The set-up of a research study in which one or more elements being studied (e.g., exposure to a test ad) are controlled or manipulated among randomly assigned groups so that observed differences between the groups can be attributed to those elements (e.g., the test ad) rather than other possible variables.