Curriculum·G702 Pedagogical Fundamentals·about 32 min
Metacognition: teaching people to know what they know
By the end of this lesson you can
- →Describe the 2008 experiment in which learners rated the worse technique as better after seeing their own results, and state why the judgment resisted the evidence
- →Distinguish confidence from calibration, and explain why only calibration can be measured and taught
- →Compute a learner's calibration from a set of recorded predictions, and read the curve for over-confidence, under-confidence and no discrimination
- →Run a lesson so that every learner records a confidence before an answer, and use the Academy's calibration items and baseline as the instrument
Graduate · enrolled learners
This lesson opens with Learning Concepts and Categories: Is Spacing the Enemy of Induction? 2008.
- What happened
- Nate Kornell and Robert Bjork had participants study six paintings by each of twelve artists in order to learn to recognize each artist's style. In Experiment 1a, six artists' paintings were presented consecutively, massed, and the other six were interleaved with other artists' work, spaced. On a test of classifying new paintings by artist, with feedback on each answer, spacing produced better performance (61 percent correct against 35 on the first test block). After the test, participants were asked which had helped them learn more: 78 percent did better with spacing, but 78 percent said massing was as good as or better than spacing. A follow-up found participants could not reliably remember which artists had been massed, which suggests the judgment rested on how the study phase had felt. The authors had expected spacing to hamper induction; the data said it helped, and the participants said the opposite.
- The decision point
- The learners were not uninformed. They had just taken a test, with feedback on every answer, that their own performance in it contradicted, and they still judged the smoother condition to have taught them more. The feeling of fluent learning during massed study was a stronger signal to them than the test they had just taken. That is a calibration failure: confidence and accuracy came apart, and confidence won. A learner who cannot tell what they know from what feels familiar will make the confident wrong decision at the moment it matters, and per this Academy's design that is why calibration is measured from the first hour and re-measured at the Senior gate rather than assumed.
What you will be able to answer
- →What did the 2008 painting study show?
- →Confidence versus calibration?
- →Over-confidence versus no discrimination on a curve?
- →Where is calibration measured in this Academy?
Orientation and Year One are open: anyone can read them without an account. From Year Two onward the lessons are for enrolled learners, because progress through the later years only means anything if it is tracked against a record.
It is free. We do not sell the list and there is nothing to buy at the end of it.
Sources and review
- https://journals.sagepub.com/doi/abs/10.1111/j.1467-9280.2008.02127.x
- https://pmc.ncbi.nlm.nih.gov/articles/PMC3978334/
- https://www.semanticscholar.org/paper/Learning-Concepts-and-Categories:-Is-Spacing-the-of-Kornell-Bjork/8c5a0879246f62b96f5f05746b6a0b6e4180e929
- https://bjorklab.psych.ucla.edu/wp-content/uploads/sites/13/2016/07/Kornell_Bjork_2008_PsychScience.pdf
Confidence high·Volatility low·Reviewed 2026-09-14·Owner unassigned
Contested
The spacing advantage is not universal. Kornell and Bjork themselves report a word task in which massing was better (34 percent recall against 22), and a 2014 replication, Verkoeijen and Bouwmeester's 'Is spacing really the friend of induction?', found a smaller spacing advantage on a recognition test (hit rates of 0.77 against 0.67). That replication again found most participants, 71 percent, judging massing most beneficial. The lesson relies on the metacognitive finding.
Participants were not shown a score per condition. They had item-by-item feedback during the test, and a follow-up found they could not reliably recall which artists had been massed, so 'after their own test performance had demonstrated the opposite', the paper's phrase, means after a test they had just taken, not after reading their own breakdown.
The calibration arithmetic in the worked example uses this Academy's own scorer, which implements the reading O100-04 teaches. The scorer's per-band curve and Brier score are one way of measuring calibration among several.