Why does Fox keep oversampling Dems? - Politics Forum.org | PoFo

Wandering the information superhighway, he came upon the last refuge of civilization, PoFo, the only forum on the internet ...

Political issues and parties in the USA and Canada.

Moderator: PoFo North America Mods

Forum rules: No one line posts please.
#15041562
Some guys on the internet have recently been arguing that Fox New's impeachment poll oversampled Democrats by almost 20%, without which support for impeachment would be basically unchanged from before the Ukraine thing started up.

This got me thinking about the polls in general. How is it that Trump's average approval rating (which includes many left-leaning polls) continues to be higher than Obama's was at this point in his Presidency, yet he consistently loses to random Democrats in a head-to-head matchup?

Fortunately, it appears that Fox News publishes all of their polling data online at Scribd. So anyone can go and read it. As it turns out, they are also oversampling Democrats in their Presidential head-to-head matchups by about 10% (edit: I did some double-checking and originally I had this at 20%).

When adjusted for this, it suggests that Trump could be approaching 50% of the vote. To put it very briefly, recently Elizabeth Warren was polled as getting like 51% of the vote (which doesn't pass the smell test) to Trump's 44%. But Democrats were oversampled by 8%. If you keep it simple and subtract 8% from 51%, you get 43% of the vote. It is not really that simple though because as the Republican and Independent ratio of votes increase, with Trump having an over 90% approval rating within the Republican Party, his voting numbers would go up as the Democrat's numbers go down. Furthermore, Republicans enjoy an electoral college advantage which can give them a 1% or more increase to relative vote counts. Finally, the usual registered vs. likely voter difference further inflates Republican votes beyond whatever a registered voter poll suggests. With these things in mind, Trump would cruise comfortably into re-election.

One way to read this information is that Fox News is leading us along towards another shocking ratings bonanza. Literally every Democrat polls above Trump because they're getting a 10% or more bump at the sampling level, meanwhile, most of those candidates don't seem like they're even trying to win a general election.

Although it's easy to point these things out, I feel that it's irresponsible to give people a false impression of reality, which is what the entire news media seems content to do these days, albeit for slightly different reasons.

Edit: I tweaked the numbers a bit.
Last edited by Hong Wu on 13 Oct 2019 02:04, edited 2 times in total.
#15041565
It's simple. There are more Dems(and independents) than Repubs.
When they say stuff you agree with, you don't complain about it. Why the fuss, now? :?:
#15041602
I'd really hesitate to put a lot of confidence in the accuracy of polls. The disappearance of the telephone landline as a fixture of modern life has thrown an enormous monkey wrench into the methodology of scientific polling.

Landlines skew heavily towards the elderly and rural areas with limited cell service. Post-boomer generations are almost universally cellphone centric. They rarely answer calls from numbers they don't recognize, and often they have installed blocking apps. My phone is set to block all numbers not in my contact list, for instance. The barriers to obtaining a truly representative sample are very high. As a result, polling organizations have been forced to post-weight their data to reflect these built-in limitations, and the accuracy of such weighting is impossible to assess in advance.
#15041620
quetzalcoatl wrote:I'd really hesitate to put a lot of confidence in the accuracy of polls. The disappearance of the telephone landline as a fixture of modern life has thrown an enormous monkey wrench into the methodology of scientific polling.

Landlines skew heavily towards the elderly and rural areas with limited cell service. Post-boomer generations are almost universally cellphone centric. They rarely answer calls from numbers they don't recognize, and often they have installed blocking apps. My phone is set to block all numbers not in my contact list, for instance. The barriers to obtaining a truly representative sample are very high. As a result, polling organizations have been forced to post-weight their data to reflect these built-in limitations, and the accuracy of such weighting is impossible to assess in advance.

This all may be true but it doesn't get to the heart of my question, which is why is Fox oversampling their opposition.

To get directly to one of the things I want to say, if I, as a hypothetical Trump supporter, saw a bunch of polls from MSNBC which said that Trump is going to win and then I went into their methodology and saw that they were oversampling his base by around 10% and that it actually looks like he's going to lose, I would be concerned, not pushing the poll. I think there's a lot that can be extrapolated from this situation, about both Fox and the people who hate Fox yet are accepting Fox's polls because they're superficially good for them.
#15041625
I'm not gonna read the dumb OP that I'm sure meanders around something simple that Tucker Carlson already said, but Fox News spends its time bitching about Democrats because the entire ideology of MAGA idiots and the Republican Party (same thing) is hating whatever rich Hollywood liberals are doing. If they didn't have bitterly disagreeing with meaningless shit like drag queens reading books or TV shows being cancelled they wouldn't have an ideology. The puppet show of Fox News watchers vs. Late Night comedy watchers is pathetic and both sides ( ;) ) are neutered idiots, but the former are the only ones I'm concerned about murdering people for their twitter-mania.
#15041645
quetzalcoatl wrote:I'd really hesitate to put a lot of confidence in the accuracy of polls.

Why they got the result right in 2016. Clinton won the popular vote as predicted. I believe the final average was off by about 2%. That's fairly good polling. American polling is fairly good overall if you know how to analyse it. British polling is overall not as good as American polling, but its still a brilliant source of data.

Polling is an art and a science, but very often the problem is that people don't trust the polls enough, including the pollsters, which can lead to herding. Outliers are vital to improving the overall polling averages.
#15041779
Rich wrote:Why they got the result right in 2016. Clinton won the popular vote as predicted. I believe the final average was off by about 2%. That's fairly good polling. American polling is fairly good overall if you know how to analyse it. British polling is overall not as good as American polling, but its still a brilliant source of data.

Polling is an art and a science, but very often the problem is that people don't trust the polls enough, including the pollsters, which can lead to herding. Outliers are vital to improving the overall polling averages.


Okay, but what is it they are actually polling then? Popular vote results or election results? Popular vote polling was fine 50 years ago when the two were usually the same. Today, demographic changes means that popular vote and electoral votes will usually be the opposite (except in very rare cases where Dems have a >7% national advantage).

Methodology will need to change to more heavily weight close counties and close states, if it is to correctly predict national election results. IMO

Hong Wu wrote:This all may be true but it doesn't get to the heart of my question, which is why is Fox oversampling their opposition.


Small technical question: How is Fox oversampling the opposition? Exactly how would that work? Sure you could concentrate your polling in certain demographics, but wait for it HW...

...They're already doing that! Every pollster has to pre-weight data to achieve any kind of accurate result. You're alleging they are skewing the pre-weighting process more than they normally would, to favor Dems. Do you have some evidence of this, besides the fact you disagree with the result?
#15041789
This came up before, and the guy that brought it up also posted everything that was available.

The numbers tracked.

IOW, the assertion that the pollsters made about limits was prob accurate. In any case, definitely not 20%, that's just the daily dose of crap from the Right.

When people are addicted to excuses, they are Excuse Junkies. The Junkies needed their fix bad after that Fox poll...

Over this coming week, follow Real Clear Politics and see how the new polls track.

https://www.realclearpolitics.com/epoll ... -6956.html
#15041804
quetzalcoatl wrote:Okay, but what is it they are actually polling then? Popular vote results or election results? Popular vote polling was fine 50 years ago when the two were usually the same. Today, demographic changes means that popular vote and electoral votes will usually be the opposite (except in very rare cases where Dems have a >7% national advantage).

Methodology will need to change to more heavily weight close counties and close states, if it is to correctly predict national election results. IMO

I go to fivethirtyeight. They give percentage predictions for every State and every House and Senate seat. In 2012 they called every State right in the Presidential race. However Nate Silver said afterwards that was luck and they couldn't expect to repeat that. Going in to 2016 I think they had Trump at about a third chance to win the electoral college. The Republicans have a significant inbuilt advantage in the House but in the electoral college the Republican advantage is quite small. On an equal Presidential vote it could easily go to the Democrats, just because the college split for the Republicans in 2016, doesn't mean it will in 2020.
#15041812
Red_Army wrote:I'm not gonna read the dumb OP that I'm sure meanders around something simple that Tucker Carlson already said, but Fox News spends its time bitching about Democrats because the entire ideology of MAGA idiots and the Republican Party (same thing) is hating whatever rich Hollywood liberals are doing. If they didn't have bitterly disagreeing with meaningless shit like drag queens reading books or TV shows being cancelled they wouldn't have an ideology. The puppet show of Fox News watchers vs. Late Night comedy watchers is pathetic and both sides ( ;) ) are neutered idiots, but the former are the only ones I'm concerned about murdering people for their twitter-mania.


Wong Ju has no idea what he's talking about by the looks of it; he refuses to even provide a link, the braindead amoebas of fox news are a problem but there are braindead amoebas on all sides..not just those two, who want to force their quarter-braincell thought-driven, soulless hypocrisy on others when they've got snake eyes, and probably have never fought anything in their lives, with red eyes. :)
#15041846
As I pointed out in the OP, Fox posts their data to Scribd online. You can compare this to many different Web pages that agree on what voter registration rates are.

Honestly the users of this forum seem to get more childish by the day.

Here's a story discussing an analysis by a pollster from Princeton: https://nypost.com/2019/10/12/fox-news- ... -analysis/

The thing is, with no disrespect intended to that guy, you don't need to go to Princeton to read the data and look up voter registration rates.
#15041962
"Why is my blatant propaganda network not telling me the lies I want to hear?" is quite possibly the dumbest, least introspective question ever asked on this forum. Good job, Hong Wu.

But Democrats were oversampled by 8%. If you keep it simple and subtract 8% from 51%, you get 43% of the vote.


Good job, genius. You're a statistician now

Also thanks for continually talking about how the poll is on Scribd and then not linking it. I'm sure even the most cursory glance at it explains their methodology and further highlights how the above quote is fucking moronic, hence why you wouldn't want to link it.

This thread is shit and it's dumb as fuck.
#15041982
Hong Wu wrote:Honestly the users of this forum seem to get more childish by the day.

You can assume some of the new posters are paid political activists. The Democrats are going to lose again and they're hiding the fact that they are pissed off about it so they can get paid to do whatever gaslighting the DNC is paying them to do. Like 2016, it's not going to work. You can take on some lily livered Republicans that way, but not Trump.

Hong Wu wrote:The thing is, with no disrespect intended to that guy, you don't need to go to Princeton to read the data and look up voter registration rates.

Yes, and presidential popular polls do not weigh out the 50 state polls and weight them accordingly, which is why they're always skewed.
#15041991
blackjack21 wrote:You can assume some of the new posters are paid political activists. The Democrats are going to lose again and they're hiding the fact that they are pissed off about it so they can get paid to do whatever gaslighting the DNC is paying them to do. Like 2016, it's not going to work. You can take on some lily livered Republicans that way, but not Trump.


Yes, and presidential popular polls do not weigh out the 50 state polls and weight them accordingly, which is why they're always skewed.


Coming from you who makes monosyllabic replies to threads?

"Trump's actions released ISIS terrorists"

You: "Yawn"

Or people who gloat on about polls then don't even link the stats?

Or people who write about how "no one cares" whether Trump says stupid things.

What planet do you inhabit I wonder? I'm not even a yank and I don't like the Dems, but feel free to keep rabbitting fella..
#15042151
late wrote:Coming from you who makes monosyllabic replies to threads?

The usual complaint around here is that I reply with a wall of text...which should be your first clue that I read every sentence of what people write. In your case, I just gave you the reply the post deserved. A little sincerity and a bit less trolling is always appreciated.
#15042152
You inject liberal amounts of trolling into nearly every post you do. Don't pretend otherwise.

I do sometimes write sarcy replies, but that ISIS thread is completely truthful and backed up with referenced facts. And being angry about ISIS terrorists being released (read the article I linked about 4 militants running free) is fair enough. It's just silly to say that a bit of angry finger pointing is 'trolling' - it's not. And we are right to be angry with Trump.

I witnessed the direct aftermath of two terror attacks in my life, one was perpetrated against a Jewish charity shop in 1994 and the other in 2013, both in London. I'll leave you to go digging if you want. I don't find it "yawn" worthy or worthy of trolling, do you? Terrorists going free and coming to blow up Europe is NOT some trivial joke.
#15042154
https://www.google.com/amp/s/www.politi ... ach-041942

Poll: Half of voters support impeaching and removing Trump
A new POLITICO/Morning Consult poll is the latest survey to show public backing for Congress' intensifying impeachment inquiry.

Half of voters support the impeachment and removal of President Donald Trump, according to a new POLITICO/Morning Consult poll.

The poll, conducted Monday and Tuesday, shows that 50 percent of registered voters surveyed would support the Senate’s removing Trump from office, while 43 percent oppose the president’s removal. Seven percent of voters were undecided.
#15042155
:excited: Jackpot @Wong Ju

As of Oct. 10, an average of impeachment polls calculated by FiveThirtyEight, a polling analysis website, shows 49.3% of respondents supporting impeachment and 43.5% not supporting it.

Here is what the individual polls say about impeachment, separated by each question.



Approve of the House opening an inquiry?

A poll conducted by NPR/Marist/PBS NewsHour from Oct. 3 to Oct. 8 shows 52% of adults support House Democrats starting an impeachment inquiry into Trump — 43% disapprove.

This is an increase from their earlier polling of the same question the night of Sept. 25, one day after Pelosi announced the impeachment inquiry. The poll showed Americans supported the opening of an inquiry at 49% to 46%.

Independents' views also appeared to have shifted in the early October poll: 54% of independents supported the opening of an inquiry, and 41% disapproved. In contrast, 50% disapproved and 44% approved in the Sept. 25 poll.

Polling by The Washington Post/Schar School released on Oct. 8 and conducted from Oct. 1-6 mirrored the finding: 58% of adults said they supported the opening of an inquiry and 38% opposed it.

Voting to impeach?
A USA TODAY/Ipsos poll released on Oct. 3 and conducted from Oct. 1-2, just one week after the beginning of the inquiry, found 45% of Americans approved of a vote in the House to impeach Trump, and 38% disapproved of a vote. Forty-four percent of people in the same poll approved of the Senate voting to convict the president, and 35% did not.

In June, previous polling conducted for USA TODAY by Suffolk University found that only 31.8% of registered voters thought the House "should seriously consider impeaching President Trump," with 60.6% opposing it. That poll had been conducted prior to special counsel Robert Mueller's testimony before the House, and far before the emergence of details of Trump's Ukraine controversy.
.......

Views on Trump and Ukraine?
The NBC/WSJ poll conducted from Oct. 4-6 asks about impeachment by asking adults what actions they think should be taken about Trump's dealings with Ukraine.

Adults were asked whether there was sufficient evidence to impeach and remove Trump, whether to hold an impeachment inquiry, whether there was insufficient evidence to hold an inquiry, or if they were not sure.

Twenty-four percent of adults said there was sufficient evidence to impeach and remove Trump, 31% said Congress should hold an inquiry, and 39% said there was insufficient evidence.

This meant a majority of Americans saw the Ukraine controversy as meriting action on impeachment.

When the NBC/WSJ poll had been conducted in July, adults were asked what Congress should do given discussion of impeachment and impeachment hearings.

At the time, 21% of Americans said there was sufficient evidence to begin impeachment hearings, 27% called for Congress to investigate to see if there was enough evidence for impeachment hearings in the future, and 50% said Congress should not hold impeachment hearings.



https://www.google.com/amp/s/amp.usatod ... 3933117002

:D

Btw even assuming the fox crapola is true, still gotta take this into consideration..



When Fox had asked the same question in July of this year, 42% of voters supported impeachment and removal, 5% supported impeachment but not removal, and 45% opposed impeachment.



Remember 1 in 8 repubs support impeachment even with the fox poll. ;) (the recent one)

@FiveofSwords Edwards' critique does not con[…]

Russia-Ukraine War 2022

70% of Americans view Ukraine as an ally or frien[…]

World War II Day by Day

April 19, Friday Allied troops land on Norway co[…]

My prediction of 100-200K dead is still on track. […]