The extent to which the observed results represent the truth in the population we are studying and, thus, are not due to methodological errors.