The help page for ansari.test() says (emphasis added):
By default (if exact is not specified), an exact p-value is computed if
both
samples contain less than 50 finite values **and there are no ties**.
Otherwise, a normal approximation is used.
However, this does not appear to be the case in R 2.4.1 or R 2.5.0. In fact,
even
example(ansari.test)
produces a warning
Warning message:
cannot compute exact p-value with ties in: ansari.test.default(ramsay,
jung.parekh)
although the "exact" parameter is omitted.
Patch:
Index: src/library/stats/R/ansari.test.R
===================================================================
--- src/library/stats/R/ansari.test.R (revision 40902)
+++ src/library/stats/R/ansari.test.R (working copy)
@@ -29,7 +29,7 @@
TIES <- (length(r) != length(unique(r)))
if(is.null(exact))
- exact <- ((m < 50) && (n < 50))
+ exact <- ((m < 50) && (n < 50) && (!TIES))
if(exact && !TIES) {
pansari <- function(q, m, n) {
Regards
Thomas Friedrichsmeier
pgp4plLqZm7WH.pgp
Description: PGP signature
______________________________________________ R-devel@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-devel
