Hi! A reviewer of my paper requested a correlation table in which missing data are treated with FIML. So far, I have a correlation table with manifest variables and casewise deletion (from SPSS). Instead of a pairwise deletion, I should now use FIML (but not a listwise deletion). I tried to estimate the correlations. However, the correlations in Mplus differ between those requested by "sampstat" under "output" and those requested by "with" under "model". How can I estimate the bivariate correlations of my manifest variables in mplus with fiml (the estimator is ML)? Thanks!