The Enemy Release Hypothesis (ERH) proposes that non-native plants escape their co-evolved herbivores and benefit from reduced herbivory in their introduced ranges. Numerous studies have tested this hypothesis, with conflicting results, but previous studies focus on average levels of herbivory and overlook the substantial within-population variability in herbivory, which may provide unique insights into the ERH. We tested differences in mean herbivory and added a novel approach to the ERH by comparing within-population variability in herbivory between native and non-native plant populations. We include several covariates that might mask an effect of enemy release, including latitude, regional plant richness, plant growth form and plant cover. We use leaf herbivory data collected by the Herbivory Variability Network for 788 plant populations (616 native range populations and 172 introduced range populations) of 503 different native and non-native species distributed worldwide. We found no overall differences in mean herbivory or herbivory variability between native and non-native plant populations. Taken together, our results indicate no evidence of enemy release for non-native plants, suggesting that enemy release is not a generalized mechanism favoring the success of non-native species.