Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2021 May 1.
Published in final edited form as: Prev Sci. 2020 May;21(4):508–518. doi: 10.1007/s11121-019-01072-8

Neighborhood structural factors and proximal risk for youth substance use

Christopher Cambron 1,2, Rick Kosterman 3, Isaac C Rhew 4, Richard F Catalano 3, Katarina Guttmannova 4, J David Hawkins 3
PMCID: PMC7166144  NIHMSID: NIHMS1546978  PMID: 31853720

Abstract

This study examined associations of neighborhood structural factors (census-based measures socioeconomic disadvantage and residential stability), self-reported measures of general and substance use-specific risk factors across neighborhood, school, peer, and family domains, and sociodemographic factors with substance use among 9th grade students. Data drawn from the Seattle Social Development Project, a theory-driven longitudinal study originating in Seattle, WA, were used to estimate associations between risk factors and past month cigarette smoking, binge drinking, marijuana use, and polysubstance use among students (N=766). Results of logistic regression models adjusting for neighborhood clustering and including all domains of risk factors simultaneously indicated that neighborhood socioeconomic disadvantage was associated with a significantly higher likelihood of cigarette smoking, binge drinking, and polysubstance use, but not marijuana use. In fully controlled models, substance use-specific risk factors across neighborhood, school, peer, and family domains were also associated with increased likelihood of substance use and results differed by the outcome considered. Results highlight substance-specific risk factors as an intervention target for reducing youth substance use and suggest that further research is needed examining mechanisms linking neighborhood socioeconomic disadvantage and youth substance use.

Keywords: neighborhood context, youth substance use, risk factors for substance use

Introduction

Preventing substance use can improve both short- and long-term health and well-being for youth (Catalano et al., 2012; Office of Surgeon General, 2016). Youth substance use is linked to both immediately preventable harms such as unintentional injury and motor vehicle crashes as well as later life problems including substance use disorders, academic failure, interpersonal violence, and mental health issues (Office of Surgeon General, 2016). Additionally, substance use prevention among youth has been identified as an important feature of improving health equity across the life course for youth from socioeconomically disadvantaged backgrounds (Viner et al., 2012). Preventing substance use requires acting upon the range of modifiable factors that place youths at increased risk for substance use (Catalano et al., 2012; Hawkins, Catalano, & Miller, 1992). While researchers have identified key risk factors for substance use across families, peer groups, and school environments (Jenson & Bender, 2014), the role that neighborhood structural factors play in youth substance use remains an open and important question (Gardner, Barajas, & Brooks-Gunn, 2010; Jackson, Denny, & Ameratunga, 2014).

Neighborhood Context and Youth Substance Use

Social ecological theories of development posit that human behavior is best understood situated within multiple, overlapping social contexts including neighborhoods, schools, peer groups, and families (Bronfenbrenner, 1977; Catalano & Hawkins, 1996). A broad range of scholarship has employed social ecological theories to examine the potential impact of neighborhoods on health (Diez Roux & Mair, 2010). Social disorganization theorists have long hypothesized that neighborhood structural features such as high poverty, low educational attainment, and lack of stability among residents lead to a higher concentration of residents engaging in crime and other deviant behaviors (Sampson, Morenoff, & Gannon-Rowley, 2002). The impact of neighborhood structural factors has been theorized to be transmitted through neighborhood social processes (Sampson et al., 2002). Systematic reviews have indicated that living in more socioeconomically disadvantaged neighborhoods is associated with poorer youth mental and physical health and is an important driver of health inequities (Diez Roux & Mair, 2010). Yet, empirical studies examining the role of neighborhood structural factors in youth cigarette smoking, alcohol use, or marijuana use have largely reported inconsistent findings. Systematic reviews have reported neighborhood socioeconomics to be associated with both increased or decreased youth substance use, while many studies have found no association (Gardner et al., 2010; Hanson & Chen, 2007; Jackson et al., 2014; Karriker-Jaffe, 2011). Neighborhood scholars, suggesting sources for these inconsistencies, have noted that insufficient control variables for family-level characteristics may over- or underestimate the effect of neighborhood socioeconomic factors on health behaviors. For instance, studies including area-level measures of educational attainment or single-parent families but failing to model parental education and family structure among study participants run the risk of conflating the effects of neighborhood and family socioeconomic factors (Wodtke, Elwert, & Harding, 2016). Others have suggested that modeling neighborhood socioeconomics, but not explicitly considering the social processes or the experiences of youth within their neighborhoods, may also overestimate the impact of neighborhoods on youth behavior (Sampson et al., 2002). Yet, studies considering both neighborhood socioeconomics and social processes have also yielded inconsistent findings in predicting youth substance use (Duncan, Duncan, & Strycker, 2002; Fagan, Wright, & Pinchevsky, 2015; Jackson, Denny, Sheridan, Zhao, & Ameratunga, 2016).

Proximal Risk Factors for Youth Substance Use

Over the past 30 years, prevention scientists have identified factors that increase risk of substance use located in proximal contexts of youth life including schools, peer groups, and families (Hawkins et al., 1992; Jenson & Bender, 2014). More recently, researchers have highlighted the utility of differentiating general and substance use-specific risk factors within each of these contexts (Bailey, Hill, Meacham, Young, & Hawkins, 2011). For instance, while meta-analyses have highlighted poor family management practices or weak parent-child bonds as risk factors for substance use (Ryan, Jorm, & Lubman, 2010), studies have also shown unique risks emanating from exposure to substance use by family members or permissive substance use norms (Epstein, Hill, Bailey, & Hawkins, 2013; Hill, Hawkins, Catalano, Abbott, & Guo, 2005). Similar findings have been demonstrated across peer and school domains whereby the general features of the environment (i.e. deviant peer groups, weak bonds to school) and substance specific features of the environment (i.e. alcohol use among close friends, high percentage of students in a school using alcohol) uniquely predict substance using behavior (Kosterman, Hawkins, Guo, Catalano, & Abbott, 2000; Lee et al., 2012).

Social processes within neighborhoods can also be categorized as general or substance specific (Fagan et al., 2015) and represent a more proximal measure of risk in relation to neighborhood structural features (Furr-Holden et al., 2015). We located only two studies that have modeled general and substance specific neighborhood factors in conjunction with neighborhood structural features. Duncan and colleagues (2002) specified a multilevel path model from neighborhood structural factors (census measure of neighborhood poverty) to social cohesion (self-report general neighborhood environment) to youth drug and alcohol problems in the neighborhood (self-report substance specific neighborhood environment) ultimately predicting police records of youth drug and alcohol arrests. While each path significantly predicted the next in the sequence, these analyses did not consider direct effects of objective or general neighborhood features on youth substance use or control for family-level SES. Fagan and colleagues (2015) examined an extensive set of neighborhood structural features, general and substance-specific social processes, and family-level controls in multilevel models predicting youth cigarette, alcohol, and marijuana use. Results indicated little evidence for an impact of neighborhood structural factors or neighborhood social processes on youth substance use leading the authors conclude that social ecological or disorganization theories may not be applicable to youth substance use.

Researchers have also examined differences in youth substance use by a range of sociodemographic factors. Results from studies employing nationally representative samples have found that African American youth engage is less smoking and alcohol use compared to Whites (Blum et al., 2000; Hoffmann & Johnson, 1998). These same studies have also shown that youths in single-parent households are at increased risk for substance use after controlling for other race/ethnicity and family socioeconomic status (SES). Studies employing multi-ethnic, longitudinal samples have similarly observed less smoking and alcohol use among Asian American and African American youth when compared to Whites after controlling for differences in family SES, family structure, and gender (Hill et al., 2005). While epidemiological surveillance and systematic reviews have reported higher rates of smoking among lower SES adults, it remains less clear if this relationship extends to youth (Hanson & Chen, 2007; Hiscock, Bauld, Amos, Fidler, & Munafò, 2012). Systematic reviews reporting associations between family SES and youth marijuana or alcohol use have identified positive, negative, and null effects of lower family SES on youth use (Hanson & Chen, 2007).

The Current Study

The goal of this study was to examine associations between neighborhood structural factors (e.g., socioeconomic disadvantage and residential stability) and youth substance use. Previous research has not reached a consensus on the existence or direction of associations between neighborhood structural factors and youth substance use (Fagan et al., 2015; Karriker-Jaffe, 2011). These inconsistent findings paired with the consistency of findings demonstrating the importance of school, peer, and family contexts for youth substance use warrant concurrent examination of neighborhood structural factors in conjunction with proximal risk factors for youth substance use. First, associations between neighborhood structural factors and youth cigarette smoking, binge drinking, marijuana use, and polysubstance use were assessed after controlling for a broad range of sociodemographic factors. Second, we tested a series of domain-specific models examining associations between self-reported general and substance use-specific risk factors across neighborhood, school, peer, and family domains and youth cigarette smoking, binge drinking, marijuana use, and polysubstance use. Based on prior research, we expected that both general and substance specific risk factors within schools, peer groups, and families would be associated with increased likelihood of substance use among youths in the 9th grade. Finally, we estimated fully-controlled models to examine if any observed associations between neighborhood structural factors and each substance use outcome persisted after the inclusion of self-reported risk factors. These models included neighborhood structural factors, all self-reported general and substance use-specific factors across neighborhood, school, peer, and family domains, and sociodemographic factors. Figure 1 provides a visual representation of the measures employed in fully-controlled models organized from most distal (i.e., neighborhood structural factors) to most proximal (i.e., individual-level characteristics). [Figure 1]

Figure 1.

Figure 1.

Open in a new tab

Neighborhood, school, peer, family, and individual domains of general and substance use specific factors associated with youth substance use. Domains are shown from most distal (neighborhood structural factors) at the top to most proximal (individual level factors) at the bottom and examples of constructs measured are provided within each domain.

Methods

Sample

Data were drawn from the Seattle Social Development Project (SSDP), a longitudinal, theory-driven study originating in 18 Seattle elementary schools and over-representing schools serving high crime neighborhoods. SSDP conducted paper and pencil interviews in 1985 with 808 students in the 5th grade when most participants were 10 years old (M = 10.3, SD = .52). Of the 1,053 5th grade students invited into the study, 77% of parents consented to youth participation and 97% of youth remained in the sample at grade 9. A detailed history of the SSDP sample can be found elsewhere (Hill, Woodward, Woelfel, Hawkins, & Green, 2016). The current study employed data from 766 participants with available home address data at grade 9.

Measures

Descriptive statistics are presented in Table 1. All constructs are measured at the 9th grade except for early initiation of substance use measured at the 5th grade. Youth enrolled in the study responded to questions about their experiences with neighborhoods, schools, peers, and families while parents reported on their own substance use, family substance use norms, and family socioeconomic indicators. Measures of substance use and risk factors across domains are primarily drawn from the social development model (Catalano & Hawkins, 1996) and have been employed in prior studies with the SSDP sample (Catalano, Kosterman, Hawkins, Newcomb, & Abbott, 1996; Hill et al., 2005; Kosterman et al., 2000; Lonczak et al., 2001). Similar measures have been commonly utilized across a range of studies examining risk factors for youth substance use (Arthur, Hawkins, Pollard, Catalano, & Baglioni Jr, 2002; Hawkins et al., 1992; Ryan et al., 2010; Sampson et al., 2002). All items were coded such that higher scores theoretically increased risk for substance use where applicable and count variables where log transformed to reduce skew where necessary (Lonczak et al., 2001).

Table 1.

Sample descriptive statistics

Variable M SD Min Max Variable M (%) SD Min Max
Neighborhood Structural (census) Past Month Substance Use
Socioeconomic Disadvantage .00 1.00 −1.40 5.67 Cigarette Smoking 18% 0 1
Residential Stability .00 1.00 −2.81 2.24 Binge Drinking 8% 0 1
Neighborhood Experience Marijuana Use 9% 0 1
Negative NH General 1.83 .58 1.00 4.00 Polysubstance Use .33 .68 0 3
NH Alcohol Env. 1.23 1.02 .00 5.53 1 substance only 14%
NH Drug Env. .62 .79 .00 4.86 2 substances 14%
NH Polysubstance Env.1 .93 .81 .00 5.02 3 substances 2%
School Early Initiation Substance Use
Negative School General 2.10 .41 1.00 3.60 Cigarette Smoking 17% 0 1
School 1.93 .90 .00 3.50 Alcohol Use 37% 0 1
Alcohol Env.
School Drug Env. 1.24 .90 .00 3.50 Marijuana Use 5% 0 1
chool Polysubstance Env.1 1.59 .83 .00 3.50 Polysubstance Use .57 .80 0 3
Peer Sociodemographics
Negative Peer General 1.44 .33 1.00 2.98 Two Parent Household 61% 0 1
Peer Alcohol Env. .67 .48 .00 2.51 Parent College Education 25% 0 1
Peer Drug Env. .42 .52 .00 3.26 Family Income 4.99 1.85 1 7
Peer Polysubstance Env.1 .55 .46 .00 2.89 Family Housing Instability 2.16 1.11 1 5
Family Female
Negative Family General 2.26 .44 1.00 3.80 African American 26% 0 1
Family Smoking Env. .50 .37 .00 1.67 Asian American 22% 0 1
Family Alcohol Env. .73 .44 .00 2.42 European American 47% 0 1
Family Drug Env. .28 .33 .00 2.78 Native American 5% 0 1
Family Polysubstance Env.1 .50 .31 .00 1.75 0 1
Open in a new tab

Notes. N = 766, M = mean, SD = standard deviation, NH = neighborhood, Env. = environment, 1 = polysubstance env, are the mean of available measures of alcohol, drug, and smoking environments within each domain.

Cigarette smoking, binge drinking, marijuana use.

Youth responded to three questions indicating how often in the past month they smoked cigarettes, had 5 or more drinks in a row, or smoked marijuana (0 = no past month use, 1 = any past month use). Past month polysubstance use was indicated by the sum smoking, binge drinking, and marijuana use. Early initiation of cigarette smoking, alcohol use, and marijuana use by grade 5 were drawn from three questions (0 = never used, 1 = ever used) concerning lifetime use (Kosterman et al., 2000).

Neighborhood.

Principal components analysis (PCA) summarizing 10 block group-level variables from the 1990 census were used to measure neighborhood structural factors. Participants were located in 419 block groups with an average of 1.8 students per block group. Two factor scores from the PCA represented socioeconomic disadvantage and residential stability. Identical measures derived from the census have been employed by other studies (Cambron, Kosterman, Catalano, Guttmannova, & Hawkins, 2018) and PCA results are provided in the supplementary material. Higher scores on the neighborhood structural factors indicated neighborhoods with lower socioeconomic resources and more stable resident populations. Youth experiences with their neighborhood environment were constructed from self-report data. A mean scale of negative general neighborhood environment was drawn from 5 items (internal consistency = .87) concerning the presence of gangs, crime, rowdy or undesirable people, or kids who got in trouble (Sampson et al., 2002). Neighborhood alcohol environment was measured by a single item of how many adults the youth knew that have been drunk in the last year. An index of neighborhood drug environment was constructed from the mean of three items concerning how many adults the youth knew who sold drugs, how many adults the youth knew who used marijuana or other drugs, and if there was drug selling in their neighborhood.

School.

A mean scale of negative general school environment was constructed from youth answers to 18 items concerning opportunities and involvement with classroom or extracurricular activities, positive or negative feedback from teachers, and strength of connections with school and teachers (internal consistency = .85). An index of school alcohol environment was measured as the mean of two items gauging youth perceptions of the percent of students who drank alcohol in the past year and whether people at school thought it was acceptable for students to drink alcohol. An index of school drug environment was measured as the mean of two items gauging youth perception of the percent of students who smoked marijuana in the past year and whether people at school thought it was acceptable for students to smoke marijuana.

Peer.

A mean scale of negative general peer environment was constructed from youth responses on 14 items (internal consistency = .76) about antisocial behavior of their three best friends (e.g. getting in trouble with teachers, gang involvement, and knowing other kids in gangs) and opportunities for antisocial behavior (e.g., being asked to do illegal things or to join a gang). An index of peer alcohol environment was created from the mean of 7 items about alcohol use among the youth’s three best friends and how many kids the youth knew who drank alcohol. An index of peer drug environment was created from the mean of 4 items about marijuana use among the youth’s three best friends and how many kids the youth knew who smoked marijuana or used other drugs.

Family.

Family demographics were reported by parents of youth enrolled in the study. Income was measured by a seven-category question with response options ranging from less than $5,000 to greater than $40,000 annual income. Two parent household was coded as 1 for two acting parents present and 0 for two acting parents not present. Parent college education was indicated by both mother and father education (0 = neither parent graduated 4-year college; 1 = at least one parent graduated 4-year college). A mean scale of negative general family environment was constructed from youth responses to 24 items (internal consistency = .86) concerning family management (e.g., parental monitoring and communication), family conflict (e.g., problem solving strategies), involvement (e.g., shared activities between parent and child), and bonding (e.g., youth’s connection with parents and sibling). An index offamily smoking environment was created from the mean of 4 questions. Parents reported their perceived harm from smoking cigarettes, their hypothetical response to their child smoking cigarettes, and if the child has ever smoked a cigarette with their permission. Youth reported if they had any siblings who smoked cigarettes. An index offamily alcohol environment was constructed from the mean of 11 questions. Parents reported the quantity and frequency of alcohol consumption for themselves and their spouse, their perceived harm from alcohol use, and their hypothetical response to their child drinking alcohol at a party. Youth reported if their parents had given them permission to drink alcohol or if they had sibling who drank alcohol. An index of family drug environment was constructed from the mean of 6 questions. Parents reported the frequency of marijuana use for themselves and their spouse, perceived harm from marijuana use, and their hypothetical response to their child smoking marijuana. Youth reported if they had any siblings who used marijuana or other drugs. [Table 1]

Analytic Strategy

Logistic regression was used to examine associations between covariates and the likelihood of past month cigarette smoking, binge drinking, and marijuana use; ordinal logistic regression was used for polysubstance use. Cluster-robust standard errors were employed to account for multiple youth living in some neighborhoods (Muthén & Muthén, 2013). All non-binary covariates across neighborhood, school, peer, and family domains were standardized prior to analysis. For each substance use outcome, six domain-specific models independently considered structural neighborhood, self-reported neighborhood, school, peer, and family experiences, and sociodemographic factors. A fully controlled model including all domains simultaneously, sociodemographic factors, and early initiation of substance use was also estimated for each outcome. For each substance use outcome, the corresponding substance use-specific environmental risk factors and measure of early initiation were employed. Measures assessing neighborhood, school, or peer cigarette smoking environments were not available in our data and averages of alcohol and drug environments were used as a proxy in models predicting cigarette smoking. Missing data were handled via the multiple imputation procedure in Mplus. Data were present for over 98% of possible data points across the 35 variables used in these analyses (26,307 out of 26,810). Data were analyzed using the maximum likelihood estimator for robust standard errors (MLR) for categorical outcomes in Mplus 7.1 and reported results were combined across 40 imputed datasets according to Rubin’s rules (Muthén & Muthén, 2013). Sensitivity tests using multilevel models produced substantive similar results. By 9th grade over 43% of participants had dispersed outside Seattle schools, for which we did not have school codes. As such, we were not able to examine clustering by school, but were reassured by the substantial dispersion of youth.1

Results

Full reporting of odds ratios and 95% confidence intervals for both domain-specific and all domain models are presented in Table 2 (graphical comparisons and bivariate correlations are available in an online supplement). The following text summarizes the significant findings for domain-specific and all domain models for each substance. Each model controlled for sociodemographic factors and early initiation of substance use. Odds ratios for those covariates are reported in Table 2. Table 3 provides the proportion of variance explained by domain-specific and all domain models described in Table 2. For past month cigarette smoking, domain-specific models found that the odds of smoking rose with greater neighborhood disadvantage, more negative general environments in school and family domains, and increased substance use-specific environments in neighborhood, school, peer, and family domains. In a fully controlled model considering all domains of risk simultaneously, odds of cigarette smoking rose only with greater neighborhood disadvantage and increased substance use-specific environments in peer and family domains. For past month binge drinking, domain-specific models found that the odds of binge drinking rose with greater neighborhood disadvantage, more negative general environments in peer and family domains, and increased alcohol-specific environments in neighborhood, school and peer domains. In a fully controlled model considering all domains of risk simultaneously, odds of binge drinking rose only with greater neighborhood disadvantage, more negative general peer environments, and increased alcohol-specific environments in school and peer domains. For past month marijuana use, domain-specific models found that the odds of marijuana use rose with more negative general school, peer, and family environments and increased drug-specific environments across neighborhood, school, peer, and family domains. In the fully controlled model considering all domains of risk simultaneously, odds of marijuana rose only with increased drug-specific environments in school and peer domains. For past month polysubstance use, domain-specific models found that the odds of using any additional type of substance rose with greater neighborhood disadvantage, more negative general environments across school, peer, and family domains, and increased substance use-specific environments across neighborhood, school, peer, and family domains. In a fully controlled model considering all domains of risk simultaneously, odds of using any additional type of substance rose with greater neighborhood disadvantage and increased substance use-specific environments in neighborhood, peer, and family domains. [Tables 2, 3]

Table 2.

Odds ratios [95% confidence intervals] for domain-specific and all models

Variable Cigarette Smoking Binge Drinking Marijuana Use Polysubstance Use
Domain-Specific All Domains Domain-Specific All Domains Domain-Specific All Domains Domain-Specific All Domains
Neighborhood Structural (census)
Socioeconomic Disadvantage 1.35 1.30 1.34 1.53 1.22 1.11 1.34 1.38
[1.11 - 1.64] [1.03 - 1.64] [1.03 - 1.74] [1.02 - 2.32] [.94 - 1.58] [.76 - 1.61] [1.13 - 1.60] [1.12 - 1.72]
Residential Stability .91 .87 .89 .86 1.15 1.10 .99 .94
[.73 - 1.13] [.69 - 1.11] [.65 - 1.22] [.59 - 1.24] [.85 - 1.54] [.74 - 1.63] [.82 - 1.20] [.74 - 1.18]
Neighborhood Experience
Negative NH General 1.23 1.08 1.10 .69 .89 .72 1.09 .85
[1.00 - 1.52] [.84 - 1.40] [.82 - 1.46] [.46 - 1.04] [.65 - 1.23] [.47 - 1.11] [.90 - 1.31] [.67 - 1.08]
NH Substance Use Env. 1.91 1.35 1.99 1.22 2.58 1.33 2.33 1.41
[1.50 - 2.42] [.99 - 1.85] [1.51 - 2.62] [.87 - 1.70] [1.96 - 3.39] [.92 - 1.93] [1.87 - 2.91] [1.05 - 1.89]
School
Negative School General 1.31 1.15 1.18 .94 1.34 1.10 1.31 1.14
[1.07 - 1.60] [.89 - 1.49] [.90 - 1.56] [.66 - 1.34] [1.02 - 1.77] [.75 - 1.59] [1.08 - 1.60] [.90 - 1.44]
School Substance Use Env. 1.54 .84 2.95 1.77 3.31 1.83 2.10 1.15
[1.20 - 1.97] [.62 - 1.15] [2.00 - 4.35] [2.00 - 4.35] [2.45 - 4.49] [1.24 - 2.71] [1.65 - 2.67] [.87 - 1.51]
Peer
Negative Peer General 1.11 1.00 1.78 1.60 1.53 1.24 1.36 1.21
[.85 - 1.46] [.74 - 1.34] [1.31 - 2.4] [1.12 - 2.27] [1.12 - 2.08] [.84 - 1.83] [1.10 - 1.70] [.96 - 1.53]
Peer Substance Use Env. 2.20 2.03 3.64 3.48 2.61 2.08 2.54 2.13
[1.71 - 2.83] [1.51 - 2.72] [2.34 - 5.66] [2.13 - 5.69] [1.97 - 3.44] [1.52 - 2.84] [2.02 - 3.20] [1.65 - 2.75]
Family
Negative Family General 1.29 .92 1.45 1.08 1.60 1.18 1.36 .98
[1.02 - 1.63] [.69 - 1.23] [1.10 - 1.92] [.75 - 1.57] [1.20 - 2.12] [.80 - 1.74] [1.10 - 1.67] [.76 - 1.27]
Family Substance Use Env. 1.89 1.64 1.33 .93 1.64 1.12 1.75 1.26
[1.57 - 2.28] [1.35 - 1.99] [.99 - 1.77] [.65 - 1.32] [1.33 - 2.02] [.87 - 1.44] [1.46 - 2.10] [1.03 - 1.53]
Sociodemographics
Two Parent Household .92 1.02 .89 1.09 1.23 1.33 .96 1.12
[.58 - 1.48] [.63 - 1.67] [.47 - 1.69] [.53 - 2.25] [.62 - 2.40] [.57 - 3.10] [.63 - 1.48] [.69 - 1.83]
Parent College Education .74 .83 .91 .97 .44 .36 .62 .67
[.43 - 1.28] [.46 - 1.49] [.43 - 1.93] [.39 - 2.44] [.19 - 1.01] [.13 - .99] [.37 - 1.04] [.37 - 1.21]
Family Income .59 .73 .71 .71 .74 .99 .60 .69
[.46 - .76] [.55 - .98] [.50 - 1.01] [.46 - 1.09] [.52 - 1.05] [.61 - 1.62] [.47 - .76] [.52 - .93]
Family Housing Instability 1.09 1.00 1.19 1.11 1.21 1.19 1.16 1.07
[.90 - 1.32] [.79 - 1.26] [.94 - 1.51] [.82 - 1.48] [.98 - 1.50] [.91 - 1.57] [.99 - 1.35] [.89 - 1.27]
Female .93 .88 .64 .52 1.13 1.02 1.00 .87
[.62 - 1.4] [.53 - 1.46] [.36 - 1.14] [.25 - 1.05] [.65 - 1.96] [.51 - 2.03] [.69 - 1.44] [.56 - 1.37]
African American .36 .24 .56 .34 1.04 .62 .51 .31
[.19 - .66] [.11 - .51] [.25 - 1.25] [.11 - 1.05] [.52 - 2.05] [.25 - 1.55] [.32 - .83] [.16 - .60]
Asian American .26 .36 .61 1.43 .21 .58 .26 .41
[.14 - .50] [.16 - .79] [.25 - 1.52] [.49 - 4.22] [.07 - .64] [.13 - 2.62] [.14 - .49] [.20 - .87]
Native American 3.72 1.96 4.05 2.71 2.37 1.66 3.66 1.69
[1.71 - 8.13] [.68 - 5.66] [1.56 - 1.47] [.73 - 10.08] [1.01 - 5.55] [.52 - 5.28] [1.84 - 7.27] [.70 - 4.08]
Early Initiation Substance Use 2.78 1.83 2.50 2.08 3.16 2.23 1.60 1.22
1.72 - 4.52] [1.06 - 3.17] [1.35 - 4.62] [.97 - 4.47] [1.35 - 7.41] [.81 - 6.19] [1.32 - 1.95] [.98 - 1.51]
Open in a new tab

Notes. N=766; bold indicates p < .05; NH= neighborhood; Env. = environment; substance use-specific environmental risk factors and early initiation measures in each model reflect constructs relevant to that outcome (i.e. NH substance use environment = NH alcohol environment for the binge drinking model and NH substance use environment = NH drug environment for the marijuana use model); neighborhood, school, and peer alcohol and drug environment measures are used as a proxy for the cigarette smoking; each domain-specific model accounts for sociodemographic factors and early initiation; all domain models include all covariates simulataneously; all neighborhood, school, peer, and family risk factors were standardized prior to analysis; race/ethnicity variables are compared to European Americans.

Table 3.

Proportion of variance explained in past month substance use by domain

Models Cigarette Smoking Binge Drinking Marijuana Use Polysubstance Use
Domain-specific
Neighborhood Structural (census) .24 .17 .23 .24
Neighborhood Experience .31 .25 .31 .33
School .27 .35 .43 .34
Peer .33 .51 .40 .40
Family .31 .20 .30 .29
All Domains .41 .58 .48 .45
Open in a new tab

Notes. Maximum proportion = 1; each domain-specific model included general and substance use-specific factors for that domain only, sociodemographic factors, and early initiation of substance use; all domain models included all covariates simultaneously.

Discussion

Understanding the role of neighborhood structural factors in youth substance use is key for both reducing the detrimental effects of youth substance use and increasing health equity (Office of Surgeon General, 2016; Viner et al., 2012). The results of this study provide important information on both of these fronts. We are not aware of any study that has examined neighborhood structural factors in conjunction with as extensive a set of modifiable general and substance use-specific risk factors across neighborhood, school, peer, and family domains while also controlling for family-level sociodemographic factors and early initiation of substance use. As such, fully controlled models provide a very conservative estimate of associations between each domain of covariates and substance use. Domain-specific models found a number of significant associations of general and substance use-specific risk factors across neighborhood, school, peer, and family with youth substance use even after accounting for sociodemographic differences and early initiation of substance use (see Table 2). Domain-specific models showed that the peer domain accounted for the highest proportion of variance in cigarette smoking, binge drinking, and polysubstance use (see Table 3). Fully controlled models highlighted the importance of exposure to substance use-specific environments in the etiology of youth cigarette smoking, binge drinking, marijuana use, and polysubstance use and the most consistent relationships were noted for peer substance use environments. The consistency of these results were not unexpected given the long history of theoretical and empirical work highlighting the central importance of peer group influences on youth norms, intentions, and substance using behaviors (Dodge, Dishion, & Lansford, 2006; Van Ryzin, Fosco, & Dishion, 2012).

The primary goal of this study was to clarify associations between neighborhood structural factors (i.e., socioeconomic disadvantage and residential stability) and youth substance use; an important goal for health equity research (Viner et al., 2012). Even after accounting for the impact of key general and substance use-specific risk factors and family-level sociodemographic factors, living in a more socioeconomically disadvantaged neighborhood was associated with a higher likelihood of past month cigarette smoking, binge drinking, and polysubstance use, but not marijuana use. Comparing estimates for neighborhood structural factors alone to estimates from all domain models showed little evidence that domain-specific risk factors were accounting for the association between neighborhood disadvantage and youth substance use. Domain-specific models indicated that neighborhood structural factors accounted for the smallest proportion of variance in each substance use outcome as compared to other domains suggesting that neighborhood structural factors represent the most distal contextual factor for substance use in this study (see Figure 1). These models found no evidence of an association between residential stability or negative general neighborhood environment and youth substance use despite some moderate bivariate correlations among these constructs (see Table 3). Other studies, however, have found direct effects of more unstable resident populations or other general features of neighborhood environment (i.e. low social cohesion or collective efficacy) on increased youth problem behavior (Sampson et al., 2002) and, more specifically, greater alcohol use among young adolescents (Jackson et al., 2016).

To achieve health equity goals, further research is needed to understand the mechanisms connecting neighborhood socioeconomic disadvantage and youth substance use. Recent analyses from the Moving to Opportunity study suggest that youth moving out of lower SES neighborhoods experienced reduced exposure to negative peer groups and, in turn, engaged in less substance use (Rudolph et al., 2018). Other studies have shown that youth living in more disadvantaged neighborhoods are often exposed to greater alcohol and tobacco advertising and point of sale retail outlets (Lee, Henriksen, Rose, Moreland-Russell, & Ribisl, 2015; Moore, Jones-Webb, Toomey, & Lenk, 2008). Greater exposure to alcohol and tobacco advertising and outlets has been linked with increased risk of youth substance use (Anderson, De Bruijn, Angus, Gordon, & Hastings, 2009; Bryden, Roberts, McKee, & Petticrew, 2012; Henriksen, Schleicher, Feighery, & Fortmann, 2010; Jackson et al., 2014). Increased exposure to outlets may offer more opportunities for youth to obtain alcohol or cigarettes or normalize use within a youth’s neighborhood (Jackson et al., 2014). Thus, exposure to outlets may be one mechanism connecting neighborhood disadvantage with youth cigarette smoking and alcohol use. This hypothesis is consistent with the findings of the current study given that cigarette smoking and binge drinking were significantly associated with neighborhood disadvantage, but marijuana use was not. The study period for these analyses predated laws in many states allowing for storefront sales of medical and recreational marijuana (Cambron, Guttmannova, & Fleming, 2017). Therefore, youth in this study were precluded from exposure to legal marijuana-related advertising or outlets. Recent research has reported that greater exposure to legal medical marijuana advertising or outlets is associated with increased youth marijuana use and intentions to use (D’Amico, Miles, & Tucker, 2015). Other studies have noted that marijuana-related outlets are concentrated in higher poverty neighborhoods (Morrison, Gruenewald, Freisthler, Ponicki, & Remer, 2014). Researchers have suggested that higher density point of sale outlets may serve as a proxy for neighborhood substance use-specific environments (Jackson et al., 2014) and may be most relevant when considering the accumulation of neighborhood risk factors along with socioeconomic disadvantage and neighborhood social processes (Chilenski & Greenberg, 2009; Duncan et al., 2002). Future research should explicitly examine exposure to advertising and outlets for cigarettes, alcohol, and marijuana as a potential mechanism relating neighborhoods to youth substance use and consider the potential impact of accumulated neighborhood risks for substance use. A better understanding of these links could help inform prevention policies designed reduce opportunities for youth to obtain substances illegally, limit youth exposure to substance use-specific advertising, and facilitate the development and appropriate placement of media campaigns promoting healthy youth behavior.

Some limitations to this study should be noted. The SSDP sample included a small number of students per block group and, as such, the perceptions of youth in this study may not be representative of all youth in their neighborhoods. Measures of neighborhood, school, or peer cigarette smoking and neighborhood laws, policing, or enforcement strategies were not available and may, as well, represent important omitted variables (Toumbourou et al., 2007). It is also worth considering that there is likely overlap across domains for some substance use-specific environment measures. For instance, respondents may be describing some the same individuals when answering questions about peer and school substance use environments. As well, when asked about substance use and drug selling among adults they know, respondents may be including some parents and siblings. Analyses of multicollinearity, however, suggested that constructs across domains contributed unique variance to the models.

Importantly, these data represent youth experiences in 1990 and do not reflect any environmental changes over the past decades. As mentioned above, the SSDP sample predates current trends with marijuana policy liberalization (Cambron et al., 2017) and, as a result, youth in lower SES neighborhoods in this study did not experience greater exposure marijuana advertising and outlets (Morrison et al., 2014). Additionally, recent public health policy successes with restrictions on tobacco use in public and a decline in tobacco use require further consideration. While these trends have likely reduced youth exposure to tobacco in the general population, health equity researchers have consistently reported that lower SES neighborhoods and families have not realized the benefits of changing tobacco policies and norms (Drope et al., 2018; Zhang, Martinez-Donate, Kuo, Jones, & Palmersheim, 2012). As a result, cigarette smoking has become increasingly concentrated in lower SES populations (Drope et al., 2018). Given the higher concentration of cigarette smoking and recent addition of marijuana outlets in lower SES neighborhoods, it is possible that youth currently living in lower SES neighborhoods experience similar or greater risk for tobacco and marijuana use.

It is also important to note that the dynamics of school, peer, and family environments have likely changed in recent decades due, in part, to technological innovation and the ubiquity of digital communication. Recent studies have suggested that technology can be used both to engage youth healthy decision making processes around substance use (Dunne, Bishop, Avery, & Darcy, 2017) or to spread pro-substance use norms via social media (Cavazos-Rehg, Krauss, Sowles, & Bierut, 2016; Elmore, Scull, & Kupersmidt, 2017). Further research is needed to understand how these rapidly evolving technology-based influences may shape risk and protective factor models for youth substance use.

Despite these limitations, this study examined a comprehensive set of neighborhood, school, peer, and family contextual risk factors for youth cigarette smoking, binge drinking, marijuana use, and polysubstance use. As such, these results provide important information for prevention scientists. While deviant and substance using peers offer a well-established intervention target (Dodge et al., 2006; Hawkins et al., 1992), malleable neighborhood, school, and family factors all remained relevant in fully controlled models suggesting multiple potential points of intervention. Programs leveraging the results of this and other studies can help ameliorate the damaging and costly impacts of substance use on youth (Hawkins et al., 2015). Most importantly, we wish to implore researchers to continue examining the unique role of neighborhood structural factors in youth substance use. Studies that bring prevention science’s extensive etiological and programmatic knowledge of risk and protective factors to bear on the central questions of health equity can undoubtedly improve both short- and long-term outcomes for youth from socioeconomically disadvantaged backgrounds (Viner et al., 2012).

Supplementary Material

11121_2019_1072_MOESM1_ESM

Acknowledgments

Funding

Data collection for this study was supported by grants from the National Institute on Drug Abuse (5R01DA003721 and 5R01DA033956). Support came from a National Poverty Research Center Dissertation Fellowship awarded by the Institute for Research on Poverty at the University of Wisconsin-Madison with funding from the Office of the Assistant Secretary for Planning and Evaluation, U.S. Department of Health and Human Services, Cooperative Agreement number AE00103. Support for this study also came from a Eunice Kennedy Shriver National Institute of Child Health and Human Development research infrastructure grant, No. R24HD042828; and training grant No. T32HD007543 to the Center for Studies in Demography and Ecology at the University of Washington. The opinions and conclusions expressed herein are solely those of the authors and should not be construed as representing the opinions or policy of any agency of the Federal government.

Footnotes

Publisher's Disclaimer: This Author Accepted Manuscript is a PDF file of a an unedited peer-reviewed manuscript that has been accepted for publication but has not been copyedited or corrected. The official version of record that is published in the journal is kept up to date and so may therefore differ from this version.

1

Some SSDP participants received a social development intervention during elementary school (Hill et al., 2016). Sensitivity tests controlling for the intervention in all domain models produced substantively similar results.

Conflicts of Interest

R. F. Catalano is on the board of Channing Bete Company, distributer of prevention programs. Other authors have no conflicts of interest to report.

Ethical approval

All procedures performed in studies involving human participants were in accordance with the ethical standards of the University of Washington and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards.

Informed Consent

Informed consent was obtained from all individual participants included in the study.

References

  1. Anderson P, De Bruijn A, Angus K, Gordon R, & Hastings G (2009). Impact of alcohol advertising and media exposure on adolescent alcohol use: a systematic review of longitudinal studies. Alcohol and alcoholism, 44(3), 229–243. [DOI] [PubMed] [Google Scholar]
  2. Arthur MW, Hawkins JD, Pollard JA, Catalano RF, & Baglioni A Jr (2002). Measuring risk and protective factors for use, delinquency, and other adolescent problem behaviors: The Communities That Care Youth Survey. Evaluation review, 26(6), 575–601. [DOI] [PubMed] [Google Scholar]
  3. Bailey JA, Hill KG, Meacham MC, Young SE, & Hawkins JD (2011). Strategies for characterizing complex phenotypes and environments: General and specific family environmental predictors of young adult tobacco dependence, alcohol use disorder, and co-occurring problems. Drug and alcohol dependence, 118(2), 444–451. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Blum RW, Beuhring T, Shew ML, Bearinger LH, Sieving RE, & Resnick MD (2000). The effects of race/ethnicity, income, and family structure on adolescent risk behaviors. American Journal of Public Health, 90(12), 1879. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bronfenbrenner U (1977). Toward an experimental ecology of human development. American psychologist, 32(7), 513. [Google Scholar]
  6. Bryden A, Roberts B, McKee M, & Petticrew M (2012). A systematic review of the influence on alcohol use of community level availability and marketing of alcohol. Health & place, 18(2), 349–357. [DOI] [PubMed] [Google Scholar]
  7. Cambron C, Guttmannova K, & Fleming CB (2017). State and National Contexts in Evaluating Cannabis Laws: A Case Study of Washington State. Journal of drug issues, 47(1), 74–90. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Cambron C, Kosterman R, Catalano RF, Guttmannova K, & Hawkins JD (2018). Neighborhood, family, and peer factors associated with early adolescent smoking and alcohol use. Journal of youth and adolescence, 47(2), 369–382. doi: 10.1007/s10964-017-0728-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Catalano RF, Fagan AA, Gavin LE, Greenberg MT, Irwin CE, Ross DA, & Shek DT (2012). Worldwide application of prevention science in adolescent health. The Lancet, 379(9826), 1653–1664. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Catalano RF, & Hawkins JD (1996). The social development model: A theory of antisocial behavior In J. D. Hawkins (Ed.), Delinquency and crime: Current theories (pp. 149–197). New York: Cambridge University Press. [Google Scholar]
  11. Catalano RF, Kosterman R, Hawkins JD, Newcomb MD, & Abbott RD (1996). Modeling the etiology of adolescent substance use: A test of the social development model. Journal of drug issues, 26(2), 429–455. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Cavazos-Rehg PA, Krauss MJ, Sowles SJ, & Bierut LJ (2016). Marijuana-related posts on Instagram. Prevention Science, 17(6), 710–720. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Chilenski SM, & Greenberg MT (2009). The importance of the community context in the epidemiology of early adolescent substance use and delinquency in a rural sample. American journal of community psychology, 44(3–4), 287–301. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. D’Amico EJ, Miles J, & Tucker JS (2015). Gateway to curiosity: medical marijuana ads and intention and use during middle school. Psychol Addict Behav, 1(1). [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Diez Roux AV, & Mair C (2010). Neighborhoods and health. Annals of the New York Academy of Sciences, 1186(1), 125–145. [DOI] [PubMed] [Google Scholar]
  16. Dodge KA, Dishion TJ, & Lansford JE (2006). Deviant Peer Influences in Intervention and Public Policy for Youth Social Policy Report. Volume 20, Number 1 Society for Research in Child Development. [Google Scholar]
  17. Drope J, Liber AC, Cahn Z, Stoklosa M, Kennedy R, Douglas CE, … Drope J (2018). Who’s still smoking? Disparities in adult cigarette smoking prevalence in the United States. CA: a cancer journal for clinicians, 68(2), 106–115. [DOI] [PubMed] [Google Scholar]
  18. Duncan SC, Duncan TE, & Strycker LA (2002). A multilevel analysis of neighborhood context and youth alcohol and drug problems. Prevention Science, 5(2), 125–133. [DOI] [PubMed] [Google Scholar]
  19. Dunne T, Bishop L, Avery S, & Darcy S (2017). A review of effective youth engagement strategies for mental health and substance use interventions. Journal of Adolescent Health, 60(5), 487–512. [DOI] [PubMed] [Google Scholar]
  20. Elmore KC, Scull TM, & Kupersmidt JB (2017). Media as a “super peer”: how adolescents interpret media messages predicts their perception of alcohol and tobacco use norms. Journal of youth and adolescence, 46(2), 376–387. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Epstein M, Hill KG, Bailey JA, & Hawkins JD (2013). The effect of general and drug-specific family environments on comorbid and drug-specific problem behavior: A longitudinal examination. Developmental psychology, 49(6), 1151. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Fagan AA, Wright EM, & Pinchevsky GM (2015). A multi-level analysis of the impact of neighborhood structural and social factors on adolescent substance use. Drug and alcohol dependence, 153, 180–186. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Furr-Holden CDM, Lee MH, Johnson R, Milam AJ, Duncan A, Reboussin BA, … Ialongo NS (2015). Neighborhood environment and marijuana use in urban young adults. Prevention Science, 16(2), 268–278. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Gardner M, Barajas RG, & Brooks-Gunn J (2010). Neighborhood influences on substance use etiology: Is where you live important. Handbook of drug use etiology: Theory, methods, and empirical findings, 423–441. [Google Scholar]
  25. Hanson MD, & Chen E (2007). Socioeconomic status and health behaviors in adolescence: a review of the literature. Journal of behavioral medicine, 30(3), 263. [DOI] [PubMed] [Google Scholar]
  26. Hawkins JD, Catalano RF, & Miller JY (1992). Risk and protective factors for alcohol and other drug problems in adolescence and early adulthood: implications for substance abuse prevention. Psychological bulletin, 112(1), 64. [DOI] [PubMed] [Google Scholar]
  27. Hawkins JD, Jenson JM, Catalano R, Fraser MW, Botvin GJ, Shapiro V, … Leslie LK (2015). Unleashing the power of prevention: Institute of Medicine of the National Academies. [Google Scholar]
  28. Henriksen L, Schleicher NC, Feighery EC, & Fortmann SP (2010). A longitudinal study of exposure to retail cigarette advertising and smoking initiation. Pediatrics, 126(2), 232–238. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Hill KG, Hawkins JD, Catalano RF, Abbott RD, & Guo J (2005). Family influences on the risk of daily smoking initiation. Journal of Adolescent Health, 37(3), 202–210. [DOI] [PubMed] [Google Scholar]
  30. Hill KG, Woodward D, Woelfel T, Hawkins JD, & Green S (2016). Planning for long-term follow-up: Strategies learned from longitudinal studies. Prevention Science, 17(7), 806–818. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Hiscock R, Bauld L, Amos A, Fidler JA, & Munafò M (2012). Socioeconomic status and smoking: a review. Annals of the New York Academy of Sciences, 1248(1), 107–123. [DOI] [PubMed] [Google Scholar]
  32. Hoffmann JP, & Johnson RA (1998). A national portrait of family structure and adolescent drug use. Journal of Marriage and the Family, 633–645. [Google Scholar]
  33. Jackson N, Denny S, & Ameratunga S (2014). Social and socio-demographic neighborhood effects on adolescent alcohol use: A systematic review of multi-level studies. Social science & medicine, 115, 10–20. [DOI] [PubMed] [Google Scholar]
  34. Jackson N, Denny S, Sheridan J, Zhao J, & Ameratunga S (2016). The role of neighborhood disadvantage, physical disorder, and collective efficacy in adolescent alcohol use: a multilevel path analysis. Health & place, 41, 24–33. [DOI] [PubMed] [Google Scholar]
  35. Jenson JM, & Bender K (2014). Preventing child and adolescent problem behavior: Evidence-based strategies in schools, families, and communities. [Google Scholar]
  36. Karriker-Jaffe KJ (2011). Areas of disadvantage: A systematic review of effects of area-level socioeconomic status on substance use outcomes. Drug and alcohol review, 30(1), 84–95. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Kosterman R, Hawkins JD, Guo J, Catalano RF, & Abbott RD (2000). The dynamics of alcohol and marijuana initiation: patterns and predictors of first use in adolescence. American Journal of Public Health, 90(3), 360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Lee JG, Henriksen L, Rose SW, Moreland-Russell S, & Ribisl KM (2015). A systematic review of neighborhood disparities in point-of-sale tobacco marketing. American Journal of Public Health, 105(9), e8–el8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Lee JO, Hill KG, Guttmannova K, Bailey JA, Hartigan LA, Hawkins JD, & Catalano RF (2012). The effects of general and alcohol-specific peer factors in adolescence on trajectories of alcohol abuse disorder symptoms from 21 to 33 years. Drug and alcohol dependence, 121(3), 213–219. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Lonczak HS, Huang B, Catalano RF, Hawkins JD, Hill KG, Abbott RD, …Kosterman R (2001). The social predictors of adolescent alcohol misuse: a test of the social development model. Journal of studies on alcohol, 62(2), 179–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Moore H, Jones-Webb R, Toomey T, & Lenk K (2008). Alcohol advertising on billboards, transit shelters, and bus benches in inner-city neighborhoods. Contemporary Drug Problems, 35(2–3), 509–532. [Google Scholar]
  42. Morrison C, Gruenewald PJ, Freisthler B, Ponicki WR, & Remer LG (2014). The economic geography of medical cannabis dispensaries in California. International Journal of Drug Policy, 25(3), 508–515. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Muthén LK, & Muthén BO (2013). Mplus 7.11. Los Angeles: Muthén & Muthén. [Google Scholar]
  44. Office of Surgeon General. (2016). Facing Addiction in America: The Surgeon General’s Report on Alcohol, Drugs, and Health. Washington, D.C. [PubMed] [Google Scholar]
  45. Rudolph KE, Sofrygin O, Schmidt NM, Crowder R, Glymour MM, Ahern J, & Osypuk TL (2018). Mediation of neighborhood effects on adolescent substance use by the school and peer environments. Epidemiology, 29(4), 590–598. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Ryan SM, Jorm AF, & Lubman DI (2010). Parenting factors associated with reduced adolescent alcohol use: a systematic review of longitudinal studies. Australian and New Zealand Journal of Psychiatry, 44(9), 774–783. [DOI] [PubMed] [Google Scholar]
  47. Sampson RJ, Morenoff JD, & Gannon-Rowley T (2002). Assessing” neighborhood effects”: Social processes and new directions in research. Annual review of sociology, 443–478. [Google Scholar]
  48. Toumbourou JW, Stockwell T, Neighbors C, Marlatt G, Sturge J, & Rehm J (2007). Interventions to reduce harm associated with adolescent substance use. The Lancet, 369(9570), 1391–1401. [DOI] [PubMed] [Google Scholar]
  49. Van Ryzin MJ, Fosco GM, & Dishion TJ (2012). Family and peer predictors of substance use from early adolescence to early adulthood: An 11-year prospective analysis. Addictive Behaviors, 37(12), 1314–1324. [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Viner RM, Ozer EM, Denny S, Marmot M, Resnick M, Fatusi A, & Currie C (2012). Adolescence and the social determinants of health. The Lancet, 379(9826), 1641–1652. [DOI] [PubMed] [Google Scholar]
  51. Wodtke GT, Elwert F, & Harding DJ (2016). Neighborhood Effect Heterogeneity by Family Income and Developmental Period 1. American journal of sociology, 121(4), 1168–1222. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Zhang X, Martinez-Donate AP, Kuo D, Jones NR, & Palmersheim KA (2012). Trends in home smoking bans in the USA, 1995–2007: prevalence, discrepancies and disparities. Tobacco Control, 21(3), 330–336. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

11121_2019_1072_MOESM1_ESM
NIHMS1546978-supplement-11121_2019_1072_MOESM1_ESM.xlsx (596KB, xlsx)

RESOURCES