The Guttman effect: its interpretation and a new redressing method

Abstract

The Guttman effect is introduced as a result of factor analyses of band data tables. The problems raised are discussed, in particular with reference to vegetation analysis. Two possible investigation pathways resulted: one aiming at using different analysis techniques, based on the Gaussian model (Ihm and van Groenewoud, 1975; Johnson and Goodall, 1980) and the other based on the redressing of the scatter plots to remove this effect (Gauch and Hill, 1980; Delicado and Aluja, 2003). In this paper, the various methods are described and a new simple redressing method is introduced, based on the interpretation of the scatter plots and their parabolic pattern and some found properties.