To introduce students to Gaussian linear mixed effects modelling in general and to the active use of modern mixed effects modelling software.

20 lectures (two lectures each week for 10 weeks)

5 tutorials (fortnightly)

2 two-hour laboratory sessions

Stats4045 Linear Mixed Models

Assessment

120-minute, end-of-course examination (100%)

It is the default expectation that all courses will offer opportunities for reassessment or deferred assessment. Where it is not possible to offer this in some assessment components, the grade achieved at the first attempt will be counted towards the final course grade, and any exceptions for this course are described below.

[No exceptions]

To introduce students to Gaussian linear mixed effects modelling in general and to the active use of modern mixed effects modelling software.

By the end of the course students will be able to:

■ explain the notion of a random effect, why and when it is useful and, in particular, how it differs from a fixed effect;

■ demonstrate detailed understanding of the theory underpinning simple mixed effects models for balanced designs;

■ demonstrate general understanding of the theory underpinning general (e.g. unbalanced) mixed effects models;

■ apply their knowledge of mixed effects modelling to practical situations through the use of general mixed effects modelling software such as nlme;

■ check the assumptions of mixed effects models, both graphically and by hypothesis testing based model comparison;

■ explain when to use restricted maximum likelihood (REML) when fitting mixed-effects models;

■ extend the treatment to cover scenarios involving general covariance structures;

■ describe the basic ideas of multilevel models and of generalised estimating equations as applied especially to models for longitudinal data;

■ translate fluently between verbal, mathematical and computational descriptions of models;
critically interpret analyses based on mixed effect models;

■ interpret the output of R procedures for linear mixed-effects models.

■ demonstrate detailed understanding of generalised linear mixed models (GLMMs) and general understanding of other methods for data with discrete responses, eg generalised estimating equations (GEEs).