The positive relationship between biodiversity and ecosystem functioning is mainly derived from studies concerning primary producers, whereas a generalization of this relationship for higher trophic levels is more difficult. Furthermore, most evidence of the biodiversity–ecosystem functioning relationship is derived from experiments manipulating only one trophic level and, as a consequence, interactive diversity effects at multiple trophic levels have mostly been ignored. Here, we performed a mesocosm experiment in which we manipulated functional group diversity at two trophic levels (primary and secondary consumers) applying a full-factorial design. More specifically, we asked whether 1) predator functional diversity affects prey mortality rates, 2) prey functional diversity affects prey mortality rates, 3) whether there are interactive effects of simultaneous diversity changes at both trophic levels. For each trophic level we used two functional groups, i.e. organisms belonging to two different habitat domains: at the higher trophic position 1) a ground foraging spider species and 2) a spider species foraging in the vegetation canopy and at the lower trophic position 3) a ground living cricket species and 4) leafhoppers living in the vegetation canopy. Increasing predator functional group diversity increased prey mortality by 53%, and increasing prey functional group diversity increased prey mortality by 24%. Further, prey mortality was highest at the uppermost level of functional group diversity (142% increase in prey mortality compared to single prey and predator functional diversity), most likely due to resource partitioning between the predators. This finding demonstrates that a multi-trophic perspective is necessary, and that previous studies focusing on only one trophic level have most likely underestimated the strength of the relationship between biodiversity and ecosystem functioning.