R help - Jan 2008 - How to apply the wilcox_test function to subsets ?

Dear R-forumites,

 

I want to apply a Wilcoxon test on subsets of the data frame mydata,
splitted using the myindice variable. When I send :

 

wilcoxtest <- by(mydata, mydata$myindice, function(x)
{wilcox_test(x$value~x$fact)})

 

I get :

 

Error in eval(expr, envir, enclos) : object "x" not found

 

whereas 

 

wilcoxtest <- by(mydata, mydata$myindice, function(x)
{wilcox.test(x$value~x$fact)})

 

works !

 

Why does it fail with the wilcox_test only? And, most important, how can I
apply the wilcox_test function to subsets of mydata splitted according to
mydata$myindice ?

 

Thanks in advance,

 

Jean-Luc Dupouey

Forest Ecology and Ecophysiology Unit

INRA

54280 Champenoux

France

 

mail: dupouey@nancy.inra.fr

 


	[[alternative HTML version deleted]]

"Dupouey Jean-Luc" <dupouey at nancy.inra.fr> wrote in
news:000001c8605e$c987c650$0b01a8c0 at PHYTOD420: 

On my system wilcox.test() exists, whereas wilcox_test() did not.  
Help.search("wilcox_test") does not produce any clues. If you are
using
a test from another package, you should specify which package the test 
comes from. Web searching produces links to package coin.

Have you tried?

wtest <- by(mydata, mydata$myindice, function (x) {
                     wilcox_test(value~fact, data=x)})

----slight modification to the wilcox_test example
data------
> mydata
     pd         age grp
1  0.80     At term   1
2  0.83     At term   2
3  1.89     At term   1
4  1.04     At term   2
5  1.45     At term   1
6  1.38     At term   2
7  1.91     At term   1
8  1.64     At term   2
9  0.73     At term   1
10 1.46     At term   2
11 1.15 12-26 Weeks   1
12 0.88 12-26 Weeks   2
13 0.90 12-26 Weeks   1
14 0.74 12-26 Weeks   2
15 1.21 12-26 Weeks   1
16 1.10 12-26 Weeks   2
> wilcoxtest <- by(mydata, mydata$grp, function(x)
+ {wilcox_test(pd~age, data=x)}
+ )
> wilcoxtest
        Asymptotic Wilcoxon Mann-Whitney Rank Sum Test

data:  pd by age (12-26 Weeks, At term) 
Z = -0.4472, p-value = 0.6547
alternative hypothesis: true mu is not equal to 0 

---------------------------------------------------

        Asymptotic Wilcoxon Mann-Whitney Rank Sum Test

data:  pd by age (12-26 Weeks, At term) 
Z = -1.3416, p-value = 0.1797
alternative hypothesis: true mu is not equal to 0 

-- 
David Winsemius

> I want to apply a Wilcoxon test on subsets of the data frame mydata,
> splitted using the myindice variable. When I send :
> 
> wilcoxtest <- by(mydata, mydata$myindice, function(x)
> {wilcox_test(x$value~x$fact)})
> 
> I get :
> 
> Error in eval(expr, envir, enclos) : object "x" not found
> 
> whereas 
>
> wilcoxtest <- by(mydata, mydata$myindice, function(x)
> {wilcox.test(x$value~x$fact)})
> works !
> Why does it fail with the wilcox_test only? And, most important, how
> can I apply the wilcox_test function to subsets of mydata splitted
> according to mydata$myindice ?
>