Interesting. The separate question on “sexual assault” was asked of 1915 “adults” but only 546 likely Republocans. So in other words, fully 2/3 of the population for that question were either Democrats, independents, non-citizens, or others whereby their voting status could not be determined. Talk about skewing a poll.
Maybe because when some respondents head it phrased as “sexual assault”, they hung up and invalidated the poll, so sampling may have become lopsided.