Pair t-test with non-independent samples

#1
Dear list,

In my experiment, each subject of my sample has three time-series of data: tc_A, tc_B and tc_C. Using Pearson correlation within subject I calculate correlations between time-series: tc_A vs. tc_B, tc_A vs. tc_C and tc_B vs. tc_C. So, for each subject I get three correlation values. My research question is whether correlation between tc_A and tc_B is different than correlation between tc_A and tc_C. The question is whether it would be correct to use a paired t-test across subjects for correlation values to test that. I am asking because tc_A time series participated in both correlation pairs (i.e., tc_A vs. tc_B and tc_A vs. tc_C), so the two samples are not independent.

Your help will be really appreciated!

John
 
Last edited: