Presented by Sandro Leidi & James Gallagher
This course is running online, via Zoom.
Mixed models are a modern powerful data analysis tool to analyse clustered data, typically arising in studies where the levels of a factor are a random selection from a wider pool, or in the presence of a multi-level nested structure with different levels of variability.
Potential benefits of mixed models are greater generalisability of results and accommodation of missing values. In particular, mixed models have been used in clinical trials to analyse repeated measures, where measurements taken over time naturally cluster according to patient.
The course will illustrate medical and health related applications of mixed modelling, such as multi-centre trials, cross-over trials, and the analysis of repeated measures.
The course focuses on the linear mixed model, assuming normally distributed data, and on how to fit it and interpret its results.
Only essential theoretical aspects of mixed models will be summarised.
Day 1
Session 1
- Random effects and variance components: concept of random vs fixed effects.
- The variance components model for analysing clustered data.
- The -mixed- command and its options.
Session 2
- Blocking: fixed or random?
- The mixed model for the analysis of incomplete block designs.
- Comparison with a fixed effect analysis; benefits of considering a block effect as random.
Session 3
- Modelling hierarchies: multi-level models for data whose design has multiple levels of variation in a nested structure.
- incorporation of fixed effects.
Session 4
- Maximum likelihood and REML methods for fitting a mixed model.
- Model selection: inference for fixed effects, problems in small unbalanced datasets and the Kenward-Roger method.
Day 2
Session 5
- Model checking: use of residuals to check the assumptions of a mixed modelling
Session 6
- Multi-centre trials.
- Analysis of data from a clinical trial that follows a single protocol but is conducted at several medical institutions.
- Comparison with a fixed effects analysis.
Session 7
- Cross-over studies
- Modelling data from a classic design in which subjects are allocated a sequence of treatments.
- Advantages over a fixed effects analysis.
Session 8
- Analysis of repeated measures data.
- The random coefficients model: motivation, definition and use.
- The marginal model: selection of covariance patterns for the correlation between successive measurements.
It is assumed that delegates are Stata users and are familiar with the practical use of linear models, covering both regression and ANOVA models.
Pre-course reading: West BT, Welch KB, Galecki AT (2014) Linear Mixed Models. A practical guide using statistical software. 2nd edition, CRC Press.
Post-course reading: Brown H, Prescott R (2015) Applied Mixed Models in Medicine. 3rd edition, Wiley.
Terms & Conditions
- Student registrations: Attendees must provide proof of full time student status at the time of booking to qualify for student registration rate (valid student ID card or authorised letter of enrolment).
- Additional discounts are available for multiple registrations.
- Delegates are provided with temporary licences for the principal software package(s) used in the delivery of the course. It is essential that these temporary training licenses are installed on your computers prior to the start of the course.
- Payment of course fees required prior to the course start date.
- Registration closes 1 calendar day prior to the start of the course.
- 100% fee returned for cancellations made more than 28-calendar days prior to start of the course.
- 50% fee returned for cancellations made 14-calendar days prior to the start of the course.
- No fee returned for cancellations made less than 14-calendar days prior to the start of the course.
The number of attendees is restricted. Please register early to guarantee your place.