The efficiency of biodiversity assessments and biomonitoring studies is commonly challenged by limitations in taxonomic identification and quantification approaches. In this study, we assessed the effects of different taxonomic and numerical resolutions on a range of community structure metrics in invertebrate compositional data sets from six regions distributed across North and South America. We specifically assessed the degree of similarity in the metrics (richness, equitability, beta diversity, heterogeneity in community composition and congruence) for data sets identified to a coarse resolution (usually family level) and the finest taxonomic resolution practical (usually genus level, sometimes species or morphospecies) and by presence-absence and relative abundance numerical resolutions. Spearman correlations showed highly significant and positive associations between univariate metrics (richness and equitability) calculated for coarse- and finest-resolution datasets. Procrustes analysis detected significant congruence between composition datasets. Higher correlation coefficients were found for datasets with the same numerical resolutions regardless of the taxonomic level (about 90%), while the correlations for comparisons across numerical resolutions were consistently lower. Our findings indicate that family-level resolution can be used as a surrogate of finer taxonomic resolutions to calculate a range of biodiversity metrics commonly used to describe invertebrate community structure patterns in New World freshwater wetlands without significant loss of information. However, conclusions on biodiversity patterns derived from datasets with different numerical resolutions should be critically considered in studies on wetland invertebrates.