Metacognition
Does expecting a clear signal make you feel more sure?
Open access · cc by · source: Europe PMC
When people expected a strong visual signal, they felt more confident and saw identical moving dots as clearer, even though they were no more accurate.
Study at a glance
- Design
- Human experiment — Within-subject cueing experiments: colour cues were first paired with strong or weak dot-motion coherence, then followed by identical medium-coherence probes; Exp 1-2 measured confidence, Exp 3 measured rated clarity, with a one-parameter learning model fitted to Exp 3
- N
- Three separate online samples: 34 in Experiment 1, 34 new participants in Experiment 2, and 62 new participants in Experiment 3; no exclusions
- Population
- Adult online participants recruited via Prolific with normal or corrected vision and no psychiatric or neurological history
- Outcome
- Confidence level, accuracy, d', meta-d' and M-ratio (Exp 1-2); subjective visibility ratings of motion clarity (Exp 3)
Structured fields used in claim comparison tables when every cited study has a complete layer.
Key findings
On the identical medium probes, people were more confident when the colour predicted a strong signal, in both Experiment 1 and the replication, while accuracy, sensitivity and response times did not differ. Metacognitive sensitivity (how well confidence tracked accuracy) was also unchanged, so expectations produced a bias rather than better insight. In Experiment 3, the same medium clouds were rated as clearer when a strong signal was expected. The model reproduced this effect, its predicted effect sizes correlated strongly with each person's actual effect, and people who weighted their prior more heavily showed larger biases.
Methodology
Online participants judged whether clouds of moving dots drifted left or right. During training, one dot colour always signalled strong motion and another always signalled weak motion, without participants being told. In the test phase, half the trials were medium-strength probes shown in either colour, so any difference between colours could only come from expectations. Experiments 1 and 2 (the second preregistered, with a harder medium level) measured confidence; Experiment 3 (preregistered) instead asked how clear the motion looked, and a simple learning model that blends prior expectation with current evidence was fitted to those ratings.
Limitations
The effects were modest in size, and the confidence effect in Experiment 2 was only just significant. Visibility ratings might still be shaped by hidden judgments about whether one could answer correctly, which the authors say future designs should separate from confidence. The learning model is not itself Bayesian, so the data do not show that the brain weights priors by their uncertainty, and the work does not test whether expectations change low-level perception or later behaviour such as information seeking. Links to hallucinations and psychosis are speculative; no clinical groups were tested.
How this study connects
Role on claims
Each row is a claim on a concept or method page where this paper supports, challenges, or qualifies the statement. Roles are hand-checked — not a model guess.
Not yet placed on a claim. This paper has study layers, but no concept page yet cites it as support, challenge, or qualifier.
Related papers in this topic
Same topic cluster — not a recommendation engine.