Theory testing and small-N design

In political science, a growing body of literature maintains that case studies do provide empirical ground for testing theory as far as procedures apply that clearly expose the probative power of evidence. We scrutinize argumentative, Bayesian, and set-theoretic procedures, and find that the former two can discipline discretion, yet not truly prevent biased conclusions about the fitting of cases to theory when used as stand-alone techniques -- while the latter relegates discretion to design issues, but cannot apply to N=1 studies.