Randomization analysis of replicated complete block designs

摘要

One of the important notions in experimental design is that of randomization. This applies not only to a single execution of an experiment but also to situations where the experiment is repeated. We consider here various forms of a replicated randomized complete block design in an agronomic trial. We are using randomization theory arguments to develop models and associated analyses for the following situations concerning the blocks and types of randomizations within each replication: (i) Distinct blocks, independent randomizations; (ii) distinct blocks, single randomization; (iii) identical blocks, independent randomizations; (iv) identical blocks, single randomization. The emphasis will be on examination of the expected mean squares in the analysis of variance table for each situation and on testing hypotheses concerning treatments and replicate-treatment interactions.