A controlled experiment in which participants are randomly assigned to view and respond to one of several versions of a stimulus, product design or ad. This technique allows for responses to stimuli in different versions to be directly compared, because the versions were administered to parallel randomized groups.