How does REML estimate correlation in Multivariate Methods with missing values?
Sep 24, 2015 6:50 AM(1222 views)
I'm trying to understand how JMP estimates correlation and covariances for data with missing values using REML in the Multivariate Methods toolbox. I read the user guide, but they don't provide details much details. Do they use imputation? Do they drop the values? Do they use the mean or other estimation? Thanks in advance.