age differences in future orientation and delay...

Child Development, January/February 2009, Volume 80, Number 1, Pages 28 – 44

Age Differences in Future Orientation and Delay Discounting

Laurence SteinbergTemple University

Sandra GrahamUniversity of California – Los Angeles

Lia O’BrienTemple University

Jennifer WoolardGeorgetown University

Elizabeth CauffmanUniversity of California – Irvine

Marie BanichUniversity of Colorado

Age differences in future orientation are examined in a sample of 935 individuals between 10 and 30 years usinga delay discounting task as well as a new self-report measure. Younger adolescents consistently demonstrateaweaker orientation to the future than do individuals aged 16 and older, as reflected in their greaterwillingness toaccept a smaller reward delivered sooner than a larger one that is delayed, and in their characterizations ofthemselves as less concerned about the future and less likely to anticipate the consequences of their decisions.Planning ahead, in contrast, continues to develop into young adulthood. Future studies should distinguishbetween future orientation and impulse control, which may have different neural underpinnings and followdifferent developmental timetables.

According to popular stereotype, young adolescentsare notoriously shortsighted, oriented to the immedi-ate rather than the future, unwilling or unable to planahead, and less capable than adults at envisioning thelonger term consequences of their decisions andactions. This myopia has been attributed to a varietyof underlying conditions, among them, the lack offormal operational thinking (Greene, 1986), limita-tions in working memory (Cauffman, Steinberg, &Piquero, 2005), the slow maturation of the prefrontalcortex juxtaposedwith the increase in reward salienceconcomitant with the hormonal changes of puberty(Steinberg, 2008), and the fact that, relative to theamount of time they have been alive, an extension oftime into the future is subjectively experienced by anadolescent asmore distant than is the same amount oftime to an adult (i.e., 10 years into the future is nearlytwice one’s life span to a young adolescent but onlyone fourth of one’s life span to a 40-year-old;W. Gard-ner, 1993). Whatever the cause, youthful shortsight-edness has been implicated as a cause of the poorjudgment and risky decision making so often evincedby young people, used as a rationale to place legalrestrictions on the choices adolescents are permitted

to make, and suggested as one explanation for thegeneral ineffectiveness of educational interventionsdesigned to persuade adolescents to avoid varioushealth-compromising behaviors, such as smoking,binge drinking, or unprotected sex (Steinberg, 2007).

Developmental psychologists interested in youth-ful shortsightedness have generally studied it underthe rubric of ‘‘future orientation.’’ This term has beenused to refer to a collection of loosely related affective,attitudinal, cognitive, and motivational constructs,including the ability to imagine one’s future lifecircumstances (Greene, 1986; Nurmi, 1989a), thelength of time one is able to project one’s imaginedlife into the future (‘‘temporal extension’’; Lessing,1972), the extent to which one thinks about or con-siders the future (‘‘time perspective’’; Cauffman &Steinberg, 2000), the extent to which one is optimisticor pessimistic about the future (Trommsdorff &Lamm, 1980), the extent to which one believes thereis a link between one’s current decisions and one’sfuture well-being (Somers & Gizzi, 2001), the extentto which one believes he or she has control over hisor her future (McCabe & Barnett, 2000), and the ex-tent to which one engages in goal setting or plan-ning (Nurmi, 1989b). These varied but potentiallyinterrelated definitions indicate that future

This research was supported by the John D. and Catherine T.MacArthur Foundation Research Network on AdolescentDevelopment and Juvenile Justice.

Correspondence concerning this article should be addressed toLaurence Steinberg, Department of Psychology, Temple University,Philadelphia,PA19122.Electronicmailmaybesent to [email protected].

# 2009, Copyright the Author(s)Journal Compilation# 2009, Society for Research inChildDevelopment, Inc.All rights reserved. 0009-3920/2009/8001-0006

orientation, as it has been operationalized in devel-opmental studies, has components that are cognitive(e.g., the extent to which one thinks about the future),attitudinal (e.g., the extent to which one prefers long-term, as opposed to short-term, goals), and motiva-tional (e.g., the extent to which one formulates plansto achieve long-term goals; see also Nurmi, 1991;Nuttin, 1985). Although, as noted above, some re-searchers also have included an evaluative dimensionof future orientation in theirmeasures ormodels (e.g.,the extent to which an adolescent is optimistic orpessimistic about the future; Trommsdorff & Lamm,1980), this strikes us as an entirely different phenom-enon and one that is more likely linked to differencesbetween individuals in their personalities and lifecircumstances (e.g., their degree of depression, theiravailable resources) than to developmental factors.

Studies of future orientation among adolescentshave examined both age and individual differences inone or more components of the phenomenon. Withrespect to age differences, although the literature issurprisingly sparse, the suggestion that adolescentsbecome more future oriented as they get older gener-ally has been supported across studies that havevaried considerably in their methodology (see re-views in Furby & Beyth-Marom, 1992; Greene, 1986;Nurmi, 1991). First, compared to young adolescents,older adolescents think more and report planningmore about the developmental tasks of late adoles-cence and young adulthood, such as completing theireducation and going to work, and older adolescentsare better able than younger ones to talk about future-oriented emotions such as fear or hope (Nurmi, 1991).Second, questionnaire-based studies of future orien-tation that include items such as, ‘‘I often do thingsthat don’t pay off right away but will help in the longrun,’’ generally find that individuals’ tendency tothink about and consider the future increase withage (e.g., Cauffman & Steinberg, 2000). Finally, a fewstudies have examined adolescents’ attentiveness tofuture considerations when making decisions. Forexample, in a study examining age differences in legaldecisionmaking inwhich individualswere presentedwith hypothetical dilemmas (e.g., how to respond toa police interrogation when one has committeda crime), Grisso et al. (2003) found that youngeradolescents (11- to 13-year-olds) were significantlyless likely to recognize the long-term consequences ofvarious decisions than were adolescents 16 and older.

On the other hand, studies have not found consis-tent age differences in future orientation when theconstruct is operationalized in terms of individuals’ability to project themselves in the future (i.e., ‘‘tem-poral extension’’). In an early study of the phenome-

non, Lessing (1972) reported no age differences intemporal extension in a sample of 9- to 15-year-oldgirls. In contrast, Greene (1986) found an age-relatedincrease in the number of future events mentioned bythe adolescents and college students in her sample butno age differences in their length of time extension.Trommsdorff, Lamm, and Schmidt (1979) found anincrease in temporal extension over adolescence withrespect to personality and occupation (i.e., olderindividuals were able to project their personalityand occupation over a relatively longer period), butnot with respect to other domains, such as physicalhealth or appearance; this finding is consistent withother reports that the extent and nature of age differ-ences in future orientation vary as a function of thespecific aspects of life asked about and the way inwhich future orientation is operationalized (Nurmi,1992). InNurmi’s (1991) study of 10- to 19-year-olds inFinland, for example, although therewere age-relatedincreases in reports of future hopes, fears, and goals,therewere decreaseswith age in individuals’ extensionof themselves into the future. As Nurmi, Poole, andKalakoski (1994) suggest, as individuals age, theycome to understand the difficulty in making rea-listic long-term predictions about their future, andtheir projections become more conservative. As anyparent or teacher will attest, a 4-year-old is likely to bemuch more confident about what her life will be likewhen she is an adult than is a 16-year-old. Temporalextension, it seems to us, is a poor measure of futureorientation.

Studies of individual differences, as opposed to agedifferences, have examined the links between futureorientation and a diverse array of factors. Three broadthemes emerge from this work. First, individualsfrom more advantaged backgrounds (as indexed bysocioeconomic status [SES] or level of education)score higher on measures of future orientation thantheir less advantaged counterparts (Nurmi, 1987,1992). Second, there are few consistent genderdifferences in future orientation, and those that arefound are typically domain specific (e.g., in theprojection of future occupational vs. family roles;Nurmi et al., 1994; Poole & Cooney, 1987; Somers &Gizzi, 2001). Finally, individuals whose behaviorwould suggest a weaker orientation to the future(e.g., delinquent youth, youth who engage in rela-tively more risky activity) score lower on measuresof future orientation than their peers (Cauffmanet al., 2005; Somers & Gizzi, 2001; Trommsdorffet al., 1979). Probably the safest conclusion one candraw from this literature is that differences amongindividuals in their attitudes, motives, and beliefsabout the future are considerable and vary a great

Age Differences in Future Orientation and Delay Discounting 29

deal as a function of factors in addition to age ordevelopmental stage.

On the whole, then, with the exception of thespecific phenomenon of temporal extension, it ap-pears that individuals become more oriented to thefuture as they mature. There are three importantlimitations to the extant literature, however. First,the self-report measures employed often do notsystematically distinguish among the cognitive, atti-tudinal, and motivational aspects of future orienta-tion, a distinction that is important for understandingage differences in the phenomenon. In the presentstudy, we analyze data derived from a self-reportmeasure of future orientation designed to distinguishamong three related, but different, phenomena: timeperspective (whether one thinks about the future),anticipation of future consequences (whether onethinks through the likely future outcomes of one’sdecisions before deciding), and planning ahead(whether one makes plans before acting).

A second limitation in the existing literature con-cerns the age ranges studied. With only a few ex-ceptions (e.g., Cauffman & Steinberg, 2000; Grissoet al., 2003; Lewis, 1980; Nurmi, 1991), studies of agedifferences in future orientation rarely involve sam-ples that span a very wide age range. Given the factthat some have speculated that adolescents’ relativelystronger orientation to the immediate than that ofadults is due to developmental changes in rewardprocessing at puberty and to the gradual maturationof self-regulatory competence that is ongoing into themid-20s (Steinberg, 2008), it is important to studypreadolescents, adolescents, and young adults simul-taneously. In the present study, we examine agedifferences in future orientation over two decades ofthe life span, in a sample ranging in age from 10 to 30.

A third limitation of extant studies of futureorientation is that they rely mainly on individuals’self-characterizations. Although older individualsdescribe themselves as more oriented to the futurethan younger ones, this may reflect developmentaldifferences in self-perceptions. Given the widely heldstereotype of adolescents as exceedingly short-sighted, it is entirely plausible that adolescents’descriptions of themselves as relativelymore orientedto the immediate than the long term are largelya reflection of their internalization of this culturalbelief. It is therefore important to ask whether agedifferences in self-reports of future orientation areparalleled in age differences in behavior. In thepresent study, in addition to our use of a new self-report measure of future orientation, we assess theconstruct using a behavioral paradigmknownas delaydiscounting.

The delay discounting paradigm is widely usedby behavioral economists and social and clinicalpsychologists in the assessment of individuals’ pref-erence for future versus immediate outcomes. Inthis paradigm, the respondent is asked to choosebetween an immediate reward of less value (e.g.,$400 today) and a variety of delayed rewards ofmorevalue (e.g., $700 1 month from now? $800 6 monthsfrom now?), and the outcome of interest is the extentto which respondents prefer the delayed and morevaluable reward over the immediately available butless valuable one. In other words, the task assessesindividuals’ ability to pit long-term benefits againstimmediate gains, a decision that individuals ofdifferent ages face frequently (e.g., Should I go tothe party or study for my SATs? Should I get a jobafter high school or enroll in college and go into debt?Should I save for my retirement or enjoy all of myearnings now?).

It is not clear whether performance on delaydiscounting tasks reflects impulse control, orientationtoward a future goal, or a combination of both. Al-though the task is usually described as one designedto measure impulsivity (e.g., Green, Myerson, &Ostaszewski, 1999), recent work on the neural un-derpinnings of delay discounting task performanceindicate that the task actually activates two differentbrain systems: one that mediates impulsivity andreward seeking (localized mainly in limbic andparalimbic areas) and one that mediates the sort ofdeliberative and abstract reasoning presumed toundergird future orientation (localized mainly inlateral prefrontal cortical areas; McClure, Laibson,Loewenstein, & Cohen, 2004). One model for under-standing delay discounting behavior is that of a com-petition between these two systems, in whichpredominance of the former leads to preference forimmediate rewards and predominance of the latterleads to preference for delayed ones. If this is the case,delay discounting behavior should be correlatednegatively with measures of impulsivity and posi-tively with measures of future orientation. Givenrecent evidence of developmental change in bothbrain systems during adolescence (Casey, Getz, &Galvan, 2008; Steinberg, 2008), the age differences infuture orientation reported by previous investigatorsare explicable, but whether these differences are dueto decreases in impulsivity during adolescence, in-creases in orientation to a future goal, or a combina-tion of the two is not clear.

In addition to revealing how strongly oriented anindividual is to immediate versus future outcomes,the delay discounting paradigm can also be used toestimate the rate at which a future reward is

30 Steinberg et al.

discounted, known as the discount function, by repeat-ing the trials of reward choices, holding the value ofthe delayed reward constant, but varying the dura-tion of the delay (e.g., 1 day, 1 week, 1 month, 3months, 6 months, 1 year). Studies of humans, non-human primates, and other animals indicate that theshape of the discount function as applied to manytypes of decision making is not linear, but hyperbolic,with smaller delayed rewards at shorter intervalsdiscounted much more steeply than larger rewardsat longer intervals. For example, the differencebetween waiting 1 versus 2 days for a reward of $1versus $2 is subjectively experienced as greater thanthe difference between waiting 364 versus 365 daysfor a reward of $999 versus $1,000, even though ineach case, the difference in delay periods (1 day) andrewards ($1) are identical. Researchers find thatdiscount functions are not significantly different forreal as compared to hypothetical rewards, whichunderscores the relevance of the delay task to actualdecisionmaking (e.g., Johnson&Bickel, 2002; Lagorio& Madden, 2005; Madden, Begotka, Raiff, & Kastern,2003).

Although the hyperbolic delay function is foundacross individuals, the steepness with which delayedrewards are discounted varies among them. Rela-tively steeper discounting indicates that the point atwhich the participant prefers the immediate rewardto the delayed reward occurs at lower values of theimmediate reward and at shorter delay times—inother words, less preference for the larger delayedreward. Figure 1 shows two theoretical discountfunctions (one steep and the other less so) froma hypothetical study in which participants are askedto choose between a reward of $1,000 whose deliveryis delayed by some specified time period (graphedalong the x-axis) and a smaller reward deliveredimmediately. The lines show the discounted value of$1,000 (i.e., the smaller amount the participant wouldsettle for) at various delay periods, ranging from1dayto 365 days. An individual whose pattern of choices isdescribed by the steeper function is relatively moredrawn to immediate rewards, even if they are smaller,than is an individual whose pattern follows the lesssteep function; as the graph illustrates, the formerindividual discounts the value of $1,000 quite a bitafter just a short delay period.

There is a large empirical literature, mainly involv-ing adults and often with clinic samples, examiningindividual differences in discount rates. For example,a number of studies have found steeper discountingfunctions among various substance abusers, such asalcoholics, drug addicts, or heavy smokers, thanamong those of matched control groups (e.g., Bickel,

Odum, & Madden, 1999; Madden, Petry, Badger, &Bickel, 1997; Petry, 2002; Vuchinich & Simpson, 1998).Studies also find steeper discount functions amongimpulsive individuals and among individuals ofrelatively lower intelligence (de Wit, Flory, Acheson,McCloskey, & Manuck, 2007). Individuals’ discountrates show significant short-term stability (Ohmura,Takahashi, Kitamura, & Wehr, 2006).

Given the literature on age differences in futureorientation discussed earlier, one would hypothesizethat relative to adults, adolescents would evincea lower indifference point and a steeper discountfunction in delay discounting task performance, re-flecting their putatively weaker orientation to thefuture. To our knowledge, however, only one studyhas examined age differences in delay discountingduring childhood and adolescence (Scheres et al.,2006), and only one program of work has examinedage differences in delay discounting during adoles-cence and adulthood. In the Scheres et al. (2006) study,children (ages 6 through 11 years) evinced steeperdiscounting than adolescents (ages 12 through 17years). In thework comparing adolescents and adults,12-year-olds (N5 12) showed a steeperdiscount func-tion than college students (average age5 20,N5 12),who in turn showed a slightly steeper function thanolder adults (average age 5 68, N 5 12), but all agegroups showed the same, widely reported hyperbola-like discount function, leading the researchers to con-clude that age differences in the discount functionwerequantitative rather thanqualitative (Green, Fry,&Myerson, 1994; Green et al., 1999). A second study,comparing younger adults (average age 33, N 5 20)and older adults (average age 71,N5 20), did not find

$300

$400

$500

$600

$700

$800

$900

0 100 200 300

Delay (Days)

Indi

ffer

ence

Poi

nt

Steep Function

Less Steep Function

Figure 1. Illustration of steep and less steepdiscount functions overdelays from 1 to 365 days.


age differences in the discount rate, however (Green,Myerson, Lichtman, Rosen, & Fry, 1996). Together,these studies suggest that developmental differencesin delay discounting might be more apparent inchildhood and adolescence than in adulthood, butmore research, with larger samples, is clearly needed.

In the present study, we examine age differences infuture orientationusing both a self-reportmeasure andthe delay discounting paradigm. We hypothesize thatrelative to adults, adolescents will report a weakerorientation to the future and evince both a lowerindifference point and a steeper discount function onthe delay discounting task. Furthermore, we hypoth-esize that age differences in delay discounting arelinked to age differences in self-reported future orien-tation. Finally, we askwhether age differences in delaydiscounting are mediated by age differences in impul-sivity, in orientation to the future, or both.

Method

Participants

The present study employed five data collec-tion sites: Denver (Colorado), Irvine (California),Los Angeles, Philadelphia, and Washington, DC.The sample includes 935 individuals between theages of 10 and 30 years, recruited to yield an agedistribution designed both to facilitate the examina-tion of age differences within the adolescent decadeand to compare adolescents of different ages withthree specific groups of young adults: (a) individualsof traditional college age (who in some studies ofdecision making behave in ways similar to adoles-cents; M. Gardner & Steinberg, 2005); (b) individualswho are no longer adolescents but who still are at anage during which brain maturation is continuing,presumably in regions that subserve orientationtoward long-term goals (Giedd et al., 1999); and (c)individuals who are older than this putatively still-maturing group. Six individuals were droppedbecause of missing data on one or more key demo-graphic variables. For purposes of data analysis, agegroups were created as follows: 10 – 11 (N 5 116),12 – 13 (N 5 137), 14 – 15 (N5 128), 16 – 17 (N 5 141),18 – 21 (N 5 148), 22 – 25 (N 5 136), and 26 – 30 (N 5123) years, yielding a sample for the present analysisof N 5 929.

The sample was evenly split between males (49%)and females (51%) and was ethnically diverse, with30% African American, 15% Asian, 21% Latino(a),24% White, and 10% Other. Participants were pre-dominantly working and middle class. Each sitecontributed an approximately equal number of par-

ticipants, although site contributions to ethnic groupswere disproportionate, reflecting the demographicsof each site.

Procedure

Prior to data collection, all site project directorsand research assistants met at one location for severaldays of training to ensure consistent task adminis-tration across data collection sites. The project coor-dinators and research assistants conducted on-sitepractice protocol administrations prior to enrollingparticipants.

Participants were recruited via newspaper adver-tisements and flyers posted at community organiza-tions, Boy’s and Girl’s clubs, churches, communitycolleges, and local places of business in neighborhoodstargeted to have an average household education levelof ‘‘some college’’ according to 2000 U.S. Census data.Individuals who were interested in the study wereasked to call the research office listed on the flyer.Members of the research team described the nature ofthe study to the participant over the telephone andinvited those interested to participate. Given thisrecruitment strategy, it was not possible to know howmany participants saw the advertisements, what pro-portion responded, andwhether thosewho respondedare different from those who did not.

Data collection took place either at one of theparticipating university’s offices or at a location inthe community where it was possible to administerthe test battery in a quiet and private location. Beforebeginning, participants were provided verbal andwritten explanations of the study, their confidentialitywas assured, and their written consent or assent wasobtained. For participants who were younger than 18years, informed consent was obtained from eithera parent or a guardian.

Participants completed a 2-hr assessment thatconsisted of a series of computerized tasks designedto measure several executive functions (Towerof London, Stroop, Iowa Gambling Task), a set ofcomputer-administeredself-reportmeasures,ademo-graphic questionnaire, and several tests of generalintellectual function (e.g., digit span, verbal fluency).The computerized tasks were administered in indi-vidual interviews. Research assistantswere present tomonitor the participant’s progress, reading aloud theinstructions as each new task was presented andproviding assistance as needed. To keep participantsengaged in the assessment, participantswere told thatthey would receive $35 for participating in the studyand that they could obtain up to a total of $50 (or, forthe participants younger than 14 years, an additional

32 Steinberg et al.

prize of approximately $15 in value) based on theirperformance on the video tasks. In actuality, we paidall participants aged 14 – 30 years the full $50 and allparticipants aged 10 – 13 years received $35 plus theprize. This strategy was used to increase the motiva-tion to perform well on the tasks but ensure that noparticipants were penalized for their performance.All procedures were approved by the institutionalreview board of the university associated with eachdata collection site.

Measures

Of central interest in the present analyses are ourdemographic questionnaire, the assessment of IQ,a self-report measure of impulsivity, a self-reportmeasure of future orientation, and the delay discount-ing task. Given the findings of previous researchlinking future orientation to risk taking, we alsodescribe the measure used to assess various types ofrisk behavior, but we use data from this measure onlyto validate the future orientation instrument.

Demographics. Participants reported their age, gen-der, ethnicity, and household education. Individualsyounger than 18 years reported their parents’ educa-tion, whereas participants 18 and older reported theirown educational attainment, both ofwhichwere usedas a proxy for SES. The age groups did not differ withrespect to gender or ethnicity but did differ (mod-estly) with respect to SES. As such, all subsequentanalyses controlled for this variable.

Intelligence. The Wechsler Abbreviated Scale ofIntelligence (WASI) Full-Scale IQ Two-Subtest(Wechsler, 1999) was used to produce an estimate ofgeneral intellectual ability based on two (VocabularyandMatrix Reasoning) of the four subtests. TheWASIcan be administered in approximately 15 min and iscorrelated with the Wechsler Intelligence Scale forChildren (r5 .81) and theWechsler Adult IntelligenceScale (r 5 .87). It has been normed for individualsbetween the ages of 6 and 89 years. Because thereweresmall but significant differences between the agegroups in IQ and because performance on the delaydiscounting task has been found to vary with IQ, thisvariable was controlled in all subsequent analyses.

Impulsivity. A widely used self-report measure ofimpulsivity, the Barratt Impulsiveness Scale, Version11 (Patton, Stanford, & Barratt, 1995), was part of thequestionnaire battery; themeasure has been shown tohave good construct, convergent, and discriminantvalidity. Based on inspection of the full list of items(the scale has six subscales comprising 34 items) andsome exploratory factor analyses, we opted to useonly 18 items (a 5 .73) from three 6-item subscales:

motor impulsivity (e.g., ‘‘I act on the spur of themoment’’), inability to delay gratification (e.g., ‘‘Ispend more money than I should’’), and lack ofperseverance (e.g., ‘‘It’s hard for me to think abouttwo different things at the same time’’); correlationsamong the three subscales in this sample are r 5 .36for motor impulsivity and lacks delay of gratification,r 5 .32 for lacks delay of gratification and lacksperseverance, and r 5 .51 for motor impulsivity andlacks perseverance). Each item is scored on a 4-pointscale (rarely/never, occasionally, often, almost always/always), with higher scores indicative of greaterimpulsivity. The three subscales we elected not tousemeasure attention (e.g., ‘‘I am restless atmovies orwhen I have to listen to people’’), cognitive complex-ity (‘‘I amagreat thinker’’), and self-control,which theinstrument developers describe as assessing ‘‘plan-ning and thinking carefully’’ (Patton et al., 1995, p.770). We could not replicate the six-factor structure ofthe scale in our sample—which is not surprising,given that the psychometrics of this version of thescale were derived from data pooled from samplesvery different from ours in age and circumstances:introductory psychologyundergraduates, psychiatricpatients, and prisoners. In addition, we concludedthat ‘‘attention’’ and ‘‘cognitive complexity’’ were notcomponents of impulsivity as we conceptualized theconstruct, and that ‘‘self-control,’’ as operationalizedin this scale, overlapped too much with the planningsubscale of our future orientation measure (the sub-scales are in fact correlated at r 5 .33, p , .001) andwould thus make it difficult to examine the indepen-dent relations of impulsivity and future orientationto delay discounting. Confirmatory factor analysisindicated that a one-factor 18-item scale provided anadequate fit to the data within each age category andfor the sample as a whole (normed fit index [NFI] 5.912, comparative fit index [CFI]5 .952, and rootmeansquare error of approximation [RMSEA]5 .033).

Risk behavior. Our battery also included a self-report measure of risk processing adapted from onedeveloped by Benthin, Slovic, and Severson (1993).Respondents are presented with eight potentiallydangerous activities (i.e., riding in a car with a drunkdriver, having unprotected sex, smoking cigarettes,vandalism, shoplifting, going into a dangerous neigh-borhood, getting into a fight, and threatening orinjuring someone with a weapon) and asked abouttheir experience with the activity, as well as a series ofquestions concerning their perceptions of the activ-ity’s riskiness, dangerousness, and costs and benefits.For purposes of validating the future orientationinstrument, we use only the items that ask aboutactual experience and created an unweighted item


composite indicating degree of engagement in riskybehavior.

Future orientation. A 15-item self-report measure(a 5 .80) of future orientation was developed for thisprogramof research. Itemswere generated by a groupof developmental psychologists with expertise inadolescent psychosocial development and pilottested with small samples of high school studentsand college undergraduates. Slight revisions inword-ing were made on the basis of these pilot studies.Following a format initially developed by Harter(1982) to minimize socially desirable responding,the measure presents respondents with a series of 10pairs of statements separated by the word BUT andasks them to choose the statement that is the bestdescriptor. After indicating the best descriptor, therespondent is then asked whether the description isreally true or sort of true. Responses are then coded ona 4-point scale, ranging from really true for onedescriptor to really true for the other descriptor andaveraged. Higher scores indicate greater future ori-entation. Future orientation, with IQ controlled, isunrelated to performance on any of the executivefunction tasks administered to the present sample.

Items were grouped into three, 5-item subscales:time perspective (e.g., ‘‘Some people would rather behappy today than take their chances on what mighthappen in the future BUT Other people will give uptheir happiness now so that they can get what theywant in the future’’; a 5 .55); anticipation of futureconsequences (e.g., ‘‘Some people like to think about allof the possible good and bad things that can happenbefore making a decision BUT Other people don’tthink it’s necessary to think about every little possi-bility before making a decision’’; a 5 .62); andplanning ahead (e.g., ‘‘Some people think that planningthings out in advance is a waste of time BUT Otherpeople think that things work out better if they areplanned out in advance’’; a5 .70). The relatively lowalpha coefficients for these subscales is likely due tothe small numbers of items that compose each;nonetheless, appropriate caution should be takenwhen using them as separate measures. However,a confirmatory factor analysis indicated that a modelwith these three intercorrelated 5-item subscalesprovided a satisfactory and better fit to the data (CFI5 .959, NFI5 .924, RMSEA5 .033) than one with one15-item factor (CFI 5 .899, NFI 5 .865, RMSEA 5.052), despite the higher alpha of the 15-item scale(the test of the difference between the models issignificant, v2diff 5 137.488, dfdiff 5 3, p , .001). Thefull instrument, along with scoring instructions, isreprinted in the Appendix. Examination of the inter-correlations among the subscales supports the view

that these three aspects of future orientation arerelated but not identical (time perspective with antic-ipation of future consequences, r5 .44; time perspec-tive with planning ahead, r 5 .44; anticipation offuture consequences with planning ahead, r 5 .55).

Patterns of correlations between scores on thefuture orientation scale and other self-report instru-ments in the study battery support the validity of themeasure. For example, future orientation scores aresignificantly correlated with responses to items fromthe Zuckerman Sensation-Seeking Scale that concernplanning ahead (e.g., ‘‘I very seldom spend muchtime on the details of planning ahead,’’ r 5 !.40, p ,.001; ‘‘Before I begin a complicated job, I make carefulplans,’’ r 5 .42, p , .001) but not with items thatconcern thrill seeking (e.g., ‘‘I like to have new andexciting experiences and sensations even if they area little frightening,’’ r 5 !.01, ns). Similarly, futureorientation scores are significantly correlated withitems from the Barratt Impulsivity Scale that concernplanning and thinking about the future (e.g., ‘‘I planwhat I have to do,’’ r5 .41, p, .001; ‘‘I try to plan formy future,’’ r5 .44, p, .001; ‘‘I like to think about howmy life will be in the future,’’ r5 .44, p, .001), but notwith items that index nonchalance (‘‘I am carefree andhappy-go-lucky,’’ r 5 .02, ns), inattentiveness (e.g., ‘‘Idon’t pay attention,’’ r5 .00, ns), or fickleness (e.g., ‘‘Ichange my friends often,’’ r 5 !.05, ns). In addition,and consistent with other studies, in our sample,future orientation scores are positively correlatedwith SES (r 5 .09, p , .01) and negatively correlatedwith risk taking (r 5 !.22, p , .001). The negativecorrelation between future orientation and risk takingis even stronger (r5!.32, p, .001) among the adultsin our sample (aged 18 and older), who presumablyhave relatively more opportunities to engage incertain risky behaviors, such as smoking, riding incars driven by intoxicated drivers, and engagingin unprotected sex. In the present sample, the corre-lation between future orientation and IQ is quitemodest, suggesting that the measure is not simplya proxy for intelligence (r 5 .10, p , .01).

Delay discounting. The delay discounting taskwas administered on a laptop computer (The tasktook about 10 min to complete.). In our adaptation ofthe task, the amount of the delayed reward was heldconstant at $1,000. We varied the time to delay in sixblocks (1 day, 1 week, 1 month, 3 months, 6 months,and 1 year), presented in a random order. For eachblock, the starting value of the immediate rewardwas$200, $500, or $800, randomly determined for eachparticipant. The respondent was then asked to choosebetween an immediate reward of a given amount anda delayed reward of $1,000. If the immediate reward

34 Steinberg et al.

was preferred, the subsequent question presented animmediate reward midway between the prior oneand the zero (i.e., a lower figure). If the delayedreward was preferred, the subsequent question pre-sented an immediate reward midway between theprior one and the $1,000 (i.e., a higher figure). Partic-ipants then worked their way through a total of nineascending and descending choices until their re-sponses converged and their preference for the imme-diate and delayed reward are equal, at a valuereflecting the ‘‘discounted’’ value of the delayed re-ward (i.e., the subjective value of the delayed rewardif it were offered immediately; Green, Myerson, &Macaux, 2005), referred to as the ‘‘indifference point’’(Ohmura et al., 2006). An individual with a relativelylower indifference point and/or a relatively higher(steeper) discount rate is relatively more orientedtoward the immediate than the future. For eachindividual, we computed the indifference point foreach delay interval, the average indifference point,and the discount rate. As in previous studies, delaydiscounting is correlated with intelligence: In thepresent sample, the correlation between IQ andindividuals’ average indifference point is r 5 .31,p, .001, and the correlation between IQ and individ-uals’ discount rate is r5!.27, p, .001. The correlationbetween individuals’ average indifference point andtheir discount rate is, as expected, negative, r 5 !.41,p, .001, given that individualswith a higherdiscountrate are more likely to prefer smaller, immediaterewards. With IQ controlled, delay discounting per-formance is unrelated or only marginally related, r".10 to performance on the measures of executivefunctioning included in the test battery.

Results

Age Differences in Self-Reported Future Orientation

A multiple analysis of covariance, with age, gen-der, and ethnicity as independent variables; IQ andSES as covariates; and the three subscales of the futureorientation measure as intercorrelated outcomes re-vealed significant main effects for age, gender, andethnicity, but no significant two- or three-way inter-actions among these predictors (we set the alpha levelfor tests of interactions for this and all other analysesat p , .01, given the large sample size).

As expected, future orientation increases with age,multivariate, F(18, 2508) 5 3.42, p , .001, withsignificant differences seen on planning ahead, F(6,836) 5 6.56, p , .001, g2

p 5 .04; time perspective, F(6,836) 5 2.62, p , .05, g2

p 5 .02; and anticipation of

future consequences, F(6, 836) 5 4.63, p , .001, g2p 5

.03. Table 1 presents the mean and standard error fortotal future orientation scores and for each of the threesubscales for each age group, adjusted for groupdifferences in IQ and SES. Mean scores for the threesubscales are presented in Figure 2.

As Figure 2 shows, the pattern of age differences isnot identical across the three subscales. Multipleregression analyses were used to examine linear andcurvilinear (quadratic) relations between the futureorientation subscales and the age. On the measure ofplanning ahead, there is a significant linear trend (b5.194, t5 5.925, p, .001), as predicted, but a significantcurvilinear trend as well (b5 .469, t5 3.037, p, .01),indicating a decline in planning between ages 10 and15 (r5!.12, p, .05), but an increase in planning fromage 15 on (r 5 .21, p , .001). On the subscalesmeasuring time perspective and the anticipation offuture consequences, only the linear trend is

Table 1

Mean Future Orientation Subscale and Total Scale Scores by Age Group

Dependent variable Age group M SE

Planning 10 – 11 2.825abc .082

12 – 13 2.662a .073

14 – 15 2.626a .099

16 – 17 2.763ab .068

18 – 21 2.911abc .068

22 – 25 3.138abc .066

26 – 30 3.039abc .071

Temporal orientation 10 – 11 2.495ac .080

12 – 13 2.644abc .071

14 – 15 2.683abc .096

16 – 17 2.722abc .066

18 – 21 2.810bc .065

22 – 25 2.855bc .064

26 – 30 2.731abc .069

Anticipation of future consequences 10 – 11 2.859ac .078

12 – 13 2.780ac .069

14 – 15 2.875abc .094

16 – 17 2.963abc .064

18 – 21 3.018abc .064

22 – 25 3.116ab .062

26 – 30 3.199b .067

Future orientation total scale 10 – 11 2.726a .064

12 – 13 2.695a .057

14 – 15 2.728ac .077

16 – 17 2.816abc .053

18 – 21 2.913abc .053

22 – 25 3.036b .051

26 – 30 2.990bc .055

Note. Scores adjusted for groupdifferences in IQ and socioeconomicstatus. Cells not sharing a subscript are significantly different atp , .05 or better.


significant (b 5 .105, t 5 3.174, p , .01 and b 5 .204,t 5 6.322, p , .001, respectively). Post hoc pairwiseBonferroni-adjusted comparisons were conducted inorder to examine the curvilinear trend in the planningdata. These comparisons indicate significantly lowerplanning scores among adolescents between 12 and15 than among younger or older individuals; theseage differences did not vary as a function of gender orethnicity.

Although not a central focus of this report, we alsofind small but significant gender differences on allthree subscales, with females outscoring males. Eth-nic differences are significant only on the planningsubscale, withAfricanAmericans outscoring all othergroups, whose scores do not differ. Scores on the timeperspective and anticipation of future consequencessubscales (but not the planning ahead subscale) aresignificantly correlated with IQ; none of the threesubscales is correlated with SES.

Age Differences in Delay Discounting

Indifference point. As noted earlier, the indifferencepoint is the amount at which the subjective value ofthe short-term reward is equivalent to that of thedelayed reward. A repeated measures analysis ofcovariance (ANCOVA) was conducted with age,gender, and ethnicity as between-subjects factors; IQand SES as covariates; and individuals’ indifferencepoints at the six time intervals (1 day, 1week, 1month,3months, 6months, and 1 year) as thewithin-subjectsfactor. As in virtually all studies of delay discounting,we find a main effect of the repeated time factor,Greenhouse –Geisser F(5, 3555)5 6.64, p, .001, g2

p 5.01, indicating that individuals’ indifference pointsdecline as the delay interval increases. Although thereare significant age differences in average indifferencepoints, F(6, 832) 5 5.90, p , .001, g2

p 5 .04, withyounger individuals demonstrating lower indiffer-ence points than older ones, we do not find a signifi-cant interaction between the repeated time factor andage, nor do we find significant interactions betweenthe repeated time factor and gender, ethnicity, IQ, orSES, suggesting that individuals of different ages,sexes, and ethnic groups, socioeconomic back-grounds, and intelligence levels show comparablepatterns of discounting over the delay intervalsexamined here. Average indifference points are pos-itively related to IQ but not SES or gender. We alsofind small but significant ethnic differences in averageindifference points, with African Americans demon-strating a relatively lower indifference point, andAsian Americans demonstrating a relatively higherindifference point, than other groups.

As Table 2 and Figure 3 indicate, there is a near-consistent break point across all delay intervals inaverage indifference points around age 14 or 15, withindividuals aged 13 and younger generally reporting

2.4

2.5

2.6

2.7

2.8

2.9

3.0

3.1

3.2

3.3

10-11 12-13 14-15 16-17 18-21 22-25 26-30

Age

Sub

scal

e S

core

PlanningAhead

TimePerspective

Anticipation ofConsequences

Figure 2. Age differences in planning ahead, time perspective, andanticipation of future consequences.

Table 2

Discounted Value (Indifference Point) of $1,000 at Varying Delay Intervals and Average Indifference Point as a Function of Age

Age

Delay interval

1 day* 1 week* 1 month* 3 months* 6 months 1 yeary Average value*

10 – 11 years $745a $711a $631ab $492a $421ac $391abc $565a12 – 13 years $756a $707a $570a $485a $447a $331ab $549a14 – 15 years $890b $796ab $606abc $569ab $460ac $343abc $611ab16 – 17 years $905b $832b $702bc $614b $498ac $423ac $662bc18 – 21 years $930b $828b $742abcd $636b $530bc $452ac $686bc22 – 25 years $919b $822b $719abcd $607b $479abc $424ac $662bc26 – 30 years $884b $813b $691abcd $637b $547bc $442ac $669bc

Note. Values not sharing subscripts with other values in the same column are significantly different at p , .05 or better.yp , .10 for comparison of age groups. *p , .001 for comparison of age groups.

36 Steinberg et al.

lower indifference points than individuals aged 16and older and individuals aged 14 and 15 years fallingsomewhere in between. That is, there are no signifi-cant age differences in average indifference pointsamong individuals aged 16 and older, and no differ-ences between the 10- and 11-year-olds and the12- and 13-year-olds, based on Bonferroni-adjustedpairwise comparisons.

Discount rate. The discount rate (k)was computedusing the standard equation, V5A/(1 + kD), whereVis the subjective value of the delayed reward (i.e., theindifference point), A is the actual amount of thedelayed reward, D is the delay interval, and k isthe discount rate. In the current study, as is usuallythe case, the distribution of k was highly positivelyskewed (4.47); accordingly, we employed a naturallog transformation to reduce skewness to an accept-able level (!.271).

An ANCOVA, with age, gender, and ethnicity asindependent variables; IQ and SES as covariates; andthe discount rate (k) as the outcome resulted ina significant effect for age, F(6, 832) 5 6.62, p , .001,g2p 5 .05, with significant differences, as indicated by

Bonferroni-adjusted pairwise comparisons, betweenindividuals aged 13 and younger, on the one hand,

and those aged 16 and older, on the other; as withaverage indifference points, individuals aged 14 and15 fell between the two extremes and did not differfrom younger or older individuals (see Figure 4).Younger individuals evince a relatively steeper func-tion, indicating a relatively weaker future orientation.As in the analysis of indifference points, there are nosignificant age differences in the steepness of individ-uals’ discount rate after age 16. Thus, adolescentsaged 13 and younger will accept a smaller rewardthan individuals aged 16 and older in order to obtainthe reward sooner, and, relative to older individuals,younger adolescents are more likely to do so at loweramount in order to get the reward even sooner. Thispattern of age differences did not vary as a function ofgender or ethnicity. Discount rate is negatively relatedto IQ but unrelated to SES or gender. In addition, wefound a small but significant difference in discountrates between Asian and African American individ-uals, with Asians evincing a less steep (i.e., morefuture oriented) discount function than AfricanAmericans.

Consistent with previous research, we find thata hyperbolic discount function fits the data well.Across all age groups, R2 for the hyperbolic functionbased on the median indifference point at each delayis .987, p, .001. Also consistentwith past research,wefind that a hyperbolic function fits the data well foreach age group (the R2s for the hyperbolic functionbased onmedian indifference points at each delay are.984, .990, .982, .905, .971, .971, and .991 for ages 10 – 11,12 – 13, 14 – 15, 16 – 17, 18 – 21, 22 – 25, and 26 – 30,respectively); Green et al. (1999; Table 1) report R2sof .995 and .996 for children (average age 12) andyoung adults (average age 20), respectively. As haveother researchers, we also find that an exponentialfunction also fits the data well, both for the sample asa whole and for each age group (with virtuallyidentical R2s to those derived from a hyperbolic

Figure 3. Age differences in discount function.

Figure 4. Discount rate as a function of age.


function) and that both the hyperbolic and the expo-nential functions provide a superior fit than a linearfunction for each age group. Thus, age differences indiscounting are more a mater of quantity (i.e., in thesteepness of the curve) than quality (i.e., in the shapeof the curve; see also Green et al., 1994).

Relation Between Self-Reported Future Orientation andDelay Discounting

Because this is the first study to simultaneouslyassess future orientation via self-report and behav-ioral measures, we were interested in examining therelation between our measure of future orientationand the measures derived from the delay discountparadigm. Table 3 shows the bivariate and partialcorrelations (with age and IQ controlled, in order toensure that any significant correlations are not attrib-utable to the fact that the measures are significantlycorrelated with these variables) between the overallfuture orientation score, the three subscales of thefuture orientation questionnaire, and the two primaryoutcomes derived from the delay discounting task:the average indifference point and the discount rate.As the table indicates, scores on the self-report mea-sure of future orientation and behavioral measuresare significantly, albeit modestly, correlated, althoughmore strongly andmore consistently for the subscalesmeasuring time perspective or the anticipation offuture consequences than for the planning subscale.Indeed, in a multiple regression analysis in which thethree future orientation subscales are used to simul-taneously predict individuals’ indifference point(with age and IQ controlled), the temporal orientationand anticipation of future consequences scores arepredictive, but scores on the planning ahead subscaleare not (b5 .126, t5 3.39, p, .001; b5 .161, t5 4.03,p , .001; and b 5 !.059, t 5 1.49, ns, respectively). Ina comparable analysis predicting discount rate, tem-poral orientation and anticipation of consequences

scores predict discount rates in the expected direction(more future-oriented individuals discount delayedvalues less steeply), whereas planning scores areinexplicably related to discount rate in the oppositedirection (more planful individuals discount delayedvalues more steeply; b5 !.164, t5 4.39, p, .001; b5!.130, t5 3.25, p, .001; and b5 .096, t5 2.41, p, .05,respectively).

Do Age Differences in Future Orientation or ImpulsivityExplain Age Differences in Delay DiscountingPerformance?

We noted in the Introduction that performance onthe delay discounting task has been traditionallylinked to impulsivity but that recent neurobiologicalevidence paints a more complicated picture, withdelay discounting linked both to brain systems regu-lating impulse control and those regulating the sort ofabstract, deliberative reasoning implied by an orien-tation to the future. Studies of brain developmentshow continued maturation during adolescenceof brain systems that subserve both processes(Steinberg, 2008). Because adolescence is a time ofincreases in impulse control (Galvan, Hare, Voss,Glover, & Casey, 2007; Leshem & Glicksohn, 2007)and, as documented in the present report, in futureorientation, it is possible that age differences in delaydiscounting aremediated by age differences in impul-sivity, age differences in orientation to the future,or both.

In order to examine whether future orientation orimpulsivity mediates age differences in delay dis-counting, we conducted two hierarchical multipleregression analyses, in which we regressed eitherthe average indifference point or the discount rateon age, impulsivity, and future orientation, enteredinto the equation in that order, and controlling for IQand SES. (Given the correlation of .33 between impul-sivity and future orientation scores, it was important

Table 3

Zero-Order and Partial Correlations (With Age and IQ Controlled) Between Future Orientation Scale and Subscales and Delay Discounting Measures

Average indifference point Discount rate

Zero order Partial Zero order Partial

Future orientation

(full scale)

.18*** .14*** !.15*** !.12***

Planning ahead .08* .05 !.05 !.01

Temporal orientation .17*** .16*** !.18*** !.16***

Anticipation of consequences .18*** .15*** !.15*** !.11***

Note. Ns vary from 919 to 924.*p , .05. ***p , .001.

38 Steinberg et al.

to examine the impact of future orientation aftercontrolling for impulsivity, so as to isolate that aspectof future orientation that may be related to impulsecontrol rather than a genuine orientation to thefuture.) Our interest was in testing whether therelation between age and delay discounting wasdiminished when impulsivity was added into theequation and whether the relation between age anddelay discounting was further diminished by theaddition of future orientation into the model, withimpulsivity left in the equation so as to test themediating role of future orientation above andbeyond any aspect of future orientation that is relatedto impulsivity. In each case, a Sobel test was used toassess the extent and statistical significance of medi-ation (MacKinnon, Lockwood, Hoffman, West, &Sheets, 2002).

The results of these analyses indicate that agedifferences in delay discounting are significantlymediated by differences in orientation to the futurebut not by differences in impulsivity. With impulsiv-ity added to the equation, the decline in the impact ofage on the average indifference score is trivial (fromb5 .208 to b5 .199, z5 1.73, ns), as is the decline in theimpact of age on discount rate (from b5!.212 to b5!.203, z5 1.52, ns). In contrast,with future orientationadded to the equation (i.e., controlling for impulsiv-ity), the impact of age on average indifference pointdeclines significantly from b 5 .208 to b 5 .177 (z 53.12, p, .01). Similarly, with future orientation addedto the equation, the impact of age ondiscount rate alsodeclines significantly from b5!.212 to b5!.187 (z53.652, p , .001). Thus, age differences in aspects ofpsychological functioning that the future orientationmeasure captures which are unrelated to impulsivitypartially explain why adolescents discount delayedrewards more than adults.

Discussion

By their own self-characterization, and as reflected intheir performance on a standard delay discountingtask, adolescents are less oriented to the future thanare adults. This relatively weaker future orientation isapparent across several different indices, includingindividuals’ self-reported likelihood of planningahead, the extent to which they say that they thinkabout the future, and their reported inclination toanticipate the future consequences of their actionsbefore acting, as well as in their preferences forsmaller rewards that are delivered sooner over largerones delivered at amore distant time point. These agedifferences, which do not vary by gender or ethnicity,

are consistent both with stereotypic portrayals ofyouthful myopia and with prior work on self-reported future orientation using a variety of oper-ationalizations of the construct.

Recent studies documenting continuedmaturationinto the mid-20s of brain regions associated withforesight and planning (Casey, Tottenham, Liston, &Durston, 2005) have prompted new interest in study-ing the development of these psychological phenom-ena. The current study indicates, however, that theconstruct, ‘‘future orientation,’’ is multidimensionalandwarrants amoredifferentiated operationalizationthan that employed in many previous studies. Scien-tists interested in charting the course of this phenom-enon should bear in mind that different aspects offuture orientation may follow different developmen-tal trajectories and reach adult levels of maturity atdifferent ages.

To our knowledge, this is the largest study toexamine age differences in delay discounting andthe first to look simultaneously at delay discountingand self-reported future orientation. Our delay dis-counting findings support and extend those reportedin previous smaller scale studies, which have re-ported age differences in discounting between veryyoung adolescents (age 12) and both young adults(age 20) and the elderly (age 68; Green et al., 1994,1999) but not between adults in their 30s and those intheir 70s (Green et al., 1996); as in these other studies,age differences in discounting are more quantitative(i.e., the steepness of the curve) than qualitative (i.e.,the hyperbolic shape of the curve). According to ourfindings, differences between adolescents and adults,both in their indifference points and in their discountrate, are limited to those between individuals aged 13and younger versus those aged 16 and older, suggest-ing that the period between 13 and 16 may beespecially important for the development of thespecific capacities that underlie discounting behavior,and, as we shall argue, affect individuals’ relativepreference for longer termversus immediate rewards.The development of a relative preference for larger,delayed rewards appears to follow a timetable that ismore similar to the development of time perspectiveor the anticipation of future consequences than to thedevelopment of planning ahead, although furtherresearch using more varied behavioral tasks is indi-cated. This account is consistent with our finding thatrelative preference for a larger but delayed reward iscorrelated with temporal orientation and the antici-pation of future consequences but not with planningahead. It is also consistent with recent work suggest-ing that different aspects of future orientation havedifferent neural bases (Fellows & Farah, 2005).


Our findings also bear on discussions of whetheradults’ more future-oriented performance on delaydiscounting tasks reflects better impulse control,a stronger future orientation, or a combination of thetwo. In particular, our examination of whether futureorientation or impulsivitymediated agedifferences inindifference points and discount rates suggesteda pattern in which future orientation, but not impul-sivity, partially explains differences in delay discount-ing between adolescents and adults. The suggestionthat developmental differences in delay discountingperformance are more closely linked to the develop-ment of future orientation than to impulse controlmust be tempered somewhat by the fact that ourmeasures of impulsivity and future orientation arebased on individuals’ self-reports, and, aswell, by thefact that the correlations between future orientationscores and delay discounting performance are smallin magnitude. It would appear, however, thatalthough individual differences in delay discountingmay be linked both to differences in impulse controland in orientation to the future, age differences indelay discounting during adolescence and adulthoodare due, at least in part, to differences in the ways inwhich adolescents and adults evaluate immediateversus longer term rewards and not to differences inimpulse control. It is also important to note that theamount of variance in delay discounting performanceexplained by the combination of the future orienta-tion and impulsivity measures is small, indicatingthat other factors (e.g., reward sensitivity, attentionalprocesses) also contribute to task performance. Fur-ther research on age differences in delay discountingshould examine other possible mediators.

The differential relation between delay discount-ing and future orientation versus impulsivity is help-ful in drawing a distinction between delaydiscounting and delay of gratification, a paradigmwhich, in contrast to delay discounting, has been usedoften in prior developmental research (e.g., Mischel &Ebbesen, 1970). In the conventional delay of gratifi-cation paradigm, there is but one choice—between animmediate reward and a delayed one; no attempt ismade to systematically vary delay or reward amountswithin the same participant’s testing in order tocalibrate the degree and rate with which a delayedreward is discounted. Delay of gratification tasks areclearly considered to measure impulse control (termssuch as self-control, self-discipline, and self-regulationpervade writings on delay of gratification). Althoughthere has been some discussion in the literature as towhether the source of self-control that permits someindividuals to wait longer than others in order toreceive a larger reward is mainly a function of

individual differences in impulse regulation (andwhere adaptive functioning is neither undercon-trolled nor overcontrolled) as opposed to differencesin the cognitive competencies employed in order toresist a tantalizing reward (and where more compe-tence is clearly better; see Funder & Block, 1989, fora discussion), the overarching theme in these studiesconcerns individuals’ abilities to manage themselves.Our findings indicate that the delay discounting tasklikelymeasures something different from self-manage-ment, however, and suggest that developmentalists,especially those interested in thedevelopment of futureorientation, might profitably incorporate this para-digm into their work. It would also be useful to studydelay of gratification and delay discountingwithin thesame sample, to specifically test the hypothesis that theformer, but not the latter, is linked to impulsivity.

Because the present findings are based on cross-sectional data, one must be cautious about drawingconclusions about patterns of development over time.Nevertheless, the differential pattern of age differencesobserved here with respect to planning ahead, whereage differences are seen well into early adulthood,versus temporal orientation, the anticipation of futureconsequences, anddelaydiscounting,whereagediffer-ences are not seen beyondmiddle adolescence, as wellas the stronger association between delay discountingand future orientation as opposed to impulsivity, isconsistent with an emerging consensus concerninga dual systemsmodel of adolescent brain development(Casey et al., 2008; Steinberg, 2008). According to thismodel, changes in reward seeking (including changesin relative preferences for immediate vs. longer termrewards) aremediated by a ‘‘socioemotional’’ network,which undergoes extensive remodeling early in ado-lescence and which is localized in limbic and paralim-bic areas of the brain, including the amygdala, ventralstriatum (nucleus accumbens), orbitofrontal cortex,medialprefrontal cortex, and superior temporal sulcus.In contrast, changes in impulse control and planningare mediated by a ‘‘cognitive control’’ network, whichis mainly composed of the lateral prefrontal andparietal cortices and those parts of the anterior cingu-late cortex towhich they are interconnected, andwhichmatures more gradually and over a longer period oftime, into early adulthood (Steinberg, 2007). The pres-ent study suggests that adolescents’ relatively greaterpreference for immediate rewards, as indexed by theirperformance on the delay discounting task, as well astheir weaker orientation to the future and their lessersensitivity to the longer term consequences of theiractions may be more strongly related to arousal of thesocioemotional network than to immaturity in cogni-tive control, a finding consistent with the observation

40 Steinberg et al.

that the neural underpinnings of impulsivity differfrom the underpinnings of temporal discounting(Fellows & Farah, 2005). Future research, using brain-imaging techniques, should compare adolescents’ andadults’ patterns of brainactivityduringdelaydiscount-ing task performance in order to better distinguishbetween these two underlying processes.

Questions about whether and to what extent ado-lescents and adults differ in their orientation to thefuture have been central in discussions about the legalregulation of adolescents and critical to debatesabout such diverse phenomena as adolescents’ rightsto autonomous medical decision making (Scott &Woolard, 2004) and the constitutionality of the juve-nile death penalty (Steinberg & Scott, 2003). Yet,although ‘‘future orientation’’ is invoked in discus-sions of different policy questions, a close inspectionof theway inwhich the term is used shows that it mayrefer to very different capacities in different contexts.When future orientation is raised in discussions ofadolescents’ abortion rights, for example, it is largelywith reference to adolescents’ ability to rationallyanticipate the future consequences of their decisions(American Psychological Association, 1989). When itis applied to discussions of the criminal culpability ofjuveniles, though, it often refers to their capacity forpremeditation or planfulness (Roper v. Simmons, 2005).To the extent that these aspects of future orientationare not the same thing, it is important to be carefulabout how arguments that make reference to it areframed. If future orientation is a multidimensionalconstruct composedof a loosely linked set of capacitiesthatdevelop alongdifferent timetables, the applicationof research on its development must be sufficientlynuanced to link specific aspects of future orientation tospecific policy questions (see Steinberg, Cauffman,Woolard, Graham, & Banich, in press).

References

American Psychological Association. (1989). Amicus Curiaebrief in Hodgson v. Minnesota, 497 U.S. 417 (1990).

Benthin, A., Slovic, P., & Severson, H. (1993). A psycho-metric study of adolescent risk perception. Journal ofAdolescence, 16, 153 – 168.

Bickel,W., Odum,A., &Madden, G. (1999). Impulsivity andcigarette smoking: Delay discounting in current, never,and ex-smokers. Psychopharmacology, 146, 447 – 454.

Casey, B. J., Getz, S., & Galvan, A. (2008). The adolescentbrain. Developmental Review, 28, 62 – 77.

Casey, B. J., Tottenham, N., Liston, C., & Durston, S. (2005).Imaging the developing brain: What have we learnedabout cognitive development? Trends in Cognitive Science,9, 104 – 110.

Cauffman, E., & Steinberg, L. (2000). (Im)maturity of judg-ment in adolescence: Why adolescents may be less culpa-ble than adults. Behavioral Sciences and the Law, 18, 741–760.

Cauffman, E., Steinberg, L., & Piquero, A. (2005). Psycho-logical, neuropsychological, and psychophysiologicalcorrelates of serious antisocial behavior in adolescence:The role of self-control. Criminology, 43, 133 – 176.

deWit, H., Flory, J., Acheson, A., McCloskey, M., &Manuck,S. (2007). IQ and nonplanning impulsivity are indepen-dently associated with delay discounting in middle-agedadults. Personality and Individual Differences, 42, 111 – 121.

Fellows, L., & Farah, M. (2005). Dissociable elements ofhuman foresight: A role for the ventromedial frontallobes in framing the future, but not in discounting futurerewards. Neuropsychologia, 43, 1214 – 1221.

Funder, D., & Block, J. (1989). The role of ego-control, ego-resiliency, and IQ in delay of gratification in adolescence.Journal of Personality and Social Psychology, 57, 1041 – 1050.

Furby, M., & Beyth-Marom, R. (1992). Risk-taking inadolescence: A decision making perspective. Develop-mental Review, 12, 1 – 44.

Galvan, A., Hare, T., Voss, H., Glover, G., & Casey, B. J.(2007). Risk-taking and the adolescent brain: Who is atrisk? Developmental Science, 10, F8 – F14.

Gardner, M., & Steinberg, L. (2005). Peer influence on risktaking, risk preference, and risky decision making inadolescence and adulthood: An experimental study.Developmental Psychology, 41, 625 – 635.

Gardner, W. (1993). A life-span rational choice theory ofrisk taking. In N. Bell & R. Bell (Eds.), Adolescent risktaking (pp. 66 – 83). Newbury Park, CA: Sage.

Giedd, J. 1, Blumenthal, J., Jeffries, N., Castellanos, F., Liu,H., Zijdenbos, A., et al. (1999). Brain developmentduring childhood and adolescence: A longitudinal MRIstudy. Nature Neuroscience, 2, 861 – 863.

Green, L., Fry, A., & Myerson, J. (1994). Discounting ofdelayed rewards: A life-span comparison. PsychologicalScience, 5, 33 – 36.

Green, L., Myerson, J., Lichtman, D., Rosen, S., & Fry, A.(1996). Temporal discounting in choice between delayedrewards: The role of age and income. Psychology andAging, 11, 79 – 84.

Green, L., Myerson, J., & Macaux, E. (2005). Temporaldiscounting when the choice is between two delayedrewards. Journal of Experimental Psychology, 31, 1121 – 1133.

Green, L., Myerson, J., & Ostaszewski, P. (1999). Discount-ing of delayed rewards across the life span: Age differ-ences in individual discounting functions. BehaviouralProcesses, 46, 89 – 96.

Greene, A. (1986). Future time perspective in adolescence:The present of things future revisited. Journal of Youthand Adolescence, 15, 99 – 113.

Grisso, T., Steinberg, L., Woolard, J., Cauffman, E., Scott, E.,Graham, S., et al. (2003). Juveniles’ competence to standtrial: A comparison of adolescents’ and adults’ capacitiesas trial defendants. Law and Human Behavior, 27, 333 – 363.

Harter, S. (1982). The Perceived Competence Scale forChildren. Child Development, 53, 87 – 97.


Johnson, M., & Bickel, W. (2002). Within-subject compar-ison of real and hypothetical money rewards in delaydiscounting. Journal of the Experimental Analysis of Behav-ior, 77, 129 – 146.

Lagorio, C., & Madden, G. (2005). Delay discounting ofreal and hypothetical rewards III: Steady-state assess-ments, forced-choice trials, and all real rewards. Behav-ioural Processes, 69, 173 – 187.

Leshem, R., & Glicksohn, J. (2007). The construct ofimpulsivity revisited. Personality and Individual Differ-ences, 43, 681 – 691.

Lessing, E. (1972). Extension of personal future timeperspective, age and life satisfaction of children andadolescents. Developmental Psychology, 6, 457 – 468.

Lewis, C. (1980). A comparison of minors and adultspregnancy decisions. American Journal of Orthopsychiatry,50, 446 – 452.

MacKinnon, D., Lockwood, C., Hoffman, J., West, S., &Sheets, V. (2002). A comparison of methods to testmediation and other intervening variable effects. Psy-chological Methods, 7, 83 – 104.

Madden, G. J., Begotka, A. M., Raiff, B. R., & Kastern, L. L.(2003). Delay discounting of real and hypotheticalrewards. Experimental and Clinical Psychopharmacology,11, 139 – 145.

Madden, G., Petry, N., Badger, G., & Bickel, W. (1997).Impulsive and self-control choices in opioid-dependentpatients and non-drug-using control participants: Drugand monetary rewards. Experimental Clinical Psychophar-macology, 5, 256 – 262.

McCabe, K., & Barnett, D. (2000). The relation betweenfamilial factors and the future orientation of urban,African American sixth graders. Journal of Child andFamily Studies, 9, 491 – 508.

McClure, S., Laibson, D., Loewenstein, G., & Cohen, J.(2004). Separate neural systems value immediate anddelayed monetary rewards. Science, 306, 503 – 507.

Mischel, W., & Ebbesen, E. (1970). Attention in delay ofgratification. Journal of Personality and Social Psychology,16, 329 – 337.

Nurmi, J. (1987). Age, sex, social class, and quality offamily interaction as determinants of adolescents’ futureorientation: A developmental task interpretation. Ado-lescence, 22, 977 – 991.

Nurmi, J. (1989a). Development of orientation to the futureduring early adolescence: A four-year longitudinalstudy and two cross-sectional comparisons. InternationalJournal of Psychology, 24, 195 – 214.

Nurmi, J. (1989b). Planning, motivation, and evaluation inorientation to the future: A latent structure analysis.Scandinavian Journal of Psychology, 30, 64 – 71.

Nurmi, J. (1991). How do adolescents see their future? Areview of the development of future orientation andplanning. Developmental Review, 11, 1 – 59.

Nurmi, J. (1992). Age differences in adult life goals,concerns, and their temporal extension: A life courseapproach to future-oriented motivation. InternationalJournal of Behavioral Development, 15, 487 – 508.

Nurmi, J., Poole, M., & Kalakoski, V. (1994). Age differ-ences in adolescent future-oriented goals, concerns, andrelated temporal extension in different socioculturalcontexts. Journal of Youth and Adolescence, 23, 471 – 487.

Nuttin, J. (1985). Future time perspective and motivation. Hill-sdale, NJ: Erlbaum.

Ohmura, Y., Takahashi, T., Kitamura, N., & Wehr, P. (2006).Three-month stability of delay and probability discount-ing measures. Experimental and Clinical Psychopharmacol-ogy, 14, 318 – 328.

Patton, J., Stanford, M., & Barratt, E. (1995). Factor struc-ture of the Barratt Impulsiveness Scale. Journal of ClinicalPsychology, 51, 768 – 774.

Petry, N. (2002). Discounting of delayed rewards in sub-stance abusers: Relationship to antisocial personalitydisorder. Psychopharmacology, 162, 1432 – 2072.

Poole, M., & Cooney, G. (1987). Orientations to the future:A comparison of adolescents in Australia and Singa-pore. Journal of Youth and Adolescence, 16, 129 – 151.

Roper v. Simmons, 541 U.S. 1040 (2005).Scheres, A., Dijkstra, M., Ainslie, E., Balkan, J., Reynolds, B.,Sonuga-Barke, E., et al. (2006). Temporal and probabilisticdiscounting of rewards in children and adolescents:Effects of age and ADHD symptoms. Neuropsychologia,44, 2092 – 2103.

Scott, E., & Woolard, J. (2004). The legal regulation ofadolescence. In R. Lerner & L. Steinberg (Eds.), Handbookof adolescent psychology (2nd ed., pp. 523 – 550). NewYork: Wiley.

Somers, C., & Gizzi, T. (2001). Predicting adolescents’ riskybehaviors: The influence of future orientation, schoolinvolvement, and school attachment. Adolescent andFamily Health, 2, 3 – 11.

Steinberg, L. (2007). Risk-taking in adolescence: Newperspectives from brain and behavioral science. CurrentDirections in Psychological Science, 16, 55 – 59.

Steinberg, L. (2008). A social neuroscience perspective onadolescent risk-taking. Developmental Review, 28, 78 – 106.

Steinberg,L.,Cauffman,E.,Woolard, J.,Graham,S.,&Banich,M. (in press) Are adolescents less mature than adults?Minors’ access to abortion, the juvenile death penalty, andthe alleged APA ‘‘flip-flop’’. American Psychologist.

Steinberg, L., & Scott, E. (2003). Less guilty by reason ofadolescence: Developmental immaturity, diminishedresponsibility, and the juvenile death penalty. AmericanPsychologist, 58, 1009 – 1018.

Trommsdorff, G., & Lamm, H. (1980). Future orientation ofinstitutionalized and noninstitutionalized delinquentsand nondelinquents. European Journal of Social Psychol-ogy, 10, 247 – 278.

Trommsdorff, G., Lamm, H., & Schmidt, R. (1979). A longi-tudinal study of adolescents’ future orientation (timeperspective). Journal of Youth and Adolescence, 8, 131 – 147.

Vuchinich, R., & Simpson, C. (1998). Hyperbolic temporaldiscounting in social drinkers and problem drinkers.Experimental and Clinical Psychopharmacology, 6, 292 – 305.

Wechsler, D. (1999). Wechsler Abbreviated Scale of Intelli-gence. San Antonio, TX: Psychological Corporation.

42 Steinberg et al.

Appendix: Future Orientation Scale


Scoring: All items are scored left to right on a scale of 1-4. Reverse score items 1, 3, 4, 6, 8, 11, and 14, so that higher scores indicate a strongerfuture orientation. Future Orientation total score is the unweighted average of all 15 items. Planning Ahead is the unweighted average of items1, 6, 7, 12, and 13. Time Perspective is the unweighted average of items 2, 5, 8, 11, 14. Anticipation of Future Consequences is the unweightedaverage of items 3, 4, 9, 10, 15.

44 Steinberg et al.

age differences in future orientation and delay...

Documents