Empowerment theory provides both a value orientation for community-based research and practice, and a conceptual framework for understanding and evaluating interventions. One critical question in empowerment theory involves whether a reciprocal (bidirectional) relationship exists between community participation (CP) and psychological empowerment (PE). This study applied structural equation modeling (SEM) with two waves of survey data from a cross-lagged panel design to test reciprocal and unidirectional causal relations between latent variables representing CP and PE. Participants (n = 474) were randomly selected neighborhood residents and organizational members from the United States. Four models were tested using SEM: (a) a baseline model with autoregressive paths, (b) a model with autoregressive effects and CP predicting future PE, (c) a model with autoregressive effects and PE predicting future CP, and (d) a fully cross-lagged model. Results indicated that CP influenced future PE; however, reciprocal causality was not found to occur between the variables. Implications of the study for empowerment theory and community-based practice are described, and directions for future research discussed. (Contains 2 tables and 2 figures.)