STATISTICAL TOOLS TO ANALYZE DATA FROM BIOMONITORING STUDIES

Authors

  • Adriano Sanches Melo Universidade Federal do Rio Grande do Sul
  • Luiz Ubiratan Hepp Universidade Regional Integrada do Alto Uruguai e das Missoes

Keywords:

planning, taxonomic resolution, sampling effort, statistical tests.

Abstract

Apart from multimetric indices and predictive models, there are statistical methods that can be employed in biomonitoring or bioassessment studies. In this revision we emphasized the need for planning researches beforehand and discussed issues to this regard: clear definition of objectives, correct making of replications, simplification of data sets (through using genus/family, binary presence/absence results, ignoring uncommon species), standardization of sampling procedures, transformation and standardization of data. Moreover, we presented a brief overview of univariate and multivariate statistical analyses, based on the identification of traits of variables, like number, type (if response or predictor variable) and nature (categorical or quantitative variables). During the description of univariate analyses, we presented examples of the use of blocks and covariates and highlight the relevance of interacting explanatory variables. In the multivariate analyses, we considered the objectives and logics driving ordination and classification analyses and distance-based MANOVA.



Author Biography

Adriano Sanches Melo, Universidade Federal do Rio Grande do Sul

 

Published

2009-12-05