Herbivore attack has important consequences on plant performance and on plant community composition, and understanding plant responses to herbivory is important for improving our ability to predict community dynamics and in developing biological control programs. The first objective of our study was to assess the response of purple loosestrife (Lythrum salicaria) to various levels of tissue loss (0, 10, 20, and 60–70%) caused by feeding by a chrysomelid beetle (Galerucella calmariensis) introduced as a biological control agent to North America. Leaf beetle herbivory changed flowering phenology, caused significant reductions in plant growth and inflorescence mass, and increased the branching pattern even in the lowest herbivory treatment. In general, plant performance measures (height, total length of inflorescence) decreased linearly with increasing defoliation level. A second objective was to evaluate whether commonly used artificial herbivory techniques (hole punching and clipping) can reproduce a plant’s response to beetle herbivory at different attack levels. Studies investigating plant–herbivore interactions often favor artificial over natural herbivory because of the control over amount and pattern of tissue loss. However, use of simulated herbivory rests on the assumption that plants respond in similar ways to natural and artificial herbivory. In the case of purple loosestrife (and other biocontrol programs), allowing artificial manipulation of aboveground herbivory may allow easier studies of the impact of multiple herbivores and potential competitive interactions among biocontrol agents. However, differences in growth, reproduction, and plant architecture of purple loosestrife in response to leaf beetle feeding could not be reproduced using simulated herbivory.