This is an advanced exercise which considers how nonignorable missing data should be taken into account in the analysis of growth in mathematics achievement over grades 7, 8, 9, and 10 in U.S. public schools for a sample of 3,102 students. The data are from the Longitudinal Study of American Youth (LSAY). The data structure is multilevel with students clustered within schools, but for the purpose of this assignment this complication can be ignored. The variables in the attached LSAY data set are shown in the attached Mplus input for a BASIC analysis.
The exercise consists of using Mplus to carry out an analysis of nonignorable missing data using a pattern-mixture approach to growth modeling of math in grades 7 – 10. For advanced modelers familiar with latent class (mixture) modeling, the Roy (2003) latent class pattern-mixture approach (see reference below) can also be used. Compare the results to a regular MAR growth analysis. This exercise may be approached by including binary missing data indicators in the analysis. Such nonignorable missing data modeling is facilitated by the Mplus DATA MISSING command (see the Mplus Version 5 User’s Guide at www.statmodel.com), which automatically creates the missing data indicators. Latent classes measured by the missing data indicators can be specified for the patterns of missing data.