The quantitative ecology of communities: Relationship between some statistical methods
- Plant Physiology Division (Substation), D.S.I.R., Lincoln
In most ecological problems we are endeavouring to define various aspects of plant/animal/environment interactions. The object of taking measurements is to:
(1) find the magnitude of particular effects or interactions and;
(2) to determine their reliability.
Statistical or probabilistic methods are required whenever a decision has to be made whether a particular conclusion can justifiably be deduced from a particular set of data after the variability, chance effects or errors inherent in the data have been considered. Having accepted that, most ecologists may be bewildered by the variety of statistical methods available. The following is an attempt to rationalize the differences between some of the methods and thereby to assist in selecting the one most suitable for a particular problem.