USPOL Fall: Biden the hand that feeds you

The Something Awful Forums > Discussion > Debate & Discussion > USPOL Fall: Biden the hand that feeds you

ewiley: Jul 9, 2003; More trash for the trash fire

Arglebargle III posted:

Can you quote this for those of us who don't support the neoliberal death cult? /s

quote:

The transition away from live interview telephone polling is well underway, and online polls now make up the principal source of data on national public opinion.

But the alternatives to traditional polling are not fully mature, and the absence of a clear set of standards for online polling research has opened the floodgates to unproven surveys of uncertain quality.

Many of the new pollsters have become familiar enough that journalists and readers have started to cite them with confidence. It�s not yet clear they all deserve it.

Of course, traditional live-interview telephone polls have been facing their own challenges. They have become very expensive, and they haven�t necessarily been faring that much better than many online surveys.

But even today, there is substantial evidence that it remains easier to construct a high-quality live-interview telephone poll than an online poll of equal quality. Just as important, it�s easier to judge the quality of a traditional poll than a nontraditional poll.

All of this makes it more crucial than ever for analysts to rely on an average of public polls, rather than citing individual surveys. And it raises the burden of proof on innovative firms to demonstrate the effectiveness of their approach.

The clear standards are gone
A decade ago, it was fairly easy to tell the difference between a good poll and a bad poll. Most pollsters assumed that high-quality survey research required a high-quality survey design: a so-called probability sample, meaning that everyone in the public had an equal chance of being included in a poll. A poll was either a good probability sample (from random dialing of telephone numbers, for example) or it was not (like an online poll).

That divide is no longer so clear. Even high-quality probability designs don�t yield representative samples, because fewer and fewer people respond to telephone surveys. Instead, pollsters rely on statistical adjustments, known as weighting, to correct their initial samples. A high-quality design without proper adjustment might be fatally flawed, as with Gallup�s results during the 2012 presidential election season, or with many state polls in the 2016 election.

At the same time, the advent of nearly universal internet access has inaugurated a new era of experimentation in public opinion research. The online firms, scorned a decade ago, have developed techniques to improve the representativeness of their samples. They have applied statistical methods to iron out whatever other issues remain.

Online pollsters tend to take a model-based view of survey research. They don�t think polls work simply because of their design � though a good design helps � but instead because demographics and other characteristics can predict vote choice, and polls are weighted to match the country�s demographics.

But the model-based view doesn�t offer the clarity of the old, exclusively design-based way of thinking. Taken to the extreme, the model-based view could allow a good model to overcome the weakness of even the lowest-quality sample, in effect turning dirt into gold. A study, for instance, did a decent job of estimating the 2012 election results based on an Xbox poll.

Online polling remains difficult
Just because it�s possible to turn dirt into gold doesn�t mean every nontraditional pollster has figured out how to do so. It�s quite difficult, and it�s also hard to tell whether a pollster has unlocked the dirt-to-gold alchemy.

What we know for sure is that much of the data collected by nontraditional means is of the dirt variety. For example, a 2016 study from Pew Research found there were substantial differences in the quality of online nonprobability samples: All but two of those samples fared worse than Pew�s online probability panel, which was recruited from Pew telephone surveys.

The online samples that fared well raise some concerns about the other online pollsters, if only by implication. YouGov, for instance, consistently fared best in the Pew test in part by using a distinctively sophisticated method of sample selection, called synthetic sampling. YouGov selects individuals from its panel of respondents, one by one, to match the demographic profile of individual Americans and thus match the country�s demographics as a whole.

Another survey has succeeded without probability samples, but by means that also raise doubts about other online surveys: the VoteCast survey, a new competitor to the exit poll fielded by NORC at the University of Chicago and sponsored by The Associated Press and other news media outlets.

The VoteCast polling of the midterms was conducted with unusual transparency, and it was a rare foray into this type of research by a well-regarded, traditional organization. The study combined a traditional telephone survey of 40,000 respondents with a large nonprobability sample of 110,000 respondents. The online-only element was calibrated in part by using the telephone survey data, and the overall results were reasonable.

But without calibration, the online-only element of the VoteCast survey would have been a mess. The nonprobability estimates, after traditional weighting, overestimated the Democrats by an average of about five percentage points across 71 races. The noncalibrated nonprobability data would have been the only pre-election poll to show Heidi Heitkamp ahead in the 2018 North Dakota Senate race over the final six months of the race, and it would have given her a comfortable seven-point lead. (She lost by 11 points)

This kind of calibration based on probability samples is not easy. A large, simultaneous telephone survey (the VoteCast approach) won�t be an option for most pollsters. Other methods could require extensive statistical modeling that goes beyond the training of most survey researchers.

Little evidence most online firms have pulled it off
The experiences of VoteCast, Pew and YouGov suggest that applying traditional sampling and weighting techniques to nontraditional data � the way most online pollsters do � might not be enough to produce high-quality results.

The online pollsters have, for the most part, put forward relatively little evidence to demonstrate the effectiveness of their approach. Perhaps most surprising, many of the most prolific online pollsters didn�t provide surveys about the 2018 congressional elections.

Those that did had decent but not exceptional results. On average, online polls had an error of 5.3 percentage points in 2018 � a tally very close to their average error in 2016 and 2017, as reported by FiveThirtyEight.

YouGov, it should be noted, posted an impressive 2.5-point average error in its final 2018 polls; no other online-only firm had an average error beneath five points.

The overall 5.3-point average error is better than that of the uncalibrated AP/VoteCast, perhaps because most online firms weight by either self-reported party identification or by vote choice in the 2016 election. This is a statistical bludgeon; it can force even the worst data into the ballpark of an election result. But it comes with the risk of systematic bias: If a pollster guesses wrong on party identification, the poll will probably be wrong as well.

And if these polls rely on party weighting to achieve the level of accuracy that they do, they might be even less effective during a primary season, for instance, when everyone is in the same party, or in a changing partisan environment, when public polling is most valuable. In recent weeks, the online pollsters have shown widely divergent results on the Democratic race.

Either way, it is a plain step back from the historical level of accuracy of the so-called gold standard polls that have adhered to traditional best practices. This is not an exhaustive list, but the pollsters that have used those practices include ABC/Washington Post, CBS/NYT, NBC/WSJ, Pew Research, the Selzer poll, Monmouth University or the Upshot/Siena polls. Those pollsters, typically with brand names, have had an average error, combined, of 3.5 points in nonpresidential general elections since 1998.

But traditional live-interview polls have also had struggles in recent cycles, particularly in state polling. In 2018, for instance, live-interview surveys had an average error of four points. That�s better than the online polls, on average, but not by so much. (FiveThirtyEight has a listing of all polls and whether they�re online-only or not.)

What does it mean for poll watchers?
In general, the simple solution for these unproven online polls is to average them, rather than pick apart the methodological details of individual surveys. But in some cases, people will have reason to consider the merits of a single poll.

One important consideration is what kind of error is tolerable for the question at hand.

It probably doesn�t make a huge difference, for instance, whether 60 percent or 64 percent of Americans think the country is on the right track. A five-plus-point survey error could be quite acceptable for that purpose.

But a five-plus-point error might be intolerably high for many kinds of election analyses, particularly in close races. Here it�s important to focus on an average, or at least on higher-quality polls.

What helps distinguish the higher-quality online polls? In general, it is reasonable to look for a transparent methodology with signs of a more sophisticated weighting or sampling approach, like YouGov�s synthetic sampling, rather than a simple draw of nonprobability sample weighted by census demographics. It is also reasonable to look for a sample drawn from a panel with diverse sources, rather than from a single place on the internet. But this will not be easy for most poll consumers to discern, and even the experts struggle to make sense of it.

It seems reasonable to be relatively skeptical of online state or district polls; amassing a sufficiently diverse sample for a representative poll can be difficult.

The online firms could help earn confidence from poll watchers by releasing more evidence. For instance, they could publish individual-level microdata; they could publish more thorough methodologies; they could publish research, like the Pew study, that shows off their product.

Until then, it will be hard to identify the online pollsters that might be roughly on par with the live-interview polls, and those that might grossly underperform.

Nate Cohn is a domestic correspondent for The Upshot. He covers elections, polling and demographics. Before joining The Times in 2013, he worked as a staff writer for The New Republic. @Nate_Cohn

e: to clarify i fully support the neoliberal death cult, tyvm

# ? Nov 23, 2019 23:21

Adbot: ADBOT LOVES YOU

# ? Jun 6, 2024 02:27

JonathonSpectre: Jul 23, 2003; I replaced the Shermatar and text with this because I don't wanna see racial slurs every time you post what the fuck; Soiled Meat

Feldegast42 posted:

If that happens Trump gets to appoint their replacements and the Republic officially ends as they back him when he inevitably contests the 2020 election

That's the quickest way we will end up a fascist dictatorship

That's assuming there is no political fallout whatsoever from the entire military command structure resigning in protest over the incompetence of the commander-in-chief.

The trap this would put every elected Republican in is absolutely fiendish in its inescapability. Now it's the entire military, the ones you've glorified for, oh, EVER, going against your circus sideshow carnival barking con man. I guess the military are the traitors now? All of them? This is such whiplash up-is-down insanity it might even end up... >gasp< MATTERING.

# ? Nov 23, 2019 23:23

BigBallChunkyTime: Nov 25, 2011; Kyle Schwarber: World Series hero, Beefy Lad, better than you.; Illegal Hen

Wait are people really mad that the Navy DOESN'T want to let the war criminal back in?

# ? Nov 23, 2019 23:29

The Glumslinger: Sep 24, 2008; Coach Nagy, you want me to throw to WHAT side of the field?; Hair Elf

BigBallChunkyTime posted:

Wait are people really mad that the Navy DOESN'T want to let the war criminal back in?

Trump is very pro-war crimes. It was a major campaign pledge to commit more of them

# ? Nov 23, 2019 23:30

Arglebargle III: Feb 21, 2006

Performative cruelty is like a pillar of conservative popularity around the world my dude.

# ? Nov 23, 2019 23:32

BigBallChunkyTime: Nov 25, 2011; Kyle Schwarber: World Series hero, Beefy Lad, better than you.; Illegal Hen

The Glumslinger posted:

Trump is very pro-war crimes. It was a major campaign pledge to commit more of them

No, I meant some Goons in this thread.

# ? Nov 23, 2019 23:32

Feldegast42: Oct 29, 2011; COMMENCE THE RITE OF SHITPOSTING

Telsa Cola posted:

Yeah thats not going to happen for various reasons.

That's what we all said about Trump getting elected

My fears of a coup have gone down but complacently assuming that the institutions of our society will always do the right thing in the end got us to this point in the first place

Also Gallagher should be hung high for what he did, my whole point is that giving the GOP a chance to purge and replace the top levels of the military right before what is going to be a very contested election one way or another would be a Very loving Bad Idea

I guess the compromise solution would be to obey Trump's very immoral order and then stick Gallagher and his team on latrine duty until they are discharged or we get a president that will throw him back into Leavenworth

Feldegast42 fucked around with this message at 23:46 on Nov 23, 2019

# ? Nov 23, 2019 23:34

TVs Ian: Jun 1, 2000; Such graceful, delicate creatures.

BigBallChunkyTime posted:

No, I meant some Goons in this thread.

Not really.

One side says the officers involved should flat out refuse the order (if given in non-tweet form) rather than resign.

The other side says that refusing a legal (if immoral) order is a bad precedent, because the military could then refuse to do stuff if someone like Bernie were to get elected. They're not saying that psycho seal should be reinstated, they're saying that quitting would be the preferable option if that does happen.

# ? Nov 23, 2019 23:42

Medium Chungus: Feb 19, 2012

Unoriginal Name posted:

Every person who argues that the military must follow orders should consider what a nation with a military that unconditionally follows Trump looks like.

Ignoring amoral orders is 100% OK

And in fact something we are REPEATEDLY told we have a duty to do.

# ? Nov 23, 2019 23:46

oxsnard: Oct 8, 2003

https://twitter.com/AdamSerwer/status/1198326504792494082?s=20

# ? Nov 23, 2019 23:51

Young Freud: Nov 26, 2006

https://twitter.com/JordanUhl/status/1198065708053204992?s=20
(warning: speakers)

# ? Nov 23, 2019 23:57

Stereotype: Apr 24, 2010; College Slice

About polls: it�s worth noting that you can legally straight up fabricate everything about a poll, from the methodology to the raw data to the interpreted results, and as long as you just copy what the results of the other guys are for awhile and continuously release new polls regularly you will eventually be taken seriously by some news organization. At that point you can just start skewing things however you want. This is why lots of campaigns do internal polling since at least they have some sort of look into the internals and can trust it a bit.

There is basically no scientific rigor in public polling and it�s astounding how much people rely on it.

# ? Nov 23, 2019 23:58

SirPablo: May 1, 2004; Pillbug

mistaya posted:

Apache Junction being called a suburb of Phoenix is misleading. It's way out in the boonies. There used to be a chunk of empty desert between it and East Mesa but I don't know if it's been built out yet by now. I guess you could technically call it a suburb but it's a good 40-50 minute drive from Phoenix to Apache Junction, and that's essentially as the crow flies since you just take 60 east and it's a completely straight road.

Apache Junction is where you go to live if you don't mind an hour commute twice a day, can't afford a house in gilbert/chandler, and think mesa is too scummy. (Anything north and west of Phoenix is retirement communities and Tempe is the college town.)

AJ absolutely is a suburb of Phoenix. Just like Buckeye is, or Anthem or I'd even include Maricopa.

That area of Phoenix is very conservative and a hot bed of Mormon activity. Mostly represented at the US Congress by Andy Biggs and Paul Gosar (the guy who's siblings were in an ad saying don't vote for our brother). It's Chudtown out there.

# ? Nov 24, 2019 00:34

canepazzo: May 29, 2006

TRUTH ALERT

https://twitter.com/RudyGiuliani/status/1198352974000644097

# ? Nov 24, 2019 00:35

The Glumslinger: Sep 24, 2008; Coach Nagy, you want me to throw to WHAT side of the field?; Hair Elf

canepazzo posted:

TRUTH ALERT

https://twitter.com/RudyGiuliani/status/1198352974000644097

RICO chart

# ? Nov 24, 2019 00:36

sexpig by night: Sep 8, 2011; by Azathoth

it's ok guys the admirals are already backing down like giant pussies, the war criminal remains protected by daddy

https://www.businessinsider.com/navy-leaders-threat-quit-trump-block-expulsion-navy-seal-gallagher-2019-11

quote:

The secretary of the Navy and the admiral overseeing the Navy SEALs threatened to resign or be fired if President Donald Trump blocks the expulsion of a member of the Navy SEALs, The New York Times reported on Saturday, citing administration officials.

Navy Secretary Richard V. Spencer, however, denied the report at the Halifax International Security Forum on Saturday afternoon.

"Contrary to popular belief, I'm still here. I didn't threaten to resign," Spencer said during a panel on Arctic security issues. "But let us just say that we're here to talk about external threats, and Eddie Gallagher is not one of them."

# ? Nov 24, 2019 00:36

nine-gear crow: Aug 10, 2013

canepazzo posted:

TRUTH ALERT

https://twitter.com/RudyGiuliani/status/1198352974000644097

Wai wh wha hua uh t- turn on M- Mr. Trum-mp? M-me? Nooooooooooo. Never. Nevernevernvere. I'd never say that. Ever. You must be confused boys. Wait. Why are we going up to the roof? NOOOOOOOOOOOOOOO!!!

# ? Nov 24, 2019 00:37

ewiley: Jul 9, 2003; More trash for the trash fire

canepazzo posted:

TRUTH ALERT

https://twitter.com/RudyGiuliani/status/1198352974000644097

USPol Fall: RICO Chart

# ? Nov 24, 2019 00:37

sexpig by night: Sep 8, 2011; by Azathoth

RICO CHART

DONNY NEEDS MONEY

RICO CHART

DONNY NEEDS MONEY

# ? Nov 24, 2019 00:38

T. Bombastus: Feb 18, 2013

BigBallChunkyTime posted:

Wait are people really mad that the Navy DOESN'T want to let the war criminal back in?

This is an aggressively uncharitable reading of what people itt are saying.

# ? Nov 24, 2019 00:39

sexpig by night: Sep 8, 2011; by Azathoth

T. Bombastus posted:

This is an aggressively uncharitable reading of what people itt are saying.

I mean, they are handwringing over the idea that 'naval officers could and should refuse to follow unethical commands done for ignorant and selfish reasons' being exactly the same as 'what if entire military rebels against Bernie and shoots him in the street would you like that huh???' so they're not a bright bunch to begin with

# ? Nov 24, 2019 00:41

Piell: Sep 3, 2006; Grey Worm's Ken doll-like groin throbbed with the anticipatory pleasure that only a slightly warm and moist piece of lemoncake could offer; Young Orc

gently caress
https://twitter.com/adamliptak/status/1198384873838403584

# ? Nov 24, 2019 00:43

Ate My Balls Redux: Aug 2, 2018; by LITERALLY AN ADMIN

BigBallChunkyTime posted:

Wait are people really mad that the Navy DOESN'T want to let the war criminal back in?

I don't know if you know this, but the Trump cult is so strong that if he told them to kill themselves to own the libs, at least half of them would

# ? Nov 24, 2019 00:45

BigBallChunkyTime: Nov 25, 2011; Kyle Schwarber: World Series hero, Beefy Lad, better than you.; Illegal Hen

canepazzo posted:

TRUTH ALERT

https://twitter.com/RudyGiuliani/status/1198352974000644097

Rudy getting paranoid.

# ? Nov 24, 2019 00:48

SpaceCadetBob: Dec 27, 2012

Piell posted:

gently caress
https://twitter.com/adamliptak/status/1198384873838403584

This is the worst ticking time bomb.

# ? Nov 24, 2019 00:50

mcmagic: Jul 1, 2004; If you see this avatar while scrolling the succ zone, you have been visited by the mcmagic of shitty lib takes! Good luck and prosperity will come to you, but only if you reply "shut the fuck up mcmagic" to this post!

Piell posted:

gently caress
https://twitter.com/adamliptak/status/1198384873838403584

I can see Mitch McConnell's hardon from here.

# ? Nov 24, 2019 00:50

Charliegrs: Aug 10, 2009

canepazzo posted:

TRUTH ALERT

https://twitter.com/RudyGiuliani/status/1198352974000644097

Why wait Rudy? Oh that's right you probably don't have poo poo.

# ? Nov 24, 2019 00:50

Stereotype: Apr 24, 2010; College Slice

canepazzo posted:

TRUTH ALERT

https://twitter.com/RudyGiuliani/status/1198352974000644097

Oh okay, so when he said that he had an insurance policy against Trump throwing him under the bus, he of course meant that he had a bunch of info that was actually super HELPFUL to Trump that he would release if Trump ever turned on him. This clears everything up.

God drat how is this poo poo real.

# ? Nov 24, 2019 00:53

Charliegrs: Aug 10, 2009

mistaya posted:

Apache Junction being called a suburb of Phoenix is misleading. It's way out in the boonies. There used to be a chunk of empty desert between it and East Mesa but I don't know if it's been built out yet by now. I guess you could technically call it a suburb but it's a good 40-50 minute drive from Phoenix to Apache Junction, and that's essentially as the crow flies since you just take 60 east and it's a completely straight road.

Apache Junction is where you go to live if you don't mind an hour commute twice a day, can't afford a house in gilbert/chandler, and think mesa is too scummy. (Anything north and west of Phoenix is retirement communities and Tempe is the college town.)

Yeah no AJ is totally a suburb of Phoenix. And it's hellhole coal rollin Trump country.

# ? Nov 24, 2019 00:54

ImpAtom: May 24, 2007

SpaceCadetBob posted:

This is the worst ticking time bomb.

I genuinely want to punch every single person who said "oh four years of Trump won't be that bad" in the face because even if he gets impeached tomorrow the Supreme Court is hosed for the rest of our lives.

# ? Nov 24, 2019 00:54

Furnaceface: Oct 21, 2004

nine-gear crow posted:

Why do I feel this story needs a �Nunes then turned and walked face first into a telephone pole� post-script?

This is from forever ago (last night) but Im kind of sad you missed a perfect opportunity to post this:

# ? Nov 24, 2019 00:54

Doctor Butts: May 21, 2002

Never figured Rudy for a Starship Troopers fanatic. That Rico chart must be pretty detailed.

# ? Nov 24, 2019 00:55

Stereotype: Apr 24, 2010; College Slice

Honestly if Ginsberg dies and Trump gets to appoint another SC justice I�m definitely going to just go off on Twitter.

# ? Nov 24, 2019 00:56

oxsnard: Oct 8, 2003

This is morbid to talk about, but RBG needs to hang on for 6 months. After that, Dems should be able to obstruct until the election

# ? Nov 24, 2019 00:56

Stereotype: Apr 24, 2010; College Slice

ImpAtom posted:

I genuinely want to punch every single person who said "oh four years of Trump won't be that bad" in the face because even if he gets impeached tomorrow the Supreme Court is hosed for the rest of our lives.

People should be constantly forcing every Dem candidate to address how they are going to fix the fact that the Supreme Court is wholly illegitimate. It is easily the most important issue of our time.

# ? Nov 24, 2019 00:57

WoodrowSkillson: Feb 24, 2005; *Gestures at 60 years of Lions history*

sexpig by night posted:

it's ok guys the admirals are already backing down like giant pussies, the war criminal remains protected by daddy

https://www.businessinsider.com/navy-leaders-threat-quit-trump-block-expulsion-navy-seal-gallagher-2019-11

Lmao the military are all such cowards

# ? Nov 24, 2019 00:58

The Super-Id: Nov 9, 2005; "You know it's what you really want."; Grimey Drawer

I feel like RBG getting replaced with a Trump appointee would make stacking the courts much more likely from the next dem admin. It�s certainly a scary thought though.

# ? Nov 24, 2019 01:03

DebtBeat: Oct 17, 2016

Piell posted:

gently caress
https://twitter.com/adamliptak/status/1198384873838403584

Don't worry, RBG just needs to survive until the new year, remember that according to Mitch McConnell we can't confirm new SCOTUS justices during an election year

# ? Nov 24, 2019 01:04

Scott Forstall: Aug 16, 2003; MMM THAT FAUX LEATHER

DebtBeat posted:

Don't worry, RBG just needs to survive until the new year, remember that according to Mitch McConnell we can't confirm new SCOTUS justices during an election year

She'll die on NYE and trumo will be giddy

# ? Nov 24, 2019 01:08

Adbot: ADBOT LOVES YOU

# ? Jun 6, 2024 02:27

i am harry: Oct 14, 2003

Hellworld, friends. Never forget where we are.

# ? Nov 24, 2019 01:11

The Something Awful Forums > Discussion > Debate & Discussion > USPOL Fall: Biden the hand that feeds you

«‹›1346 »