I have 2 sets. I formed 5000 random groups of sizes ranging from 2 to 100 per group from the 1st set (sample size:4991)and then computed the average time series correlation between the groups (i.e the average correlation of the correlations between each pair of groups) . this correlation was 0.27.For the second set (sample size:7511) the average of the time series correlation between groups 0.224. However, when I took the time series correlation between the average correlations of sets 1 and 2the correlation was too high 0.88. Does it make sense?. How come that the within-group correlations are low but the cross-average correlations is high?