Investigating the "last holdouts" for anything is usually interesting. "The last 7%" is a heterodox group by definition. That said, I don't think the way this

Investigating the "last holdouts" for anything is usually interesting. "The last 7%" is a heterodox group by definition. That said, I don't think the way this data is presented tells us much about who they are. It just tells us how many they are (spoiler: few) within each demographic group. Besides being backwards, I'd also posture that sampling and surveying a "last 7%" would be very hard. Things get quirky at the tail.