In article <[log in to unmask]>, Zahava
<[log in to unmask]> says:
>
 We are planning to survey some Smithsonian WWW users later this fall - when
>we are comfortable with a sampling methodology. Again, we'll keep
>Museum-L informed.

I think we need to ask ourselves whether the question we want to answer is,
"Having visited our museum, what is the chance that people have seen our WWW
page?" or is it, "Having seen our WWB page, what is the probability that
they'll visit our museum?" Surveying visitors answers the former question.