Comparing to zero

[Eric Maris]

Hi Litvak (and others),


> What would be the best way to find clusters that are significally
> different from zero (or any other constant value for this matter)? I once
> saw in SnPM of Nichols an option to have a single input with the
> permutations done my toggling the sign, but there is no such option in the
> current version of clusterrandanalysis.

As far as I know, there is no randomization/permutation test by means of
which
one can test the null hypothesis that some random variable has expectation
zero.
The randomization tests that are implemented in clusterrandanalysis all test
null
hypotheses about the effect of an independent variable. Now, if you have a
two-level
within-subjects design, then the null hypothesis of exchangeability between
these two levels (our specification of the null hypothesis of no effect) is
equivalent
to the following null hypothesis: the probability of the difference between
the two
levels is equal to the probability of minus this difference. In this case,
randomly
toggling the sign of the difference produces the randomization/permutation
distribution. However, if you want to produce this randomization
distribution using
clusterrandanalysis, you have to feed it the data of both levels (and not
just their
difference).


Eric Maris