The polling methodology is so far off they can't even be called polls, just guesses.
A poll calls people at random, gets their answers, and then gets their demographic information so they can adjust the numbers to represent the total population.
If you call 100 people across the country and by chance 80 percent of them happen to be black, you can't just put that out there as raw data. You have to adjust that number to the nationwide black population of 13%. It's called weighting, and it's necessary.
But what happens if your assumptions about your weights are incorrect? What if you look at 2016 and see that 60% of Registered Republicans showed up to vote and you base your weights on that number? What if it was 65% of Registered Democrats?
What happens to your poll when 2020 rolls around and 75% of Registered Republicans show up and only 55% of Registered Democrats?
Well, then your polls were all way off because you weighted them wrong. It happened in 2016 and it's going to be even more extreme this year. The number of people who went from "never voted" or "registered voter" to "likely voter" skyrocketed on the Republican side and has gone down on the Democrat side.
Every poll is wrong because their weighting algorithms are shitty.
The polling data is right.
The polling methodology is so far off they can't even be called polls, just guesses.
A poll calls people at random, gets their answers, and then gets their demographic information so they can adjust the numbers to represent the total population.
If you call 100 people across the country and by chance 80 percent of them happen to be black, you can't just put that out there as raw data. You have to adjust that number to the nationwide black population of 13%. It's called weighting, and it's necessary.
But what happens if your assumptions about your weights are incorrect? What if you look at 2016 and see that 60% of Registered Republicans showed up to vote and you base your weights on that number? What if it was 65% of Registered Democrats?
What happens to your poll when 2020 rolls around and 75% of Registered Republicans show up and only 55% of Registered Democrats?
Well, then your polls were all way off because you weighted them wrong. It happened in 2016 and it's going to be even more extreme this year. The number of people who went from "never voted" or "registered voter" to "likely voter" skyrocketed on the Republican side and has gone down on the Democrat side.
Every poll is wrong because their weighting algorithms are shitty.