The role of anchoring information in judgments of learning.

Abstract

This study examined informative and uninformative anchoring effects on judgments of learning (JOLs), focusing on two hypotheses: the optimistic/pessimistic and differential-scaling hypotheses. The optimistic/pessimistic hypothesis states that anchoring information changes subjective confidence in memory, whereas the differential-scaling hypothesis states that anchoring information elicits a scaling bias in the conversion process of subjective internal confidence into scale JOLs (i.e., 0-100% responses). Experiment 1 focused on binary JOLs (i.e., Yes/No predictions). The results confirmed that the informative anchoring effect occurred (i.e., binary JOLs in the high anchor condition were higher than those in the low anchor condition), whereas the uninformative anchoring effect did not. Experiment 2 evaluated whether the difference in response scales between anchoring information and JOLs elicited the anchoring effect, demonstrating that the informative anchoring effect occurred when different response scales were used for the anchoring information (i.e., the number of words correctly recalled) and JOLs (i.e., 0-100% scale), and the uninformative anchoring effect did not. Experiment 3 examined whether the uninformative anchoring effect can be explained by numeric priming rather than scaling bias, demonstrating that anchoring information unrelated to test performance using a 0-100% scale did not elicit the uninformative anchoring effect. These findings suggest that the informative anchoring effect supports the optimistic/pessimistic hypothesis, whereas the uninformative anchoring effect supports the differential-scaling hypothesis. Thus, the nature of anchoring information affects the process of forming JOLs. Specifically, the uninformative anchor elicits only scaling bias, whereas the informative anchor changes subjective confidence in memory.