Local community structure and interactions have been shown to depend partly on landscape context. In this paper we tested the hypothesis that the spatial scale experienced by an organism depends on its trophic level. We analyzed plant-herbivore and herbivore-parasitoid interactions in 15 agricultural landscapes differing in structural complexity using the rape pollen beetle (Meligethes aeneus), an important pest on oilseed rape (Brassica napus), and its parasitoids. In the very center of each landscape a patch of potted rape plants was placed in a grassy field margin strip for standardized measurement. Percent non-crop area of landscapes was negatively related to plant damage caused by herbivory and positively to the herbivores’ larval mortality resulting from parasitism. In a geographic scale analysis, we quantified the structure of the 15 landscapes for eight circular sectors ranging from 0.5 to 6 km diameter. Correlations between parasitism and non-crop areas as well as between herbivory and non-crop area were strongest at a scale of 1.5 km, thereby not supporting the view that higher trophic levels experience the world at a larger spatial scale. However, the predictive power of non-crop area changed only slightly for herbivory, but greatly with respect to parasitism as scales from 0.5 to 1.5 km and from 1.5 to 6 km diameter increased. Furthermore, the effect of non-crop area tended to be stronger in parasitism than herbivory suggesting a greater effect of changes in landscape context on parasitoids. This is in support of the general idea that higher trophic levels should be more susceptible to disturbance.