ABSTRACT
If in a randomized block design the number of experimental units in a block is less than the number of treatments, then the design is called an incomplete block design. An incomplete block design in such as experiment could result in the following arrangement of data. An alternative method that can be used to analyze data from an incomplete block design is to reparametrize the model to a full rank regression model and then use regression techniques to carry out the analysis. This approach is particularly useful when one does not have access to computer software that handles general linear models. In some cases due to constraints on the design an experimenter will be unable to have every treatment present in each block. In such a case a design that may be appropriate, having hand computable formulas, is a balanced incomplete block design.
