Specifically because their methodology for sample selection is entirely hidden. If you do a generic online survey with self-selected respondents then 6k is not

Specifically because their methodology for sample selection is entirely hidden. If you do a generic online survey with self-selected respondents then 6k is not very many people from which to be drawing sweeping conclusions. Edit: also, I disagree that sample size is entirely worthless in non-random sample selection. If your sample size is a big enough percentage of the target population you are studying, you can draw some conclusions regardless of randomness. You just end up having to use absurdly large samples to get meaningful data.