Course leader
Jonas Björk, professor in Epidemiology, jonas.bjork@med.lu.se
Aldana Rosso, aldana.rosso@med.lu.se

Target group
The course is intended for doctoral and postdoctoral students at the medical faculty with research projects that include repeated measurements of empirical data. All participants must have knowledge in applied statistics corresponding to the courses at level I and II. Access to and substantial skills in one of the following statistical software packages are also required: SPSS, SAS, STATA or R. Ten of the 20 places on the course will be reserved for researches affiliated with the MultiPark or EpiHealth network.

Before you can apply to this course you have to take Applied Statistics I and II OR our former course Statistical methods for Medical research.

Time:

Autumn 2021:

Lectures:

Tuesday: 16 /11  8:45 – 16:00

Tuesday: 23 /11  8:45 – 16:00

Tuesday: 30 /11  8:45 – 16:00

Take home examination: Friday 3/12

The participants have to make time equivalent of two days, for self-studies and work with the home exam. The course workload is 40 hs.

Credits
1,5 h

Number of participants:
20

Location
The course will be given online.

Content of the course
The course is held in English and is based on the following three themes:
 
1) Introduction to repeated measurements
Statistical dependence
Hierarchical data structures
What goes wrong if the repeated nature of the data is ignored in the statistical analyses?
Simple but valid statistical methods based on summarized repeated measurements

2) Introduction to mixed linear regression models
Linear vs. mixed linear regression models for continuous outcome variables
Fixed and random effects
Covariance structures
Model validation

3) More advanced topics
Study design
Handling of time effects
Interaction
Presentation of results
Comparisons with ANOVA for repeated measurements
Missing data

This level III course in applied statistics provides the participant with necessary tools in order to being able to plan, analyze and evaluate empirical studies based on repeated measurements of continuous outcome variables. The course starts with a broad introduction to repeated measurements, problems and common pitfalls, and easy analytical approaches that are valid.
 
Focus is on repeated measurements on the individual level, but other types of clustered measurements will also be discussed, such as families, work places and neighborhoods. The course will emphasize mixed linear regression analysis as a model of the statistical dependence that repeated measurements generate. Some more advanced aspects will also be covered, regarding study design, handling of time effects, interaction, presentation of results, model validation and comparisons with ANOVA.

Course literature
Papers and lecture material will be made available before the course starts and during the course. 

When you apply for the course, please make a note about the statistics package you use, for exampel (SPSS, STATA, SAS or R).