Answers are grounded in this paper and cite their sources.
Principal component scores are used widely in summarizing information from ecological data sets, but little attention has been given to the scaling methods producing them. We describe the most common scales and how their properties can contribute dramatically to pattern interpretation. We applied morphological null models to present a case where a method commonly applied by community ecologists shows contradictory results (rejection or not) depending on the particular scaling choice in producing PC scores. Our intention is not to condemn the use of PC scores but to call attention to the fact that different scaling methods emphasize different aspects of the data.The contradictions in results found by our null models, and possibly in the literature, are simply the consequence of different hypotheses being tested. In order to provide general guidelines, we discuss the adequacy of different scaling methods when analyzing particular ecological situations.
No takes yet. Share an insight, caveat, or question.
Peres‐Neto et al. (2001) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: