Introduction to Analysing Repeated Measures Data
Training session with Dr Helean 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:
Non-normal data
- Mixed models based on adaptions to fixed effects approaches available:
--- Logistic regression
--- Poisson regression
--- Ordinal logistic regression
--- Generalised linear models

Example: Repeated Measures Data With a Non-Normal Distribution
- Trial of anti-convulsant therapy for epilepsy versus placebo
- Run-in period then 4 x 2 week visits post-treatment
- Outcome: Number of seizures
--- A count outcome suggests a Poisson distribution
Model:
- General linear mixed model (GLMM) with log link function
- Assess various correlation patterns for repeated measurements…

Models with different correlation patterns

Model 6 with and without outlier

Consider Analysing as Ordinal Data
Categorical outcome formed from seizure counts:
1 = 0 attacks
2 = 1-3 attacks
3 = 4-10 attacks
4 = 11+ attacks

Software