More than one way to violate independence: Reply

T. Curran and D. L. Hintzman's (1997) analysis of nonindependence in process dissociation did not confound aggregation bias with process dependence, as L. L. Jacoby and P. E. Shrout (1997) claimed. This reply shows that the numerical example presented as undermining Curran and Hintzman's arguments is entirely consistent with their equations and reinforces their analysis of the effect of subject-item variance in inducing aggregation bias. Process dependence and aggregation bias can be understood as distinct sources of nonindependence in data. However, process-dissociation estimates are affected by both.