Development of empowerment theory has focused on defining the construct at different levels of analysis, presenting new frameworks or dimensions, and explaining relationships between empowerment-related processes and outcomes. Less studied, and less conceptually developed, is the nature of empowerment as a higher-order multidimensional construct. One critical issue is whether empowerment is conceptualized as a superordinate construct (i.e., empowerment is manifested by its dimensions), an aggregate construct (i.e., empowerment is formed by its dimensions), or rather as a set of distinct constructs. To date, researchers have presented superordinate models without careful consideration of the relationships between dimensions and the higher-order construct of empowerment. Empirical studies can yield very different results, however, depending on the conceptualization of a construct. This paper represents the first attempt to address this issue systematically in empowerment theory. It is argued that superordinate models of empowerment are misspecified and research that tests alternative models at different levels of analysis is needed to advance theory, research, and practice in this area. Recommendations for future work are discussed.