Multidimensional correlations, with the verification of assumptions

Although Prof. Novák demanded the calculation of correlations, it is really more appropriate to use multidimensional analysis – namely the canonical correlation analysis because there will be too many correlation coefficients (4 x 8 = 32) to be interpreted easily and, moreover, the p-values calculated for these correlation coefficients would have to be corrected for a multiple comparison. In addition, the correlation coefficients of each variable from one group with each variable from the other group cannot capture internal correlations between the variables within each group.

However, you must verify the assumptions before performing the canonical correlation analysis.