Using International Databases to Address Substantive Issues in Education, 24-26 June 2014
Multilevel Regressional Modeling

Multilevel Regression Modeling


Multilevel regression models are statistical models designed for the analysis of nested data structures. Nesting arises from cross-sectional hierarchical data structures (e.g., students nested in classrooms, teachers nested within schools), longitudinal data structures (multiple measures nested within individuals), or both (multiple measures nested within students and students nested within classrooms). Nesting is seen in the data from cross-sectional studies such as PIRLS. When data are nested, there are both statistical and substantive advantages to using multilevel regression modeling procedures and the approach provides a framework for exploring the relationships between group or organizational characteristics (e.g., classroom, teacher, or school characteristics) and individual outcomes (e.g., student learning). This institute provides a comprehensive introduction to multilevel regression modeling as it is used to analyze international large-scale data.