Mixed linear models include both fixed effects and random effects and hence allow flexible modelling of longitudinal and repeated measures data. The important issues of statistical inference about mixed models comprise the estimation of fixed effects, estimation of variance components, and prediction of the random effects and/or a linear combination of fixed and random effects. The prediction problems have found applications e.g. in small area estimation and animal breeding. In this talk, an overview about the results concerning prediction in univariate mixed linear models will be given.  Further, the concept of equivalent mixed linear models will be discussed.

When: September 21, at 1-2 pm

Where: Room B705, Department of Statistics