A field experiment was conducted over 3 yr to assess the effects of transgenic field corn expressing stacked lepidopteran-active Bacillus thuringiensis (Bt)-derived VIP3A and Cry1Ab proteins on nontarget arthropods. Effects were compared with an isogenic conventional corn hybrid treated with a foliar insecticide and an isogenic control without insecticides. Plant inspections, pitfall traps, sticky cards, and emergence traps determined the abundance and diversity of plant, aerial, and surface-dwelling communities. More than 500,000 arthropods were enumerated, representing 203 taxonomic groups in 112 families and 13 orders. In terms of abundance, 70% of the arthropods were saprovores, 13% were herbivorous insects, 14% were predators, and 3% were parasitoids. Overall biodiversity and community-level responses were not significantly affected by expression of the stacked VIP3A and Cry1Ab proteins. Densities of most nontarget taxa exposed to the Bt hybrid showed no differences compared with the isogenic control. However, significant changes in certain taxa did occur in the Bt plots, which were indirectly related to plant-mediated factors, prey density responses, and the absence of plant injury. Arthropod communities in the insecticide-treated plots displayed both negative and positive changes in the abundances of individual taxa. Together, effects observed in the Bt plots were significantly less than the community disturbances caused by insecticide applications. Changes in nontarget communities in plots previously exposed to insecticides and the Bt hybrid did not carryover to the following growing season. Major arthropods as bioindicators, routes of exposures, and food web interactions relative to the observed effects are discussed.