Trophic level (TL)-based indicators have been widely used to examine fishing impacts in aquatic ecosystems and the induced biodiversity changes. However, much debate has ensued regarding discrepancies and challenges arising from the use of landings data from commercial fisheries to calculate TL indicators. Subsequent studies have started to examine survey-based and model-based indicators. In this paper, we undertake an extensive evaluation of a variety of TL indicators across 9 well-studied marine ecosystems by making use of model- as well as surveyand catch-based TL indicators. Using detailed regional information and data on fishing history, fishing intensity, and environmental conditions, we evaluate how well TL indicators are capturing fishing effects at the community level of marine ecosystems. Our results highlight that the differences observed between TL indicator values and trends is dependent on the data source and the TL cut-off point used in the calculations and is not attributable to an intrinsic problem with TLbased indicators. All 3 data sources provide useful information about the structural changes in the ecosystem as a result of fishing, but our results indicate that only model-based indicators represent fishing impacts at the whole ecosystem level.