I have a 3-factor CFA model with 16 items. I conducted 8 groups comparison.For one group, Amos analytic result showed " The following covariance matrix is not positive definite." The sample size for this group is n=31 (1) Might the "not positive definite problem" happen because the sample size is too small, n= 31?(2) How to fix the about question in order to accomplish 8 groups comparison? Thank you very much and I look forward to hearing from you.