ABSTRACT
Multilevel models tackle the analysis of data that have been collected from experiments with a complex design. For example, multilevel models are typically used to analyze data from the students’ performance at different tests. A typical problem in multilevel modeling is assessing the variances of the different group effects. In addition to the grouped or nested random effects, multilevel data may contain values of the covariates at different levels. For example, in the case of measuring students’ performance, socio-economic information about the students is available at the individual level, as well as information about the class, school and region. The interest is on how the production method affects the production of penicillin. However, production methods use a raw material which is quite variable and may affect the production. Multilevel models have traditionally been associated with education research on measuring students’ performance. J. Carpenter et al. describe a multilevel model to analyze the performance of children using literacy and numeracy scores.
