Introspective access or retrospective inference? Mind wandering reports are shaped by performance feedback

Most mind wandering (MW) paradigms use self-reports following task performance, but the extent to which these reports are confounded by performance cues is unknown. In two experiments, we examined whether self-reports and confidence therein are influenced by performance indicators during visual metronome response tasks. In Experiment 1, we found that sham feedback modulated reports independently of behavioural performance with participants more likely to report MW after incorrect than correct sham feedback. In Experiment 2, we replicated this pattern using a more implicit manipulation of perceived performance, a surreptitious delay in the onset of response targets. Participants were more likely to report MW after delay than control trials. In both experiments, confidence in on-task reports was lower when the corresponding indicator (falsely) implied poor performance. These findings suggest that MW reports and experiential state confidence are partly confounded by performance monitoring and have implications for a variety of experience sampling methodologies.