True enough.
And polling – particularly the generic D vs. R polls – have done reasonably well in terms of the midterms (w/some notable missed), particularly considering that midterms:
Apples and oranges
Nationwide polls have little value in local elections. There are many, many people who might think poorly of a political party as a whole but will swear their representative is an exception.
Also, poll of the general public or poll of likely voters? Or bogus mailer poll of the type you described? Generalizing all polls based on biased GOP polls you got in the mail is not sensible
OK, let me ask you, what is your polling experience? The ones I mentioned were the only times I was asked for my opinion. They were either weighted towards a particular ideology, or were too lengthy and monotonous to maintain my interest. I don’t answer calls that aren’t from my contact list, I don’t look for polls online to fill, nor do I talk to pollsters after I finish voting. I’m sure I’m not the only citizen who feels this way.
There was a brief period (a year or so) where I was getting some poll calls and I took them all and answered them. None of them took a particularly long time, and they seemed reasonable to me. Not particularly slanted.
I think I made it clear that I was slanted against Republicans and Trump in particular, though. Which is fine; that’s what polls try to determine I guess.
Well now, I just want to take a moment to acknowledge that I was wrong and @ParallelLines was correct. Apparently bullying and threatening your political allies DOES work. If you’re Trump.
Sens. Bill Cassidy, R-La., and Rand Paul, R-Ky., had previously voted to progress the resolution giving Congress the ability to halt the war.
But after a contentious closed-door lunch Wednesday, in which Cassidy said he had “lost my temper” and Trump said he raised his voice as well, Republicans held a late-night vote in which the two senators backed away from their support for the resolution.
I think I forgot to solve for the “Republicans are inherently spineless fucking cowards” variable. My mistake. 
There are probably other citizens who feel the same. But there are others who don’t. Generalizing your personal experience to be representative of all Americans is ludicrous on the face of it.
There’s a lot of study that’s been done on polling design and how to make them representative. The better ones do a good job on the whole.
But there are a few issues otherwise.
One is there are some polling firms that aren’t interested in accurately gauging public opinion. These tend to be focused more on conservatives. They’ve already got an idea of the answer they want to hear.
Another is that people generally are horrendously bad at actually understanding poll results. Or really anything involving numbers at all. People are good at seeing/hearing what they want rather than what is actually there. Example: the meteorologist predicts a 70% chance of rain. It doesn’t rain in the neighborhood (maybe it does somewhere else in town, maybe it doesn’t). This means the weather man was “wrong” somehow and terrible at his job. Lots of people think this way.
A few pollsters are better than others at explaining them but this is a fundamental problem with humans than with the polls themselves.
Whenever something like this happens, I just assume Trump used his mob connections to threaten their families.
By “his mob connections” I assume you mean “his MAGA followers”. I doubt the actual mob wants anything to do with him these days - too much risk, too much attention. But his cult members will still send death threats to anyone not sufficiently worshipping their Dear Leader at the drop of a hat.
This. It’s exactly what happened to Marjorie Taylor Greene once she defected from the Trump Adoration Society.
I love pointing out that when she was dissing Democrats she’d just threaten she carried a gun. When she defected she hired a bodyguard.
So it’s not that polling is flawed, but human understanding of them is? That further drives home the point that polling numbers are unreliable. So if Trump gets an approval rating of X%, we should derive other conclusions because we haven’t been researching how polls work properly? That just further erodes their veracity.
Let’s just add this to the many times I and others have been wrong in thinking “surely, there’s a bottom to this” rather than any predictive abilities on my part. You’re in good company.
How so?
Again, if the local weather forecaster says there’s a 70% chance of rain and somebody concludes that means it absolutely will rain, that’s not the fault of weather forecasting or the forecaster. That’s just somebody being a dumbass
Speaking as a professional market researcher: increasingly, it’s both.
Survey research in general has increasingly struggled with getting a true, unbiased sample over the last 10-15 years.
When I started in the industry, in the late '80s, the default research media were random-digit-dialing to landline phones, mail surveys, and shopping-mall intercepts; all three of those actually worked reasonably well, to give a pretty good sample, because the vast majority* of people did answer their phones, did read and respond to mail surveys, and did go to shopping malls.
As cell phones replaced landlines, as “number portability” came into play (meaning that the area code on your cell phone may have zero relevance to where you live), and as the explosion in spam calls means that most people won’t answer the phone if they don’t recognize who is calling, getting a good sample in phone research is very challenging.
Online research has become the norm now, but that, too, struggles with getting a truly representative sample, as not everyone is online, and even those who are are online may not want to participate in a research study; researchers know that “non-response error” is not evenly distributed across demographic groups, which means it’s harder still to make sure that all groups are represented in your sample.
tl;dr: political polling will always give you numbers, but due to increasing, systemic issues with how surveys and samples are constructed, it’s even more questionable now than it used to be as to how well those survey results represent the opinions of the entire voting population.
Post-Amazon and post-COVID, not as many people in malls, either.
Absolutely, and those who do go to malls now are not nearly as close to being a representative sample of Americans as it was back in the '80s and '90s.
I forgot to add the footnote:
*- Yes, I know that some of you will insist that, even back in the '80s and '90s, you didn’t automatically answer your phones, you didn’t respond to that sort of mail, and if you did go to shopping malls, you refused to talk to the “survey ladies.” The point still stands that most people did.
Dopers gonna dope. Back in the 80s and 90s, we pretty much answered every single phone call, at least until we got Caller ID, but my family wasn’t big on answering polls either by phone or through the mail. The only surveys I participate in are with the US Census with whom I happily took part in several surveys a few years back.
I never understood not answering the phone because you don’t know the number. I always answer the phone, if only out of curiosity. If it’s someone I don’t want to talk to, I say “no thank you” and hang up.