Objectives. This study investigated whether a reasoning bias (‘jumping to conclusions’; JTC) found to be associated with higher levels of conviction in delusions is also associated with high‐conviction beliefs in obsessive‐compulsive disorder (OCD). Design. The experimental design was mixed‐effects, with one between‐subjects factor of group and one within‐subjects factor of task. Methods. Participants were 16 people with high‐conviction OCD (≥50%), 16 people with low‐conviction OCD (<50%), 16 people with delusions (≥50% conviction), and 16 non‐clinical controls. JTC was assessed using a neutral probabilistic reasoning task (beads task) and a version involving words of differing emotional salience (words task). Results. There was no statistically significant difference in draws to decision on the beads task between the high‐conviction OCD group, the low‐conviction OCD group, and the non‐clinical controls. The delusions group made significantly fewer draws to decision compared to the non‐clinical controls on the words task but not the beads task and significantly fewer draws on both tasks compared with the combined OCD group. Emotionally salient material (words task) did not affect draws to decision in any group. Conclusions. JTC is associated with delusions but not high‐conviction beliefs in OCD. The cognitive processes associated with high‐conviction beliefs in OCD and psychosis may not be trans‐diagnostic. This is consistent with the view that high‐conviction OCD should not be classified as a psychotic disorder.