13 Comparison of methods of adjustment
The course has given several ways of adjusting a measure of association for a confounder: direct and indirect standardisation (Section 8), inverse-variance pooling (Section 9.3), the Mantel–Haenszel estimators (Section 10) and maximum likelihood in regression models (Section 12). This section applies them all to the same data, explains why they agree when they do and differ when they do not, and ends with guidance on choosing among them.
13.1 Every summary is a weighted average
13.2 A homogeneous example
13.3 A heterogeneous example
13.4 Sparse strata
13.5 Choosing a method
13.6 Problems
13.2 A homogeneous example
13.3 A heterogeneous example
13.4 Sparse strata
13.5 Choosing a method
13.6 Problems
Questions on this section
Stuck on something here? Ask below and it stays attached to this topic.