Related Experiment Videos
[Multilevel model applications to the analysis of longitudinal data].
María Victoria Zunzunegui1, María Jesús García de Yébenes, Mathieu Forster
1Centro Universitario de Salud Pública, Universidad Autónoma de Madrid. vzunzunegui@taiss.com
Revista Espanola De Salud Publica
|June 18, 2004
Summary
This study introduces hierarchical linear models for analyzing longitudinal data with repeated measurements. These models help describe average population changes and individual variations over time, with applications in social sciences and health research.
Area of Science:
- Longitudinal data analysis
- Statistical modeling
- Social sciences research methods
Context:
- Hierarchical linear models (HLMs) are widely used in social sciences for multilevel data.
- HLMs have evolved to analyze change in populations and individual trajectories.
- Understanding within-subject and between-subject variability is crucial for longitudinal studies.
Purpose:
- To introduce a two-stage modeling framework using hierarchical linear models for repeated measurement analysis.
- To explain core concepts including fixed and random effects, and linear/quadratic growth models.
- To illustrate model application with a real-world cohort study on cognitive function in older adults.
Summary:
- The work presents a two-stage hierarchical linear model approach for longitudinal studies.
- Key concepts such as between- and within-subject variability, person-specific trajectories, and growth models are detailed.
- The study demonstrates model fitting using the 'Aging in Leganés' cohort data and discusses extensions to non-linear outcomes.
Impact:
- Provides a foundational understanding of advanced statistical techniques for analyzing change over time.
- Enables researchers to describe average population changes and investigate factors influencing individual variability.
- Offers insights into potential generalizations for dichotomous, nominal, or ordinal outcome data in longitudinal research.