Invasive alien plants are likely to be released from specialist herbivores and at the same time encounter biotic resistance from resident generalist herbivores in their new ranges. The Shifting Defense Hypothesis predicts that this will result in evolution of decreased defense against specialist herbivores and increased defense against generalist herbivores. To test this, we performed a comprehensive meta-analysis of 61 common garden studies that provide data on resistance and/or tolerance for both introduced and native populations of 32 invasive plant species. We demonstrate that introduced populations, relative to native populations, decreased their resistance against specialists, and increased their resistance against generalists. These differences were significant when resistance was measured in terms of damage caused by the herbivore, but not in terms of performance of the herbivore. Furthermore, we found the first evidence that the magnitude of resistance differences between introduced and native populations depended significantly on herbivore origin (i.e. whether the test herbivore was collected from the native or non-native range of the invasive plant). Finally, tolerance to generalists was found to be higher in introduced populations, while neither tolerance to specialists nor that to simulated herbivory differed between introduced and native plant populations. We conclude that enemy release from specialist herbivores and biotic resistance from generalist herbivores have contrasting effects on resistance evolution in invasive plants. Our results thus provide strong support for the Shifting Defense Hypothesis.