This article examines four ways in which the ‘two faces of power’ argument might be, or has been, subjected to empirical test. These include taking into account the impact of structure and socialization; re-defining ‘non-decisions' as observable acts; substituting the question ‘who benefits' for ‘who governs’; and, finally, comparisons between cities. The analysis concludes that each poses difficulties of empirical verification, some of which can be overcome, some not. Even if overcome however, they would not nullify existing findings on power in communities.