This is the mysterious skew in the polls not accounted for Non-voters/Not Likely voters who equally mistrust both parties coming out for Trump. Many of these polls can’t be counting them. The numbers of of unlikely to vote/non voters are coming out for Trump and no one has even seen such in any election and so have not taken such a possibility into account!
I saw that Trump had a huge disadvantage among suburban women.
I also saw that the nonresponse rate is way up.
Possibly it’s genuine - but I wonder if they’re oversampling single women? They might be more likely to answer the phone.