Adolescent risk taking describes decisions and actions undertaken when outcomes are uncertain and some possible outcomes carry meaningful costs. Developmental psychology studies why the value assigned to novelty, reward, acceptance, autonomy, and future consequences may shift across adolescence, and why the same young person can make careful decisions in one setting and hazardous choices in another. Sensation seeking, social influence, available opportunities, and developing regulation all contribute. Their effects vary across individuals and contexts. Defoe et al., 2015 National Academies, 2019
Adolescent risk taking is an ordinary subject of developmental research, not a psychiatric diagnosis. It includes potentially harmful choices and constructive actions such as trying a demanding skill, initiating a difficult conversation, resisting an unfair group norm, or entering a competitive activity with the possibility of failure. The central question is how adolescents learn to evaluate uncertainty while gaining access to consequential decisions and social independence. Duell and Steinberg, 2019 Duell and Steinberg, 2021
What Adolescent Risk Taking Means
Risk, uncertainty, ambiguity, and opportunity
Researchers distinguish risk from several neighboring concepts. In a decision under known risk, possible outcomes and their probabilities are specified or reasonably understood. Under ambiguity, possible outcomes may be known while their probabilities remain uncertain. Under broader uncertainty, both the relevant possibilities and their likelihoods can be learned only through exploration. Everyday adolescent choices often combine all three conditions. A person deciding whether to join a new friendship group may know little about the group's expectations; someone considering an unsafe ride may underestimate the likelihood or seriousness of injury. These decisions should not be assumed to involve the same psychological calculations. van den Bos and Hertwig, 2017 Blankenstein et al., 2021
A hazardous outcome and a willingness to tolerate uncertainty are related but different. Some actions with uncertain outcomes create opportunities for learning, independence, social contribution, and identity exploration. Others impose severe, irreversible, or disproportionate costs. A developmental account examines expected benefits, perceived costs, who bears those costs, whether choice is voluntary, and whether a safer alternative can achieve a similar goal. Judging adolescents solely by whether they accepted uncertainty overlooks the quality of information, the resources available, and the consequences that followed. Duell and Steinberg, 2021 National Academies, 2019
Risk propensity, behavior, and harm are different variables
Risk propensity means a measured tendency to prefer or accept uncertain options in particular tasks or situations. Risk behavior means what a person actually does. Harm refers to a negative outcome, which may occur despite a reasonable choice or may be avoided despite a hazardous choice. An adolescent can report strong novelty seeking while adhering to safety rules, or experience an accident without showing any enduring risk-seeking disposition. The distinction matters because laboratory games, self-report questionnaires, school records, and health surveillance capture different portions of the phenomenon. Defoe et al., 2015 van Duijvenvoorde et al., 2022
Risk-related measures are also domain specific. Financial lotteries, speed choices, social confrontation, sexual decisions, sports challenges, online disclosure, and substance use involve different incentives, skills, social rules, and opportunities. A laboratory preference for an uncertain reward cannot be translated directly into an individual prediction of real-world dangerous behavior. Researchers need the relevant behavioral context and repeated observations before drawing strong conclusions. Blankenstein et al., 2021 Defoe et al., 2015
How Risk Preferences Change Across Adolescence
Age trends are distributions, not personal destinies
Population studies often find increasing reward-oriented exploration from childhood into adolescence and strengthening behavioral regulation into young adulthood. The curves vary by construct and method. In a multi-country study of more than 5,000 people aged 10 to 30, sensation seeking generally rose toward late adolescence while self-regulation improved more gradually; substantial variation remained across people and societies. These average trends do not provide a timetable for any individual teenager. Steinberg et al., 2018
A meta-analysis comparing experimental risk tasks found that adolescents took more risks than adults on average, with an effect size of about g = 0.37, while children and adolescents did not differ reliably overall. Age differences depended on how tasks presented choices and feedback. The finding challenges a simple narrative of a unique adolescent spike in every form of risk preference. Opportunities and supervision also change with age: an older adolescent may have access to a vehicle, money, or unsupervised travel unavailable to a younger child. Defoe et al., 2015
Early, middle, and late adolescence
Early adolescence often introduces pubertal change, a widening peer world, and new occasions to test abilities and social roles. Middle adolescence may involve expanding independent mobility, stronger peer affiliation, and decisions with more immediate social consequences. Later adolescence typically brings additional practical experience and more opportunities to plan for work, education, relationships, and adult responsibilities. These labels describe useful social-developmental periods rather than universal fixed stages, and timing differs greatly across countries, families, schools, and individuals. National Academies, 2019
Puberty can influence reward-related motivation and social experience, but it does not singularly determine risk taking. Calendar age, pubertal timing, access to opportunities, socioeconomic position, temperament, family expectations, and prior learning are often intertwined. An early-maturing adolescent may encounter older peers or adult-like expectations earlier; interpreting any corresponding behavior as a direct hormonal effect would overlook such changes in social exposure. Steinberg, 2008 National Academies, 2019
Sensation Seeking and Reward Motivation
Why novelty and intensity can be attractive
Sensation seeking refers to a preference for varied, novel, complex, or intense experiences and a willingness to accept some risks to obtain them. It is a motivational disposition, not a diagnosis or a synonym for poor self-control. A young person may seek intensity through music, athletics, art, performance, travel, science competitions, or social adventure. The consequences depend on context, preparation, safeguards, and available channels for exploration. Steinberg et al., 2008 Duell and Steinberg, 2019
Reward sensitivity concerns how strongly rewards are anticipated, noticed, valued, or learned from. It can include material outcomes, pleasure, social approval, mastery, and reputation. Some experimental studies show developmental differences in reward processing and affective reactions to feedback, but results depend on task design, reward magnitude, the distinction between anticipating and receiving outcomes, and the measurement of neural activity. The same neural response cannot be read as a direct measure of recklessness. Steinberg, 2008 van Duijvenvoorde et al., 2022
Sensation seeking is not impulsivity
Impulsivity is often used broadly, but scientific measures separate acting without deliberation, difficulty withholding a response, preference for immediate rewards, and poor persistence. Someone may deliberately plan an ambitious, exciting experience and still have excellent inhibitory control. Another person may act hastily in a mundane situation without enjoying novelty. These dimensions show partly distinct developmental trajectories. Treating them as one personality defect produces weak explanations and misguided support strategies. Steinberg et al., 2008 Steinberg et al., 2018
Learning from reward and loss
Decision makers update expectations as outcomes unfold. Reward learning may encourage repeating actions that proved successful, while negative feedback can discourage particular choices. In real life, feedback can be delayed or misleading: a dangerous behavior may be followed by peer approval and no immediate injury, whereas a cautious choice may be mocked despite protecting long-term interests. Such mismatches between immediate social reinforcement and later consequences help explain why providing factual warnings alone may have limited influence on behavior. The specific learning process depends on context and should be tested rather than assumed. Albert et al., 2013 National Academies, 2019
Theories of Adolescent Risk Taking
Dual-systems models
Dual-systems accounts propose that changes in incentive-related motivation and improvements in cognitive control follow partly different developmental courses. In some contexts, especially salient rewards or social evaluation, powerful incentives may be present before a young person has accumulated the same consistency of regulatory skills and life experience that typically accompany adulthood. This framework captures useful average patterns, including the interaction of social motivation and regulation. Steinberg, 2008 Shulman et al., 2016
The framework is a research model rather than a literal picture of two independently functioning brain compartments. Neurodevelopment is distributed and interconnected, and statistical tests of a simple numerical 'imbalance' face conceptual and analytic limitations. Some real-world outcomes peak later than straightforward developmental predictions. A scientifically defensible reading treats dual systems as one explanatory hypothesis to compare with task-specific learning, social opportunity, memory, valuation, and environmental accounts. Meisel et al., 2019 Shulman et al., 2016
Social-affective and contextual approaches
Social-affective approaches ask how feelings, relationships, status, and immediate incentives change the subjective value of choices. A decision can appear unattractive in isolation but worthwhile in the presence of admired peers, particularly when taking it signals courage or group loyalty. This does not require ignorance of the risks: a teenager can understand a possible loss and still assign greater weight to an immediate social reward. Studies in which peers observe decisions help isolate some of these influences experimentally. Gardner and Steinberg, 2005 Albert et al., 2013
A contextual approach also asks whether young people had a meaningful choice. Family income, neighborhood safety, transport options, disability accommodation, unequal school discipline, discrimination, available recreation, and institutional constraints can shape exposure to both hazards and opportunities. Actions classified as risky from the perspective of a well-resourced observer may reflect constrained alternatives, responsibility for siblings, or a response to environmental danger. Social conditions therefore belong inside a developmental explanation rather than serving as an afterthought. National Academies, 2019
Fuzzy-trace and qualitative judgment
Fuzzy-trace approaches emphasize developmental shifts in how people represent the essential meaning of a decision, rather than solely calculating exact probabilities. A person may increasingly rely on a general principle such as avoiding an outcome with catastrophic potential, even when the mathematical chance appears small. This perspective offers a complementary explanation for age patterns that vary with task format and whether a sure, safe alternative is visible. Its applicability to any specific behavior requires direct evidence. Defoe et al., 2015
Ecological and developmental-learning models
Ecological models consider changes in the person's competence and the opportunities, demands, norms, and constraints of the surrounding environment. Adolescents encounter novel social and practical problems as they acquire independence. Their decisions develop through experience, feedback, coaching, and access to safe practice, not merely as a byproduct of age. An accurate model needs the full interaction among person, task, moment, and setting. National Academies, 2019 Defoe et al., 2015
What Brain Research Does and Does Not Show
Plasticity, valuation, and regulation
Adolescence includes continuing changes in neural systems involved in motivation, attention, learning, emotion, and behavioral regulation. These systems operate together, and their activity depends heavily on what the situation demands. Research on reward-related circuits has identified developmental variation, but neuroimaging group averages cannot diagnose an individual adolescent's judgment, maturity, or future conduct. Brain development is compatible with considerable decision-making competence and adaptive learning throughout adolescence. Steinberg, 2008 National Academies, 2019
The widespread phrase 'the teenage prefrontal cortex is not finished' is too coarse to explain specific choices. Development has multiple time scales and is influenced by experience; task performance may already be adult-like under many conditions. A developmental difference in one laboratory neural signal neither establishes an inability to reason nor identifies a fixed age at which all decisions become competent. National Academies, 2019 van Duijvenvoorde et al., 2022
Puberty, hormones, and individual variation
Pubertal changes can affect emotional salience, physical capacities, social attention, and opportunities. Attributing risk behavior to a single hormone or to a universal pubertal sequence overstates current science. Researchers must distinguish age from pubertal status and consider gendered expectations, cultural meanings, and differences in timing. The resulting developmental associations describe groups and cannot identify a personal cause from behavior alone. Steinberg, 2008 Steinberg et al., 2018
How Peers Change Decisions
Presence, observation, norms, and explicit pressure
Peer effects are not a single process. Presence means others are nearby; observation means the adolescent believes others can see the choice; expressed norms communicate what the group values; explicit pressure consists of direct encouragement, teasing, demands, or threats. These conditions may have different effects. Classic experiments found increased risky choices among younger participants making decisions with peers, stimulating extensive work on how social audiences change reward valuation. Gardner and Steinberg, 2005
A later experiment showed that adolescents who believed an unknown peer was observing them gambled more often even when win and loss probabilities were explicitly displayed. This supports an influence of social context beyond a basic lack of probability knowledge. Yet laboratory observation does not replicate every aspect of real friendship, group power, or actual safety consequences; conclusions must remain tied to the tested condition. Smith et al., 2014
What the meta-analyses say
A meta-analysis of direct peer observation synthesized 53 research reports and more than 5,000 participants. Across 186 effect sizes, the average increase in adolescent risky choices from observation was small, g = 0.16, with high between-study heterogeneity. Effects were more pronounced when peers expressed pro-risk preferences. Thus, 'a peer is watching' does not predict a large change by itself; conveyed social norms are crucial. Powers et al., 2022
A separate preregistered meta-analysis published in 2026 synthesized 79 independent experimental reports involving more than 9,500 participants and found a larger overall peer-influence effect across behavior types, g = 0.47, with very high heterogeneity. Peer norms were a significant moderator. These pooled estimates are not interchangeable: the studies examined different manipulations and outcomes. Their shared implication is that the precise form and meaning of social influence must be specified. Meehan and Hubbard, 2026
Selection, influence, and shared environments
Peers can resemble one another because they influence each other, choose similar friends, or inhabit the same circumstances. A cross-sectional association between friends' behavior and an adolescent's behavior cannot by itself separate these processes. A study of roughly 1,700 friendship dyads illustrated the importance of controlling for friendship selection and shared factors when estimating behavioral influence. Longitudinal meta-analysis provides stronger evidence that peers predict subsequent behavioral change, while showing that average effects depend on outcomes and design. Jaccard et al., 2005 Laursen and colleagues, 2021
Peer influence can support constructive as well as hazardous choices. Young people may be encouraged to study, train, volunteer, help another person, or resist discrimination; they may also be encouraged toward dangerous stunts or harmful experimentation. Groups differ in norms, relationships, supervision, and the opportunities they create. A psychologically accurate account asks which behavior is being normalized and who benefits from it. Meehan and Hubbard, 2026 Duell and Steinberg, 2019
Belonging, reputation, and identity
Social standing can turn a seemingly trivial decision into a meaningful one. An adolescent may fear exclusion, hope for admiration, or want to present an identity as competent, brave, loyal, independent, or morally committed. The subjective stakes of the choice are therefore relational. This is especially relevant in settings where a public refusal carries reputational cost, or where a risky act provides a shortcut to visibility. It remains essential to distinguish freely chosen affiliation from coercion or exploitation. Albert et al., 2013 National Academies, 2019
Decision Making Under Known Risk and Ambiguity
Why task instructions matter
Laboratory tasks can show all possible payoffs, conceal probabilities, allow repeated sampling, provide immediate feedback, or require decisions after social cues. Each design measures a somewhat different process. A person choosing between a guaranteed small reward and a lottery with known odds may behave differently from someone exploring an unfamiliar environment. Variations in developmental findings often emerge because apparently similar studies manipulate these conditions differently. Defoe et al., 2015 Blankenstein et al., 2021
One experimental study involving 105 participants ages 8 to 22 separated decisions under described risk, ambiguity, and experience-based uncertainty. Adolescents showed greater acceptance of ambiguity and shorter search before some choices than comparison groups. The modest sample and task-specific design limit generalization, but the results demonstrate why responses to missing information deserve attention alongside responses to clearly stated odds. van den Bos and Hertwig, 2017
Immediate feedback versus delayed consequences
Laboratory rewards often arrive instantly. Everyday costs such as injury, damaged trust, academic disruption, or financial debt may be delayed, uncertain, and unevenly distributed. A teenager can accurately describe future harm while responding strongly to immediate incentives. Decisions made in the heat of a public moment may differ from private deliberation, especially if stopping to think seems costly. Education can therefore combine accurate information with realistic rehearsal of consequential decisions rather than assuming that facts alone will determine action. Steinberg, 2008 Albert et al., 2013
Relationships Among Autonomy, Identity, and Self-Regulation
Autonomy as practiced decision-making
Autonomy involves progressively directing choices and taking responsibility within relationships, institutions, and community life. It develops through supported opportunities to decide, evaluate outcomes, renegotiate commitments, and learn from mistakes. A blanket prohibition on all uncertainty can deprive young people of useful practice, while exposure to severe hazards offers little developmental justification. The educational task is to match responsibility and safeguards to circumstances and capacity. National Academies, 2019
Identity also matters: preferences, aspirations, group memberships, and moral commitments influence which risks appear meaningful. For some adolescents, speaking publicly, applying to a selective program, or reporting misconduct requires more courage than a physical challenge. Such choices can involve real social uncertainty with long-term significance. Interpretations of 'good' risk vary culturally, and the possible costs to other people require attention. Duell and Steinberg, 2021
Self-regulation is flexible and trainable
Self-regulation involves directing attention and behavior in relation to goals, standards, and circumstances. Its expression depends on sleep, stress, practice, incentives, emotional arousal, and available support. Adolescents are capable of planning, restraint, and reflective decisions; sustained regulatory performance can improve with experience and context. Individual differences may be as important as mean age trends for explaining a specific decision. Steinberg et al., 2018 National Academies, 2019
Family, School, and Community Environments
Family communication and supportive boundaries
Caregivers influence both exposure to hazardous situations and the quality of decision-making practice. Clear expectations, open communication, predictable support, and developmentally appropriate supervision can create conditions in which a young person brings uncertainty or mistakes to a trusted adult. Simply increasing surveillance may undermine disclosure when privacy and dignity are ignored. The relevant practice is responsive and proportionate: establish nonnegotiable safeguards for serious dangers while expanding choice in matters where trial, error, and feedback are reasonable. National Academies, 2019
Adults should ask what a risky choice was expected to accomplish. Curiosity, peer acceptance, relief from stress, excitement, and resistance to perceived unfairness have different implications. Understanding motives can guide a safer route to a similar developmental goal. The principle follows from motivational and contextual research, although any particular family strategy should be evaluated in its own circumstances. van Duijvenvoorde et al., 2022 National Academies, 2019
School climates and realistic opportunities
Schools can supply protected opportunities for challenge: debate, sport with appropriate safety protocols, creative performance, internships, community service, and demanding academic projects. These contexts make exploration visible, structured, and supported. They can also create opportunities to practice dissenting from a group, evaluating persuasive claims, and considering consequences. Policies that treat every departure from conformity as misconduct risk missing the developmental value of agency and problem solving. Duell and Steinberg, 2019 National Academies, 2019
Inequality and unequal risk exposure
The same decision has different consequences depending on housing, health care access, discrimination, neighborhood conditions, school resources, transportation, and family finances. A missed shift or unsafe route may be unavoidable for a student with few alternatives; a severe accident may produce greater consequences where resources for recovery are limited. Research therefore separates an individual's preference for uncertainty from structural exposure to harm. Developmental explanations and prevention efforts should examine both levels. National Academies, 2019
Digital Environments and Risk
Online contexts combine social visibility, potentially rapid feedback, imperfect information, and audiences that extend beyond a close friendship group. Public challenges, oversharing, deceptive opportunities, reputation management, and group dynamics may create new settings for familiar psychological processes. A viral clip can amplify perceived social rewards while obscuring how many peers actually approve. The scale and persistence of online exposure may alter consequences, but a generalized claim that all digital engagement increases risky behavior is not supported by this article's evidence. National Academies, 2019 Meehan and Hubbard, 2026
Risk-related online behavior should be studied through concrete exposures and outcomes: the type of platform affordance, nature of peer feedback, degree of anonymity, availability of privacy controls, and consequences in offline relationships. Cross-sectional correlations cannot establish whether exposure preceded behavior, whether preexisting preferences guided platform use, or whether both arose from common conditions. This is particularly important when discussing algorithms, social comparisons, and any new AI-mediated social setting; each requires direct evidence for the claimed mechanism. National Academies, 2019
Positive Risk Taking and Developmental Opportunity
Constructive forms of uncertain action
Positive risk taking describes socially constructive or developmentally beneficial choices with meaningful uncertainty and manageable potential costs. Examples include trying out for a team, speaking up against bullying, interviewing for a first job, sharing original creative work, joining a new community, or taking the social risk of asking for help. Positive risks may nurture skills, efficacy, relationships, and civic contribution. The construct is a developing research area, and beneficial effects should be evaluated rather than assumed. Duell and Steinberg, 2019 Duell and Steinberg, 2021
Positive and harmful risk behaviors are not opposite ends of one scale. An adolescent may show both, and constructive activities do not automatically cancel exposure to serious danger. Programs that simply replace hazardous behavior with supervised excitement require outcome evaluation. What matters is whether accessible opportunities provide the belonging, agency, challenge, or achievement that the adolescent seeks without imposing disproportionate harm. Duell and Steinberg, 2021
Prosocial risk and moral courage
A prosocial risk can involve accepting uncertainty or disapproval to help another person. Reporting harassment, refusing group cruelty, intervening safely as a bystander, or defending an excluded peer may entail reputational consequences. Neurodevelopmental perspectives suggest overlap between systems engaged by rewards and social decisions, but the study of prosocial risk is relatively young. This distinction broadens developmental discussion beyond the prevention of hazardous acts. Do et al., 2017 Duell and Steinberg, 2019
Research Methods and Evidence Quality
Experiments and causal inference
Random assignment to peer-observed versus solitary task conditions can establish whether that specific manipulation caused a difference in measured choices under the tested conditions. It cannot establish that the same effect size applies to substance use, driving, interpersonal violence, or every online situation. Ecological realism, participants' understanding of the task, and the direction of expressed peer norms all affect interpretation. Gardner and Steinberg, 2005 Powers et al., 2022
Longitudinal designs and reciprocal pathways
Longitudinal studies track earlier characteristics and later outcomes and can help separate temporal ordering from simultaneous association. Stronger peer-influence designs account for an adolescent's baseline behavior and obtain peers' behavior from external sources rather than relying only on the adolescent's perceptions. Even then, measured and unmeasured common causes may remain. Repeated reciprocal associations can reflect both adolescents shaping their environments and environments shaping them. Laursen and colleagues, 2021 Jaccard et al., 2005
Surveys, surveillance, and selection effects
Public-health surveys monitor the prevalence and distribution of behaviors and their correlates. The US Youth Risk Behavior Surveillance System provides repeated population-level indicators, including school-based patterns that can inform prevention priorities. These data are not a clinical assessment of an individual and do not identify the psychological cause of a particular choice. Population coverage, school attendance, reporting accuracy, question wording, and changing social conditions all matter for trend interpretation. CDC Youth Risk Behavior Surveillance System
Heterogeneity and measurement comparability
Meta-analytic averages conceal differences among tasks, peer conditions, samples, and definitions. For example, the 2022 peer-observation synthesis found high heterogeneity around a small average effect, and the broader 2026 peer-influence synthesis also found high heterogeneity despite a larger pooled estimate. This variation is scientifically informative: the relevant social signal and behavioral domain may matter more than a general adolescent-versus-adult label. Claims about universal susceptibility should therefore be replaced with conditional, measurable predictions. Powers et al., 2022 Meehan and Hubbard, 2026
Practical Implications for Adolescents and Adults
For adolescents: evaluating the decision
A useful approach is to describe the specific decision, what is known and unknown, the worst plausible consequence, who will be affected, and what safer alternative would preserve the desired reward. Asking whether an action would remain appealing without an audience can reveal the importance of public approval. Seeking another perspective is an exercise of judgment and autonomy, particularly when decisions involve irreversible consequences. This framework applies developmental evidence to everyday reasoning rather than functioning as a validated clinical screening instrument. Albert et al., 2013 National Academies, 2019
For parents and caregivers: support competent choice
Conversations are more useful when adults identify the teenager's goal, explain specific consequences, model evaluation of uncertainty, and preserve room for age-appropriate choice. Establishing clear safety boundaries for driving, exploitation, intoxication, and other serious dangers is compatible with encouraging challenging but manageable experiences elsewhere. Invite disclosure without automatic humiliation or catastrophic interpretation. These recommendations express a developmentally informed application of the research, and results vary with relationships and circumstances. National Academies, 2019
For educators and youth programs: design the environment
An effective developmental setting makes constructive challenges accessible and hazardous choices less attractive or less available. Supervised skill-building, peer-led positive norms, truthful risk information, equitable resources, and opportunities to practice disagreement are plausible components. Programs must still test outcomes; a well-intentioned lesson or exciting activity cannot be presumed effective because it fits a theory. Particular care is required when inequalities constrain participation. Duell and Steinberg, 2021 National Academies, 2019
For clinicians and youth specialists: examine function and safety
Risk taking alone is not a mental disorder. Professionals assessing dangerous behavior examine frequency, severity, context, functioning, coercion, relevant symptoms, environmental hazards, and immediate safety. Repeated episodes may warrant assessment when serious harm, loss of control, intoxication, marked impairment, or other concerning symptoms are present. Specialized assessment and treatment of substance use, self-harm, suicidal behavior, mood episodes, or other disorders belong to their clinical guidelines and diagnostic frameworks; no general developmental model substitutes for such evaluation. National Academies, 2019
Frequently Asked Questions
Are all teenagers naturally reckless?
No single age category predicts the behavior of every individual. On average, some forms of sensation seeking and experimental risk preference change during adolescence, but differences depend on task, opportunity, relationships, and prior experience. Many adolescents make careful and socially responsible decisions, and considerable risk-taking variance exists within every age group. Defoe et al., 2015 Steinberg et al., 2018
Is adolescent risk taking caused by an immature brain?
Brain development is one part of a broader interaction among motivation, learning, regulation, and context. The phrase 'immature brain' compresses complex trajectories into an unhelpful single label. Neuroimaging findings do not establish individual incompetence. Adolescents can reason effectively and benefit from meaningful participation in decisions affecting their lives. National Academies, 2019 Meisel et al., 2019
Do peers always make risky decisions more likely?
Peer influence depends on the behavior, social norm, relationship, and measurement method. The effect of simply being observed is small on average, and stronger effects appear when peers communicate pro-risk expectations. Peers can also reinforce constructive choices. Powers et al., 2022 Meehan and Hubbard, 2026
Can a teenager know the consequences and still choose a risk?
Yes. Knowledge of the odds and the subjective value of the options are different psychological variables. Experimentally, peer observation has increased gambling even when probabilities were provided explicitly. The result indicates that information is necessary in many contexts but cannot fully specify the incentives that drive a particular decision. Smith et al., 2014
Is sensation seeking the same as impulsivity or ADHD?
Sensation seeking is a preference for novelty or intensity. Impulsivity encompasses different aspects of response control and decision timing. ADHD is a clinical diagnostic category evaluated through established criteria and an appropriate professional assessment. A teenager who enjoys exciting experiences cannot be classified as having ADHD or another disorder on that basis. Steinberg et al., 2008
Does participation in positive risks prevent harmful risk behavior?
Constructive exploration can offer important developmental opportunities, but current evidence does not show that positive risk taking automatically substitutes for harmful risk behavior. Evaluation must consider what the experience provides and whether it reduces serious hazards under real conditions. Duell and Steinberg, 2021
When should risky behavior prompt professional assessment?
A pattern involving repeated dangerous incidents, clear functional impairment, coercion, substantial substance-related harm, or other serious psychiatric or safety concerns merits individualized professional assessment. If there is an immediate threat of severe injury or suicide, local emergency or crisis services are appropriate. This article explains developmental mechanisms and cannot determine an individual's diagnosis or level of imminent risk. National Academies, 2019
Conclusion
Adolescent risk taking emerges from changing incentives, learning, opportunities, relationships, and self-regulation. Sensation seeking can motivate exploration; peer norms can make a choice more or less attractive; known probabilities and ambiguous situations elicit different responses; and community resources influence the consequences of the options available. The strongest research therefore favors a contextual, multidimensional understanding of decision-making over a universal account of reckless teenagers. Supporting adolescents means protecting them from serious preventable harm while creating meaningful opportunities to practice judgment, responsibility, moral courage, and autonomous growth. Defoe et al., 2015 Meehan and Hubbard, 2026 National Academies, 2019
Related Articles
Adolescent Psychology and Identity Development
Peer Pressure in Adolescence: Social Influence and Resistance
Adolescence and Psychological Development: Early, Middle, and Late Stages
Adolescent Identity Development: Theories, Processes, and Evidence
Adolescent Brain Development: Plasticity, Reward, and Self-Regulation
Adolescent Self-Regulation: Impulses, Goals, and Development
References
Albert, D., Chein, J., and Steinberg, L. (2013). Peer influences on adolescent decision making. Current Directions in Psychological Science. https://doi.org/10.1177/0963721412471347
Blankenstein, N. E., Huettel, S. A., and Li, R. (2021). Resolving ambiguity: Broadening the consideration of risky decision making over adolescent development. Developmental Review. https://doi.org/10.1016/j.dr.2021.100987
Centers for Disease Control and Prevention. (2026). Youth Risk Behavior Surveillance System. https://www.cdc.gov/yrbs/
Defoe, I. N., Dubas, J. S., Figner, B., and van Aken, M. A. G. (2015). A meta-analysis on age differences in risky decision making: Adolescents versus children and adults. Psychological Bulletin. https://doi.org/10.1037/a0038088
Do, K. T., Guassi Moreira, J. F., and Telzer, E. H. (2017). But is helping you worth the risk? Defining prosocial risk taking in adolescence. Developmental Cognitive Neuroscience. https://doi.org/10.1016/j.dcn.2016.11.008
Duell, N., and Steinberg, L. (2019). Positive risk taking in adolescence. Child Development Perspectives. https://doi.org/10.1111/cdep.12310
Duell, N., and Steinberg, L. (2021). Adolescents take positive risks, too. Developmental Review. https://doi.org/10.1016/j.dr.2021.100984
Gardner, M., and Steinberg, L. (2005). Peer influence on risk taking, risk preference, and risky decision making in adolescence and adulthood: An experimental study. Developmental Psychology. https://doi.org/10.1037/0012-1649.41.4.625
Jaccard, J., Blanton, H., and Dodge, T. (2005). Peer influences on risk behavior: An analysis of the effects of a close friend. Developmental Psychology. https://doi.org/10.1037/0012-1649.41.1.135
Meehan, Z. M., and Hubbard, J. A. (2026). A meta-analysis of susceptibility to peer influence effects in adolescence: The role of method of influence and type of behavior. Journal of Adolescence. https://doi.org/10.1002/jad.70214
Meisel, S. N., Fosco, W. D., Hawk, L. W., and Colder, C. R. (2019). Mind the gap: A review and recommendations for statistically evaluating dual systems models of adolescent risk behavior. Developmental Cognitive Neuroscience. https://doi.org/10.1016/j.dcn.2019.100681
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
Powers, K. E., Schaefer, L., Figner, B., and Somerville, L. H. (2022). Effects of peer observation on risky decision-making in adolescence: A meta-analytic review. Psychological Bulletin. https://doi.org/10.1037/bul0000382
Shulman, E. P., Smith, A. R., Silva, K., Icenogle, G., Duell, N., Chein, J., and Steinberg, L. (2016). The dual systems model: Review, reappraisal, and reaffirmation. Developmental Cognitive Neuroscience. https://doi.org/10.1016/j.dcn.2015.12.010
Smith, A. R., et al. (2014). Peers increase adolescent risk taking even when the probabilities of negative outcomes are known. Developmental Psychology. https://doi.org/10.1037/a0035696
Steinberg, L. (2008). A social neuroscience perspective on adolescent risk-taking. Developmental Review. https://doi.org/10.1016/j.dr.2007.08.002
Steinberg, L., et al. (2008). Age differences in sensation seeking and impulsivity as indexed by behavior and self-report: Evidence for a dual systems model. Developmental Psychology. https://doi.org/10.1037/a0012955
Steinberg, L., et al. (2018). Around the world, adolescence is a time of heightened sensation seeking and immature self-regulation. Developmental Science. https://doi.org/10.1111/desc.12532
van den Bos, W., and Hertwig, R. (2017). Adolescents display distinctive tolerance to ambiguity and to uncertainty during risky decision making. Scientific Reports. https://doi.org/10.1038/srep40962
van Duijvenvoorde, A. C. K., van Hoorn, J., and Blankenstein, N. E. (2022). Risks and rewards in adolescent decision-making. Current Opinion in Psychology. https://doi.org/10.1016/j.copsyc.2022.101457