new CA cnn poll and some problems in the details

[floppyboo]

The big summary #: Biden 21, Bernie 20
Some questions I have about how accurate this is:
The sample # is just over 500 dem likely voters.
The sample size for >$50K is insufficient for inclusion in the poll.
Bernie has the most individual donors of ALL time, and the majority are service workers, Amazon and Walmart being at the top.
The 99% is a real thing.
How is this poll accurate to 3/4%?

Second:
+/- 65% are most concerned with replacing Trump
+/- 25% are most concerned with policy

Bernie is ahead in the top 4 policy concerns.

If I am reading all this right, a small butterfly wing shift in confidence in Joe leads to an inevitable Sanders victory.
The media until very recently has been ok with the inevitability of a Biden nominee, but that seems a little more shaky of late. Will Bloomberg be the new Joe?

[floppyboo]

1. forgot link to poll.

https://cdn.cnn.com/cnn/2019/images/12/11/rel1_ca.pdf

[booley]

2. Biden seems immune to gaffes right now

Not that he doesn't' make any. He's had a reputation for gaffes his entire career.

Telling Democrats to vote for Trump isn't' a good look. And he's done that at least 3 times this primary to my recollection.

But no real consequences that I can see.

That could change, however. Bidens' main strength is that he is a safe " choice". If it seems trump has too many ways to attack Biden in the general, that could change.

Personally I think the final two in the primaries will be Biden and Sanders.

Anyway, I have learned to distrust anytime the media says someone is "inevitable"

[The Valley Below]

5. Would you be happier if Joe Biden encouraged all the disaffected conservatives to join our party?

Be careful what you wish for.

[booley]

9. How do you mean?

Personally, if a candidate who could get Republicans to embrace Democratic ideals that would be great.

Oh wait, we do have a guy who has been doing that. And it ain't Joe Biden.

[The Valley Below]

15. It ain't the guy who ain't a Democrat who is drawing in disaffected Republicans--

especially working people to vote for him. It's Joe Biden.

And if you wish for more moderates and moderate-conservatives to join the Democratic party, I'd tell you again to be careful what you wish for.

[LanternWaste]

6. I get the same mistrust when someone gaffes that the US created Russia as an enemy

"Anyway, I have learned to distrust..."

I get the same mistrust when someone states the US created Russia as an enemy without providing any objective evidence to support it as such.

Historical revisionism for the sake of party, indeed.

[booley]

10. Hard to say what you mean without context

But I can give you an objective example

The United States expanding NATO into the warsaw pact countries after it had said it would not do that.

The economic Shock Doctrine being led by the US and ultimately led to them electing a strong man probably didn't' help either.

Though again since you didn't provide any context or citations I cant' tell if the above examples apply to what you vaguely refer.

But hey, nice Whataboutsim! Kind of ironic if you think about it.

[Demsrule86]

3. Wishful thinking...haha.

[George II]

4. We have no idea how many contributors Sanders has or what they do for a living....

....he's only released the names of less than 100,000 contributors.

[floppyboo]

14. Third quarter donation break-down from

https://www.commondreams.org/news/2019/10/01/fueled-teachers-and-average-donation-18-sanders-raised-record-253-million-third
Fueled by Teachers and Average Donation of $18, Sanders Raised Record $25.3 Million in Third Quarter

"Bernie is proud to be the only candidate running to defeat Donald Trump who is 100 percent funded by grassroots donations—both in the primary and in the general."

The Sanders campaign said the average third-quarter donation was just $18, "teacher" was the most common profession of donors, and Starbucks, Amazon, and Walmart were the most common employers of donors.

[floppyboo]

16. Second quarter donation breakdown from

https://www.businessinsider.com/walmart-employees-donations-to-bernie-sanders-campaign-2019-7

Sanders' campaign said Tuesday that it raised $18 million during the second fundraising quarter of the year, and that Walmart was the top employer of his donors for the period.

Other leading employers of Sanders' donors included Amazon, Starbucks, Target, and the US Postal Service. More than 99% of donations were $100 or less, according to the campaign.

[George II]

20. Now this is an interesting excerpt:

More than 99% of donations were $100 or less, according to the campaign.

Of course, that just glosses over all those contributors who give dozens or even hundreds of individual contributions all "$100 or less". It gives the aura of a so-called "grass roots campaign", but that's not entirely accurate.

The most extreme is one woman in Michigan who in the first three months of his campaign made 825 individual contributions, more than 800 of them exactly $3.00 each. So, that raises the "%" of small contributions and lowers the "average" contribution. But do you know how much that woman gave in aggregate? $2,821 (actually $21 over the allowable limit)

Now, is that woman a "grass roots contributor" or a "big donor"? What's the difference between a person who gives 825 contributions that total $2800 and a person who gives only one contribution of $2800?

Finally, the Sanders campaign has been bashing Walmart and Amazon, along with the "financial industry" and "pharmaceutical industry". So what's the difference between accepting contributions from "Walmart and Amazon" (apparently good) and the "financial and pharmaceutical industries" (apparently bad)?

[George II]

17. We have no idea how and from where that break down was generated. He's only reported....

....the names, locations, occupations, and employers of less than 100,000 contributors to the FEC. If his campaign has more detail available, why not release it?

[floppyboo]

18. Has anyone asked him?

I'm not a 'member' of his campaign, but maybe on his web site?
Is this normal reporting to the FEC?
What are Biden and Buttigieg's names, locations, occupations and employers?
That seems a bit over the top to give such detail. Maybe more than 100,000 would be considered collecting meta-data? Maybe its illegal?
Interesting questions now occupy my little brain.

[George II]

22. One would think that if a candidate is boasting about a set of contributions that candidate....

....would be willing or even eager to release the details of those contributions.

I don't know what the Federal level is for which the details must be reported to the FEC, I think it's $100. But a campaign should maintain accurate records of all contributors regardless of the amount given.

For example, let's say a person gives $50 and the campaign doesn't record the name, address, occupation, and employer. Then later in the year the same person gives another $50, putting that person at the level that requires full details to be reported to the FEC. But if the details of the first $50 aren't kept, how does the campaign know that the person reached the reporting threshold?

However, going through the FEC filings, details of some "small" contributors HAVE been reported. Its curious why some small contributions are reported in detail and others aren't.

[NYMinute]

27. Do contributions in rubles need to be tabulated? nt

[TexasTowelie]

7. The margin of error formula is

near the end of this link.

https://www.statisticshowto.datasciencecentral.com/probability-and-statistics/hypothesis-testing/margin-of-error/

It takes into account what the confidence level is, the size of the sample, and the proportion of the sample. Some background in statistics helps to understand the formula. If the for the people earning over $50K is under-represented then it probably means that Biden's numbers in the poll are likely less than what the entire population will vote for.

I don't see any good news in the CNN poll for Bernie. It appears that most of the O'Rourke voters in Texas favor Biden now and it is at least possible that Biden will sweep all of the Texas delegates. If that occurs, then I expect that the California won't be important because the candidates will be split almost evenly. 38% of the delegates will be known by Super Tuesday.

There is some good information at this link:

https://www.270towin.com/2020-democratic-nomination/

[floppyboo]

11. that was under $50K

not sure what you mean that 'over $50K is under-represented then it probably means that Biden's numbers in the poll are likely less'.

Is there a CNN poll today for Texas? Missed it.

I only have under-grad stats, so any more learning is appreciated!

edit: From your link, if there is no valid statistical data (in this case for less than $50K) how can you calculate a margin of error?

The margin of error I mentioned in the OP was for the poll in general, as cited by CNN.

[TexasTowelie]

24. Okay, my comment was based upon your OP where you indicated

the sample size for those over 50K rather than under 50K so please disregard.

The margin of errors mentioned on these polls are related to the normal distribution curve (aka bell-curve). Since calculating the area under the normal distribution curve involves some fairly high level calculus, statisticians use z scores along with the mean (average) and the standard deviation to determine confidence levels (usually at the 90%, 95%, and 99% levels). There are also other considerations as to whether it is a one-sided or two-sided tails tests (is the level of uncertainty on one side of the mean or on both sides). For polls, the consideration is normally a two-sided tail, but other mathematical applications it may be appropriate to use a one-sided poll.

About seven years ago, a college buddy who studied computer science and calculus, but had no statistical background was trying to work on a marketing project for his employer. He kept relaying messages to his employer mentioning about the project, but he didn't really understand anything about "confidence levels". The people that he was pitching the project to wanted to see a lot of statistical data so he turned to me and I prepared a short document that condensed a couple hundred pages from a textbook into a two page document.

I was able to locate the document as an attachment that I sent in my email and would be happy to send it to you (for free) if you provide me an email address in a personal message on DU. I'll warn you that it is "dry" reading, touches upon theory, and the Greek mathematical symbols for the mean and standard deviation. I'm not going to guarantee that it will help and I've become a bit rusty with that area of mathematics so I don't know if I could answer a lot of questions about it.

However, even if there is missing data for a selected subgroup of a data sample, in theory that is accounted as an "uncertainty" in the margin of error and confidence level calculations.

FWIW, my college buddy never paid me and I didn't want to let it a relatively small amount of money destroy our friendship. His project eventually floundered and died like many other projects that I have experience with in the computer science and theoretical mathematics fields. That's why I've developed healthy skepticism when I see startup companies delve into things where the personnel are winging it.

[Mr.Bill]

8. Another factor in the California primary

is the Democratic party is one of three parties who allow independent voters to receive a ballot with the Democratic presidential candidates on it. I'm thinking these voters may be more likely to be Sanders voters than Biden voters. And the poll was of likely Democratic voters, not independents. So this poll may be missing a piece of the puzzle.

This coupled with the California primary being held in March instead of June as it has been in the past should make for a larger turnout than usual. When out primary was in June, in past elections things were often already decided by then.

Should be a very exciting and more dynamic than usual primary.

[TidalWave46]

12. Incorrect.

"If I am reading all this right, a small butterfly wing shift in confidence in Joe leads to an inevitable Sanders victory."

That is not how polling works. That is not how the delegate count works. That isn't how any of it works.

People who try to unskew polls should learn about statistics and the process the participants of the poll are taking place in.

[floppyboo]

13. You are correct to criticize my wording. Let me try again:

...shift in confidence could lead to Sanders winning well more than 25% and the majority of delegates for the California primary.

As far as not understanding, I beg to differ.

[LongtimeAZDem]

19. Bernie'e endorsement of Uygur is going to cost him big in California

LA County Democratic Party Says Cenk Uygur ‘Does Not Belong in Congress’

[boomer_wv]

21. This is silly.

They weigh samples to represent what they expect the voter profile to be.

The only logic that Sanders supporters seem to apply is that if Sanders isnt leading the poll then there must be some diabolical methodology at play, designed specifically to tamper his support. It is the same thing every time a new poll is released.

It's a 1 point topline difference. You can just argue that Sanders and Biden have about the same chance to win CA based on the poll, which is true.

[floppyboo]

23. what's silly is this has nothing to do with what I was referring to

I wasn't referring to the margin of error - just added it in because some people seem to think its important.
What I was referring to is more of the fragility of Biden as I see it from the numbers in this poll.
And, yes, I did point out that the sample size of people making under $50K, out of a total # of some 500+, was too small to come up with a number, which is problematic.
Not diabolical.

[boomer_wv]

25. .

It is exactly what you are trying to get at, even if you are being coy about it. You want to say that they didn't sample enough people of x demographic and because of that, Sanders is actually better then the poll shows.

California isn't exactly a Biden stronghold. This is a state where somebody like Warren or Sanders should be able to make a lot of hay. How can you argue that Biden if fragile when he's leading against them on their turf? Even if a higher number of sub-50kers turn out, that's not going to have a negative impact on Biden. He performs really well with that group, especially when you consider that his base is minorities and non-college educated whites, which you will find a lot of in that group.

Diving into the crosstabs of a poll is usually a fools errand anyway. Out of 500 people, if you get, say 15% in a given demo, that's only n=75. That means that the MoEs for the sub-samples are going to be outrageous compared to what you want for a full sample. You aren't going to get much from those other than generalities and likely a lot of noise. We can't speak about trends within groups without a lot of polling to pull from.

[NYMinute]

26. bwaaaaahahahahahahahahahahahahaha nt