When the raw data are expressed as distances rather than values at individual points, a correlogram can also be constructed using the Mantel coefficient (Sokal et al. 1987). For this analysis, the Mantel correlation is calculated between the data distance matrix and the weight matrix. In this case, wij contains a 0 if points i and j are connected, and a 1 if they are not. The reversed weighting scheme is used because the data are expressed as distances; therefore high values (i.e., 1) indicate dissimilarity, while low values (i.e., 0) indicate similarity. The Mantel correlation is then plotted versus distance as in any other correlogram.
| Menu: | Analysis 2→Scattered Data→Correlograms→Mantel Correlograms |
| Button: | |
| Batch: | MantelCorrelogram |
Mantel Correlograms window.
Explain options and give example