Vivian Martin (VBMARTIN@AOL.COM) wrote: < snip ... >

: For me, all roads led to SPSS, but it appears, from manual reading, that : there is an assumption of randomness for all the tests, even ones that some : textbooks would indicate can be used for purposive nonprobability samples : with nominal and rank orders, which is where I fit in. Am I misreading the : manual, or are there manuals/Internet articles that might clarify this for

-- I am just a statistician, though I am one who happens to answer a lot of questions that other people ask... and I want to ask you what you mean by this? -- "purposive nonprobability samples" that have some special need for nominal or rank order tests. Maybe, you could cite a page of your textbook?

Non-probability samples can leave you at a loss for good estimates of means, etc., that depend on relative sizes of groups; but the simple tests of equality are not more problematic, I think, than rank order tests. So, I don't know what you or your text referred to.

Rich Ulrich, biostatistician wpilib+@pitt.edu http://www.pitt.edu/~wpilib/index.html Univ. of Pittsburgh