Raw data graphing: an informative but under-utilized tool for the analysis of multivariate abundances

In exploring the relationship between multivariate abundance data and environmental variables, a rarely used approach is to graph raw data separately for each different taxon. It is proposed that such raw data graphs become part of the standard toolset for graphing and analysing multivariate abundances. The key advantage of this approach is that axis scales have quantitative interpretations, enabling quantitative interpretation of patterns in abundance. In contrast, ordinations only present qualitative information. Ordinations are useful for inferring overall, qualitative patterns and raw data graphing is a complementary tool of greater use for answering more specific questions, aimed at a deeper understanding the ecology of a community. It is demonstrated using some well-known examples that our understanding of the nature of associations can be considerably improved by using raw data graphs, even when only plotting a subset of variables. One example describes how an often-cited dataset has been misinterpreted in key methodological papers, because data were interpreted from ordinations alone, with no consideration of plots of the raw data.