So what does this have to do with 8 million people?

If the survey was of 2,079 people, then how does this extrapolate to 8 million? They say 23% of the respondents represent 8 million people, but that does not mean that 8 million people actually used crypto.