Implementation theory and the theory-driven approach to validity

Abstract

The emphasis on validity, a traditionally important question in science, has recently been heavily criticized. The Chen and Rossi approach attempts to answer some of these criticisms by redressing the balance between internal and external validity. The approach opens up the "black box" of process evaluation. Thus, implementation becomes a crucial factor in assessing the validity of program interventions. Inadequate attention to how a program is being or has been implemented poses threats to the four kinds of validity identified by Chen and Rossi. But there are at least four theories of implementation (top-down, bottom-up, adaptive, and evolutionary). These theories and their relationship to validity are discussed. The paper concludes that it may not be possible to achieve external validity, particularly if the adaptive and evolutionary theories of implementation are correct, and that it may be better to move away from causal to other forms of theorizing to get out of the validity dilemma.