Beads task vs. box task: The specificity of the jumping to conclusions bias

- We investigated the Jumping to Conclusions (JTC) bias in high- and low-delusion-prone groups.
- We used the 'beads task' and a conceptually similar 'box task' to assess JTC.
- All participants requested significantly less information on the beads task relative to the box task.
- High-delusion-prone participants were significantly more conservative than low-delusion-prone group.

Background and objectivesPrevious research involving the probabilistic reasoning 'beads task' has consistently demonstrated a jumping-to-conclusions (JTC) bias, where individuals with delusions make decisions based on limited evidence. However, recent studies have suggested that miscomprehension may be confounding the beads task. The current study aimed to test the conventional beads task against a conceptually simpler probabilistic reasoning “box task”MethodsOne hundred non-clinical participants completed both the beads task and the box task, and the Peters et al. Delusions Inventory (PDI) to assess for delusion-proneness. The number of 'draws to decision' was assessed for both tasks. Additionally, the total amount of on-screen evidence was manipulated for the box task, and two new box task measures were assessed (i.e., 'proportion of evidence requested' and 'deviation from optimal solution').ResultsDespite being conceptually similar, the two tasks did not correlate, and participants requested significantly less information on the beads task relative to the box task. High-delusion-prone participants did not demonstrate hastier decisions on either task; in fact, for box task, this group was observed to be significantly more conservative than low-delusion-prone group.LimitationsNeither task was incentivized; results need replication with a clinical sample.ConclusionParticipants, and particularly those identified as high-delusion-prone, displayed a more conservative style of responding on the novel box task, relative to the beads task. The two tasks, whilst conceptually similar, appear to be tapping different cognitive processes. The implications of these results are discussed in relation to the JTC bias and the theoretical mechanisms thought to underlie it.