Sure, but that is why polling (or good pollsters, of which Marist College is reputedly one) attempt to avoid bias through large sample sizes, randomizing, stratifying, combining multiple methods of reaching people, looking at likely vs. unlikely voters and on and on.
Instead of dismissing all polls out of hand as inherently inaccurate and unreliable because they can’t possibly account for every factor, what is the particular problem with this poll, other than it’s early and things might change? Is there some obvious flaw with it?
So was my soil sampling. I sampled shallow and deep, took thousands of samples; I sampled at multiple times of the day and night, on different days of the week and during different phases of the moon. I sent half of my samples to an independent lab to make sure i wasn’t introducing errors with my lab equipment. Still, all of my samples were wrong because of where I sampled.
There is a strong cultural bias in the US with political polling because a significant chunk of American voters will never answer a poll. And the chunk that will answer has a bias towards the conservative and towards the credulous. There is no amount of sampling that can correct for that bias.
And to @danimagoo’s point, every biased poll that is published reinforces that bias.
Right, but pollsters KNOW this and they try to design polls so that they account for those things. They use multiple methods of reaching people and stratify their sampling so that they get enough people in a variety of demographics. It will never be perfect, so I guess this argument is really about how much imperfection we’re willing to live with.
It’s a bit strange that (often young people and the left) bemoan that polling is broken because it doesn’t represent the population at large, yet they also choose not to participate in the polls which makes the pollster’s job harder and the polls less accurate.
Garbage in = garbage out. Finding a partially edible rind of parmesean in a bag with coffee grounds, rotting vegetable peels, and food-stained cardboad doesn’t get you a fine dinner.
Putting the onus of this problem on people who don’t respond to polls is shitty victim-blaming. Look at how we got here - shitty biased polls have been around a lot longer than responder hesitancy. Your cause and effect are inverted.
Okay, so, show me where the garbage is in this poll. How would you make it better?
You say polling is broken. You say one of the reasons is that people (often left leaning) don’t respond to the polls in a random manner. I see no other way to fix that than for those people to start responding, unless you are saying that the pollsters need to do more to reach them, in which case, what should they do?
I also don’t see how anyone is a “victim” here, like, who is being hurt by the polling?
I suspect the people who refuse a poll are more “suspicious of the government or political organizations” rather than young or left vs right. Though perhaps there are more motivated people among the extremists. I know that I would personally be willing to participate but the opportunity has never arisen. I’ve never thought to actively pursue it either though as I was under the impression this wasn’t how it was done.
It bugs me too. The Democrats use Act Blue to generate these emails. I think Act Blue is set up to receive payment and to vet employment status per the FECs requirement, which makes it easier than to have a 585 + individual websites.
Each time I make a donation, I get the same solicitations, but Act Blue has been really good about removing me from further emails when I’ve requested it.
There are likely other ways to be able to get the information they need, other than a method that has proven to be highly problematic in the past few years…
The polls didn’t elect Trump though. If people were complacent and didn’t vote because they thought Clinton had a comfortable lead, that is on them (and the media that misinterpreted the polls). The polls aren’t responsible for voter suppression efforts by Republicans either. If anything, polls are one of the better efforts to detect those efforts.
The polls in 2016 weren’t wrong, despite this take by Politico. They showed that a Trump victory was within the realm of possibility, especially with his electoral college advantage.
I didn’t say that. I said that the pollsters failed miserably at doing their part (a role they themselves claim) in supporting democracy in America.
Using Nate Silver of all people to back up that particular claim isn’t going to help your case.
Silver isn’t a pollster or academically trained statistician, just a pundit whose entire career depends on the public’s belief that polls are a pinpoint reflection of reality.
