A comparison of standard and compositional data analysis in studies addressing group differences in sedentary...

Area of Science:

• Occupational health
• Biostatistics
• Physical activity epidemiology

Background:

• Time spent in physical activity, sedentary behavior, and sleep are compositional data, summing to 24 hours.
• Standard statistical methods applied to compositional data can yield misleading results.
• Compositional data analysis (CoDA) offers appropriate analytical procedures for such data.

Purpose of the Study:

• To compare time spent sedentary and in physical activity between age groups and sexes.
• To investigate the differences in results obtained using CoDA versus standard analytical procedures.
• To highlight the importance of using appropriate methods for analyzing compositional data in physical activity research.

Main Methods:

• Accelerometry was used to measure time spent sedentary, standing, and in physical activity (walking, running, stair climbing, cycling) during work and leisure in 677 blue-collar workers over 1-4 days.
• Differences between sexes and age groups were analyzed using MANOVA with both standard and CoDA approaches (isometric log-ratio transformed data).

Main Results:

• CoDA yielded a 15% larger effect size than standard analysis for sex differences in work-time activities, with both methods showing statistical significance.
• For age group differences during work, CoDA produced a 60% larger significant effect size compared to the non-significant standard analysis.
• During leisure time, both standard and CoDA analyses showed similar results for age and sex differences.

Conclusions:

• Inferences regarding age and sex-based differences in sedentary time and physical activity at work varied significantly between CoDA and standard analysis.
• The study recommends the use of CoDA for studies involving compositional data on physical activity and sedentary behavior to ensure accurate interpretation.
• Accurate analysis of physical activity and sedentary behavior data is crucial for understanding health-related differences across demographics.