Delay of Gratification:

Impulsive Choices and Problem Behaviors

in Early and Late Adolescence

Edelgard Wulfert
Jennifer A. Block
Elizabeth Santa Ana
Monica L. Rodriguez
Melissa Colsman
The University at Albany, State University of New York

ABSTRACT High school students were offered a monetary incentive for

participating in research. They were given a choice between a smaller fee
immediately or a larger fee one week later. Compared to students who
delayed gratification, those who chose the immediate fee showed more self-
regulatory deficits. They showed greater involvement with cigarettes,
alcohol, and marijuana, had a poorer self-concept and underperformed
academically. A replication study with middle-school students and different
reward parameters yielded equivalent results. Younger adolescents who
chose the immediate monetary incentive showed a similar pattern of
problem behaviors as the high school students. The findings indicate that
this simple choice-delay procedure yields an unobtrusive behavioral
measure of self-regulation and offers a developmentally appropriate

Edelgard Wulfert, Jennifer A. Block, Elizabeth Santa Ana, Monica L. Rodriguez and
Melissa Colsman, University at Albany, State University of New York.
Monica L. Rodriguez was supported by a William T. Grant Faculty Scholars Award.
Correspondence concerning this article should be addressed to Edelgard Wulfert,
Department of Psychology, The University at Albany, State University of New York.,
SS112, Albany, New York 12222. Electronic mail may be sent to e.wulfert@
albany.edu.
Journal of Personality 70:4, August 2002.
Copyright # 2002 by Blackwell Publishing, 350 Main Street, Malden, MA 02148,
USA, and 208 Cowley Road, Oxford, OX4 1JF, UK.
534 Wulfert et al.

extension of the delay-of-gratification paradigm for use with older children

and adolescents.

The ability to self-regulate is a developmental process with broad

implications not only for the individual but for society at large. Failure
to exert self-control has been linked to numerous problem behaviors
including substance abuse (Madden, Petry, Badger, & Bickel, 1997),
risky sex (Wulfert, Safren, Brown, & Wan, 1999), aggression
(Baumeister, Heatherton, & Tice, 1994) and violence (Cherek, Moeller,
Dougherty, & Rhoades, 1997), suggesting that self-regulatory diffi-
culties lie at the root of many problems afflicting contemporary society.
Individual differences in self-control are readily apparent early in
life and tend to be fairly stable across the life span. Children with self-
regulatory deficits show more psychopathology (Krueger, Caspi,
Moffit, White, & Stouthamer-Loeber, 1996; White et al., 1994), are
more frustrated and aggressive (Mischel, Shoda & Rodriguez, 1989;
Rodriguez, Shoda, Mischel, & Wright, 1999), pay less attention in
school and have lower grades than children who exercise self-control
(Funder & Block, 1989; Funder, Block, & Block, 1983). As
adolescents, such children have lower academic, cognitive, and social
competencies (Mischel, Shoda, & Peake, 1988; Shoda, Mischel, &
Peake, 1990), deficits that persist into adulthood (Ayduk, et al., 2000).

Delay of Gratification as a Measure

of Self-Regulation
In young children, self-regulation has typically been assessed with
Mischel’s (1974; Mischel, Ebbesen, & Zeiss, 1972) delay-of-
gratification procedure. It entails leaving young children alone with
two attractive reward objects (e.g., one versus two marshmallows)
after instructing them that they will receive the larger reward if they
wait for the experimenter to return, or the smaller reward, if they ring a
bell to call the experimenter back. Self-control is measured in time the
child delays ringing the bell.
Few studies on adolescent delay of gratification exist (Funder &
Block, 1989; White et al., 1994; Krueger et al., 1994), even though it is
at this very time when real-life difficulties arise from young people’s
inability to control their impulses. One reason may be that Mischel’s
delay procedure is difficult to extrapolate in a meaningful way
to research with adolescents. First, it is hard to find feasible and
Delay of Gratification 535

non-trivial incentives for this age group. Second, meaningful delay

intervals for adolescents could span days or weeks rather than minutes
used for young children’s delay tasks. Given these limitations,
noncontroversial measures of adolescents and adults’ self-regulatory
ability are rarely reported in the literature. Many researchers have
relied on questionnaire assessments, which raises the specter of social
desirability and other self-presentational biases. Others have used
analogue measures such as choices between hypothetical money
rewards (Madden et al., 1997).
A few studies have used monetary incentives. Funder & Block
(1989), for example, employed a complex procedure involving six
assessment sessions spread over several weeks. They offered
adolescents repeated choices between immediate payments of $4 after
each assessment or a $28 lump sum payment including ‘‘interest’’
paid at the end of the study. While this procedure yields a good
behavioral measure of impulsivity, it is expensive and the requirement
of repeated assessments makes it unsuitable for studies involving only
a single contact between experimenter and research participant.
White et al. (1994) also used monetary reinforcers. They presented
adolescents with a computerized task involving choices between
pressing a button immediately for a 40% chance of winning a nickel
versus delaying button presses for an 80% chance of winning a nickel.
Alas, chances of winning nickels may no longer be a sufficiently
attractive incentive in our consumer-oriented society to guarantee a
legitimate assessment of adolescents’ ability to delay gratification.

Adolescents’ Choice Behavior and Its Correlates

To produce a non-trivial, but also uncomplicated and feasible experi-
mental assessment of adolescents’ ability to delay gratification, two
things are necessary. First, the adolescents must be offered choices that
are sufficiently attractive to engage their motivation. Money as a
generalized reinforcer fulfills this requirement.
Second, if a choice between two attractive incentives is offered, care
must be taken to start from a preference equilibrium because there is
ample evidence from human and animal studies that amount and delay
of reinforcement are not independent (Chapman, 1996; Green &
Snyderman, 1980; Kirby, 1997; Logue, Forzano, & Tobin, 1992;
Rachlin & Green, 1972; Stevenson, 1993). To illustrate, people
typically state a preference for $115 available in 53 weeks over $100
536 Wulfert et al.

available in 52 weeks but would rather receive $100 today than $115
in one week (Herrnstein, 1990). Therefore, when money choices are
used to assess adolescents’ self-regulatory abilities, pretesting
different amounts and delays can ensure that no strong preference
exists for the immediate or the delayed reward. Any observed
preferences following an experimental manipulation can then be
attributed to individual differences in the ability to delay gratification.
To test the possibility of using money choices as an indicator of
adolescents’ self-regulation, we invited middle and high school students
to participate in a confidential survey on academic achievement and
substance use. As it is customary to compensate research participants
for their time and effort, we used this ‘‘subject fee’’ to produce an
unobtrusive and uncontrived measure of delay of gratification by
offering the adolescents an immediate and smaller or a delayed and
larger fee. As self-regulatory deficits often manifest themselves in a
cluster of problem behavior (Donovan, Jessor, & Costa, 1988; Jessor,
1987), we predicted, based on previous research, that adolescents who
chose the immediate, smaller reward would show greater involvement
with substances (Brown, Myers, Mott, & Vik, 1994; Clapper, Buka,
Goldfield, Lipsitt, & Tsuang, 1995; Gallucci, 1997; Henderson, Galen,
& DeLuca, 1998), have low self-esteem, and underperform academi-
cally (Dawes, Tarter, & Kirisci, 1997; Donovan et al., 1988).

STUDY 1
METHOD
Research Participants
The research participants (N = 69) were 43 male and 26 female students
from an urban high school (grades 9 through 12). Their mean age was 15.9
years (range 14 to 18); 94% were Caucasian and most students, socio-
economically, came from a middle-class background (the school had 8.1% of
free lunch students). Based on conduct records from school, we designated
36 (23 males and 13 females) as problem students; the remaining 33 were
non-problem students.

Measures
Academic achievement was obtained from school records. English and
Mathematics grades were averaged. This grade point average, GPA, had a
possible range from 0–100.
Delay of Gratification 537

Self-esteem was assessed with the Rosenberg Self-Esteem Scale (RSE;

Rosenberg, 1965). It consists of 10 items (e.g., I feel that I have a number of
good qualities) that are endorsed on 4-point Likert-type scales (from 1 =
strongly agree to 4 = strongly disagree; possible range 10–40). The RSE has
adequate psychometric properties (Wylie, 1989) and is the most widely used
measure to assess global self-esteem.

Problem behaviors were assessed in two areas. First, school guidance

counselors identified ‘‘problem’’ and ‘‘non-problem’’ students from school
records, depending on whether students had received disciplinary sanctions
during the present school year. Second, students completed a confidential
questionnaire about any substances they had used during the past month.
Questions were adapted from the Health Behavior Questionnaire (Jessor,
Donovan, & Costa, 1992). On 7-point Likert-type scales (range 0–6) students
reported their experience with various substances, including how many
cigarettes they smoked on an average day (from 1 = one cigarette a day or
less to 6 = two packs or more); how often they drank alcohol (from 1 = once
to 6 = almost every day); how much they typically drank each time they
consumed alcohol (from 1 = one drink or less to 6 = six drinks or more); and
how often they smoked marijuana (from 1 = once to 6 = six or more times).
For each question, the option ‘‘never or not applicable,’’ scored zero, was
also available.
Delay of gratification was measured by presenting students with a choice
between an immediately available payment of $7 or a one-week delayed
payment of $10 for participating in the study. The dollar amounts were
selected based on pretesting to ensure that neither value was intrinsically
preferred. For the pretest, 45 college freshmen were given a series of
hypothetical choices, such as receiving $5 now versus $10 one week later. No
particular preference was apparent when the immediate incentive reached
70% of the value of the delayed incentive, holding the delay interval constant
at one week.

Procedure
The study was approved by the review boards of the university and the
participating high school. Guidance counselors identified problem students
who during the past year had been written up for at least one disciplinary
violation (e.g., truancy, verbal aggression, fighting) and non-problem students
who matched the former in gender and grade level. The school forwarded a
letter from the researchers to the parents, asking them to call the school if they
did not wish the researchers to contact them. No parent called. The researchers
then contacted parents by phone to explain the project and request written
permission for the adolescents to participate. With parental approval, the
538 Wulfert et al.

adolescents were contacted to obtain their assent. Five parents (two of

problem students and three of non-problem students) did not allow their
children to be part of the study. For 74 students contacted, written parental
consent and adolescent assent was obtained for a total of 69 students.
A guidance counselor assisting the researchers reviewed school records,
verified the students’ problem status, and obtained their English and Math
grades. Parents were asked to complete a brief biographical information sheet
and mail it to the researchers. Adolescents completed a confidential attitude
and behavior questionnaire and were led to expect a $7 honorarium for
participating in the study. After data collection was completed, the
experimenter thanked each student individually, placed $7 plus an additional
$3 and an envelope with the student’s name in front of him or her, and said:
Now, I promised that I would pay you $7, but you actually have a choice.
You can either have $7 right now or, if you want to wait one week, you can
have $10. If you choose the $10, we will put the money in this envelope
with your name on it, and you can pick it up in exactly a week from today
in Mrs. X’s (school secretary) office.
Mrs. X was chosen as the contact person because she was friendly,
exceptionally well-liked by students, and generally perceived as helpful and
trustworthy. Depositing the money with her assured the students that the money
would in fact be available one week later, should they choose the delayed, larger
amount. Students who were unwilling to state a preference were paid $7.

RESULTS

Of the 69 participants, two male problem students did not state a

preference for immediate or delayed payment. They received $7, and
their data were excluded from the analyses. The data of one female
student were sufficiently incomplete that they were also eliminated.
The final sample consisted of 41 male and 25 female students; of
those, 34 (21 males, 13 females) were problem students and 32 (20
males, 12 females) were non-problem students.

Initial Analyses
As shown in Table 1 (upper part), exactly one half of the students (21
males and 12 females) chose the $7 whereas the other half (20 males
and 13 females) decided to wait one week for the $10. Of those who
chose the immediate reward, 27 pertained to the problem and 6 to the
non-problem group. Of those who delayed gratification, 7 were
problem and 26 were non-problem students.
 Table 1
Comparison of High School Students Choosing the Immediate Versus Delayed Reward (Study 1)

Immediate reward Delayed reward F statistic p value r value

N (%) N (%)
Gender
Males 21 (63.6) 20 (60.6)
Females 12 (36.4) 13 (39.4)
Problem Status
Problem students 27 (81.8) 7 (21.2)
Non-problem students 6 (18.2) 26 (78.8)

Mean (SD) Mean (SD)

Gradesa 64.00 (18.4) 78.82 (16.4) F 11.96 .001 ( .40)
Conduct problems in schoolb 0.82 (0.39) 0.21 (0.42) F 37.21 <.001 (.61)
Substance use (composite)c 8.24 (6.26) 3.91 (4.91) F 9.80 .003 (.36)
Cigarettes (frequency of use)c 1.70 (1.59) 0.64 (1.25) F 9.10 .004 (.35)
Alcohol (frequency of use)c 1.42 (1.20) 0.97 (1.02) F 2.76 .101 (.20)
Alcohol (typical amount)c 3.55 (2.88) 1.64 (2.13) F 9.35 .003 (.34)
Marijuana (frequency of use)c 1.85 (2.27) .70 (1.55) F 5.81 .019 (.29)
Self-esteemd 29.70 (5.50) 33.39 (4.66) F 8.68 .004 ( .33)

SD = Standard deviation; possible range of variables: a0–100; b0–1; c0–6; d10–40.

540 Wulfert et al.

A logistic regression analysis was performed to examine whether

gender or problem status were associated with delay of gratification.
Results showed that adolescents who made impulsive choices were much
more likely to come from the problem student group than those who
delayed (odds ratio [OR] = 16.83, 95% confidence interval [CI] = 4.77–
59.37); there was neither a significant gender difference (OR = 0.79, 95%
CI = 0.23–2.79)nor aninteractionof genderwithproblemstatus regarding
the ability to delay gratification (OR 1.01, 95% CI = 0.29–3.56).
As parental income was known for 30 problem and 25 non-problem
students (i.e., 55 of the 66 participants), we examined with logistic
regression whether possible differences in parental income might have
made money a more powerful reinforcer for adolescents in the
problem group. The analysis showed that neither parental income (OR
= 0.85, 95% CI = 0.60–1.21) nor the interaction between problem
status and parental income (OR = 1.05, 95% CI = 0.74–1.49) was
significantly associated with failure to delay gratification.

Classification of Students by Type of Choice

Next, we tested the hypothesis that impulsive choices are indicative of a
pattern of problem behavior, which Jessor (1987) described as deviance
syndrome. First, the intercorrelations among choices and hypothesized
correlates (discipline problems in school, academic achievement,
substance use, and self-esteem) were examined (for exploratory
purposes, gender was also included in the matrix). As expected, all
variables but gender were related, but in each of these relationships a
fair portion of the variance remained unexplained (Table 2).

Table 2
Correlation Matrix of Variables Assessed in High School Students

1 2 3 4 5 6
1. Impulsiveness 1.00
2. Problem status 0.61 1.00
3. GPA 0.40 0.71 1.00
4. Substance use 0.36 0.67 0.59 1.00
5. Self-esteem 0.33 0.46 0.38 0.49 1.00
6. Gender 0.03 0.01 0.03 0.06 0.13 1.00
Delay of Gratification 541

A discriminant analysis was performed to classify students who

delayed versus those who chose the immediate reward. F-tests
(Table 1, lower part) showed that the four predictors achieved
univariate significance: GPA, F (1, 64) 11.96, p = .001 (r = .40),
misconduct at school, F (1, 64) 37.21, p < .001 (r = .61), self-
esteem, F (1, 64) 8.68, p = .004 (r = .33), and substance use, F
(1, 64) 9.80, p = .003 (r = .36). (Due to sample size constraints,
frequency of cigarette, alcohol, and marijuana use, and amount of
alcohol typically consumed were converted to z-scores and averaged
to yield a composite substance use index. Non-transformed values
for individual variables are presented in Table 2). The discriminant
function yielded a Wilks’ lambda of .620, c2 (4) = 29.65, p < .001.
The four just-described variables together correctly classified 80.3%
of the adolescents (78.8% of students who had chosen the delayed
reward and 81.8% of those who opted for the immediate reward).
This classification rate significantly exceeded the 50% hit rate
expected by chance, z = 4.92, p < .001 (Huberty, 1994).
As the presence or absence of disciplinary violations at school had
been used to preselect research participants, and this categorization
significantly overlapped with their ability to delay, we calculated a
second discriminant function excluding this variable. The function
yielded a Wilks’ lambda of .795 c2 (3) = 14.32, p = .002. GPA,
substance use, and self-esteem together achieved a correct classifica-
tion of 77.3% of the high school students, including78.8% of those
who delayed and 75.8% of those who did not delay gratification.
This classification rate significantly exceeded chance level, z = 3.94,
p < .001.

DISCUSSION

By extending the delay-of-gratification paradigm originally developed

for young children, we produced a behavioral measure suitable for
assessing high school students’ delay behavior. Mischel and
colleagues used primary reinforcers (candy) and delay times suited
for small children (minutes). We implemented parameters more
appropriate for adolescents, including a generalized reinforcer
(money) and a one-week delay. This simple behavioral measure of
willingness to delay gratification in a research setting was significantly
associated with self-regulation in other life areas.
542 Wulfert et al.

As a group, high school students who opted for the immediate,

albeit smaller, reward differed in several attributes from those who
waited for the delayed, larger reward. Their grades were lower,
suggesting that they did not concentrate on, or devote enough time and
attention to, school-related matters. Their self-esteem was lower,
perhaps in part as a reaction to their lackluster performance in the
classroom. They were more likely to have transgressed against school
norms, and they admitted to a greater involvement with cigarettes,
alcohol, and marijuana. Follow-up analyses showed that the increased
alcohol consumption was not so much apparent in how often they
imbibed but rather in how much they drank on a typical occasion. This
is not surprising because many adolescents experiment with alcohol
during their high school years. Experimentation in itself is not
necessarily a sign of maladjustment (Shedler & Block, 1990).
However, drinking large amounts of alcohol can become problematic
as early-onset heavy drinking is a strong predictor of later alcohol
abuse (Caetano, 1986; Clark & Midanik, 1981).
In summary, the adolescents’ money choices proved to be a good
behavioral measure of their ability to delay gratification. Unwilling-
ness to postpone gratification in a research setting was simply the
manifestation of a more general pattern of self-regulatory deficits in
other life areas.

STUDY 2

The purpose of Study 2 was to examine the generality of the findings

from Study 1. The procedures employed in the previous study were
replicated with younger adolescents who attended middle school. To
demonstrate that the effects observed in Study 1 were not particular to
the specific incentive values manipulated, we used smaller monetary
rewards in the replication. The delay interval was held constant across
the two studies, so as not to confound amount of reward and delay.

METHOD
Research Participants
The research participants (N = 48) were 22 boys and 26 girls from an urban
middle school (grades 6 through 8). Exactly one half (11 boys and 13 girls)
were designated problem students because of disciplinary violations at
Delay of Gratification 543

school; the other half were non-problem students. The students’ mean age
was 12.6 years (range 11 to 14); 93.8% were Caucasian, and, socio-
economically, most came from a middle-class background (the school had
9.3% free lunch students).

Procedure
All selection, consent, and testing procedures were identical to those
described in Study 1. Six parents did not grant permission for their children to
participate (four of the children were problem students, and two were non-
problem students). Of 54 potential research participants, parental consent and
adolescent assent was obtained for a total of 48.
In the delay-of-gratification manipulation, one difference to the previous
study was the amount of money offered as an incentive for participating in
the project. Each adolescent was presented individually with a choice of
receiving five dollars immediately or seven dollars one week later.
Analogous to Study 1, the value of the immediate reward was approximately
70% of the value of the delayed reward.

Measures
The same achievement, self-esteem, and behavior measures as described in
Study 1 were administered, with the following exception. Given the younger
age of the participants, it seemed implausible that they would be heavily
involved in substance use. Cigarette smoking, alcohol, and drug use were
assessed with the deviant behavior scale of Jessor et al.’s (1992) Health
Behavior Questionnaire, but a time frame of six months was used. The
adolescents were asked how often they had used cigarettes, alcohol, and
marijuana and how often they had consumed more than five drinks at one
time (which is a widely accepted measure of binge drinking). Their answers
were recorded on 5-point Likert-type scales (0 = never, 1 = once, 2 = twice,
3 = three to four times, 4 = five or more times).

RESULTS

Initial Analyses
As shown in Table 3 (upper part), 13 boys and 11 girls chose the
immediate reward of $5 whereas the 9 boys and 15 girls decided to
wait one week for the delayed $7 reward. When considering problem
status, 20 problem and 4 non-problem students chose the immediate
reward, whereas 20 non-problem and only 4 problem students decided
to wait one week for the larger reward.
 Table 3
Comparison of Middle School Students Choosing the Immediate Versus Delayed Reward (Study 2)

Immediate reward Delayed reward F statistic p r

N (%) N (%)
Gender
Males 13 (54.2) 9 (37.5)
Females 11 (45.8) 15 (62.5)
Problem Status
Problem students 20 (83.3) 4 (16.7)
Non-problem students 4 (16.7) 20 (83.3)

Mean (SD) Mean (SD) (one-tailed)

Gradesa 70.03 (12.80) 86.60 (9.44) F 26.07 <.000 7( .60)
Conduct problems in schoolb 0.83 (0.38) 0.17 (0.38) F 36.80 <.000 (.68)
Substance use (composite)c 2.72 (1.64) 1.28 (0.48) F 16.97 <.000 (.52)
Cigarettes (frequency of use)c 1.88 (1.87) .42 (1.14) F 10.64 .002 (.43)
Alcohol (frequency of use)c 1.79 (1.81) .75 (1.07) F 5.85 .020 (.34)
Alcohol (binge)c 1.54 (1.84) 1.08 (0.28) F 14.71 <.000 (.49)
Marijuana (frequency of use)c 2.63 (1.93) .00 (0.00) F 17.02 <.000 (.52)
Self-esteemd 26.75 (5.72) 30.76 (4.41) F 6.90 .012 ( .32)

SD = Standard deviation; possible range of variables: a0–100; b0–1; c0–5; d10–40.

Delay of Gratification 545

As in Study 1, a logistic regression analysis was performed to

examine whether gender or problem status were associated with delay
of gratification. Adolescents who did not delay were significantly
more likely to come from the problem student group than those who
delayed (OR = 32.66, 95% CI = 5.86–182.09); there was neither a
gender difference (OR = 0.27, 95% CI = 0.05–1.52) nor an interaction
of gender with problem status (OR = 0.82, 95% CI = 0.15–4.55).

Classification of Students by Type of Choice

As in the previous study, we tested the hypothesis that unwillingness
to delay gratification is associated with a cluster of problem behaviors.
An inspection of the intercorrelations among variables (for exploratory
purposes, gender was also included in the matrix) revealed a good deal
of overlap, except for gender. However, as in Study 1, a fair portion of
the variance remained unexplained in each relationship (Table 4).
In the discriminant analysis, all variables achieved univariate
significance, including problem status, F (1, 46) 36.80, p < .001
(r = .67); GPA, F (1, 46) 26.07, p < .001 (r = .60); self-esteem,
F (1, 46) 5.10, p = .029 (r = .31); and substance use, F (1, 46) 15.73,
p < .001 (r = .52). (A composite substance-use variable was computed
by averaging the scores for smoking, alcohol use, binge drinking,
and marijuana use; individual values for the four variables are
presented in Table 3, lower part). The discriminant function yielded
a Wilks’ lambda of .516, which was statistically significant, c2 (4) =
29.09, p < .001. All variables together achieved a correct
classification of 83.3% of the adolescents, regardless of whether
they chose the immediate reward or delayed gratification. This

Table 4
Correlation Matrix of Variables Assessed in Middle School Students

1 2 3 4 5 6
1. Impulsiveness 1.00
2. Problem status 0.68 1.00
3. GPA 0.60 0.84 1.00
4. Substance use 0.52 0.55 0.51 1.00
5. Self-esteem 0.32 0.40 0.26 0.22 1.00
6. Gender 0.17 0.00 0.06 0.03 0.04 1.00
546 Wulfert et al.

classification rate significantly exceeded the hit rate expected by

chance, z = 4.91, p < .001 (Huberty, 1994).
As in Study 1, we also conducted a second discriminant analysis
that excluded problem status, given that receiving disciplinary
sanctions for transgressions at school had been a selection criterion
for the study participants. The discriminant function yielded a Wilks’
lambda of.556, c2 (3) = 26.18, p < .001. It correctly classified 87.5%
of the adolescents who delayed gratification and 79.2% of those who
chose the immediate reward. The overall correct classification rate of
83.3% significantly exceeded a hit rate expected by chance, z = 5.20,
p < .001.

DISCUSSION

The findings with high school students (Study 1) proved robust and
were successfully replicated with younger adolescents and different
reward parameters. In both studies, an unwillingness to delay
gratification was related to self-regulatory deficits in other life areas.
Compared to students who delayed gratification, those who chose the
immediate reward were more likely to report lower self-esteem,
transgress against school norms, earn lower grades, and lack self-
restraint in their use of cigarettes, alcohol, and other drugs. One
difference between Studies 1 and 2 was that the middle school students
who chose the immediate reward drank alcohol both more frequently
and in larger quantities than those who delayed, whereas among high
school students the frequency of drinking did not differentiate between
groups. This finding probably reflects the fact that well-adjusted middle
school students do not yet use alcohol, whereas high school students,
regardless of their adjustment, begin to experiment with alcohol, but
only those who are maladjusted use alcohol in large quantities.
One additional point deserves mention. In the middle school sample,
proportionately more girls than boys delayed gratification, although
the gender difference was not statistically significant. In contrast, in
the high school sample there was no indication of a gender difference.
Because of the small sample size, we cannot be sure whether boys and
girls differ in their self-regulatory ability during late childhood. If they
do, the difference apparently vanishes with increasing age. This may
reflect differences in socialization. Perhaps parents inhibit girls’
impulsive tendencies more than boys’, but girls ‘‘catch up’’ as they
become increasingly independent during adolescence. The possibility
Delay of Gratification 547

of different developmental trajectories for boys and girls’ self-

regulatory abilities should be examined in future research.

GENERAL DISCUSSION

The delay of gratification paradigm (Mischel, 1974) is useful for

assessing individual differences in young children’s self-regulatory
abilities but due to methodological limitations it is not suitable for
studying older age groups. Our subject-fee-based choice procedure
bridges the existing gap and offers a simple preparation that is well-
suited for use with adolescents. The procedure yields a behavioral
marker of adolescents’ self-regulatory ability in that impulsive choices
in an experimental situation are indicative of a lack of self-control in
other life areas (e.g., substance use, low achievement). As these
problem behaviors are often the long-term sequelae of self-regulatory
deficits first apparent in childhood (Ayduk et al., 2000; Mischel et al.,
1989), it is reasonable to assume that our delay procedure taps into the
same self-regulatory process as Mischel’s delay-of-gratification
paradigm. Thus, this procedure offers a simple yet developmentally
appropriate extension of Mischel’s paradigm and opens the door to
studying the developmental trajectory of self-regulation across a much
broader age range.
Our findings also build on Block and Funder’s (1989) work. These
researchers demonstrated that adolescents who did not contain their
impulse for immediate gratification were rated as less intelligent and
more rebellious, self-indulgent, and hostile than those who delayed.
Our study found that adolescents who made impulsive money choices
had low self-esteem, underperformed in school, and manifested a set
of undesirable behaviors including cigarette smoking, drinking,
marijuana use, and fighting. Together, these studies indicate that
adolescents who are unwilling to delay gratification show a broad
range of personality attributes, attitudes and behavior that might be
described as signs of general undercontrol.
A difference between our procedure and the traditional delay
paradigm (as used by Mischel and Funder and Block) is that children
in the latter can change their mind any time during the delay interval,
whereas, in our procedure a decision cannot be reversed. The
traditional paradigm is useful for elucidating influences on young
children’s ability to delay gratification by manipulating variables
such as the salience of the reward objects or the use of cognitive
548 Wulfert et al.

coping strategies. However, by age six children become unresponsive

to such manipulations because by that time they have acquired basic
strategies to facilitate delay (Rodriguez, Mischel, & Shoda, 1989;
Shoda et al., 1990). A simple discounting procedure such as ours is
therefore well-suited for adolescents and adults. It mimics many real-
life choices that involve irreversible decisions with consequences
occurring at different times, such as how much time a student spends
today studying for an exam versus what grade she will earn at the end
of the semester.
The specific incentive values and the delay time used in the present
studies were based on pretesting with 18-year-old college freshmen.
To further our understanding of adolescent self-regulatory processes,
we recommend that additional delay and incentive value parameters be
examined in future research. For example, it would be important to
identify at what point a monetary incentive might be considered too
trivial to support behavior and what minimum or maximum delay
intervals are necessary to separate impulsive from non-impulsive
choices. A rigorous examination of such parameters could provide
information about the possibility of specific discounting biases that
might underlie impulsive choices.

Limitations and Implications

There is some need for caution when interpreting the findings from the
present investigation. Students were drawn from middle-class,
ethnically homogeneous schools. Participants were not randomly
selected and only students were included whose parents gave
permission. In both studies, sample sizes were modest, limited by the
availability of students with behavior problems in the schools from
which we sampled. These factors limit the generalizability of the
results.
A criticism of our delay procedure might be that we have not
measured impulsivity independently from the delay task. Hence, we
cannot show whether delay is selectively correlated with impulsivity or
globally with many aspects of psychopathology. To explore this
possibility, future studies might include additional measures such as the
Stroop Color and Word Association Test (Stroop, 1935) or Eysenck’s
Impulsiveness Scale (S.B.G. Eysenck, Eastling, & Pearson, 1984;
S.B.G. Eysenck & Eysenck, 1977) and examine their relationship to our
delay procedure.
Delay of Gratification 549

Another criticism of our studies might be that we have not

conclusively demonstrated that the adolescents made impulsive
choices. An alternative interpretation might be that those opting for
the immediate payment did not trust the experimenters and wanted to
secure the money owed them. Although we did not independently
assess the adolescents’ level of trust, we sought to minimize trust
issues by choosing individuals as guardians of the money with whom
the children were familiar and whom they perceived as trustworthy.
Further, since ‘‘impulsive’’ choices were predictive of a cluster of
problem behaviors commonly viewed as markers of deficits in self-
regulation, the bulk of the evidence favors our interpretation.
Despite these limitations, three findings strengthen our confidence
that our incentive manipulation was an accurate indication of
adolescents’ self-regulatory tendencies. First, the pattern of results
obtained with high school students was replicated with younger
adolescents at the middle school level, suggesting that the findings
were robust. Second, the results were consistent with theoretical
assumptions that self-regulatory failures express themselves in a
pattern of problem behavior. Third, our studies fit in with a number of
empirical investigations that have linked self-control deficits to a
range of problem behaviors, including substance abuse, low self-
esteem, and academic failure. Taken together, these facts make it
implausible that extraneous factors accounted for the observed
relationship between the adolescents’ reward choices and their self-
regulatory behavior in other life areas.
In conclusion, the incentive manipulation used in the present studies
with students ranging in age from 12 to 18 provides a simple and easy
way to administer assessment of self-regulatory ability suitable for this
older age group. As a behavioral measure, it presents a significant
improvement over self-report or analogue measures, it is simpler and
more economical than procedures requiring repeated assessments, and
it provides a ready tool for identifying adolescents who are most in
need of intervention.

REFERENCES

Ayduk, O., Mendoza-Denton, R., Mischel, W., Downey, G., Peake, P. K., &
Rodriguez, M. (2000). Regulating the interpersonal self: Strategic self-regulation
for coping with rejection sensitivity. Journal of Personality and Social Psychology,
79, 776–792.
550 Wulfert et al.

Baumeister, R. F., Heatherton, T. F., & Tice, D. M. (1994). Losing control: How and
why people fail at self-regulation. San Diego, CA: Academic Press.
Brown, S. A., Myers, M. G., Mott, M. A., & Vik, P. W. (1994). Correlates of success
following treatment for adolescent substance abuse. Applied and Preventive
Psychology, 3, 61–73.
Caetano, R. (1986). Drinking and Hispanic-American family life. Alcohol, Health and
Research World (pp. 26–35).
Chapman, G. B. (1996). Temporal discounting and utility for health and money. Journal
of Experimental Psychology: Learning, Memory, and Cognition, 22, 771–791.
Cherek, D. R., Moeller, F. G., Dougherty, D. M., & Rhoades, H. (1997). Studies of
violent and nonviolent male parolees: II. Laboratory and psychometric measures of
impulsivity. Biological Psychiatry, 41, 523–529.
Clapper, R. L., Buka, S. L., Goldfield, E. C., & Lipsitt, L. P., & Tsuang, M. T. (1995).
Adolescent problem behaviors as predictors of adult alcohol diagnoses. Interna-
tional Journal of the Addictions, 30, 507–523.
Clark, W., & Midanik, I. (1981). Alcohol use and alcohol problems among U.S.
adults: Results of the 1979 National survey. In: Alcohol Consumption and Related
Problems (pp. 4–52). Washington, D.C.: U.S. Government Printing Office.
(Alcohol and Health Monograph 1).
Dawes, M. A., Tarter, R. E., & Kirisci, L. (1997). Behavioral self-regulation:
Correlates and 2-year follow-ups for boys at risk for substance abuse. Drug and
Alcohol Dependence, 45, 165–176.
Donovan, J. E., Jessor, R., & Costa, F. M. (1988). Syndrome of problem behavior in
adolescence: A replication. Journal of Consulting and Clinical Psychology, 56,
762–765.
Eysenck, S. B. G., Eastling, G., & Pearson, P. R. (1984). Age norms for
impulsiveness, venturesomeness and empathy in children. Personality and
Individual Differences, 5, 315–321.
Eysenck, S. B. G., & Eysenck, H. J. (1977). The place of impulsiveness in a
dimensional system of personality description. British Journal of Social and
Clinical Psychology, 2, 46–55.
Funder, D. C., & 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.
Funder, D. C., Block, J. H., & Block, J. (1983). Delay of gratification: Some
longitudinal personality correlates. Journal of Personality and Social Psychology,
44, 1198–1213.
Gallucci, N. T. (1997). Correlates of MMPI-A substance abuse scales. Assessment, 4,
87–94.
Green, L., & Snyderman, M. (1980). Choice between rewards differing in amount and
delay: toward a choice model of self-control. Journal of the Experimental Analysis
of Behavior, 34, 135–147.
Henderson, M. J., Galen, L. W., & DeLuca, J. W. (1998). Temperament style and
substance abuse characteristics. Substance Abuse, 19, 61–70.
Herrnstein, R. J. (1990). Rational choice theory: Necessary but not sufficient.
American Psychologist, 45, 356–367.
Huberty, C. J. (1994). Applied discriminant analysis. New York: Wiley.
Delay of Gratification 551

Jessor, R. (1987). Problem behavior theory, psychosocial development, and adolescent

problem drinking. British Journal of Addiction, 82, 331–342.
Jessor, R., Donovan, J. E., & Costa, F. M. (1992). Health Behavior Questionnaire:
High School Form. Colorado: University of Colorado.
Kirby, K. N. (1997). Bidding on the future: Evidence against normative
discounting of delayed rewards. Journal of Experimental Psychology: General,
126, 54–70.
Krueger, R. F., Caspi, A., Moffit, T. W., White, J., & Stouthamer-Loeber, M. (1994).
Delay of gratification, psychopathy, and personality: Is low self-control specific to
externalizing problems? Journal of Personality, 64, 107–129.
Logue, A. W., Forzano, L. B., & Tobin, H. (1992). Independence of reinforcer amount
and delay: The generalized matching law and self-control in humans. Learning and
Motivation, 23, 326–342.
Madden, G. J., Petry, N. M., Badger, G. J., & Bickel, W. K. (1997). Impulsive and
self-control choices in opioid-dependent patients and non-drug-using control
participants: Drug and monetary rewards. Experimental and Clinical Psychophar-
macology, 5, 256–262.
Mischel, W. (1974). Processes in delay of gratification. In L. Berkowitz (Ed.),
Advances in experimental social psychology (Vol. 7, pp. 249–292). San Diego, CA:
Academic Press.
Mischel, W. (1996). From good intentions to willpower. In P. M. Gollwitzer & J. A.
Bargh (Eds.), The psychology of action: Linking cognition and motivation to
behavior (pp. 197–218). New York: Guilford Press.
Mischel, W., Ebbesen, E. B., & Zeiss, A. R. (1972). Cognitive and attentional
mechanisms in delay of gratification. Journal of Personality and Social
Psychology, 21, 204–218.
Mischel, W., Shoda, Y., & Peake, P. K. (1988). The nature of adolescent competencies
predicted by preschool delay of gratification. Journal of Personality and Social
Psychology, 54, 687–696.
Mischel, W., Shoda, Y., & Rodriguez, M. L. (1989). Delay of gratification in children.
Science, 244, 933–938.
Rachlin, H., & Green, L. (1972). Commitment, choice and self-control. Journal of the
Experimental Analysis of Behavior, 37, 15–22.
Rodriguez, M. L., Mischel, W., & Shoda, Y. (1989). Cognitive person variables in the
delay of gratification of older children at-risk. Journal of Personality and Social
Psychology, 57, 358–367.
Rodriguez, M. L., Shoda, Y., Mischel, W., & Wright, J. C. (1999). Self-regulatory
competencies and social adaptation in aggressive children at risk. Manuscript under
review.
Rosenberg, M. (1965). Society and the adolescent self-image. Princeton, NJ: Princeton
University Press.
Shedler, J., & Block, J. (1990). Adolescent drug use and psychological health: A
longitudinal inquiry. American Psychologist, 45, 612–630.
Shoda, Y., Mischel, W., & Peake, P. K. (1990). Predicting adolescent cognitive and
self-regulatory competencies from preschool delay of gratification identifying
diagnostic conditions. Developmental Psychology, 26, 978–986.
Stevenson, M. K. (1993). Decision making with long-term consequences. Temporal
552 Wulfert et al.

discounting for single and multiple outcomes in the future. Journal of Experimental
Psychology: General, 122, 3–22.
Stroop, J. P. (1935). Studies of interference in serial verbal reactions. Journal of
Experimental Psychology, 18, 643–662.
White, J. L., Moffitt, T. E., Caspi, A., Bartusch, D. J., Needles, D. J., & Stouthamer-
Loeber, M. (1994). Measuring impulsivity and examining its relationship to
delinquency. Journal of Abnormal Psychology, 103, 192–205.
Wulfert, E., Safren, S., Brown, I., & Wan, C. (1999). Cognitive, behavioral, and
personality correlates of HIV-positive persons’ unsafe sexual behavior. Journal of
Applied Social Psychology, 29, 223–244.
Wylie, R. C. (1989). Measures of self-concept. Lincoln: University of Nebraska Press.