The internals on the web page say telephone poll of adults 18 or older. The phones 70% cellphones, 30% landline. Random dialed numbers. Long time since my college days but this is NOT a scientific poll. They need to study analysis of variance at Gallop.
Could you post the internals? I really could not find who they sampled.
When it’s a poll of “adults”, not registered voters, it’s always suspicious, because they don’t know the attitudes and leanings of the responders.