STATIS ordinations vs. the Juhasz-Nagy models: the predictability of an exploratory tool

Exploratory analysis, confirmatory analysis, and modeling are treated as a three-step model of research, each step using particular statistical and data analysis tools. Sampling requires different techniques in each step, and the problem of optimal plot size becomes evident at least in the second step. P. Juhasz-Nagy's model and the STATIS procedure are described for the analysis of three-way data tables. The ability of the former in detecting optimal plot size of a community is discussed, and the use of STATIS as an exploratory tool for the same purpose is suggested to overcome the problem of large sample size required by Juhasz-Nagy's models. -Author