Many attempts to ascertain the physicochemical processes governing the emergence of life have focused on studying the synthesis of particular biomolecules and their precursors or on designing simple systems that manifest life-like properties. In this paper, we present the methodological components of an experimental framework designed to generate and detect spontaneously forming chemical systems capable of collective propagation and adaptive evolution. The chemical ecosystem selection paradigm involves incubating complex mixtures of organic compounds with populations of mineral grains to promote the appearance of autocatalytic sets that interact with the mineral surface and perform serial transfers to favor surface-associated systems that are better at being transmitted from grain to grain. This approach has the potential to serve as a novel tool for screening a vast array of experimental conditions and determine the likelihood that they will produce life-like chemical systems.