11.Repeated Measures Data

Introduction to Mixed Models
Training session with Dr Helen Brown, Senior Statistician, at The Roslin Institute, March 2016.
************************************************
These training sessions were given to staff and research students at the Roslin Institute. The material is also used for the Animal Biosciences MSc course taught at the Institute.
************************************************
*Recommended YouTube playback settings for the best viewing experience: 1080p HD
************************************************
Content:
Repeated measures data
Repeated measurements on same individual or ‘item’
--- Eg repeated measurements on crabs of consumption of different diets
Multiple measures on the same crab are not independent

Repeated measures data – Possible approaches
Analyse one measure for each individual
--- Mean
--- Area under curve
--- Maximum value
--- Time to maximum value
--- One chosen timepoint
Analyse all timepoints separately and adjust for multiple testing
If data are complete could use a GLM with adapted test for comparing treatments
--- Allows test of treatment, time and treatment*time interaction
--- Data need to be complete
Use a mixed model:
--- Suitable for complete or incomplete data

Mixed model for repeated measures data
Repeated measurements on same individual or ‘item’
--- Eg repeated measurements on crabs of consumption of different diets
Crab may be fitted as a ‘random’ effect
--- Account now taken of correlation between multiple measurements on the same animal
Pattern of correlation may be defined for Error terms
Model suitable even if some data are missing

Alternative approach for repeated measures data: Random coefficient/slopes model