CNN Poll Showing Biden In The Lead Sampled Statistically Insignificant Number Of Voters Under 50

[uawchild]

Cenk Uygur, John Iadarola, and Maytha Alhassen, hosts of "The Young Turks," discuss a CNN poll conducted by research organization SSRS that showed Joe Biden with a resounding lead against Sen. Bernie Sanders in the Democratic presidential primary. However, it turns out that the poll had too few respondents in both the 18-34 and 35-49 age cohorts to be deemed statistically significant, and therefore polled mostly voters over 50.

Just days after announcing his candidacy, Joe Biden apparently got a big bump in the polls -- or did he? Because the headlines all seem pretty clear that he shot up, but when you look closer it's not as clear as the headlines make it seem.

On Tuesday there was a CNN poll showing Biden at 39, Sanders at 15, and Warren at 8. That is not only a big lead for Biden over Sanders, it is also a big jump for Biden, who had generally been in the high 20s, occasionally 30-31.

That's a big jump, but I want to show you the actual demographic breakdown in the poll. The text is very small and there's a lot of stuff there, but what you will notice immediately is there's a lot of N/As. And the N/As are in the category of people between the ages of 18-34 and 35-49. Now, a lot of people noticed that. That does not mean that they didn't have any respondents in those ages, they did. It just means that there weren't enough of them to consider it to be a representative sample for opinions held by those people, there just weren't enough of them. Overall it means that the results are going to be a good bit biased because what you are really polling is not people, it is the oldest age groups in America.

https://www.realclearpolitics.com/video/2019/05/05/tyts_john_iadarola_cnn_poll_showing_biden_in_the_lead_sample_statistically_insignificant_number_of_voters_under_50.html

I had no idea this was true. Makes me want to take a closer look at all the polls.

[NanceGreggs]

1. You lost me at ...

... Cenk Uygur.

[BannonsLiver]

2. It's basically an attempt at "unskewing" the poll.

[still_one]

4. Yup, you noticed that to. Gee, perhaps the OP thinks that Biden will drop out of the race because

Uygur thinks so. Maybe we can get a quote from Glenn Greenwald while we are at it.

This is just pure flamebait

[NurseJackie]

16. Has HA Goodman checked in on this?

[still_one]

19. +++++++++++++

[NYMinute]

28. ...

[KitSileya]

34. Or Walter Bragman?

He of the "A Liberal Case for Trump" infamy?

[True Blue American]

5. Same here!

Cenk will do anything to appear relevant.

[MrsCoffee]

23. +1

[elleng]

3. Too early to poll anyway.

[NurseJackie]

17. How so? Who says? Why is it "too early" to poll?

How so? Who says? Why is it "too early" to poll?

Personally, I believe that ANY time is the perfect time to poll if someone wants to take a snapshot of how things are at that moment, and whether someone's campaign is surging... or whether someone else's campaign is cratering.

What's wrong with that? How does that hurt anyone?

[MrsCoffee]

22. Because Bernie isn't in front.

[NurseJackie]

25. I believe you've nailed it. That's what I was thinking too.

[qazplm135]

30. eh, it is pretty early for polling

you get a general sense of who has strength and who doesn't...but if we believed polling this far out, Clinton beats Obama fairly handily in 2008 and Jeb Bush is the republican nominee in 2016.

Lotta time for stuff to happen.

[elleng]

32. Thanks for the practical response, qazplm135.

[elleng]

33. Baloney

[Cha]

36. I don't think so.

We can see where the candidates are now. And, of course it will change.. telling us how far they've come or not.

[InAbLuEsTaTe]

6. Hmmmmmm... interesting!!!

Bernie & Elizabeth 2020!!!
Welcome to the revolution!!!

[MrsCoffee]

21. Roflmfao.

[NYMinute]

29. You need a gif to go with that nt

[watoos]

7. I ignore all polls whether good or bad,

they are flat out propaganda.

No way Trump's favorability went up to 46%. I laughed my ass off at that poll.

[TwilightZone]

14. 46% is within the MOE of several other polls.

It's also within the MOE of his long-term average.

He's also at 46% in three other polls.

https://www.realclearpolitics.com/epolls/other/president_trump_job_approval-6179.html

"they are flat out propaganda"

No, they're not. Some have inherent biases, but polls like Quinnipiac have a proven record of reliability. They have him at 41%, which is about his long-term average. None of the results are really a surprise, since they're all within a few points of his long-term average.

[Hortensis]

8. You STILL have "no idea this is true." I couldn't find any analysis

Last edited Tue May 7, 2019, 05:49 AM - Edit history (1)

that critiqued this poll. I wish I had. But what you posted is just a statement from a partisan group published without analysis or comment by RCP.

The Hill points out that one reason for the high 39% number on that poll is in part due to Biden's currently strong, 50%, support by nonwhites. Biden polled 29% among white Democrats, and presumably all of that 29% will actually vote for a Democrat in both the primary and GE, unlike that gathered to Sanders in his 14%.

Btw, that last? Poll analyses should identify the character and quality of support. Biden's black support is believed sincere but fluid as candidates like Booker and Harris gain in appeal. Otoh, much of Sanders' 14% is believed ultracommitted to him alone, but what actual percentage? I.e., how much of it would shift to Democratic candidates as his candidacy failed?

[lapucelle]

9. Cenk Uygur and Tim Hains...

Donuts for everyone!

[ucrdem]

10. Cenk has a certificate

[NYMinute]

11. Cenk Uygur?

[dsc]

12. I didn't watch the entire video

but I think I get the gist of their argument. I teach math for a living and have taught AP stats (though admittedly not in about a decade). There are two issues they are bringing up. One is that there aren't enough young people polled relative to the percent they are of the voting population (they polled only septegenarions) and two is that the young people they polled are too small a sample to be representative. To evaluate the first we would need to know the relative numbers who vote and how many were polled. Clearly they polled fewer young people but we don't know how many fewer they polled nor do we know how many fewer of them vote. The reason they put the N/A in those columns is that they couldn't find small enough confidence intervals given the size of those samples. That doesn't mean that the number of people in that sample wasn't the proper number given the percent who vote.

As to the representativeness of the sample, they might have a point or they might not. We do know that the confidence interval would have been large (though not how large) and that due to that they didn't produce one nor list percentages. That is all we know. It should be noted though that provided the percent of young people in the sample was correct as to the percent of the voting population, then even if this were an unrepresentative sample it wouldn't have had a huge effect (since it was small). In short, the first issue is the bigger problem and they do nothing at all to tell us if it exists or doesn't.

[NYMinute]

27. Very good post

I may add that modern software like PASS17 strives to make a sample valid by eliminating both type 1 and type 2 errors. If the sample is not representative a warning appears that the MOE is higher than +-5% and which means the observations will be beyond mean +/- 2 standard deviations.

The second point is - who cares? The poll's findings have been replicated by TWO other polls and that is statistically better.

Cenk is panicking. The polls have been devastating for some candidates' supporters, particularly one who was expecting to waltz into the nomination.

[TwilightZone]

13. There are four other polls showing him with similar leads.

https://www.realclearpolitics.com/epolls/2020/president/us/2020_democratic_presidential_nomination-6730.html

Uygur, et al, are in denial.

[WeekiWater]

15. Cenk is a deceptive right winger with a very sordid history of ratfucking and sexism.

He uses the Fox news approach when deceiving his audience.

[R B Garr]

18. Bingo!

Cenk is a RWer, funded by a Republican. Went from a piggish GOPer to a new-found “progressive” dedicated to bashing leading Democrats for money. Hmmm.

[NYMinute]

26. But... but... he is consoling and soothing to some ...

[Cha]

37. I guess the OP didn't

know.. Thank You!

[highplainsdem]

45. And lately he's praising Tucker Carlson:

https://www.democraticunderground.com/100212099667

[MrsCoffee]

20. Screw Cenk and the Russians he rode in on.

ShitStirrers (being nice here) Incorporated. All they do is spew propaganda.

[dalton99a]

24. +1

[R B Garr]

35. Exactly! So perfectly stated!

[Kahuna7]

31. Three polls came out that same day..All were trending largely in Biden's favor...

They pick out the one poll that didn't show a breakdown by age. The Quinnipiac poll included under 50 breakdown but they don't want to talk about that because it showed Biden ahead with that group. Most recent polls are showing Biden ahead with younger voters.

[Cha]

38. But Cenk drones on.

[Gothmog]

39. From Nate Silver

Link to tweet

[highplainsdem]

40. Thanks, Gothmog! I like Nate's description of their nonsense: "their full-of-shit unskewing."

[Gothmog]

41. I trust Nate on polling issues

[Cha]

42. Hey! Nate Silver.. "To start with the basics, The Young Turks are full of shit about this,"

Link to tweet

Thank You, Goth!

[Gothmog]

43. Again I trust Nate on polling issues

[Cha]

44. I like his

way with words! lol