Dissociation is regarded as a possible psychological mechanism in nonepileptic seizures (NES), although existing evidence for this is equivocal. It has been suggested that the contradictory findings in this area reflect the use of measures that conflate qualitatively distinct types of dissociation, and provide inadequate coverage of the aspects of dissociation most closely related to NES. The study described here addressed this shortcoming by measuring the occurrence of two different types of dissociation, "detachment" (measured using the Cambridge Depersonalisation Scale) and "compartmentalization" (measured using the Somatoform Dissociation Questionnaire), in patients with NES (n=32) and epilepsy controls (n=37). As predicted, patients with NES scored significantly higher on the measure of compartmentalization only; contrary to prediction, however, this difference was no longer significant when anxiety and depression were controlled for. The conceptual and methodological implications of the study are discussed.