ABSTRACT
These two objectives are typically in conflict. For example, as we saw in Chapter 2, if all responses have the same variance, then precision of estimation and the power of the associated tests are optimised by designs allocating equally to all treatments.
