Mobbing represents a well-known anti-predatory behavior, where potential prey display aggressively against a predator. Despite considerable experimental and descriptive work, no models predict species participation in mobbing assemblages. Here, we aimed to understand why some bird species engage in this behavior, while others do not, and what factors can be used to predict mobbing engagement within an avian community. We investigated whether certain functional traits, such as body size, foraging guild, foraging mode, and strata, as well species abundance and evolutionary relatedness, are important mobbing predictors. To address these goals, we simulated the presence of the Ferruginous Pygmy-Owl (Glaucidium brasilianum) by broadcasting its voice in 230 experiments conducted in 115 points, systematically distributed in a dry forest of northeastern Brazil. We compared these results to 162 avian surveys (point counts) conducted in the same area. Our avian surveys detected 108 bird species (local avian community), whereas our playback experiments attracted 72 species (mobbing assemblage). In general, small, canopy insectivorous or frugivorous birds dominated the mobs. The best mobbing predictors were body mass and guild, whereas species abundance, foraging mode, and strata were not retained in the best models. We found a strong phylogenetic component in body mass and mobbing propensity (almost 90% of the species and individuals participating in the mobs were passerines). At the community level, we found significant differences in the functional and phylogenetic structure of the mobbing assemblage in relation to the avian community. Our results suggest that mobbing behavior is tightly associated to predation risk and the capacity of individual species to find and detect predators, and that functional and phylogenetic features can predict species participation in this complex animal behavior.