Metacognitive illusions can be reduced by monitoring recollection during study

Abstract

Prior work has shown that judgments of learning (JOLs) are prone to an auditory metacognitive illusion such that loud words are given higher predictions than quiet words despite no differences in recall as a function of auditory intensity. The current study investigated whether judgments of remembering and knowing (JORKs)—judgments that focus participants on whether or not recollective details will be remembered—are less susceptible to such an illusion. In Experiment 1, participants studied single words, making item-by-item JOLs or JORKs immediately after study. Indeed, although increased volume elevated judgement magnitude for both JOLs and JORKs, the effect was significantly attenuated when JORKs were elicited. Experiment 2 replicated this finding and additionally demonstrated that participants making JORKs were less likely than participants making JOLs to choose to restudy quiet words relative to loud words. Taken together, these results suggest that JORKs are impacted less—in terms of both metacognitive monitoring and control—by irrelevant perceptual information than JOLs. More generally, these data support the contention that metacognitive illusions can be attenuated by simply changing the way metacognitive judgments are solicited, an important finding given that subjective experiences guide self-regulated learning.

Keywords:

• Metamemory
• Judgments of learning
• Recollection
• Metacognitive illusions

We thank Megan Cain, Caitlin Chase, Marie Dobson, Lauren Ferrell, Rachel Lam, Kira McCormack, and Jordan Peters for their assistance with data collection.

We thank Megan Cain, Caitlin Chase, Marie Dobson, Lauren Ferrell, Rachel Lam, Kira McCormack, and Jordan Peters for their assistance with data collection.

Notes

¹ Imposing a delay between studying and making JOLs for items has been shown to dramatically increase the relative accuracy of JOLs, a finding termed the delayed-JOL effect (Nelson & Dunlosky, 1991; for a review, see Rhodes & Tauber, 2011). However, of current interest are immediate judgments.

² A long-standing debate concerns whether episodic memory is based on a single process of memory strength or on dual processes composed of recollection (often equated with remembering) and familiarity (often equated with knowing; for a review, see Yonelinas, 2002). We are simply interested in the subjective experiences of remembering and knowing, and make no contention whether these states accurately map onto the processes of recollection and familiarity, respectively.

³ JOLs are typically made on a 0%–100% scale; however, given that JORKs were made on a 3-point scale (Recollect, Know, or Forget?), JOLs also needed to be for the two judgments to be directly comparable. Also, for JORKs the term ‘recollect’ was used instead of ‘remember’ in order to avoid potential confusion among participants regarding the differences between the subjective states of remembering and knowing.

⁴ We also calculated judgment-recall gammas, which did not differ between JOLs (G = .38; SE = 10) and JORKs (G = .41; SE = .09) (p < .05). Thus, in this case, JORKs did not predict overall memory better than JOLs, highlighting the possibility that the previously reported JORK advantage may hinge on the type of memory test being administered (see McCabe & Soderstrom, 2011, for more on this issue). Nevertheless, of current interest was whether JORKs were less sensitive to irrelevant cues than JOLs.

⁵ The observed differences in restudy choices between JOL and JORK participants provide evidence against the argument that volume was an equally powerful cue in both conditions, but that high-confidence predictions were simply distributed differently across JOLs and JORKs. If this was the case, both conditions should have chosen to restudy a similar proportion of loud vs. quiet items.