The Two Meanings of External Invalidity

A distinction is proposed between two different senses of the treatment by background interaction (termed 'external invalidity' by Campbell & Stanley, 1966). The distinction parallels the dichotomy between universalistic (or theory centered) and particularistic (or context centered) objectives of behavioral research. In its universalistic sense the interaction above suggests a redefinition of the abstract meaning attached to the independent variables effective with respect to a given phenomenon of interest. In its particularistic sense this interaction defines the concrete contexts over which a given outcome will hold. The implications of the proposed distinction are discussed in reference to the phases of (1) designing the experiment and (2) evaluating the obtained results.