Irony, part 1 - I click on @cmyk 's article on teens using iPhones, and the ad on top is for Senior internet.
Irony, part 2 - in a thread whos content is saying telephone polls are broken, the centerpiece of the article is statistics based on a telephone poll from a biased source. (c’mon, you’re not going to convince me that a site called “Apple Insider” isn’t going to be biased towards Apple products).
I’m not interested in any Apple bias, one way or the other. The actual company that surveyed the kids isn’t Apple affiliated. They’re called Piper|Sandler, and apparently they survey teens semi-annually for many years now. Their site says they survey a geographically diverse subset of high schools across the US. Not sure if that’s by phone or leaflet or what. It would certainly put a ding in my theory if it was by phone.
In short, what I guess I’m trying to say is a LOT of teens/GenZs have iPhones. Most of those will be screening calls by default. Cold-calls from pollsters will be greatly culled.
It’s very much not just teens/Gen Zs, either. From the NYT’s methodology disclosure for the May 11-13 NYT/Siena poll:
We spoke with 4,097 registered voters from April 28 to May 9, 2024 …
Voters are selected for the survey from a list of registered voters. The list contains information on the demographic characteristics of every registered voter, allowing us to make sure we reach the right number of voters of each party, race and region. For this set of polls, we placed nearly 500,000 calls to about 410,000 voters.
There’s no way that call-screening on smart-phones is strictly a young-people thing.
ISTM that the group “people who’d respond to a cold-call poll” is substantially different in age, engagement, and otherwise from the group “people who’d never respond to a cold-call poll”.
Oh, definitely agreed. Myself being a fairly tech-savvy, GenX, 51 yo; I’ve also been screening my calls for the last few years. I can imagine this trend will only continue. I contrast that with my extremely conservative, Boomer parents, and they still have and call from a landline. They still get telemarketers and cold calls from everywhere. They do have smartphones, but even on those they get spammed constantly. My parents simply come from a generation and time where if the phone rings, you pick it up.
It’s not only that most GenXers, Millennials and GenZs don’t pick up an unrecognized call — their phones literally don’t ring.
I’m 66 and I’ve had my iPhone set to send any call not from a contact (and I have few contacts) straight to voicemail with no notification for as long as it’s been an option. I love my iPhone, but I rarely use it to make or receive a call.
@SyncoSmalls , I’ve cited your recent post in the “Who’s to blame?” thread over here to avoid a hijack of that thread.
SyncoSmalls:
… there is no reason to think polling procedures are suddenly less reliable than they were in 2020.
Thinking more that our jumping-off points are (a) the reliability of the 2022 mid-term polls, and (b) the media coverage thereof.
My apologies in advance if you’ve already read through the five posts linked below.
(Start with “Rosenberg himself, FWIW …” in the link above. Preceding text is the quote of another poster.)
…
Information aimed at 2024:
This is as good a place/time as any to mention this:
Toward the end of this podcast, they mention how the “margin of error” in these polls is the standard statistical one (one deviation, based on sample size), but this doesn’t include several other types of error — that the true margin is probably typically about twice this.
Just a good thing to keep in mind.
(ABC stopped posting 538 podcasts, so you have to go to Apple or Spotify.)
It’s tougher than that – the margin-of-error applies to each candidate’s poll percentage. That means a percentage difference has to be twice the margin-of-error to be statistically meaningful.
Despite my action in this thread, it’s fine to me for others to consider Biden as not leading in the polls. What I rail against are people treating Trump’s victory, in spring 2024, as a fait accompli. Done deal. All downhill for Trump because The PollsTM say so! Not so fast.
A post was merged into an existing topic: SyncoSmalls trocking thread
Poling has success: the people polled are idiots led by Trump and Foxnews to think we are going downhill fast. Majority of Americans wrongly believe US is in recession – and most blame Biden | US economy | The Guardian
Simon, 5/22/2024, writing about the differences between the April and May Morning Consult battleground polls:
Here are the results of new Bloomberg/Morning Consult polling of the battleground states, Biden-Trump:
- AZ 44-49 Biden gains 2 since last poll
- GA 44-47 Biden gains 3, within margin of error
- MI 46-45 Biden gains 1
- NV 46-46 Biden gains 8
- NC 42-49 Biden gains 2
- PA 46-48 Trump gains 1, within margin of error
- WI 46-47 Biden gains 3, within margin of error
These are good polls for us. For Trump to win he needs to win AZ, GA, NV, keep NC, and win one of MI, PA, WI. He lost all these states except NC in 2020. So he has to go get a lot of stuff he didn’t have in 2020. In this poll, and in the three other recent battleground state polls (NYT Likely Voters, Not Registered Voters) he is not ahead, outside of the margin of error, in MI, PA or WI, or any combination of states getting to 270. The map is hard for him, and today, simply and plainly, he is not leading or ahead in the 2024 election. He is not where he wants to be right now.
Our path to 270 is much clearer. Assuming we win the single Nebraska Electoral College vote we just need to win MI, PA, WI - all states we won last time, and all states where we have strong Dem governors who won in 2022. There are polls now with Biden ahead in AZ, MI, WI and tied in NV (yes this NV result is embarrassing for the New York Times). All four of these battleground polls have PA within margin of error, and a new poll about to be released in North Carolina has the Biden-Trump race there within margin of error, 43-45 Biden-Trump. This is the third poll taken in NC in recent weeks showing the election a toss up in North Carolina. Also note for those worried about Michigan that there are more polls showing us ahead in there than any other battleground state.
My big point here is that not all the data in front of us is pointing in the same direction - thus folks need to be cautious about jumping to conclusions or letting a single influential poll dictate our understandings.
The blue wall states - MI, PA, WI - are clearly within margin of error, tied now. We have recent polls showing AZ, GA, NC, NV also within margin of error, tied. Given all this it is simply impossible, wrong and inaccurate to argue that Trump leads or is favored. The election is close and competitive. Our path to 270 is easier. We have enormous financial and organizational advantages now, a better candidate and far, far better arguments. Senate and House polling remains encouraging, and we’ve made meaningful gains in the Congressional Generic. We’ve been winning elections of all kinds across the country since Dobbs and they keep struggling, a dynamic we’ve seen show up in voting in 2024 too
Michael Cohen is testifying under oath about Red Finch and gaming polls back in 2016.
EDIT: More reading.
EDIT 2: I see upon closer reading that Red Finch wasn’t asked to goose political polling. Still … too close.
Here’s an interesting analysis of why the polls have shown Trump leading.
It seems that Trump’s lead comes from those voters who identify as Democratic but are unlikely to vote. It also shows Trump does better among low information voters, voters who get their news from social media rather than traditional national and local media, and those who have not been paying any attention to the Trump trial.
I wonder if that means a lot of people who haven’t been paying attention to the trial are going to start paying attention once he’s found guilty — or once he’s found not guilty, and starts shouting from the rooftops that he’s been vindicated.
I think it means that those three things, being low information, getting news from social media, and not paying attention to the trial, are correlated. From what I gather, he people who fit in those three categories but who still identify as Democrats are the ones driving Trump’s polling advantage. I think if they had looked into the demographics of those particular people, it would match with the one’s that we tend to think of as being demographics Biden should have in the bag but is having trouble with, especially young people and Black men.
ETA: To be fair, I’m not blaming the people in those demographic groups. I’m blaming the people running the campaign for not doing a good job reaching out to those groups.
For example, there was a linked article in one of these P&E threads about a Black man in Detroit who is a local leader and stated he wasn’t sure who he out the other Black men he knows were going to vote for. People like that shouldn’t be in an article talking about how no one has reached out to him. They should be the local Democratic leader organizing their particular community.
The answer to the conundrum posed in that article is surely as simple as adding one more question to the poll: “do you intend to vote?”
I wonder if pollsters have determined that any kind of “do you intend to…” question is much less predictive than “well, did you actually vote in the last election?”
I have all sorts of intentions; my path is paved with them!
This all has to do with likely voter screening. Pollsters vary in the questions they ask to try to determine if the respondent will vote. General opinion among pollsters is that likely voter screening only works as you get closer to election day.
At present, I am noticing that almost all the national Biden-Trump polling is still being done without a likely voter (LV) screen, but that a lot of the swing state polling has started putting in the LV screen. Like everything with polling, I’m sure there is a good reason for this, but I haven’t come across it.
A likely voter screen will reduce sample size.
The numbers in the polls for Kennedy only reflect independents and naive people who do not understand that Trump and Biden will be the candidates. They are still voting for “someone else” so that in the end “they will replace Biden.”
I predict that it’s both; He’ll be found guilty on some charges, and therefore shout from the rooftops that he’s been vindicated.
Unless my fantasy comes true, and once the guilty verdict is announced, Trump goes right off the rails, threatens the judge, jury and prosecutor with death, and then gets charged with direct criminal contempt and is immediately escorted to his 30 day jail term (before sentencing for the guilty verdict)