Replacement or additive designs for competition studies?
Replacement and additive designs are compared for their ability to provide valid and interpretable measures of (a) resource complementarity, i.e. relative yield total (RYT), (b) competitive ability and (c) severity of competition in binary mixtures. In replacement designs, the density of one component of the mixture is confounded with that of the other, so that competition between plants of one component is confounded with competition between plants of different components; the design if therefore statistically invalid. Because of this confounding, values of RYT, competitive ability and severity of competition in replacement are affected by: the density of each component to density in its pure-stand; the pattern of response of each to density; and the proportions of the components in mixtures; interpretation of these indices is therefore difficult or impossible. By contrast, additive designs give valid and interpretable values of each of the indices of competition, regardless of density or proportions. Various misconceptions about additive designs are considered. The value of binary factorial designs in competition studies is explored and methods of analysing the resulting data are briefly considered.