1. Many developing countries harbor inordinate numbers of species that face little-understood, gradual changes in their environment, such as chronic anthropogenic disturbance (CAD, a high frequency but low intensity form of disturbance). These countries also lack the resources to study each species, so conservation practices have had to be generalized assuming that complete taxonomic groups (e.g., cacti or cetaceans) may be managed in the same way. This could be justified if closely related species respond similarly to threatening factors. We assessed if relatives respond similarly to CAD and whether this affects phylogenetic diversity. To determine if such patterns occur in different systems, we studied a semiarid grassland and tropical dry forest in Mexico. 2. Species’ identity and abundance was recorded in 59 sites differing in CAD intensity. Disturbance response indices were calculated for each species. Nested ANOVAs coupled with null models were applied to determine if generalizations within taxa are justified and which taxonomic level (using APG IV system) explained the most variation in disturbance response among species. We obtained phylogenetic trees using molecular data for the grassland and published phylogenies for the tropical forest. Phylogenetic signal was measured with Pagel’s λ, whereas community mean phylogenetic distance was regressed on CAD. 3. Higher taxonomic levels explained more variation among species than expected by chance, indicating that related species respond similarly to CAD. However, only species in the same genus behaved similarly enough (explaining over 80 % of the accumulated variance) to make generalizations reliable. This is the result of an underlying, if modest, phylogenetic signal in CAD responses. 4. Mean phylogenetic distance decreased with CAD in the grassland but not the tropical dry forest. This suggests that CAD was a stronger environmental filter in the grassland but weak in the tropical dry forest, where it was less intense. Policy implications. Species inform us about how their relatives respond to CAD, supporting the idea that generalizations in management are possible; however, this procedure is seemingly reliable only within genera, and not within orders or families. CAD may reduce phylogenetic diversity, perhaps leading to ecosystem function loss.27-Feb-2018