CNN not happy with its own pollster's results

[fanfanois]

Democrats are ahead by 3% in the generic ballot, up from a tie in June.
But in their quest to attracting right wing nuts by becoming more like Fox, CNN says that in “competitive districts” Republicans are here by 5%.

Several questions go unanswered.
What was the margin in “competitive districts” in June? What qualifies as a competitive district?

Read and throw up.
https://www.cnn.com/2022/10/13/politics/cnn-poll-voter-preferences/index.html

[FBaggins]

1. They are correct to signal caution

I doubt it has anything to do with the audience they’re trying to attract.

Note that their last generic congressional poll before the 2020 election showed us up by 12. Being up by three now is an improvement over their most recent poll, but still worthy of caution.

[fanfanois]

5. Ok but more caution is warranted when you're tied then when you're ahead by 3

The CEO has purge liberals and wants right wing readers and viewers.

[FBaggins]

8. Who says that they're urging more caution than when their poll was tied?

You haven't provided any parallel commentary from that last polling report.

[Sympthsical]

2. There's a chart in the second paragraph

https://www.cnn.com/election/2022/house-race-ratings

Competitive districts are those that President Biden or Trump won in 2020 by less than 5%.

[Johnny2X2X]

3. And all districts have been redrawn

Typically, the polls are really hard to compile after redistricting, Biden or Trump +/5 is harder to predict now.

[Sympthsical]

4. I think prediction in general this year is rough

Yes, for the reason you stated. Also Dobbs and Trump things.

The axiom has always been that the general congressional preference always skews towards us because urban population densities don't reflect diffuse districts and gerrymandering. So if we're up by five on the general question, we might as well be tied.

That's usually how it is. But I don't think this is a usual year. Beats the hell out of me how this will go.

[Johnny2X2X]

6. Looking at the 538 predictor model

Dem +3 in generic ballot ends with Dems holding the House more times than not. This is going to be a barn burner. But I think Roe is the difference, it means 5 hidden points for Dems.

[onenote]

9. Why would the impact of Roe be "hidden" in poll results?

Particularly in a poll that specifically asked those polled how important they viewed abortion as an issue.

[Sympthsical]

10. New registrations

A lot of models use prior behavior as indicative of future actions. That's why you can end up with those odd differences between polls of "likely voters" or "polls of adults" etc. They're kind of accounting for different things.

20 states offer same-day voter registration. So we may not know if Dobbs is lurking in all these current numbers.

[Polybius]

25. Are you also factoring in the uninformed that blame Biden for rising gas prices and a bad economy?

There are a lot of those out there too.

[Sympthsical]

30. That's why I say any other year, I think we'd lose

Handily. But it's things like Dobbs and Trump that have me very uncertain. Mix in some baffling candidate choices like Walker and Oz.

I knew there was a solid chance we'd lose in 2016. I knew 2018 would go heavily for us. And I knew 2020 would be a lot closer in the battleground states than anyone was allowing.

I honestly do not know about this one. My instinct says it goes better than one would ordinarily think.

But I wouldn't put any of my own money on it.

[Johnny2X2X]

11. Turnout models

There's very little way to quantify how many voters will turn out because of Roe, it gets messy, so pollsters "likely voters" models really aren't taking it into account. That's what you saw Kansas beat the polls by 20 full points, and NY-19 (abortion was what the winning candidate ran on almost exclusively) by 10 points.

It's statistically messy to absorb such an event so pollsters just end up ignoring them and saying they aren't as certain.

[karynnj]

14. Good point

I think one idea behind the "hidden" impact is that it will have an unusual impact on who turns out to vote. The polls that try to predict likely voter need some model to adjust what an estimate from a sample of registered voters predicts. However, as you note, this sample directly asks about abortion. So it should pick up some i,pact.

[FBaggins]

17. D+3 is what we ended up with in 2020

And, of course, we won a very narrow margin.

However - that isn't the same thing as saying D+3 in the CNN poll results in D+3 at the ballot box.

[Polybius]

26. Yeah, and all the polling at the time in 2020 had us up by a lot more than 3 points

Maybe this year will be the opposite, hard to tell.

[FBaggins]

16. Redrawing districts doesn't mean that they can't tell what the 2020 results would have been

Precinct-level results are still available.

[brush]

7. Are pollsters polling cell phones now? At one point not too...

long ago they just polled landlines, which certainly skewed their results.

[Sympthsical]

12. Yes

Source: My phone, and I don't know how they get my number. I don't use it to sign up for anything.

They even poll by text now. Which . . . right. I barely reply to work via text. Like I'm gonna take the time to go take a poll.

[Genki Hikari]

21. Good grief

They've been calling cell phones for at least 2 decades now.

They've also used online polling methods for quite a long time now.

Stop listening to things ill-informed people say about polling, and instead pay attention to things like methodologies in polls. Every single pollster is required--by industry standards--to list the methodology behind every single poll they take. These days, it only takes following a link to access how they conduct a poll--from method of contact, to population breakdowns in their sample, and even the precise wording of the questions they asked.

The methodology of polls over the past 20 years reveal that no modern poll relies solely or even primarily on landline polling. It also would have revealed the dwindling share of landllne contact over time, which is now around 30% for most polls. What they don't tell you is that landline contact persists because rural people are more likely to have lousy cellphone or internet access.

[brush]

24. Two decades? I don't think that's true as many people...

didn't even have cell phones two decades ago so pollsters relied on calling landlines which were much more plentiful. I think you're exaggerating a bit but you're right about cell phones being polled now more than they used to be.

"Cell phones have been around since the late 1980s. Why has it taken pollsters so long to catch on?

“There’s a reason,” Dimock said. “Just two years ago (2008), I think it was just 12 percent of the adult population was cell-only. The latest figures from the government show that 16 percent of adults in the country are cell-only. This was the first election cycle where many of the major pollsters said, ‘We’ve got to do something about this.'”

https://www.ecommercetimes.com/story/have-pollsters-been-dialing-the-right-numbers-65030.html

[Celerity]

27. That article is completely out of date (it is from 2008)

By Jeff Meisner •TechNewsWorld•ECT News Network November 3, 2008 1:58 PM PT

[brush]

28. Yeah, I inserted 2008 in the text. The poster I was...

responding to said pollsters have been call cell phones for two decades, an exaggeration IMO. I don't think that's true as many didn't have cell phones two decades ago but did have landlines. So of course pollsters called the landlines to reach people.

And as the article said, it wasn't until a good percentage of people lived in a household with a cell phone only that pollsters started polling cell numbers also.

[Celerity]

29. Yes, not for 2 decades, agreed on that

[brooklynite]

13. I don't have a problem with the story...

The generic poll is just that. A GENERIC national average which says little about district by district results. Generic results have been rising, so the chance of individual wins improves BUT the number of competitive seats Democrats have to hold is greater than the number of pickup opportunities.

[Just A Box Of Rain]

15. The link doesn't support the false conspiracy theory that was initiated by right-wing sources.

If we are to feel "nauseated," it is when this sort of right-wing nonsense is perpetuated on DU.

Please stop with the baseless nonsense.

[AntivaxHunters]

18. I don't know what CNN is doing except I think I might

It really is a bad move and it's a bad move from a business POV too.
CNN has a huge problem with being a brand that's seen as toxic with "fake news!" for the right. It makes absolutely no sense what their executives & CEO's are doing financially. They're completely ditching their base of viewers who see CNN as a factual news source and claiming they want a more "centrist approach".

I don't believe that's the truth or the reality of the situation and I have a working theory about this.

What's being fed us is nothing but sheer unadulterated gaslighting, lies, & bullshit.
That's not what this is really all about. Nope, not one bit.
CNN's new CEO is a Trumper. Literally. He's a huge Republican. His mission isn't what he saying. His mission I believe is to completely destroy CNN entirely; literally taking them completely out. Think about it.....Trump has called CNN "fake news" for years. To objectively destroy the station entirely would only benefit 1 thing -- The GOP.

Eliminating the station is, I believe, what the true mission is here. CNN full well knows that the right will NEVER watch the network and if they won't then nobody can. "Burn it all down" is the logic here because CNN has done massive damage to the GOP & Trump.

Anyways, good morning

[Just A Box Of Rain]

19. None of this is true.

Christ Licht, the new CEO of CNN is NOT a Trumper.

And the CEO of Warner Discovery (CNN's parent company) is a Democrat who contributes to Democrats.

We may be being fed "sheer unadulterated gaslighting, lies, & bullshit," but of a different kind that you suggest.

[Midnight Writer]

20. Democrats nearly always get more total votes for the House and the Senate.

It is the distribution of the votes that will decide the results.

For example, a Congressional District in Chicago may get 75% Democratic votes, and win that seat. Meanwhile, a suburban District may go to the Republican by 2%. The total votes are overwhelmingly for Democrats, but each Party will end up winning the same number of seats; one apiece.

[Kaleva]

23. But rural areas going 70% or more Republican will balance that out

As districts are divided by population Each House district represents a little over 700k people.

[Kaleva]

22. CNN continues to live rent free in the heads of some.