What cognitive development means during adolescence
Cognitive development in adolescence is the continued change in how young people acquire, represent, evaluate, and use information. It includes increasingly complex reasoning, deliberate control of attention and behavior, awareness of the strengths and limits of one's own thinking, and the ability to interpret events from other people's viewpoints. These capacities operate together when an adolescent compares evidence, revises an opinion, plans a difficult project, recognizes uncertainty, or responds to a disagreement. Research describes broad developmental tendencies, but the pace and expression of change depend on the person, the task, previous learning, and social opportunities. [National Academies, 2019] [Folker et al., 2025]
Reasoning concerns how a person draws inferences, tests explanations, and evaluates conclusions. Metacognition concerns knowledge and monitoring of one's own cognitive activity, including judgments about learning, accuracy, and confidence. Perspective taking concerns representing another person's knowledge, beliefs, intentions, or viewpoint. Executive functions—such as working memory, inhibitory control, and flexible switching—help coordinate these activities. Because researchers measure each construct in different ways, improvement on one laboratory test should never be treated as proof of an across-the-board increase in intelligence or maturity. [Weil et al., 2013] [Hollarek and Lee, 2022] [Tervo-Clemmens et al., 2023]
Adolescence is defined differently across research traditions, educational systems, and societies. The teenage years provide a useful reference, while the developmental transition often extends into the twenties. Age should be distinguished from pubertal timing, grade level, access to education, responsibility, and lived experience. A young person may reason exceptionally well about a familiar scientific problem yet need substantial support with an unfamiliar bureaucratic decision. Such variation is a central feature of developmental science rather than an anomaly. [National Academies, 2019]
From childhood reasoning to complex and reflective thought
Historical theories and what contemporary research adds
Jean Piaget's account of formal operational thought placed hypothetical and abstract reasoning at the center of adolescent cognitive change. Its enduring contribution was to identify the capacity to reason about possibilities, propositions, and systematic combinations as an important research problem. Contemporary evidence treats these skills as multidimensional and strongly shaped by instruction, language, practice, and domain knowledge. Performance does not follow a universal, irreversible shift into one mode of thinking; people of different ages can use sophisticated reasoning in familiar areas and simpler heuristics when demands become unusual or high. [National Academies, 2019] [Dumontheil et al., 2010a]
One way to see the difference is to compare explaining why a familiar machine stopped working with designing a fair experiment that separates several possible causes. The latter requires keeping alternative hypotheses in mind, identifying which observations distinguish them, and changing one's conclusion when the evidence changes. Adolescents can learn increasingly complex versions of this reasoning through science, mathematics, debate, and everyday problem solving. Success depends on both conceptual understanding and executive resources: knowing how to test a claim and having the attention and working memory to carry out that test are related achievements. [Dumontheil et al., 2010a] [National Academies, 2019]
Deductive, inductive, and probabilistic reasoning
Deductive reasoning examines what follows necessarily from stated premises. Inductive reasoning uses observations to support generalizations that remain revisable. Probabilistic reasoning considers uncertain outcomes, their relative likelihood, and the quality of information available. In real adolescent decisions, these forms often interact: a student evaluating a rumor may ask whether a conclusion follows logically, how representative the examples are, and what further evidence would change the judgment. Developmental progress involves learning when each form of inference is appropriate, not replacing everyday intuition with formal logic in every situation. [National Academies, 2019]
Reasoning also involves identifying hidden assumptions, distinguishing the credibility of a source from the appeal of its message, and recognizing that two people may draw different conclusions from incomplete evidence. When a young person learns to say 'this explanation fits some observations, but there is another plausible account,' the achievement includes cognitive flexibility and a more explicit representation of uncertainty. Educational access determines how often such habits can be practiced. Argumentation, inquiry-based instruction, and feedback therefore form part of the developmental environment, although the effect of any one teaching method varies with implementation. [National Academies, 2019] [Dunlosky et al., 2013]
Relational reasoning and integration of multiple conditions
Relational reasoning involves combining relationships rather than merely remembering isolated facts. A problem might require comparing the relationship between two quantities with the relationship between two other quantities, or coordinating several rules whose consequences interact. In a multimethod study, Dumontheil and colleagues observed non-linear age-related variation in relational reasoning, alongside complex differences in brain activation and structure. Their behavioral study involved female participants aged 7 to 27, with a smaller imaging sample. Its results caution against reducing development to a smooth curve or attributing every performance difference to one brain region. [Dumontheil et al., 2010a]
More advanced relational reasoning is useful in interpreting an experiment, understanding an analogy, evaluating competing social obligations, or predicting how a change in one part of a system affects another. Improvement may reflect more precise representations, better integration, greater familiarity with problem formats, and strategic use of working memory. Laboratory tasks isolate selected components of this capacity; they do not provide a complete scale for judging a teenager's judgment in family, legal, academic, or clinical settings. [Dumontheil et al., 2010a] [National Academies, 2019]
Executive functions and the coordination of thought
Working memory, inhibition, and cognitive flexibility
Working memory allows information to remain accessible while it is transformed or compared. Inhibitory control supports selection among competing responses and resistance to an irrelevant but tempting answer. Cognitive flexibility helps people change rules, reframe a question, or shift strategies when conditions change. These functions support reasoning and reflective thinking, but test performance also depends on processing demands, motivation, prior experience, and measurement choices. They are related components of goal-directed cognition rather than a single switch that turns on at a particular birthday. [Tervo-Clemmens et al., 2023] [Folker et al., 2025]
A large analysis combining four datasets and 10,766 participants aged 8 to 35 found substantial executive-function development in late childhood and mid-adolescence, with performance stabilizing toward adult levels in late adolescence on the measures studied. The authors described a reproducible, nonlinear developmental pattern. This establishes an informative group-level trajectory across selected laboratory assessments; it does not mean that every nineteen-year-old has equivalent judgment in every context, that later learning ends, or that individual differences become unimportant. [Tervo-Clemmens et al., 2023]
A complementary longitudinal project followed 1,093 participants from eight sites in seven countries at ages 10, 14, 17, and 21, using working-memory, Stroop, and Tower of London tasks. Average task performance improved through adolescence and into young adulthood, while the degree of change differed considerably across sites and individuals. Longitudinal follow-up strengthens inference about within-person change and illustrates the importance of cultural and contextual heterogeneity. Differences in tasks and sampling also explain why two credible studies need not identify identical ages of stabilization. [Folker et al., 2025]
Planning and adapting to new goals
Planning combines an imagined future outcome with a sequence of feasible actions, monitoring of progress, and revision of the plan when obstacles appear. Adolescents increasingly encounter goals with delayed rewards: completing a research assignment, applying for training, maintaining relationships, or saving for a purchase. Goal-directed learning research finds varied developmental results. Some experiments show steady gains; others find adult-like adolescent performance or advantages for adolescents under particular uncertainty or reward conditions. These differences are meaningful evidence about the design of tasks and learning contexts. [Wilbrecht and Davidow, 2024]
This literature supports a view of adolescence as a period of opportunity for adaptive experimentation and learning. Practice can help adolescents discover which actions reliably produce desired outcomes, while feedback can encourage strategy changes. It also highlights why the same person may plan carefully for a personally meaningful project yet make a poorly informed choice in a novel setting. Neither observation alone captures a stable trait. An accurate interpretation asks what the adolescent knew, what incentives were present, what alternatives were actually available, and how much decision-making experience the situation required. [Wilbrecht and Davidow, 2024] [National Academies, 2019]
Metacognition: knowing and regulating one's own thinking
Monitoring, confidence, and calibration
Metacognition includes knowledge about cognitive strategies and processes, monitoring of current performance, and control of subsequent action. A student uses metacognitive monitoring when asking whether a paragraph was really understood or whether an answer is merely familiar. Metacognitive control occurs when that judgment leads to a new action: rereading a confusing explanation, testing memory without notes, seeking feedback, or changing a study method. These processes require both the opportunity to reflect and reasonably informative cues about actual performance. [Weil et al., 2013] [Dunlosky et al., 2013]
Confidence is an important but imperfect signal. A person can answer correctly while reporting low confidence or feel very certain about an incorrect inference. Researchers therefore distinguish accuracy from metacognitive sensitivity, the extent to which confidence differentiates correct from incorrect decisions. In a study of 56 participants aged 11 to 41, Weil and colleagues found improvement in this aspect of visual-decision metacognition during adolescence even when basic task performance was relatively stable. Because the study was cross-sectional and modest in size, its findings speak to one measurable ability, not a universal timetable for self-awareness. [Weil et al., 2013]
Metacognitive development also includes recognizing the limits of a strategy across different subjects. Being able to recall facts about a familiar topic is different from understanding a concept well enough to apply it in a new situation. Young people benefit from comparing predictions with results, explaining errors, and receiving feedback that connects an outcome with a specific decision process. The goal is more accurate self-evaluation and adaptive strategy selection, rather than constant self-surveillance or an expectation of perfect certainty. [Dunlosky et al., 2013] [Muncer et al., 2022]
Metacognition and academic learning
A meta-analysis of research involving adolescents aged 11 to 16 found a positive association between metacognition and mathematics performance. Across the quantitative studies the pooled correlation was approximately .37, with a 95% confidence interval from .29 to .44; the analysis also found considerable heterogeneity. Correlation does not establish whether metacognitive skills improved achievement, stronger academic performance improved monitoring, or additional factors influenced both. Differences in measurement and task complexity mattered for estimated effect sizes. [Muncer et al., 2022]
Research on learning techniques provides a complementary practical perspective. Dunlosky and colleagues identified practice testing and distributed practice as relatively robust methods for improving long-term learning across several conditions and learner groups. These findings do not imply that one technique works identically for every student or subject. Their importance for adolescent cognition lies in the relationship between monitoring and action: an effective learner can inspect what is known, choose a method suited to the material, and return later to test whether understanding persists. [Dunlosky et al., 2013]
Metacognitive instruction can be embedded in ordinary teaching: ask learners to explain the reasoning behind an answer, identify which evidence they would need to change their minds, compare a confidence judgment with subsequent performance, or describe why a strategy succeeded. These approaches cultivate reflection while keeping attention on the actual content being learned. Educational outcomes depend on the quality of the activity, teacher guidance, learning resources, and the student's prior knowledge. No isolated questionnaire score should be used to label a teenager as globally self-aware or unaware. [Muncer et al., 2022] [National Academies, 2019]
Perspective taking and adolescent social cognition
Representing another person's knowledge and beliefs
Cognitive perspective taking is the ability to represent how a situation may appear to someone with different information, beliefs, experiences, or intentions. It is related to theory of mind, mentalizing, and cognitive empathy, though their definitions and measures are not fully interchangeable. A person may know that someone holds a mistaken belief, understand the reasons for that belief, and still disagree with it. Perspective taking is therefore a reasoning skill about minds and situations; it does not require approving every viewpoint or experiencing the same emotion. [Hollarek and Lee, 2022]
Basic false-belief understanding often appears much earlier in development. Adolescence introduces more demanding applications: tracking several viewpoints simultaneously, interpreting indirect language, managing misunderstandings, and taking account of what a conversational partner can see or knows. In one computerized communication study, participants had to move objects according to directions from another person whose visual access differed from their own. The older adolescent group was still less accurate than adults in critical perspective-taking trials even though performance on a comparison condition had converged. The study included only female participants, so the results should be interpreted in that sampling context. [Dumontheil et al., 2010b]
That example shows why possession of a social concept and rapid application of it are distinct. An adolescent may correctly explain, in a calm discussion, why two friends interpreted an event differently, yet fail to update that interpretation while responding immediately to a tense message. Online use of perspective information competes with attention, memory, emotional demands, and social goals. The relevant developmental question is how these processes become coordinated under different conditions, not whether teenagers categorically possess or lack empathy. [Dumontheil et al., 2010b] [Hollarek and Lee, 2022]
What longitudinal research reveals
A systematic review by Hall and colleagues examined 21 prospective or longitudinal articles addressing adolescent perspective taking. The synthesis described a general tendency toward improvement across ages 13 to 18 and identified associations with parent and teacher relationships, self-esteem, prosocial attitudes, and aspects of psychological functioning. The included studies differed in measures, populations, and developmental trajectories; the results support continued development, not a single normative score or a fixed age at which all adolescents achieve adult perspective-taking competence. [Hall et al., 2021]
A subsequent review by Hollarek and Lee brought together longitudinal and cross-sectional findings and emphasized refinement of higher-level perspective-taking abilities during adolescence. It also highlighted how parent and peer interactions, school environments, and individual characteristics may contribute to variation. The authors called for better longitudinal designs and developmentally appropriate measures. Associations between supportive relationships and perspective taking may reflect influences in both directions: sophisticated perspective taking can support relationships, while repeated constructive interactions may provide opportunities to practice it. [Hollarek and Lee, 2022] [Hall et al., 2021]
Cognitive empathy, affective empathy, and disagreement
Cognitive empathy commonly refers to understanding another person's mental or emotional state; affective empathy concerns sharing or responding emotionally to that state. They may cooperate but need not change in parallel. Perspective taking can make conflict resolution more informed by revealing the other person's constraints and interpretations, yet it can also be used strategically in persuasion or competition. Research should therefore distinguish an ability to infer mental states from moral concern, prosocial action, and emotion regulation. An adolescent who understands why a peer is angry may still need support to negotiate a fair response. [Hollarek and Lee, 2022] [Hall et al., 2021]
In everyday life, perspective taking may appear when a young person notices that a teacher lacks information about a family problem, when a friend recognizes why a message sounded insulting, or when classmates compare experiences shaped by culture and opportunity. These examples require social knowledge as well as inferential skill. Differences in communication style, disability, language, or cultural norms can affect how a task is understood and expressed. Evaluating real competence requires attention to context rather than using one socially preferred style of eye contact, speech, or emotional expression as a universal benchmark. [Hollarek and Lee, 2022] [National Academies, 2019]
How cognitive and social development inform identity and autonomy
Reflection allows adolescents to compare present roles with possible futures, evaluate inherited beliefs, and connect experiences across time. Autobiographical reasoning describes the process of making causal, thematic, and developmental links between life events and a person's understanding of self. Habermas emphasized that these connections develop through memory, language, narrative practice, and culturally situated conversations, particularly in late childhood and adolescence. The capacity to reason about one's life can support identity exploration, but the content and timing of that exploration vary substantially across individuals and communities. [Habermas, 2011] [National Academies, 2019]
More complex reasoning may support autonomy by helping an adolescent anticipate consequences, articulate preferences, and understand how choices affect others. Autonomy also depends on real opportunities to participate, trustworthy information, material resources, and adults who offer meaningful responsibility. Cognitive skill cannot by itself remove institutional barriers or guarantee independence. A scientifically useful account considers capacities and environments together: a young person can exercise sophisticated judgment while still requiring assistance with unfamiliar or high-stakes choices. [National Academies, 2019] [Wilbrecht and Davidow, 2024]
Developing the ability to question a belief is compatible with maintaining close relationships and values. Family discussions, friendship disagreements, and classroom debates provide occasions to compare reasons, acknowledge uncertainty, and build a defensible position. A disagreement may signal engagement with a new perspective rather than a failure of development. Researchers should nevertheless avoid treating every conflict as an educational benefit: the quality of dialogue, safety of the setting, and availability of respectful alternatives matter. [Hall et al., 2021] [National Academies, 2019]
Brain development and the mechanisms of cognitive change
Adolescent cognitive change is supported by ongoing development of distributed brain networks involved in attention, memory, monitoring, learning, and social understanding. Studies of executive function examine coordinated activity across frontal, parietal, and additional systems; relational reasoning tasks recruit regions that support integration of multiple relationships. Brain imaging identifies correlations between neural measures, task demands, and development, but an activation difference does not by itself prove what caused a behavior or reveal a person's overall capacity for sound judgment. [Dumontheil et al., 2010a] [Tervo-Clemmens et al., 2023]
Neurocognitive development should be interpreted alongside changes in experience. Practice in complex reasoning, social responsibilities, physical development, schooling, stress, sleep, and access to supportive environments can influence how cognitive resources are recruited. Wilbrecht and Davidow's review of goal-directed learning describes variability across tasks and contexts, showing that adolescents may use different learning strategies under different conditions. Broad claims that an adolescent brain is uniformly unfinished or inevitably driven by impulse conceal this adaptive and heterogeneous profile. [Wilbrecht and Davidow, 2024] [National Academies, 2019]
Sensitive-period theories propose that experience may have particularly strong developmental effects within certain windows. Blakemore and Mills advanced a foundational account of adolescence and sociocultural processing. Cheng, Mills, and Pfeifer later concluded that relatively few studies directly test the strongest sensitive-period claims using designs that compare the effects of equivalent experiences across different ages. Plasticity, opportunity, and context-sensitive learning are well-supported ideas; strict claims about a unique or universal window for acquiring perspective taking remain contested. [Blakemore and Mills, 2014] [Cheng et al., 2024]
Why performance changes across situations
Motivation, emotional significance, and social attention
Adolescents may reason differently about an identical logical structure when it is embedded in an abstract worksheet, a personally meaningful goal, or an emotionally charged peer interaction. Performance depends on what information attracts attention, the immediate value of an outcome, social expectations, and the time available to deliberate. An observed change across settings may reflect recruitment of different cognitive and motivational processes rather than loss of a previously acquired skill. Experimental work on adolescent goal-directed learning underscores how reward structure and uncertainty can shift which strategies are effective. [Wilbrecht and Davidow, 2024]
The quality of thinking is also affected by how a question is posed. A multiple-choice test, spontaneous conversation, private written reflection, and open-ended group discussion make different demands on language, response speed, memory, and social interpretation. Researchers may obtain divergent developmental curves because their tasks emphasize different components. For parents and educators, this means that a young person's performance in one high-pressure moment should be interpreted with information about the setting and not immediately generalized into a statement about character or overall capacity. [Folker et al., 2025] [Hollarek and Lee, 2022]
Individual differences and the social distribution of opportunity
Individual differences arise from many interacting influences, including prior learning, accessibility of instruction, language experience, neurodevelopmental variation, family and peer relationships, disability-related supports, cultural expectations, and socioeconomic resources. The seven-country executive-function study found substantial differences between sites and individuals in developmental change. This evidence supports attention to local conditions and discourages the assumption that a developmental average represents a universal timetable for every student. [Folker et al., 2025] [National Academies, 2019]
A fair developmental assessment asks whether a person has encountered the concepts required by a task, can understand its instructions, has enough time to respond, and can demonstrate competence in an accessible format. This is particularly important when comparing multilingual learners, students with disabilities, or communities with different schooling opportunities. Cross-cultural comparisons can reveal variation, but a difference in scores may reflect differences in measurement meaning or institutional experience rather than an inherent difference in cognitive potential. [National Academies, 2019] [Hall et al., 2021]
Cognitive development in digital environments
Digital environments provide repeated occasions to evaluate claims, infer an absent person's intentions, manage attention, and reflect on how one's statements are interpreted. They also vary enormously in design: private conversation, collaborative learning, passive exposure to content, competitive gaming, and algorithmically recommended material place different cognitive demands on users. A broad measure such as daily screen time is too coarse to establish whether a specific adolescent's reasoning or perspective taking is improving or declining. The scientifically relevant variables include activity, content, social context, prior characteristics, and the direction of possible effects. [National Academies, 2019] [Hollarek and Lee, 2022]
Consider an ambiguous text exchange. The writer knows their own intention, while the recipient sees only a short message and may infer a very different tone. Productive interpretation requires representing incomplete information and considering alternative explanations. Similarly, an adolescent assessing an online claim can use metacognitive questions: What is the source? What evidence supports it? What would disconfirm it? How confident should I be? These are applications of established cognitive principles; they should not be mistaken for evidence that every particular app has a measured developmental benefit or harm. [Dunlosky et al., 2013] [Hollarek and Lee, 2022]
Conversational AI and other automated systems add a new source of plausible explanations and feedback. They can be used as materials for comparison and critique: a learner may ask for multiple solutions, check a claim against primary evidence, and explain why one answer is stronger. Such practices still require external verification, awareness of model errors, and meaningful human teaching or support. Current cognitive-development findings do not establish a universal causal effect of AI interaction on adolescent reasoning, self-awareness, or social understanding. Claims about lasting consequences require direct longitudinal or experimental investigation of clearly defined uses. [Wilbrecht and Davidow, 2024] [National Academies, 2019]
How scientists study cognitive change
Cross-sectional, longitudinal, and experimental methods
Cross-sectional research compares people of different ages at one time. It is useful for identifying age-associated patterns and developing hypotheses, but differences may reflect schooling, recruitment, or cohort effects rather than within-person development. Longitudinal research follows participants across time and is better suited to studying trajectories and temporal ordering, while remaining vulnerable to attrition, practice effects, confounding, and changing measurement meaning. The contrast between small experimental tasks, large multi-dataset analyses, and repeated international assessments shows why interpretation should combine designs rather than privilege one method in isolation. [Weil et al., 2013] [Folker et al., 2025] [Tervo-Clemmens et al., 2023]
Experimental tasks can isolate particular mechanisms. Relational-reasoning paradigms vary the number of relationships to integrate; metacognitive tasks compare confidence with actual accuracy; perspective-taking tasks distinguish the participant's information from what another person can know. Randomized instructional interventions can test whether a teaching strategy improves selected outcomes. Even a well-controlled experiment may have limited real-world generalizability if its stimuli, time pressures, or incentives differ sharply from ordinary social decisions. [Dumontheil et al., 2010a] [Dumontheil et al., 2010b] [Weil et al., 2013] [Dunlosky et al., 2013]
Measurement validity, statistical averages, and uncertainty
A single task score is not the same as an entire cognitive construct. Researchers must consider reliability, ceiling effects, age-appropriate complexity, cultural validity, and whether two groups interpret instructions similarly. Repeated administration may improve scores through task familiarity rather than general cognitive growth. Average developmental curves are useful descriptions of samples; they do not define thresholds for individual competence, legal responsibility, or clinical diagnosis. Strong conclusions emerge when converging results survive changes in task, population, analytic method, and time. [Tervo-Clemmens et al., 2023] [Hall et al., 2021] [National Academies, 2019]
The evidence is strongest for continued refinement in several executive, metacognitive, and social-cognitive abilities across adolescence. Evidence for exact age boundaries or a universal stage sequence is weaker. Associations between relationships and cognitive outcomes do not independently identify causal direction. Claims that contemporary digital environments alter long-term development remain dependent on the quality of direct, activity-specific studies. Maintaining these distinctions is a scientific requirement, because developmental mechanisms are probabilistic and sensitive to context. [Folker et al., 2025] [Hollarek and Lee, 2022] [Cheng et al., 2024]
Implications for adolescents, families, and educators
Supporting reasoning and intellectual autonomy
Adolescents benefit from opportunities to reason about genuine questions, compare alternatives, defend a conclusion with evidence, and revise a position after encountering stronger information. Adults can model how to separate facts, assumptions, and predictions without demanding agreement as proof of maturity. Asking 'what information would help decide?' often creates a clearer learning opportunity than supplying an immediate judgment. The amount of scaffolding should reflect the task's difficulty and consequences, not a blanket assumption about what teenagers are capable of understanding. [National Academies, 2019] [Wilbrecht and Davidow, 2024]
A developmentally responsive classroom can make uncertainty discussable. Teachers may request a claim, its supporting evidence, a plausible counterexample, and a revised conclusion. Students can compare how the same data support different interpretations while learning that some interpretations are better supported. Such routines help make reasoning visible, but effective implementation depends on disciplinary knowledge and constructive feedback. Explicit teaching of subject content and logical structure remains essential; open discussion by itself cannot substitute for learning the underlying material. [Dunlosky et al., 2013] [National Academies, 2019]
Strengthening metacognitive learning
A practical approach to metacognition connects prediction, action, and feedback. Before a task, students estimate what they know and identify a strategy; afterward, they compare their confidence with performance and decide what to practice next. Retrieval practice and spacing across days are well-supported approaches for many learning goals. Keeping a record of recurring errors may help direct attention to misunderstood concepts, while recognizing correct reasoning reinforces calibration. Feedback should identify a changeable process rather than attach a global intelligence label to the learner. [Dunlosky et al., 2013] [Muncer et al., 2022]
Making room for perspective taking
Perspective-taking opportunities arise in literature, history, civic discussion, cooperative projects, and personal relationships. A useful exercise asks participants to identify what each person knows, what each person wants, what assumptions are being made, and how additional information could change the interpretation. Respectful disagreement makes these distinctions easier to explore. Positive interaction may support social-cognitive growth, although existing longitudinal associations cannot prove that any single conversation routine causes a generalized improvement in empathy. [Hall et al., 2021] [Hollarek and Lee, 2022]
Parents and caregivers can invite adolescents into decisions with increasing complexity, explain constraints transparently, and acknowledge when a young person has a stronger argument. For high-stakes choices, shared deliberation can combine the adolescent's preferences with adults' experience and institutional knowledge. This approach recognizes developing agency while maintaining appropriate support. Decisions involving safety, serious health concerns, or legal consequences may require specialist advice irrespective of apparent cognitive sophistication. [National Academies, 2019]
Developmental variation and appropriate assessment
Reasoning style, school performance, self-confidence, and perspective taking differ among adolescents without constituting a mental disorder. An isolated reasoning error, a disagreement with adults, or difficulty predicting another person's reaction cannot establish a diagnosis. When learning difficulties or changes in functioning are persistent and significant, an appropriate evaluation considers developmental history, educational access, language, sleep, stressors, disability accommodations, and relevant health factors. Specialized diagnostic assessment belongs to qualified clinicians and educational professionals using validated methods. [National Academies, 2019]
Research tasks used to study executive functions, confidence, or theory of mind are not interchangeable with clinical screening instruments. Their group-level findings should not be used to identify a particular adolescent as having a disorder or to infer capacity from age alone. Support can be personalized through accessible instruction, collaborative problem solving, and observation across meaningful settings. This remains consistent with the central scientific finding: adolescent cognition reflects developing capacities interacting with available contexts, and development continues through experience. [Tervo-Clemmens et al., 2023] [Folker et al., 2025]
Frequently asked questions
At what age do adolescents develop abstract reasoning?
Abstract reasoning develops through a long period with substantial variation across skills, situations, and educational experience. Many adolescents become increasingly proficient at coordinating hypotheses and relations, but a fixed birthday or a universal formal-operations milestone is not supported as a complete description of individual competence. Familiarity with a topic and opportunities for structured practice influence what a young person can demonstrate. [Dumontheil et al., 2010a] [National Academies, 2019]
What is the difference between metacognition and intelligence?
Metacognition refers to monitoring and controlling one's thinking, whereas intelligence is a broader and differently operationalized family of abilities. Someone may solve a problem accurately while judging their confidence poorly, or may recognize an error and use feedback effectively despite difficulty with a particular task. Their measures overlap in some contexts but should not be treated as identical. Metacognitive calibration can continue to refine during adolescence. [Weil et al., 2013]
Does adolescent perspective taking continue to improve after childhood?
Yes. Review evidence and experimental studies indicate continued development in advanced perspective-taking tasks during adolescence, with individual variation and some improvements extending toward adulthood. Recognizing that a person can hold a false belief is less demanding than using that perspective rapidly in a complex interaction. Family, peers, and school provide meaningful settings in which these skills are expressed and practiced. [Hall et al., 2021] [Dumontheil et al., 2010b]
Why can a teenager make a thoughtful argument and still make an impulsive choice?
Reasoning quality depends on more than whether an abstract rule is understood. A time-pressured or highly meaningful choice changes demands on attention, incentives, learned habits, and opportunities to reflect. Adolescents may demonstrate adult-like performance in one goal-directed task and a different strategy in another. A sound developmental explanation examines the situation and available supports before attributing the contrast to stable individual incapacity. [Wilbrecht and Davidow, 2024] [National Academies, 2019]
Can reflection and study strategies improve adolescent thinking?
Instruction can help students monitor understanding, select effective learning practices, and test conclusions against evidence. Practice testing and distributed practice have comparatively robust support for improving learning outcomes, while the transfer of a specific strategy to every form of judgment is uncertain. Cognitive development is supported by meaningful opportunities, accurate feedback, and accumulated knowledge. [Dunlosky et al., 2013] [Muncer et al., 2022]
Conclusion
Cognitive development in adolescence consists of interacting changes in reasoning, executive coordination, metacognitive monitoring, and social perspective taking. Research documents continued refinement while also revealing striking heterogeneity across tasks, populations, and contexts. The most informative scientific model recognizes active learning, the value of emerging autonomy, and the importance of environments that permit young people to evaluate evidence, correct errors, understand others, and use their capacities in meaningful decisions. [Folker et al., 2025] [Hollarek and Lee, 2022] [National Academies, 2019]
Related Articles
Cluster Adolescent Psychology and Identity Development
Emotion Regulation in Adolescence: Development and Challenges
Adolescence in the Age of AI: Identity, Social Comparison, Learning, and AI Companions
Adolescent Development and Criminal Responsibility
Adolescence and Psychological Development: Early, Middle, and Late Stages
Puberty and Psychological Development: Emotions, Identity, and Social Change
Adolescent Moral Development: Values, Judgment, and Prosocial Behavior
References
Blakemore, S.-J., and Mills, K. L. (2014). Is adolescence a sensitive period for sociocultural processing? Annual Review of Psychology, 65, 187–207. https://doi.org/10.1146/annurev-psych-010213-115202
Cheng, T. W., Mills, K. L., and Pfeifer, J. H. (2024). Revisiting adolescence as a sensitive period for sociocultural processing. Neuroscience and Biobehavioral Reviews, 164, 105820. https://doi.org/10.1016/j.neubiorev.2024.105820
Dumontheil, I., Apperly, I. A., and Blakemore, S.-J. (2010). Online usage of theory of mind continues to develop in late adolescence. Developmental Science, 13, 331–338. https://doi.org/10.1111/j.1467-7687.2009.00888.x
Dumontheil, I., Houlton, R., Christoff, K., and Blakemore, S.-J. (2010). Development of relational reasoning during adolescence. Developmental Science, 13, F15–F24. https://doi.org/10.1111/j.1467-7687.2010.01014.x
Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., and Willingham, D. T. (2013). Improving students’ learning with effective learning techniques: Promising directions from cognitive and educational psychology. Psychological Science in the Public Interest, 14, 4–58. https://doi.org/10.1177/1529100612453266
Folker, A., Bertrand, C., Hong, Y., et al. (2025). A longitudinal study of adolescent-to-young adult executive function development in seven countries. Developmental Science, 28, e70040. https://doi.org/10.1111/desc.70040
Habermas, T. (2011). Autobiographical reasoning: Arguing and narrating from a biographical perspective. New Directions for Child and Adolescent Development, 2011(131), 1–17. https://doi.org/10.1002/cd.285
Hall, H. K., Millear, P. M. R., Summers, M. J., and Isbel, B. (2021). Longitudinal research on perspective taking in adolescence: A systematic review. Adolescent Research Review, 6, 125–150. https://doi.org/10.1007/s40894-021-00150-9
Hollarek, M., and Lee, N. C. (2022). Current understanding of developmental changes in adolescent perspective taking. Current Opinion in Psychology, 45, 101308. https://doi.org/10.1016/j.copsyc.2022.101308
Muncer, G., Higham, P. A., Gosling, C. J., Cortese, S., Wood-Downie, H., and Hadwin, J. A. (2022). A meta-analysis investigating the association between metacognition and math performance in adolescence. Educational Psychology Review, 34, 301–334. https://doi.org/10.1007/s10648-021-09620-x
National Academies of Sciences, Engineering, and Medicine. (2019). The Promise of Adolescence: Realizing Opportunity for All Youth. National Academies Press. https://doi.org/10.17226/25388
Tervo-Clemmens, B., Calabro, F. J., Parr, A. C., Fedor, J., Foran, W., and Luna, B. (2023). A canonical trajectory of executive function maturation from adolescence to adulthood. Nature Communications, 14, 6922. https://doi.org/10.1038/s41467-023-42540-8
Weil, L. G., Fleming, S. M., Dumontheil, I., Kilford, E. J., Weil, R. S., Rees, G., Dolan, R. J., and Blakemore, S.-J. (2013). The development of metacognitive ability in adolescence. Consciousness and Cognition, 22, 264–271. https://doi.org/10.1016/j.concog.2013.01.004
Wilbrecht, L., and Davidow, J. Y. (2024). Goal-directed learning in adolescence: Neurocognitive development and contextual influences. Nature Reviews Neuroscience, 25, 176–194. https://doi.org/10.1038/s41583-023-00783-w