OBJECTIVE: Deficits in object relations (OR) and reality testing (RT) functions are found in schizophrenia but have never been investigated in bipolar disorder. In the current study, we examine if there are OR and RT differences in schizophrenia and bipolar disorder compared to healthy controls and to what extent differences in clinical characteristics mediates the putative effect of diagnosis. METHODS: We used the Bell Object Relation and Reality Testing Inventory (BORRTI) to measure OR and RT in schizophrenia (n = 55), bipolar disorder (n = 51) and healthy controls (n = 158). Diagnoses and the life time presence of psychotic symptoms were evaluated based on the Structured Clinical Interview for DSM-IV. We used the Positive And Negative Symptom Scale to measure current symptoms. RESULTS: Analyses of variance with post hoc tests showed statistically significant differences in OR and RT between the Schizophrenia (SCZ), Bipolar Disorder (BD), and Healthy Control (HC) groups. Multiple regression analyses indicated that a lifetime history of psychotic symptoms contributed significantly to the variance in one BORRTI subscale (Social Incompetence) while Positive And Negative Symptom Scale components (either the positive component and emotional discomfort component) contributed significantly to the variance in all BORRTI subscales except one (Uncertainty of Perception). CONCLUSIONS: OR and RT deficits are present both in SCZ and BD, but differences appears to be mediated by differences in current positive and depressive symptoms.