Dear all, In the context of maximum likelihood factor analysis, it often construct the likelihood ratio test for the following hypothesis: H0: covarianace matrix is a m-factor model H1: a general covariance matrix to select the number of factors (m). I would like to know whether the following likelihood ratio test for the following hypothesis is valid or not. H0: m-factor model H1: (m+1)-factor model Sometimes told me that the H0 is not nested in H1 but I cannot figure it out. Thanks. Philip [log in to unmask] %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%