Abstract
Purpose: To examine the correlates of cigarette smoking among African-American, Hispanic, and white adolescents in a cross-sectional national sample.
Methods: A total of 1795 mother–child dyads from the 1992 National Longitudinal Survey of Youth were selected for analyses. Measures of adolescents cigarette smoking and family, individual, peer, and sociodemographic risk factors were analyzed.
Results: White youths reported the highest rates of lifetime, current, and persistent smoking, and initiated smoking at a significantly earlier age than African-Americans and Hispanics. Except for maternal cigarette smoking and substance use, African-Americans and Hispanics experienced a disproportionately larger number of purported risk factors than whites. Multivariate analyses revealed common and ethnic-specific correlates of adolescent lifetime and current smoking, with many more significant associations among whites than minorities. Common correlates included youth’s age across all three ethnic groups, problem behaviors and delinquency among whites and African-Americans, and perceived peer pressure to smoke among whites and Hispanics. Ethnic-specific correlates included maternal smoking, maternal cocaine use, low maternal religiosity, and negative scholastic attitudes, which increased smoking for whites; and positive parenting, which reduced smoking for African-Americans.
Conclusions: The lack of effects of maternal smoking and perceived peer pressure to smoke on African-American adolescents compared with whites suggests that role modeling and interpersonal influence may be more important determinants of smoking for white than African-American adolescents. The differential impact of family and peer factors on the smoking of adolescents of different ethnicity warrants further investigation.
Keywords
Many studies report substantially higher prevalence of licit and illicit drug use among white than Hispanic, and especially, African-American adolescents
1
, 2
, 3
, 4
, 5
, 6
. These ethnic differences are particularly striking with respect to cigarette smoking. The differential between African-American and white adolescents steadily became wider through 1992, over a 20-year period spanning a general decrease in adolescent drug use 7
. Beginning in 1993, rates of smoking started to increase, especially among African-American males 5
, 8
. Despite these marked differences in the prevalence of cigarette smoking between African-American and white youths, much remains to be understood about the risk factors associated with ethnic differences in adolescent smoking.There are relatively few systematic analyses of ethnic differences in the correlates of adolescent smoking. The majority of studies have been based exclusively on white or mixed ethnic samples that contain too few youths of different ethnicity to obtain reliable comparisons across groups. Findings reported for whites have not always been replicated for other groups, nor have consistent patterns been documented. The contradictory results across investigations vary as a function of the ethnic composition of the sample, whether it is mixed (multiethnic) or exclusively African-American, its representativeness, and the nature of the data (cross-sectional vs. longitudinal). Etiologic explanations advanced to account for ethnic differences have emphasized differential exposure to risk and protective factors, and differential vulnerability to risk factors
9
. Common and ethnic-specific effects on adolescent smoking have been reported, although cigarette smoking has not always been disaggregated from the use of other substances.Factors related to adolescent smoking belong to five domains of influence: family, youth characteristics, peer group, community factors, and mass media
10
, 11
, 12
. We describe family, individual, peer, and sociodemographic risk factors for smoking identified in ethnically undifferentiated samples and ethnic-specific samples, when available. Community and media influences are not discussed because relevant data were not available in the data set analyzed.U.S. Department of Health and Human Services. Preventing tobacco use among young people: A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994.
Family factors and adolescent smoking
Family factors are important predictors of adolescent cigarette smoking and substance use more generally. Parenting practices, including lack of parent–child closeness, weak or excessive controls, inconsistent discipline, and ineffective monitoring are positively associated with cigarette initiation and current smoking among white, African-American, and Hispanic youths
13
, 14
, 15
. The negative relationship of adolescent substance use with maternal–child closeness, attachment, and support has been reported to be stronger among African-Americans than whites, while the reverse has been reported for proactive parenting 16
, 17
. Other ethnic-specific family correlates of current cigarette smoking include parent–child conflict in African-Americans 3
and lack of parental norms against smoking 18
, 19
, 20
, disruptive family events, frequent accidents, illnesses, and relocations in whites 21
. Taken together, these studies suggest that across ethnic groups, the quality of the parent–child relationship is an important correlate of adolescent cigarette smoking, but the specific quality indicator varies across investigations and by ethnicity. Attachment and normative prescriptions against smoking emerge as important factors against smoking among African-Americans 16
, 17
, 18
, 19
, 20
. Family religiosity is related to lower rates of cigarette smoking and general substance use in all ethnic groups 12
, but particularly African-Americans U.S. Department of Health and Human Services. Preventing tobacco use among young people: A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994.
2
.There is a positive association between parental and adolescent smoking. However, the majority of findings are based predominantly on white and ethnic aggregated-samples, and have relied almost exclusively on the child’s perceptions of parental smoking behavior rather than on parental self-reports
14
, 22
. The effects of parental smoking on adolescent smoking are greater in white than African-American families, or absent altogether in African-American families 3
, 23
, 24
, 25
, 26
. Maternal smoking appears to be more influential than paternal smoking 14
. Patterns of parental influence on adolescent smoking depend on parental level of use, child gender, and age 22
.Individual factors and adolescent smoking
Across ethnic groups, behavior problems and delinquency are consistently related to adolescent smoking
12
. In multiethnic samples, cigarette smoking is predicted by behavior problems in whites and risk-taking in African-Americans U.S. Department of Health and Human Services. Preventing tobacco use among young people: A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994.
3
, 26
, and is associated with aggression in whites and Hispanics 27
, and delinquency across ethnic groups 28
. In an exclusively African-American sample, males rated as aggressive or shy/aggressive in the first grade were most likely to have used cigarettes by age 17 years 29
. In a predominantly white sample, childhood conduct problems predicted daily smoking at age 15 years 30
.In comparison with youth behavior problems, ethnic variations in the relationships among depression, self-esteem, academic functioning, and adolescent cigarette smoking have been examined less extensively and have yielded mixed findings. Depression is correlated with smoking in adolescence in representative samples
31
and current cigarette smoking among Hispanics 3
. Low self-esteem predicts the initiation and continued use of cigarettes in representative 12
and ethnically differentiated samples U.S. Department of Health and Human Services. Preventing tobacco use among young people: A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994.
32
.Commitment to the goals of conventional institutions such as school and church is related to lower rates of smoking. Poor school performance, negative scholastic attitudes, and low educational aspirations are associated with adolescent cigarette use in ethnic undifferentiated samples
10
, 11
and among whites and Hispanics in multiethnic samples 26
, 27
. In exclusively African-American samples, poor school grades and low scholastic esteem are associated with lifetime and current smoking and intentions to smoke 24
.Peers and adolescent smoking
Peer influence, as measured by peer smoking
21
, 23
, 33
, 34
, 35
, 36
, 37
, 38
, peer approval 33
, and normative 35
and social pressure 36
, is strongly associated with adolescent cigarette smoking 10
and other forms of substance use. Peer influence for smoking appears to be stronger among whites and Hispanics than African-Americans 3
, 15
, 23
, 25
, 26
, 37
, 38
.Sociodemographic characteristics and adolescent smoking
Age is positively associated with cigarette smoking across ethnic groups
12
. For males, the rates of lifetime and current smoking are equivalent to or slightly higher than those of females U.S. Department of Health and Human Services. Preventing tobacco use among young people: A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994.
12
. Ethnic differences appear in relation to socioeconomic status. Low family socioeconomic status and low parental education are associated with higher rates of lifetime and current smoking, and predict smoking onset in predominantly white U.S. Department of Health and Human Services. Preventing tobacco use among young people: A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994.
10
, 12
but not African-Americans samples U.S. Department of Health and Human Services. Preventing tobacco use among young people: A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994.
24
. Living in a single-parent household is associated with higher rates of lifetime cigarette use among whites and African-Americans 12
, U.S. Department of Health and Human Services. Preventing tobacco use among young people: A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994.
24
.The common and ethnic-specific risk factors for adolescent cigarette smoking identified to date and the lack of consistent findings underscore the need for further exploration of risk factors in multiethnic samples within a multivariate framework. The white–African-American differential in adolescent smoking persists after statistical adjustment for family background and lifestyle variables
39
. The present study examines family, individual, peer, and sociodemographic factors uniquely associated with cigarette smoking in a cross-sectional national sample of African-American, Hispanic, and white adolescents and their mothers.Methods
Subjects
The analyses are based on 1795 mother–child dyads, with children aged 10 years and older, from 1238 families in the 1992 wave of the National Longitudinal Survey of Youth (NLSY). The NLSY is a multistage stratified area probability sample of over 12,000 youths representative of persons born between 1957 and 1964 in the coterminous United States. African-Americans, Hispanics, and economically disadvantaged white youths have been oversampled. Since 1979, respondents have been interviewed annually through personal household interviews with relatively low attrition rates; 72% of the initial sample of women were reinterviewed in 1992. Beginning in 1986, the children of all female respondents have been assessed and followed biennially. The NLSY children interviewed to date are born to young mothers and are not a nationally representative sample born to the full age spectrum of mothers
40
. The 1992 survey included three child-related supplements. One probed mothers about their child’s behavioral and social functioning; a second supplement, orally administered by the interviewer, assessed youth’s perceived competence; a youth self-administered component inquired about family relations, scholastic expectations and attitudes, mood, substance use, and deviance. Although the 1994 data were recently released, the analyses are based on the 1992 survey, since it contains the most comprehensive battery of family, child, and peer assessments for the total sample of youth.In 1992, the average age of the 1795 children 10 years and older (910 males, 885 females) was 12.4 years [standard deviation (SD) = 1.9], with 60.1% 10–12 years old. The 1795 youth included all the eligible youth from 1238 households; 791 families included one child, 355 two children, 79 three children, 10 four children, 2 five children, and 1 seven children. The mothers were on average 32.4 years old (SD = 2.1). The unweighted ethnic distribution was 41.7% African-American, 23.7% Hispanics, and 34.6% white. The mothers had 11.8 years of education (SD = 2.0). The average family income from all sources was $31,200 (SD = $21,150); 25.7% of the families lived in poverty and 51.3% had a biological father absent. The analytical sample of 1238 families constitutes 88.1% of the 1405 families of mothers with children aged 10 years and older interviewed in 1992. In 167 families (11.9%), 284 age-eligible youth did not complete the self-administered supplement and were excluded. While the total 1992 sample of 3056 interviewed mothers with children of all ages represents 91.9% of mothers eligible to be interviewed, the completion rate for families with children aged 10 years and older was not available from the Center for Human Resource Research (CHRR) at Ohio State.
Measures
All measures were assessed in 1992, unless otherwise specified.
Family characteristics
Socialization practices
Home Observation and Measurement of the Environment–short form (HOME-SF)
The total score of the HOME-SF (for children aged 10 years and older) measured the quality of the youth’s home environment. Based on maternal reports and interviewer impressions, the total score (27 items) sums the relevant dichotomized items from the cognitive and emotional subscales, which the latter assesses family involvement and maternal disciplinary techniques. Internally normed, age-standardized scores for the total sample (x̄ = 100, SD = 15) were used
40
. Internal consistency was low (∝ = 0.37).Maternal–child closeness, monitoring, rules, and shared activities
Mothers and children were independently asked about the frequency of behaviors reflecting levels of closeness, overall satisfaction in their relationship, and parental monitoring of the child. Combined mother–child scales were constructed to decrease individual response bias and to obtain more reliable estimates of parenting behavior. Youth were asked about the presence of family rules and shared family activities. Closeness averaged three 4-point items: the youth’s perceptions of: (a) how well mother and child shared/talked about important issues and (b) the closeness of the relationship, and (c) the mother’s evaluation of the quality of the relationship: (∝ = 0.58). Monitoring averaged three 4-point items tapping the mother’s knowledge of (a) her child’s friends, (b) with whom her child spends time when away from home, and (c) the youth’s perception of the mother’s knowledge in this domain (∝ = 0.44). Rules summed four items regarding youths’ reports of family rules for television viewing, dating, attending parties, and keeping parents informed of their whereabouts (∝ = 0.37). Shared activities summed eight items asking youth if they engaged in the activities of moviegoing, shopping, going to church or out for dinner with a parent during the past month, and if they had done schoolwork, played a game or sport, and built, sewn, or cooked with a parent in the past week (∝ = 0.60).
Importance of religion/religiosity
The mother’s perceived importance of providing religious training to her child was determined (1 = “not at all important” to 3 = “very important”).
Maternal substance use
Maternal cigarette smoking
Two dichotomous smoking variables indicated the mother’s (a) lifetime and (b) current smoking.
Maternal use of other substances
In 1984, 1988, and 1992, mothers were asked if they had ever used marijuana or cocaine; and in 1992, they were asked about the frequency and quantity of alcohol use in the past 30 days. Two dichotomous variables indicated maternal lifetime use of (a) marijuana or (b) cocaine, defined as self-reported use in any of the three assessments. A continuous quantity–frequency index was calculated to estimate the mothers’ current average daily intake of alcohol (drinks/day).
Youth characteristics
Cigarette smoking
Youth were asked if they had smoked ever, or in the past 3 months, and their age of first smoking. Two dichotomous measures of smoking were defined: (a) lifetime and (b) current smoking.
Behavior problems
Maternal reports of the frequency, range, and type of behavior problems exhibited by their children in the past 3 months were obtained from a 28-item behavior problem assessment
40
. Mothers rated how likely each behavior was of her child (0 = “not true”; 1 = “sometimes true”; 2 = “often true”). The analyses are based on 24 dichotomized (0 vs. 1 and 2) and summed items comprising the Total Behavior Problem scale. Four items measuring dependency were excluded because they were not asked for youth 12 years and older. The scale was standardized within age for the total sample of youth [mean (M) = 100, SD = 15].Delinquency
A modified version of the Elliot Delinquency scale
40
summed child involvement in the eight delinquent behaviors of curfew violations, lying, truancy, stealing, and physically damaging persons or property in the last year, scored 0 = “never” to 3 = “3 or more times.” The delinquency scores ranged from 0 to 24 (∝ = 0.72).Negative mood
A six-item Negative Mood scale averaged youth’s ratings of the frequency with which (1 = “hardly ever” to 3 = “often”) they experienced feelings of sadness, anxiety, stress, boredom, loneliness, and excessive fatigue (∝ = 0.58).
Attitudes toward school and educational expectations
The School Attitudes scale averaged nine 4-point items (1 = “not at all true” to 4 = “very true”) measuring the ability to learn, make friends, and access teacher support; perceptions of teacher competence, educational stimulation and autonomy, school disorder, and safety; and overall school satisfaction (∝ = 0.65). One item assessing educational expectations ranged from 1 = “will not graduate high school” to 5 = “postgraduate education.”
Perceived scholastic competence
The scholastic competence subscale of the Self-Perception Profile
40
summed six items assessing doing well at schoolwork and completing it quickly, accurately, and well, remembering what one learns, and being smart, and was scored from 1 = “low perceived competence” to 4 = “high perceived competence” (∝ = 0.69).Peer group characteristics
Peer pressure to smoke
Youth indicated if they experienced peer pressure to try cigarettes (0 = “no”; 1 = “yes”). No information was available on peer smoking. Peer pressure for smoking is correlated with perceived and self-reported peer use
33
, 35
, 36
.Sociodemographic characteristics
Race/ethnicity
The race and ethnicity of the sample was African-American, Hispanic, and white.
Family structure
For the sample, biological father present = 0, and absent = 1.
Family poverty
Poverty was defined as earning less than the Health and Human Services established guidelines of income level for one’s family size and state of residence (0 = “no”; 1 = “yes”).
Maternal education
The educational level of mothers in the sample ranged from 1 to 20 years.
Youth age and gender
Age was coded in years; gender 1 = “male”, and 2 = “female.”
General overview of statistical strategies
Three analyses were conducted. (a) The prevalence rates of adolescent lifetime, current, and persistent smoking were examined by ethnicity and gender. (b) Means in family socialization practices, youth, and peer characteristics were ascertained by ethnicity and gender, controlling for youth’s age. Significant mean differences were tested by Student–Newman–Keuls posthoc comparison for multiple groups. Sociodemographic characteristics and rates of maternal substance use were examined by ethnicity. (c) Multivariate logistic regressions were estimated to identify the unique factors associated with adolescent lifetime and current smoking for each ethnic group, controlling for other factors.
Since the analyses were based on the total number of paired mother–child dyads, siblings in the same family were nonindependent observations. The inclusion of more than one child per family may lead to observations that are correlated because of siblings’ shared inheritance or experience, and increases the likelihood of biased parameter estimates and Type I errors. To correct for this clustering, both total sample and ethnic-specific familial correlations were estimated using intraclass correlations (ICC) for a subset of seven child-reported variables (adolescent lifetime and current smoking, delinquency, negative mood, positive school attitudes, scholastic competence, and educational expectations). The median ICC for the total sample and for each ethnic group were selected as familial correction factors, and the total and ethnic nominal sample sizes were adjusted to the effective sample sizes to calculate the statistical tests of significance for all analyses.
All analyses used the 1992 child sampling weights provided by the CHRR, which adjusted the unweighted data for: (a) sample attrition of mothers and children between the initial 1979 survey and 1992, (b) sample reduction owing to the exclusion of the military and disadvantaged white oversample, and (c) the overrepresentation of African-American and Hispanic youths
40
.Results
Prevalence of adolescent cigarette smoking, by ethnicity and gender
The rates of cigarette smoking were highest among white and lowest among African-American youths, and only slightly higher among Hispanics than African-Americans (Table 1). White adolescents were almost twice as likely to be lifetime smokers as African-Americans, three times as likely to be current smokers as African-Americans, and twice as likely as Hispanics. White adolescents were also more likely to be persistent smokers (i.e., currently smoking among those who ever smoked) than African-Americans, and started smoking at an earlier age than African-Americans or Hispanics. Although there were no gender differences in the rates of adolescent smoking, males began smoking at an earlier age than females. Although not statistically significant, white males began smoking 2 years earlier than African-American males.
Table 1Prevalence of Cigarette Smoking Among Adolescents Aged 10 Years and Older: 1992 National Longitudinal Survey of Youth (n = 1773)
| Smoking Pattern | Total | African-American | Hispanic | White | Significance |
|---|---|---|---|---|---|
| Lifetime | |||||
| Total (%) | 24.1 | 16.0 | 20.6b,c | 27.3 | 22.24, |
| Males (%) | 23.9 | 18.8 | 19.3 | 26.3 | |
| Females (%) | 24.2 | 13.1 | 21.9 | 28.2 | |
| Last 3 mo | |||||
| Total (%) | 10.0 | 4.4 | 5.9 | 12.5 | 25.81, |
| Males (%) | 9.4 | 4.7 | 6.3 | 11.5 | |
| Females (%) | 10.6 | 4.0 | 5.6 | 13.5 | |
| Total (n) | 1773 | 738 | 422 | 613 | |
| Total male (n) | 895 | 373 | 217 | 305 | |
| Total female (n) | 878 | 205 | 365 | 308 | |
| Persistent | |||||
| Total (%) | 41.6 | 27.2 | 28.7b,c | 45.9 | 10.32, |
| Males (%) | 39.2 | 24.9 | 32.4 | 43.7 | |
| Females (%) | 43.9 | 30.8 | 25.4 | 47.9 | |
| Age of onset | |||||
| Total years | |||||
| x̄ | 10.9 | 11.8 | 11.5b,c | 10.7 | 7.65, |
| SD | 2.7 | 3.2 | 2.6 | 2.5 | (2,335) |
| Males years | |||||
| x̄ | 10.3 | 11.6 | 10.9 | 9.9 | 30.01, |
| SD | 3.0 | 3.2 | 3.2 | 2.8 | (1,335) |
| Females years | |||||
| x̄ | 11.6 | 12.1 | 12.0 | 11.4 | |
| SD | 2.1 | 3.2 | 1.9 | 1.8 | |
| Total smokers (n) | 341 | 107 | 81 | 153 | |
| Male smokers (n) | 176 | 62 | 41 | 73 | |
| Female smokers (n) | 165 | 45 | 40 | 80 |
a Weighted estimates, unweighted n’s.
b Proportions and means with different superscripts are significantly different from each other.
c Proportions and means with different superscripts are significantly different from each other.
d Current smoking among lifetime users.
e Among lifetime users.
f χ2test.
g F test, main effect of ethnicity.
h F test, main effect of gender (M < F).
∗ p ≤ 0.05;
∗∗ p ≤ 0.01;
∗∗∗ p ≤ 0.001.
Distribution of sociodemographic, family, youth, and peer characteristics, by ethnicity
To identify differences in levels of variables to be examined in relation to smoking, 3 × 2 Ethnicity × Gender analyses of covariance were conducted, with youth’s age as the covariate. Sociodemographic characteristics and rates of maternal substance use were examined by ethnicity. Significant ethnic differences appeared for most factors.
Sociodemographic characteristics
Compared with whites, African-American and Hispanic mothers were younger (p < 0.01) and had older children (p < 0.001) and lower family incomes (p < 0.001), and were more likely to live in poverty (p < 0.001). African-American mothers were most likely to live in households with the child’s biological father absent (p < 0.001). Hispanic mothers had the fewest years of formal education (p < 0.001) (data not presented).
Family characteristics
Socialization practices
African-Americans and Hispanics scored lower than whites on the HOME Total scale (Table 2). There were no significant ethnic differences in maternal–child reported closeness and child-reported shared activities and family rules, but African-American youths and their mothers reported lower rates of monitoring than Hispanics and whites. Compared with female youths, males were monitored less (p < 0.001) and reported fewer family rules (p < 0.001). African-American mothers valued the importance of providing religious training to their children more than whites and Hispanics. There were no significant Ethnicity × Gender interactions on family socialization practices.
Table 2Means and Proportions of Family Socialization Practices and Child and Peer Characteristics by Ethnicity, Controlling for Youth’s Age: Adolescents Aged 10 Years and Older, 1992 National Longitudinal Survey of Youth
| Ethnicity | F Tests (df) (Source of Variation) | |||
|---|---|---|---|---|
| African-American | Hispanic | White | ||
| Family socialization practices | ||||
| HOME Total | ||||
| x̄ | 92.42 | 97.10 | 103.55 | 101.03 |
| SD | 15.42 | 15.16 | 13.30 | 2, 1754 |
| Closeness | ||||
| x̄ | 3.24 | 3.31 | 3.23 | 1.25 |
| SD | 0.66 | 0.59 | 0.58 | 2, 1777 |
| Monitoring | ||||
| x̄ | 3.19 | 3.25 | 3.40 | 28.84 |
| SD | 0.62 | 0.63 | 0.49 | 2, 1775 |
| Rules | ||||
| x̄ | 2.84 | 2.85 | 2.91 | 0.90 |
| SD | 0.99 | 1.02 | 0.90 | 2, 1770 |
| Shared activities | ||||
| x̄ | 3.95 | 4.12 | 3.84 | 1.80 |
| SD | 2.16 | 1.99 | 1.85 | 2, 1769 |
| Importance of religious training | ||||
| x̄ | 2.74 | 2.60 | 2.44 | 37.48 |
| SD | 0.49 | 0.55 | 0.64 | 2, 1750 |
| Child characteristics | ||||
| Total behavior problems | ||||
| x̄ | 107.50 | 105.43 | 105.94 | 3.78 |
| SD | 11.70 | 10.72 | 10.19 | 2, 1766 |
| Delinquency | ||||
| x̄ | 4.64 | 4.06 | 3.56 | 11.52 |
| SD | 4.32 | 4.19 | 3.92 | 2, 1729 |
| Negative mood | ||||
| x̄ | 1.86 | 1.83 | 1.88 | 1.35 |
| SD | 0.38 | 0.38 | 0.40 | 2, 1725 |
| Scholastic | ||||
| Positive school attitudes | ||||
| x̄ | 3.10 | 3.17 | 3.08 | 2.66 |
| SD | 0.45 | 0.45 | 0.48 | 2, 1730 |
| Scholastic competence | ||||
| x̄ | 16.89 | 16.56 | 17.39 | 4.27 |
| SD | 4.16 | 4.48 | 4.21 | 2, 1777 |
| Educational expectations | ||||
| x̄ | 3.28 | 3.50 | 3.69 | 18.69 |
| SD | 1.22 | 1.20 | 1.13 | 2, 1707 |
| Peer characteristics | ||||
| Peer pressure to try cigarettes (%) | 12.0 | 10.8 | 12.5 | 0.30 |
| Total n ≥ | 612 | 373 | 567 | 2, 1742 |
a Weighted estimates; unweighted n’s.
b Means with different superscripts are significantly different from each other.
c Means with different superscripts are significantly different from each other.
d Means with different superscripts are significantly different from each other.
e Combined mother–interviewer reports.
f Combined mother–child reports.
g Child reported.
h Mother reported.
∗ p ≤ 0.05;
∗∗ p ≤ 0.01;
∗∗∗ p ≤ 0.001.
Maternal cigarette smoking
Highly significant ethnic differences emerged in rates of lifetime and current maternal smoking, which paralleled the differences observed among the children. The rates were highest among whites compared with African-Americans (p < 0.001) and Hispanics (p < 0.001): 66.3%, 50.1%, and 40.4%, respectively, for lifetime smoking; and 50.7%, 38.6%, and 29.0% for current smoking. While there were no significant ethnic differences in persistence of smoking, there were differences in extensiveness of smoking. Whites smoked significantly (p < 0.001) more cigarettes per day (x̄ = 19.5, SD = 9.2) than African-Americans (x̄ = 14.4, SD = 8.7) and Hispanics (x̄ = 15.1, SD = 9.8).
Maternal use of other substances
There were significant ethnic differences in maternal lifetime use of marijuana and cocaine, and current alcohol use. Lifetime marijuana and cocaine use was higher among white than African-American and Hispanic mothers (73.4%, 67.3%, and 53.0%, respectively, for marijuana; and 23.0%, 18.0%, and 15.5%, for cocaine). (All differences are significant at p < 0.05, except for marijuana use between whites and Hispanics p < 0.001.) White mothers reported higher levels of current alcohol use than African-American (x̄ = 1.12, SD = 1.76 vs. x̄ = 0.87, SD = 1.43; p < 0.05), but not Hispanic mothers (x̄ = 1.16, SD = 2.73). Maternal cigarette smoking was significantly correlated with lifetime cocaine use among African-American (r = 0.20, p < 0.001) and Hispanic (r = 0.18, p < 0.001), but not white mothers (r = 0.07); and with lifetime marijuana use among African-American, Hispanic, and white mothers (r = 0.25, 0.16, and 0.22, respectively; p < 0.001).
Youth characteristics
African-American adolescents had the highest levels of self-reported delinquency and maternal-reported total behavior problems (Table 2). Males exhibited higher levels of self-reported delinquency (p < 0.001) and maternal-reported total behavior problems (p < 0.01) than females. There were no ethnic differences with respect to negative mood; females reported higher negative mood than males (p < 0.001).
African-American adolescents reported lower educational expectations than Hispanics and whites; African-Americans and Hispanics reported lower perceived scholastic competence than whites (Table 2). There were no ethnic differences in positive school attitudes. Females had more positive school attitudes (p < 0.001) and higher educational expectations than males (p < 0.001). There were no Ethnic × Gender interactions in youth characteristics.
Peer characteristics
Across all ethnic and gender groups, adolescents reported similar rates of peer pressure to try cigarettes (Table 2).
Family, youth, peer, and sociodemographic characteristics, and adolescent smoking
Four strategies were used to reduce the number of covariates included in the multivariate logistic regression analyses. (a) Only variables significantly correlated at the zero-order level for any ethnic group with adolescent lifetime, current, or maternal current smoking were candidates for the multivariate analyses. (b) Logistic regressions were estimated to retain those family socialization practices significantly associated with adolescent lifetime and current smoking, controlling for the other practices and sociodemographic covariates. Maternal–child closeness and monitoring were retained as measures of family socialization practices; child-reported family rules and activities, and the HOME Total scale were eliminated. (c) The maternal-reported Total Behavior Problems and youth self-reported Delinquency scales were z-transformed and averaged to create a combined Total Problem Behavior-Deviance score. (d) Maternal-reported current cigarette smoking was selected as an indicator of mothers’ use because it was contemporaneous to child use.
Following these elimination procedures, 18 of 22 variables were retained (Table 3, Table 4). Missing values for model variables ranged from 0.6% to 4.2%, except for family poverty (13.5%) and maternal current smoking (6.4%), for which missing categories were included. Listwise deletion resulted in the loss of 109 cases (14.7%) for African-Americans, 57 cases (13.5%) for Hispanics, and 63 cases (10.3%) for whites.
Table 3Unadjusted Odds Ratios (UORs) and Adjusted Odds Ratios (AORs) From Logistic Regressions Predicting Adolescent Lifetime Smoking, by Ethnicity, From Sociodemographic, Family, Youth, and Peer Characteristics: Adolescents Aged 10 Years and Older, 1992 National Longitudinal Survey of Youth
| Predictors | African-American | Hispanic | White | |||
|---|---|---|---|---|---|---|
| UOR | AOR | UOR | AOR | UOR | AOR | |
| Child’s sex (male = 1, female = 2) | 0.6 | 0.7 | 1.2 | 1.3 | 1.1 | 1.4 |
| Child age (yr) | 1.4 | 1.3 | 1.5 | 1.6 | 1.5 | 1.7 |
| Maternal education (yr) | 0.9 | 1.1 | 1.1 | 1.0 | 0.8 | 1.0 |
| Family poverty (vs. not in poverty), | 1.0 | 0.8 | 0.6 | 0.5 | 1.5 | 1.1 |
| Biological father absent (vs. present) | 1.7 | 1.4 | 1.5 | 1.3 | 2.0 | 1.3 |
| Importance of religious training | 0.8 | 1.1 | 0.7 | 0.6 | 0.6 | 0.7 |
| Maternal ever use marijuana (vs. never used) | 1.4 | 1.1 | 2.0 | 2.4 | 2.2 | 1.1 |
| Maternal ever use cocaine (vs. never used) | 0.9 | 0.5 | 1.4 | 0.9 | 2.2 | 2.3 |
| Maternal current alcohol use | 1.0 | 1.1 | 1.0 | 1.0 | 1.1 | 0.9 |
| Maternal current smoking (vs. does not smoke), | 1.1 | 1.0 | 1.1 | 0.7 | 2.3 | 1.0 |
| Positive parenting | 0.5 | 0.7 | 0.7 | 1.5 | 0.6 | 1.3 |
| Child problem behaviors/deviance | 2.5 | 2.0 | 2.2 | 1.8 | 4.3 | 4.4 |
| Child negative mood | 1.1 | 0.8 | 1.7 | 0.8 | 2.3 | 0.6 |
| Child positive school attitudes | 0.7 | 1.1 | 0.3 | 0.3 | 0.3 | 0.5 |
| Child perceived scholastic competence | 1.0 | 1.0 | 1.0 | 0.9 | 0.9 | 1.0 |
| Child educational expectations | 0.8 | 1.0 | 1.0 | 1.4 | 0.8 | 1.1 |
| Perceived peer pressure to smoke (vs. no pressure) | 3.1 | 1.4 | 8.8 | 6.1 | 13.2 | 20.4 |
| Constant | −6.37 | −2.78 | −4.14 | |||
| Model χ2 | 52.36 | 43.26 | 514.16 | |||
| df | 19 | 19 | 19 | |||
| n | 629 | 365 | 550 | |||
a Exponentiated unstandardized weighted coefficients, unweighted n’s.
b Category for missing included.
c Mother reported.
d Child reported.
e Combined mother–child reports.
∗ p ≤ 0.05;
∗∗ p ≤ 0.01;
∗∗∗ p ≤ 0.001.
Table 4Unadjusted Odds Ratios (UORs) and Adjusted Odds Ratios (AORs) From Logistic Regressions Predicting Adolescent Past 3 Months Smoking, by Ethnicity, From Sociodemographic, Family, Youth, and Peer Characteristics: Adolescents Aged 10 Years and Older, 1992 National Longitudinal Survey of Youth
| Predictors | African-American | Hispanic | White | |||
|---|---|---|---|---|---|---|
| UOR | AOR | UOR | AOR | UOR | AOR | |
| Child’s sex (1 = male, 2 = female) | 0.9 | 0.9 | 0.9 | 0.8 | 1.2 | 1.5 |
| Child age (yr) | 1.4 | 1.2 | 1.5 | 1.6 | 1.6 | 1.5 |
| Maternal education (yr) | 0.7 | 0.8 | 1.1 | 0.9 | 0.8 | 0.9 |
| Family poverty (vs. not in poverty), | 2.3 | 1.0 | 0.6 | 0.5 | 0.9 | 0.6 |
| Biological father absent (vs. present) | 2.0 | 1.0 | 1.7 | 1.4 | 2.1 | 1.2 |
| Importance of religious training | 0.9 | 1.0 | 0.6 | 0.4 | 0.7 | 0.9 |
| Maternal ever use marijuana (vs. never used) | 1.0 | 0.5 | 2.9 | 6.5 | 2.3 | 1.2 |
| Maternal ever use cocaine (vs. never used) | 1.0 | 1.0 | 2.6 | 2.2 | 1.8 | 1.4 |
| Maternal current alcohol use | 0.9 | 0.8 | 1.1 | 1.0 | 1.0 | 0.9 |
| Maternal current smoking (vs. does not smoke), | 1.3 | 1.5 | 0.4 | 0.3 | 3.1 | 1.8 |
| Positive parenting | 0.5 | 0.8 | 0.6 | 1.2 | 0.7 | 1.2 |
| Child problem behaviors/deviance | 3.7 | 2.3 | 2.8 | 2.0 | 4.4 | 3.5 |
| Child negative mood | 2.3 | 1.8 | 3.2 | 2.8 | 2.5 | 0.8 |
| Child positive school attitudes | 0.6 | 0.9 | 0.7 | 1.7 | 0.3 | 0.6 |
| Child perceived scholastic competence | 0.9 | 1.0 | 1.0 | 1.0 | 0.9 | 1.0 |
| Child educational expectations | 0.6 | 0.9 | 1.0 | 1.3 | 0.7 | 0.9 |
| Perceived peer pressure to smoke (vs. no pressure) | 3.3 | 1.5 | 9.9 | 4.8 | 4.3 | 3.5 |
| Constant | −5.00 | −17.41 | −5.85 | |||
| Model χ2 | 27.74 | 22.69 | 255.94 | |||
| df | 19 | 19 | 19 | |||
| n | 629 | 365 | 550 | |||
a Exponentiated unstandardized weighted coefficients, unweighted n’s.
b Category for missing included.
c Mother reported.
d Child reported.
e Combined mother–child reports.
∗ p ≤ 0.05;
∗∗ p ≤ 0.01;
∗∗∗ p ≤ 0.001.
Estimation problems were encountered. Among whites, maternal–child closeness and monitoring were negatively associated with adolescent lifetime and current smoking at the bivariate level, but became positively associated in the multivariate models. To minimize this problem, a combined maternal–child closeness and monitoring variable was constructed. Each scale was z-transformed and averaged, and all models were reestimated with the new positive parenting measure. This strategy, however, was not completely successful (see below).
Overall, a larger number of factors were associated with lifetime (Table 3) and current (Table 4) smoking among white than African-American and Hispanic adolescents, and remained statistically significant with controls for other correlates. Youth’s age had a statistically significant effect on lifetime smoking for all three ethnic groups and on current smoking for whites only.
Maternal self-reported smoking had a statistically significant unique effect on current smoking for white adolescents only; the effect on lifetime smoking observed at the univariate level was not retained when controlling for other covariates. The distribution of adolescent smoking according to maternal smoking is displayed in Table 5. White adolescents were twice as likely to report having ever smoked if their mothers smoked lifetime or currently, four times as likely to be current smokers if their mother ever smoked, and three times as likely to be current smokers if their mother also currently smoked. The effect was present among both sons and daughters, but stronger among daughters than sons (data not presented). Among African-Americans and Hispanics, however, there was no association between maternal and child smoking. The rates of smoking among these youths, whether lifetime or past year, were similar to those of white children whose mothers did not smoke.
Table 5Adolescent Cigarette Smoking by Maternal Lifetime and Current Cigarette Smoking and Ethnicity: Adolescents Aged 10 Years and Older: 1992 National Longitudinal Survey of Youth
| Adolescents Who Smoke Cigarettes (%) | Mother Lifetime Smoking | Mother Current Smoking | ||
|---|---|---|---|---|
| No (%) | Yes (%) | No (%) | Yes (%) | |
| African-American | ||||
| Lifetime | 15.2 | 16.5 | 14.9 | 16.1 |
| Past 3 mo | 4.2 | 4.3 | 3.8 | 5.0 |
| Total n | 368 | 352 | 436 | 253 |
| Hispanic | ||||
| Lifetime | 18.0 | 25.1, | 20.0 | 22.1, |
| Past 3 mo | 6.0 | 6.4 | 7.8 | 3.1 |
| Total n | 252 | 150 | 286 | 108 |
| White | ||||
| Lifetime | 15.8 | 33.7c, | 19.8 | 36.7c, |
| Past 3 mo | 4.0 | 16.8c, | 6.8 | 18.5c, |
| Total n | 210 | 390 | 292 | 284 |
a Weighted estimates, unweighted n’s.
b Means with different superscripts are significantly different from each other.
c Means with different superscripts are significantly different from each other.
∗ p ≤ 0.001.
Although for whites maternal lifetime marijuana and cocaine use were strongly associated with adolescent lifetime smoking and maternal lifetime marijuana use was associated with adolescent current smoking, only maternal lifetime cocaine use retained a statistically significant effect on adolescent lifetime smoking in the multivariate model. Even when maternal cocaine, marijuana, and alcohol use were excluded from the multivariate models, maternal current cigarette smoking did not retain a statistically significant effect on lifetime smoking among whites (data not presented).
Positive parenting, indexed by maternal–child closeness and monitoring, was associated with reduced lifetime smoking among African-American youths only; the univariate effect observed on current smoking was no longer significant in the multivariate model. Among whites, positive parenting, which was negatively associated with lifetime and current smoking at the univariate level, became positively associated with lifetime smoking in the multivariate models despite the redefinition of this variable. Maternal perceived importance of religious training in her child’s life decreased the odds of adolescent lifetime smoking only among whites.
The combined maternal and youth reported measure of child deviance was positively and significantly associated with lifetime smoking among African-American and white adolescents, and current smoking among whites only. The adjusted odds for lifetime smoking among whites were twice those of African-Americans, and significantly different.
Controlling for all other covariates, adolescent negative mood was not associated with smoking. Positive attitudes toward school were associated with reduced lifetime and current smoking among whites only. Across all ethnic groups, scholastic self-esteem and educational expectations were unrelated to either lifetime or current adolescent smoking, with control for other covariates. Hierarchical regressions indicated that youth’s age and behavior problems might partially account for the elimination of the associations of negative mood and low scholastic competence in the multivariate models (data not presented).
Perceived peer pressure to try cigarettes was positively and strongly associated with lifetime smoking among whites and Hispanics, and with current smoking only among whites. By contrast, at the univariate level, among all adolescents, the prevalence of lifetime and current smoking was much higher among youths who reported peer pressure to try cigarettes than among those who did not. Lifetime smoking was four times higher among whites (77.1% vs. 20.3%) and Hispanics (62.2% vs. 15.8%) who reported peer pressure to smoke, but only twice as high among African-Americans (33.2% vs. 14.9%). Current smoking was three times higher among whites (31.9% vs. 9.9%) and African-Americans (10.8% vs. 3.6%), and eight times higher among Hispanics (26.5% vs. 3.5%), if adolescents experienced peer pressure to smoke.
Discussion
Ethnic differences in the rates and correlates of adolescent cigarette smoking reflect eight major patterns. (a) African-American youths report lower rates of lifetime and current cigarette smoking than whites and Hispanics. (b) Although African-Americans and Hispanics experience a disproportionate number of assumed risk factors compared with whites, the majority of these factors are associated with cigarette smoking only among whites both at the univariate and multivariate levels. (c) Controlling for other factors, the correlates of adolescent lifetime and current cigarette smoking are ethnic-specific, except for child age. (d) Behavior problems are positively associated with lifetime smoking among African-Americans and white youths, but not Hispanics. (e) Maternal smoking, as reported by mothers, is a significant predictor of the youth’s current smoking for whites only. (f) Peer pressure to smoke, as perceived by adolescents, is a significant correlate of lifetime and current smoking for whites, and of lifetime smoking for Hispanics. Neither maternal smoking nor peer pressure is significant for African-Americans. (g) High quality of parent–child interactions reduces smoking among African-Americans only. (h) Attitudes reflecting a positive orientation toward social institutions, such as religiosity and positive school attitudes, are more important correlates of not smoking for white than minority adolescents. This cross-sectional examination of correlates of adolescent smoking confounds causes and consequences. Whether the same patterns of common and ethnic-specific factors predate the initiation and continued use of cigarettes is not known. Selected results are discussed further.
Consistent with the literature
1
, 3
, 12
, U.S. Department of Health and Human Services. Preventing tobacco use among young people: A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994.
15
, 41
, the prevalence rates of cigarette smoking were lower among African-American than Hispanic and white adolescents. The ethnic differences increased with recency of smoking: Whites were almost twice as likely to have ever smoked, but three times as likely to be current smokers as African-Americans. White adolescents, especially males, reported an earlier age of onset of smoking than African-Americans and Hispanics. Evidence from concurrent biological assays indicate that the lower smoking rates reported by minorities are probably not due to reporting bias 42
.Minority adolescents, particularly African-Americans, scored more negatively than whites on many factors potentially associated with cigarette smoking, such as higher levels of behavior problems and delinquency, lower educational expectations and perceived scholastic competence, lower maternal monitoring, lower family income, and higher absence of biological father. However, white mothers were more likely to smoke cigarettes and to use other substances, and less likely to value religious training for their children than Hispanic and, especially, African-American mothers. Despite the fact that African-American adolescents, and to a lesser extent Hispanics, experienced disproportionate risks, many more significant associations of these risk factors with cigarette smoking were observed for white than African-American and Hispanic adolescents
43
. This is consonant with analyses conducted by Giovino 41
on high school seniors from Monitoring the Future 1
, 5
showing that over a 20-year period, there was consistently greater differentiation of smoking rates by individual attributes among white than African-American adolescents.Maternal smoking was an important factor only for white adolescents, and was somewhat stronger among daughters than sons
14
, 44
. Even with control for maternal use of other substances, such as alcohol, marijuana, and cocaine, maternal cigarette smoking retained a unique and exclusive effect on current smoking by white adolescents. Compared with African-American and Hispanic mothers, white mothers reported more extensive smoking. These ethnic differences parallel those observed among mothers in the same historical period in another national sample, the National Household Survey of Drug Abuse (NHSDA). We identified women aged 27–35 years with children in the household in the 1992 NHSDA to match mothers in the NLSY sample. The rates of lifetime smoking were higher among whites (80%) than African-Americans (67%) and Hispanics (48%), although the rates of past month smoking were the same among whites and African-Americans (36%) and lower among Hispanics (19%). Differences in methods of data collection (self-administration vs. oral answers) may partially account for the differences in current prevalence rates across the two studies 45
. As in the NLSY, white mothers in the NHSDA were more likely than African-Americans or Hispanics ever to have smoked daily and to be heavy smokers. Sixty percent of whites reported smoking at least a pack of cigarettes per day in the last 30 days, compared with 38% of African-American and 24% of Hispanic mothers; 6% of whites versus 1% and 1% of the other two groups reported smoking at least two packs of cigarettes a day. The higher rates of smoking by white than African-American mothers and the stronger influence of white than African-American mothers on smoking by their children could partially explain ethnic differences in rates of smoking by adolescents.While African-American, Hispanic, and white adolescents reported similar rates of peer pressure to smoke, with controls for other covariates the peer variable was consistently significant only for white youths, and only for lifetime smoking among Hispanics. These findings are consistent with other multiethnic samples that document an influence of perceived peer pressure and peer substance use on the smoking, drinking, and use of illicit substances by white and Hispanic adolescents, but not African-Americans
2
, 3
, 15
, 23
, 25
, 26
, 27
, 37
, 38
. Why African-American youths should be less vulnerable than whites or Hispanics to the negative impact of peer influence on cigarette smoking and substance use remains to be clarified. The lower importance of peer approval reported by African-American adolescents than whites 46
may reduce their susceptibility to negative peer influences. Adolescent problem behaviors emerged as a common correlate of lifetime cigarette smoking among African-Americans and whites, but not Hispanics 29
, 30
. This association underscores that adolescent tobacco use co-occurs with behavior problems and delinquency across ethnically diverse youths 47
.Several unexpected findings emerged. Adolescents’ positive attitudes toward school were associated with lower prevalence of lifetime and current smoking among whites, and lower prevalence of lifetime smoking among Hispanics, but not African-Americans. However, measures of academic performance, such as school grades, which have been found to be associated with cigarette smoking and substance use among African-Americans
24
, were not ascertained. The finding that maternal importance of religious training decreased smoking only among whites was also unexpected, given the association between religiosity and decreased substance use reported for African-Americans 2
.Differences in the role of the family on smoking by white and African-American adolescents: interpretations and research implications
The ethnic-specific impact of maternal smoking on adolescent cigarette use observed in this study gains additional power from the fact that maternal smoking was based on mothers’ self-reports rather than adolescents’ perceptions. Compared with African-Americans, white adolescents appear to be differentially exposed and differentially influenced by maternal smoking. Potential mechanisms include ethnic differences in children’s modeling of maternal cigarette use, variations in parental norms proscribing smoking, and variations in effective parenting. Based on focus groups and a telephone survey of white and African-American smoking parents, Clark and colleagues
18
found that African-American parents held much stronger negative norms against smoking by their children and used more effective disciplinary techniques for curtailing and preventing their children’s smoking than white parents. Consonant with these findings, other investigators have reported that African-American parents hold much stronger negative norms against cigarette smoking 19
, 20
and alcohol use than white parents 4
.Similarly, we found that positive parenting, indexed by an aggregate measure of maternal–child closeness and monitoring, significantly reduced the likelihood of lifetime smoking only among African-Americans. This finding is consistent with the documentation of the importance of parenting factors, particularly parent–child closeness, support, and monitoring, in reducing cigarette, alcohol, and general substance use among African-American adolescents
2
, 4
, 16
, 17
. The anomalous finding observed among whites that positive parenting had a positive effect in the multivariate analyses remains unexplained. Although African-American mothers were less likely than white mothers to monitor their children, they were as close and as likely to provide rules. The absence of ethnic differences in levels of maternal closeness is congruous with other general population samples regarding parenting practices in African-American and white families 2
.Social contexts favorable to cigarette smoking, particularly maternal use of cigarettes and perceived peer pressure to smoke, appear to constitute greater risk factors for smoking among white than minority adolescents, especially African-Americans. Quality of the parent–child interaction, however, seems to be a protective factor, especially for African-American adolescents. Future research on ethnic differences in the role of the family and peers on adolescent smoking would need to consider adolescent smoking within the broader context of adolescent development. Such research would need to understand the differential impact on adolescent smoking of parental and peer role modeling, differential parenting and socialization practices related to smoking and to other domains of child behavior, and the use of substances other than tobacco by family members.
The findings regarding maternal effects on adolescent smoking have important policy implications, and highlight that the role of the family on teen smoking deserves greater emphasis in smoking prevention efforts. These efforts need to be culturally sensitive and take into account the fact that different aspects of the family context may be important for children of different ethnicities. Parental behaviors are especially important in white families, where antismoking norms need to be strengthened. Prevention efforts need to target the family as one of the important contextual components of relevance to smoking by adolescents, in addition to peers, the school, the community, the marketplace, and the media.
Acknowledgements
Work on this manuscript was partially supported by research grant U48/CCU209663-RFS from the Centers for Disease Control and Research Scientist award DA00081 to Denise Kandel from the National Institute on Drug Abuse. Partial support for computer costs was provided by Mental Health Clinical Research Center Grant MH30906 from NIMH to the New York State Psychiatric Institute.
References
- Racial/ethnic differences in smoking, drinking, and illicit drug use among American high school seniors, 1976–1989.Am J Public Health. 1991; 81: 372-377
- Family influences on alcohol abuse and other problem behaviors among black and white adolescents in a general population sample.J Res Adolesc. 1994; 4: 183-201
- Cultural diversity in the predictors of adolescent cigarette smoking.J Behav Med. 1994; 17: 331-346
- Disentangling the effects of parental drinking, family management, and parental alcohol norms on current drinking by black and white adolescents.J Res Adolesc. 1994; 4: 203-227
Johnston LD. Cigarette Smoking Continues to Rise Among American Teenagers in 1996. Ann Arbor, MI: University of Michigan, press release, December 19, 1996.
- National Household Survey of Drug Abuse. Substance Abuse and Mental Health Services Administration, Rockville, MD1996
- Trends in cigarette smoking among U.S. adolescents, 1974–1991.Am J Public Health. 1995; 85: 34-40
CDC. Tobacco use and usual source of cigarettes among high school students—United States, 1995. MMWR 1996;45:413–8.
- Drug use etiology among ethnic minority adolescents.in: Botvin G.J Schinke S Orlandi M Drug Use Prevention With Multiethnic Youth. Sage, California1995: 59-80
- Why children start smoking cigarettes.B J Addict. 1992; 87: 1711-1724
- Social psychological contributions to the understanding and prevention of adolescent cigarette smoking.Pers Soc Psychol Bull. 1990; 16: 133-151
U.S. Department of Health and Human Services. Preventing tobacco use among young people: A report of the Surgeon General. Atlanta, GA: U.S. Department of Health and Human Services, Public Health Service, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, Office on Smoking and Health, 1994.
- Parent monitoring and the incidence of drug sampling in urban elementary school children.Am J Epidemiol. 1995; 141: 25-31
- The contributions of mothers and fathers to the intergenerational transmission of cigarette smoking in adolescence.J Res Adolesc. 1995; 5: 225-252
- Are the correlates of cigarette smoking initiation different for black and white adolescents?.Am J Public Health. 1990; 81: 854-857
- Ethnic differences in family factors related to early drug initiation.J Stud Alcohol. 1992; 53: 208-217
Salem DA, Zimmerman MA, Notaro PC. Effects of family structure, family process, and father involvement on psychosocial outcomes among African-American adolescents. Presented at the biennial meeting of the Society for Research in Child Development, Washington, DC, April 1997.
Clark P, Scarisbrick-Hauser A, Gautam SP, et al. Anti-tobacco socialization in the homes of African-American and white, and smoking and nonsmoking parents. J Adolesc Health.
- Health behaviors in minority families.Family and Community Health. 1990; 13: 35-43
Zbikowski SM, Robinson LA, Klesges RC. Parental attitudes toward smoking: Ethnic and gender differences. Presented at the Third Annual Meeting of the Society for Research on Nicotine and Tobacco, Nashville, TN, June 1997.
- A multivariate model of determinants of cigarette smoking among adolescents.J Health Soc Behav. 1987; 28: 273-289
- Parent substance use as a predictor of adolescent use.Annal Behav Med. 1996; 18: 157-164
- The influence of friends’ and parental smoking on adolescent smoking behavior.J Appl Soc Psychol. 1995; 25: 2018-2047
- Predictors of cigarette smoking among inner-city minority youth.Dev Behav Pediatr. 1994; 15: 67-73
- Predictors of risk for different stages of adolescent smoking in a biracial sample.J Consulting and Clinical Psychol. 1997; 65: 653-662
Sussman S, Dent CW, Flay BR, et al. Psychosocial predictors of cigarette smoking onset by white, black, Hispanic, and Asian adolescents in Southern California. MMWR 1987;36:11–6S.
- Ethnic and gender differences in risk for early adolescent substance use.J Youth Adolesc. 1994; 23: 195-213
- The relationship of alcohol, tobacco, marijuana, and other illegal drug use to delinquency among Mexican-American, black, and white adolescent males.Adolescence. 1990; 25: 171-181
- Sex differences in antecedents of substance use among adolescents.J Soc Issues. 1982; 38: 25-42
- Childhood conduct problems, attention deficit behaviors, and adolescent alcohol, tobacco, and illicit drug use.J Abnorm Child Psychol. 1995; 23: 281-302
- Epidemiology of depressive mood in adolescents.Arch Gen Psychiatry. 1982; 39: 1205-1212
- Moderators of peer social influence in adolescent smoking.Pers Soc Psychol Bull. 1992; 18: 163-172
- Differential influence of parental smoking and friends’ smoking on adolescent initiation and escalation of smoking.J Health Soc Behav. 1994; 35: 248-265
- Processes of adolescent socialization by parents and peers.Int J Addict. 1987; 22: 319-342
- Peer influence in adolescent cigarette smoking.Addict Behav. 1990; 15: 247-255
- Smoking onset among teens.Addict Behav. 1985; 10: 1-13
- Sociobehavioral influences on smoking initiation of Hispanic adolescents.J Adolesc Health. 1997; 20: 46-50
- Close friend and group influence on adolescent cigarette smoking and alcohol use.Dev Psychol. 1997; 33: 834-844
- Explaining racial/ethnic differences in adolescent drug use.Soc Prob. 1991; 38: 333-357
- NLSY Child Handbook. Center for Human Resource Research, Columbus, Ohio1993
Giovino GA. Epidemiology of tobacco use among young people in the United States. Presented at the meeting of the Society for Research on Nicotine and Tobacco, Washington, DC, March 1996.
- The validity of self-reports of smoking.Am J Public Health. 1997; 87: 56-61
- Predictors of smoking onset in adolescent African American boys and girls.J Health Education. 1995; 26: 85-91
- Changes in peer and parent influence during adolescence.Dev Psychol. 1986; 22: 327-334
National Household Survey of Drug Abuse. Population Estimates 1994. Rockville, MD: Substance Abuse and Mental Health Services Administration, 1995.
- The family and peer relations of black adolescents.J Marr Fam. 1993; 55: 277-287
- Relationship between drug use and other problem behaviors in urban adolescents.J Consult Clin Psychol. 1992; 60: 705-712
Article info
Publication history
Accepted:
November 23,
1997
Identification
Copyright
© 1998 Society for Adolescent Medicine. Published by Elsevier Inc.
User license
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0) | How you can reuse 
Elsevier's open access license policy
Creative Commons Attribution – NonCommercial – NoDerivs (CC BY-NC-ND 4.0)
Permitted
For non-commercial purposes:
- Read, print & download
- Redistribute or republish the final article
- Text & data mine
- Translate the article (private use only, not for distribution)
- Reuse portions or extracts from the article in other works
Not Permitted
- Sell or re-use for commercial purposes
- Distribute translations or adaptations of the article
Elsevier's open access license policy
