8% for alternate candidates is too high with a 10% sample. The standard model used to be D+4 (in a decent year for RATS) and D+2 (in a good year from GOP). They work these poll numbers to create a “race,” remember pollsters need to be needed and media needs to sell ads, neither of which are needed if a blowout is expected.
I’ve no reason to believe they’re fudging the samples.
Let’s not forget the ‘Unskewed Polls’ disaster in 2012 lol.
Bingo. Media hires polling firm. Poling firm creates "results" needed to pump up clicks and encourage ad spending.