Bayesian analysis of the role of metacognition in cognitive control

Metacognitive theories of hypnotic responding, such as the cold-control theory, assert that people engage in strategies to create the experience requested by a suggestion (e.g., percept of something that is not present). This act is accompanied with the feeling of involuntariness due to a disrupted metacognition that makes suggestible people (highs) unaware of their intention to be engaged in the strategy. The presented research investigated predictions drawn from the cold-control theory focusing on the claim that hypnotic suggestions should not provide highs with any special abilities, since hypnotic responses are implemented by traditional cognitive control processes. In several experiments, the word-blindness suggestion (suggestion to see meaningless words in the Stroop task) was applied. The suggestion halves the extent of the Stroop interference when it is given to highs, posing a challenge to the metacognitive account of hypnosis and so providing a unique opportunity to test its assumptions. In addition, as the experimental usage of hypnosis requires offline screening to identify potential participants, I investigated whether it is possible to conduct the screening online to ease the costs of recruitment. Throughout this thesis, the Bayes factor was applied for hypothesis-testing. In the last chapter, a case with a 2x2 design (a design that was frequently used in this thesis) is presented to demonstrate how Bayes factors with cut-offs of good enough evidence relate to the old inferential mistake of the neglect of the test of interaction. This thesis presents empirical evidence that responses to the word-blindness suggestion are not produced by the disruption of reading. First, the extent of the effect is influenced by the proportion of incongruent Stroop trials of the experimental blocks implying that the suggestion alleviates response competition whereas it does not de-automatise reading. Second, when highs are asked to use simple visual strategies, such as blurring and looking-away, to reduce the Stroop interference, the pattern of results produced by the strategies are not in harmony with those of the word-blindness suggestion deeming it unlikely that highs use these strategies when they respond to the suggestion. This thesis also examines whether a voluntary act to imagine meaningless words is sufficient to reduce the Stroop interference. Consistent with the core idea of cold control theory, we found a positive correlation between the extent to which highs reduced interference in the volitional request and in the suggestion conditions. Nonetheless, the current strength of evidence is not good enough to conclude whether or not voluntary and hypnotic responses can reduce the interference to the same extent. The experiment comparing offline and online hypnotic screening demonstrated evidence supporting the notion that hypnotic suggestibility (measured via the SWASH) can be assessed online. Finally, different scenarios of a case study were presented to help researchers develop the right intuition on the issue of why Bayesian evidence for H1 in one group and Bayesian evidence for H0 in another group does not mean Bayesian evidence for the difference between the two by itself.