Skip to content
PaperFren

Does knowing how well you're doing improve during adolescence?

Open paper intelligence

The ability to judge accurately when you are right or wrong improved with age across adolescence and then levelled off in adulthood, even though task performance itself was held constant.

Source

The development of metacognitive ability in adolescence

Weil LG, Fleming SM, Dumontheil I, et al. · Consciousness and cognition · 2013

doi.org/10.1016/j.concog.2013.01.004Read the full paper ↗158 citationscc by

Study at a glance

Design
Cross-sectional — Adolescents and adults compared on a two-interval contrast pop-out task with confidence ratings; performance clamped near 71% by staircase; metacognition measured as area under the type II ROC.
N
N=56 · 28 adolescents aged 11-17 and 28 adults aged 19-41 included after four exclusions.
Population
Healthy adolescents and adults recruited in the UK through schools and word of mouth
Outcome
Metacognitive ability (Aroc, how well confidence tracks accuracy) and confidence bias (Broc)

Structured fields used in claim comparison tables when every cited study has a complete layer.

What they did

Adolescents and adults watched two brief displays of striped patches and decided which display contained the patch with higher contrast, then rated their confidence on a six-point scale. A staircase adjusted difficulty so that everyone got about the same proportion correct, separating metacognition from raw ability. Metacognitive ability was measured as how well each person's confidence distinguished their correct from incorrect answers, and age effects were tested within and between age groups.

What they found

Accuracy was held near 71% for everyone, and task difficulty and perceptual sensitivity were unrelated to age. Metacognitive ability rose with age within the adolescent group and did not change within adults; surprisingly, adolescents as a group scored slightly higher than adults. Confidence bias followed a similar rise-then-fall pattern, but controlling for it did not remove the metacognition effect. Females showed somewhat better metacognition than males.

The limits

What it doesn't show

This is a small cross-sectional sample, so it compares different people of different ages rather than following individuals as they develop, and the within-adolescent correlation was only just significant. It measures metacognition for one visual task, so it may not reflect metacognition about memory, reasoning or learning, and no link to school learning or mentalising was tested. The unexpected adolescent advantage over adults and the sex difference were not predicted and need replication in larger samples.

Key terms

Metacognition
Thinking about one's own thinking, including monitoring how well one is performing.
Staircase procedure
A method that makes a task harder after correct answers and easier after errors to keep accuracy at a fixed level.
Type II signal detection
An analysis of how well confidence ratings discriminate one's own correct from incorrect responses.
Aroc
The area under the type II ROC curve; 0.5 means confidence carries no information about accuracy, higher means better metacognition.
Confidence bias
An overall tendency to give high or low confidence ratings regardless of accuracy.

Flashcards

1 / 10

0 of 10 answers reviewed

Research intelligence for this paper

See its role on concept claims, tensions it is part of, placement history, and related discoveries.

Open paper intelligence

Quiz yourself

1 / 5

Why did the researchers use a staircase procedure?

Common questions

Why hold performance constant?

When people get more answers right they tend to know it, so without clamping accuracy, better metacognition could just reflect better performance.

Did teenagers actually do better than adults?

As a group adolescents had slightly higher metacognitive scores, but the key pattern was improvement with age within adolescence, and the group difference was unexpected.

Does this tell us about studying and exams?

Not directly; the task was a visual judgement, and the authors only speculate that better metacognition could support self-regulated learning.

More on Metacognition