Chapter 10: Multilevel Measurement Modeling

Multilevel modeling can be utilized for psychometric analyses, and such a use of multilevel modeling techniques is referred to as multilevel measurement modeling (MMM) (e.g., Beretvas & Kamata, 2005). Typically, traditional psychometric models, including classical test theory (CTT) and item response theory (IRT) models, do not consider a nested structure of the data, such as students nested within schools. However, data in educational research frequently have such a nested data structure, especially when data are collected by multi-stage sampling. The strength of MMM becomes important when we analyze psychometric data that have such a nested structure. MMM appropriately analyzes data by taking into account both within- and between-cluster variations of the data. Also, since multilevel modeling is essentially an extension of a regression model to multiple levels, the flexibility of MMM offers the opportunity to incorporate covariates and interaction effects. As discussed in previous chapters of this book, another advantage of a multilevel approach is that it can accommodate unbalanced data, using all of the available information in the data.