Multilevel modelling is a development of a common statistical technique known as ‘regression analysis’. It explores the relationship between a measure of interest (‘dependent variable’) and the values of one or more related measures. For example, we may wish to predict institutions’ average test performance in literacy given some background factors, such as size of student body and location (these are sometimes called ‘independent variables’).
Multilevel modelling takes account of data which is grouped into similar clusters at different levels, such as individual pupils grouped within schools. Incorporating this hierarchical structure into our analysis improves the accuracy of its findings, and avoids drawing false or misleading conclusions from the data.
NFER statisticians have been using multilevel techniques since their development over twenty years ago. They have since maintained their position as leading experts, applying and interpreting results from analysis in a range of educational settings.
Please contact Simon Rutt if you would like to discuss how NFER could help you with multilevel modelling.