So lets do some History ?


#882

fuck that noise. i wasn’t bowled over or impressed by anyone…excpet that lady that teared up. she made me tear up a bit.

fuck rand paul too.


#883

I was surprised by Leoffler.

But she may have wanted to get home quick and get a jump on boxing up her Amazon returns of all the flannel shirts and jeans she bought to look folksy during her campaign.


#884

All I want is for Trump to end like Saddam Hussein. Hung in some nondescript industrial building, while shouting something about making the country great again. That’s not too much to ask, is it?


#885

After a trial, of course.


#886

Nearly half of Republicans support the invasion of the US Capitol


#887

Will we ever stop paying for the sins of Ronald Reagan?


#888

Do they have a % that were responding? Who did they ask? How many people? Are they office-holders or just people from the street?

I am calling BS on that chart.


#889

#890

His administration started and ended with “alternative facts.”

I made the mistake of taking in dive into Ashley (the airforce vet shot) Twitter feed.
https://twitter.com/Ashli_Babbitt

There’s also an Instagram called homegrownterrorists, which was DOXing nearly every MAGAT but I think the FBI took it down.


#891

With 328 million people in the US, that poll got 1,397 responses. Not a chance that it is anywhere near accurate, since they asked .00000425914634146341 of the population.


#892

You’re joking right? It’s a random sample of 1,397 responses. It has a margin of error of +/-3.3% that is completely independent of the size of the population. You can prove mathematically that the margin of error of a random sample has nothing to do with how large the population is. It would be the same whether it’s 1,397 out of 3,000 or 30,000 or 328 million people. That’s the great thing about inferential statistics.


#893

Anyone else getting a little turned on?


#894

Yes, where is @Nebulance???


#895

Here’s some porn for you, @Ely_Plains:


#896

Sadly, it’s not panda porn.


#897


#898

Jesus


#899

Think back–how many points ahead was Clinton in 2016? Biden this year? How did that turn out?

We’re depending on yougov to be reputable. Are they? Were they accurate in 2016 and 2020 elections? Asking because I don’t know, but in most jobs if you had such world-changing, enormous gaps in your analysis you’d be bumped down to selling fries at Burger King, but we seem to be giving these folks a pass they likely haven’t earned.


#900

Absolutely, we depend yougov to be reputable. Fivethirtyeight gives their methodology a B, for what it’s worth.

Assessing past performance in presidential polling is important, but it’s also misleading, because this is a poll that is significantly less complex than election forecasting. I also completely agree that such things as question wording need to be considered separately from the margin of sampling error, which doesn’t take other errors (such as whether the question may be formulated in a misleading way) into account.

The tricky thing about the presidential election is that national polls have performed pretty well, in most cases within the margin of error. The problem is that polling at the electoral college level is much harder. There is a good report by AAPOR, the national association of survey researchers, on the performance of polls in the 2016 election: https://www.aapor.org/Education-Resources/Reports/An-Evaluation-of-2016-Election-Polls-in-the-U-S.aspx


#901

That is well-considered and well-written, so don’t take my response as being directed toward you.

It is admittedly my own problem. Despite my use of casual language and folksy examples, I understand the extrapolation of small sample sizes to address larger groups, but I look at it this way.

They had one job. And failed miserably. If I hire a guy to build a deck and my uncles wheelchair always rolls off toward the stairs on it, I don’t hire that guy again. If I am at a party and he is discussing his building projects, I rightly assume he is full of shit and should shut up.

One job. Any polling agency that ‘got it wrong’ by such a huge amount doesn’t deserve to remain in the game.

Again, my problem for having strict standards and expectations of professionals. Perhaps I missed the part of their website that laid out how they held an independent accounting of their practices and what actions they are taking to address them and improve.

Bring on the panda porn!