Abstract
ABSTRACTKoriat's Self-Consistency Model of subjective confidence proposes that the consistency and accessibility with which a response is retrieved are cues to confidence in the accuracy of that response. The Self-Consistency Model, however, has only been assessed using two-alternative forced-choice questions. Consideration of open-ended questions suggested that response cardinality—the number of unique response options that might come to mind—should directly reduce response consistency, and thus indirectly reduce confidence. Australian undergraduate students (N = 389) completed a 20-item open-ended general-knowledge test with confidence ratings attached to each answer. Replicating previous Self-Consistency Model findings, but in an open-ended format for the first time, consistency and accessibility were independent positive predictors of confidence, though the prediction of accessibility did not reach statistical significance. Extending the model, the number of unique responses given by participants, a measure of response cardinality, was found to be a strong negative predictor of consistency and indirect negative predictor of confidence. Implications of the results therefore include the use of simple strategies such as prompting or providing varying numbers of response options as potentially effective for manipulating cardinality and thus confidence.
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have