[R] about the 95%CI around the median...

[Rolf Turner]

On 5/08/2008, at 1:31 AM, Frank E Harrell Jr wrote:

	<snip>

> I wonder why we don't just use the exact nonparametric confidence  
> interval for the median, which is just as easy to compute.  Also,  
> it will be asymmetric if the data are skewed, as it should be.

	<snip>

The boxplot.stats() function gives the same ``confidence interval''  
for the median as the
boxplot() function. This isn't however *really*  a confidence interval.

The help on boxplot() (see the ``notch'' argument) indicates that  
this ``confidence interval''
actually provides a rough *hypothesis test* of the equality of *two*  
medians. The null hypothesis
is rejected if the notches (confidence intervals) don't overlap.

	cheers,

		Rolf

######################################################################
Attention:\ This e-mail message is privileged and confid...{{dropped:9}}