Randomization analysis for multitiered experiments

Dr Chris Brien with Professor Rosemary Bailey

Some suggest that the only basis for statistical inference in experiments is the randomization employed in them. A theory that uses the randomization as a means of testing hypotheses without making the assumption of normality has been developed for experiments in which there is just a single randomization. We are developing a randomization theory for experiments that involve multiple randomizations. The relationship between the analysis of variance and REML-estimates for multitiered experiments will be described.


Professor Rosemary Bailey, School of Mathematical Sciences, Queen Mary, University of London, UK.


Bailey, R & Brien, C, 2016, 'Randomization-based models for multitiered experiments: I. A chain of randomizations', Annals of Statistics, vol. 44, no. 3, pp. 1131-1164.

Areas of study and research

+ Click to minimise