Bioassessment protocols usually call for collection of very large benthic samples to proportionally represent field conditions. They often rely on field-composite (FC) sampling methods where many small samples are collected and combined in the field to form one large, representative sample. FC samples are generally believed to homogenize the within-site variance of biological metrics and to generate more extensive taxa lists than smaller samples. However, subsampling procedures are universal among bioassessment protocols because indentifying large samples is usually cost-prohibitive.  I evaluated the assumptions underlying FC samples (500-organism-subsample) by comparing them with electronically-composited (EC) single-Surber samples (200-organism).  Significant within-site variation persisted among sites using FC sampling. EC-data produced 2x-higher taxa richness estimates than FC data, but could be made comparable through rarefaction analysis. Rarefaction showed that FC under-sampling can obscure differences among sites.  Estimating taxon-specific richness (e.g., EPT-richness) using rarefaction tended to generate lower estimates than FC-data, but corrective measures can be implemented. Analysis of eight EC-samples required laboratory cost similar to processing five FC-samples, but requires less field expense, is applicable in a wider range of stream-sizes, allows more analytical/diagnostic analysis, statistical hypothesis-testing power, and generates more complete taxa lists.