Longitudinal and Correlated Data Analysis
Enables students to apply appropriate methods to the analysis of data arising from longitudinal (repeated measures) epidemiological or clinical studies, and from studies with other forms of clustering (cluster sample surveys, cluster randomised trials, family studies) that will provide non-exchangeable outcomes.
This course is offered in conjunction with the Biostatistics Collaboration of Australia (BCA).
Distance Education - Callaghan
- Semester 1 - 2015
1. Recognise the existence of correlated or hierarchical data structures, and describe the limitations of standard methods in these settings
2. Develop and analytically describe an appropriate model for longitudinal or correlated data based on course matter considerations
3. Be proficient at using a statistical software package (eg Stata or SAS) to properly model and perform computations for longitudinal data analyses, and to correctly interpret results
4. Express the results of statistical analyses of longitudinal data in language suitable for communication to medical investigators or publication in biomedical or epidemiological journal articles.
The concept of hierarchical data structures will be developed, together with simple numerical and analytical demonstrations of the inadequacy of standard statistical methods. Beginning with the normal-theory model, more appropriate statistical procedures involving mixed linear models will be developed and explored using the SAS or Stata statistical software packages. Extensions to non-normal outcomes will be demonstrated in which emphasis will be placed on the clinical research question. Using a set of case studies, generalised estimating equations and generalised linear mixed models will be developed and contrasted. The limitations of traditional repeated measures analysis of variance will be demonstrated, together with an introduction to non-exchangeable models.
This course is only available to students enrolled in the Graduate Diploma in Medical Statistics or Master of Medical Statistics programs.
Epidemiology (EPID6420); Mathematical Background for Biostatistics (BIOS6040); Probability and Distribution Theory (BIOS6170); Principles of Statistical Inference (BIOS6050); Linear Models (BIOS6070); Categorical Data; GLMs (BIOS6020).
Written Assignment: Essays / Written Assignments
Self-Directed 6 hour(s) per Week for Full Term
Suggest 8-12 hours of time commitment per week (guide only).