Mathematical Statistics [chapter]

2013 Mathematical Statistics and Stochastic Processes  
A general linear model has a response variable and a number of possible explaining variables. The explaining variables can either be fixed effects that can be estimated or random effects that come from a distribution. Often a study include both fixed and random effects and the model fitted is then called a linear mixed model. When an effect is included as random the measurements within the same effect can not be considered independent and the correlation between measurements has to be
more » ... in some way. This thesis is a study of mixed models and their use in repeated measurements. An example of repeated measurements is a cross-over study where at least two different treatments are given to each individual. The individual effects can then be included in the model but since the patients will probably be a random sample from a bigger population the individual effects can be fitted as random. We will look at a cross-over study where the measurements are repeated within periods and within days within periods and how we can use mixed models to analyse data in this study.
doi:10.1002/9781118562024.part1 fatcat:64ujgklhebfs5mzeo5z7jmntgm