Evaluation of Multiple Imputation for Longitudinal Ordinal Data under MCAR and MAR Missing-Data Mechanisms

[[abstract]]Multiple imputation can be used to solve the problem of missing data that is a common occurrence in longitudinal studies. An imputation strategy proposed by Demirtas and Hedeker (Statistics in Medicine 2008; 27, 4086-4093) is to deal with incomplete longitudinal ordinal data, which converts discrete outcomes to continuous outcomes by generating normal values, employs multiple method based on normality, and reconverts to binary scale as well as ordinal one. The performance of multiple imputation in terms of standardized bias, root-mean-squared error and coverage percentage under missing completely at random (MCAR) and missing at random (MAR) was discussed by various configurations. The simulated results indicated this mutation strategy is suitable for most of incomplete data under these two missing-data mechanisms.[[journaltype]]國外[[incitationindex]]EI[[ispeerreviewed]]Y[[booktype]]紙本[[countrycodes]]JP