Development of software for the assessment
of multilevel structural equation modelling
Social context characteristics like classroom climate, classroom management and the social composition of a school or class are key factors in school learning results. However, the evaluation of these relationships poses some methodological challenges. This is because context characteristics in educational research are often to be found in a multilevel structure. The aforementioned factors therefore need to be observed on several levels. If students are from various classes, they are taught by different teachers, for instance. So a multilevel analysis is needed to be able to relate the context factors to student performance. The project addresses this aspect and aims to improve the evaluation of such multilevel data.
Established multilevel structural equation models are very well suited for the purpose of improving data evaluation. However, the project is also aimed at developing a new approach, focusing on how these models can be employed when dealing with data that is difficult to evaluate, such as cases in which only a few different classes are surveyed. Bayesian statistical methods like the Markov chain Monte Carlo (MCMC) method, for which special software is to be programmed, will be applied. This programme is going to be available for free (package in R) so that other researchers can use it.