The overall objective of this study was to investigate the longitudinal association between diffusion of responsibility and pro-aggressive bystander behavior across three time points in upper elementary education. This three-wave longitudinal study included 1905 Swedish students who completed a questionnaire in at least one of the three waves: the fourth (M[subscript age] = 10.56), fifth (M[subscript age] = 11.55), and/or sixth grades (M[subscript age] = 12.58). Both traditional and random intercept cross-lagged panel models revealed a reciprocal relationship between pro-aggressive bystanding and diffusion of responsibility from the fourth to fifth grades, whereas the only significant cross-lagged path from the fifth to sixth grades was from pro-aggression to diffusion of responsibility in the traditional cross-lagged panel model. Thus, this study provides evidence for bidirectional longitudinal associations between diffusion of responsibility and pro-aggressive bystander behavior but did not support a full cross-lagged bidirectional model.