Data Aggregation in Multilevel Analysis: A Review of Conceptual and Statistical Issues

Understanding that the behavior of people takes place within a context, over the past 20 years research in education and the sport sciences has witnessed an increasing development of multilevel frameworks that are both conceptually and methodologically sound. Despite these advances, the use of multilevel models and research designs in education still could be improved. As noted by recent editorial commentaries and reviews, confusion often exists about appropriate levels of analysis both at the conceptual and statistical level. The purpose of this article is threefold. First, we explain how multilevel conceptualizations can aid our understanding of behavior, especially within a given context. Second, we review three aggregation methods for analyzing multilevel data: within group agreement, intraclass correlations, and within and between analysis. Third, we provide a conceptual and statistical example to demonstrate how these aggregation techniques could be utilized to answer questions in sport and physical activity.