Re: Understanding response protocols

Keith Moore <moore(_at_)cs(_dot_)utk(_dot_)edu> writes:

> you don't have firm evidence either, because you don't have anything
> close to a random sample.  statistics that aren't representive of what
> you are trying to measure are meaningless.
Good point; I can't really use evidence of the case that I think is most
typical to prove to you that it's the most typical.  :)

-- 
Russ Allbery (rra(_at_)stanford(_dot_)edu)             
<http://www.eyrie.org/~eagle/>