On 19 Jan 2000 08:49:46 -0800, [EMAIL PROTECTED] (Muriel Strand)
wrote:
> i gather that a collection of events which is analyzed with statistics
> must have sufficient similarity (between each event) for the analysis to
> be accurate/precise. how similar is sufficient? can anyone recommend
> refs (preferably books) that discuss this issue, and provide guidelines
> for assuring sufficient similarity? does this consideration affect the
> appropriate choice of model?
After reading another answer, I think I get a notion of what this
question might be asking. It reminded me of a bit of Philosophy of
Science. So, then I did a search (google) on "nomological net" and
came up with a 119 matches -- about 118 more than I had hoped for.
Paul Meehl and L. Cronbach made a bigger splash with the idea than I
had ever noticed. A few of the hits were dead-links to a course at
MIT. The others that I could glance at seemed to have a high degree
of seriousness; I was impressed by the results. Their 1955 paper is
available, in its entirety --
http://www.yorku.ca/dept/psych/classics/Cronbach/construct.htm
Most relevant: Bill Trochim has a web site that I want to recommend,
with regard to some other questions that often arise in the stats
groups. It has some fine definitions and short discussions of
validity and related ideas --
http://trochim.human.cornell.edu/kb/nomonet.htm
--
Rich Ulrich, [EMAIL PROTECTED]
http://www.pitt.edu/~wpilib/index.html