Polling and direct democracy

One of the things that I was wondering about since the polling errors seem so high in a lot of places is how misleading a lot of the general questions we see from polls are as well - Donald Trumps approval rating for example. It seems likely that many more people may approve of him, but just dont get weighted appropriately or polled at all (for various reasons). Similar thinking for Covid-19 handling, M4A desire, etc. I even wonder if Trump supports and their distrust of media make them less likely to do exit polls as well and how misleading those might be.

It seems like direct voting from the public at large is the best way to get these opinions. any other ideas?

People are not going to respond honestly if they feel shame or even remotely vulnerable, even if anonymous.

It’s why the estimation of LGBT population % has been so difficult for decades. There are methods to get around it (like flip a coin, if heads respond a gay, if tails respond honestly, and back out the answer probabilistically), but on a large scale that’s way too complicated for the average person.

Many polls are done via “random phone numbers” . . . guess where most numbers are likely to be?

Also, how do these poll account for those who really don’t want to participate in such polls?

Many ruralites I know really don’t care to participate in those “big city tom-foolery” functions.


Agreed. That’s where this question is stemming from.

Yeah the margin of error only makes sense for a truly randomized poll. That said, I believe the statisticians that are behind these polling methods are doing their best and adjusting for as many factors as possible, but clearly not enough if the margin of error is 5% and they’re getting >5% error rates in back to back elections.

I’m in favor of more direct democracy. I’d like to see a national referendum amendment to the US constitution.

I’ve pitched it before – referendum, not initiative. Reps who can’t get bills passed through the normal process can get a limited number onto a regular Nov ballot.

Yep. The polling story is a big deal. After 2016, they were really trying to do better and seem to have done worse. 538’s last national poll average had Biden ahead by 8.4%. He’s currently leading by 2.4%. IIRC, that’s a much bigger error than 2016.

Is it a techy thing, just not getting a cross section? Or is it a deliberate choice by a certain set of voters to not participate?

We never respond to phone surveys. I don’t think that’s related to our political leanings, but ??