Technically, the investigation of the way in which the INDEPENDENT VARIABLE in an EXPERIMENT affects the DEPENDENT VARIABLE, but the term is used more generally for the whole procedure of an experiment, from the selection of subjects to the statistical analysis of the results.