Metacognition

Monitoring or control of one's own cognition. Metacognitive monitoring estimates aspects of current or future performance, while metacognitive control uses such estimates to guide behaviour. Metamemory applies these functions to remembering; uncertainty responses provide a way to study some of them without verbal reports. Successful behaviour alone need not establish introspection or subjective awareness.

An individual's expressed certainty in a judgment. Its relation to objective correctness is measured separately from the certainty itself. Metacognitive calibration concerns numerical agreement with outcomes; metacognitive sensitivity concerns discrimination between better and worse performance.

A computation that evaluates an action using partially shared evidence about the action-producing system. First-order decisions and confidence need not have identical internal samples; calling the computation second order does not imply conscious verbal reasoning.^[ref]

Confidence inferred from agreement among sampled representations supporting a response. Internally consistent evidence can systematically favour a wrong answer, separating confidence from correctness.^[ref]

The ability of confidence to distinguish a person's correct from incorrect responses, or to rank items by later success. It is distinct from metacognitive calibration. A type 2 receiver operating characteristic measures discrimination in a binary-correctness task; other metamemory designs may need other measures.

The first-order sensitivity index that, in a fitted signal detection theory model, would generate the observed confidence discrimination. Model assumptions and sampling uncertainty matter; raw confidence–accuracy association alone does not estimate it.^[ref]

A model-based comparison of meta-d-prime with first-order discrimination, often their ratio. Interpretation becomes unstable when the denominator is close to zero. Efficiency, metacognitive calibration and overall confidence are different quantities.

If confidence $C$ is replaced by a strictly increasing function $g(C)$, the attainable threshold partitions have the same ordering and ties. Consequently the type 2 receiver operating characteristic and its area under the ROC curve are unchanged. Equivalently, $P(C_{\mathrm{correct}}>C_{\mathrm{error}})+\frac{1}{2}P(C_{\mathrm{correct}}=C_{\mathrm{error}})$ is invariant. Calibration need not be preserved: adding a positive constant to probabilities within their allowed range changes their absolute values while preserving ranks. A merely nondecreasing transformation can merge distinct reports and requires additional care.

A receiver operating characteristic treating high confidence as a prediction that a first-order answer was correct. Varying confidence threshold $t$ traces the false-positive and true-positive probabilities in the displayed expression. Raw discrimination can depend on first-order task performance; it is not automatically a performance-independent measure of monitoring quality.

For correctness indicator function $Y$ and stated probability $C$, ideal calibration means that answers assigned confidence $c$ are correct with conditional probability $c$. Overall overconfidence is $\mathbb{E}[C-Y]>0$; a zero overall difference does not establish calibration at every confidence level. Estimated calibration depends on sample size and how reports are grouped.

Confidence evaluated after making a response. It may use evidence, further processing and information about the action that were not all available at the moment of the response.

A normative Bayesian inference model of retrospective confidence. For two exhaustive alternatives with equal prior probabilities, let $L=\log [p(E\mid H_1)/p(E\mid H_0)]$ be the log-likelihood ratio. Bayes' theorem gives posterior odds $e^L$. Choosing the alternative with the larger posterior yields $P(\text{correct}\mid E)=(1+e^{-|L|}{)}^{-1}$. Unequal priors add prior log odds; choosing a different action changes which posterior probability is appropriate. This calculation does not establish that an observer knows the true likelihoods.

Metacognitive monitoring and metacognitive control applied to memory encoding, memory retention and memory retrieval. Different judgments predict different outcomes; a prediction of future recall need not agree with confidence in an answer currently retrieved.

A theory in which metamemory judgments infer likely performance from available cues, rather than necessarily inspecting a memory trace directly. Cue use and cue validity need not coincide.^[ref]

An aspect of the learner's current processing experience, such as retrieval ease, used in judging memory.

A property of the learning conditions, such as study time or repetitions, used in judging memory.

A property of the material, such as the difficulty or relatedness of paired words, used in judging memory.

An estimate that currently unrecalled information could be recognized or recovered later. It differs from confidence in an answer already produced and from a prediction made while initially studying it.

A prediction, made during or after study, of whether material will be remembered on a later test. Specify the delay, the test and whether the target remains visible; these determine what information the learner can use.

Improved itemwise predictive accuracy of judgments of learning made after a delay under suitable conditions. The available cue matters: delaying a cue-alone judgment can be useful even when delaying a cue–target judgment gives little benefit.^[ref]

Regulation of cognition using information about one's own processing: allocating further study, seeking information, answering or declining a test. A change in a monitoring report does not by itself demonstrate a beneficial change in control.

An action permitting a subject to decline a difficult test, seek information or select an explicit uncertainty option. Reward, delay and task contingencies influence the action; an uncertainty response is an operational behaviour rather than a direct observation of subjective awareness.

Under linear expected utility, let a correct answer yield $r$, an error yield $-w$, and declining yield $s$, with $r+w>0$. Answering has expected payoff $pr-(1-p)w$, so it is preferred exactly when $p>(s+w)/(r+w)$. Equality gives indifference. Different utility, time costs or beliefs change this criterion; behaviour across differently rewarded tasks cannot be compared as if the threshold were fixed.

Estimating aspects of one's own cognition, such as whether an answer is correct or whether material can later be recalled. Metacognitive monitoring should be distinguished from metacognitive control, which changes behaviour in response to such estimates.

Awareness of one's current mental state, such as noticing that attention has drifted. Off-task thought and awareness of that thought can be measured separately.^[ref]