VIDTAXA: An algorithm for the identification of statistically different groups based on variability obtained in factorial analyses

Abstract Factorial analyses are frequently used in ecological studies to identify different groups, which is a valid approach if those variables which show the highest variability also best differentiate among groups. Here we present VIDTAXA, an algorithm designed to identify statistically different groups in Principal Components and Correspondence analyses, without previous knowledge of the various potential groups present in the dataset. VIDTAXA is freely available as a function of the VARSEDIG package, which is available as an RWizard application and as an R package on CRAN. As a demonstration of VIDTAXA's potential, we used the algorithm in an example of phenetic taxonomy, for the identification of statistically different taxa in a number of marine Scorpaeniform species. This algorithm. however, it may be used with any kind of data, whether qualitative or quantitative, for the identification of statistically different groups in ecological studies.