I have collected some data from different subjects in 2 different ways. In the first way, each subject performed one long trial, I segmented the data captured and and calculated a variable value for each segment (15 in total, for each subject). In the second way, each subject performed 15 short trials and I calculated the variable value for each trial (again, 15 in total for each subject).
Now I want to comment on the variability of the first way, vs the second way. I can't use standard deviation since each subject's mean will be different. Could I use coefficient of variation(cv) and/or ICC? If I were to use ICC, how would I go about that, and what would it provide that cv would not?