Course content
Time-use epidemiology is a multidisciplinary field of research, defined as “the study of determinants, incidence, distributions, and effects of health-related time-use patterns in populations and of methods for preventing unhealthy time-use patterns and achieving the optimal distribution of time for population health.”1
Durations and patterns of sedentary behavior, physical activities, and sleeping refer to the use of time. Thus, they conceptually fall under time-use research. They are also parts of a finite whole, typically expressed in terms of percentages of a 24-hour day or some other reference total.2–5 Consequently, proportions of time spent on specific activities during the day represent relative information and are intrinsically co-dependent and collinear.6 That is, more time spent on one activity will necessarily replace time spent on at least one other activity. Analysis of constrained data forming parts of a whole, known in the literature as compositional data, requires particular procedures; compositional data analysis (CoDA).6,7
CoDA is well-known within several research fields, including geology, economy, chemistry, genetics and nutrition. However, CoDA has only recently been introduced within the field of physical activity epidemiology. Pedišić presented the rationale for using CoDA with physical activity data in 2014,8 and since then, a few studies have used CoDA to analyze and test data on sedentary behavior, physical activity and sleep.9–12
The present course will give an introduction to the theory of compositional data, the procedures used for analyzing data in CoDA, and the interpretation of results obtained by CoDA. The course will have a focus on practical hands-on aspects of CoDA, addressing explorative and multivariate analyses, including regression techniques, cluster analysis and ANOVA. The course aims at enabling students with an interest in time-use epidemiology related to sedentary behavior and physical activities in occupational and public health to use CoDA in future studies. Participants will not need any particular mathematical background, but they are recommended to have completed a basic statistical course, including multivariate methods.
Expected learning outcomes
Theoretical and practical knowledge of compositional data analysis at a level that makes you able to analyse own 24 hour physical activity data.
Course fee
Free of charge for PhD students from the Faculty of Health Science, SDU.
For PhD students from other Danish Universities that have joined the "Open market agreement" the course fee is DKK 1180-,.
For other participants the course fee is DKK 7211.25-,.