SAMPLE

1
AIM OF STUDY The purpose of this study was to
develop trajectories of marijuana use among
adolescents and young adults over a 16 year span.
Differences in predictors and outcomes within
these trajectory groups were examined.
DESIGN Year of Test Age at
Each Test Time (Birth Year)
T1 T2 T3 T4
(79-81)
(82-84) (85-87)
(92-94) ________________________________________
__________________________ I. Youngest (1967 -
69) 12---------------15---------------18---------
-----25
II. Middle (1964 - 66)
15---------------18---------------21------------
--28
III. Oldest (1961 -
63) 18---------------21---------------24---------
-----31 Eligible adolescents were recruited
through a random sampling of telephone numbers.
Between 1979-1981 successive rounds of telephone
calls were carried out to fill specified quotas
of 200-225 males and females aged 12, 15, or 18.
An initial anonymous telephone interview served
to identify households with eligible adolescents
and to obtain demographic information. Following
the telephone survey, field interviewers visited
prospective participants in their homes to gather
additional demographic data, interview parents,
and enroll adolescents in the study.
Subsequently, participants came to the test site
for a full day of data collection.
SAMPLE The initial sample of participants (698
males, 682 females) was predominantly white (89)
compared to 83 of the NJ population (U.S. Bureau
of Census, 1981). Half of the participants were
Catholic, 30 were Protestant, 9 are Jewish, and
the remaining 11 had another or no religion,
analogous to the religious composition of NJ. The
median family income of participants parents at
T1, (between 20,000- 29,000), was also
comparable to that of the entire state at that
time (U.S. Bureau of Census, 1981). Most of the
participants lived with both biological parents
(79), 10 lived with a single parent, and 11
lived in another household arrangement, which is
consistent with Census data for that time (U.S.
Bureau of Census, 1981). Overall, participants
were comparable to those who refused to
participate on demographic characteristics and
selected behaviors that were assessed during the
initial telephone interview, except that
participants displayed slightly higher levels of
parental income and education. Yet, both
variables exhibited adequate variation in the
sample. The sample is most representative of
white, working- and middle-class youth living in
a metropolitan area of the Eastern United States.
(For more detail on sample and design, see
Pandina, Labouvie, and White, 1984.) The
participants were tested initially between 1979
and 1981 (Time 1, T1) at the ages of 12 (youngest
cohort), 15 (middle cohort), and 18 (oldest
cohort) (N1380). These participants returned
three years later in 1982-1984 (Time 2, T2),
again in 1985-1987 (Time 3, T3), and finally in
1992-94 (Time 4, T4). Ninety-one percent of the
original participants returned at T4. Those who
dropped out were more likely to be male and to be
older. We tested for attrition bias on all of the
variables used in the analyses and none of the
differences was statistically significant.
Therefore, sample attrition did not affect the
results presented.
MEASURES Subjects marijuana consumption
Marijuana use was measured as the product of last
year frequency times typical quantity. These data
were collected at all four points in time and
were log transformed. Subjects marijuana related
problems A cumulative measure of problems
associated with the use of marijuana, spanning
the domains of legal, social, interpersonal,
interpersonal, physical and mental health. These
data were collected at all four points in time
and were log transformed. Subjects alcohol
consumption Alcohol use was measured as the
product of last year frequency times typical
quantity. These data were collected at all four
points in time and were log transformed. Subjects
alcohol related problems A cumulative measure
of problems associated with the use of alcohol,
spanning the domains of legal, social,
interpersonal, interpersonal, physical and mental
health. These data were collected at all four
points in time and were log transformed. Negative
affect A sum of items reflecting depression and
hostility from the SCL-90 (Derogatis, 1977), and
stress (Dohrenwend Dohrenwend, 1981 Moos,
1986). Arousal needs A sum of items reflecting
sensation seeking and risk taking attitudes and
behaviors (Zuckerman, 1979). Educational
attainment Highest grade completed at T4.
Subjects were classified as either completing
some/all of high school or as attending and/or
completing college. Occupational attainment
Full-time employment at T4. Based upon the
distribution, subjects were classified as either
administrative/professional or another
occupation. Marital status Marital status at
T4. Subjects reported that they either were never
married, married, living together as married,
divorced, separated, or widowed. Depression
Inventory to Diagnose Depression at T4. Parents
alcohol use Participants reported on their
parents frequency of alcohol use. In addition,
parents completed questionnaires (at T1)
regarding their own frequency of beer, wine, and
hard liquor. We took the maximum value for beer,
wine, and hard liquor to create the measure for
frequency of alcohol use. Parental hostility
Subscale derived from factor analysis of the
Streit (1978) Family Perception Inventory to
measure parental hostility/control. For each
parent, the participant responded to how often
parents engaged in certain behaviors. The
parental hostility scale includes 17 items
measured at baseline.
ANALYSIS TO DEVELOP TRAJECTORIES We analyzed
the data using a growth mixture model, which is a
semi-parametric latent-class based modeling
technique (see Muthen Shedden, 1999 Roeder,
Lynch, Nagin, 1999). This mixture model method
allows for cross-group differences in the shape
of developmental trajectories and is, therefore,
especially suited for identifying heterogeneity
in types of developmental trajectories (Nagin
Tremblay, 1999). This technique is based on the
assumption that the population is composed of a
mixture of distinct groups defined by their
developmental trajectories. The approach allows
for identification of population heterogeneity in
the level of a behavior at a given time, as well
as in the development of the behavior over time.
Further, the approach allows one to fit censored
normal distributions to the longitudinal data,
which often reflect the distribution of substance
use in adolescent samples. Finally, the approach
makes full use of the data to determine parameter
estimates and does not lose data through listwise
deletion of cases (Hill,, White, Chung, Hawkins
Catalano, 2000). The analyses were restricted to
users only. Since subjects were followed from
mid-adolescence (age 15) into adulthood (age 31),
we could model both the development of and the
maturation out (cessation) of marijuana related
problems. We conducted these analyses using both
consumption and problems and found parallel
results.
2
RESULTS 1. More than 1/4 of the marijuana users
in our sample followed trajectory 3, where
adolescent limited (ages 15-24) use-related
problems was evident. Approximately 16 of the
users in our sample exhibited heightened levels
of use-related problems over longer periods of
time (groups 6 7 ages 18-31). 2. Lifetime
alcohol use-related problems The number of
alcohol related problems is directly related to
the number of marijuana use-related problems. 3.
Lifetime alcohol use Levels of alcohol
consumption were highest in trajectory group 7,
but could not be differentiated in groups 4, 5 or
6. Trajectory group 7 exhibited a continuous
heightened level of consumption as well as
problems related to use. 4. Lifetime levels of
negative affect Subjects in groups 1 and 3
exhibited lowest levels of negative affect than
all of the other groups. It appears that low
levels of negative affect act as a buffer against
the development of marijuana use behaviors. 5.
Lifetime arousal needs Subjects in groups 1, 2
or 3 could not be differentiated in terms of
arousal needs. Subjects in group 6, while
attaining similarly high levels of arousal needs
as group 7, dropped significantly in the level of
marijuana consumption and problems by age 31,
while subjects in group 7 remained high. It
appears that heightened levels of arousal needs
serve as a risk for heightened levels of
marijuana use behaviors, even into adulthood. 6.
Parental hostility atT1 Subjects in groups 5,
6 and 7 exhibited the highest levels of parental
hostility. High levels of perceived parental
hostility are a risk for the development of
higher levels of marijuana use behaviors, even
into adulthood. 7. Parental alcohol use
Subjects in groups 1, 2 and 3 exhibited lowest
levels of parental alcohol use. Low levels of
parental alcohol use are associated with lower
levels of subjects marijuana use. 8. Divorced
at T4 Among subjects aged 31, group 7 exhibited
the highest percentage of divorces. 9.
Educational attainment at T4 Groups 1 and 7
exhibited the lowest percentage of college
attendees. The fact that trajectory 1 subjects
had relatively few who attended college is
interesting and other mediating or moderating
factors (school performance, intelligence) for
this finding should be explored. 10. Depression
at T4 Trajectory groups 1 and 3 exhibited the
lowest percentage of depressed subjects at T4.
The finding that trajectory group 5 (moderate
continuous users) exhibited the highest
percentage of depressed subjects should be
explored more fully. Other mediating or
moderating factors influencing the relationship
of marijuana use to depression may include
reasons for use and the propensity to self report
instances or symptoms of depression.
REFERENCES Adreason, N.C., Endicott, J.,
Spitzer, R., Winoder, G. (1977). The family
history method using diagnostic criteria.
Archives of General Psychiatry, 34,
1229-1235. Cloninger, C.R. (1987).
Tridimensional Personality Questionnaire (Version
4). Unpublished manuscript. Derogatis, L.R.
(1977). SCL-90-R (questionnaire form).
Administration, Scoring and Procedures Manual,
Vol. 1, Baltimore Johns Hopkins University,
School of Medicine. Dohrenwend, B.S.,
Dohrenwend, B.P. (1981). Stressful Life Events
and Their Contents. Reseda, CA Watson. Hill,
K., White, H.R., Chung, I-J., Hawkins, J.D.,
Catalano, R.F. (2000). Early adult outcomes of
adolescent alcohol use Person- and
variable-centered analyses of binge drinking
trajectories. Alcoholism Clinical and
Experimental Research, 24, 892-901. Moos, R.H.
(1986). Life Stressors and Social Resources
Inventory. Palo Alto, CA Social Ecology
Laboratory, Stanford University. Muthen, B.,
Shedden, K. (1999). Finite mixture modeling with
mixture outcomes using the EM algorithm.
Biometrics, 55, 463-469. Nagin, D.S.,
Tremblay, R.E. (1999). Trajectories of boys
physical aggression, opposition, and
hyperactivity on the path to physically violent
and non violent juvenile delinquency. Child
Development, 70, 1181-1196. Pandina, R.J.,
Labouvie, E.W., White, H.R. (1984). Potential
contributions of the life span developmental
approach to the study of adolescent alcohol and
drug use the Rutgers Health and Human
Development Project, a working model. Journal of
Drug Issues, 14, 253-268. Roeder, K., Lynch,
K.G., Nagin, D.S. (1999). Modeling uncertainty
in latent class membership A case study in
criminology. Journal of the American Statistical
Association, 94, 766-776. Streit, F. (1978).
Technical Manual Youth Perception Inventory.
Highland Park, NJ Essence Publications. U.S.
Bureau of the Census (1981). Current Population
Survey Money, Income and Poverty of Families and
Persons in the United States 1980. Current
Population Reports, No. 127. Zimmerman, M.,
Coryell, W. (1987). The inventory to diagnose
depression (IDD) A self-report scale to diagnose
major depressive disorder. Journal of Consulting
and Clinical Psychology, 55, 55-59. Zuckerman,
M. (1979). Sensation Seeking Beyond the Optimal
Level of Arousal. Hillsdale, NJ Lawrence Erlbaum
Associates, Inc.
ACKNOWLEDGEMENT Supported by NIDA 03395 NIAAA
05823,11699,11594.