Dear R-ians, I am working with princomp and I now want to manually recalculate my original values. I want to do it to completely understand the procedure of principal components. I tried it with a test data set (2 dimensions) and I was able to calculate my original values (of a random point of my dataset) using the output of princomp: test.data$scores[1,]%*%matrix(data=test.data$loadings,ncol=2)+test.data$center it seemed that result is identical with test[1,] When I tried it with my 4 dimensional real dataset, it did not work: pca.data$scores[1,]%*%matrix(data=pca.data$loadings,ncol=4)+pca.data$center is not equal to data[1,] Do I misunderstand the procedure somewhere, or what else do I do wrong? Thanx in advance and kind regards, Stef