This chapter explores the relative strengths and weaknesses of program theory as a tool for inferring causality and outlines a five-stage approach that makes increased use of inductively built program theories and takes more deliberate account of the varying levels of certainty that are required for evaluative conclusions.