First, the relative usefulness of various mechanisms for police control cannot be assessed by any empirical research thus far conducted, and there are reasons for believing that such a test is impossible. Second, the mechanisms for producing police accountability do not vary by forms of government, a surprising finding considering the differences between democratic and nondemocratic governments. Third, the structure of policing within a country has no effect on the way that accountability is achieved and probably little on accountability itself. Centralization, for example, does not indicate that accountability is weak, nor does decentralization indicate that it is strong. Other lessons are as follows: (1) changes in modes of control over the police should be undertaken for symbolic rather than instrumental reasons; (2) accountability is an effort to balance responsiveness to community opinion with the rule of law; (3) police discretion causes police performance to vary according to the situation, so practices to ensure accountability must also vary; (4) the primary objective of independent supervisory bodies is to enhance the police capacity for effective self-regulation; (5) independent bodies are necessary to act as a check on police internal discipline and to reassure the public that the police are not a closed operation; and (6) the art of achieving accountability, especially when there is demand for reform, is to enlist the support of the police in a commitment to disciplinary activities.