Transforming the canons of John Stuart Mill from philosophy to replicative, empirical research: The Common Cause research design

Abstract

When an element or factor is common to a set of circumstances that element may be causal in its relationship to particular dependent variables. This premise was stated by John Stuart Mill more than 170 years ago, and Mill's canon, the Method of Agreement, is used here as a basis to create the "Common Cause" (CC) research design. The CC design is particularly relevant when a set of multiple circumstances can be represented by alternative theories of change or competing explanations. We consider several potential applications of the design and elaborate its structure within the validity framework of Shadish, Cook, and Campbell. We discuss threats to validity controlled by the CC design (e.g., selection bias, the bane of applied researchers, is not relevant) and illustrate possible analytic strategies using simulated data. We explicitly compare the CC design to four quasi-experimental designs in terms of the validity threats that they eliminate. Design weaknesses are addressed and ways to enhance the design's inferential power discussed. The CC design itself represents a proof of concept suggesting that other research designs can be created from philosophical principles.