AntecedentsIn a previous study (Senín-Calderón et al., 2010) we observed that the REF scale of referential thinking (Lenzenweger et al., 1997) didn’t discriminate among different mental disorders.Objectives and hypothesesWe try to verify if self-references in various disorders are related to the severity of psychopathology (patients from public hospital and a private clinical). We predict that there will be differences between patients and controls, but not between the clinical samples. Psychotic disorders will be characterized by a significantly greater presence of self-references.MethodsParticipants: 287 subjects, 47 patients from a private clinical center, 57.4% women (mean age = 35.02, SD = 12.69), 30 patients from a public hospital, 53.3% women (38.36 years, SD = 9.53), and 210 controls selected from the general population, 50.5% women (33.80 years, SD = 11.79). Cross-sectional design, correlation method. All analysis were accepted at p < .05.ResultsThere are significant differences in self-references between patients and controls in frequency (t (285) = 2.33, p = . 021) and intensity (t (83.98) = 3.59, p = . 001). No significant differences between patients groups (p>.05) (REF-intensity without homogeneity, p < .05). No significant differences in self-references between types of diagnoses except psychotic patients versus adjustment disorder (frequency and intensity).ConclusionsSelf-references are highlighted in psychosis but, with the exception of adjustment disorders, doesn’t discriminate between personality, mood or anxiety disorders. Differences are more related to the clinical severity (BPRS) than with referential thinking.