The pollsters are starting to cover their asses

SpeedRacerXXX's Avatar
Speed,
Over 156 mill voters expected for for
Nov 2020
1 mill is an easy call for worthlessness. Especially when the questions are loaded.
Next time use the professional version of sampling explanation, not the high school version Originally Posted by Unique_Carpenter
Polls cost money. Lots of it. If you can contact 1 or 2 thousand likely voters and have the results be accurate within 4% at the 95% confidence level, why contact more?

In 2016 the polls got the national vote almost perfect -- 3% versus the actual results of 2.1%. So would polling a million people have gotten more accurate results at an astrnomical cost?

Unfortunately for the pollsters, when the vote was broken down into states, historical voting patterns, on which polling is based, changed. Turnout for white men was higher than normal. and turnout for black women was lower than normal. Maybe those inaccuracies were addressed for 2020. Maybe not. We'll know in about 3 1/2 weeks.

If you can find anything on sampling theory that contradicts anything I've said, please post it.
Unique_Carpenter's Avatar
Speed, you're missing a key point where most political polls have loaded questions. Thus, easily disregarded unless a very large sample.
There's huge differences between a "balanced" sample vs the stuff fed to newsies that's worthless.
SpeedRacerXXX's Avatar
Speed, you're missing a key point where most political polls have loaded questions. Thus, easily disregarded unless a very large sample.
There's huge differences between a "balanced" sample vs the stuff fed to newsies that's worthless. Originally Posted by Unique_Carpenter
No they do not. That is an excuse that is used when polls do not come out in favor of one's candidate or POV.

If you go to FiveThirtyEight and look at the polls and click on the poll, in most cases the questions asked and the margin of error are both presented. Here is a FOX News poll as an example. Read it over and let me know the question bias in it. I did not cherry-pick the survey.
https://static.foxnews.com/foxnews.c...-7-Release.pdf
texassapper's Avatar
No they do not. That is an excuse that is used when polls do not come out in favor of one's candidate or POV.

If you go to FiveThirtyEight and look at the polls and click on the poll, in most cases the questions asked and the margin of error are both presented. Here is a FOX News poll as an example. Read it over and let me know the question bias in it. I did not cherry-pick the survey.
https://static.foxnews.com/foxnews.c...-7-Release.pdf Originally Posted by SpeedRacerXXX
You really want to go down that road???? Okay.

This is the first one I grabbed.

Some people say they are less likely to vote because they are concerned it might expose
them to coronavirus. Others say they intend to vote no matter what.
How about you --
how likely are you to vote in the presidential election on a scale from zero to 10, where
10 means you’re certain you will vote and zero means there is no chance you will vote.
The phrasing of this question is all about ChiComCOVID. The question is posed to remind the voter about COVID before they make their decision.

Why not just ask:
how likely are you to vote in the presidential election on a scale from zero to 10, where
10 means you’re certain you will vote and zero means there is no chance you will vote?
Because the goal isn't to ask you if you'll vote... it's meant to remind you about ChiComCOVID first... and that influences the answer.


And this:
Supreme Court Justice Ruth Bader Ginsburg’s recent death has sparked a debate over
how to fill the vacancy on the nation’s highest court. Taking into consideration that it is
an election year,
which of the following is closer to your view?
I'm sorry...what does the election year have to do with nominating and seating a justice? Does the Constitution mention this at all? And sparked a debate? no it sparked complaints from Democrats who lost the Senate and now want to cry that its not fair... "Elections have Consequences, bitches" Obama. The setup of this question is entirely subtle framing meant to elicit a response that is hesitant to make a quick decision... especially if they are uncertain of the impact.

A better phrasing would be:
Should the President nominate and the Senate proceed with the confirmation process?

Polling is as much psychology as it is anything. Anyone with a background in science knows how difficult it is to construct unbiased polling even for STRICTLY science stuff..

this crap is marketing voodoo bullshit. It worked for a time when the electorate was tied to 3 channels and a landline.

If it were my Republic, I'd outlaw polling, make the election on a Saturday, Midnight to midnight, finger dipped in ink. in person only with DL or passport. Its YOUR DUTY to vote, and that might involve some inconvenience.... but if COVID has shown us anything it's that we can all bear a little inconvenience in the pursuit of the common good right? So it's not too much for you to get a photo ID and show up with it. The only exception is govt. duty where you can vote at the embassy, consulate, or military base etc.
Why_Yes_I_Do's Avatar
...Polling is as much psychology as it is anything. Anyone with a background in science knows how difficult it is to construct unbiased polling even for STRICTLY science stuff..

this crap is marketing voodoo bullshit... Originally Posted by texassapper

Great observations. I think it is more about conditioning, influencing, leading, shaping, steering etc. than just basic marketing though. Either way, many polls ensure they inject the fear-porn, before you ask the question.



How many days before the election will you stop beating your wife?
Do you expect to beat your wife more before or after the election?
Do you have less fear in life after Muh Carona killed people?
Since Muh Carona has impacted the world and killed many, do you believe it was intentionally released against the US, as Donald Trump claims?
rexdutchman's Avatar
Yupper , catastrophizing everything