Can anyone point me please ti discussions of sampling
(stratification/clustering etc.) when *all *costs are included?
Generally the costs considered are those of the survey body, not those of
respondents. Yet presumably if e.g. some categories of respondent would
incur greater resource costs than others in providing equivalent data
(whatever that may mean), then the respondents having greater costs would
receive a lesser sampling fraction.
I am currently considering a membership organisation which exists for the
benefit of its members. If they are being surveyed, surely their costs
should be included in deciding 'optimality'? However, some of these issues
arose also in a survey I discussed for NHSID some time ago. But generally
they are ignored, are they not?
Another issue is that of 'overload': if the survey is one of a series, then
layover effects should enter the formulae. But how?
Thanks for any thoughts you may have on these matters.
JOHN BIBBY
You may leave the list at any time by sending the command
SIGNOFF allstat
to [log in to unmask], leaving the subject line blank.
|