Hi,
It sounds like you have a non-significant result, probably because of the low number of subjects.
This is just the result that you have, and there is no "z-statistic" option as randomise is giving you a full non-parametric statistical test and any parametric test (that would normally involve z-statistics) will not be valid.
You are only likely to improve your result by getting more data. Depending on the type of study, you can report non-significant values as trends, but this p-value is really very weak, although fairly typical of only having a single subject in one group.
All the best,
Mark
On 2 May 2013, at 13:47, Niyati Roychowdhury <[log in to unmask]> wrote:
> Dear FSL experts,
> I am a masters student and new to FSL. I have a DTI data from 3 healthy subjects and 1 patient.
> When I performed my TBSS analysis on these 2 groups, I am getting some difference in 2 areas but only after adjusting the p value at 0.25 (i.e 1-p at 0.75).
> I had 2 questions..
>
> 1) This result is not statistically significant, so should I use z statics, since I have only 1 patient against more healthy subjects? If I need to use z stats, could you tell me how can I use it?
> 2) Does this data make sense if I find difference at p value 0.25?
> I will have more healthy subjects in coming days, but only 1 patient data.
>
> Your help will be greatly appreciated.
> Thanks,
> Niyati
|