Statistical methods for compensating for missing longitudinal data in a cluster randomised trial
SWAT: 52
Objective of this SWAT:
To investigate whether multiple imputation or maximum likelihood-based methods perform better as a statistical approach to handling missing data in a longitudinal cluster randomised trials. To use real data, calibrated by simulations, to determine the relative power and type I error of the trial, and to identify biases that arise in effect estimates.
Status:
In Progress
Resources:
Please click the links below to access the registered protocol or SWAT download bundle which includes: 1) SWAT Abstract
