Psychology Ph.D. Dissertations

Assessing Metacognitive Illusions: Fluency, Timing, and Judgments-of-Learning

Date of Award


Document Type


Degree Name

Doctor of Philosophy (Ph.D.)



First Advisor

Richard Anderson (Advisor)

Second Advisor

Lynn Darby (Other)

Third Advisor

Dale Klopfer (Committee Member)

Fourth Advisor

Laura Leventhal (Committee Member)


The present study aimed to reconcile two hypothetical mechanisms driving JOL delay effects. The first hypothesis is the monitoring-dual-memories (MDM) hypothesis proposed by Dunlosky and Nelson (1992), which states that increased accuracy of delayed judgments of learning (JOLs) occurs because delayed JOLs activate the same memory storage system as the memory task itself (i.e., long-term memory). The second hypothesis is the accessibility model proposed by Koriat (1993) which states that delayed JOLs are more accurate because they increase retrieval fluency by reinforcing memory activation. Fluency research (e.g., Ball, Klein, & Brewer, 2014; Mueller, Dunlosky, Tauber, & Rhodes, 2014; Reber & Greifeneder, 2017) has not previously applied the accessibility model, but the model may explain fluency’s effects on metacognitive illusions, such that increased processing leads to increased encoding fluency creating a false sense of knowing. This dissertation presents two experiments and a combined analysis in which I investigated the effects of fluency and JOL delay on the size of metacognitive illusions measured in ways that replicated previous research and in ways that are novel in learning research. Through an interaction between JOL timing and fluency, the MDM hypothesis explains the retrieval side of the memory process whereas the accessibility model explains the encoding side of the memory process. The remainder of the findings generally supported the MDM hypothesis. The present results also established a new avenue for investigating metacognitive illusions and call into question the findings of previous research. Specifically, participants’ prediction of their future memory performance may not be as poor as previously thought. Implications for these findings and future directions are discussed.