HomeLatest ThreadsGreatest ThreadsForums & GroupsMy SubscriptionsMy Posts
DU Home » Latest Threads » Forums & Groups » Retired » Retired Forums » 2016 Postmortem (Forum) » Important UPDATE to CNN/O...

Thu Sep 15, 2016, 01:59 AM

Important UPDATE to CNN/ORC Ohio poll: only 18% of sample was under 50 years old.

Last edited Thu Sep 15, 2016, 02:45 AM - Edit history (1)

By contrast, in the last 3 election cycles, people under 50 comprised 44-51% of voters.

http://www.politicususa.com/2016/09/14/cnn-trump-lead-ohio-polling-50-years.html

Update: It turns out that CNN did poll voters under 50, but the reason why their crosstabs showed no data was because they polled too few younger voters. Younger voters have made up 44%-51% of those who voted in the last three election cycles. CNN’s sample was made of 18% younger voters. In essence, CNN cut the number of younger voters in Ohio in half.

Donald Trump will definitely win Ohio if no one under the age of 50 comes out to vote, but that is not going to happen. It may just be an error, but if CNN shaped their numbers to get a newsworthy headline, it would serve as a perfect example how news networks use their polling to make the news instead of reporting on it.


_________________________________________

Also, compared to 2015 PEW figures for party affiliation, this CNN/OR study may have over-sampled Republicans. (I say “may” because party affiliation can change over time.)

A total of 1,006 Ohio adults were interviewed by telephone statewide by live interviewers calling both landline and cellphones. All respondents were asked questions concerning basic demographics, and the entire sample was weighted to reflect statewide Census figures for gender, race, age, education and region of the state.

This sample includes 606 interviews among landline respondents and 400 interviews among cell phone respondents.
Respondents were asked questions about whether they are registered to vote, their likelihood of voting, past voting behavior, and interest in the campaign. Based on the answers to those questions, 769 respondents were classified as likely voters. Among those likely voters, 31% described themselves as Democrats, 32% described themselves as Republicans and 37% described themselves as independents or members of another party.


http://www.pewresearch.org/data-trend/political-attitudes/party-identification/

Ohio party affiliation in 2015

23.7 Republican
30.4 Democratic
40.1 Independent

24 replies, 2159 views

Reply to this thread

Back to top Alert abuse

Always highlight: 10 newest replies | Replies posted after I mark a forum
Replies to this discussion thread
Arrow 24 replies Author Time Post
Reply Important UPDATE to CNN/ORC Ohio poll: only 18% of sample was under 50 years old. (Original post)
pnwmom Sep 2016 OP
DemonGoddess Sep 2016 #1
davidn3600 Sep 2016 #2
Divine Discontent Sep 2016 #4
ProudToBeBlueInRhody Sep 2016 #5
pnwmom Sep 2016 #7
Norbert Sep 2016 #10
NewJeffCT Sep 2016 #12
davidn3600 Sep 2016 #17
alcibiades_mystery Sep 2016 #16
herding cats Sep 2016 #23
Divine Discontent Sep 2016 #3
Iliyah Sep 2016 #6
Bernardo de La Paz Sep 2016 #8
MyNameIsKhan Sep 2016 #9
Hortensis Sep 2016 #11
whistler162 Sep 2016 #13
BumRushDaShow Sep 2016 #15
uponit7771 Sep 2016 #14
triron Sep 2016 #18
triron Sep 2016 #19
pnwmom Sep 2016 #21
triron Sep 2016 #20
triron Sep 2016 #22
triron Sep 2016 #24

Response to pnwmom (Original post)

Thu Sep 15, 2016, 02:08 AM

1. Surprise surprise

they tailored the poll to fit their horse race. urrrgh

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Thu Sep 15, 2016, 02:20 AM

2. In my experience, those who are worried about sample distribution are losing

 

Complaining about who is being polled is typically what the right wing always complains about, and end up wrong.

With the youth vote, it's difficult to gauge. It's harder to reach young people to poll. However, it's even trying to figure out if young people are actually going to vote or not. It's been said for quite some time that young people aren't all that thrilled about this election. So they are not going to turn out in the numbers you saw with Obama. The pollsters seem to agree with that.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to davidn3600 (Reply #2)

Thu Sep 15, 2016, 02:33 AM

4. I could see that for under 30, but under 50? It seems a clear under-representation, Ohio isn't

Ft Myers FL. Ya know?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to davidn3600 (Reply #2)

Thu Sep 15, 2016, 02:35 AM

5. 18% would be a historic low

No legit pollster can project that.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to davidn3600 (Reply #2)

Thu Sep 15, 2016, 02:42 AM

7. I guess it's good to hear that somebody considers under 50's to be "young people."

Reply to this post

Back to top Alert abuse Link here Permalink


Response to davidn3600 (Reply #2)

Thu Sep 15, 2016, 03:41 AM

10. In 1932 Hoover was primed for an easy reelection over FDR

based on phone surveys. Pollsters back then didn't worry about sample distribution or who actually had telephones on their homes back in the Great Depression.

How did that work out for Herbert Hoover?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to davidn3600 (Reply #2)

Thu Sep 15, 2016, 05:00 AM

12. hmm - I was worried about that in 2012

and that turned out okay. I've posted on here numerous times - back in 2012 and again this year as a reminder - but, after the VP debate in 2012, most people felt that Biden easily beat Paul Ryan. However, the CNN poll had it as a tie and the next day Republican co-workers were happy that "liberal" CNN called it a draw.

However, if you looked at the sample they used, they sampled 8% more Republicans than Democrats - something like 48% Republicans and 40% Democrats. The reality is that Republican turnout never beats Democratic turnout in terms of percentages - even in the 1984 Reagan landslide, Democrats made up 39% of the voters and Republicans 35%. (Reagan won independents handily and got a lot more of the Democratic vote than Mondale did of the Republican vote) - the actual vote turnout in 2012 was something like 40% Democrat and 32% Republican, after being 41%/32% in 2008.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to NewJeffCT (Reply #12)

Thu Sep 15, 2016, 08:40 AM

17. And did CNN calling that debate tied change anything?

 

No. So is it worth getting upset over it?

You cant wake up every morning and get hysterical because a poll here or there says something you dont agree with. My statistics professor in college always used to say that "statistics is as much an art as it is a math."

And of course there is the famous saying, "there are three types of lies: lies, damn lies, and statistics."

But many people, especially here, are guilty of confirmation bias. They only trust the polls that show what they think the results should be. And then they attack and scrutinize only the polls they disagree with. Meanwhile they don't pay any attention that the poll they do agree with over-sampled Democrats.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to davidn3600 (Reply #2)

Thu Sep 15, 2016, 07:08 AM

16. 35-49 is not a "youth vote"

 

Reply to this post

Back to top Alert abuse Link here Permalink


Response to alcibiades_mystery (Reply #16)

Fri Sep 16, 2016, 12:32 AM

23. Hush.

I'm still basking in the glow of my rediscovered youth. Don't ruin it for me just yet.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Thu Sep 15, 2016, 02:31 AM

3. THANK YOU, pnwmom! Very important info that they thought was accurate... it's laughable...

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Thu Sep 15, 2016, 02:35 AM

6. Ohio, NC, FL, NV are considered toss ups

Fl tied pe se, and Oh tu, so with 2 polls showing 7 - 5 for Oh for t-rump, HRC looks to be fairly in good shape even with constant trashing by our lovely worthless media.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)


Response to pnwmom (Original post)

Thu Sep 15, 2016, 02:55 AM

9. Ridiculous poll Ann Selzer and CNN

5 point loss in Ohio means 500k votes ... Obama won 2012 by 200k votes... If we extrapolate 25percent of those voters will stay home not vote trump .., this is just not possible ...


46 percent of voters in Ohio less then 50 voted in 2012.., 18 percent in CNN is a joke .., that down sampling about 150 voters in 750 sample ...

JEEZ

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Thu Sep 15, 2016, 04:36 AM

11. When poll results seem strange, be suspicious.

There are a lot more bad polls out there than good ones. And a whole bunch are presented deceptively -- as in not telling people only left-handed widows over 70 were polled.

This poll bucked the rest of the picture and should never have been just swallowed whole.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Thu Sep 15, 2016, 05:52 AM

13. They need to run a survey on

how many people use https://www.nomorobo.com/ using a robocall system!

I geet at least 6 or more single ring, robocalls, a week all of which are dropped. I wonder how many other households have the same set up?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to whistler162 (Reply #13)

Thu Sep 15, 2016, 06:04 AM

15. I use a digitone

http://www.digitone.com/

Best thing I could have ever purchased for this sort of thing. And the "single-ring" only happens on the phones that don't have the device on it so I have the device on my bedroom phone to avoid hearing even the single-ring of illegal/scam robocalls at night while I'm asleep.

I may consider getting the upgrade model but the one I have right now is doing the job.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Thu Sep 15, 2016, 05:56 AM

14. See, this is why I don't trust the LV models with no standards.... they can skew to irrational ends

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Thu Sep 15, 2016, 11:38 PM

18. In yougov poll

over 40% were under 45. Very hard to understand a 12 pt swing in 2days but this is one mechanism. Yougov got party affiliation about right according to 2012 data it seems.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Thu Sep 15, 2016, 11:41 PM

19. What

about internals of nbc poll. How much of this supposed tightening is really there?

Reply to this post

Back to top Alert abuse Link here Permalink


Response to triron (Reply #19)

Thu Sep 15, 2016, 11:56 PM

21. Much of the "tightening" was due to the switch from registered to likely voters in most of the polls

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Thu Sep 15, 2016, 11:53 PM

20. kick

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Fri Sep 16, 2016, 12:15 AM

22. In yougov poll

Clinton won the 18-30 age group by 51-19 margin and 30-45 by 50-30 margin
Yet in the CNN poll according to cross tabs she only won the under 50 by 1. Seems extremely incongruous with the yougov poll. Their data is very suspect since they list N/A for both subgroups under 50. Seems contrived somehow.

Reply to this post

Back to top Alert abuse Link here Permalink


Response to pnwmom (Original post)

Fri Sep 16, 2016, 12:38 AM

24. My suspicion

is Hillary is probably ahead of Trump in Ohio. Likely she is ahead in Florida as well. Same reasons as I have indicated for Ohio (i.e. Yougov had her ahead in Florida as well, although closer)

Reply to this post

Back to top Alert abuse Link here Permalink

Reply to this thread