Co-Principal Investigator: Steve Raudenbush (University of Chicago)

Researchers will work to develop methodology and user-friendly software that is broadly accessible to education researchers for unbiased and efficient analysis of multivariate hierarchical models with missing data. This project will expand the work completed in a previous IES grant (Development of Accessible Methodologies and Software in Hierarchical Models with Missing Data) on software to handle missing data in univariate two- and three-level multilevel. The current project will expand the existing univariate software to handle binary and ordinal missing data. In addition, researchers plan to develop the multivariate software to handle binary and ordinal missing data.

The hierarchical models to be addressed by this project can have outcome variables defined at single or multiple levels that are discrete or continuous, or both. They may also involve auxiliary variables that are not of direct interest but are highly correlated with outcomes and covariates that could be missing data. Covariates and outcomes as well as auxiliary variables may be subject to missing data with a general missing pattern at any level of the model.

A key feature of the final software product will be that it will require users only to know and to input the model they intend to analyze, with the rest of the analysis steps automated by the software. Additional products of the grant include peer-reviewed journal articles, conference presentations, and workshops for training researchers on the use of the new software.

Related IES Projects: Development of Accessible Methodologies and Software in Hierarchical Models with Missing Data (R305D090022)

Products and Publications

Book chapter

Shin, Y. (2013). Efficient Handling of Predictors and Outcomes Having Missing Values. In L., Rutkowski, M., von Davier, D. Rutkowski,(Eds.), A Handbook of International Large-Scale Assessment Data Analysis: Background, Technical Issues, and Methods of Data Analysis (pp. 451–479).