Hierarchical Linear Modeling in Educational Research

Catalog description

Statistical techniques for the analysis of multilevel data such as in nested designs or hierarchical data. EDPSY 557 Hierarchical Linear Modeling in Education Research (3) Hierarchical Linear Modeling (HLM) models are particularly important when analyzing data for school settings. This course is designed as an applied statistics course specifically geared to analyzing data from educational settings and using data sets from educational research. Data collected in these ecological contexts with nested designs, such as students enrolled in classrooms, classrooms in schools, and schools within school districts, must be analyzed carefully as relations between and among variables could change given a particular level (e.g., student-level, classroom-level) for analysis. The topics of this course highlight the importance of studying random versus fixed effects for data collected in multilevel educational research settings. Two-level HLM models, growth-curve models, three-level HLM models, and Hierarchical Generalized Linear Models with binary and ordinal outcomes are the four primary types of models that will be the focus of the class. Students will also learn how to use HLM software to analyze their data given the four types of models. Other topics covered in this class will include: a) centering of independent variables; b) restricted maximum likelihood estimation; c) effect sizes and power analysis; and d) the relevance of educational theory and psychometric analysis in variable selection, and model specification.