Researchers in criminology have discovered the benefits of studying a phenomenon cross culturally, and a revival of the method is currently underway. However, criminologists currently use the method to determine the scope and generalizability of their theories, which were originally developed for use in one particular culture. Due to problems in this approach, cross-cultural research has met with varying success. In this study, two methods of constructing cross-cultural theories in criminology are examined. The evaluative approach determines a theory's scope and generalizability while the generative approach constructs explanatory models designed to transcend cultural boundaries. The current use of the evaluative approach has repeatedly shown the cultural blindness of current theories. The discovery that single-culture theories lack generalizability has led to a reexamination of their completeness. Consideration of scope conditions is the key to the success of the generative approach since it can explain behavior under a wide variety of systemic conditions. Generative models can thus increase the understanding of crime cross culturally, as well as enhance the understanding of crime within a culture. Notes, a diagram, and 17 references are included.