chapter  5
The Internal Validity of Research Designs
Pages 11

If the project involves manipulating one or more independent variables and noting the effects of these manipulations on one or more dependent variables, one primary objective is to be able to attribute changes or differences observed in the dependent variable to the manipulations either produced or observed by the experimenter. To justify the conclusion that one set of events (intervention and manipulation) produced another set of events (changes or differences in the dependent variable or variables), the experimental procedures must be arranged to eliminate the possibility that some influence other than the observed or intended one is responsible for the differences or changes in the dependent variable. Campbell and Stanley (1963)

have called this procedure eliminating plausible, rival explanations for the research findings.