A Note on Null Models: Justifying the Methodology

The use of random, or null, models to determine whether species co-occurrences differ from chance expectations is an established technique in community ecology. Several methods have been used to draw conclusions regarding species co-occurrence patterns. We compared a sequential method derived previously with a recently proposed new method. Despite similarities in the methods, significant mathematical differences exist. The two methods produce significantly different results on a randomly generated example. While one method was supported by mathematical theory, the proposed new method cannot be justified mathematically because the statistical test was based on sequential steps that were not independent.