This paper attempts to illuminate the problems of outcome validity in psychological inquiry by reference to the two fundamental characterizations of research results: (1) their context of applicability, and (2) their causal meaning. The failure of a research proposition to hold within its intended context of generalization is labelled as context invalidity and is characterized as one sense of a statistical interaction between the treatment and the background of a research. A second sense of such interaction is discussed in relation to the notion of treatment invalidity. Specifically, this sense is defined as the implication that an effect might be causally attributed to some composite of the treatment and background rather than to the treatment as such. The progression of causal explication is characterized as the successive establishment of conclusion, treatment and construct validities. Misclassifications of validity types, the twofold implication of the perils to ‘external validity’, the priority issue between ‘internal’ ‘and external’ validities, and misconstruals of the requirement of representative sampling are discussed in light of the context-meaning analysis above.