Multivariate analysis of ecological communities

P. G. N. Digby & R. A. Kempton

The development of statistical methods in field biology has, until recently, been dominated by the requirements of agricultural scientists working with highly controlled systems of experiments. Here the experimental factors (e.g. plant variety or nutrient level or site) generally cover only a limited range, while the variation in uncontrolled factors is made as small as possible, for example, by spraying to control disease. In consequence, the range of measured responses and the unexplained error in those responses are also fairly small. This has led to the development of a large statistical theory in which the pattern of response, possibly after transformation, is described by the effects of addictive factors and a residual response whose distribuition approximates to some known mathematical form (e.g. a normal distribution).

Fonte: Digby, P. G. N. & Kempton, R. A. 1987. Multivariate analysis of ecological communities. Londres, Chapman.