You might want to try using a non-parametric test, such as wilcox.test.
How about some modification of the following:
d=data.frame(grp=rep(1:2,e=5),replicate(10,rnorm(100))); head(d)
lapply(d[,-1],function(.column)wilcox.test(.column~grp,data=d))
David Freedman
stephen sefick wrote:
>
> Up a
Up and down are the treatments. These are replicates within date for
percent cover of habiat. This is habitat data for a stream
restoration - up is the unrestored and dn is the restored. I have
looked at the density plots and they do not look gaussian - you are
absolutely right. Even log(n+1) t
stephen sefick wrote:
I would like to preform a t.test to each of the measured variables
(sand.silt etc.)
I am a big fan of applying t.test()s, but in this case: Are you really
sure? The integers and particularly boxplot(x) do not indicate very well
that the variables are somehow close to G
I would like to preform a t.test to each of the measured variables
(sand.silt etc.) with a mean and sd for each of the treatments (up or
down), and out put this as a table I am having a hard time
starting- maybe it is to close to lunch. Any suggestions would be
greatly appreciated.
Stephen S
4 matches
Mail list logo
