Multilevel models (MLM), also known as hierarchical linear models, are statistical techniques used to analyze nested data, data structured on multiple levels, and longitudinal data. Traditional methods like regression or analysis of variance require independent observations to meet their underlying statistical assumptions. However, MLMs can handle situations where these assumptions are violated by incorporating random effects into the analysis. 

A common example of MLM usage is in analyzing student achievement. Here, students (Level 1) are nested within classrooms (Level 2), which are further nested within schools (Level 3). MLMs allow researchers to simultaneously consider factors at multiple levels, such as student-level demographics and school-level characteristics. 

Similarly, longitudinal data can be analyzed using MLMs. For instance, in a study where patients take different blood pressure medications and have their blood pressure recorded weekly for ten weeks, the assumption of independent observations is violated since multiple measurements come from each individual. MLMs account for this using a technique called a growth curve model. 

How to Request Help with MLM 

OIT Research Computing Support statisticians have extensive experience performing multilevel models (MLMs). If you need assistance performing MLMs for your research or would like to meet with a statistical consultant, please contact the OIT HelpDesk by calling 865-974-9900 or visiting help.utk.edu.