Accurate detection of organisms is crucial for the effective management of threatened and invasive species because false detections directly affect the implementation of management actions. The use of environmental DNA (eDNA) as a species detection tool is in a rapid development stage; however, concerns about accurate detections using eDNA have been raised. We evaluated the effect of sampled water volume (0.25 to 2 L) on the detection rate for three macroinvertebrate species. Additionally, we tested (depending on the sampled water volume) what amount of total extracted DNA should be screened to reduce uncertainty in detections. We found that all three species were detected in all volumes of water. Surprisingly, however, only one species had a positive relationship between an increased sample volume and an increase in the detection rate. We conclude that the optimal sample volume might depend on the species−habitat combination and should be tested for the system where management actions are warranted. Nevertheless, we minimally recommend sampling water volumes of 1 L and screening at least 14 μL of extracted eDNA for each sample to reduce uncertainty in detections when studying macroinvertebrates in rivers and using our molecular workflow.