Selecting Statistical Procedures for Multi-Level Data: Examining Relationships between Principal Authenticity and Teacher Trust and Engagement

  • Chuang Wang University of North Carolina at Charlotte
  • James J. Bird
Keywords: unit of analysis, conceptual fallacy, multi-level data analysis, hierarchical linear models.


Using a case study of the relationship between principal authenticity and teacher trust and engagement, this article presents the theoretical rationale and statistical procedures to use hierarchical linear models (HLM) to analyze multi-level data. The dilemma of unit of analysis when working with multi-level data, the estimation method associated with HLM, and the conceptual fallacy during the interpretation of relationships at the individual level based upon aggregated data are also discussed.
Methods Forum Articles