On 06 Aug 2003 23:59:13 +0200, "Nils Chr. Framstad" <[EMAIL PROTECTED]> wrote:
> > Rich Ulrich <[EMAIL PROTECTED]> writes: > > > > > On 06 Aug 2003 15:58:28 +0200, "Nils Chr. Framstad" <[EMAIL PROTECTED]> > > wrote: > > > > > > > > Is there a good (encyclopaedic) source on this, including the > > > underlying assumptions? Should cover at least the gamma, the Tau's and > > > Somers' D. > > > > Here is a review online of 2x2 tables. > > > > http://www.stir.ac.uk/staff/psychology/rrm1/conchi.html > > Thanks... However I am not really interested in 2x2, but in many x > very many (possibly continuum.) You wrote that you were interested in underlying assumptions; I think you should find that the 2x2 case offers the initial, deep discussions of relevant issues that won't be explicated in as much detail when you read about the R x K. Maurice Kendall wrote the classic book (1948) on Rank correlation methods. I assume that the 1990 edition (with a co-author) is comprehensive, but I have not seen it. > > Will browse the reference list. I have Goodman & Kruskal on my desk > right now. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html "Taxes are the price we pay for civilization." Justice Holmes. . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
