It has been previously demonstrated that a cognitive bias against disconfirmatory evidence (BADE) is associated with delusions. However, small samples of delusional patients, reliance on difference scores and choice of comparison groups may have hampered the reliability of these results. In the present study we aimed to improve on this methodology with a recent version of the BADE task, and compare larger groups of schizophrenia patients with/without delusions to obsessive–compulsive disorder (OCD) patients, a population with persistent and possibly bizarre beliefs without psychosis.

A component analysis was used to identify cognitive operations underlying the BADE task, and how they differ across four groups of participants: (1) high-delusional schizophrenia, (2) low-delusional schizophrenia, (3) OCD patients and (4) non-psychiatric controls.

As in past studies, two components emerged and were labelled ‘evidence integration’ (the degree to which disambiguating information has been integrated) and ‘conservatism’ (reduced willingness to provide high plausibility ratings when justified), and only evidence integration differed between severely delusional patients and the other groups, reflecting delusional subjects giving higher ratings for disconfirmed interpretations and lower ratings for confirmed interpretations.

These data support the finding that a reduced willingness to adjust beliefs when confronted with disconfirming evidence may be a cognitive underpinning of delusions specifically, rather than obsessive beliefs or other aspects of psychosis such as hallucinations, and illustrates a cognitive process that may underlie maintenance of delusions in the face of counter-evidence. This supports the possibility of the BADE operation being a useful target in cognitive-based therapies for delusions.