[R] paired t-test vs pairwise t-test

[Jack Tanner]

On Thu, 2004-08-19 at 14:42, Liaw, Andy wrote:
 > Perhaps other stat packages do it differently?  Does SPSS manuals detail
 > what its t-test procedure does, including which t-test(s) it does and 
when
 > it's appropriate?

SAS documentation, which Marc pointed out, is much more detailed than SPSS. 
For completeness, here's SPSS's:

username Guest, password Guest (case sensitive):
http://support.spss.com/tech/stat/algorithms/algorithms.htm

specifically, for the t-test, see
http://support.spss.com/tech/stat/Algorithms/12.0/t_test.pdf

 > That might make it easier on users, but I still think the
 > users should learn the appropriate use of statistical procedures
 > elsewhere...

Hell, I took my share of stats classes, but it's not like the terminology is 
always standard, anyway. I much prefer a walked-through example to an 
abstract explanation like "Calculate pairwise comparisons between group 
levels with corrections for multiple testing." That seems terse to me. It 
sorta makes sense only after the fact.