Generalization. Stimulus and response generalization were programmed as described above. Generalization was assessed by conducting probes in novel settings, with novel instructors, and with play sets similar, but not identical, to toy sets depicted in videos. First, baseline probes (without video modeling) were conducted with Generalization 1 play sets (i.e., Gen 1 Circus, Gen 1 House). When scripted play actions were low (i.e., generalized responding was not observed), video modeling was used to teach the participants to play with the Generalization 1 play sets. That is, the original videos depicting the teaching sets were shown, followed by access to the Generalization 1 play sets. When scripted play actions increased with the Generalization 1 play sets following this teaching, baseline probes (without videos) were conducted with Generalization 2 play sets (i.e., Gen 2 Circus, Gen 2 House) to assess generalized responding with these novel toys.