Objective. Several studies have provided evidence for the claim that a subgroup of (schizophrenic) patients with current delusions share a jumping to conclusions (JTC) bias. The primary aim of the present study was to investigate whether currently deluded and non-deluded schizophrenic patients perform differently on three tasks tapping probabilistic reasoning. Method. Probabilistic reasoning was assessed in 31 schizophrenic patients, 28 psychiatric controls, and 17 healthy controls. In addition to the traditional draws to decision procedure, we employed two tasks for which participants had to judge, at each stage, the likelihood that beads come from either container (graded estimates procedure). Reaction times were recorded for the graded estimates procedure. Results. AJTC bias was displayed by 42% of the schizophrenic patients in the draws to decision condition, while 7% of the psychiatric patients and none of the healthy controls reached a decision after only one bead. A similar pattern of results was observed for the graded estimates procedure. This bias was more pronounced in deluded schizophrenic patients, although currently non-deluded patients also showed evidence for earlier decisions. A bias to over-adjust when confronted with potentially disconfirmatory evidence was confined to deluded schizophrenic participants. There was also evidence for an increase in JTC in the deluded group over the course of the tasks. No substantial group differences occurred with respect to reaction time parameters indicating that results are not attributable to impulsivity. Discussion. The findings provide further evidence for state and trait characteristics of abnormal reasoning in paranoid schizophrenia. Results are discussed in light of several competing explanations for JTC in schizophrenia.