[In future, please press reply all as there might be others in the mailing list who can help you] Without a (short) reproducible example, it would be difficult to say. BTW, I think you have mistyped the error/warning message. Does it actually say "cannot compute _correct_ p-value" or "cannot compute _exact_ p-value". See example below.
wilcox.test( 1:3, 1:10 )
Wilcoxon rank sum test with continuity correction data: 1:3 and 1:10 W = 4.5, p-value = 0.08964 alternative hypothesis: true mu is not equal to 0 Warning message: Cannot compute exact p-value with ties in: wilcox.test.default(1:3, 1:10)
On Thu, 2004-11-04 at 00:31, fang lai wrote:
Dear Adaikalavan, Many thanks for your suggestions. The message "cannot compute correct p-values with ties in" sometimes shows up as an error (in most cases I have small group for the test), and sometimes shows up as waring when I typed warning() (in most cases the group I am testing is relatively large). But all the data are from the same data set. So what is the distinction between these two? Many thanks, Fang --- Adaikalavan Ramasamy <ramasamy at cancer.org.uk> wrote:
Reading the posting guide will tell you (among other things) to : a) use a sensible subject line b) give a reproducible example when possible See other comments below. On Wed, 2004-11-03 at 19:04, fang lai wrote:
Hi there, When I ran the wilcox.exact test, it always shows "number of items to replace is not a multiple of replacement length" However, according to the help it seems to be
able to
handle sample of different size.
I think the error refers to incorrect lengths during assignment. See example below.
aaa <- matrix( 1:4, nr=2 ) aaa
[,1] [,2] [1,] 1 3 [2,] 2 4
aaa[1, ]
[1] 1 3
aaa[1, ] <- 1:3
Error in "[<-"(`*tmp*`, 1, , value = 1:3) :
number of items to replace is not a multiple of
replacement length
Or you might be trying something like
aaa <- matrix( 1:4, nr=2 ) aaa[1, ] <- wilcox.test( 1:3, 1:10 )
Error in "[<-"(`*tmp*`, 1, , value =
wilcox.test(1:3, 1:10)) :
number of items to replace is not a multiple of
replacement length
Wilcoxon and all other two-group test statistics
(when pairing is not
involved) can handle groups of different size.
Try wilcox.test( 1:3, 1:10 ).
To extract p-value only, do wilcox.test( 1:3, 1:10
)$p.value
What does this message indicate, and will it cause
a
problem of the resulting p-value?
Not directly. You could be expecting p-values to be in, say the 5th column, but it might be stored elsewhere. But in your case I think the results will not be stored anyway.
Also, when I run the ks.test(), it shows
sometimes
that cannot compute correct p-values with ties in: ks.test(income.perm[(data[, 3] == 59) & (data[, 2]
==
1)], income.perm[(data[, ...
Is this a warning or an error ? There is a distinction.
And sometime it shows results and sometimes it
just
return NULL. When it shows the results, does it
mean
that it had some kind of continuity correction or other adjustment so that I could believe the
results?
Many thanks,
Fang
=====
Lai, Fang
PhD candidate
University of California, Berkeley
Department of Agricultural and Resource Economics
314 Giannini Hall, Berkeley, CA 94720-3310
tel: (510) 643 - 5421(O)
(510) 847 - 9811(Cell)
fax: (510) 643 - 8911
email: lai at are.berkeley.edu
http://www.are.berkeley.edu/jobmarket/fang.html
______________________________________________ R-help at stat.math.ethz.ch mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide!
=====
Lai, Fang
PhD candidate
University of California, Berkeley
Department of Agricultural and Resource Economics
314 Giannini Hall, Berkeley, CA 94720-3310
tel: (510) 643 - 5421(O)
(510) 847 - 9811(Cell)
fax: (510) 643 - 8911
email: lai at are.berkeley.edu
http://www.are.berkeley.edu/jobmarket/fang.html
__________________________________________________