You are on page 1of 6


...... ...

The Risks for Late Adolescence of

Early Adolescent Marijuana Use

Judith S. Brook, EdD, Elinor B. Balka, BA, and Martin Whiteman, PhD

This study builds on the work of others for a causal interpretation. Given this topic's
who have examined the impact of marijuana significance, we used 2 measures of risky
use on psychosocial functioning by incorpo- sexual behavior: number of sexual partners
rating several desirable features gleaned from and condom use.
previous research'-5: (1) the longitudinal Both the use of marijuana or other illicit
design necessary for time ordering of vari- drugs and the combination of legal and illegal
ables and more confident predictions; (2) a drug use have been found to increase the like-
5-year interval enabling examination of long- lihood of later deviant behavior, including
term rather than short-term and more transi- theft, vandalism, drug-related crimes, assault,
tory associations; (3) a time interval within and suicide.6">'4 The present study included
important developmental periods, early and theft, vandalism, and violence measures. Our
late adolescence; (4) a focus on marijuana's unconventionality measure, frequency of
associations with problem behaviors and atti- church attendance, has been associated with
tudes, drug problems, and sibling and peer both marijuana and general drug use.2"0"5
behavior; (5) a difficult-to-access inner-city Previous studies suggest that early ado-
African American and Puerto Rican sample; lescent marijuana use is associated with such
and (6) controls on early problems in exami- drug-related attitudes as tolerance ofdeviance
nations of the relationship of early marijuana and risk taking and the perception of drugs
use with later problems. as not harmful.6 Other investigators report
In the review to follow we focus on stud- that adolescents who experiment with and
ies in which marijuana use is the only inde- later increase their use ofmarijuana minimize
pendent variable. These are supplemented by the adverse physical and psychological
and distinguished from studies of drug use, effects ofmarijuana. 16
some combining legal and illegal drugs, that Research indicates considerable stability
have implications for this investigation. We in marijuana and general drug use behav-
expect early adolescent marijuana use to be ior.4,5,17-19 Studies have shown that adoles-
related to an increase in problem behaviors cents who report ever being "stoned" are
such as lowered educational achievement. likely to continue using marijuana. The
Previous research regarding this relationship strongest predictor of any substance use is
is not conclusive. Some investigators report earlier use of that substance; marijuana use
that earlier marijuana use is associated with also predicts use of other illicit drugs.'3 This
subsequent lowered achievement motivation stability may increase the risk of problems
among high school students.6'7 Others have with specific drugs.
not found that marijuana use affects educa- Drug use has consequences not only for
tional achievement.2 However, because the the user. Studies suggest that drug use by a
temporal order of the relationship between child serves as a behavioral model for his or
marijuana use and achievement motivation
has been questioned, we used a measure of
actual educational attainment rather than one The authors are with the Department of Commu-
that explicitly includes motivation.8'9 nity and Preventive Medicine, Mount Sinai School
In adolescence, marijuana use may have of Medicine, New York City.
an impact on sexual problem behavior. New- Requests for reprints should be sent to Judith
S. Brook, EdD, Department of Community and Pre-
comb and Bentler'0 used "latent-variable ventive Medicine, Mount Sinai School of Medicine,
structural modeling" to analyze the effects of Box 1044A, 1 Gustave L. Levy P1, New York, NY
adolescent marijuana use on young adult 10029-6574.
functioning and, in addition, met the criteria This article was accepted April 15, 1999.

American Journal of Public Health 1549

Brook et al.

her siblings and predicts siblings' associating incentive. Those who did not included, for different persons have you had sex with?"
with deviant peers.2024 In a longitudinal example, youths in prisons. There, the authori- The time frame implied was "ever." The
investigation of African American youth, high ties disallowed any incentive. group defined as at high risk reported more
levels of gateway drug use were related to We compared the participants seen at than 1 sexual partner. Condom use was also
increases in peer drug models.25 Among high both time 1 and time 2 with those who took measured with a single item, "How often
school students, regular marijuana users had part only at time 1. There were no differences were condoms (rubbers) used when you and
more friends who exhibited deviant behavior between these time 2 participants and non- your partner had sex?" The group defined as
and used drugs and fewer high-achieving participants in terms of age, sex, and ethnic- at high risk reported rates of condom use
friends at follow-up than did nonusers or infre- ity at time 1. A few of the variables were not ranging from "never" to "often"; the low-risk
quent users.6 The present study assessed the assessed at time 1. On those that were, the time group "always" used them.
relationship of earlier marijuana use with later 2 participants and nonparticipants differed Incapacitation at school or work was
sibling drug use and deviant peer associations. only in regard to time 1 peer deviance. This measured with a 3-item scale involving the
Illicit drug use also increases the chances difference was probably a chance finding. question "How often in the last 6 months
of the user's being a victim of violence.14 This were you high, drunk, or stoned while at
study examined as well the relationship of Measures school or work [on different substances]?"32
marijuana use to being an assault victim. The group defined as at high risk reported
In summary, this longitudinal study Relationships between drug use and being stoned on at least 1 substance.
investigated some of the psychosocial risks many measures have been established with Drug-related attitudes. Tolerance of
of early adolescent marijuana use among other samples.lS15l282'29 The scales used in this deviance and risk taking was measured by a
African Americans and Puerto Ricans. These study were based on item intercorrelations and 6-item scale regarding (1) the extent to which
later correlates include other problem behav- reliabilities and were grouped as follows: (1) activities such as faking an excuse note are
iors, drug-related attitudes and problems, and problem behaviors, (2) drug-related attitudes, wrong and (2) how well certain behavioral
sibling and peer problem behavior. We exam- (3) drug problems, and (4) sibling and peer styles, such as liking to live dangerously,
ined these relationships as a function of sex, problem behavior. Several scales were adapta- described participants.33'34 Percepfion of drug
ethnicity, and age. tions of measures used in previous studies. The risk was assessed with a 3-item scale involv-
scales were adapted to ensure their linguistic ing the question "How much do people risk
and culturl relevance. The alpha coefficients harming themselves if they sometimes use
Methods for the multi-item scales were as follows: self- [a specific drug or group of drugs]?"35 The
deviancy, 0.83; violence toward others, 0.79; group defined as at high risk indicated "no
The data for the present study were incapacitated at school or work, 0.70; toler- risk" in all 3 instances. Others' intolerance of
obtained in a 5-year follow-up of 695 African ance of deviance-risk taking, 0.67; perception drug use was measured with a 4-item scale in
American and 637 Puerto Rican adolescents of drug risk, 0.51; others' intolerance of drug which participants were asked how friends or
who were initially seen in 1990 (time 1), use, 0.77; peer deviancy, 0.79; and violence family would feel if the participants experi-
when the sample was drawn from 11 schools toward the subject, 0.69. In the subsequent mented with marijuana or other illegal drugs
serving the East Harlem area of New York descriptions, scales without a reference origi- once or twice. On each item, the group
City. At time 1, the sample included 713 girls nated with the authors. defined as at high risk reported that others
and 619 boys, and their mean age was Marijuana use. At time 1, frequency of would think it is "okay," the most tolerant
14 years (SD = 1.31). (For more details ever using marijuana was measured with a response possible.
regarding time 1 procedures, see Brook single item. For another of our samples, the Drug problems. This measure consisted
et al.26'27) The 1995 (time 2) sample included R value of this item, measured twice over of 4 single-item scales asking if the partici-
627 African American and 555 Puerto Rican approximately the same number of years and pant had ever had a problem with each sub-
older adolescents and young adults, of whom age span, was 0.44. The response options stance (J. Endicott, PhD, written communica-
648 were female and 534 were male. At time ranged from "never used" (1) to "used once a tion, November 1993).
2, their mean age was 19 years (SD = 1.48). week or more" (5). In this study, the mean Sibling and peer problem behavior.
We used extensive tracking procedures level of use was 1.16 (SD = 0.63). The group The 2 sibling measures and 2 peer measures
to locate the original participants. A passive defined as at high risk reported using mari- were single-item scales in which partici-
consent format was used with parents of par- juana about once a month (a scale response pants were asked how many of their siblings
ticipants younger than 18 years. Trained inter- of 3) or more often. and how many of their peers had ever used
viewers from the National Opinion Research Problem behavior. We determined level marijuana and how many had ever used
Center located and interviewed the youths of education by asking for participants' cur- other illegal drugs. The respective high-risk
with a structured questionnaire. Virtually all rent grade. At time 1, all participants were 7th groups had 1 or more siblings or peers who
of the participants were interviewed by an to 10th graders (mean= 8.27, SD= 1.04). At were users. In the 3-item peer deviancy
interviewer of the same sex and ethnicity. The time 2, those not attending school reported the scale, participants were asked how many of
respondents, following along with their own last year they had completed. The high-risk their friends had engaged in theft, aggres-
copy of the questionnaire, answered aloud group had obtained an 11th-grade education sion, or behavior that led to police involve-
after the interviewer read the question. For sec- or less. Self-deviancy, the participants' delin- ment.30 The 5-item violence toward the sub-
tions regarding drug use, the respondent and quent behavior, was measured with a 9-item ject scale assessed how often the subject had
the interviewer exchanged booklets, and the scale emphasizing frequency of theft and been assaulted in various ways.3'
respondent marked his or her own responses aggression.30 The subject's violence toward
directly on the questionnaire. Puerto Rican others was measured with a 4-item scale con- Data Analytic Procedure
youths were given the option to complete the ceming frequency of weapon use.31
interview in either English or Spanish. The Number of sexual partners was mea- We used logistic regression analysis to
great majority of respondents received a $25 sured with the question "About how many assess the association between early adoles-

October 1999, Vol. 89, No. 10

1550 American Journal of Public Health
Early Adolescent Marijuana Use

TABLE 1 -Longitudinal Odds Ratios for Time 2 Psychosocial Factors Associated With Time 1 Marijuana Use
OR With Control on
OR With Control on Time 1 Psychosocial Factors
Time 2 Psychosocial Factor Covariates (95% Cl) and Covariates (95% Cl)
Problem behavior
Level of education (low) 1.91* (1.05, 3.47) 2.00* (1.09, 3.66)
Self-deviance 3.51** (1.59, 7.76) 1.96* (1.03, 3.73)
No. of sexual partners 4.44** (1.27, 15.53) 3.61** (1.02,12.78)
Less condom use 4.23** (1.45,12.34) 3.58** (1.22,10.55)
Incapacitated at school or work 2.37* (1.13, 4.98) ...a
Violence toward others 2.59** (1.19, 5.62) ...a
Church attendance (less) 2.73** (1.29, 5.81) 2.34** (1.07, 5.15)
Drug-related attitudes
Tolerance of deviance-risk taking 1.99* (1.08, 3.64) 1.41 (0.75, 2.65)
Perception of drug risk 0.41* (0.20, 0.80) 0.49* (0.24, 0.97)
Others' intolerance of drug use 0.39** (0. 1 6, 0.98) 0.57 (0.28,1.16)
Drug problems
Cigarette problems 3.53** (1.59, 7.85) ...
Alcohol problems 2.55* (1.21, 5.38) ...a
Marijuana problems 2.49* (1.24, 5.02) ...a
Other illegal drug problems 2.69 (0.60,12.16) ...a
Sibling and peer problem behavior
Sibling marijuana use 2.09* (1.19, 3.69) 1.15 (0.62, 2.12)
Sibling illegal drug use 0.75 (0.27, 2.13) 0.46 (0.15,1.39)
Peer deviance 2.34** (1.08, 5.07) 1.91* (1.04, 3.49)
Peer marijuana use 3.71** (1.18, 11.63) 1.95 (0.79, 4.76)
Peer illegal drug use 1.02 (0.46, 2.24) 0.99 (0.44, 2.19)
Violence toward subject 2.81** (1.27, 6.22) ...

Note. The covariates are age, sex, and ethnicity. OR = odds ratio; Cl = confidence interval.
*p <.05; **P <.01.
aVariable not available at time 1.

cent (time 1) marijuana use and late adoles- lier problems, in part because such problems both with and without control on time 1
cent (time 2) psychosocial factors. Odds were more likely only after longer or more church attendance.
ratios estimated the increase in odds or risk of regular use.
time 2 problem behavior associated with an Drug-Related Attitudes
increase in time 1 marijuana use. An odds
ratio greater than 1.0 represented a positive Results Time 1 marijuana use also was related to
relation between adolescent marijuana use time 2 attitudes and perceptions about prob-
and problem behavior. An odds ratio less than Problem Behavior lem behavior. Thus, time 1 marijuana use
1.0 depicted an inverse relation. Statistical was associated with subsequent increased
significance was demonstrated if the confi- The odds of late adolescent problem tolerance of deviance and risk-taking behav-
dence intervals were above 1.0 for the former behavior given early adolescent marijuana ior and with a lesser likelihood of perceiving
and below 1.0 for the latter. use, with controls on age, sex, and ethnicity, that drugs are harmful. The relationship
Review of the response distribution for are presented in Table 1. An increase in ado- between early marijuana use and the per-
the continuous measure of marijuana use lescent marijuana use was associated with a ceived harmfulness of drugs at time 2 was
indicated that dichotomizing it to form a decrease in the likelihood of attaining at least significant with control on this measure at
high-risk group (those scoring in the top a high school education, a more than tripling time 1. Finally, time 1 marijuana use was
25%) and a low-risk group (the remaining ofthe risk of self-deviancy, and an increase in related to an increased odds of older adoles-
subjects) provided sufficient numbers for the risk of all of the other problem behaviors, cents' perceiving others to be tolerant of
the analysis. The psychosocial measures including violence toward others, sexual their drug use.
were treated similarly. Fewer than 5% of the problem behaviors, and being incapacitated
interactions of the variables with age, sex, at school or work. This indicates a signifi- Drug Problems
and ethnicity were significant, so no interac- cant, long-term association between time 1
tion term was included in the regressions. marijuana use and time 2 problem behavior. Time 1 marijuana use was associated
Each time 2 outcome was regressed against In addition, time 1 marijuana use was related with an almost 4-fold increase in the likeli-
age, sex, and ethnicity. In additional analy- to an increased odds of time 2 self-deviancy, hood of problems with cigarettes and a more
ses, the time 2 outcome was regressed having more than 1 sexual partner, and not than doubling of the odds of alcohol and mar-
against its time 1 measure (when available) always using condoms and decreased odds of ijuana problems. However, there was no
and against the 3 demographic variables. In finishing high school, despite control on increase in the risk of time 2 problems with
analyses of the relation of time 1 marijuana these behaviors at time 1. Time 1 marijuana other illegal drugs. Because time 1 data
use to later problems with the use of various use was associated with an increased odds of were not available, we could not determine
substances, there were no controls for ear- less frequent church attendance at time 2 whether the impact of time 1 marijuana use

October 1999, Vol. 89, No. 10 American Journal of Public Health 1551
Brook et al.

was sustained when time 1 problems were associated with a later decrease in the effort young adults and with her work regarding
controlled. necessary to stay in school through gaduation "stage of drug use," although we focused on
beyond whatever achievement and motiva- problem use, not just use, of drugs.44"45 The
Sibling and Peer Problem Behavior tional limitations existed earlier. In this way, absence of problems with other illegal drugs
manjuana use may contribute to limiting the suggests that between time 1 and time 2, a
We examined the relationship of mari- acquisition of skills that maximize employ- sufficient number of marijuana users had not
juana use with siblings' use of drugs. Time 1 ment opportunities in young adulthood.'0"8 In used enough other illegal drugs to have
marijuana use was associated with time 2 sib- the present study, some of the neuropsycho- developed problems attributable to them.
ling marijuana use without control on time 1 logical consequences of marijuana use (e.g., However, the results of other studies suggest
sibling use. However, time 1 marijuana use impairment of attention and short-term mem- that marijuana users may be at risk for
was not related to other illegal drug use by ory) may have mediated the development of becoming users ofother illicit drugs as young
siblings. Also, marijuana use was related to educational problems, leading time 1 users to adults.2
an increased risk of having friends who were become dropouts.9 28'39 This possibility indi- Time 1 marijuana use did not affect time
delinquent and used marijuana but not to an cates the need for experimental studies ofmar- 2 sibling marijuana use after earlier sibling
increased risk ofhaving peers who used other ijuana's impact. use had been controlled. Perhaps the associa-
illegal drugs. Of these relationships, only the The association of time 1 marijuana use tion with sibling modeling reflects such fac-
association with having delinquent peers was with both number of sexual partners and con- tors as siblings' previous dfinking or having
sustained with control on its time 1 measure. dom use at time 2 also has wider implica- deviant peers.2O'22 Time 1 marijuana use was
Finally, time 1 marijuana use was associated tions. These sexual behaviors involve a risk related to having more deviant friends at time
with an increased risk of the participants' of sexually transmitted diseases and preg- 2 than earlier. This relationship suggests that
being the victims ofviolence at time 2. nancy. Among this study's respondents, hav- young people select peers whose unconven-
ing more than 1 sexual partner and not tional lifestyles are similar to theirs.6,23,24
always using condoms might heighten the Finally, our finding of the likelihood of time
Discussion risk for contracting and transmitting HIV and 1 marijuana users' becoming assault victims
having shortened lifetimes. is consistent with the instances of violent
This longitudinal study assessed the Contrary to the results of Newcomb and deaths suffered by alcohol and illicit drug
relation of using marijuana at least monthly Bentler,'0 our findings suggest that early mar- users.'4
at time 1 to aspects of time 2 psychosocial ijuana use is correlated with later deviance A family interactional perspective
functioning in a large inner-city sample. The beyond the effect of earlier similar behavior. focusing on the association ofearlier problem
time interval suggests that the relationships Thus, this relationship, previously found with behavior with later drug use has guided much
are more than transitory, yet the interval is a predominantly White sample, has been gen- of our work.'5 Examining this interrelation-
short enough to detect the impact on aspects eralized to African American and Puerto ship over time indicated that the antecedents
of time 2 functioning pertinent to developing Rican youth.'2 Engaging in multiple problem were often also the effects of drug use.'2 This
independence. behaviors in adolescence heightens the risk study of marijuana use further substantiates
Our results suggest that time 1 mari- of criminality, alcohol abuse, and drug abuse those effects.
juana use was related to certain time 2 prob- in adulthood.4 This study has several limitations. First,
lem behaviors, attitudes, and peer character- Time 1 marijuana use's association with it relied on unconfirmed self-reports. Studies
istics, all of which are risk factors for further time 2 church attendance may reflect the indicate that self-report is a reasonably valid
drug use.36 Many of these relationships were user's decreasing attachment to conventional means of investigating drug use.46 Moreover,
sustained despite control on early functioning institutions. Along with marijuana's other this study's validity was enhanced by (1)
(e.g., delinquency). Instances in which they relationships, this association suggests the ensuring confidentiality, (2) having partici-
were not sustained may have been due to a loosening of bonds to social forces that pro- pants mark their own answers to the drug
controlled mediating factor. For example, tect against drug use and unconventional items, (3) using unfamiliar interviewers, and
time 1 marijuana use was related to time 1 affiliations (e.g., delinquent peers), both of (4) using a school-based sample rather than
tolerance of deviance, which in turn was which do not support developing sound inde- those accused of criminal acts.46'47 Also,
associated with time 2 tolerance of deviance. pendence.37'41 including peer self-reports rather than relying
The loss of the relationship between time 1 Our significant finding for the associa- on the accuracy of participants' perceptions
marijuana use and time 2 tolerance of deviance tion of marijuana use with a later perception of peer behavior would have enhanced this
with control on time 1 tolerance of deviance that drugs are not harmful (after controlling study. Still, our results are consistent with
supports this assumption. However, the rela- for this attitude at time 1) substantiates oth- those of most of the investigators cited.
tionship between time 1 use and the time 2 ers' comparable results.'6 This relationship is Second, there may be an underlying
psychosocial measure may also stem from also consistent with Bem's self-perception variable, such as self-control or attraction to
their both being related to an antecedent theory, which suggests that people modify antisocial values, that is related to both mari-
rather than a mediating factor. their later attitudes to accord with their earlier juana use and psychosocial factors. This vari-
Among the problem behaviors, the result behavior.42'43 This attitude change enables able may have contributed to the relation-
for educational attainment is consistent with youth to avoid confronting the realistic con- ships found. However, controlling for the
those of other studies.7'9 Our finding extends sequences of marijuana use. Bailey et al. time 1 measure may also have controlled for
to Puerto Rican youth and confirms for implied that denial of risk can result in con- this underlying variable (e.g., controlling for
African American youth what has been found tinued use.'6 time 1 self-deviance controlled for time 1
with predominantly African American sam- Time 1 marijuana use was related to an antisocial values). For some of our dependent
ples. The longitudinal impact on education of increased risk of time 2 problems with ciga- variables, we did not have time 1 measures.
even moderate marijuana use is suggestive of rette, alcohol, and marijuana use. This fmd- Such findings need replication with controls.
an amotivational syndrome.37 Such use may be ing is consistent with Kandel's results for Finally, only future research will indicate

1552 American Journal of Public Health October 1999, Vol. 89, No. 10
Early Adolescent Marijuana Use

whether the associations between early and and white adolescents. Genet Soc Gen Psychol in adolescent risk for alcohol use and abuse. In:
late adolescence are sustained into young Monogr. 1989;1 15:349-369. Brody GH, ed. Sibling Relationships: Their
adulthood. 7. Friedman AS, Bransfield S, Kreisher C. Early Causes and Consequences. Norwood, NJ:
Because the present study was correla- teenage substance use as a predictor of educa- Ablex; 1996:1-30.
tional-vocational failure. Am JAddict. 1994; 23. Kandel DB. On processes of peer influences in
tional, we have minimized the use of causal 3:325-336. adolescent drug use: a developmental perspec-
language. However, the longitudinal design 8. Sussman S, Stacy AW, Dent CW, Simon TR, tive. AdvAlcohol SubstAbuse. 1985;4: 139-163.
and the controls on prior psychosocial factors Johnson CA. Marijuana use: current issues and 24. Kaplan HB, Johnson RJ, Bailey CA. Deviant
and demographic characteristics suggest new research directions. J Drug Issues. 1996; peers and deviant behavior: further elaboration
causality and thus inferences for public 96:695-733. of a model. Soc Psychol Q. 1987;50:277-284.
health. Early marijuana use may have an 9. Krohn MD, Thornberry TP, Collins-Hall L, 25. Farrell AD, Danish SJ. Peer drug associations
influence on later problems stemming from Lizotte AJ. School dropout, delinquent behav- and emotional restraint: causes or consequences
cigarette, alcohol, and marijuana use; number ior, and drug use. In: Kaplan HB, ed. Drugs, of adolescents' drug use? J Consult Clin Psy-
Crime, and Other Deviant Adaptations: Longi- chol. 1993;61:327-334.
of sexual partners; and failure to always use tudinal Studies. New York, NY: Plenum Press; 26. Brook JS, Hamburg BA, Balka EB, Wynn PS.
condoms. 1995:163-183. Sequences of drug involvement in African-
Hence, assessments of marijuana use 10. Newcomb MD, Bentler PM. Consequences of American and Puerto Rican adolescents. Psy-
should be incorporated into clinical practice Adolescent Drug Use. Newbury Park, Calif: chol Rep. 1992;71:179-182.
with adolescents. More information is needed Sage Publications; 1988. 27. Brook JS, Whiteman M, Balka EB, Hamburg
regarding treatment for early users and their 11. Brook JS, Whiteman M, Balka EB, Win PT, BA. African-American and Puerto Rican drug
concurrent problems to prevent the later Gursen MD. Similar and different precursors to use: personality, familial, and other environ-
development of problem behavior. Under- drug use and delinquency among African mental risk factors. Genet Soc Gen Psychol
Americans and Puerto Ricans. J Genet Psychol. Monogr 1992;18:417-438.
standing which early users are most vulnera- 1998;159: 13-29.
ble to subsequent difficulties would help in 28. Johnson RJ, Kaplan HB. Stability of psycholog-
12. Brook JS, Whiteman M, Finch SJ, Cohen P. ical symptoms: drug use consequences and
identifying those who would benefit most Young adult drug use and delinquency: child- intervening processes. J Health Soc Behav.
from treatment. [ hood antecedents and adolescent mediators. 1990;31:277-291.
JAm Acad Child Adolesc Psychiatry. 1996;35: 29. Jessor R, Jessor SL. Problem Behavior and
1584-1592. Psychosocial Development: A Longitudinal
Contributors 13. Kandel DB, Davies M, Karus D, Yamaguchi K. Study of Youth. New York, NY: Academic Press
J. S. Brook planned the study, supervised the data The consequences in young adulthood of ado- Inc; 1977.
analysis and interpretation, and wrote the paper. lescent drug involvement. Arch Gen Psychiatry. 30. Gold M. Undetected delinquent behavior. JRes
E. B. Balka helped plan the study, interpret the data, 1986;43:746-754. Crime Delinquency. 1966;3:27-46.
and write the paper and made most of the revisions. 14. Rivara FP, Mueller BA, Somes G, Mendoza CT, 31. Chavez EL, Oetting ER, Swaim RC. Dropout
M. Whiteman also assisted with the planning of the Rushforth HB, Kellerman AL. Alcohol and and delinquency: Mexican-American and Cau-
study and the interpretation of the results. illicit drug abuse and the risk ofviolent death in casian non-Hispanic youth. J Clin Child Psy-
the home. JAMA. 1997;278:569-575. chol. 1989;23:47-55.
15. Brook JS, Brook DW, Gordon AS, Whiteman 32. Newcomb MD. Drug Use in the Workplace:
M, Cohen P. The psychosocial etiology of
Acknowledgments adolescent drug use: a family interactional
Risk Factors for Disruptive Substance Use
This research was supported by grant DA05702 and Among Young Adults. Dover, Mass: Auburn
approach. Genet Soc Gen Psychol Monogr. House; 1988.
research scientist award K05DA00244 from the 1990; 116:111-267.
National Institute on Drug Abuse. 33. Jackson DN. Personality Research Form.
16. Bailey SL, Flewelling RL, Rachal JV Predict-
We are grateful to Pe Thet Win for running the ing continued use of marijuana among adoles- Goshen, NY: Research Psychologists Press;
programs and for statistical advice. We also appreci- 1974.
cents: the relative influence of drug-specific
ate the help of our research assistants, Erica Johnson and social context factors. J Health Soc Behav. 34. Jessor R, Graves TD, Hanson RC, Jessor SL.
and Michal Gursen. 1992;33:51-65. Society, Personality, and Deviant Behavior: A
17. Guy SM, Smith GM, Bentler PM. Conse- Study of a Tri-Ethnic Community. New York,
quences of adolescent drug use and personality NY: Holt, Rinehart & Winston; 1968.
References factors on adult drug use. J Drug Educ. 1994; 35. Johnston JD, Bachman JG, O'Malley PM.
24:109-132. Monitoring the Future Study, 1986. Ann Arbor,
1. Eggert LL, Herting JR. Drug involvement Mich: Survey Research Center, Institute for
among potential dropouts and "typical" youth. 18. Kaplan HB, Johnson RJ, Bailey CA. Explaining
adolescent drug use: an elaboration strategy for Social Research, University of Michigan; 1987.
JDnugEduc. 1993;23:31-55. 36. Brook JS, Balka EB, Gursen MD, Brook DW,
2. Jessor R, Donovan JE, Costa FM. BevondAdo- structural equations modeling. Psychiatry.
lescence: Problem Behavior and Young Adult 1988;51:142-163. Shapiro J, Cohen P. Young adults' drug use: a
19. Windle M. A longitudinal study of antisocial 17-year longitudinal inquiry of antecedents.
Development. New York, NY: Cambridge Uni-
versity Press; 1991. behaviors in early adolescence as predictors of PsycholRep. 1997;80:1235-1251.
late adolescent substance use: gender and eth- 37. Baumrind D, Moselle KA. A developmental
3. Newcomb MD. Prospective dynamics of intoxi-
cation in the workplace: personal and job- nic group differences. JAbnorm Psychol. 1990; perspective on adolescent drug abuse. Adv
related predictors and consequences. Exp Clin 99:86-91. Alcohol SubstAbuse. 1985;4:41-67.
20. Ary DV, Tildesley E, Hops H, Andrews J. The 38. Brook JS, Newcomb MD. Childhood aggression
Psychopharmacol. 1995;3:56-74. and unconventionality: impact on later academic
4. Newcomb MD, Jack L. Drug use, agency, and influence of parents, siblings, and peer model-
communality: causes and consequences among ing and attitudes on adolescent use of alcohol. achievement, drug use, and workforce involve-
adults. Psvlchol Addict Behav. 1995;9:67-82. IntJAddict. 1993;28:853-880. ment. J Genet Psychol. 1995; 156:393-413.
5. Stein JA, Smith GM, Guy SM, Bentler PM. 21. Brook JS, Whiteman M, Gordon AS, Brook 39. Pope HG, Gruber AJ, Yurgelun-Todd D. The
Consequences of adolescent drug use on young DW. The role of older brothers in younger residual neuropsychological effects of can-
adult job behavior and job satisfaction. JAppl brothers' drug use viewed in the context of par- nabis: the current status of research. DrugAlco-
Psvchol. 1993;78:463-474. ent and peer influences. J Genet Psychol. 1990; hol Depend. 1995;38:25-34.
6. Brook JS, Gordon AS, Brook A, Brook DW. 15 1:59-75. 40. Stattin H, Magnusson D. Antisocial develop-
The consequences of marijuana use on intraper- 22. Conger RD, Rueter MA. Siblings, parents, and ment: a holistic approach. Dev Psychopathol.
sonal and interpersonal functioning in black peers: a longitudinal study of social influences 1996;8:617-645.

October 1999, Vol. 89, No. 10 American Journal of Public Health 1553
Brook et al.

41. Brunswick AF, Rier DA. A structural strain: 43. Stacy AW, Newcomb MD, Bentler PM. Cogni- of research methods. In: Harrison L, Hughes A,
drug use among African-American youth. In: tive motivation and drug use: a 9-year longitudi- eds. The Validity of SelfReported Drug Use:
Taylor RL, ed. African-American Youth: Their nal study. JAbnorm Psychol. 1991;100:502-515. Improving the Accuracy of Survey Estimates.
Social and Economic Status in the United 44. Kandel DB. Marijuana users in young adult- Rockville, Md: National Institute on Drug
States. Westport, Conn: Praeger; 1995:225-246. hood. Arch Gen Psychiatry. 1984;41:200-209. Abuse; 1997:17-36.
42. Bem DJ. Self-perception theory. In: Berkowitz 45. Kandel DB. Stages in adolescent involvement 47. Mensch BS, Kandel DB. Under-reporting of
L, ed. Cognitive Theories in Social Psychology. in drug use. Science. 1975;190:912-914. substance use in a national longitudinal youth
New York, NY: Academic Press Inc; 1978: 46. Harrison L. The validity of self reported drug cohort: individual and interviewer effects. Pub-
221-282. use in survey research: an overview and critique lic Opinion Q. 1988;52:100-124.

This is the 9th in a series of manuals oftestsfor syphilis, first published by APHA in conjunction with Centers
for Disease Control & Prevention in 1990. This practical guide for the laboratory-assisted diagnosis of
syphilis contains sections on laboratory safety, specimen collection and clinical diagnosis using antibody aind
antigen detection methods. The format complies with the CLIA-88 regulations. This edition is imiore com-
prehensive than the previous ones because it includes additional tests beyond those considered Standard
Status Tests. Procedures are included for commercially-available provisional tests as well as tests for which
many of the individual components may be purchased.

OF TESTS Highlights of Contents

.* Laboratory Biosafety and Quality Control
FOR .. Collection of Blood and Cerebrospinal Fluid
*: Diagnostic Tests:
SYPH ILIS .. Darkfield Microscopy for the Detection and Identification of Treponema palidum
*: Direct Fluorescent Antibody Test for Treponema pallidum
Edi d by*: Direct Fluorescent Antibody Tissue Test for Treponema paUidum
Edited by *:* Venereal Disease Research Laboratory, Unheated Serum Reagin (USR) Test
*:- Rapid Plaisma Reagin (RPR) 18-mm Circle Card Test
Sandra A. Larsen *> Rapid Plasma Reagin 18 mm Circle &ird Test (RPR), Toluidine Red Unheated Serum Test (TRUST)
Victoria Pope Fluorescent
and FTA-ABS Treponemal
Double Staining (FTA-ABS DS)(FTA-ABS)
Antibody-Absorption Tests
Robert E. Johnson
Robert **.:.~ ~ e Treponemal iand Nontreponemal
Treponemal Western Blot Enzyme Immunoassays (EIA)
Edward J. Kennedy, Jr. Stock no. 087553-234-9/SYAD98 * 370 Pages Softcover
$ 48.00 Non-nwmebrs O $ 34.00 Members

American Public Health Association * Publications Sales

PO Box 753, Waldorf, MD 20604-0753
Tel: 301/893-1894; Fax: 301/843-0159; Web:

1554 American Journal of Public Health October 1999, Vol. 89, No. 10