Dear SPMers,
apologies for the following stats-question not directly related to SPM.
I wonder how I would statistically compare effect sizes within one sample. For example, I find volume/activation/metabolism in ROI A to be reduced a significant 10% in a sample of patients compared to the mean of the control group, and ROI B shows a significant 20% reduction. How would I then test if ROI B is significantly more reduced than ROI A.
In previous work I calculated ROC curves and compared the AUC values of the ROIs using ROCKIT software (http://metz-roc.uchicago.edu/), but I wonder if there is a more straightforward way of adressing this question.
Sorry if this is basic stats, but somehow I cannot find a satisfying answer in the vastness of the www.
Thanks in advance,
Michel
|