# Bio-Env vs BVStep

BEST amalgamated the earlier (PRIMER 5) BIOENV and BVSTEP procedures (hence BEST = <u>B</u>io-<u>E</u>nv + <u>St</u>epwise) since they had an identical purpose – to search for high matrix correlations, rank-based, between a fixed sample <u>similarity</u> matrix (typically from a species assemblage) and resemblance matrices generated from different variable subsets of a supplied <u>data</u> matrix (usually a transformed and normalised suite of environmental variables presumed to include those ‘driving’ the assemblage structure). The only difference in operation is that BIOENV carries out a complete search of all possible combinations of variables from the datasheet, whereas BVSTEP caters for the common situation in which there are too many variables to do an exhaustive search, and a forward-stepping and backward-eliminating stepwise procedure is necessary to arrive at a (possibly) optimal set. Within **Analyse>BEST**, the first choice is therefore of Method•BIOENV or Method•BVSTEP. (BVSTEP will be discussed in Section [14](https://learninghub.primer-e.com/books/primer-v7-user-manual-tutorial/chapter/14-further-matching-of-multivariate-patterns-relate-2stage-best-mvdisp), where it becomes essential for use on biotic matrices).