Multinomial processing-tree modeling has had a major impact on process-dissociation theory. Buchner, Erdfelder, and Vaterrodt-Plünnecke (1995) added guessing parameters to the original model of Jacoby (1991) and created a class of process-dissociation models. Furthermore, Erfelder and Buchner (1998) formulated criterion values of the dual-process signal-detection model ( Yonelinas, 1994 ) as multinomial parameters. Buchner, Erdfelder, Steffens, and Martensen (1997) suggested a new approach by proposing a multinomial source-monitoring model for process-dissociation data. Two experiments described here demonstrated that dual-process signal-detection theory must assume different levels of familiarity in inclusion and exclusion tests. Similarly, in some cases the source-monitoring model must assume different levels of recognition guessing in the two tests. Reasons are given for preferring the source-monitoring model.