The 35/48 number is what they wanted to make him look vulnerable. Too bad the other 17% undecided are overwhelmingly R-leaning. I mean, why didn't they poll a Democratic opponent? Maybe they just innocently forgot lol
Btw, pay no attention to their makeup of 2020 results, this is a sleazy trick they use to make their polls seem fair. I recall many polls of Iowa, Texas, Florida, etc. where the 2016 makeup was spot on but the poll was 7-10 points off.
I suspect that too. I remember Tillis having approval ratings like this in early 2019, and we all know how that turned out.