Ecosystem engineers facilitate beneficiary species by ameliorating physical habitat. The stress‐gradient hypothesis (SGH) predicts the importance of facilitation in communities should increase with physical stress but has rarely been tested in freshwater. Bluehead Chubs (Nocomis leptocephalus) build gravel nests for spawning, which can reduce negative effects of sedimentation for lithophilic species including invertebrates and other taxa. Our goal was to test the SGH using chubs and benthic assemblages as a model system. We surveyed assemblages in chub nests, paired unmodified substrate and reach‐wide samples in 10 Piedmont streams in South Carolina, USA, placed across a gradient of sedimentation. Based on the SGH, we predicted benthic assemblage diversity in chub nests would show no relationship to increasing embeddedness but that diversity in unmodified substrate should decrease with embeddedness. We found that taxa counts, richness and Shannon diversity were higher in chub nests than unmodified substrate but were lower than reach‐wide samples. Canonical correspondence analyses indicated benthic assemblages differed between nests and unmodified substrate, but assemblages in both microhabitats were nested subsets of the more diverse reach‐wide assemblage. Contrary to our hypotheses, diversity in both microhabitats decreased significantly with substrate embeddedness but was consistently higher in nests. While substrate modification by chubs clearly facilitated benthic diversity at the microhabitat scale, it was not sufficient to overcome the worst effects of sedimentation. This study provides mixed evidence for SGH in streams; chub nesting appears to be facilitative at the microhabitat scale but may not have reach‐wide effects on benthic assemblages in this system.