Heightened Adolescent Risk-Taking? Insights From Lab Studies on Age Differences in Decision-Making

Abstract

Surveys concur that adolescents disproportionately engage in many real-world risk behaviors, compared with children and adults. Recently researchers have employed laboratory risky decision-making tasks to replicate this apparent heightened adolescent risk-taking. This review builds on the main findings of the first meta-analysis of such age differences in risky decision-making in the laboratory. Overall, although adolescents engage in more risky decision-making than adults, adolescents engage in risky decision-making equal to children. However, adolescents take fewer risks than children on tasks that allow the option of opting out of taking a risk. To reconcile findings on age differences in risk-taking in the real-world versus the laboratory, an integrative framework merges theories on neuropsychological development with ecological models that emphasize the importance of risk exposure in explaining age differences in risk-taking. Policy insights and recent developments are discussed.

Keywords

risky decision-making risk-taking age differences meta-analysis adolescence policy

An up-to-date review extends meta-analytic findings on age differences in risky decision-making in the lab: Overall, teens and children take equal risks—which has implications for policy.

Key Points

Controlling for task characteristics, adolescents and children engage equally in risky decision-making, but this is not consistent with real-world findings.

However, when adolescents can decide to opt out of taking a risk, and choose a safe option instead, they choose the safe option more often than children do.

If and when children encounter the same risk exposure as adolescents in the real world, they may actually end up engaging in more risk behavior than adolescents.

Early adolescents take more laboratory risks than mid-late adolescents, which is inconsistent with real-world surveys finding the opposite (older adolescents engage in more risk behavior); perhaps the real-world increase in risk-taking with age is partially due to increased risk exposure with age.

Finally, overall results showed that adolescents engaged in more risky decision-making than adults, and especially given an immediate consequence of the risk-taking.

These results could be relevant for policy makers involved in the juvenile justice system, particularly when deciding developmentally appropriate categorization of adolescents, for example, establishing at what age youth should be treated as adults.

Introduction

I wish that there were no age between ten and three and twenty or that youth would simply sleep out the rest; for there is nothing in between but getting wenches with child, wronging the ancestry, stealing and fighting.

—Shakespeare, The Winter’s Tale, 1600 a.d.

If we fast-forward to today, the same real-world risk behaviors (e.g., delinquency) still show growth and/or peaks during adolescence. The adverse consequences of such risk behaviors on youth mental health, education, and career prospects can be severe. For instance, such heightened adolescent risk behavior could start a vicious cycle that by interrupting a youth’s education, subsequently limits adult career prospects (Caspi, Wright, Moffitt, & Silva, 1998). Hence, knowledge about the causes of adolescent risk behavior is of scientific, policy, and societal relevance.

One interpretation of trends in adolescent risk behavior is that it reflects increases in the propensity for risk-taking (making potentially harmful choices). That is, when adolescents engage in risky behavior, such as drug use or unprotected sex, this expresses a preference for risk—reflected in risky decision-making that results in risk behavior. Although scientists and lay people often use the terms risk behavior and risk-taking interchangeably, we will clarify this distinction.

The Ecology of Risk Exposure

Surveys and laboratory studies report conflicting patterns of risk-taking with age. This review aims to resolve this discrepancy and its policy implications, by examining the ecology of risk exposure.

Scientific studies on risk-taking primarily use self-reports of engaging in risk behaviors. This methodology obscures true propensity to engage in risks, due to age-dependent access to risk-conducive situations (i.e., risk exposure) in the real world. For example, access to alcohol depends on age, because of minimum-age laws. Indeed, the transition to adolescence usually accompanies increasing independence and thus greater access to risk-conducive situations (e.g., parties where alcohol is served). Such a transition—by itself—might account for adolescents’ increased engagement in risky behavior (compared with children). Similarly, individuals experience even more independence when they transition from adolescence to emerging adulthood. Hence, this could at least partially explain why emerging adults (e.g., college students) sometimes engage in certain types of risks (e.g., substance use) to a greater extent than adolescents.

However, this ecological aspect of risk exposure (or risk opportunity) and how it relates to age differences in risk-taking is often absent in survey-based correlational studies (e.g., epidemiological studies). Survey data make it challenging to assess someone’s true propensity to take risks because age is confounded with risk opportunity. Accordingly, such data cannot answer, Would children engage in equal or more risk behavior than adolescents if they were given the same opportunity to do so?

In contrast to surveys’ correlational data, laboratory studies can make risk exposure equal for all participants, regardless of age. In such settings, children versus adolescents versus adults have equal opportunity to take risks, which should isolate age differences in risky decision-making unconfounded by differential exposure to risky situations. Thus, such controlled settings will come closer to reflecting true propensity for taking risks.

The issues with quantifying age differences in real-world risk-taking motivated a meta-analysis (i.e., Defoe, Dubas, Figner, & van Aken, 2015) of age differences in risky decision-making in more controlled settings, such as the laboratory. Such an analysis summarizes in a quantitative fashion the results of many laboratory studies conducted with differing age groups from childhood to adulthood. Furthermore, a meta-analysis of such laboratory studies could identify the direction and effect size of any age trends in risky decision-making. Finally, a meta-analysis can identify theoretical factors that moderate age differences in risky decision-making, thereby explaining age differences in risk behavior.

The first meta-analysis (i.e., Defoe et al., 2015) of its kind to pursue such questions yielded some surprising findings. The current review extends the main findings of this meta-analysis and extrapolates some policy insights. Before proceeding, specifying some controversial terms is essential: the definition of “adolescence” and “risk.”

Adolescence

The traditional age span of adolescence, namely 11 to 19 years, is more or less the period in which many Western youth begin and finish (junior-) high-school. Adolescence, long recognized as a heterogeneous period, consists of early, middle, and late adolescence. Hence, Defoe et al. (2015) divided adolescence into early (11-13 years) versus mid-late adolescence (14-19 years). Children in the meta-analysis were ages 5-9; adults were ages 20-65.

Risk

Various definitions describe what risk entails. Moreover, cultures differ in perceiving and engaging in risk. On one hand, the economic definition describes engaging in risks as choosing the option with the largest possible variation in outcomes (Figner & Weber, 2011). For example, in choosing whether to take an option with a certain outcome (e.g., to win US$1) versus one that could win US$2 half the time but nothing the other half, the latter would be the riskier choice, and thus be a case of risky decision-making. This objective definition for risk-taking is the typical outcome measure for “risk behavior” in laboratory decision-making (e.g., gambling tasks). Furthermore, one of the possible outcomes of a risky choice could possibly incur a loss (Figner & Weber, 2011), which comes closer to the more common lay definition of “risk.” In economic sciences, both risk-taking and risky decision-making refer to engaging in risks.

On the other hand, psychologists typically refer to engaging in risks as behavior that could risk a negative consequence (e.g., health complications, legal system encounters, school setbacks, psychological problems); this definition also overlaps with the lay and clinical definitions for engaging in risks. In psychology, and in everyday life, the terms risk-taking and risk behavior typically describe potential negative consequences, and someone who engages in risks is typically considered a “risk taker.” Classic work (Jessor & Jessor, 1977) defined risk engagement as “behavior that is socially defined as a problem, a source of concern, or as undesirable by the norms of conventional society and the institutions of adult authority, and its occurrence usually elicits some kind of social control response” (p. 33). A more recent comprehensive review of the development of risk engagement summarized, “Risk-taking is defined in the developmental literature as engagement in behaviors that are associated with some probability of undesirable results” (Boyer, 2006, p. 291).

As noted, cultural context influences these lay definitions of risk. For example, in the Caribbean (e.g., St. Maarten), most islands adhere to a minimum drinking age of 18 years for alcoholic beverages, whereas a few years ago in the Netherlands, the legal drinking age was 16 years. Thus, drinking alcohol at age 16 would have been considered adolescent risk-taking in St. Maarten compared to the Netherlands. Accordingly, if a 16-year-old adolescent in St. Maarten made the decision to drink alcohol, such a risk behavior could be considered a consequence of risky decision-making (Petraitis, Flay, & Miller, 1995; Reyna & Farley, 2006; Reyna & Rivers, 2008). The current review uses “risk behavior” to refer to lay conceptualizations of engaging in such risks.

For pure references to the economic conceptualization of risk engagement, we will use the phrase risky decision-making. Given that real-world risky behavior generally involves making decisions under a range of contexts, scientists have developed risky decision-making tasks for laboratory study to control extraneous variables. Hence, such laboratory tasks are also experimental tasks. In general, laboratory tasks involve making decisions between multiple options that differ in levels of risks, and sometimes a non-risk option that is a “sure/safe” option. An example of such a task that is used in the laboratory to measure risky decision-making is the Framing Spinner Task.¹ This is also one of the tasks that was used in the studies included in the meta-analysis of Defoe et al. (2015).

The Meta-Analysis

The central questions concerned, first, whether adolescents take more risks than children and/or adults on risky decision-making tasks under controlled experimental settings, and second, whether theoretically derived factors might moderate these age differences. More specifically, four independent but related meta-analyses examined age-differences in risky decision-making tasks between (a) early adolescents versus children (meta-analysis 1a; 12 studies), (b) adolescents versus children (meta-analysis 1b; 21studies), (c) early adolescents versus mid-late adolescents (meta-analysis 2; 14 studies), and (d) adolescents versus adults (meta-analysis 3; 23 studies). In the studies included in the meta-analysis, all participants (e.g., regardless of age) had equal opportunity to engage in risks (i.e., each participant in a given study encountered the same tasks and instructions).

Finally, informed by theory, we examined cognitive and affective/motivational task characteristics that could moderate the main findings. These moderators could account for variability in the hypothesized age differences in risky decision-making (i.e., moderators can explain potential heterogeneity in the effect sizes). For example, a task’s motivational aspect could be whether the task provides outcome feedback after each round. Moderators are more principled and diagnostic when they come from a priori theory, as opposed to being ad hoc inventions. Such moderators in the current meta-analysis drew on two contemporary theories of heightened adolescent risk-taking: namely, fuzzy-trace theory (FTT) and neurodevelopmental imbalance models.

FTT

FTT differentiates between verbatim-based decision-making and gist-based decision-making (Reyna & Rivers, 2008; Reyna, Weldon, & McCormick, 2015). That is, decisions may use concrete memory for literal information or abstract memory for its meaning. According to this theory, cognitive maturation increases reliance on what is called gist-based intuition, which favors risk aversion (i.e., preference for the sure gain). Verbatim-based decision-making decreases with age, and this type of decision-making is associated with more risk-taking. Gist-based intuition increases with experience as children age. Extrapolating from FTT derives the following age differences in risky decision-making:

Adolescents take fewer risks than children.

Early adolescents take more risks than late adolescents.

Adolescents take more risks than adults.

Neurodevelopmental Imbalance Models

Neurodevelopmental imbalance models differentiate between cognitive control versus socioemotional (reward) brain systems (e.g., Casey, Galván, & Somerville, 2016; Shulman et al., 2016; Somerville, Jones, & Casey, 2010; Steinberg, 2007). These models further posit that the reward system of adolescents is hyperresponsive—as reflected in a reward-seeking peak observed in adolescents, perhaps due to ongoing pubertal-maturational changes in the brain (Somerville et al., 2010; Steinberg, 2007; for a review, see Crone & Dahl, 2012). Conversely, adolescents’ cognitive control system is still slowly developing. Accordingly, in emotionally arousing “hot” contexts (e.g., a situation with a potential reward), adolescents’ hyperresponsive reward processing system overrides the cognitive control system, ultimately increasing risk-taking. Thus, these models posit that emotions affect risk-taking, particularly for adolescents, and that adolescents are less capable of top-down cognitive control than adults but more emotionally influenced than both children and adults. Taken together, extrapolating from neurodevelopmental imbalance models derives the following age differences in risky decision-making:

Adolescents take more risks than children.

Early adolescents take more risks than late adolescents (e.g., extrapolating from: Crone & Dahl, 2012; Somerville et al., 2010).

Adolescents take more risks than adults.

Results of the General Meta-Analyses

This section concerns the general meta-analyses, which held task characteristics equal, that is, not considering moderators. In contrast to the two theoretical frameworks and real-life accounts of risk behavior, early adolescents and children engaged in equal levels of risky decision-making (meta-analysis 1a). Next, consistent with both neurodevelopmental imbalance models and FTT, but in contrast to real-world reports of risk behavior, early adolescents engaged in more risky decision-making than mid/late adolescents. Finally, and in accordance with both theoretical frameworks, and with real-life accounts, adolescents engaged in more risky decision-making than adults. Put another way, children = early teens > late teens > adults.

Moderation Analyses Inspired by the Two Theoretical Frameworks

Follow-up meta-analyses (i.e., meta-regression analyses) tested whether moderator variables qualified the reported age results and nonresults. No significant moderators (e.g., task characteristics) emerged for the comparison between early adolescents versus children. As noted, early adolescents (11-13 years) and children (5-10 years) engage in equal levels of risky decision-making, and none of the investigated moderators altered these findings. Thus, contrasting the early adolescents versus children did not support the FTT or imbalance models.

Despite no overall age differences in risky decision-making when comparing all adolescents (ages 11-19) with children (ages 5-10), moderation emerged: age differences depended on task characteristics (moderators). Namely, adolescents took fewer risks than children on tasks that provided a sure/safe option, which is consistent with FTT (Reyna & Rivers, 2008). Specifically, although adolescents and children generally engaged in equal levels of risk, the results changed, given an opportunity to reason qualitatively about risks: When the task provided a sure option (a “no risk” option) versus a “some risks” option, adolescents chose the sure option more often compared with children.

No significant moderators informed the comparisons of the early adolescents versus mid/late adolescents, however. That is, overall, early adolescents take more risk than mid/late adolescents, and theory-based moderators could not explain these age differences.

Finally, adolescents particularly took more risks than adults on tasks with immediate outcome feedback on rewards and losses. Reasoning from neurodevelopmental imbalance models, feedback on rewards could be driving these age differences in risk-taking, as adolescents might be hyperresponsive to rewards, heightening their risky decision-making (Somerville & Casey, 2010; Steinberg, 2007).

To conclude, neurodevelopmental imbalance models and FTT cannot fully explain age differences in risky decision-making in the lab. As one possibility, taking risk exposure into account might help bridge the findings of age differences in risky decision-making in the lab versus risk behavior in the real-world. Specifically, whereas existing theories might explain age differences in risky decision-making between adolescents and adults, they do not completely capture the complexity of age differences in risk-taking between children and adolescents, which may particularly be confounded with risk exposure in the real world.

A Hybrid Approach

Accordingly, inspired by the meta-analytic findings, a hybrid theory, the Developmental Neuro-Ecological Risk-taking Model (DNERM), appeared as a conclusion of the meta-analysis (Defoe et al., 2015). This model emphasizes that an interaction with a crucial “risk exposure” (whether physical or social) factor could explain age differences in risk-taking—in addition to the already described individual developmental factors such as cognitive- and emotional-control. Furthermore, risk exposure could be both physical and social (Defoe, 2016). Namely, physical risk exposure could include access to risk-conducive situations, and social risk exposure could include affiliation with deviant friends. Ongoing research will test these hypotheses.

Other Moderators: Recent Developments

Besides the hypotheses put forward by DNERM, at least two issues related to the context of a risk-taking scenario in the lab could further moderate age differences in risk-taking. The first relates to risk ambiguity and individual differences, while the second relates to the influence of peer presence.

Adaptive exploration

After the meta-analysis was published, a new theory emerged, the life span wisdom model, which posits that neurodevelopmental imbalance only characterizes a subset of adolescents: Imbalance leads to maladaptive risk-taking (e.g., risk-taking that leads to addiction and other antisocial behavior) in adolescents with cognitive control difficulties that were already evident in childhood (Romer, Reyna, & Satterthwaite, 2017).

In contrast, the risk-taking that is driven by a peak in sensation-seeking during adolescence is a form of adaptive exploratory behavior that is essential for learning and wisdom development rather than a result of imbalanced neurocognitive development. Finally, consistent with FTT, this model suggests that adolescents generally take more risks than adults because they apply less gist-based decision-making than adults, as a result of their lack of experience (or wisdom) compared with adults.

According to this model, when the probabilities of risky choices are ambiguous (i.e., not provided in the lab), adolescents will generally engage in more risky decision-making than children or adults because adolescents will view such a scenario as novel and worthy of exploration, factors that are tied to more adaptive risk-taking (Romer et al., 2017).

Future studies could also consider investigating the effects of task difficulty (see Romer et al., 2017) and framing a risk-taking scenario in terms of losses versus gains (see, for example, Reyna & Brainerd, 2011), as these issues could further moderate age differences in risky decision-making (Romer et al., 2017).

Peer presence

The second issue that could moderate age differences on risky decision-making tasks concerns social emotions: whether risky decision-making takes place under conditions that are cognitively cold and static (i.e., the typical lab contexts) versus emotionally charged social situations, where real-world risk-taking among adolescents often occurs (Gardner & Steinberg, 2005).

Social neurodevelopmental imbalance models focus on the relationship between peers and perceived rewards in adolescence. These models predict that adolescents’ hypersensitivity to rewards becomes even stronger when adolescents are among peers, leading to more risk-taking (Steinberg, 2010). The first lab study to demonstrate the significant effects of peer presence reported that when adolescents (13-16 years) performed a risky driving task in the presence of peers (versus alone), their risky choices increased—compared with when emerging adults (18-22 years) and adults (24+) performed the same task with peers (Gardner & Steinberg, 2005; see also Chein et al., 2011). However, like most studies on peer presence effects on risk-taking, Gardner and Steinberg (2005) did not compare adolescents versus children. Thus, peer presence might or might not also amplify risk in children.

At the time of the meta-analysis, not enough studies tested whether age differences in risk-taking were moderated by peer presence. Subsequent studies have sometimes replicated peer presence effects on risk-taking, although results are conflicting as to how peers influence risk-taking (e.g., via peer presence, peer pressure) and on which risk-taking behavior. Although one study found an increase in risk-taking in the presence of peers (Shepard, Lane, Tapscott, & Gentile, 2011), other studies did not (e.g., Kretsch & Harden, 2014), while still other studies find a peer effect only when the peer displayed a clear preference for risk-taking (Bingham et al., 2016; Centafanti, Modecki, MacLellan, & Gowling, 2014; but not Ouimet et al., 2013).

Peer presence effects have also shown mixed effects with other tasks. Peer presence effects have emerged among adolescents on gambling (Smith, Chein, & Steinberg, 2014), although other studies have reported mixed findings (Somerville et al., 2018). Finally, the Balloon Analogue Risk Task (BART) tests risk-taking by participants inflating a cartoon-balloon as much as possible without bursting it. Similar to many of the results for risky driving, adolescents took more risks on the BART, when encouraged by a peer, compared with either being alone or in a peer’s presence but without encouragement (Reynolds, MacPherson, Schwartz, Fox, & Lejuez, 2014). Although, in general, BART showed no effects of peer presence in a sample of 13- to 16-year-old males (Kessler, Hewig, Weichold, Silbereisen, & Miltner, 2017), if anything the adolescents became more cautious after having a success in the peer condition.

Thus, the effects of peer presence on risk-taking behavior depend on the specific situational context. Taken together, these lab studies suggest that the effects of peers might only occur for some adolescents who may be more vulnerable to risky behavior or more reactive to peer presence or given direct pressure by peers to engage in risk behavior. Another study of real-world teen driving used in-car video monitored naturalistic teenage driving (Simons-Morton, Ouimet, & Zhang, 2011). Risky driving was actually lower in the presence of teenage passengers, compared with being alone; however, risky driving more than doubled when driving with a risk-prone friend and was substantially less when driving with an adult (for reviews, see Romer, Lee, McDonald, & Winston, 2014; Romer et al., 2017).

Considered together, peer presence might function as social risk exposure particularly if the peers are deviant (e.g., if they have a preference for risks). Hence, as suggested by DNERM, future studies should investigate whether such social risk exposures (and physical risk exposures) might interact with adolescent neuropsychological development (cognitive and emotional control) to predict heightened risk-taking. In any case, context must be taken into account when examining adolescent risk-taking. Moreover, while several studies focused on younger or middle adolescents, many of the reviewed studies focused on late adolescents and still found peer effects—despite increasing capacities to resist peer pressure at these ages. In addition, interventions aimed at reducing risk-taking behavior should focus on how peers can encourage increased participation in positive activities (Reynolds et al., 2014). In fact, several studies have also shown that peer presence effects equally exist for prosocial behavior (e.g., van Hoorn, van Dijk, Meuwese, Rieffe, & Crone, 2014).

Summary

(Early) Adolescents’ Versus Children’s Risky Decision-Making

When task characteristics were held equal (no moderators were considered), (early) adolescents and children engaged in equal levels of risks. This finding is surprising, as we are not aware of theories that predict this, at least in part because in the real-world adolescents typically take more risks than children. However, when comparing adolescents with children, one ecological factor that is perhaps evident in all cultures is that parents supervise their adolescents less than their younger children. Adolescents have more exposure to risk-conducive situations (i.e., more “risk opportunities” or “risk exposure”), such as more affiliation with deviant friends and/or access to substances than do children. Prevailing developmental neuropsychological theories often overlook this ecological factor. Laboratory studies show that, holding constant such risk exposure differences, children and adolescents become equally susceptible to engaging in risks. These meta-analytic results cannot say for sure why that is. Presumably, the factors heightening risk-taking for adolescents and for children would differ (Defoe et al., 2015), as children and adolescents differ in their neuropsychological development (as predicted by FTT and neurodevelopmental imbalance models), but they additionally differ in their physical and social ecology too (as further predicted by DNERM).

Also, when adolescents (11-19 years) have the option of opting out of taking a risk, and choosing a safe option instead, they choose to do the latter more often than children. Thus, these results suggest that if children would encounter the same amount of risk-conducive situations as adolescents in the real world, they might actually end up engaging in more risks than adolescents (i.e., children would choose the “safe” option to a lesser extent than adolescents).

Early Adolescents Versus Mid-Late Adolescents

Often when lay persons talk about adolescents (or “teenagers”), they lump 13 to 19 year olds together. However, scientific research distinguishes between earlier versus later adolescence, or even makes three distinctions, namely early versus middle versus late adolescence. For example, mid-late adolescents engage in more delinquency than early adolescents and emerging adults (Farrington, 1986), whereas emerging adults use more substances than early and mid-late adolescents (Loeber, Farrington, & Petechuk, 2013).

The meta-analysis (Defoe et al., 2015) showed that early adolescents took more risks than mid-late adolescents. These results are generally in line with both neurodevelopmental imbalance models and FTT, but they do not mirror real-world risk-taking. Because age differences in risk-taking in the real-world are actually the opposite of what the meta-analysis found, differing risk exposure might also explain why risk behavior is more prevalent among mid-late adolescents versus early adolescents in the real world, whereas in the laboratory studies (wherein risk exposure was equal for all participants) an opposite pattern emerged.

Adolescents Versus Adults

The meta-analysis revealed that in the lab, adolescents (ages 11-19) take more risks than adults in general, and especially on tasks with immediate outcome feedback on rewards and losses. Thus, generally, a similar pattern of age differences in risk-taking between adolescents versus adults appears in the lab and the real-world. This is perhaps because risk exposure is not such a decisive factor across these two age groups. Specifically, in the real-world, especially older adolescents increasingly encounter or create more opportunities to engage in risk behaviors, and perhaps to a similar extent as adults (Defoe, 2016). Thus, at least up until the age of 19 years, adolescents appear to be more inclined to take risks than adults when they have the same opportunity.

Next, why are adolescents particularly susceptible to taking more risk than adults on decision-making tasks that provide immediate outcome feedback? Extrapolating from neurodevelopmental imbalance models, feedback on rewards may drive heightened adolescent risk-taking on these tasks. These findings further imply that prevention and intervention programs might be more effective if they immediately reward adolescents when they engage in non-risky real-world behaviors or by encouraging healthy forms of risk-taking, such as taking a challenging course or engaging in sports.

Conclusion

The current review could have implications for policy makers involved with the welfare of youth. Namely, the meta-analytic findings imply that not only do legal measures need to protect (early) adolescents from tempting risk-conducive situations but also that equal (or even more) efforts need to continue protecting children from engaging in heightened risks. At the same time, the meta-analytic findings also underscore that younger adolescents and older adolescents are two distinctive developmental groups when it comes to risky decision-making, as the former engage in more risks than the latter. This finding runs counter to age differences in risk-taking in the real-world, where older adolescents engage in more risks. However, in the real world, older adolescents clearly have more freedom in choosing or creating their environments, and minimum-age laws permit them the opportunity to engage in certain risk behaviors (e.g., to purchase and drink alcohol at the age of 16 or 18). Consequently, older adolescents might encounter more risk-conductive situations than younger adolescents; minimum ages for substances are typically 16, 18, or 21 years. If the minimum age were 13 (early adolescents) for alcohol, for example, perhaps such early adolescents would be drinking more than late adolescents, everything else being equal. Thus, legal minimum ages should be set to the upper bound as much as possible.

The social context of youth should also be considered. Namely, recent peer presence studies on heightened adolescent risk-taking further suggest that policies targeted at reducing contexts known to heighten hazardous risk-taking (such as driving with certain types of peers) and increasing contexts that are known to lessen such risk-taking (such as driving with a parent or other adult) have greater potential to reduce risks and fatalities (Lambert, Simons-Morton, Cain, Weisz, & Cox, 2014).

Finally, according to the meta-analysis, the risky decision-making of adolescents and adults differs in general (even without feedback), and especially given an immediate consequence of the risk behavior. Hence, this could be relevant information for policy makers involved in the juvenile justice system, particularly when deciding on developmentally appropriate lawful treatments for adolescents (Defoe, 2016). This includes pertinent current debates in the United States such as the appropriate legal age boundaries deciding at what age youth should be treated as adults (e.g., legal minimum age for the purchasing/use of substances such as cannabis; see also Cauffman, Donley, & Thomas, 2017).

Footnotes

Acknowledgements

The authors would like to thank Dr. Bernd Figner and Dr. Marcel van Aken for their contribution to the meta-analysis, on which the current review is based.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: Annenberg Public Policy Center, University of Pennsylvania. The meta-analysis was part of Research Project 404-10-152 was financed by the Netherlands Organization for Scientific Research.

Notes

ORCID iD

Ivy N. Defoe

References

Bingham

C. R.

Simons-Morton

B. G.

Pradham

A. K.

Almani

Falk

E. B.

. . . Albert

P. S.

(2016). Peer passenger norms and pressure: Experimental effects on simulated driving among teenage males. Transportation Research Part F: Traffic Psychology and Behavior, 41(Pt. A), 124-137.

Boyer

(2006). The development of risk-taking: A multi-perspective review. Developmental Review, 26, 291-345. doi:10.1016/j.dr.2006.05.002

Casey

B. J.

Galván

Somerville

L. H.

(2016). Beyond simple models of adolescence to an integrated circuit-based account: A commentary. Developmental Cognitive Neuroscience, 17, 128-130.

Caspi

Wright

B. R. E.

Moffitt

T. E.

Silva

P. A.

(1998). Early failure in the labor market: Childhood and adolescent predictors of unemployment in the transition to adulthood. American Sociological Review, 63, 424-451.

Cauffman

Donley

Thomas

(2017). Raising the age of majority: Raising the age, raising the issues. Criminology & Public Policy, 16, 73-81.

Centifanti

L. C. M.

Modecki

K. L.

MacLellan

Gowling

(2014). Driving under the influence of risky peers: An experimental study of adolescent risk taking. Journal of Research on Adolescence, 26, 207–222. doi:10.1111/jora.12187

Chein

Albert

O’Brien

Uckert

Steinberg

(2011). Peers increase adolescent risk taking by enhancing activity in the brain’s reward circuitry. Developmental Science, 14(2), F1-F10.

Crone

E. A.

Dahl

R. E.

(2012). Understanding adolescence as a period of social–affective engagement and goal flexibility. Nature Reviews Neuroscience, 13(9), 636.

Defoe

I. N.

(2016). The puzzle of adolescent risk taking: An experimental-longitudinal investigation of individual, social and cultural influences (Doctoral thesis). Utrecht University, The Netherlands.

10.

Defoe

I. N.

Dubas

J. S.

Figner

van Aken

M. A.

(2015). A meta-analysis on age differences in risky decision-making: Adolescents versus children and adults. Psychological Bulletin, 141, 48-84.

11.

Farrington

D. P.

(1986). Age and crime. Crime and Justice, 7, 189-250. doi:10.1086/449114

12.

Figner

Weber

E. U.

(2011). Who takes risks when and why? Determinants of risk-taking. Current Directions in Psychological Science, 20, 211-216. doi:10.1177/0963721411415790

13.

Gardner

Steinberg

(2005). Peer influence on risk taking, risk preference, and risky decision making in adolescence and adulthood: An experimental study. Developmental Psychology, 41(4), 625-635.

14.

Jessor

S. L.

(1977). Problem behavior and psychosocial development: A longitudinal study of youth. New York, NY: Academic Press.

15.

Kessler

Hewig

Weichold

Silbereisen

R. K.

Miltner

W. H. R.

(2017). Feedback negativity and decision making behavior in the Balloon Analogue Risk Task (BART) in adolescents is modulated by peer presence. Psychophysiology, 54, 260-269.

16.

Kretsch

Harden

K. P.

(2014). Pubertal development and peer influence on risky decision making. The Journal of Early Adolescence, 34(3), 339-359.

17.

Lambert

A. E.

Simons-Morton

B. G.

Cain

S. A.

Weisz

Cox

D. J.

(2014). Considerations of a dual-systems model of cognitive development and risky driving. Journal of Research on Adolescence, 24, 541-550.

18.

Loeber

Farrington

D. P.

Petechuk

(2013). Bulletin 1: From juvenile delinquency to young adult offending (study group on the transitions between juvenile delinquency and adult crime). Washington, DC: National Institute of Justice, U.S. Department of Justice.

19.

Ouimet

M. C.

Pradham

A. K.

Simons-Morton

B. G.

Divekar

Mehranian

Fisher

D. L.

(2013). The effect of male teenage passengers on male teenage drivers: Findings from a driving simulator study. Accident Analysis & Prevention, 58, 132-139.

20.

Petraitis

Flay

Miller

T. Q.

(1995). Reviewing theories of adolescent substance abuse: Organizing pieces of the puzzle. Psychological Bulletin, 117, 67-86. doi:10.1037/0033-2909.117.1.67

21.

Reyna

V. F.

Brainerd

C. J.

(2011). Dual processes in decision making and developmental neuroscience: A fuzzy-trace model. Developmental Review, 31, 180-206. doi:10.1016/j.dr.2011.07.004

22.

Reyna

V. F.

Estrada

S. M.

DeMarinis

J. A.

Myers

R. M.

Stanisz

J. M.

Mills

B. A.

(2011). Neurobiological and memory models of risky decision making in adolescents versus young adults. Journal of Experimental Psychology: Learning, Memory, and Cognition, 37(5), 1125-1142.

23.

Reyna

V. F.

Farley

(2006). Risk and rationality in adolescent decision making. Psychological Science in the Public Interest, 7, 1-44. doi:10.1111/j.1529-1006.2006.00026.x

24.

Reyna

V. F.

Rivers

S. E.

(2008). Current theories of risk and rational decision making. Developmental Review, 28, 1-11. doi:10.1016/j.dr.2008.01.002

25.

Reyna

V. F.

Weldon

R. B.

McCormick

(2015). Educating intuition: Reducing risky decisions using fuzzy-trace theory. Current Directions in Psychological Science, 24, 392-398.

26.

Reynolds

E. K.

MacPherson

Schwartz

Fox

N. A.

Lejuez

C. W.

(2014). Analogue study of peer influence on risk-taking behavior in older adolescents. Prevention Science, 15, 842-849.

27.

Romer

Lee

McDonald

C. C.

Winston

F. K.

(2014). Adolescence, attention allocation, and driving safety. Journal of Adolescent Health, 54, S6-S15.

28.

Romer

Reyna

V. F.

Satterthwaite

T. D.

(2017). Beyond stereotypes of adolescent risk taking: Placing the adolescent brain in developmental context. Developmental Cognitive Neuroscience, 27, 19-34.

29.

Shepherd

J. L.

Lane

D. J.

Tapscott

R. L.

Gentile

D. A.

(2011). Susceptible to social influence: Risky “driving” in response to peer pressure. Journal of Applied Social Psychology, 41(4), 773-797.

30.

Shulman

E. P.

Smith

A. R.

Silva

Icenogle

Duell

Chein

Steinberg

(2016). The dual systems model: Review reappraisal, and reaffirmation. Developmental Cognitive Neuroscience, 17, 103-117.

31.

Simons-Morton

B. G.

Ouimet

M. C.

Zhang

(2011). The effect of passengers and risk-taking friends on risky driving and crashes/near crashes among novice teenagers. Journal of Adolescent Health, 49, 587-593.

32.

Smith

A. R.

Chein

Steinberg

(2014). Peers increase adolescent risk taking even when the probabilities of negative outcomes are known. Developmental Psychology, 50(5), 1564-1568.

33.

Somerville

L. H.

Casey

B. J.

(2010). Developmental neurobiology of cognitive control and motivational systems. Current Opinion in Neurobiology, 20, 236-241. doi:10.1016/j.conb.2010.01.006

34.

Somerville

L. H.

Haddara

Sasse

S. F.

Skwara

A. C.

Moran

J. M.

Figner

(2018). Dissecting “peer presence” and “decisions” to deepen understanding of peer influence on adolescent risky choice. Child Development. Advance online publication. doi:10.1111/cdev.13081

35.

Somerville

L. H.

Jones

R. M.

Casey

B. J.

(2010). A time of change: Behavioral and neural correlates of adolescent sensitivity to appetitive and aversive environmental cues. Brain and Cognition, 72, 124-133. doi:10.1016/j.bandc.2009.07.003

36.

Steinberg

(2007). Risk-taking in adolescence: New perspectives from brain and behavioral science. Current Directions in Psychological Science, 16, 55-59. doi:10.1111/j.1467-8721.2007.00475.x

37.

Steinberg

(2010). A dual systems model of adolescent risk-taking. Developmental Psychobiology: The Journal of the International Society for Developmental Psychobiology, 52(3), 216-224.

38.

van Hoorn

van Dijk

Meuwese

Rieffe

Crone

E. A

. (2014). Peer influence on prosocial behavior in adolescence. Journal of Research on Adolescence, 26, 90-100.

Heightened Adolescent Risk-Taking? Insights From Lab Studies on Age Differences in Decision-Making

Abstract

Keywords

Tweet

Key Points

Introduction

The Ecology of Risk Exposure

Adolescence

Risk

The Meta-Analysis

FTT

Neurodevelopmental Imbalance Models

Results of the General Meta-Analyses

Moderation Analyses Inspired by the Two Theoretical Frameworks

A Hybrid Approach

Other Moderators: Recent Developments

Adaptive exploration

Peer presence

Summary

(Early) Adolescents’ Versus Children’s Risky Decision-Making

Early Adolescents Versus Mid-Late Adolescents

Adolescents Versus Adults

Conclusion

Footnotes

Acknowledgements

Declaration of Conflicting Interests

Funding

Notes

ORCID iD

References