Dear everyone, A small query about PCA. PCA works best if the original variables are highly correlated. However, what if we have 4 variables (say) X1, X2, X3, Z and X4 and X4=X1+X2+X3+Z . We think about using X1 X2 X3 and X4 in a PCA but would the relationship between these 4 variables interfere with the analysis and, if so, how would we remedy this? Perhaps by omitting x4? Kind Regards, Kim