I'll have a bash...
On 14 Mar 2007, at 14:42, Ged Ridgway wrote:
> Hi all, especially Tom,
>
> It's not clear to me from p9 of [1]:
>
> "The set of N possible relabelings is reduced to a
> more manageable N' consisting of the true labeling
> and N-1 randomly chosen from the set of N 1 possible
> relabelings."
>
> whether the random selection of a subset should be with-replacement
> or without. Intuitively (but not at all knowledgeably!) I would
> have thought that without-replacement would give a better
> approximation of the null distribution for a given number of
> samples. But perhaps not?
We do it without replacement - I don't think it matters too much but
that's the way we do it, under uncle Tom's guidance.
> It seems to me from a quick glance at the code for randomise that
> both the single-sample sign-flipping and the more general
> permutations, are randomly generated each time, with no memory of
> previous ones, and hence the sampling is with-replacement. Is this
> correct, or am I missing something?
You're missing something :)
Cheers.
>
> Many thanks,
> Ged.
>
> [1] Thomas E. Nichols and Andrew P. Holmes.
> "Nonparametric Permutation Tests for Functional Neuroimaging: A
> Primer with Examples."
> Human Brain Mapping 15:1 25(2001)
------------------------------------------------------------------------
---
Stephen M. Smith, Professor of Biomedical Engineering
Associate Director, Oxford University FMRIB Centre
FMRIB, JR Hospital, Headington, Oxford OX3 9DU, UK
+44 (0) 1865 222726 (fax 222717)
[log in to unmask] http://www.fmrib.ox.ac.uk/~steve
------------------------------------------------------------------------
---
|