The ridiculous Wisconsin demos just prove what I've been saying all year. The polls are bad, the methodology is bad, and nothing is really being done to fix it because no one wants to admit the real issue: landline based samples reach a lot of weird people who are removed from the average person (who doesn't own a landline). And when they target cell phone users guess what happens: most young people don't answer the calls. So you get weird responses that result in wild, illogical samples like majority young voters supporting Trump or majority elderly voters supporting Kamala.
It's statistically impossible for there to be a 40 point shift over the course of 1-2 months without some drastic historical event (911, Pearl Harbor, etc). Yet the pollsters had no issue publishing that shyt lol.
That’s not the problem at all. Respectfully, you don’t seem familiar with modern polling. You seem to be thinking it’s just a matter of calling people and then publishing a response based on that.
This doesn’t remotely prove what you’ve been saying. This is an example of someone finding an example of something correlating to their belief and thinking they have a eureka moment or just hubris. Whatever it is, you’re misleading people and obscuring them from asking the actual questions.
The problem with the polling is that different pollsters create models and weigh different things based on what they believe to be likely outcomes or something like likely voters versus registered voters. So for example, Harris has a lot of young new voters but they wouldn’t fit into the likely voter category. That’s how Democrats beat the polls in 2022 - they had a lot of people who were registered but not “likely” show up to vote. The problems with a lot of these polls are that they are making certain assumptions that are giving them unbalanced sample sizes. It leads to doing things like overweighting white people or underweighting young people.
In other words, it’s not the process of getting responses - it is who you choose to contact and how you choose to weigh those responses. By making a choice like R + 3, you end up contacting more Republicans than Democrats and so you end up with numbers that don’t add up or if weigh whatever factors you think are more or less probable. The variance in the polls is based mostly on the formulas, not the responses. Internal polling doesn’t have this level of variance.