The discussion is based on 11 interviews with senior police officers, civil servants, and academics as well as the literature they recommended. These individuals attributed the Home Office's interest in measuring police efficiency to several factors, including a cost-conscious approach to public services dictated by economic difficulties and substantial increases in the police force, accompanied by a rise in crime rates. The Home Office, however, has chosen to rely on the judgment of its inspectors, a decision dictated by the difficulty of applying evaluation measures to the police. While respondents identified several difficulties in assessing police performance, the interviews revealed major problems in defining what police and patrol officers are trying to do and how they allocate their time, the impossibility of determining whether changes in police practices have produced changes in crime, and the lack of a common yardstick for evaluating results equivalent to a profit in business. In addition, the introduction of measures of efficiency could be seen as an attempt to control situations in which control is impossible and perhaps undesirable. In practice, measures have the most usefulness within the context of a policy, a research design, and a political debate or management system. In developing a mechanism for monitoring policy implementation, four principles are important in producing valid statistical data: knowledge of background, specificity, comparison, and robustness. Evaluations of police operations will require four sets of statistics and information: force statistics, operational statistics, statistics relevant to support functions, and personnel measures. The paper discusses each area in some detail, with attention to various techniques and possible problems. It then considers the implications of this approach for policy and research. A list of the individuals interviewed and four references are appended.