Re: 3 facets of research design?
Alan M. Zaslavsky (zaslavsk@hustat.harvard.edu)
Mon, 12 Sep 94 08:36:32 EDT
ostbye@sunsage.biostats.uwo.ca (Truls Ostbye):
> a few weeks ago an interesting post about concepts in problem solving/
> intro to stats appeared. the idea was that different designs were weak
> in some aspects, strong in others, and a tradeoff nearly always had to be
> made among them . there were at elst 3 such aspects, but i cannot remember
..
You may be thinking of Leslie Kish's "Statistical design for research",
where he talks about Realism, Randomisation, and Representativeness.
These three features are most strongly embodied in observational studies,
experiments, and surveys respectively.