ABSTRACT
The analysis of block designs presented in Chapter 1 was concerned with the estimation of treatment contrasts from comparisons made within blocks. If the blocks can be regarded as a random sample of blocks from some population then estimates of treatment contrasts may also be available from comparisons between blocks. Whether it is worth incorporating the information from this inter-block ana lysis with that from the within-block, or intra-block, analysis will depend on the extent to which the grouping of the experimental units into blocks has achieved a marked reduction in the error mean square. If this grouping is successful, i.e. block effects are large, then the amount of information which may be recovered from the inter block analysis will be small. On the other hand if block effects are small there may be substantial gains to be achieved by recovering the inter-block information.
