Another fake Democrat poll exposed

[artichoke]

Capt. Bullshit wastes his time. Numbers polls without complete data are pretty worthless if you're honest with yourself.

Pointing out that polls are biased, unscientific trash is wasting my time...because polls are...unscientific trash? Nice counterpoint there, Socrates.

[Trumpet]

Trump gets the boot in November 2020, and jail in 2021.

[artichoke]

Trump gets the boot in November 2020, and jail in 2021.

Democrat predictions are always wrong.

nyt hillary election.jpg
.
a.jpg
.
spoiler alert hillary wins.jpg
.
alarmist scaremongering.jpg
.
trump economy soars prediction.jpg

[christiefan915]

Another provably fake poll. This time, it's Quinnipac. They had to over-sample registered Democrats by 10% to manufacture their new claim about the child-molesting Alzheimer's patient now having "double-digit lead." They sampled 36% registered Democrats and 26% registered Republicans to make this happen.

This is why you ALWAYS click on the methodology and see how they got their numbers. This is the same kind of outright lie that caused people to think Hillary had a 92% chance of beating President Trump.

Attachment 15384

Biden Holds Double-Digit Lead Over Trump, New Poll Shows

Imagine the headlines conservatives could fabricate by wildly over-sampling registered Republicans like this:

-Majority of Americans want Obama administration officials prosecuted for treason

-Those who want to finish building the wall have a "double-digit lead" over those who do not

-American people overwhelmingly want power-abusing Democrat judges impeached and purged from federal judiciary

Polls?

[Grokmaster]

Nonsense. The Quinnipiac poll has a historical democratic bias of .2% Basically nothing.

Quinipiac's OWN METHODOLGY/DEMOGRAPHICS reveal the BIAS. READ THE LINK, LIAR.


PARTY IDENTIFICATION Republican 26% Democrat 36 Independent 31 Other/DK/NA 7




MAKEUP SOME MORE BULLSHIT.

[BRUTALITOPS]

Another provably fake poll. This time, it's Quinnipac. They had to over-sample registered Democrats by 10% to manufacture their new claim about the child-molesting Alzheimer's patient now having "double-digit lead." They sampled 36% registered Democrats and 26% registered Republicans to make this happen.

This is why you ALWAYS click on the methodology and see how they got their numbers. This is the same kind of outright lie that caused people to think Hillary had a 92% chance of beating President Trump.

Attachment 15384

Biden Holds Double-Digit Lead Over Trump, New Poll Shows

Imagine the headlines conservatives could fabricate by wildly over-sampling registered Republicans like this:

-Majority of Americans want Obama administration officials prosecuted for treason

-Those who want to finish building the wall have a "double-digit lead" over those who do not

-American people overwhelmingly want power-abusing Democrat judges impeached and purged from federal judiciary

This survey uses statistical weighting procedures to account for deviations in the survey sample from known population characteristics, which helps correct for differential survey participation and random variation in samples. The overall adult sample is weighted to recent Census data using a sample balancing procedure to match the demographic makeup of the population by region, gender, age, education and race.When including thedesign effect, the margin of sampling error for this study of registered voters is +/-3.3percentage points.

.

[AProudLefty]

Polls?

Where is Legion to complain about the polls?

[AProudLefty]

[BRUTALITOPS]

it's really not that hard to understand you fucking dummies.

It's called weighting.

Lets say you know the population is 500 democrats, 500 republicans.

You sample 50 people where 30 are democrat and 20 are republican. i.e. 60% democrat vs. 40% republican.

BUT - we know the population data is exactly 50/50, so what do we do if we are not a retard on JPP that doesn't understand basic stats?

We weight the responses accordingly. Thus every democrat answer is worth .8x and every republican answer is worth 1.25x and
This brings the sample into line with the known population

I'm pretty sure I learned this stuff in 4th grade.

[AProudLefty]

it's really not that hard to understand you fucking dummies.

It's called weighting.

Lets say you know the population is 500 democrats, 500 republicans.

You sample 50 people where 30 are democrat and 20 are republican. i.e. 60% democrat vs. 40% republican.

BUT - we know the population data is exactly 50/50, so what do we do if we are not a retard on JPP that doesn't understand basic stats?

We weight the responses accordingly. Thus every democrat answer is worth .8x and every republican answer is worth 1.25x and
This brings the sample into line with the known population

I'm pretty sure I learned this stuff in 4th grade.

Don't expect retards to understand it.

[Concart]

Quinipiac's OWN METHODOLGY/DEMOGRAPHICS reveal the BIAS. READ THE LINK, LIAR.


PARTY IDENTIFICATION Republican 26% Democrat 36 Independent 31 Other/DK/NA 7




MAKEUP SOME MORE BULLSHIT.

LOL, the sample is the sample. The breakdown is very likely correct. The margin of error is very small on the poll. It's like saying they oversampled people that don't like Trump. No, there are just lots of people that don't like Trump.

[Concart]

it's really not that hard to understand you fucking dummies.

It's called weighting.

Lets say you know the population is 500 democrats, 500 republicans.

You sample 50 people where 30 are democrat and 20 are republican. i.e. 60% democrat vs. 40% republican.

BUT - we know the population data is exactly 50/50, so what do we do if we are not a retard on JPP that doesn't understand basic stats?

We weight the responses accordingly. Thus every democrat answer is worth .8x and every republican answer is worth 1.25x and
This brings the sample into line with the known population

I'm pretty sure I learned this stuff in 4th grade.

I would say 'yes and no' to this. For demographics like age, gender, race, etc, there is a relatively fixed baseline. For party affiliation, that baseline changes and often changes dramatically over time. So in cases like this, they pollster may very well go with whatever breakdown the sample indicates. I know there is some kind of algorithm and likely a baseline they are working from as a starting point, but I think that type of demographic is less likely to be weighted to a static number.

[Earl]

Another provably fake poll. This time, it's Quinnipac. They had to over-sample registered Democrats by 10% to manufacture their new claim about the child-molesting Alzheimer's patient now having "double-digit lead." They sampled 36% registered Democrats and 26% registered Republicans to make this happen.

This is why you ALWAYS click on the methodology and see how they got their numbers. This is the same kind of outright lie that caused people to think Hillary had a 92% chance of beating President Trump.

Attachment 15384

Biden Holds Double-Digit Lead Over Trump, New Poll Shows

Imagine the headlines conservatives could fabricate by wildly over-sampling registered Republicans like this:

-Majority of Americans want Obama administration officials prosecuted for treason

-Those who want to finish building the wall have a "double-digit lead" over those who do not

-American people overwhelmingly want power-abusing Democrat judges impeached and purged from federal judiciary

Indeed.

[FreeSpeech]

MSM and their think tanks are direct branches of the DNC ministry of propaganda and need to be regulated as in kind campaign contributors.

[christiefan915]

Where is Legion to complain about the polls?

I suspect he's with us in spirit and "the grapevine" lets him know what's going on here.