@LiberalLibrarian I'm sure the "official" result "corrects" the oversample. But as I said below, that means Siena is creating a demographic model based on its own assumptions.

The biggest problem with the poll is that the 14% "uncommitted" are, by NYT's own admission, much more hostile to Trump than Biden. But pundits cling to "they say they won't vote or will vote 3d party so they DON'T COUNT!" Because "I'll vote 3d party!" declarations are always reliable 15 months out, right?