Cheatgrass (Bromus tectorum), Russian thistle (Salsola kali), and tall tumblemustard (Sisymbrium altissimum) are nonnative plants widely distributed throughout the desert and shrubsteppe communities of the western United States. The impact of these invaders on plant community structure, form, and function has been well documented, but investigations determining the impacts of this cumulative invasion on terrestrial vertebrates have not been undertaken. Our objective was to assess community-level rodent responses to changes in plant community features, with an emphasis on dominance of invasive plant species. We sampled rodent and plant communities in the Great Basin Desert (Utah) over 4 years. Using estimates of rodent species richness and average nightly captures (relative abundance) as our response variables, we developed generalized linear mixed models (GLMMs) to determine the effects of invasive species cover. We found that rodent richness decreased with increasing abundance of invasive plant cover. Contrary to other studies, there was a nonlinear relationship between invasive species cover and rodent abundance, where rodent captures increased with invasive plant cover, reached a threshold, and then exhibited a negative response. This nonlinear relationship provides support for the intermediate disturbance hypothesis and suggests that moderate levels of plant invasions, by way of bolstering rodent abundance and rodent biomass, could have bottom-up effects (i.e., positively influencing species that predominantly prey upon rodents). Our findings contradict previous findings on plant invasions in arid portions of the western United States and suggest that the species comprising or dominating a given rodent community may determine the impact of plant invasions.