Biological indicators used to assess stream conditions are often influenced by many sources of error. We utilized 66 sites (20 reference and 46 test) where algae, fish, and invertebrate data were collected by the National Water-Quality Assessment Program to compare the relative magnitude of sampling variability to annual and spatial variation for biological indicators. Additionally, we evaluated how sampling error influences stream assessments. Overall, annual effects across all three assemblages were negligible and spatial effects often described the highest proportion of total variance. At reference sites, total variation attributable to sampling error was highest for invertebrates (64%) and lowest for algae (16%); whereas, algae was highest (35%) and fish was lowest (20%) at test sites. Percent disagreement in site assessment due to sampling error was similar across assemblages at reference sites (15% - 16%), but nearly double for algae and invertebrates at test sites (29% - 31%). Because of sampling error, results suggest that a 15 - 16% change in indicator values for all assemblages is needed to detect a temporal change at reference sites and a 17 - 30% change is needed at test sites depending on the assemblage.