Phylogenetic comparative methods represent a major advance in integrative and comparative biology and have allowed researchers to rigorously test hypotheses about trait divergence in an evolutionary framework. However, phylogenetic comparative methods require trait data for many species, which can be impractical for certain taxonomic groups and trait types. We suggest that the philosophical principles of severity and strong inference can be employed in an integrative framework to infer adaptation in studies that compare only a few populations or species. This approach requires 1) ensuring that the study system contains species that are relatively closely related, 2) the formulation of a specific, clear, overarching hypothesis that can be subjected to integrative testing across the biological hierarchy (e.g., ecology, behavior, morphology, physiology, genetics), 3) the application of severe tests that avoid statistical underdetermination, and 4) the systematic refutation of alternative hypotheses. While difficult to collect for more than a few species, detailed and integrative data can be used to differentiate among several potential agents of selection. In this way, integrative studies of small numbers of closely related species can complement and even improve upon broad-scale phylogenetic comparative studies by revealing the specific drivers of adaptation.