Multilevel modelling has transformed social science research across a wide range of disciplines, including health sciences, economics, education and political studies.
No man is an island – we are part of a wider world and a complex society. When studying human behaviour and social interaction researchers don’t only focus on the individual, but also on the surrounding social context. For instance, how well a pupil performs at school is affected by a range of factors including classroom environment, school type and educational system.
These different factors form a ‘multi-level’ structure – a hierarchy where factors on a higher level (environment and context) influence factors on a lower level (individual people). It can be seen as a hierarchy with different levels - the pupil at the individual level, followed by ever-expanding levels such as class, school, local education authorities and national educational system.
Where traditional statistical modelling had a tendency to focus too much on the individual, multilevel modelling emphasises both individuals and their social contexts at different levels - enabling the researcher to build a more realistic model.
Modern multilevel modelling emerged in the 1970s and 80s, and has become one of the basic techniques for modelling data with complex hierarchical structures.