deliberately or out of ignorance. This probably does happen a fair bit. That's why the source of the poll is often your best guide as to its reliability; the more disinterested the pollster, the more likely the poll is to be giving an accurate estimate.

[Poopadoop]

If this is the USC polls you're referring to, this is certainly an unorthodox approach to polling.

http://cesrusc.org/election/

The 2016 USC Dornsife / LA Times Presidential Election Poll represents a pioneering approach to tracking changes in Americans' opinions throughout a campaign for the White House. Around 3000 respondents in our representative panel are asked questions on a regular basis on what they care about most in the election, and on their attitudes toward their preferred candidates. The "Daybreak poll" is updated just after midnight every day of the week.

Just from a quick scan of their methods, they seem to be doing something sensible, but in all likelihood they're overcomplicating their analysis and probably skewing their results. For example, they're asking people to give a probability estimate of whether they plan to vote and another probability estimate of who they're planning to vote for. The they combine this with demographic data to equate each person's responses to a certain number of votes for candidate X.

The first probability question is not unreasonable but combined with the latter amounts, mathematically, to counting die-hard supporters of either candidate as more than one voter. My intuition is that what this poll is good at showing is how many people have already voted in their minds.

Finally, they don't seem to have any empirical support to suggest that what they're doing is more valid than traditional polling techniques. It all seems like one big experiment based on someone's reasonably-well-thought-out-but-not-necessarily-true assumptions about how good people are at estimating their own behavior. Certainly people are going to be pretty accurate when they believe 'I'm 100% likely to vote and 100% likely to vote for X', but how accurate they are at estimating probabilities outside 100% is questionable.

Anyways it's an interesting idea.

[Poopadoop]

This method seems a lot more sensible, as it combines data from other polls. It also has a proven track record of being very accurate in previous elections.

http://pollyvote.com/en/

Hillary up by 6 points.

Sorry Wuf.

[surviva316]

What you have to understand is that when there is a huge public debate raging, many people just quietly keep their thoughts to themselves and only express their opinion in the form of voting. I think such behaviour has a tendancy to skew the polls in favour of the candidate or option that the media favours, because those who oppose the media's idea of the "positive" option are less likely to engage in public discussion or polls.

This sounds like a groundbreaking argument for a dissertation/study, and something that wouldn't be all that difficult to verify with research. Maybe your armchair brilliance is unmatched by anyone in the field ... or maybe you're just coming up with slapshod theories for why the polls with your favored results are better than other polls. I don't know, just spitballing here.

Not sure why you're so obsessed with Brexit upset. The polling average was well within the margin of error, especially considering they were consistently reporting double digits for Undecided voters. Seems a little over-reactionary to completely blow up your idea of the value of polls over that one result.

[OngBonga]

Either that, or it's direct indluence by the powers that be, to skew polls and betting odds in favour of their preferred candidate in an effort to make it seem a forgone conclusion.

While that's possible, I don't consider it likely in this case.

[OngBonga]

A Trump presidency scares me. I think he's totally clueless and narcisistic. He's not doing it for the good of the people but rather the good of his branding.

While I couldn't argue with this, there's not a chance in hell that Clinton represents the opposite.

It's cat shit or dog shit, I'm afraid. Same old story then.

[BankItDrew]

While I couldn't argue with this, there's not a chance in hell that Clinton represents the opposite.

It's cat shit or dog shit, I'm afraid. Same old story then.

Clinton certainly doesn't represent the opposite but I'm not as afraid.

[OngBonga]

Clinton certainly doesn't represent the opposite but I'm not as afraid.

I am, if afraid is the right word. I don't trust her at all, she has a really nasty past and is neck deep in corruption. They don't call her crooked for nothing. America will not change under her leadership.

The best hope is that Trump is just a genius at winning votes and he's not nearly as batshit as people seem to think. I think that's wishful thinking though.

Trump represents the best chance of change, so I guess I lean in his direction. But it's with extreme caution.

[JKDS]

They don't call her crooked for nothing.

Trump started the "crooked hillary" brand so you'd say exactly this.

[wufwugy]

crooked hillary

What can I say, the man has a passion for the truth.

[OngBonga]

Trump started the "crooked hillary" brand so you'd say exactly this.

I'm pretty sure Hillary started it by being crooked in the first place.

[BankItDrew]

It's not staged. Media interviews upon landing take place all the time.

[Poopadoop]

It's not staged. Media interviews upon landing take place all the time.

What I mean is it doesn't actually look right, like it's been done in another room and the plane was added in afterwards. Not sure why they would do that though.

[wufwugy]

Poopy answered BID's question.

As for starting WW3, look at who the real problem is, who the real problem isn't, and what Hillary and Trump are doing. The real problem is ISIS but Obama is targeting ISIS's enemies, using them as a tool to destabilize eastern state powers (Iran, Russia). Hillary is embedded into this and uses escalated rhetoric against Russia. While the Democrats are acting as if we're still in Cold War and have an agenda of total state hegemony, Trump is focused on what is good for people (instead of what's good for the US government) by deescalating rhetoric with Russia and targeting ISIS.

Hillary gets us much closer to WW3 type conflict than Trump does. Trump would get us furthest from that sort of conflict we've had in, well, since WW2. He would be the first President whose foreign policy wouldn't revolve around what's best for the US government.

As for pollyvote, his methodology is stellar when the polling parameters are good. I have bypassed this by declaring the parameters are garbage. Because they are. These polls at large use 2012 demographics (bad idea), sometimes 2008 demographics (worse idea), often blatant over-stating of women or Democrat turnout (in absurd quantities), and only polling the "youngest member of the household" (their poor methodology couldn't be more clear).

I posted the LA USC link because it was the most helpful poll I found to predict 2012.

[Poopadoop]

As for pollyvote, his methodology is stellar when the polling parameters are good. I have bypassed this by declaring the parameters are garbage. Because they are. These polls at large use 2012 demographics (bad idea), sometimes 2008 demographics (worse idea), often blatant over-stating of women or Democrat turnout (in absurd quantities), and only polling the "youngest member of the household" (their poor methodology couldn't be more clear).

I can't be bothered to check all these assertions, but...

I posted the LA USC link because it was the most helpful poll I found to predict 2012.

Lots of other polls were fairly accurate about 2012 as well.

When almost all the standard analyses point strongly in one direction and an unproven, unnecessarily complicated analysis points in the other, it's statistical folly to follow the outlier and ignore the cluster. Doesn't mean Trump can't win, just means he probably isn't winning right now.

[wufwugy]

Lots of other polls were fairly accurate about 2012 as well.

When almost all the standard analyses point strongly in one direction and an unproven, unnecessarily complicated analysis points in the other, it's statistical folly to follow the outlier and ignore the cluster. Doesn't mean Trump can't win, just means he probably isn't winning right now.

I posted that poll as an example of a really good one that tells a different story than the shitty ones. I'm well aware of why you don't use just one poll.

Also, it doesn't mean he "probably isn't winning right now" because the parameters they're using are wrong. I'm a pretty big poll-watcher. These are much different than in 2012. A composite of shit isn't an apple. It's shit.

[Poopadoop]

Either that or you want Trump to win so you've convinced yourself this poll is right and the others are shit

[wufwugy]

Either that or you want Trump to win so you've convinced yourself this poll is right and the others are shit

I base very little of my opinions on the frontline of the polls.

Keep in mind that the smart money is on Trump. Nate Hydrogen et al have gotten this cycle wrong repeatedly. Just a couple years ago they were saying all sorts of stuff about how polling is on the verge of existential crisis. You'd think that they'd use this (and the many other) as a reason to reflect, but no, they're still marching the redcoats into a guerilla battle in orderly fashion.

[Poopadoop]

Here's the problem with the USC method as I see it. I could be wrong about what they're doing and if I am then please correct me.

When you poll people about an election, you want to get at two things: 1) Are you going to vote? and 2) Who are you going to vote for? Then you extrapolate these answers to the population to get an estimate of the outcome if the votes were to be cast today.

The USC poll asks these questions, but they do so in an unnecessarily complicated way, by asking people to give a % instead of yes/no like most polls do. When you say ' what is the likelihood you are going to vote?', that's a reasonable question because it allows you to weigh their choice on the probability of them actually voting.

But when you ask them, 'what is the probability you will vote for X?', and then use that answer to weigh their response, you're doing something silly. You're treating each person like a divisible voter. This is what I meant when I said they are weighing some people as having more than one vote - though to be accurate what I really should have said is they're weighing some people as having fractions of votes.

When the person says "I'm 100% going to vote for X', these guys count that as 1 votes for X and 0 for Y. That's fine. The problem is when the person says "I'm not sure, I'm 75% likely to vote for X and 25% likely to vote for you. These guys count that as 0.75 votes for X and 0.25 votes for Y. They then extrapolate these numbers to that person's demographic, so they assume that out of that group 75% will vote for X and 25 for Y.

That's wrong. When this person says "I'm 75% for X" they don't mean if they voted today, there's some random process going on whereby 75% of the time they'd vote for X and 25% of the time they'd vote for Y. What they mean is that they would vote for X but they're not entirely convinced it's the right choice. But they would still vote for X!

When you throw this unnecessary complication into your equation, you open your polling method up to bias. If one candidate has a lot of diehard, fanatical supporters and the other has a lot of more moderate supporters (as I suspect may be the case here) , your weighing the strength of their support as if being a fanatic gives you the voter more votes than does being a moderate supporter. Obviously it doesn't. All that matters is the answer to the two simple questions 1) Would you vote? and 2) Who for?

That's my best guess as to why that particular poll disagrees so much with the majority of the others: It's giving Trump fanatics excessive weighting over Hillary's moderate supporters.

[wufwugy]

Here is one the multitudes of examples: the candidate with the most excitement always wins. Always. This one mere basic heuristic alone performs far better than the best team of the best pollsters doing their best work.

Or here's another one: an incumbent polling <50% is very dangerous. If they're anywhere around a few points below that, they typically do not win. Crooked can be mostly assumed an incumbent in this cycle, as she shares most of the relevant elements that gives somebody the incumbent appearance. The fact that she is polling so low, the low 40s and high 30s, is a very good reason to look into the idea that the polling parameters are getting things very wrong.

The funny thing is that as somebody who reads this stuff for fun, the last time I saw journalists discuss it was when they could use it to explain why their favorite incumbent (Obama) was not in as much trouble as Romney supporters thought. But now that the rationale shows their favorite candidate (Crooked) in dire straits, mum's the word.

[Poopadoop]

Here is one the multitudes of examples: the candidate with the most excitement always wins. Always. This one mere basic heuristic alone performs far better than the best team of the best pollsters doing their best work.

Is there some scientific basis for this analysis? There's a lot of ways to define 'excitement'.

Or here's another one: an incumbent polling <50% is very dangerous. If they're anywhere around a few points below that, they typically do not win.

So in other words, the less they score in the polls, the less likely they are to win. That seems to validate polls, not say they're shit.

Crooked can be mostly assumed an incumbent in this cycle, as she shares most of the relevant elements that gives somebody the incumbent appearance.

Okay...

The fact that she is polling so low, the low 40s and high 30s, is a very good reason to look into the idea that the polling parameters are getting things very wrong.

I don't follow this. What is she polling at those numbers in? The election or something else? If you mean the election then that's only true in the polls you've decided are valid.

The funny thing is that as somebody who reads this stuff for fun, the last time I saw journalists discuss it was when they could use it to explain why their favorite incumbent (Obama) was not in as much trouble as Romney supporters thought. But now that the rationale shows their favorite candidate (Crooked) in dire straits, mum's the word.

Ya fine, but that's neither here nor there. We know about the bias in the MSM, for example. So it's not surprising the MSM looks for excuses to predict Clinton will win. I thought the question was which polls are valid and which are shit.

[wufwugy]

Is there some scientific basis for this analysis? There's a lot of ways to define 'excitement'.

there isn't but there absolutely should be. I think it would be super valuable to be able to predict excitement based on data. Regardless, this is one of the most established ideas in political minds.

So in other words, the less they score in the polls, the less likely they are to win. That seems to validate polls, not say they're shit.

I'm not saying they're invalid. They're definitely valid; the question is regarding what. These polls can easily not show that well what turnout will be while also showing low level of support for Clinton. Furthermore, the polls are very biased towards Clinton, which means that they should be showing above average numbers for her, yet they're showing below average.

Okay...

Same party candidate of the incumbent, has the media advantage, has the money advantage, has the insider status, is running effectively as a third Obama term. She has most of why the incumbent advantage is considered a thing.

[wufwugy]

These are probably last two from Hydrogen himself http://fivethirtyeight.com/features/is-the-polling-industry-in-stasis-or-in-crisis/

http://fivethirtyeight.com/features/the-state-of-the-polls-2016/

IIRC Hydrogen's thesis was that while polling could be in crisis, it probably isn't. Note: in second to last paragraph of first article he writes "we could see a disastrous year for the polls if and when political coalitions are realigned." Yeah, many signs are pointing to this being what's happening. The Obama coalition will not come together for Hillary, Trump is getting remarkable results in "blue wall" states, the primaries suggest a GOP landslide, etc.. In second article he writes "polls that employ more expensive methodologies, and abide by higher levels of disclosure and transparency, tend to be more accurate than those that don’t. It may be that the best polls are roughly as accurate as ever but that the worst polls are increasingly far off the mark." The reason this is the case is because of the underlying factors her discusses that contribute to it being harder and harder to get good data.

He doesn't think it's in crisis, but that it could be. The thing is that according to his analyses from before, a state of crisis would look similar to what we have now.

[Poopadoop]

These are probably last two from Hydrogen himself http://fivethirtyeight.com/features/is-the-polling-industry-in-stasis-or-in-crisis/

http://fivethirtyeight.com/features/the-state-of-the-polls-2016/

Just skimmed the first one, but the data in every one of those graphs suggest his concern is misplaced. It seems to puzzle him but I'm going to guess that's because his grasp of statistics is on another level to the one stats actually work on. Which is a polite way of saying he's not 'getting it'.

[wufwugy]

Just skimmed the first one, but the data in every one of those graphs suggest his concern is misplaced. It seems to puzzle him but I'm going to guess that's because his grasp of statistics is on another level to the one stats actually work on. Which is a polite way of saying he's not 'getting it'.

I think you might agree with him more than you think because he's definitely quite sanguine about the state of polling. At least in those articles he shows that results have gotten (slightly) better, but elsewhere he discusses more about the underlying characteristics that are getting worse that he's worried about.

[Poopadoop]

I think you might agree with him more than you think because he's definitely quite sanguine about the state of polling. At least in those articles he shows that results have gotten (slightly) better, but elsewhere he discusses more about the underlying characteristics that are getting worse that he's worried about.

I don't have time to check the veracity of his arguments about polls getting worse, but as I said the data doesn't back it up.

A more likely possibility is that individual polls have become less reliable but that it's reflected in their variability not their mean accuracy. You might expect that to happen if average sample sizes were getting smaller, for example. In other words different polls might give more varied estimates that still tend to cluster near the true value. But as he doesn't measure their variability it's impossible to tell that either.

This is what makes me think he doesn't really understand the issue of statistical reliability very well, because he's asking one question, not getting the answer he expects, but forgetting it's not the only, or even most likely, explanation.

[Poopadoop]

Sorry to keep rambling on about this, but this might be a better way of explaining things.

There's two ways a poll can be 'bad'. It can either be biased or it can be unreliable.

A biased poll is one in which the sample misrepresents the population. For example, if I went to a Trump rally and polled the people there on their voting choices, you'd expect me to get an answer within a hair's breadth of 100% Trump. Let's say I then decide to extrapolate that result and claim that nearly 100% of Americans are going to vote Trump. Meanwhile someone equally as stupid has done their poll at a Clinton rally and is now saying nearly 100% of Americans are going to vote Clinton. Both these polls are bad because they're both so obviously biased.

An unreliable poll is one in which the sample is not biased but there are issues with accuracy nonetheless. The most common issue is having too small a sample. If I poll 100 people using perfect random selection, polling different demographics equally, etc., there should be no bias in the results of the poll. It's likely to give an answer somewhere in the neighborhood of the truth. However, the size of that neighborhood depends on my sample size. Since my sample of 100 is so small I couldn't have very much confidence in how close my results were to the truth (the neighborhood is large). If I polled 1000 people the same way I'd have more confidence (the neighborhood gets smaller), 10,000 even more confidence (neighborhood gets even smaller), and so on.

All polls are likely biased to some degree at least and all polls are unreliable to varying degrees. Bias gets introduced automatically because you're polling people who can be bothered to respond to polls and it's not impossible that may correlate with particular voting choices. Some unreliability is unavoidable as well whenever you have to extrapolate from a sample to the population. The nice thing about reliability is that it can at least can be estimated through knowledge of statistics. That's why you hear things like 'accurate to within +/-4 points 95% of the time' - the smaller the value after +/- the more reliable the poll (the smaller the neighboorhood). Further, when averaged across multiple polls, this issue tends to be self-correcting, meaning that even a number of unreliable polls will ON AVERAGE give a reliable estimate (a small neighborhood).

These are two fundamental issues with sampling from a population. The difference is that bias tends to skew your results in one direction or another (think the pollsters at the Trump or Clinton rallies), whereas unreliability tends to skew your results in a random direction.

This Nate person claims that polls seem to be becoming 'bad'. Then he queries why they're just as accurate on average as previous polls. It's because either a) he's wrong about them becoming 'bad', or b) he is equating 'bad' with 'biased', which isn't necessarily true. 'Bad' could also mean 'unreliable' and as I said, unreliable polls tend to still give reliable results when you average across enough of them.

[OngBonga]

I'm sorry poop I just got back home at it's gone midnight, I'm a bit drunk, and I read your posts through, and by the time I got to the end, I forgot what we were talking about.

I'll try again in the morning.

[BankItDrew]

Hilary is crooked. Trump is an idiot. But Hilary is crooked.

[OngBonga]

Stop using American spellings when you're British, you fucking traitor.

[Poopadoop]

Stop using American spellings when you're British, you fucking traitor.

I'm Canadian, but I use American spellings even though we don't. British spellings are a waste of a keystroke.

[OngBonga]

I'm Canadian, but I use American spellings even though we don't. British spellings are a waste of a keystroke.

Oh. If you're Canadian, then well done for not speaking French. I retract my "fucking traitor" insult.

[Poopadoop]

If you're Canadian, then well done for not speaking French.

Funny thing is they made us learn French when I was in school even though I live about 2500 miles closer to any French people now than I did then.

[OngBonga]

I think you missed one category - bullshit polls. They're they ones where they ask people who they're going to vote for, then don't even bother counting it up and just print whatever number they want. In fact, they might not even bother asking anyone, they could just pretend they did.

I think those kind of polls exist too.

[Savy]

I think you missed one category - bullshit polls. They're they ones where they ask people who they're going to vote for, then don't even bother counting it up and just print whatever number they want. In fact, they might not even bother asking anyone, they could just pretend they did.

I think those kind of polls exist too.

Surely that falls under biased?

[OngBonga]

Surely that falls under biased?

Well, I suppose, although then you'd have to have a distiction between deliberate bias and indirect bias. Someone who asks a load of Trump supporters who they intend to vote for could just be too dumb to realise their poll will be biased.

[Poopadoop]

I think you missed one category - bullshit polls. They're they ones where they ask people who they're going to vote for, then don't even bother counting it up and just print whatever number they want. In fact, they might not even bother asking anyone, they could just pretend they did.

I think those kind of polls exist too.

Probably, but that's why you don't just look at one poll and accept it. You need to look at all of them together, and if you really care, you need to go deeper and check their methods, etc..

Thankfully I don't care enough to be bothered doing all that hard work, but I still know what makes a good poll and what makes a bad one and what the different kinds of bad are.

[OngBonga]

They make us learn French too. At least they did when I was at school. Not that I paid much attention. I know how to buy milk from a shop. Prance around saying a stupid rhyme. Je voudrais un litre de lait si vous plais. I bet my spelling is woeful. I fucking hated French.

[BankItDrew]

Poop, who do you think will win and by what margin?

[Poopadoop]

Poop, who do you think will win and by what margin?

Don't know. I think Hillary is ahead now, but not by a lot. If there's a serious terrorist attack on US soil, or if she fails at masking her brain damage too many times she will prolly lose.

Part of me wants Trump to win just to see what happens. I think it'll be disappointing though (in the sense that he won't be as crazy as he seems to act sometimes). There are a lot of checks on a president's power that make it unlikely some of his more radical ideas will get implemented.

Another part of me wants Clinton to win just so we can get a better confirmation on what the fuck is wrong with her. It'd also be entertaining to watch her slowly being replaced by the world's first 3D printed copy of a country's leader.

[wufwugy]

Good write up Mr. Poopy

I think what Mr. Hydrogen discusses is bias. There are signs that it is becoming harder and harder for pollsters to maintain standards regarding bias.

My points are not so much about the stats themselves but with the political assumptions, because those are creating what other signs are strongly suggesting to be very significant bias. An easy example is how the pollsters this cycle typically weight for 2012 demographics, yet no non-stupid assessment of the political landscape would suggest that Clinton can reasonably be expected to get the 2012 turnout. The reason pollsters are doing this appears to be either (1) they don't know what else to do or (2) they're playing their part as a persuasion tool for Her Majesty. If I'm completely honest, I think some of it is the latter (and in subtle ways) but most of it has elements of the former. 2012 demos is the best of the not-hard ways to do it. Since polling is an inherently conservative practice, there is little going outside of the box.

Which brings us to my next point. It's regarding your point yesterday about your concern regarding the USC poll methodology. My interpretation is that they are thinking outside of the box in an attempt to more correctly predict "likely voters." This team's comparative advantage can be thought of as their innovation. They did some things differently in 2012 and got very fantastic results. Many signs are pointing towards the turnout modeling for 2016 likely looking nothing like 2012, so they're trying to predict that best they can.

Another point is that sometimes some pollsters have achieved fantastic results by merely not weighting for some standard variables. Ann Selzer, for example, was lauded as the best pollster alive since her polls blew all the others out of the water in the 2012 primaries. Her brief explanation for why she caught the trends better than anybody else was that she merely reported what the respondents were telling her. Instead of assuming the Iowa caucus would look similar in demos to the previous ones, she reported her poll with as few weights as she could. Lo and behold, this gave her a much more accurate measure of the turnout because it allowed her to catch what could probably be called an enthusiasm shift.

This cycle we have seen several polls with unweighted data showing Trump ahead but then the pollsters weight for 2012 demos and it shows him significantly behind. I do not know why pollsters don't release their polls with various assumptions. They should do something like release each of their polls with one iteration as 2012 demos, one a little closer to response demos, and probably one with predicted demos based on other variables. The latter would be an attempt to include things like the shy Tory effect or evaluating what the change in primary turnout means. Regarding just these two, there are very likely a bunch of blue collar men that typically do not respond to polls and don't vote, yet are going to vote this time; and a state like New Hampshire can be predicted to fall more to the right than polling suggests due to changes in its primary turnout numbers from 2008 to 2016.


My last point is going back to Mr. Hydrogen. Even though he shows that polling accuracy has slightly increased, the reason why includes techniques that create a reduction in predictive power during "off" cycles. Polls are getting more accurate by weighting more accurately, yet they're only able to weight more accurately because there has been little political realignment. We are very likely looking at a political realignment year. New Jersey and Rhode Island have polled within the margin of error ffs iirc. Pennsylvania alone is in a very unique, tip of the iceberg situation. Solid analyses have shown that the state is trending red rather swiftly even though on the surface it went for Obama twice. It just so happens that what Trump targets is exactly where the redness has been coming from. It is quite reasonable to predict that we are looking at a solid probability of the "blue wall" breaking, but as long as pollsters assume that doing the same thing they've always been doing is the way to go, they're not going to catch it before it happens.

[Poopadoop]

Ok, well I don't have any way of disputing all that without getting more deeply involved in understanding the whole polling mechanics thing than i really want to. So I'm happy to say you could be right.

[Poopadoop]

wtf is she trying to accomplish here?