ABSTRACT
In the philosophy of science, only one exception to the need to use similar cases to develop counterfactuals is routinely noted: the theory of signed causes. It is also the theory detectives use when visiting the scene of a crime. A similar theory—called theory-based evaluation—has recently been reinvented in evaluation. It too has evolved to do away with the need for a credible and valid causal counterfactual. However, the theory of signed causes makes considerable demands on the specificity and accuracy of substantive theory and also requires measures that are totally valid. Congress has regularly mandated that evaluations of job training programs be conducted with random assignment. The same is true of evaluations of early childhood education, where programs are tested to see whether the delivery of educational services to poor children actually improves their scholastic performance.
