RE: green bubble syndrome

That's my point.  We have to be persuaded that the sample was properly
chosen.  Beyond that, the relatively small sample size does not in
itself invalidate the significance of the results.

The results are only valid for the 1000 people they bothered to survey.
Claiming anything above that is pure nonsense and at best simply a very
rough guesstimate.
 Such surveys should always specifically say something like "40% of the
people surveyed" ...NOT EVER simply "40% of people", because
that's where the novice general public then confuses it with being
reality for everyone, when it is not.