Analysing time-use composition as dependent variables in physical activity and sedentary behaviour research:...

Area of Science:

• Health Sciences
• Data Science
• Behavioral Science

Background:

• Physical activity and sedentary behavior are increasingly viewed as a compositional interplay, not independent health risks.
• Compositional data analysis (CoDA) is emerging as a key methodology in this research area.
• However, analyzing time-use composition as dependent variables remains under-explored due to interpretation challenges.

Purpose of the Study:

• To present and discuss four distinct statistical approaches for analyzing time-use composition as dependent variables.
• To provide guidance on selecting appropriate data analysis methods based on research objectives.
• To encourage the adoption of compositional analysis for time-use data in physical activity and sedentary behavior research.

Main Methods:

• Exploration of four statistical methodologies for compositional dependent variables.
• Discussion of interpretation challenges associated with these methods.
• Guidance on aligning analytical approaches with research aims.

Main Results:

• Identified four distinct approaches for analyzing time-use composition as dependent variables.
• Highlighted the importance of research aims in defining dependent variables and selecting statistical methods.
• Addressed challenges in interpreting compositional data analysis results.

Conclusions:

• The choice of statistical approach should be driven by the research question when analyzing time-use composition.
• Researchers are encouraged to consider time-use components as dependent variables for a more holistic understanding of health behaviors.
• Further research and methodological development are needed to facilitate the use of CoDA in this domain.