two sample t-test residual standard deviation

Submitted by David on

Hi,

is it possible to obtain the values of the residual standard deviation from the two sample t-test? I'm trying to look for outliers in my data.

Any help appreciated :)

YAN Chao-Gan

Thu, 02/04/2016 - 09:13

Permalink

I guss a better way to look for the outliers is ploting them with quantile range.

You can get the 4D resicual by turning on the output 4D switch in y_GroupAnalysis_Image.

Best,

Chao-Gan

Thank you! I turned the switch on and get two nifti files after calculating the two sample t-test.

But I'm not sure what to look for now. Should I take a look at high values of residuals and SSE or both? And how to plot the quantile range  for every voxel? 

And do I have to look for outliers in the individual subject data (how?) or outliers in the residuals of the ttest results between the two groups? Not sure how it is done in practice.

Sorry for the many questions, I have never done this before, so any help or paper/link about this appreciated.

Hi,

You can use boxplot on the original values.

I am not sure how you can determine the outliers through the 4D residuals.

Best,

Chao-Gan

Forums
DPABI/DPABISurf/DPARSF