[R] significant test

[Prof Brian Ripley]

On Mon, 31 Oct 2005, Romain Francois wrote:

> Le 31.10.2005 11:22, Robert a écrit :
>
>> I have two groups of data and want to test how the
>> mean of one group is significant different from the
>> mean of the other group of data.
>> which R function can be used?
>> Thanks.
>
> t.test if you assume normality
> wilcox.test otherwise

Sorry, no.  The Wilcoxon test does NOT test a difference in means: its 
null hypothesis is that the two samples came from the same continuous 
distribution, a much narrower assumption.  (It is sensitive to differences 
in variances, for example, and is probably closer to testing a difference 
in medians than means where the shapes of the two samples differ)

The Welch two-sample t-test does test exactly the null hypothesis stated 
under normality, and it is pretty insensitive to quite large departures 
from normality.  It is unlikely (but possible) that means are of primary 
interest in very-non-normal distributions.

-- 
Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272866 (PA)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595