Okay but they're talking about a difference of a few percentage points, and the survey collected a lot of other demographic information. It strikes me as very

Okay but they're talking about a difference of a few percentage points, and the survey collected a lot of other demographic information. It strikes me as very unlikely that this is statistically significant. If you check every demographic split, you're likely to find a handful that show whatever phenomenon you want (in this case that crypto is more popular with group X).