Wilcox paired test error message

On Jun 23, 2013, at 22:30 , Sarah Goslee wrote:

Hi,

If you're doing exactly as described below, then you need to add the
data argument to wilcox.test so R knows where to find beforesmall and
aftersmall. But if so, that's a rather uninformative error message.

Saraj

Not really, if you use the formula interface, the rhs is supposed to be the grouping, as in

wilcox.test(extra ~ group, data=sleep, paired=TRUE)

so it's telling you that "aftersmall" does not describe two groups.

(It is unfortunate, though, that we don't have a formula interface to the parallel-vector data layout for paired tests. That is the common case, the sleep data set is a rather rare exception. A formula specification for paired columns has been talked about; something like cbind(before, after) ~ 1 should be workable, but nothing has materialized to date.)

-pd

On Sun, Jun 23, 2013 at 2:34 PM, ruth parker <parkeractonruth at gmail.com> wrote:

Hi,
I've been trying to run a wilcox paired test on some data
beforesmall aftersmall
[1,]        63.5      512.0
[2,]        54.5      237.5
[3,]        52.5      161.5
[4,]        78.0      153.5
[5,]        53.5       68.0
[6,]        50.5       65.5
[7,]        69.0       52.0
[8,]        76.0       59.0
[9,]        68.0       66.5
[10,]        75.5       66.5
[11,]        67.0       45.5
[12,]        81.0       54.5
[13,]        49.0       44.0
[14,]        51.0       42.5
[15,]        53.0       34.5

using
wilcox.test(beforesmall ~ aftersmall,paired=T)
but I get the error message:

Error in wilcox.test.formula(beforesmall ~ aftersmall, paired = T) :
 grouping factor must have exactly 2 levels

I don't have any missing values, I have the same amount of data points for
each group and I've looked everywhere trying to find the answer.
Please can somebody tell me how to make it work
thanks