Polling Perils

[g5000]

There may be a bit of truth in that in the US, but not much. For example, when asked in telephone polls, 40% of Americans have generally identified themselves as conservatives. However, in exit polls, conservatives are around 35% of the electorate. This means that, if true, only 10%-15% of conservatives, and 1 in 20 Americans, fibbed about being a conservative in the exit poll.

Nope. It could mean it is easier for a conservative to answer a telephone poll than to go and vote.

 

[The Rabbi]

It's like editors of major publications read my posts here and write about them.
Conservative Voters Give Pollsters Politically Correct Answers . . . and Then They Vote National Review Online

 

[Statistikhengst]

It's like editors of major publications read my posts here and write about them.
Conservative Voters Give Pollsters Politically Correct Answers . . . and Then They Vote National Review Online

National Review Online = Extreme RWNJ junk.

 

[g5000]

The NRO wants to believe there are more conservatives than there are, and so their article headline reflects that bias even though all of reasons cited for why polls are wrong has almost nothing to do with people "fibbing" or giving "politically correct" responses.

Here is the crux:

Perhaps it’s just been a run of bad luck. But there are lots of reasons to worry about the state of the polling industry. Voters are becoming harder to contact, especially on landline telephones. Online polls have become commonplace, but some eschew probability sampling, historically the bedrock of polling methodology. And in the U.S., some pollsters have been caught withholding results when they differ from other surveys, “herding” toward a false consensus about a race instead of behaving independently. There may be more difficult times ahead for the polling industry.

"Eschewing probability sampling". Yeah. We see that shit here on this forum all the time. Online polls are completely useless as predictive models.

 

[Statistikhengst]

The NRO wants to believe there are more conservatives than there are, and so their article headline reflects that bias even though all of reasons cited for why polls are wrong has almost nothing to do with people "fibbing" or giving "politically correct" responses.

Here is the crux:

Perhaps it’s just been a run of bad luck. But there are lots of reasons to worry about the state of the polling industry. Voters are becoming harder to contact, especially on landline telephones. Online polls have become commonplace, but some eschew probability sampling, historically the bedrock of polling methodology. And in the U.S., some pollsters have been caught withholding results when they differ from other surveys, “herding” toward a false consensus about a race instead of behaving independently. There may be more difficult times ahead for the polling industry.

"Eschewing probability sampling". Yeah. We see that shit here on this forum all the time. Online polls are completely useless as predictive models.

Online polls that are conducted unscientifically are certainly totally useless.

But the day will come when 100% of the USA is online and with the right kind of safeguards in place, polling over the net will replace telephone polling completely in the next 20-40 years. Zogby has tried it, with some really fucked-up results. YouGov has a slightly better track record. Again, the online polling firm that I saw as truly groundbreaking in 2012 was the Rand Corporation Poll. It was truly innovative.

 

[LoneLaugher]

It's like editors of major publications read my posts here and write about them.
Conservative Voters Give Pollsters Politically Correct Answers . . . and Then They Vote National Review Online

I think it is the other way around. Loser.

 

[Toro]

It's like editors of major publications read my posts here and write about them.
Conservative Voters Give Pollsters Politically Correct Answers . . . and Then They Vote National Review Online

That's an interesting article, but it doesn't really support the notion that conservatives consistently say one thing in polling then do another. What it did was point out a few instances where that was true, primarily in the UK. And even then, it noted three elections where exit polls underestimated Conservative support during a period when there have been 10 elections.

 

[Toro]

Results from the British election show one clear loser: POllsters. Every pollster virtually had the parties in a dead heat. But in the end it was a slaughter for the Torys. Even Nate Silver, favorite of the libs here, was way off, offering an explanation
What We Got Wrong In Our 2015 U.K. General Election Model FiveThirtyEight

But the truth, which no oe has brought out, is that people tend to lie to pollsters because to admit being for the conservatives is to be perceived as "mean." The same thing probably operates here: people claim to like Hillary or whatever Dem but then vote GOP. This will play out on the coming election for sure.

Guy, 60% of Brits voted against the Tories. That they couldn't agree on an oppossition doesn't mean anyone thinks Cameron has done a good job.

The scary thing is that Scotland has pretty much decided they've given up on the United Kingdom. So Conservatives will have accomplished something- the End of the Country.

Essentially, yes.

With this kind of unheard of hypermajority in Scotland, it is only a matter of time before secession comes onto the table.

It's not that clear. Canada had a similar situation in that for nearly 20 years, the party that won the most seats in Quebec during federal elections was a separatist party. And during that time, there was a referendum on sovereignty in Quebec and the separatists lost. Now it's considered to be a dormant issue

 

[Derideo_Te]

Results from the British election show one clear loser: POllsters. Every pollster virtually had the parties in a dead heat. But in the end it was a slaughter for the Torys. Even Nate Silver, favorite of the libs here, was way off, offering an explanation
What We Got Wrong In Our 2015 U.K. General Election Model FiveThirtyEight

But the truth, which no one has brought out, is that people tend to lie to pollsters because to admit being for the conservatives is to be perceived as "mean." The same thing probably operates here: people claim to like Hillary or whatever Dem but then vote GOP. This will play out on the coming election for sure.

How do you square that with the incredible accuracy the polls have for predicting the elections here? They are very accurate right before the election occurs.

Because in this country we compare pollsters to determine their accuracy.

For instance Rasmussen has a decided rightwing bias all through the elections and then in the last 2 weeks before the election they remove their bias so that their final polls are "accurate".

Essentially Rasmussen tries to sway the electorate by "predicting" the outcome they want instead of what is actually being reported to them.

 

[Statistikhengst]

Results from the British election show one clear loser: POllsters. Every pollster virtually had the parties in a dead heat. But in the end it was a slaughter for the Torys. Even Nate Silver, favorite of the libs here, was way off, offering an explanation
What We Got Wrong In Our 2015 U.K. General Election Model FiveThirtyEight

But the truth, which no one has brought out, is that people tend to lie to pollsters because to admit being for the conservatives is to be perceived as "mean." The same thing probably operates here: people claim to like Hillary or whatever Dem but then vote GOP. This will play out on the coming election for sure.

How do you square that with the incredible accuracy the polls have for predicting the elections here? They are very accurate right before the election occurs.

Because in this country we compare pollsters to determine their accuracy.

For instance Rasmussen has a decided rightwing bias all through the elections and then in the last 2 weeks before the election they remove their bias so that their final polls are "accurate".

Essentially Rasmussen tries to sway the electorate by "predicting" the outcome they want instead of what is actually being reported to them.

This, however, did not keep Rasmussen from missing the call in SIX of twelve battleground states in 2012. That's a pathetic track record.

Gesendet von meinem GT-I9515 mit Tapatalk

 

[FA_Q2]

They are very accurate right before the election occurs.

are you talking hours, minutes or seconds ?

Days to weeks.
And the facts have been posted all over this thread. Are you refuting that it is the truth?

 

[FA_Q2]

Dick Morrison?

You realize there are a lot of pollsters out there and you stated that 'most' of them were predicting a Romney win. I distinctly remember the polls in general predicting an Obama win. What Dick Morrison predicted is irrelevant.

Dick Morris isn't a "pollster", either. He's a pundit.

Correct but I am not going to get into that with Rabbi – such would be a pointless endeavor.

 

[Clementine]

Results from the British election show one clear loser: POllsters. Every pollster virtually had the parties in a dead heat. But in the end it was a slaughter for the Torys. Even Nate Silver, favorite of the libs here, was way off, offering an explanation
What We Got Wrong In Our 2015 U.K. General Election Model FiveThirtyEight

But the truth, which no oe has brought out, is that people tend to lie to pollsters because to admit being for the conservatives is to be perceived as "mean." The same thing probably operates here: people claim to like Hillary or whatever Dem but then vote GOP. This will play out on the coming election for sure.

All depends on the integrity of the pollsters. Some seek to measure public opinion while others seek to shape it. I've seen a lot of MSNBC polls that show Dems are favored, but the truth is that they are hoping all the people who haven't kept up on things will think a majority of people are voting a certain way and will follow suit.