Martin wrote
> I would have thought that a random sample would be much
> better!
I think this is why the supermarket element is objectionable --
responses will be biased towards the outlook of opinionated customers of
chain stores, rather than being those of "typical" NHS users, who include
the un-self-confident poor who only have access to back-street convenience
stores.
> My comment when I heard the plan was "They haven't
> got a statistician, have they?". Surely this is just an
> expensive publicity stunt.
I think "*cheap* publicity stunt" would be more precise: I gather
Milburn described it as a "census", which it plainly isn't -- they've left
out the horde of enumerators required to make a proper census (or even a
sensibly-designed survey) work, obviously because this would take serious
resourcing.
Julian
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
|