Despite growing interest in stage theories of health behavior, there is considerable confusion in the literature concerning the essential characteristics of stage theories and the manner in which such theories should be tested. In this article, the 4 key characteristics of a stage theory-a category system, an ordering of categories, similar barriers to change within categories, and different barriers to change between categories--are discussed in detail. Examples of stage models of health behavior also are described. Four major types of research designs that might be used for testing stage theories are examined, including examples from the empirical literature. The most commonly used design, which involves cross-sectional comparisons of people believed to be in different stages, is shown to have only limited value for testing whether behavior change follows a stage process.