Trump drops below 40% in an NBC poll

[laylalex 7,282]

13 minutes ago, Nature Boy 2.0 said:

Hmm calculus is the last thing that would make me question his sanity😄

You wound me.   

[laylalex 7,282]

6 minutes ago, Nature Boy 2.0 said:

everyone thanks to google is an expert in representative statistics,  virology,  micro bology, infectious disease,  pathogen vectoring, 

Electro mechanical engineering,  fermentation,  constitutional law, criminal justice, sociology, etc.

 

I mean other than Dr. Roy Hinkley I have never been exposed to such broad expertise in so many fields. 

I never said I was an expert in polling. But the IBD poll says it has a representative sample of X number of Ds and Y number of Rs and Z number of independents. So they must have asked the respondents what their party affiliation was, or had access to voter information that stated their registration status. The pollsters speak with more people than the actual number of responses that are reflected in the results, in order to weight representation accordingly. It isn't that complicated.

[Burnt Reynolds 5,058]

6 minutes ago, Nature Boy 2.0 said:

I mean other than Dr. Roy Hinkley I have never been exposed to such broad expertise in so many fields. 

 

22 minutes ago, Nature Boy 2.0 said:

polled

[Nature Boy 2.0 8,367]

5 minutes ago, laylalex said:

I never said I was an expert in polling. But the IBD poll says it has a representative sample of X number of Ds and Y number of Rs and Z number of independents. So they must have asked the respondents what their party affiliation was, or had access to voter information that stated their registration status. The pollsters speak with more people than the actual number of responses that are reflected in the results, in order to weight representation accordingly. It isn't that complicated.

sweet pea I wasnt really talking to you.

[Dashinka 24,110]

2 hours ago, laylalex said:

Well, there are actually more registered Democrats than Republicans in this country, so they should be polling more Democrats proportionately.

9% more?

 

Nevermind, I found my own answer.  So why poll 3% more Democrats?

 

As of May 2020, Gallup polling found that 31% of Americans identified as Democrats, 25% identified as Republican, and 40% as Independent.

 

https://www.google.com/search?q=registered+voters+by+party&rlz=1C1GCEV_enUS884US888&oq=registered+voters&aqs=chrome.3.69i57j0l7.8050j0j15&sourceid=chrome&ie=UTF-8

Edited October 5, 2020 by Dashinka

[Voice of Reason 10,298]

1 hour ago, laylalex said:

I did calculus but I have about zero recollection of it. Al does calculus for fun, which makes me question his sanity. 

There are events looming in his near future that cause me to question his sanity. but maff isn't one of them. 

[Voice of Reason 10,298]

2 hours ago, Burnt Reynolds said:

How would that work exactly?

Democrats are more populous, smarter, and richer than republicans (or so they have claimed for some time now).  It's only fair that they get more votes in the polls, right?  (I mean, they will probably get more votes in the election, so it only seems fair to me)

Edited October 5, 2020 by Voice of Reason

[Steeleballz 7,550]

20 minutes ago, Dashinka said:

9% more?

 

Nevermind, I found my own answer.  So why poll 3% more Democrats?

 

As of May 2020, Gallup polling found that 31% of Americans identified as Democrats, 25% identified as Republican, and 40% as Independent.

 

https://www.google.com/search?q=registered+voters+by+party&rlz=1C1GCEV_enUS884US888&oq=registered+voters&aqs=chrome.3.69i57j0l7.8050j0j15&sourceid=chrome&ie=UTF-8

 

   Wouldn't that be what happens with a random sampling and normal distribution, given the numbers you provided?

  

[Voice of Reason 10,298]

4 minutes ago, Steeleballz said:

 

   Wouldn't that be what happens with a random sampling and normal distribution, given the numbers you provided?

  

Completely dependent on the sample area, I believe.  

[Burnt Reynolds 5,058]

59 minutes ago, Steeleballz said:

 

   Wouldn't that be what happens with a random sampling and normal distribution, given the numbers you provided?

  

In other words.. sampling more.. that's where I was getting at with laylalex but got busy. Error is reduced and the need to play with data to try and unskew is rendered unnecessary by simply sampling more. I liken fitness in this regard to trying to use paint to resize a 10x10 image into 1000x1000. Yeah you can say it covers the area and its parts but the way it details them (in this case, accuracy in pinpointing how they vote) means a fat lot without environmental controls. 

 

VOR you describe strata which is relevance and one of my key problems (when assuming good faith) with national polling, besides other (and even larger) issues such as response rate. The plummeting response rate plays into the more bad faith polling efforts and use. 

[TBoneTX 44,672]

10 hours ago, laylalex said:

calculus for fun

And for profit, when the hygienist scrapes it from teeth.

[TBoneTX 44,672]

11 hours ago, laylalex said:

IBD poll

The "Irritable Bowel Disease" poll?

[TBoneTX 44,672]

10 hours ago, CanAm1980 said:

Math, what a concept.

       A mathematician, a biologist, and a physicist are sitting in a street café, watching people going into and coming out of the house on the other side of the street.

       First, they see two people going into the house.  Time passes.  After a while, they notice three people coming out of the house.

       The physicist:  "The measurement wasn't accurate."

       The biologist:  "They have reproduced."

       The mathematician:  "If now exactly one person enters the house, then it will be empty again."

[Dashinka 24,110]

12 hours ago, Steeleballz said:

 

   Wouldn't that be what happens with a random sampling and normal distribution, given the numbers you provided?

  

Is it random sampling?  I would think the folks paying for a poll, especially a national poll regarding a presidential election which is meaningless, might have ways to skew the sampling.

[Crtcl Rice Theory 2,432]

3 hours ago, Dashinka said:

Is it random sampling?  I would think the folks paying for a poll, especially a national poll regarding a presidential election which is meaningless, might have ways to skew the sampling.

"Skewing' the sample can be achieved many ways but that doesn't make the results wrong. They are just numbers.

 

The most useful thing about polls us to measure shift's in public opinion which you can do by using the same methods over a period of time. One would expect a lot if volatility right after that debate.

 

The polls are not meaningless, just another tool.