Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2024 Oct 1.
Published in final edited form as: Res Child Adolesc Psychopathol. 2023 Jun 5;51(10):1425–1438. doi: 10.1007/s10802-023-01090-9

Polygenic Effects on Individual Rule Breaking, Peer Rule Breaking, and Alcohol Sips across Early Adolescence in the ABCD Study

Kit K Elam 1,*, Jinni Su 2, Fazil Aliev 3, Angel Trevino 2, Jodi Kutzner 1, Dong-Chul Seo 1
PMCID: PMC10601492  NIHMSID: NIHMS1937214  PMID: 37273065

Abstract

Alcohol use emerges during early adolescence and is strongly associated with individual and peer risky, delinquent, and rule breaking behaviors. Genetic predisposition for risky behavior contributes to individual rule breaking in adolescence and can also evoke peer rule breaking or lead youth to select into delinquent peer groups via gene-environment correlations (rGE), collectively increasing risk for alcohol use. Little research has examined whether genetic predisposition for risky behavior contributes to individual and peer rule breaking behavior in developmental pathways to alcohol use in early adolescence or in large diverse racial/ethnic populations. To address this, polygenic scores for risky behavior were considered predictors of individual rule breaking, peer rule breaking, and alcohol sips using data from the Adolescent Brain Cognitive Development (ABCD) study at age 11–12 and 12–13 in a cross-time cross-lagged model. This was examined separately in European American (EA; n=5113; 47% female), African American (AA; n=1159; 50% female), and Hispanic/Latinx (Latinx; n=1624; 48% female) subgroups accounting for sociodemographic covariates and genetic ancestry principal components. Polygenic scores were positively associated with all constructs in EAs, with individual rule breaking at age 11–12 in AAs and Latinx, and with alcohol sips at age 11–12 in Latinx. Individual and peer rule breaking were associated with one another across time only in the EA subgroup. In all subgroups, peer rule breaking at 12–13 was associated with alcohol sips at 12–13. Results indicate that alcohol sips in early adolescence are associated with individual and peer rule breaking with rGE implicated in EAs.

Keywords: Polygenic, Rule breaking, Adolescence, Peer, Alcohol use, ABCD

Alcohol use and misuse are prevalent in adolescence and associated with significant personal, interpersonal, and societal costs in the United States (Sacks et al., 2015). Early initiation of alcohol use, in particular, is associated with elevated risk for developing an alcohol use disorder (Grant & Dawson, 1997). One common way that adolescents start using alcohol is through alcohol sips (Donovan & Molina, 2008). Research shows that alcohol sips in early adolescence are associated with more positive alcohol expectancies (Murphy et al., 2021), lower school engagement (Jackson et al., 2015), and higher odds of binge drinking and alcohol-related harms (Aiken et al., 2020). Across early adolescence there are transactions between individual rule breaking and peer rule breaking which exacerbate one another over time (Hoeben et al., 2016). Both individual and peer rule breaking can serve as developmental precursors to the emergence of alcohol use (Dodge et al., 2009). Genetic predisposition for risky behavior can contribute to rule breaking in adolescence and also evoke rule breaking behavior in peers or lead youth to select into delinquent peer groups via gene-environment correlations (rGE), all of which increases risk for alcohol use over time. However, very little research has examined the contribution of genetic predisposition for risky behavior to both individual and peer rule breaking behavior over time and resulting alcohol use. Even less research exists in early adolescence or in racially-ethnically diverse population. It is critical to understand etiologic genetic, individual, and peer mechanisms in developmental pathways to alcohol sips in diverse populations. The present study examines the contribution of genetic predisposition for risky behavior to transactions between individual rule breaking and peer rule breaking in developmental pathways to alcohol sips across early adolescence.

Individual and Peer Rule-breaking Behaviors as Precursors of Alcohol Sips

Risky behavior, such as rule breaking (e.g., truancy, vandalism, theft) is common and increases across early adolescence, often escalating into alcohol use (Hair et al., 2009). Concurrent to this, youth increasingly spend time with their peers across adolescence and are exposed to their peers rule breaking behaviors (Savi-Çakar et al., 2015). Individual rule breaking and peer rule breaking transact and exacerbate one another across adolescence. This can occur as youth observe, learn, and model their peers’ rule breaking behaviors, which reinforces peer behaviors (Reiter et al., 2019). Escalations in individual and peer rule breaking behavior during adolescence are strongly associated with engaging in other risky behaviors, particularly alcohol use and substance use (Dodge et al., 2009). The reciprocal transactions between individual and peer rule breaking behavior can occur, in part, due to genetic effects.

Genetic Influences on Rule-breaking Behaviors and Alcohol Use

Risk-taking behaviors are heritable in early adolescence and adolescence (Miles et al., 2001; Wang et al., 2016). Twin and molecular genetic studies have established that risky and risk-taking behaviors are one facet of the larger construct of externalizing behavior, which also includes novelty seeking, substance use, and conduct/delinquency problems (Derringer et al., 2015; Vrieze et al., 2013; Young et al., 2009). Thus, there is both unique and shared genetic etiology across these behaviors, indicating that genetic predisposition for one facet may be associated with related behaviors. For instance, one twin study by Miles et al. (2001) found covariation in the genetic influences underlying risk-taking behaviors and drug use in adolescence.

Extending this, polygenic scores (PGS) are frequently being used to capture genetic predisposition by aggregating across small effects of single nucleotide polymorphisms (SNPs). PGS are formed by leveraging summary statistics on individual SNP effect sizes from a discovery genome wide association study (GWAS) of a trait or behavior. This information is used to form PGS based on individual’s genotype in an independent sample, resulting in a cumulative measure of genetic predisposition for a trait or behavior. One large GWAS on risky behavior currently exists. Using data from the UK Biobank, Linnér et al. (2019) conducted principal component analysis on four risky behaviors: automobile speeding propensity, drinks per week, ever smoker, and number of sexual partners. They then conducted a GWAS on the first extracted principal component of the four risky behaviors (n = 315,894 individuals). Two studies have used this GWAS to create PGS for risky behavior. Ksinan et al. (2022) found that greater PGS for risky behavior was associated with more alcohol and cannabis use in European ancestry adolescents and also lower maternal closeness. In another study, greater PGS for risky behavior was associated with more lifetime suicide attempts in African American males (Rabinowitz et al., 2021). In related research, greater PGS for externalizing behavior have been found to be associated with more rule breaking, aggression, and disinhibition in adolescence, and greater alcohol use in early adulthood (Elam et al., 2022; Elam et al., 2021). Thus, PGS for risky behavior and related phenotypes increase risk for externalizing and disinhibited behaviors. However, these studies have primarily been conducted in European descent individuals during adolescence or adulthood. The current study extends the literature by examining the role of PGS for risky behavior in influencing rule breaking behavior and alcohol use in early adolescence among diverse racial/ethnic populations. We expect higher PGS for risky behavior to be associated with greater individual rule breaking behavior, as well as more alcohol sips in early adolescence, given the overlap in genetic etiology underlying risky behavior and alcohol use.

Gene-Environment Correlations (rGE) as Pathways of Risk for Alcohol Use

An rGE occurs when one’s genetic predisposition is associated with the environment (Plomin, et al., 1977; Scarr & McCartney, 1983). One type is a passive rGE, which occurs when a child’s genotype, provided by the biological parent, is associated with her/his rearing environment, which is also provided by parents. A second type is an active rGE, which occurs when an individual selects into a specific environment based on her/his genetic predisposition. Finally, an evocative rGE occurs when an individual elicits a systematic reaction from the environment. Developmentally, passive rGEs are theorized to be most prominent in childhood and wane into adolescence whereas active rGEs occur more often from early adolescence onward. Evocative rGEs are proposed to occur consistently throughout life. Research using PGS provides evidence of evocative rGE within the family (Elam et al., 2018; Elam 2017; Su et al., 2018) and also relative to peer behavior.

In a recent study, Su et al. (2021) found greater PGS for alcohol use to be associated with lower family support in adulthood, but not friend support, and family support mediated this effect on alcohol use later in life. In another study, Elam et al. (2022) found greater PGS for aggression to be associated with greater peer rejection in middle childhood, which predicted more marijuana use in early adolescence. This finding can be interpreted as support for evocative rGE, that greater genetic predisposition for aggression elicits greater rejection from peers, increasing risk for substance use later in life. Finally, Kuo et al. (2021) found that greater PGS for externalizing was associated with lower parent knowledge and more peer substance use in early adolescence, and peer substance use was associated with greater externalizing behavior in adolescence. The association between PGS for externalizing and peer substance use could indicate an active or evocative rGE as individuals with greater PGS for externalizing behavior may elicit greater peer substance use or choose to spend more time with peers who use substances.

No study, to our knowledge, has examined rGEs underlying polygenic predisposition for risky behavior across early adolescence in developmental pathways to alcohol use. Integrating across genetic, individual, and peer literatures, we expect PGS for risky behavior to be associated with individual rule breaking and alcohol sips as well as peer rule breaking (via rGE), and that individual and peer rule breaking will transact over time and be associated with alcohol sips. As the vast majority of genetics research is in European ancestry participants it is imperative to examine this in diverse populations (Popejoy & Fullerton, 2016).

The Importance of Studying Genetic Effects in Diverse Populations

The underrepresentation of racial/ethnic minority populations in genetic research is problematic, which may exacerbate existing health disparities among racial/ethnic minorities (Martin et al., 2019). Using genetically informed designs to understand the developmental etiology of alcohol use in diverse populations is critical to better understand and reduce racial/ethnic disparities in alcohol related outcomes and to ensure equal benefit from health advances related to genetic discoveries. This is highlighted by important differences in rates of rule breaking behaviors and alcohol use, related social and health consequences, as well as environmental, cultural, and genetic factors across racial/ethnic groups. For instance, racial/ethnic minorities, such as Black/African Americans and Hispanics/Latinx, on average consume less alcohol but experience similar or more negative social and health consequences related to alcohol use compared to Non-Hispanic White individuals (Chartier & Caetano, 2010; Mulia et al., 2009). Racial/ethnic minorities also often experience structural inequalities and greater sociocontextual stressors such as lower socioeconomic status and racial discrimination (Williams et al., 2010), which may increase risky behaviors and alcohol use as well as exacerbate genetic influences. In addition, there are important differences in genetic diversity, such as allele frequencies and linkage disequilibrium patterns across populations, implying potential differences in genetic effects on risky behaviors, alcohol use, and related phenotypes (Campbell & Tishkoff, 2008; Gelernter et al., 2014). To note, genetic effects are not statistically comparable across racial/ethnic subgroups due to these genetic differences in genetic ancestry (e.g., allele frequency, linkage disequilibrium patterns), which can lead to genetic indices such as PGS operating differently across different subgroups (Martin et al., 2017). Together, these literatures suggest that genetic and environmental influences related to alcohol use and risky behaviors may vary across racial/ethnic groups, highlighting the importance of studying genetically informed pathways of risk in racially/ethnically diverse populations.

Current Study

In the current study, we examined polygenic scores for risky behavior (Risk-PGS) as predictors of individual rule breaking, peer rule breaking, and alcohol sips at age 11–12 and 12–13. In addition, we examined transactions among individual rule breaking, peer rule breaking, and alcohol sips within time and across time. This was tested separately in the White/European American (EA hereafter), Black/African American (AA hereafter), and Hispanic/Latinx (Latinx hereafter) subgroups of the ABCD sample. The current study focused on individual rule breaking, peer rule breaking, and alcohol sips given their prognostic associations with substance use in adolescence and early adulthood, which remains a significant public health concern. Examination of genetic influences underlying developmental pathways in the emergence of alcohol use in diverse populations during early adolescence addresses a critical gap in the literature as past research has primarily used European decent samples during adolescence and adulthood.

We hypothesized that greater Risk-PGS would be associated with greater individual rule breaking and alcohol sips at both ages representing genetic main effects. We also hypothesized that greater Risk-PGS would be associated with higher peer rule breaking, representative of a rGE. We hypothesized that these genetic associations would be detected in the EA subgroup given that the only existing GWAS on risky behavior is in a European American sample and the limited portability of PGS based on GWAS of European samples to non-European samples (Martin et al., 2019). We hypothesized that in the AA and Latinx subgroups genetic effects would be nonsignificant or reduced in size. We also hypothesized that rule breaking, peer rule breaking, and alcohol sips would be positively associated within each age for all racial/ethnic subgroups. Finally, we hypothesized that there would be positive associations across these constructs over time, that is, transactions among individual rule breaking, peer rule breaking, and alcohol sips from age 11–12 to 12–13, within all racial/ethnic subgroups.

Method

Participants

Data for this project were drawn from the Adolescent Brain Cognitive Development (ABCD) Study (N = 11,875; youth = 9–10 years old at baseline; 47.8% female; 52.1% non-Hispanic White, 15.0% non-Hispanic Black, 20.3% Hispanic/Latinx, 2.1% Asian, and 10.5% other [e.g., biracial]). The ABCD Study is an ongoing study following youth across the United States on their behavioral, social, and neurocognitive development from the age 9–10 into their 20s.

Procedures

The ABCD recruited eligible children from a set of 21 nationally distributed study sites using probability sampling of schools within the study site areas with less than 10% of the final sample recruited using alternative procedures such as mailing lists, referrals, and twin registries (Garavan et al., 2018). As part of the ABCD study, parents provided consent for their participation and written informed consent for youth participants and youths provided assent. Baseline data collection occurred between September 1, 2016 and November 15, 2018. The ABCD study includes neurocognitive, mental and physical health assessments, environmental and cultural measures, structural and functional brain imaging, as well as whole genome genotyping and shares data with the research community through annual data releases via the NIMH Data Archive. Data used in the present study came from ABCD data release 4.0.

As part of the ABCD study, parents and children are assessed annually with smaller assessments occurring mid-year. The baseline assessment occurred when youth were 9–10 years old with annual assessments occurring at the 1 year follow up (youth aged 10–11), the 2 year follow up (youth aged 11–12), and the 3 year follow up (youth aged 12–13). The present study uses data from the 2-year follow up (hereafter referred to as Time 1 [T1]) and 3-year follow up (hereafter referred to as Time 2 [T2]) because the measure of peer rule breaking was first assessed in ABCD at the 2-year follow up (T1). The current subsample included youth who were White/European American (n = 5113; 47% female), Black/African American (n = 1159; 50% female), or Hispanic/Latinx (n = 1624; 48% female). Of this sample, youth were 11.55 years old on average at T1 and 12.42 years old at T2. The Indiana University Institutional Review Board approved the current study.

The ABCD study collected saliva samples from youth at the baseline assessment (see Uban et al., 2018). Samples were sent for genotyping to the Rutgers University Cell and DNA Repository (RUCDR). DNA was extracted from saliva samples and genotyped on the Smokescreen Genotyping Array (Baurley et al., 2016). RUCDR performed DNA quality controls based on calling signals and variant call rates, and the quality-controlled genotyping data contains 11,099 unique individuals with 516,598 genetic variants in the ABCD study. Imputation was performed via the TOPMed imputation server using mixed ancestry and Engle v2.4 phasing. Single nucleotide polymorphisms (SNPs) with a genotyping rate < 0.95 or that violated Hardy–Weinberg equilibrium (p < 10−6) or with minor allele frequency < 0.01 were excluded from analysis.

Measures

Polygenic Scores for Risky Behavior

We created our PGS based on summary statistics from Linnér et al. (2019). In this study, a principal components analysis was conducted across four risky behaviors: “automobile speeding propensity” (the tendency to drive faster than the speed limit), “drinks per week” (the average number of alcoholic drinks consumed per week), “ever smoker” (whether one has ever been a smoker), and “number of sexual partners” (the lifetime number of sexual partners). A GWAS was then performed on the first extracted principal component based in 315,894 individuals of European ancestry descent from the UK Biobank sample. Recent studies have created PGS based on this GWAS in independent samples, finding positive associations with substance use (Ksinan et al., 2022) and lifetime suicide attempts (Rabinowitz et al., 2019).

Using summary statistics from Linnér et al. (2019), we created Risk-PGS in ABCD sample based on 834,519 SNPs after filtering and matching discovery, target, and reference samples. Bayesian regression and continuous shrinkage method (PRS-CS) (Ge et al., 2019) was used to create genome-wide PGS. The PRS-CS method uses linkage disequilibrium (LD) information from 1000 Genomes Project reference panels with the matching ancestry and estimates the posterior effect sizes for SNPs in a given set of GWAS summary statistics. Empirical tests and simulations have shown improved predictive power above traditional methods of polygenic construction. Risk-PGS based on posterior PRS-CS weights were created using the score procedure in PLINK 1.9 (Chang et al., 2015).

Population Stratification and Genetic Admixture

Principal components analysis were conducted to represent population admixture in ABCD. ABCD participants are of diverse ancestry and were first assigned to empirically based ancestry groups using principal components derived from 1000 Genomes (phase 3) reference populations, as described by Peterson and colleagues (2017). Subsequently, within-ancestry principal components were calculated in each of subsamples to capture fine-grained stratification. The first 10 genetic ancestry principal components were extracted based on ancestry informative markers. These 10 genetic principal components were residualized from the Risk-PGS and the subsequent Risk-PGS were standardized for ease of interpretation.

Individual Rule Breaking at T1 and T2

Parents completed the Child Behavior Checklist (CBCL) (Achenbach & Ruffle, 2000) regarding their child’s behavior. Individual risky behavior was assessed using the CBCL rule-breaking subscale which consisted of 17 items (e.g., “my child breaks rules at home, school or elsewhere” and “my child is sets fires”) and assessed the degree to which the parent believed the item applied to their child on a three-point scale ranging from not true (scored as 0) to very true/often true (scored as 2) during the past 6 months. Items were summed so higher scores indicated higher levels of rule breaking. Internal consistency was good (alpha T1 = .73, alpha T2 = .73).

Peer Rule Breaking at T1 and T2

Youth completed the Youth Peer Behavior Profile regarding their peers’ behavior. Peer rule breaking was assessed using the 3-item Rule Breaking/Delinquent Peer scale (e.g., friends who skip school, shoplift, have been suspended from school) which is a subset of items from the Peer Behavior Profile/Peer Activities Questionnaire (Bingham, Fitzgerald & Zucker, 1995). Youth reported the proportion of their peers who engage in rule breaking behavior on a five-point scale ranging from none or almost none (scored as 0) to all or nearly all (scored as 5). Because only three items from a larger scale were included in ABCD assessments, their internal psychometric properties have not been validated. To address this, we used these items as indicators of a latent variable capturing peer rule breaking behavior. Latent variable properties are reported in the preliminary analyses section of the results.

Alcohol Sips at T1 and T2

The Timeline Follow-Back (TLFB) was used to assess quantitative estimates of substance use over time (Sobell & Sobell, 1996). Youth were asked if they had ever tried in their life ‘A sip of alcohol such as beer, wine or liquor’, which was captured as yes (scored as 1) or no (scored as 0).

Covariates

Participant age, sex, and family income were controlled for in the analysis. In addition, the first 10 principal components were residualized from the Risk-PGS in order to account for potential population stratification.

Analytic Approach

As part of preliminary analyses, study variables were examined for patterns of missingness, descriptive statistics, and bivariate correlations in SPSS v28. Significant differences among phenotypic study variables across racial/ethnic groups were examined using ANOVA and Tukey’s HSD post-hoc comparison. As part of variable creation, we examined the three peer rule breaking items as indicators of a latent variable within each wave using confirmatory factor analysis in Mplus v.8.8. A latent factor approach is advantageous in accounting for potential measurement error, and this is particularly good for measures with a small number of items.

For our main analyses, study variables were examined in a path model in Mplus v8.8 using the stratification command to account for study site and the cluster command to account for siblings within family. Figure 1 presents a conceptual model. Within the path model, the Risk-PGS was considered a predictor of individual rule breaking and alcohol sips at age T1 and T2, representing genetic main effects. The Risk-PGS was considered a predictor of peer rule breaking, representing rGE. Autoregressive paths within construct and cross-lagged paths across constructs were included over time. Correlations were included within time between rule breaking, peer rule breaking, and alcohol sips at age T1. Rule breaking and peer rule breaking were correlated at age T2 and considered concurrent predictors of alcohol sips at age T2. This was examined separately in EAs, AAs, and Latinx subgroups. To note, genetic effects are not statistically comparable across racial/ethnic subgroups due to genetic differences across ancestral groups (e.g., allele frequency, LD patterns), which can lead to genetic indices such as PGS operating differently within each group (Martin et al., 2017). Thus, we did not statistically compare genetic associations across racial/ethnic groups within our models.

Figure 1. Conceptual Model.

Figure 1.

Open in a new tab

Figure 1 represents a conceptual model of genetic main effects, rGE, and phenotypic effects among study variables. Solid black lines represent phenotypic effects. Doubled black lines from the risk polygenic score to rule breaking and alcohol sips at T1 and T2 represent genetic main effects. Solid gray lines from the risk polygenic score to peer rule breaking at T1 and T2 represent rGE effects.

In supplement to our main model, we ran three sensitivity analyses to examine the robustness of effects. In the first sensitivity analysis, we excluded items from the CBCL rule breaking subscale that assessed alcohol use (1 item) and affiliation with deviant peers (2 items) to ensure that associations between individual rule breaking with peer rule breaking and alcohol sips were due to unique effects rather than measurement overlap. In the second sensitivity analysis, we examined our main model split by gender within each racial/ethnic subgroup given some evidence that genetic effects and alcohol use can vary by gender (Salvatore, Cho, & Dick, 2017). Differences across gender were probed using the Satorra-Bentler adjusted Chi-square difference test. In the third sensitivity analysis, we ran our main model excluding youth with missing data at T2 given study attrition in order to examine whether our results are biased due to attrition.

Results

Preliminary Analyses

There was missing data across study variables, largely due to study attrition from T1 to T2 (Risk-PGS: 6.8%; individual rule breaking T1, T2: 13.1%, 48.2%; peer rule breaking T1, T2: 20.2%, 53.3%; alcohol sips T1, T2: 13.6%; 47.7%). Participants with missing data at T2 compared to those not missing data at T2 were more likely to be female, be a racial/ethnic minority, have lower income, be older, and at T1 have greater individual rule breaking, greater peer rule breaking, and less alcohol sips. Full information maximum likelihood was used to handle missing data. Given these patterns of missingness, we ran sensitivity analyses examining our main model excluding those with missing data at T2 (see below) and we also comment on the generalizability of findings in the discussion.

A latent variable for peer rule breaking was examined using confirmatory factor analysis. The factor structure was found to have configural (X2 (24) = 214.69, p < .001), metric (X2 (32) = 226.30, p < .001), and scalar (X2 (40) = 315.55, p < .001) invariance across EAs, AAs, and Latinx. Model fit was improved by including correlations for the same item across time. Final peer rule breaking factor loadings were significant and adequate for T1 and T2 in all subgroups (EA: .43 to .73; AA: .40 to .75; Latinx: .40 to .74). These latent variables were subsequently used in the main path model.

Descriptive Statistics

Descriptive statistics can be found in Table 1. Average levels of individual rule breaking, peer rule breaking, and alcohol sips were low in all subgroups at both T1 and T2. AAs had higher individual rule breaking compared to EAs and Latinx at T1 and T2. At both T1 and T2, peer rule breaking was highest in AA and lowest in EA with Latinx scoring in the middle. At both T1 and T2, alcohol sips were highest in EAs, lowest in AAs with Latinx scoring in the middle.

Table 1.

Means, Standard Deviations, and Correlations Among Primary Study Variables for European Americans, African Americans, and Hispanics

European American 1 2 3 4 5 6 7 8 9 10
1. Risk-PGS 1
2. Individual Rule Breaking T1 .08*** 1
3. Peer Rule Breaking T1 .11*** .28*** 1
4. Alcohol Sips T1 .05** .04** .11*** 1
5. Individual Rule Breaking T2 .11*** .67*** .28*** .07*** 1
6. Peer Rule Breaking T2 .11*** .28*** .95*** .11*** .30*** 1
7. Alcohol Sips T2 .06** .04* .14*** .45*** .08*** .15*** 1
8. Age −.01 .02 .09*** .06*** .01 .09*** .10*** 1
9. Sex .00 −.11*** −.11*** −.03* −.10*** −.11*** .01 .00 1
10. Income −.07*** −.20*** −.20*** .07*** −.17*** −.19*** .09*** .04** .01 1
M --- 0.95a −0.09ab 0.13ab 0.97a −0.09ab 0.16ab --- --- ---
SD --- 1.69 0.30 0.34 1.76 0.30 0.36 --- --- ---
African American 1 2 3 4 5 6 7 8 9 10
1. Risk-PGS 1
2. Individual Rule Breaking T1 .09** 1
3. Peer Rule Breaking T1 −.01 .18*** 1
4. Alcohol Sips T1 .03 .07* .07* 1
5. Individual Rule Breaking T2 .09* .69*** .17*** −.02 1
6. Peer Rule Breaking T2 .00 .18*** .96*** .07* .18*** 1
7. Alcohol Sips T2 .05 −.02 .10** .29*** .00 .16*** 1
8. Age −.02 .01 .05 .04 .05 .04 .05 1
9. Sex .09*** −.09** −.14*** −.02 −.06 −.13*** .04 −.01 1
10. Income −.02 −.10** −.24*** .06* −.07 −.23*** .05 .06* 0.02 1
M --- 1.35ab 0.29ac 0.04ac 1.32ab 0.28ac 0.04ac --- --- ---
SD --- 2.23 0.60 0.19 2.14 0.56 0.20 --- --- ---
Latinx 1 2 3 4 5 6 7 8 9 10
1. Risk-PGS 1
2. Individual Rule Breaking T1 .06** 1
3. Peer Rule Breaking T1 .05* .24*** 1
4. Alcohol Sips T1 .07** .12*** .07** 1
5. Individual Rule Breaking T2 .09** .54*** .19*** .04 1
6. Peer Rule Breaking T2 .06** .22*** .95*** .06** .21*** 1
7. Alcohol Sips T2 .04 .05 .15*** .26*** .07* .17*** 1
8. Age .01 .00 .07** .01 .02 .07** .09** 1
9. Sex −.03 −.09*** −.09*** −.03 −.12*** −.08** .02 −.01 1
10. Income .03 −.07** −.15*** .08** −.03 −.15*** .09** .04 −.02 1
M --- 1.05b 0.03bc 0.07bc 1.07b 0.04bc 0.10bc --- --- ---
SD --- 1.77 0.41 0.25 1.74 0.42 0.29 --- --- ---
Open in a new tab

Note. PGS = Polygenic Score. Sex was coded as 1 = male, 2 = female. Peer rule breaking is extracted latent factor scores. Point-biserial correlations were estimated for binary variables. Significant mean differences across European American, African American, and Hispanic subgroups are indicated by a common superscript based on Tukey’s HSD post-hoc comparison at p < .05.

***

p < .001

**

p < .01

*

p < .05.

Among EAs, the Risk-PGS was positively associated with individual rule breaking, peer rule breaking and alcohol sips at T1 and T2 (see Table 1). Individual rule breaking, peer rule breaking, and alcohol sips were positively associated within and across T1 and T2. In AAs, the Risk-PGS was associated with rule breaking at T1 and T2 as well as alcohol sips at T2 (see Table 1). Individual rule breaking, peer rule breaking, and alcohol sips were positively associated within and across T1 and T2, although individual rule breaking at T1 and T2 were not associated with alcohol sips at T2. In Latinx, the Risk-PGS was associated with individual rule breaking and peer rule breaking at T1 and T2, and alcohol sips at T1 (see Table 1). Individual rule breaking, peer rule breaking, and alcohol sips were positively associated within and across T1 and T2, although individual rule breaking at T1 was not associated with alcohol sips at T2.

Results from the Main Path Model

The main model had adequate fit (X2 (127) = 346.80, p < .001, RMSEA = .026, CFI = .96, TLI = .93). Full model results are presented in Table 2. Within the EA subgroup, greater Risk-PGS was associated with higher individual rule breaking and peer rule breaking at both times as well as alcohol sips at age T1. Rule breaking, peer rule breaking, and alcohol sips were positively associated at T1. Rule breaking and peer rule breaking were positively associated with one another from T1 to T2. There was continuity within rule breaking, peer rule breaking, and alcohol sips from T1 to T2, respectively. Rule breaking at T1 was negatively associated with alcohol sips at T2, whereas peer rule breaking at T2 was positively associated with alcohol sips at T2.

Table 2.

Main Path Model Results

Rule Breaking T1 Peer Rule Breaking T1 Alcohol Sips T1 Rule Breaking T2 Peer Rule Breaking T2 Alcohol Sips T2
European American B p-value B p-value B p-value B p-value B p-value B p-value
Risk-PGS 0.08 < .001 0.12 < .001 0.05 < .001 0.03 0.027 0.07 < .001 0.02 0.289
Rule Breaking T1 0.25 < .001 0.06 < .001 0.64 < .001 0.14 < .001 −0.05 0.018
Peer Rule Breaking T1 0.13 < .001 0.08 0.03 0.56 < .001 −0.01 0.72
Alcohol Sips T1 0.03 0.05 0.04 0.09 0.41 < .001
Rule Breaking T2 0.20 < .001 0.03 0.31
Peer Rule Breaking T2 0.17 < .001
Age 0.02 0.11 0.12 < .001 0.05 < .001 0.00 0.77 0.07 < .001 0.06 < .001
Sex −0.11 < .001 −0.13 < .001 −0.03 0.02 −0.03 0.07 −0.07 < .001 0.04 0.01
Income −0.18 < .001 −0.26 < .001 0.07 < .001 −0.03 0.14 −0.02 0.54 0.06 < .001
African American
Risk-PGS 0.09 0.003 −0.02 0.625 0.04 0.293 0.01 0.827 0.03 0.509 0.00 0.929
Rule Breaking T1 0.15 < .001 0.04 0.218 0.75 < .001 0.09 0.20 0.00 0.973
Peer Rule Breaking T1 0.08 0.081 0.06 0231 0.52 < .001 −0.18 0.081
Alcohol Sips T1 −0.03 0.213 0.04 0.381 0.26 0.001
Rule Breaking T2 0.14 0.051 −0.09 0.125
Peer Rule Breaking T2 0.41 0.003
Age 0.04 0.199 0.09 0.014 0.03 0.343 0.02 0.498 0.06 0.366 0.02 0.707
Sex −0.08 0.011 −0.15 < .001 −0.03 0.424 −0.02 0.385 0.01 0.809 0.03 0.475
Income −0.09 0.003 −0.26 < .001 0.05 .099 0.02 0.607 −0.15 0.015 0.08 0.109
Latinx
Risk-PGS 0.06 0.012 0.06 0.086 0.06 0.012 0.05 0.104 0.06 0.161 −0.01 0.85
Rule Breaking T1 0.23 < .001 0.09 0.006 0.57 < .001 0.07 0.188 −0.01 0.891
Peer Rule Breaking T1 0.09 0.038 −0.01 0.782 0.55 < .001 −0.08 0.192
Alcohol Sips T1 0.00 0.972 0.01 0.849 0.23 < .001
Rule Breaking T2 0.17 0.005 0.01 0.842
Peer Rule Breaking T2 0.25 0.002
Age −0.02 0.497 0.06 0.068 −0.01 0.659 −0.01 0.721 0.06 0.112 0.07 0.022
Sex −0.09 < .001 −0.13 < .001 −0.03 0.216 −0.07 0.008 0.00 0.92 0.06 0.057
Income −0.10 < .001 −0.21 < .001 0.06 0.011 −0.01 0.872 −0.06 0.165 0.07 0.029
Open in a new tab

Note. PGS = Polygenic score. Significant effects at p < .05 are bolded. Sex was coded as 1 = male, 2 = female.

Within the AA subgroup, the Risk-PGS was positively associated with rule breaking at T1 (see Table 2). Rule breaking and peer rule breaking were positively associated at T1, but no cross-lagged associations were detected. There was continuity within rule breaking, peer rule breaking, and alcohol sips from T1 to T2, respectively. Peer rule breaking at T2 was positively associated with alcohol sips at T2.

Within the Latinx subgroup, the Risk-PGS was positively associated with rule breaking and alcohol sips at T1 (see Table 2). Rule breaking and peer rule breaking were positively associated within T1 and T2. Individual and peer rule breaking at T1 were both associated with alcohol sips at T1. There was continuity within rule breaking, peer rule breaking, and alcohol sips from T1 to T2, respectively. Peer rule breaking at T2 was associated with alcohol sips at T2.

Sensitivity Analyses

Sensitivity analyses using the alternate individual rule breaking composite excluding alcohol and deviant peer items was largely the same as the main model within each subgroup (see Supplemental Table 1). Within the EA subgroup, the association between the Risk-PGS and rule breaking at T2 and the association between rule breaking and alcohol sips at T1 became nonsignificant, although the direction of effects remained the same. No associations became nonsignificant in either the AA or Latinx subgroups.

Sensitivity analyses within males and females were largely the same as the main model within each racial/ethnic subgroup (see Supplemental Table 2). In EA females vs. males, there were associations between the Risk-PGS and peer rule breaking at T2 as well as between rule breaking at T1 and alcohol sips at T1; however, these associations were not significantly different (X2difference (1) = .35, p = .55; X2difference (1) = .08, p = .78). In AA males vs. females, there was an association between the Risk-PGS and rule breaking at T1 and estimates were significantly different (X2difference (1) = 3.94, p = .047). In AA females vs. males, there was an association between alcohol sips at T1 to T2 which was significantly different (X2difference (1) = 4.46, p = .035) but the association between peer rule breaking and alcohol sips at T2 was not (X2difference (1) = .12, p = .72). Finally, there were several associations that were significant in Latinx males which were not significant in females but none were significantly different, such as from the Risk-PGS to rule breaking at T1 (X2difference (1) = 2.25, p = .132), from the Risk-PGS to peer rule breaking at T1 (X2difference (1) = 2.34, p = .126), between rule breaking and peer rule breaking at T1 (X2difference (1) = 3.55, p = .06) as well as alcohol sips at T1 (X2difference (1) = 2.74, p = .098), and between rule breaking and peer rule breaking at T2 (X2difference (1) = 2.42, p = .12).

Sensitivity analyses excluding those with missing data at T2 (EA n = 2989; AA n = 457; Latinx n = 805) were similar to those in the full sample, but overall, partially attenuated in magnitude likely due to reduced sample size and statistical power (see Supplemental Table 3). In addition, in EAs the association between the Risk-PGS and rule breaking at T2 became nonsignificant. In AAs the association between rule breaking and peer rule breaking at T1 became nonsignificant. In Latinx, the association between the Risk-PGS and both rule breaking and alcohol sips at T1 became nonsignificant.

Discussion

Accumulating evidence indicates that genetic influences can contribute to both individual and peer delinquent and externalizing behaviors in adolescence, leading to substance use later in life (Elam et al., 2022). Little research has examined genetic influences on individual rule breaking behavior or peer rule breaking behaviors across early adolescence and the vast majority of research has been in European ancestry populations. The current study addresses this gap by examining polygenic predisposition for risky behavior as predictive of individual rule breaking, peer rule breaking, and alcohol sips across early adolescence using racially/ethnically diverse samples from the ABCD study.

Partial support was found for our hypothesis that the Risk-PGS would be positively associated with individual rule breaking and alcohol sips in the EA subgroup, representing genetic main effects. In the EA subgroup, the Risk-PGS was positively associated with individual rule breaking at T1 and T2 and alcohol sips at T1. Conversely, there was no association between the Risk-PGS and alcohol sips at T2 which may be because there was a low prevalence and we accounted for T1 alcohol sips, indicating the Risk-PGS was not associated with change in alcohol sips across this period. In the same regard, this indicates that the Risk-PGS was associated with change in individual rule breaking across this period.

We also found partial support for our hypothesis that genetic associations in AAs or Latinx would be nonsignificant or reduced in size. The Risk-PGS was positively associated with rule breaking at T1 in the AA and Latinx subgroups, and alcohol sips at T1 within the Latinx subgroup. This may indicate that although we used a GWAS from an EA sample that there is overlap in genetic variants associated with risky behavior in AAs and Latinx. More broadly, this indicates genetic main effects on rule breaking. The genetic effects on alcohol sips in EAs and Latinx may be because alcohol sips in early adolescence represents a risky behavior or that our Risk-PGS captured some shared genetic etiology underlying broader externalizing behavior. The presence of multiple genetic associations in the EA subgroup were hypothesized as this aligns with the European descent discovery GWAS that was used to form PGS. Large GWAS on risky behavior do not currently exist for AAs or Latinx, which likely contributed to the limited predictive power of the Risk-PGS in these groups. This highlights the need for large discovery GWAS in diverse samples and new polygenic approaches to better represent diverse populations in genetics research.

We also hypothesized that the Risk-PGS would be associated with peer rule breaking in EAs, indicative of a rGE, which was supported by the finding that greater Risk-PGS was associated with greater peer rule breaking at T1 and T2. Thus, the Risk-PGS was associated with change in peer rule breaking across this period. Recent research has found polygenic predisposition for indices of externalizing behavior to be associated with peer behavior, including peer rejection and peer substance use (Elam et al., 2022; Kuo et al., 2021). This study is unable to discern which type of rGE is occurring but both evocative and active are theoretically plausible (Elam, Lemery-Chalfant, & Chassin, 2022). For example, greater genetic predisposition for risky behavior could evoke greater rule breaking in peers, indicating an evocative rGE. This may occur as youth align their behavior with those around them through learning and reinforcement (Reiter et al., 2019). Alternatively, youth may choose to affiliate with peers who are more delinquent based on their own behavior, in part, based on genetic predisposition for risky behavior, indicating an active rGE.

The limited genetic effects and lack of rGEs in AA and Latinx subgroups can be interpreted in a few ways. The relatively smaller sample sizes in AA and Latinx subgroups could contribute to a lack of genetic effects. It may also be that the limited portability of our Risk-PGS contributed to fewer associations with individual rule breaking and alcohol sips but also with peer rule breakings. Alternatively, AAs and Latinx face greater structural inequalities and are subject to higher rates of some contextual risk factors including discrimination, residential segregation, and limited access to health resources which can increase risk for externalizing behaviors and alcohol use (Scott, 2017). Thus, environmental effects not captured in the current study may be more salient, masking rGE effects (Dick et al., 2007). This highlights the need for genetic research in diverse samples that capture a range of social and contextual risk and protective factors.

There was partial support for our hypotheses that associations would be found among rule breaking, peer rule breaking, and alcohol sips within time in all racial/ethnic groups. In EAs and Latinx there were positive associations among individual rule breaking, peer rule breaking, and alcohol sips at T1 and T2, except in the Latinx subgroup individual rule breaking and alcohol sips at T2 were not associated. In AAs, individual rule breaking and peer rule breaking were positively associated at T1 as were peer rule breaking and alcohol sips at T2. This supports evidence that these risky behaviors are associated with one another, and that individual and peer delinquent behavior can increase risk for alcohol (Mundt, 2011; Kelly et al., 2012).

Across time, we found continuity in individual rule breaking, peer rule breaking, and alcohol sips in all racial/ethnic subgroups, indicating these behaviors are relatively stable across early adolescence. Conversely, we only found evidence of transactions across constructs over time in EAs. Individual rule breaking and peer rule breaking were positively associated across time with one another and individual rule breaking at T1 was positively associated with alcohol sips at T2. This is supported by previous research indicating that individual and peer delinquent behavior transact over time due to peer contagion of behavior within peer groups, and that early delinquent behavior, such as rule breaking, can lead to alcohol use (Mundt, 2011; Kelly et al. 2012). The lack of transactions over time in the AA and Latinx subgroup may be to other environmental influences not captured in this study such as buffering effects of parenting or the disproportionate effects of structural inequalities and contextual risk factors that some AAs and Latinx experience (Scott, 2017).

Integrating across hypotheses, our findings indicate that greater genetic predisposition for risky behavior contributes to greater rule breaking in early adolescence, that there is continuity in individual and peer rule breaking over time, and that concurrent peer rule breaking is associated with alcohol sips a year later. In addition, in EAs we found evidence that genetic main effects and rGEs may underlie transactions between individual and peer rule breaking over time and contribute to alcohol sips in early adolescence. Thus, genetic predisposition for risky behavior increases both individual rule breaking and may also influence youth to join deviant peer groups or evoke more rule breaking within peer groups. Over time, transactions between individual and peer rule breaking increase risk for alcohol sips.

These findings were fairly robust in sensitivity analyses. In EAs the association between the Risk-PGS and rule breaking at T2 became nonsignificant in the model with the alternate composite variable for individual rule breaking and when excluding those with missing data at T2. Additionally, in the Latinx subgroup when excluding those with missing data at T2 there were no genetic effects detected, likely due to a loss of power. In AAs the association between the Risk-PGS and rule breaking at T1 was greater in magnitude in males vs. females. Collectively, these findings bear replication given the scarcity of genetic studies in early adolescence in diverse samples. In addition, those with missing data at T2 compared to those not missing data had lower income, were older, more likely to be a racial/ethnic minority and had greater rule breaking and peer rule breaking at T1 which may limit generalizability of findings. Racial and ethnic minorities are subject to higher rates of some contextual risk factors such as discrimination, racism, and limited access to healthcare resources, which, collectively, may be more salient than genetic effects during sensitive developmental periods (Dick et al., 2007; Scott, 2017).

This study has important strengths including investigation of genetic influences on individual and peer rule breaking behavior in developmental pathways to alcohol use over early adolescence in diverse populations. We used an advanced Bayesian method, PRS-CS, to create polygenic scores. This was examined in a nationally representative sample using multiple reporters. We examined the main model across various sensitivity analyses and found effects to be fairly robust. Limitations were also present, which include the relatively smaller sample sizes in AAs and Latinx which could have contributed to detection of fewer significant effects. In addition, large GWAS on risky behavior in AAs and Latinx do not currently exist. We leveraged the largest existing GWAS on risky behavior from a European descent sample to create polygenic scores. Polygenic scores suffer from known issues of portability across race/ethnicity (Martin et al., 2017), and although we detected genetic effects in all subgroups, there was no evidence of rGE in AAs and Latinx. This highlights the need for future large GWAS in diverse samples. Another limitation is the large amount of missing data at T2. We attempted to address this by using full information maximum likelihood and through sensitivity analyses, but results bear replication. In addition, more advanced modeling methods such as random intercepts-cross lagged panel models are able to parse between and within individual effects. However, these models are only applicable when three waves of data are available making this a useful endeavor for future studies examining these constructs with more waves of data. Finally, given this was a nationally representative sample of youth in early adolescence, there were relatively low rates of individual rule breaking, peer rule breaking, and alcohol sips so results may not generalize to a high-risk sample.

Conclusion

Genetic influences can increase risk for alcohol use in adolescence by contributing to individual rule breaking via genetic main effects and peer rule breaking through evocative and/or active rGEs. These effects may contribute to transactions among individual and peer rule breaking over time, increasing risk for alcohol use in early adolescence. This has primarily been investigated in European ancestry populations during adolescence with less of a focus on younger ages. The current study found that across EA, AA, and Latinx subgroups, genetic influences contributed to individual rule breaking whereas peer rule breaking primarily contributed to alcohol sips. Evidence of rGE was evident only in EAs. These findings can help guide universal preventive interventions for alcohol use by targeting peer rule breaking in early adolescence. Further, these findings add to our understanding of developmental genetics in early adolescence by identifying how genetic influences contribute to specific individual and peer rule breaking behaviors in developmental pathways to alcohol sips.

Supplementary Material

Supplemental Materials

References

  1. Achenbach TM, & Ruffle TM (2000). The Child Behavior Checklist and related forms for assessing behavioral/emotional problems and competencies. Pediatrics in Review, 21(8), 265–271. 10.1542/pir.21-8-265 [DOI] [PubMed] [Google Scholar]
  2. Aiken A, Clare PJ, Boland VC, Degenhardt L, Yuen WS, Hutchinson D, Najman J, McCambridge J, Slade T, McBride N, De Torres C, Wadolowski M, Bruno R, Kypri K, Mattick RP, & Peacock A (2020). Parental supply of sips and whole drinks of alcohol to adolescents and associations with binge drinking and alcohol-related harms: A prospective cohort study. Drug and alcohol dependence, 215, 108204. 10.1016/j.drugalcdep.2020.108204 [DOI] [PubMed] [Google Scholar]
  3. Bingham CR, Fitzgerald HE, & Zucker RA (1995). Peer Behavior Profile/Peer Activities Questionnaire Unpublished questionnaire. Department of Psychology, Michigan State University. [Google Scholar]
  4. Campbell MC, & Tishkoff SA (2008). African genetic diversity: implications for human demographic history, modern human origins, and complex disease mapping. Annual review of genomics and human genetics, 9, 403–433. 10.1146/annurev.genom.9.081307.164258 [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Chartier K, & Caetano R (2010). Ethnicity and health disparities in alcohol research. Alcohol Research and Health, 33(1–2), 152–160. [PMC free article] [PubMed] [Google Scholar]
  6. Derringer J, Corley RP, Haberstick BC, Young SE, Demmitt BA, Howrigan DP, ... & McQueen MB. (2015). Genome-wide association study of behavioral disinhibition in a selected adolescent sample. Behavior genetics, 45(4), 375–381. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Dick DM (2007). Identification of genes influencing a spectrum of externalizing psychopathology. Current Directions in Psychological Science, 16(6), 331–335. 10.1111/j.1467-8721.2007.00530. [DOI] [Google Scholar]
  8. Dodge KA, Malone PS, Lansford JE, Miller S, Pettit GS, & Bates JE (2009). A dynamic cascade model of the development of substance-use onset. Monographs of the Society for Research in Child Development, 74(3), vii–119. 10.1111/j.1540-5834.2009.00528.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Donovan JE, & Molina BS (2008). Children’s introduction to alcohol use: sips and tastes. Alcoholism, Clinical and Experimental Research, 32(1), 108–119. 10.1111/j.1530-0277.2007.00565.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Elam KK, Lemery-Chalfant K, & Chassin L (2022). A gene-environment cascade theoretical framework of developmental psychopathology. Journal of Psychopathology and Clinical Science. Advance online publication. 10.1037/abn0000546 [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Elam KK, Chassin L, Lemery-Chalfant K, Pandika D, Wang FL, Bountress K, Dick D, & Agrawal A (2017). Affiliation with substance-using peers: Examining gene-environment correlations among parent monitoring, polygenic risk, and children’s impulsivity. Developmental Psychobiology, 59(5), 561–573. 10.1002/dev.21529 [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Elam KK, Chassin L, & Pandika D (2018). Polygenic risk, family cohesion, and adolescent aggression in Mexican American and European American families: Developmental pathways to alcohol use. Development and Psychopathology, 30(5), 1715–1728. 10.1017/S0954579418000901 [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Elam KK, Bountress KE, Ha T, Shaw DS, Wilson MN, Aliev F, Dick DM, & Lemery-Chalfant K (2022). Developmental genetic effects on externalizing behavior and alcohol use: Examination across two longitudinal samples. Development and Psychopathology, 1–10. Advance online publication. 10.1017/S0954579422000980 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Elam KK, Clifford S, & Ruof A (2022) Genotype-environment correlation by intervention effects underlying middle childhood peer rejection and associations with adolescent marijuana use. Development and Psychopathology 34:171–182. 10.1017/S0954579420001066 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Elam KK, Ha T, Neale Z, Aliev F, Dick D & Lemery-Chalfant K (2021). Polygenic effects on alcohol use in African Americans and European Americans: Age varying associations from adolescence to adulthood. Scientific Reports, 11, 22425. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Garavan H, Bartsch H, Conway K, Decastro A, Goldstein RZ, Heeringa S, Jernigan T, Potter A, Thompson W, & Zahs D (2018). Recruiting the ABCD sample: Design considerations and procedures. Developmental Cognitive Neuroscience, 32, 16–22. 10.1016/j.dcn.2018.04.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Ge T, Chen CY, Ni Y, Feng YA, & Smoller JW (2019). Polygenic prediction via Bayesian regression and continuous shrinkage priors. Nature communications, 10(1), 1776. 10.1038/s41467-019-09718-5 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Gelernter J, Kranzler HR, Sherva R, Almasy L, Koesterer R, Smith AH, Anton R, Preuss UW, Ridinger M, Rujescu D, Wodarz N, Zill P, Zhao H, & Farrer LA (2014). Genome-wide association study of alcohol dependence: significant findings in African- and European-Americans including novel risk loci. Molecular Psychiatry, 19(1), 41–49. 10.1038/mp.2013.145 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Haynie DL, & Osgood DW (2005). Reconsidering peers and delinquency: How do peers matter?. Social Forces, 84, 1109–1130. [Google Scholar]
  20. Hoeben EM, Meldrum RC, Walker DA, & Young JT (2016). The role of peer delinquency and unstructured socializing in explaining delinquency and substance use: A state-of-the-art review. Journal of Criminal Justice, 47, 108–122. [Google Scholar]
  21. Jackson KM, Colby SM, Barnett NP, & Abar CC (2015). Prevalence and correlates of sipping alcohol in a prospective middle school sample. Psychology of Addictive Behaviors, 29(3), 766–778. 10.1037/adb0000072 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Karlsson Linnér R, Biroli P, Kong E, Meddens SFW, Wedow R, Fontana MA, Lebreton M, Tino SP, Abdellaoui A, Hammerschlag AR, Nivard MG, Okbay A, Rietveld CA, Timshel PN, Trzaskowski M, Vlaming R. de, Zünd CL, Bao Y, Buzdugan L, … & Beauchamp JP. (2019). Genome-wide association analyses of risk tolerance and risky behaviors in over 1 million individuals identify hundreds of loci and shared genetic influences. Nature Genetics, 51(2), 245–257. 10.1038/s41588-018-0309-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kelly AB, Chan GC, Toumbourou JW, O’Flaherty M, Homel R, Patton GC, & Williams J (2012). Very young adolescents and alcohol: evidence of a unique susceptibility to peer alcohol use. Addictive Behaviors, 37(4), 414–419. 10.1016/j.addbeh.2011.11.038 [DOI] [PubMed] [Google Scholar]
  24. Ksinan AJ, Smith RL, Barr PB, & Vazsonyi AT (2022). The associations of polygenic scores for risky behaviors and parenting behaviors with adolescent externalizing problems. Behavior Genetics, 52(1), 26–37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Kuo SI, Salvatore JE, Barr PB, Aliev F, Anokhin A, Bucholz KK, Chan G, Edenberg HJ, Hesselbrock V, Kamarajan C, Kramer JR, Lai D, Mallard TT, Nurnberger JI Jr, Pandey G, Plawecki MH, Sanchez-Roige S, Waldman I, Palmer, AA., Externalizing Consortium, … & Dick DM. (2021). Mapping pathways by which genetic risk influences adolescent externalizing behavior: The interplay between externalizing polygenic risk scores, parental knowledge, and peer substance use. Behavior Genetics, 51(5), 543–558. 10.1007/s10519-021-10067-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Martin AR, Gignoux CR, Walters RK, Wojcik GL, Neale BM, Gravel S, Daly MJ, Bustamante CD & Kenny EE (2017). Human demographic history impacts genetic risk prediction across diverse populations. American Journal of Human Genetics, 100(4), 635–649. 10.1016/j.ajhg.2017.03.004 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Martin AR, Kanai M, Kamatani Y, Okada Y, Neale BM, & Daly MJ (2019). Clinical use of current polygenic risk scores may exacerbate health disparities. Nature Genetics, 51(4), 584–591. 10.1038/s41588-019-0379-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Miles DR, van den Bree MB, Gupman AE, Newlin DB, Glantz MD, & Pickens RW (2001). A twin study on sensation seeking, risk taking behavior and marijuana use. Drug and Alcohol Dependence, 62(1), 57–68. [DOI] [PubMed] [Google Scholar]
  29. Mulia N, Ye Y, Greenfield TK, & Zemore SE (2009). Disparities in alcohol-related problems among white, black, and Hispanic Americans. Alcoholism: Clinical and Experimental Research, 33(4), 654–662. 10.1111/j.1530-0277.2008.00880.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Mundt MP (2011). The impact of peer social networks on adolescent alcohol use initiation. Academic Pediatrics, 11(5), 414–421. 10.1016/j.acap.2011.05.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Murphy MA, Dufour SC, & Gray JC (2021). The association between child alcohol sipping and alcohol expectancies in the ABCD study. Drug and Alcohol Dependence, 221, 108624. 10.1016/j.drugalcdep.2021.108624 [DOI] [PubMed] [Google Scholar]
  32. Plomin R, DeFries JC, & Loehlin JC (1977). Genotype-environment interaction and correlation in the analysis of human behavior. Psychological Bulletin, 84, 309–322. 10.1037/0033-2909.84.2.309 [DOI] [PubMed] [Google Scholar]
  33. Peterson RE, Edwards AC, Bacanu SA, Dick DM, Kendler KS, & Webb BT (2017). The utility of empirically assigning ancestry groups in cross-population genetic studies of addiction. The American Journal on Addictions, 26(5), 494–501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Popejoy AB, & Fullerton SM (2016). Genomics is failing on diversity. In Nature (Vol. 538, Issue 7624) (pp. 161–164). Nature Publishing Group. 10.1038/538161a [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rabinowitz JA, Jin J, Kahn G, Kuo SIC, Campos A, Renteria M, ... & Cohen D. (2021). Genetic propensity for risky behavior and depression and risk of lifetime suicide attempt among urban African Americans in adolescence and young adulthood. American Journal of Medical Genetics Part B: Neuropsychiatric Genetics, 186(8), 456–468. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Ragan DT, Osgood DW, & Kreager DA (2022). Impulsivity, Peers, and Delinquency: A Dynamic Social Network Approach. Journal of Quantitative Criminology, 1–34.36532607 [Google Scholar]
  37. Reiter AM, Suzuki S, O’Doherty JP, Li SC, & Eppinger B (2019). Risk contagion by peers affects learning and decision-making in adolescents. Journal of Experimental Psychology: General, 148(9), 1494. [DOI] [PubMed] [Google Scholar]
  38. Sacks JJ, Gonzales KR, Bouchery EE, Tomedi LE,&Brewer RD. (2015). 2010 national and state costs of excessive alcohol consumption. American Journal of Preventive Medicine, 49, e73–e79. 10.1016/j.amepre.2015.05.031 [DOI] [PubMed] [Google Scholar]
  39. Salvatore JE, Cho SB, & Dick DM (2017). Genes, environments, and sex differences in alcohol research. Journal of Studies on Alcohol and Drugs, 78(4), 494–501. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Savi-Çakar F, Tagay Ö, & Ebru Ikiz F (2015). Risky behaviors of adolescents: Definitions and prevntion. In Columbus AM (Ed.) Advances in psychology research (Vol. 106) (pp. 137–167). Nova Science. [Google Scholar]
  41. Scarr S & McCartney K (1983). How people make their own environments: A theory of genotype environment effects. Child Development, 54, 424–435. 10.1111/j.1467-8624.1983.tb03884.x [DOI] [PubMed] [Google Scholar]
  42. Scott MS (2017). Commentary: Perspectives on alcohol-related gene and environment interplay in diverse populations. American Journal on Addictions, 26(5), 526–531. 10.1111/ajad.12584 [DOI] [PubMed] [Google Scholar]
  43. Su J, Kuo SI, Bucholz KK, Edenberg HJ, Kramer JR, Schuckit M, & Dick DM (2018). Understanding mechanisms of genetic risk for adolescent internalizing and externalizing problems: The mediating role of parenting and personality. Twin Research and Human Genetics: The Official Journal of the International Society for Twin Studies, 21(4), 310–321. 10.1017/thg.2018.36 [DOI] [PubMed] [Google Scholar]
  44. Sobell LC & Sobell MB (1992). Timeline follow-back: a technique for assessing self-reported alcohol consumption. In Litten RZ, Allen JP (Eds). Measuring Alcohol Consumption: Psychosocial and Biological Methods (pp. 41–72). Humana Press. [Google Scholar]
  45. Su J, Kuo SI, Aliev F, Chan G, Edenberg HJ, Kamarajan C, McCutcheon VV, Meyers JL, Schuckit M, Tischfield J, & Dick DM (2021). The associations between polygenic risk, sensation seeking, social support, and alcohol use in adulthood. Journal of Abnormal Psychology, 130(5), 525–536. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Vrieze SI, McGue M, Miller MB, Hicks BM, & Iacono WG (2013). Three mutually informative ways to understand the genetic relationships among behavioral disinhibition, alcohol use, drug use, nicotine use/dependence, and their co-occurrence: Twin biometry, GCTA, and genome-wide scoring. Behavior Genetics, 43(2), 97–107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  47. Wang XT, Zheng R, Xuan YH, Chen J, & Li S (2016). Not all risks are created equal: A twin study and meta-analyses of risk taking across seven domains. Journal of Experimental Psychology: General, 145(11), 1548. [DOI] [PubMed] [Google Scholar]
  48. Williams DR, Mohammed SA, Leavell J, & Collins C (2010). Race, socioeconomic status, and health: Complexities, ongoing challenges, and research opportunities. In Annals of the New York Academy of Sciences (Vol. 1186) (pp. 69–101). Blackwell Publishing Inc. 10.1111/j.1749-6632.2009.05339.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Young SE, Friedman NP, Miyake A, Willcutt EG, Corley RP, Haberstick BC, & Hewitt JK (2009). Behavioral disinhibition: liability for externalizing spectrum disorders and its genetic and environmental relation to response inhibition across adolescence. Journal of Abnormal Psychology, 118(1), 117. [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

Supplemental Materials
NIHMS1937214-supplement-Supplemental_Materials.docx (46.6KB, docx)

RESOURCES