Biotic interaction studies have revealed a large discrepancy among experiments in target responses to the effects of neighbours, which may in part be due to both high species-specificity of plant responses and low number of target species used in experiments. Our aim was to assess facilitative responses at the community level and the role of both functional groups and ecological attributes of target species. In a sub-alpine grassland on the eastern Tibet plateau, we assessed growth responses of all species in the community to removal of a dominant shrub. We also measured changes in the main environmental variables. Species responses were analysed by functional group and in relation to their mean regional altitudinal distribution. All significant interactions were positive and affected one-third of the total species richness of the community. All functional groups were facilitated but forbs were less strongly facilitated than in the two other groups. High-alpine species were less strongly facilitated than low-sub-alpine species, but the strength of this relationship was weaker than that reported in previous work. There was evidence of a decrease in extreme temperatures below the canopy of the shrub but no variations in soil moisture. We conclude that the highly stressful conditions induced by the dry continental climate of the eastern Tibet plateau are a main driver of the exclusive dominance of positive interactions. Assessing interactive responses at the community level is likely to provide a useful tool to better understand the role of biotic interactions in community responses to environmental changes.< Réduire