By the way,
an update from yesterday that should dispel all of the Alex Jones stuff:
"
CNN did an internal investigation and found that an operator had enlarged the font on their computer screen, cutting off Buttigieg's name, which had been at the bottom.
The polling methodology randomly cycles candidates' names so that they are read in a different order, making it impossible to determine which other candidates' names might have been left out — or how many times it might have happened.
While the polling partners are aware of only one instance of a candidate's name being omitted, the decision was made to cancel the poll because the same thing could have occurred with Buttigieg or other candidates.
"
Do they still do polls by calling landline numbers?
Is that at all supposed to be an accurate sample?
Another polling myth that just won't fucking die.
There are several approaches, and several ways to get a representative sample. In
this survey from last June, the Des Moines Register used phone numbers from the Iowa voter registration list, which is a reasonable way of getting a solid landline-to-mobile ratio.
But the short answer is mostly yes, landline responses are still a major part of getting an accurate sample of voters. Pollsters are well aware that this may not be true forever, and have been incorporating different methods into their work for years.
The super short answer is read part 1 and part 2 of
this article.