Professional Documents
Culture Documents
2056 | Research and Practice | Peer Reviewed | Ramchand et al. American Journal of Public Health | November 2007, Vol 97, No. 11
RESEARCH AND PRACTICE
Beginning at the sixth year of follow-up, were aged between 10 and 14 years, scale RESULTS
all the adolescents were asked about their values of affiliation with drug-using peers
experiences with tobacco, and beginning at and affiliation with deviant peers as well as Sample
the 10th year of follow-up, they were asked dichotomous indicators of whether the ado- Table 1 presents descriptive characteristics
about the average time they spent working lescents were in a grade level lower than of the baseline and restricted samples (those
for pay during the school year. Of the total grade 6 (as an indicator of grade retention) interviewed in both years 10 and 11). Across
799 adolescents with some information at or had any friends aged older than 15 years all samples, close to 55% were male and
grade 1, 570 were interviewed at year 10 were all included in the model.24–26 We 85% were African American. Of the total
(501 of the 678 original recruits; 69 of the also included a scale of parent monitoring, sample, a little more than half received free
additional 121), 515 were interviewed at in which higher scores indicate less moni- school lunch at the first grade (a proxy of
year 11, and 488 were interviewed in both toring, in the model.27 From year 9, di- socioeconomic status for adolescents).29 Com-
years 10 and 11 (by definition, all were chotomous indicators of whether the ado- plete information on baseline socioeconomic
from the original 678). Parents received lescents were in a grade level lower than and intervention status were available on
consent information, and adolescents com- grade 9 or had any friends aged older than adolescents at year 11 and in the restricted
pleted an assent procedure and were as- 17 years were included, as was a measure sample, which reflects the follow-up strategy.
sured of their confidentiality before each in- of parents’ highest level of educational The prevalence of work increased from year
terview. attainment, a proxy for socioeconomic sta- 10 to year 11. At year 10, 26% of the adoles-
tus.28 The difference between year 9 and cents surveyed reported working for pay; by
Analysis year 6 scale values for drug-using peer affil- year 11, close to 40% worked for pay. There
In the first stage of the analysis, we exam- iation, deviant peer affiliation, and parent were no apparent differences between the dis-
ined the link between working for pay and monitoring were included in the final tribution of work status among the separate
substance use at the 10th year of follow-up, model. We also included gender and race year-10 or year-11 samples and the restricted
when the adolescents were aged between as potential confounds that were collected sample of adolescents who were assessed in
14 and 18 years (the mean age was 16 at baseline. both years. Among those who worked in year
years). Current tobacco use was a dichoto- To assess the possibility of additional selec- 10, 30% were employed in fast-food restaurants
mous variable that indicated use in the past tion effects, we conducted survival analysis to and the remainder were employed babysitting
30 days. We collected work intensity (the examine the age at first use of tobacco strati- (11%), working in restaurants (9%), working as
average number of hours adolescents fied by the year-10 work categories specified a store clerk (12%), or in other occupations, in-
worked per week during the school year) in above. Because the proportionality assump- cluding cosmetology and other retail (20%). By
5-hour intervals; we subsequently catego- tion for this model was valid when adoles- year 11, babysitting reduced to only 4%, al-
rized these responses into 1 of 3 categories cents who worked more than 10 hours were though 30% of adolescents who worked still
of work: not working during the past year, compared with a combined category of ado- worked in fast food and the remainder worked
working an average of 1 to 10 hours per lescents who did not work and who worked at restaurants (12%), stores (15%), or in other
week (moderate work intensity), or working less than 10 hours, a Cox–proportional haz- occupations, such as construction, cosmetology,
an average of more than 10 hours per week ards regression model was used to empiri- retail, service (i.e. catering, teaching dance, etc),
(high work intensity). In separate analyses, cally quantify the difference in the hazard and maintenance and cleaning jobs (24%; data
we used other thresholds (15 hours and 20 function of smoking based on year-10 work on type of job not shown).
hours), which yielded qualitatively similar, status. The prevalence of current tobacco use in-
although less precise, results (because of the The final analytic step was to assess the creased from approximately 13% at year 10
smaller sizes of the high-intensity group). incidence of tobacco use between years 10 (which is similar to state-level estimates of
We used logistic regression to assess the re- and 11. Adolescents at risk for incident to- public school–attending 10th graders in Mary-
lationship between working for pay and cur- bacco use were defined as never reporting land) to 17% at year 11.30 Among the 488
rently using tobacco. tobacco use since the year-6 assessment adolescents assessed in both years, 50% did
To examine the possibility that adoles- through year 10. Incidence was then esti- not work in both years, 25% did not work in
cents more likely to use tobacco were also mated across the entire sample at risk and year 10 but worked in year 11, and the re-
more likely to work at year 10 (i.e., a poten- then separately across 4 work transition cat- mainder was roughly evenly distributed be-
tial selection effect), early childhood charac- egories: did not work in both years 10 and tween working in year 10 but not in year 11
teristics previously associated with adoles- 11, did not work in year 10 but worked in and working in both years (data not shown).
cent tobacco use were incorporated into the year 11, worked in year 10 but did not work
logistic model. We included teachers’ re- in year 11, and worked in both years. Relative Cross-Sectional Analysis
ports of adolescents’ early aggressive be- risk estimates were calculated with adoles- Of the 570 adolescents assessed at year
haviors as exhibited in the first grade in the cents not working in both years as the refer- 10, complete information was available
model.23 From year 6, when adolescents ence group. for 400 on the first-grade, sixth-year, and
November 2007, Vol 97, No. 11 | American Journal of Public Health Ramchand et al. | Peer Reviewed | Research and Practice | 2057
RESEARCH AND PRACTICE
TABLE 1—Descriptive Characteristics of Samples, by Year of Follow-Up: Second Generation Constructs that have been previously asso-
of the Prevention and Intervention Research Center Study, Baltimore, Md, 1993–2004 ciated with substance-using behaviors were
entered into the logistic model to assess for
Restricted a potential selection effect. Estimates from
Baseline, No. Year 10, No. Year 11, No. Sample,a No.
the multivariable model (also presented in
(Category %) (Category %) (Category %) (Category %)
Table 2) revealed that after including these
Total 799 (100.00) 570 (100.00) 515 (100.00) 488 (100.00) variables, the relationship between high work
Baseline characteristics intensity and current tobacco use was some-
Gender what attenuated and the corresponding CI
Male 435 (54.44) 315 (55.26) 283 (54.95) 263 (53.89) now included the null value (adjusted OR
Female 364 (45.56) 255 (44.74) 232 (45.05) 225 (46.11) [AOR] = 1.98; 95% CI = 0.84, 4.67). Further
Raceb analyses revealed that 1 construct was largely
African American 678 (84.86) 492 (86.32) 452 (87.77) 429 (87.91) responsible for attenuating the relationship
White 120 (15.02) 78 (13.68) 63 (12.23) 59 (12.09) between high work intensity and tobacco use:
Proxy measures of socioeconomic status changes in drug-using peers between years
Not available 6 (0.75) 4 (0.70) 5 (0.97) 4 (0.82) 6 and 9, which itself remained linked with
Received free school lunch 411 (51.44) 305 (53.51) 317 (61.55) 297 (60.86) current use of tobacco (AOR = 1.18; 95%
Received reduced-price school lunch 52 (6.51) 41 (7.19) 40 (7.77) 39 (7.99) CI = 1.07, 1.31). The only other variable in
Paid by child 209 (26.16) 151 (26.49) 153 (29.71) 148 (30.33) the multivariable model that remained linked
Missing 121 (15.14) 69 (12.11) 0 (0.00) 0 (0.00) with current use of tobacco was affiliation
Intervention status with drug-using peers in year 6 (AOR = 1.23;
Control group 261 (32.67) 181 (31.75) 169 (32.82) 157 (32.17) 95% CI = 1.05, 1.44).
Classroom-centered group 258 (32.29) 182 (31.93) 167 (32.43) 163 (33.40)
Family–school partnership group 260 (32.54) 192 (33.68) 179 (34.76) 168 (34.43) Age of First Tobacco Use
No design 20 (2.50) 15 (2.63) 0 (0.00) 0 (0.00) Kaplan–Meier (product–limit) estimates for
Year 10 follow-up characteristics age of first use of tobacco by 10th-year work
Work status intensity are presented in Figure 1. Aside
No work 419 (73.51) 361 (73.98) from a few sparse respondents who initiated
1–10 h/wk 61 (10.70) 54 (11.07) tobacco use earlier than age 7 years, the ear-
> 10 h/wk 90 (15.79) 73 (14.96) liest tobacco users initiated use at age 7
Current tobacco use 74 (12.98) 58 (11.89) years. The median failure time (age of first
Year 11 follow-up characteristics use) among adolescents who worked more
Work status than 10 hours at year 10 was at age 13 years.
No work 320 (62.14) 300 (61.48) Nonworkers at year 10 had a median failure
1–10 h/wk 54 (10.49) 52 (10.66) time at age 14 years, whereas moderate
> 10 h/wk 141 (27.38) 136 (27.87) workers had a median failure time at age 15
Current tobacco use 89 (17.28) 75 (15.37) years. Estimates from a Cox–proportional
a hazards model indicated an increased hazard
The restricted sample included only those adolescents surveyed in both years 10 and 11.
b
One student was of Hispanic ethnicity at baseline and was lost to follow-up at years 10 and 11. ratio for high-intensity workers compared
c
These students joined the classrooms during first grade. with a collapsed category of moderate-
intensity workers and nonworkers (hazards
ratio = 1.34; 95% CI = 1.01, 1.78).
ninth-year variables used in the multivariable 95% confidence interval [CI] = 1.40, 6.11), al-
logistic regression model. Univariate associa- though there was not strong evidence to link Longitudinal Analysis
tions between constructs included in the ad- moderate work with current use (OR = 1.38; The incidence of tobacco use between years
justed model and current tobacco use for 95% CI = 0.50, 3.83). There were additional 10 and 11 across the entire sample and the 4
these 400 adolescents are presented in positive associations in univariate analyses be- work transition categories is presented in Table
Table 2 (estimates are qualitatively similar tween current tobacco use and high levels of 3. A total of 488 adolescents were surveyed in
in univariate analyses that included all ado- aggression in first grade, reductions in parent both the 10th and 11th year of follow-up. Of
lescents). These results showed that high- monitoring between years 6 and 9, affiliation these, 214 had not yet reported using tobacco
intensity workers at year 10 were 3 times with peers who used drugs in year 6, and in- at year 10. Overall, the 1-year incidence of to-
more likely to report current use of tobacco creases in affiliations with peers who used bacco use in this “at risk” sample was 14%.
than were nonworkers (odds ratio [OR] = 2.93; drugs between years 6 and 9. However, there were significant differences in
2058 | Research and Practice | Peer Reviewed | Ramchand et al. American Journal of Public Health | November 2007, Vol 97, No. 11
RESEARCH AND PRACTICE
TABLE 2—Prevalence for Categorical Variables and Logistic Regression Results, Tobacco the incidence of tobacco use across the work-
Use at Year 10: Second Generation of the Prevention and Intervention Research Center transition categories. Incidence of smoking was
Study, Baltimore, Md, 1993–2004 (N = 400) 0.29 among adolescents who did not work in
year 10 but did work in year 11 versus 0.04
Current Tobacco Usea among the adolescents who did not work in
Prevalence, OR AOR both periods (relative risk = 8.0; 95% CI = 2.8,
No. % (95% CI) (95% CI) 22.9). In addition, the incidence of tobacco use
Work status among adolescents who worked in both peri-
No work (Ref) 299 8.5 1.00 1.00 ods was 0.31, which yielded a relative risk of
1–10 h/wk 44 11.4 1.38 (0.50, 3.83) 1.50 (0.46, 4.82) 8.7 (95% CI = 2.9, 15.8) compared with non-
> 10 h/wk 61 21.3 2.93 (1.40, 6.11) 1.98 (0.84, 4.67) workers. Although both relative risk estimates
Gender have corresponding 95% CIs that are wide,
Male (Ref) 215 11.6 1.00 1.00 the lower bounds for both are far from the
Female 185 9.7 0.82 (0.43, 1.55) 1.04 (0.48, 2.22) null, which signals strong evidence of an ele-
Race
vated risk.
African American (Ref) 352 9.7 1.00 1.00
White 48 18.8 2.16 (0.96, 4.83) 1.25 (0.43, 3.63)
DISCUSSION
Academic performance
The results from this study indicated a
≥ Grade 6 or above (year 6; Ref) 366 11.2 1.00 1.00
clear relationship between working for pay
< Grade 6 34 5.9 0.49 (0.11, 2.14) 0.22 (0.04, 1.32)
and adolescent tobacco use. It is noteworthy
≥ Grade 9 or above (year 9; Ref) 331 9.4 1.00 1.00
that this association held among adolescents
< Grade 9 69 17.4 2.04 (0.99, 4.20) 2.07 (0.76, 5.63)
in a primarily urban environment. Previous
First grade behaviorsb
research indicated that adolescents who find
Authority acceptance . . .c . . .c 1.54 (1.08, 2.18) 1.43 (0.94, 2.19)
c c jobs in urban areas are unique because they
Parent monitoring—year 6 ... ... 1.02 (0.94, 1.10) 0.98 (0.88, 1.11)
have found work where jobs are likely
Change in parent monitoring—from year 6 to year 9 . . .c . . .c 1.07 (1.00, 1.15) 1.03 (0.95, 1.12)
scarce.14,15,31 They also may be more likely to
Peer affiliationd . . .c . . .c
work to pay for transportation and supple-
Peers who use drugs—year 6 . . .c . . .c 1.11 (1.02, 1.21) 1.23 (1.05, 1.44)
c
mental school materials.15,32 However, we
Change in peers who use drugs—from year 6 to year 9 ... . . .c 1.14 (1.06, 1.22) 1.18 (1.07, 1.31)
found that the time adolescents report work-
Deviant peers—year 6 . . .c . . .c 1.07 (0.99, 1.16) 1.01 (0.88, 1.16)
ing for pay during the school year was posi-
Change in deviant peers—from year 6 to year 9 . . .c . . .c 1.03 (0.96, 1.10) 0.99 (0.90, 1.10)
tively associated with the likelihood of cur-
No peers aged older than 15 years—year 6 (Ref) 198 9.1 1.00 1.00
rent tobacco use. Thus, it appears that the
Peers aged older than 15 years—year 6 202 12.4 1.41 (0.74, 2.68) 1.32 (0.62, 2.82)
relationship between work and tobacco use
No peers aged older than 17 years—year 9 (Ref) 134 7.5 1.00 NA 1.00 NA
seen previously in both nationally representa-
Peers aged older than 17 years—year 9 266 12.4 1.76 (0.84, 3.68) 1.08 (0.46, 2.53) tive and community surveys of more-affluent
Parents’ educational attainment adolescents exists even in urban areas.
High school graduate (Ref) 160 10 1.00 1.00 Evidence of a selection effect is supported
Eighth grade or less 5 60 . . .e . . .e by these analyses, which indicate that the
Some high school 51 15.7 1.67 (0.67, 4.18) 1.48 (0.51, 4.27) cross-sectional relationship between work
Vocational/college/graduate 184 8.7 0.86 (0.41, 1.77) 0.92 (0.41, 2.07) hours and tobacco use was somewhat attenu-
Note. OR = odds ratio (crude); CI = confidence interval; AOR = adjusted odds ratio. All estimates were derived from respondents ated in models in which we controlled for a
with no missing data and who were assessed at the 10th year of follow-up. number of earlier childhood predictors of
a
Current tobacco use was defined as having used tobacco in the past 30 days. adolescent substance use. However, because
b
Early childhood characteristics previously associated with adolescent tobacco use were incorporated into the logistic model.
Teachers’ reports of adolescents’ early authority acceptance, an indicator of aggressive behaviors, as exhibited in the first work information was not collected until year
grade were included in the model. A scale of parent monitoring, in which higher scores indicate less monitoring, was also 10 and changes in drug-using peers between
included in the model. years 6 and 9 largely drove this attenuation,
c
Continuous variables; OR represents change per unit.
d
From year 6, when adolescents were aged between 10 and 14 years, scale values of affiliation with drug-using peers and these results should be interpreted cautiously.
affiliation with deviant peers as well as dichotomous indicators of whether the adolescents were in a grade level lower than We can not exclude the possibility that a
grade 6 (as an indicator of grade retention) or had any friends aged older than 15 years were all included in the model.24–26 young person started working between years
From year 9, dichotomous indicators of whether the adolescents were in a grade level lower than grade 9 or had any friends
aged older than 17 years were included. 6 and 9 and developed friendships with
e
Small sample sizes impeded ability to calculate estimates. coworkers who used drugs, which would
clearly implicate working as a risk factor
November 2007, Vol 97, No. 11 | American Journal of Public Health Ramchand et al. | Peer Reviewed | Research and Practice | 2059
RESEARCH AND PRACTICE
2060 | Research and Practice | Peer Reviewed | Ramchand et al. American Journal of Public Health | November 2007, Vol 97, No. 11
RESEARCH AND PRACTICE
Strengths and Limitations as fast food, other restaurant work, and retail), 4. Resnick MD, Bearman PS, Blum RW, et al. Pro-
There are 2 significant limitations of this efforts should be made to encourage these in- tecting adolescents from harm. Findings from the Na-
tional Longitudinal Study on Adolescent Health. JAMA.
study. In the 10th and 11th years of follow-up, dustries be smoke-free. Previous research has 1997;278:823–832.
adolescents were asked only limited informa- indicated that totally smoke-free workplaces 5. Bachman JG, Schulenberg J. How part-time work
tion about their work experiences. Other work- are associated with reductions in prevalence intensity relates to drug use, problem behavior, time
related constructs, such as work duration or of smoking, and we believe that these policies use, and satisfaction among high school seniors: are
these consequences or merely correlates. Dev Psychol.
work schedules, also likely impact young work- might also lead to reduced smoking incidence 1993;29:220–235.
ers’ tobacco use, but were not ascertained in among young people.43 Second, efforts should 6. Valois RF, Dunham AC, Jackson KL, Waller J.
these surveys. Additional information about be made to establish workplace-based preven- Association between employment and substance abuse
adolescents’ work experiences has been col- tion programs. These programs would ideally behaviors among public high school adolescents. J Ado-
lesc Health. 1999;25:256–263.
lected in year 12, although these data are not incorporate prevention strategies pertinent to
7. Safron DJ, Schulenberg JE, Bachman JG. Part-time
yet available for analysis. The other significant young workers, such as fiscal management
work and hurried adolescence: the links among work
limitation is temporal sequencing. Although our and ways to cope with stress while at work. intensity, social activities, health behaviors, and sub-
longitudinal study assessed adolescents annu- Finally, this research should highlight the need stance use. J Health Soc Behav. 2001;42:425–449.
ally, if adolescents moved from not working to for future studies to more carefully and sys- 8. Wu LT, Schlenger WE, Galvin DM. The relation-
ship between employment and substance use among
working between the year-10 and year-11 as- tematically evaluate the impact that working
students aged 12 to 17. J Adolesc Health. 2003;32:
sessments and also initiated tobacco use be- for pay has on the substance-using behaviors 5–15.
tween these 2 surveys, it is unclear whether of adolescents. 9. Steinberg LD, Fegley S, Dornbusch SM. Negative
working preceded smoking or vice versa. Fi- impact of part-time work on adolescent adjustment:
nally, potentially important confounders associ- evidence from a longitudinal study. Dev Psychol. 1993;
About the Authors 29:171–180.
ated with adolescents smoking in the past, such At the time of the study, Rajeev Ramchand, Nicholas S.
10. Paternoster R, Bushway S, Brame R, Apel R. The
as discretionary income and parental smoking Ialongo, and Howard D. Chilcoat were with the Department
effect of teenage employment on delinquency and
of Mental Health, Johns Hopkins Bloomberg School of
behaviors, are not included in the survey or problem behaviors. Soc Forces. 2003;82:297–335.
Public Health, Baltimore, Md.
analysis. Requests for reprints should be sent to Dr Rajeev Ramc- 11. Turner L, Mermelstein R, Flay B. Individual and
Our study carries with it strengths that off- hand, Rand Corporation, 1200 S Hayes St, Arlington, VA contextual influences on adolescent smoking. Ann N Y
22202-5050 (e-mail: rajeev_ramchand@rand.org). Acad Sci. 2004;1021:175–197.
set these limitations. First, the cohort of ado-
This article was accepted November 14, 2006. 12. Light A. High School Employment. In: Bureau of
lescents came from a well-defined population
Labor Statistics, ed. National Longitudinal Survey Dis-
of adolescents who entered first grade class- cussion Paper, report NLS 95-27. Washington, DC: US
Contributors
rooms in a specific geographic region, thereby Department of Labor; 1995.
R. Ramchand originated and performed all analyses
reducing biases inherent in convenience or and led the writing. N. S. Ialongo was the principal in- 13. Board on Children, Youth and Families. Protecting
vestigator of the larger cohort study and contributed to Youth at Work: Health, Safety, and Development of
other sampling procedures. The strengths of
the writing and analyses. H. D. Chilcoat also contributed Working Children and Adolescents in the United States.
the longitudinal structure were exploited to to the writing and analyses. All authors helped to con- Washington, DC: National Academy Press; 1998.
control for childhood constructs assessed ceptualize ideas, interpret findings, and review drafts of 14. Mortimer JT. Working and Growing Up in America.
the article.
prospectively and, thus, less subject to recall Cambridge, Mass: Harvard University Press; 2003.
bias, including measures of first-grade aggres- 15. Newman KS. No Shame in My Game: The Working
sion and age at first use of tobacco. The longi- Acknowledgments Poor in the Inner City. 1st ed. New York, NY: Knopf and
This work was supported by grants from the National the Russell Sage Foundation; 1999.
tudinal design also allowed us to look at inci- Institute on Drug Abuse (DA018013 and DA11796)
16. Newcomb MD, Bentler PM. Consequences of Ado-
dent tobacco-using behaviors specific to and from the National Institute of Mental Health
lescent Drug Use: Impact on the Lives of Young Adults.
(MH57005).
general work transition categories, which to Newbury Park, Calif: Sage Publications; 1988.
our knowledge, have never been performed 17. Staff J, Uggen C. The fruits of good work: early
elsewhere. Finally, the sample was an urban- Human Participant Protection work experiences and adolescent deviance. J Res Crime
This study was approved by the Johns Hopkins Delinquency. 2003;40:263–290.
based community cohort of primarily African Bloomberg School of Public Health institutional review
18. Bachman JG, Safron DJ, Sy SR, Schulenberg JE.
American youth. Research on work in this board.
Wishing to work: new perspectives on how adoles-
population is limited, and our study lends ev- cents’ part-time work intensity is linked to educational
idence to the notion that the link between References disengagement, substance use, and other problem
1. Healthy People 2010: Understanding and Improving behaviors. Intl J Behav Dev. 2003;27:301–315.
working and tobacco is generalizable to set-
Health. 2nd ed. Washington, DC: US Department of 19. Storr CL, Ialongo NS, Kellam SG, Anthony JC. A
tings of varying economic conditions. Health and Human Services; 2000. randomized controlled trial of two primary school in-
2. Grunbaum JA, Kann L, Kinchen S, et al. Youth tervention strategies to prevent early onset tobacco
Conclusions risk behavior surveillance—United States, 2003. smoking. Drug Alcohol Depend. 2002;66:51–60.
We recommend 3 public health responses MMWR Surveill Summ. 2004;53:1–96. 20. Furr-Holden CD, Ialongo NS, Anthony JC, Petras H,
to these findings. First, given that young work- 3. Centers for Disease Control and Prevention. Ciga- Kellam SG. Developmentally inspired drug prevention:
rette use among high school students—United States, middle school outcomes in a school-based randomized
ers in this study and in most other studies are 1991–2003. MMWR Morb Mortal Wkly Rep. 2004; prevention trial. Drug Alcohol Depend. 2004;73:
concentrated among certain industries (such 53:499–502. 149–158.
November 2007, Vol 97, No. 11 | American Journal of Public Health Ramchand et al. | Peer Reviewed | Research and Practice | 2061
RESEARCH AND PRACTICE
21. Ialongo NS, Werthamer L, Kellam SG, Brown CH, 38. Catalano R, Morrison DM, Wells EA, Gillmore MR,
Wang S, Lin Y. Proximal impact of two first-grade pre-
ventive interventions on the early risk behaviors for
later substance abuse, depression, and antisocial behav-
Iritani B, Hawkins JD. Ethnic differences in family
factors related to early drug initiation. J Stud Alcohol.
1992;53:208–217.
Shortage?
ior. Am J Community Psychol. 1999;27:599–641.
39. Clark PI, Scarisbrick-Hauser A, Gautam SP,
22. Lambert SF, Brown TL, Phillips CM, Ialongo NS. Wirk SJ. Anti-tobacco socialization in homes of African-
The relationship between perceptions of neighborhood American and white parents, and smoking and non-
characteristics and substance use among urban African smoking parents. J Adolesc Health. 1999;24:329–339.
American adolescents. Am J Community Psychol. 2004; A Warning Shot:
34:205–218. 40. Sussman S, Dent CW, Flay BR, Hansen WB,
Johnson CA; Centers for Disease Control and Preven- Influenza and the
23. Kellam S, Ialongo N, Brown H, et al. Attention tion. Psychosocial predictors of cigarette smoking onset
problems in first grade and shy and aggressive behav- by White, Black, Hispanic, and Asian adolescents in
2004 Flu Vaccine
iors as antecedents to later heavy or inhibited sub-
stance use. NIDA Res Monogr. 1989;95:368–369.
Southern California. MMWR Morbid Mortal Wkly Rep.
1987;36(suppl 4):11S–16S.
Shortage By Tim
24. Kandel D. Adolescent marihuana use: role of par- Brookes
41. Hirschi T. Crime and the family. In: Wilson J, ed.
ents and peers. Science. 1973;181:1067–1070.
Crime and Public Policy. San Francisco, Calif: Institute
25. Lloyd JJ, Anthony JC. Hanging out with the wrong for Contemporary Studies; 1983:53–68.
crowd: how much difference can parents make in an
urban environment? J Urban Health. 2003;80: 42. McMorris BJ, Uggen C. Alcohol and employment
383–399. in the transition to adulthood. J Health Soc Behav.
2000;41:276–294.
26. Bryant AL, Schulenberg J, Bachman JG, O’Malley PM,
Johnston LD. Understanding the links among school 43. Fichtenberg CM, Glantz SA. Effect of smoke-free
misbehavior, academic achievement, and cigarette use: workplaces on smoking behaviour: systematic review.
a national panel study of adolescents. Prev Sci. 2000;1: BMJ. 2002;325:188.
71–87.
27. Chilcoat HD, Dishion TJ, Anthony JC. Parent
monitoring and the incidence of drug sampling in
urban elementary school children. Am J Epidemiol.
1995;141:25–31.
28. Miech RA, Hauser RM. Socioeconomic status and
health at midlife. A comparison of educational attain-
ment with occupation-based indicators. Ann Epidemiol. This important book gives the facts
2001;11:75–84.
and the analysis necessary to under-
29. Ensminger ME, Forrest CB, Riley AW, et al. The
validity of measures of socioeconomic status of adoles-
stand what happened and how to
cents. J Adolesc Res. 2000;15:392–419. avoid similar crises in the future.
30. 2002 Maryland Adolescent Survey. Baltimore: ORDER your copy TODAY!
Maryland State Department of Education; 2003. ISBN 0-87553-049-4 •Softcover • 2005
31. Wilson WJ. When Work Disappears: The World of $13.97 APHA Members • $19.99 Nonmembers
the New Urban Poor. 1st ed. New York, NY: Knopf; 1996. plus shipping and handling
32. Entwisle DR, Alexander KL, Olson LS. Early work
histories of urban youth. Am Sociol Rev. 2000;64: American Public
279–297. Health Association
33. Johnston LD, O’Malley PM, Bachman JG, Publication Sales
Schulenberg JE. Monitoring the Future: National Survey Web: www.aphabookstore.org
Results on Drug Use, 1975–2005, Volume 1: Secondary E-mail: APHA@pbd.com
School Students. Bethesda, Md: National Institute on Tel: 888-320-APHA
Drug Abuse; 2006. NIH publication 06-5883. Fax: 888-361-APHA
34. Substance Abuse and Mental Health Services Ad-
ministration. Results From the 2004 National Survey on Don’t get caught with the Flu!!!
Drug Use and Health: National Findings. Rockville, Md:
Office of Applied Studies; 2004. NSDUH series H-28;
DHHS publication SMA 05-4062.
35. Wallace JM Jr, Bachman JG, O’Malley PM,
Schulenberg JE, Cooper SM, Johnston LD. Gender and
ethnic differences in smoking, drinking and illicit drug
use among American 8th, 10th, and 12th grade stu-
dents, 1976–2000. Addiction. 2003;98:225–234.
36. Ellickson PL, Orlando M, Tucker JS, Klein DJ.
From adolescence to young adulthood: racial/ethnic
disparities in smoking. Am J Public Health. 2004;94:
293–299.
37. Griesler PC, Kandel DB. Ethnic differences in
correlates of adolescent cigarette smoking. J Adolesc
Health. 1998;23:167–180.
2062 | Research and Practice | Peer Reviewed | Ramchand et al. American Journal of Public Health | November 2007, Vol 97, No. 11