Greetings. Probably a basic question. I'm setting up a problem for MCMC multiple imputation in NORM, and ran into a question. Say I have 3 variables, A, B, and C, where A and B are missing at random and C is missing not at random. I recognize that the imputations for C will be a problem, biasing estimates of parameters involving C (its mean, variance, and covariances). Will inclusion of C in the imputation model distort parameter estimates that don't involve C (e.g., the covariance between A and B)? Thanks, Pat Malone -- Patrick S. Malone, Ph.D., Research Associate Duke University Center for Child and Family Policy Durham, North Carolina, USA e-mail: [EMAIL PROTECTED] http://www.duke.edu/~malone/
