Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Mar 5.
Published in final edited form as: Psychol Addict Behav. 2026 Mar 2;40(3):266–280. doi: 10.1037/adb0001133

Reciprocal Associations Between Adolescent Extracurricular Activity Availability and Engagement and Alcohol-Related Cognitions and Norms: Moderation by Socioeconomic Status

Samuel N Meisel a, Ella Diab b, Samuel F Acuff c, Kristina M Jackson d
PMCID: PMC12958451  NIHMSID: NIHMS2142569  PMID: 41770165

Abstract

Objective:

The current study examined reciprocal associations between structured and unstructured extracurricular activity access and engagement and reasons against drinking and descriptive drinking norms in a predominantly alcohol naïve adolescent sample. Socioeconomic status (SES) was tested as a moderator of these bidirectional associations, considering access to and benefits of structured activities may vary by SES.

Method:

Using six waves of data from a sample of early to middle adolescents (N=1023, 52% female, 76% White, 5% Black, 12% Latine, 28% from urban school districts), preregistered latent growth curve models with structured residuals tested reciprocal associations and whether they varied by SES.

Results:

The relative availability of structured activities to total extracurricular activities (i.e., structured and unstructured) decreased across time, whereas relative engagement remained constant. Reasons against drinking decreased and descriptive norms increased across time. There was no support for preregistered bidirectional hypotheses for extracurricular access or engagement and descriptive norms. At the between-person level, adolescents who continued participating in structured activities had more reasons against drinking in early adolescence and showed slower declines in these reasons over time. Contrary to our hypotheses, for most waves, when adolescents had greater engagement in structured activities than their typical levels, they had significantly lower reasons against drinking at the next wave. Despite growth trajectories differing across SES, there was no evidence for SES moderation.

Conclusions:

Findings highlight the importance of distinguishing between- and within-person effects when studying extracurricular activities, as well as better capturing the interpersonal processes that occur during extracurricular activities.

Keywords: extracurricular activities, reasons against drinking, descriptive drinking norms, adolescence, socioeconomic status

Introduction

Delaying adolescent alcohol use initiation is a critical public health concern, as early adolescent (i.e. 10–14 years-old) drinking is associated with myriad negative outcomes, including altered brain development and increased risk for future alcohol use and mental health disorders (Magid & Moreland, 2014). Theoretical models of alcohol use and positive youth development suggest that greater access to and engagement in structured extracurricular activities play a protective role in reducing alcohol use risk (Acuff et al., 2023; Benson, 2002; Farb & Matjasko, 2012). Although reasons against drinking and descriptive drinking norms are salient predictors of adolescent drinking (or lack thereof; Meisel & Colder, 2020; Merrill et al., 2016a), the bidirectional influence of extracurricular activity availability and engagement on these cognitions remains untested. Moreover, access to, and potential protective effects of, extracurricular activities may not be uniform across socioeconomic status (SES) levels (Baldwin & O’Flaherty, 2018; Morris, 2015). Using reciprocal determinism as a theoretical basis (Bandura, 1978), the current study used a longitudinal sample of predominantly alcohol naïve early adolescents to examine reciprocal associations between extracurricular activity access and engagement and both reasons against drinking and descriptive drinking norms across six waves. Examining these associations may inform prevention programs for youth that often seek to increase access to and engagement in structured activities prior to the period of greatest drinking risk.

Extracurricular Activities and Alcohol Use

Over 80% of adolescents aged 10–15 report never having consumed a full drink of alcohol (Johnston et al., 2023; SAMHSA, 2024). Considering the low rates of use during this developmental period, examining theoretically informed antecedents of alcohol use, such as descriptive norms and reasons against drinking, can facilitate prevention efforts (Trucco & Hartmann, 2021). Multiple theoretical perspectives provide a framework for understanding how extracurricular activities relate to adolescent alcohol use (Acuff et al., 2023; Benson, 2002; Catalano & Hawkins, 1996). Positive youth development theory argues that structured activities – supervised, and goal-oriented pursuits which typically involve formal rules or a set schedule (e.g., organized sports, clubs), provide adolescents with opportunities for positive engagement and development (Benson, 2002). Specifically, the opportunities to have prosocial peer interactions (e.g., formation of high-quality friendships and sharing of norms and cognitions that do not support risk behaviors like alcohol use) under adult supervision and monitoring in structured activities are conceptualized to minimize the likelihood of engaging in alcohol use behaviors (Benson, 2002). Similarly, the Social Development Model (Catalano & Hawkins, 1996) argues that strong bonds to conventional institutions such as schools, community centers, and religious organizations (i.e., structured activities) are protective, whereas poorer institutional bonds (i.e., more time spent engaging in unstructured activities such as going to parties, playing video games, spending time on social media, watching TV) increase alcohol use risk. Collectively, these theories suggest that structured activities decrease and unstructured activities increase the risk for adolescent alcohol use.

Consistent with these theoretical models, structured activity engagement has largely demonstrated a protective effect on adolescent alcohol and other substance use (e.g., Adachi-Mejia et al., 2014; Bone et al., 2025; McCabe et al., 2016; Spillane et al., 2021; Timonen et al., 2021). In part, these protective effects are due to organized activities improving positive affect (DesRoches & Willoughby, 2014) and facilitating positive peer connections (Eisman et al., 2018). Of note, sports participation shows mixed effects on adolescent alcohol and other substance use, with sport activities that are formally organized appearing to be protective factor (for a review, see Walczak et al., 2023). In contrast to structured activities, studies examining overall engagement in unstructured activities (Meldrum & Leimberg, 2018; Spillane et al., 2020) as well as engagement in specific unstructured activities like hanging out with friends (Hoeben et al., 2020) and using social media (Curtis et al., 2018) demonstrate positive associations with adolescent alcohol use.

To date, less work has examined how extracurricular activity availability relates to adolescent alcohol use. A cross-sectional study of American Indian adolescents demonstrated that the availability of extracurricular activities was associated with lower substance use (Moilanen et al., 2014). Using two waves of data from the same study used for the current secondary analysis, Spillane et al. (2020) found that greater availability of structured activities was unrelated to risk of any cannabis use and heavy drinking across adolescence, whereas greater availability of unstructured activities was positively associated with cannabis use and heavy drinking. Understanding how activity availability is related to alcohol use may help refine prevention strategies since prevention implications may differ depending on whether structured activities are unavailable or available but minimally engaged in.

Alcohol-Related Cognitions, Norms, and Alcohol Use

The theory of planned behavior argues that attitudes and cognitions (e.g., reasons against drinking) and norms (e.g., descriptive drinking norms) play key roles in the initiation of behaviors like alcohol use (Ajzen, 1991). Reasons against drinking refer to anti-drinking attitudes and beliefs that reflect an individual’s desire to abstain from alcohol, such as religious proscriptions or family upbringing, or to limit drinking, such as preferring to feel in control and interference with sports performance (Chassin & Barrera, 1993). Multiple studies have demonstrated protective effects of reasons against drinking on adolescent alcohol use (Anderson et al., 2013; Anderson et al., 2011; Chassin & Barrera, 1993; Georgeson et al., 2024). Descriptive drinking norms, which are perceptions of peer alcohol use, are robust predictors of adolescent alcohol use behaviors and are a core prevention target (normative feedback interventions; e.g., Brooks-Russell et al., 2014; Jones et al., 2017; Meisel & Colder, 2020; Voogt et al., 2012). Reflecting the increased risk for alcohol use initiation during adolescence, descriptive norms increase and reasons against drinking decrease over the course of adolescence (Meisel & Colder, 2020; Merrill et al., 2016a).

Reciprocal Determinism

Reciprocal determinism, a tenet of Social Learning Theory (Bandura, 1978), argues that social environments in which adolescents spend their time both influence and are influenced by alcohol-related cognitions (Bandura, 1978). Greater availability of and engagement in structured extracurricular activities, relative to unstructured activities, may help lower descriptive drinking norms and increase reasons against drinking through interactions with prosocial peers and adults and less exposure to deviant peers, who may provide access to alcohol (Eisman et al., 2018; Mahoney et al., 2005). Lower perceptions of peer alcohol use and greater reasons against drinking may lead youth to seek out and spend time in social activities consistent with these norms and attitudes (i.e., structured activities) and avoid unstructured activities where alcohol use can be discussed or modeled. Examining reciprocal associations between extracurricular activity availability and engagement and alcohol-related cognitions may inform developmental changes in two core targets of prevention programming.

Socioeconomic Status

Access to and engagement in structured activities are not equally distributed across SES levels (Baldwin & O’Flaherty, 2018). Adolescents from low, relative to moderate or high, SES backgrounds have fewer opportunities to engage in structured activities (Tandon et al., 2021) and lower engagement in structured activities (Baldwin & O’Flaherty, 2018; Snellman et al., 2014). Resource compensation theory argues that the benefits from structured activities may differ as a function of SES (Morris, 2015). According to this and similar perspectives (e.g., the resilience and differential effectiveness perspectives; Roth & Brooks-Gunn, 2016; Steinberg & Simon, 2019), despite lower access and engagement, adolescents from low SES backgrounds may benefit most from structured activity availability and engagement, as these activities provide adult support and skill development that moderate or high SES youth often gain outside extracurriculars (Morris, 2015). To date, there is some evidence for moderation by SES with these findings demonstrating increased benefits from structured activity engagement among low SES youth (Mahoney & Stattin, 2000; Marsh & Kleitman, 2002; Randall & Bohnert, 2009), although moderation has not been supported in several studies (Fredricks & Eccles, 2006; Khoddam et al., 2018; Larson et al., 2006). More rigorous testing of the protective benefits across SES levels may inform which adolescents benefit most from structured activity access and engagement.

SES may also moderate associations between descriptive norms, reasons against drinking, and extracurricular activity availability and engagement. The social cognitive theory of social class suggests that individuals from a high SES background, who are more individualistic, may be less likely to align their behaviors with norms and cognitions than individuals from a low SES background, who are more communal and interdependent (Kraus et al., 2011). In contrast, differential opportunity theory suggests the differences in extracurricular affordances across SES levels alter the ease in which youth can act on risk (e.g., descriptive norms) and protective (e.g., reasons against drinking) norms and cognitions. For example, an adolescent with many reasons for why they should not drink may be better able to select extracurricular contexts supportive of these beliefs (e.g., structured activities) at high, relative to low SES levels, as structured activities are likely more readily accessible to them. Adolescents from a low SES background, in contrast, may be better able to select into more unstructured contexts, where others may be drinking. To date, very few studies have examined whether SES moderates associations between health-related norms and cognitions and behaviors (Schuz et al., 2020; Sherman et al., 2022).

Current Study

Using six waves of data spanning early to middle adolescence, the current study examines reciprocal associations between extracurricular activity availability and engagement and reasons against drinking and descriptive norms, and how these associations may differ across SES levels, using latent growth curve models with structured residuals (LCM-SR; Curran et al., 2014). LCM-SR simultaneously examines associations between latent growth factors (i.e., between-person associations) and cross-lagged associations (i.e., within-person associations). LCM-SR improve the accuracy of cross-lagged associations when constructs are changing (Curran & Hancock, 2021; Curran et al., 2014), which is important since structured activity and reasons against drinking decline while unstructured activity and descriptive norms increase during adolescence (Denault & Poulin, 2019; Fredricks & Eccles, 2006; Meisel & Colder, 2020; Merrill et al., 2016a). We examine reciprocal associations after controlling for lifetime alcohol sipping, lifetime consumption of a full drink, school setting (e.g., urban vs. suburban vs. rural), and neighborhood control and reciprocal exchange. We explicitly do not control for race as prior differences with respect to race on extracurricular activity engagement likely reflect structural inequalities associated with racism (access to financial and neighborhood resources; Baldwin & O’Flaherty, 2018).

The following pre-registered hypotheses were tested: (1) At the between-person level, greater initial levels of the availability of and engagement in structured activities, relative to unstructured activities, will be associated with greater initial levels of reasons against drinking (i.e., intercept-intercept associations; see Figure 1). Growth in the availability of and engagement in structured, relative to unstructured, activities will be associated with slower decreases in reasons against drinking (i.e., slope-slope associations). (2) At the within-person level, waves when youth have greater availability of and engagement in structured, relative to unstructured, activities than average will be associated with greater reasons against drinking at the next wave. Waves when youth have greater reasons against drinking, than average, will be associated with greater availability of and engagement in structured, relative to unstructured, activities at the next wave. (3) At the between-person level, greater initial levels of the availability of and engagement in structured activities, relative to unstructured activities, will be associated with lower initial levels of descriptive drinking norms (i.e., intercept-intercept associations). Growth in the availability of and engagement in structured, relative to unstructured, activities will be associated with slower increases in descriptive drinking norms (i.e., slope-slope associations). (4) At the within-person level, waves when youth have greater availability of and engagement in structured, relative to unstructured, activities than average will be associated with lower descriptive norms at the next wave. Waves when youth have greater descriptive norms, than average, will be associated with lower availability of and engagement in structured, relative to unstructured, activities at the next wave. Given the limited work and mixed evidence examining SES as a moderator, there were no a priori-hypotheses regarding the moderational effects of SES.

Figure 1.

Figure 1.

Open in a new tab

Hypothesized associations between the relative availability of and engagement in structured activities and reasons against drinking and descriptive drinking norms.

Note. RA =r elative availability of structured activities to total activities, RE = relative engagement in structured activities to total activities, RAD = reasons against drinking, DN = descriptive drinking norms.

Methods

Sample

Adolescents (N=1023; 52% female sex assigned at birth), recruited from six middle schools, were enrolled through five cohorts at six-month intervals from 2009–2011. Three schools were suburban (n=508), two rural (n=231) and one an urban inner-city school (n=284). Consent return rates (34% suburban to 47% rural) and enrollment rates (79% urban to 94% suburban) varied across school settings. The average age at enrollment was 12.2 (SD=0.98, range=10–15). The racial composition of the sample was 76% White, 5% Black, 8% mixed race, 12% other race, and 12% self-identified as Latine. The sample was largely representative of the schools from which they were drawn, and there was some evidence the sample was more racially diverse but less socioeconomically disadvantaged (Jackson et al., 2021).

Procedure

Participants completed visits for Wave (W) 1 to W5 every 6 months, and W6 was roughly one year after W5. At the onset of the study, participants completed a two-hour in-person group orientation held after school, which included the baseline survey. Participants completed subsequent follow-up surveys electronically which took roughly 45 minutes to complete. Compensation was $25 for the baseline survey and $20 for each follow-up. The Brown University institutional review board approved all study procedures.

Retention rates were high across W2-W6 (range=83%−92%). Chi-square tests for categorical variables and analysis of variance tests for continuous variables examined whether baseline characteristics differed among adolescents who did not complete follow-up surveys. There were no significant differences for age, school setting, structured activity engagement, or neighborhood social control. Across all waves, male (p’s<.05, φ range=0.07–0.09) and non-White (p’s<.01, φ range=0.07–0.16) participants were significantly more likely to have missing data. Adolescents with missing data at W2-W6 were significantly more likely to sip alcohol (p’s<.05, φ range=0.06–0.12) and have a full drink in their lifetime (p’s<.05, φ range=0.09–0.12). For reasons against drinking, adolescents who did complete W2-W4 or W6 had significantly lower baseline reasons against drinking (p’s<.05, η2 range=0.00–0.02). Adolescents with missing data at W4-W6 (p’s <.05, all η2=0.01) had significantly greater unstructured activity engagement at W1. Participants who did not complete W2 and W4-W6 surveys (p’s <.05, all η2=0.01) had significantly greater descriptive drinking norms at baseline. Missing data at W2 (p=.031, η2=0.00) and W4 (p=.004, η2=0.01) was associated with significantly lower SES. Neighborhood reciprocated exchange was significantly lower among adolescents with missing W2 and W4-W6 data (p’s <.05, η2 range=0.01). Overall, although there were multiple significant differences with respect to missing data, these effects were often below the threshold for a small effect size or just met cutoffs for small effects (φ small effect=0.10, η2 small effect=0.01).

Measures

Structured and Unstructured Extracurricular Activity Availability and Engagement (W1-W6):

Seventeen items were derived from the original 25-item Adolescent Reinforcement Survey Schedule (ARSS; Holmes et al., 1987), and the worded adjusted for a younger demographic (the original ARSS was developed for college students in the 1980s). An additional 8 items, such as playing computer or video games and participating in after-school clubs, were incorporated to ensure the measure’s relevance (Spillane et al., 2020). We excluded sexual pleasure and substance use activities to accommodate these new items. Participants were asked to “Rate the frequency of each item over the past month,” and response choices varied from “Less than once a week” (0) to “More than once a day” (5), with additional options for “I don’t have the opportunity to do this” (6) and “I have the opportunity to do this, but I choose not to” (7). Items were characterized based on whether they were structured (7 items) or unstructured (18 items) based on the definition of structured and unstructured activities proposed by Mahoney and Stattin (2000). To evaluate the availability of activities, items were dichotomized into available, collapsing across categories (response options 0–5 and 7=1) and not available (response option 6= 0). To evaluate activity engagement, responses were coded as follows: 0=“I have the opportunity to do this, but I choose not to”, 1=“less than once a week,” 2=“once a week,” 3=“2–4 times per week,” 4=“5–6 times a week,” 5=“every day,” and 6=”more than once a day.”

Extracurricular activity availability and engagement were operationalized as: (1) The relative availability of structured activities, which reflects a proportion of the sum total of structured activities available divided by the sum total of structured and unstructured available activities, and (2) relative engagement in structured activities, which reflects the average engagement in structured activities divided by the average engagement in all activities, structured and unstructured (Acuff et al., 2019; Herrnstein, 1974).

Reasons against drinking (W1-W6):

The 12-item reasons for abstaining and limiting drinking measure was adapted from Chassin and Barrera (1993) and Greenfield et al. (1989). The 12-item scale, including items such as “I’ve seen the negative effects of someone else’s drinking,” “drinking interferes with my studies,” and “I was brought up not to drink,” was answered using a 4-point Likert scale ranging from not true (1) to true and very important (4). The internal consistency of an overall average composite demonstrated good internal consistency (ω range=.82–.92), consistent with prior work supporting a reliable single factor this measure (Merrill et al., 2016b).

Descriptive Drinking Norms (W1-W6).

Descriptive norms were assessed using a measure developed by Wood et al. (2004) but adapted to middle and high schoolers. Adolescents were asked to rate the alcohol use frequency (10-point Likert-scale ranging from doesn’t drink [0] to twice a day or more [9]) and quantity (open-ended response) of same-age and biological sex students in their grade during the school year. The quantity item was recoded as 0=they don’t drink to 4=more than 3 drinks, as prior work with this variable indicated it is non-normally distributed. Two questions assessed adolescent perceptions of same-age and biological-sex students in their grade past month heavy episodic drinking (i.e., three or more drinks in a row), one during the school year and one during the summer. Both questions were answered on a 7-point Likert scale ranging from none (1) to (6) ten time or more. An average of these four variables demonstrated good internal consistency (ω range=.86–90).

Covariates

Alcohol use (W1-W6).

At each wave, participants reported whether they had ever had a sip or a full drink of alcohol, not including consumption as part of a religious service. Given evidence that sipping alcohol and consuming a full drink are distinct milestones (Jackson et al., 2015), variables were created to indicate whether the participant reported a lifetime sip or full drink at any point during the study.

Neighborhood Social Capital (W1):

Neighborhood social capital was assessed using eight items taken from Community Survey of the Project on Human Development in Chicago Neighborhoods (Sampson, 1997). This caregiver-report measure contains two subscales: Informal social control scale (three items) and the reciprocated exchange scale (five items). The informal social control items included the prompt “What is the likelihood that your neighbors can be counted on to intervene in the following situations?” and caregivers answered questions such as “If children were skipping school and hanging out on the street” on a 5-point Likert scale (0=very unlikely to 4=very likely). For the reciprocated exchange scale, sample items included “When a neighbor is not at home, how often do you and other neighbors watch over their property?” and “How often do you and other people in the neighborhood ask each other advice about personal things such as childrearing or job openings?” and these questions were answered on a Likert scale ranging from never (0) to often (3). Variables from each of these subscales were averaged to create the informal social control (ω=.91) and reciprocated exchange scales (ω=.89).

School Setting (W1):

Of the six schools participants were recruited from, one school was in an urban district, three schools were in suburban districts, and two schools were from rural districts. Schools were dummied coded as urban, suburban, and rural (reference group).

SES (W1):

Parents reported on their annual household income (from 1=<$5,000 to 9=$150,000), whether their child qualified for a federally sponsored free or reduced lunch at school, and their highest level of education (from 1= <high school to 5=postgraduate degree). These three variables were standardized and aggregated into a single SES scale (Roberts et al., 2017; ω=.86).)

Data Analytic Plan

The analytic plan for the current study was preregistered and can be found at https://osf.io/g64mj. Data preparation and descriptive statistics were conducted in SAS 9.4 (SAS Institute Inc, 2013). Descriptive statistics included examining mean differences in structured and unstructured activity access and engagement across SES levels as well as univariate latent growth curve models across SES levels. LCM-SR were estimated in a sequential model building approach recommended by Curran et al. (2014). First, unadjusted univariate growth curve models were estimated. Multiple forms of growth (e.g., linear, quadratic) were compared using chi-square difference tests (for normally distributed outcome variables modeled using Maximum Likelihood estimation [FIML]) or the Satorra-Bentler chi-square difference test (for non-normally distributed outcome variables modeled using Maximum Likelihood Robust estimation [MLR]; Satorra & Bentler, 2001). Estimation using either FIML or MLR allowed for the inclusion of all participants, even those with missing data. After identifying the best fitting growth models, step two involved specifying structured residuals and using chi-square difference tests to determine whether the inclusion of structured residual autoregressive paths led to a significant increment in model fit and whether constraining these paths to be equal across time led to a non-significant decrement in model fit.

In the third step, four bivariate LCM-SR were estimated: (1) relative availability of structured activities and reasons against drinking, (2) relative availability of structured activities and perceived drinking norms, (3) relative engagement in structured activities and reasons against drinking, and (4) relative engagement in structured activities and descriptive norms. Nested model tests examined whether the models supported the inclusion of covariances across intercepts and slopes, within-time residual covariances, and reciprocal effects. When supported by nested model tests, parameters were constrained to be equal across time. All models controlled for baseline age, biological sex, lifetime sipping, lifetime full drink, SES, school setting (i.e., urban, suburban, rural) and neighborhood social control and reciprocated exchange.

To date, there are no clear guidelines for conducting continuous moderation within an LCM-SR framework. Therefore, SES was converted into a categorical variable. We created three groups, low SES (<25th percentile), moderate SES (25th-75th percentile), and high SES (>75th percentile) based on scores on our standardized SES indicator. Consistent with our preregistered analytic approach, a multiple group LCM-SR examined whether bidirectional between- and within-person associations across the four bivariate models differed as a function of SES level. As with the bivariate LCM-SR, nested model tests were conducted to determine whether between-person covariances, within-person within-time covariances, and within-person lagged associations differed across SES levels.

For all LCM-SR, the Comparative Fit Index (CFI), Tucker-Lewis Index (TLI), Root Mean Square Error Approximation (RMSEA), and Standardized Root-Mean-Square Residual (SRMR) were used to evaluate model fit. Specific cut-offs for assessing “good” fit cannot be generalized across all models (Marsh et al., 2004), therefore, ranges were used to determine model fit acceptability (for CFI and TLI, <0.90 is poor, 0.90 to 0.94 is acceptable, and >0.95 is excellent; for RMSEA, 0.08 ≥ is poor, 0.05 to 0.07 is acceptable, and <0.05 is excellent; and for SRMR, ≥ 0.09 is poor, 0.06 to 0.09 is acceptable, and <0.06 is excellent). All data and LCM-SR syntax are available at https://osf.io/8wgz4/.

Results

Descriptive Statistics

The vast majority of adolescents (71.6%) reported never having a full drink of alcohol in their lifetimes across waves. Two-thirds of the sample (66.2%) reported sipping or tasting alcohol in their lifetime. In the low SES group (n=235), 100% of the adolescents received free or reduced lunch, parents’ average highest education level was less than high school, and had an average household income between $10,000-$14,999. In the moderate SES group (n=425), 24% received free or reduced lunch, parents’ average education was an associate’s or technical school degree and had an average household income between $50,000-$74,999. In the high SES group (n=277), 0% of the adolescents received free or reduced lunch, parents’ average education was a college or university degree, and their average income was between $100,000-$149,999. The Supplemental Materials contains mean comparisons of the relative availability of and engagement in structured activities across SES levels.

Measurement Invariance

Detailed information regarding measurement invariance models can be found in the Supplemental Materials. Overall, there was support for configural and metric variance and partial scalar and residual variance for the relative availability and engagement in structured activities as well as reasons against drinking and descriptive drinking norms.

Growth Trajectories and Univariate LCM-SR

Table 1 provides model fit information for the univariate LCM-SR and the Supplemental Materials provide detailed information regarding model selection (e.g., growth form) as well as model-implied growth trajectories across SES levels. A linear growth model provided the best fir for the relative availability of structured activities, which significantly declined across waves for the overall sample (M=−.01, p=.042) as well as for the high SES group (M=−.01, p=.001), but did not significantly change for the low (M=−0.01, p=.189) or moderate SES groups (linear M=−0.03, p=.064, quadratic M=0.004, p=.096; see Figure 2). A piecewise growth model provided the best fit for relative engagement in structured activities, which did not significantly change in the overall sample (W1-W4 M=−0.002, p=.129; W4-W6 M=0.002, p=.166). There was nonsignificant change in the relative engagement in structured activities for low (M=−0.003, p=.273), moderate (W1-W4 M=−.04, p=.113; W4-W6 M=.03, p=.157), and high SES groups (W1-W4 M=0.01, p=.640; W4-W6 M=0.03, p=.142). A quadratic growth model provided the best fit for reasons against drinking, which significantly declined across the study for the overall sample (linear M=−0.08, p<.001, quadratic M=0.003, p=.146), as well as for adolescents with low (M=−0.06, p<.001), moderate (linear M=−0.09, p<.001; quadratic M=0.003, p=.388), and high SES (W1-W4 M=−0.05, p<.001; W4-W6 M=−0.08, p<.001). A quadratic growth model provided the best fit for descriptive drinking norms, which significantly increased across the study for the overall sample (linear M=0.06, p=.003, quadratic M=.02, p<.001), as well as for adolescents with low (W1-W4 M=.11, p=.001; W4-W6 M=.24, p<.001), moderate (linear M=0.14, p=.002, quadratic M=.02, p<.001), and high SES (W1-W4 M=0.08, p<.001; W4-W6 M=0.27, p<.001).

Table 1.

Univariate, bivariate, and multiple group LCM-SR fit

Model χ2 df p SCF RMSEA CFI TLI SRMR
Univariate
Structured Activity Availability 17.72 16 .340 2.44 0.01 0.99 0.99 0.07
Structured Activity Engagement 11.99 12 .446 1.42 0.00 1.00 1.00 0.02
reasons against drinking 15.25 10 .123 1.41 0.02 0.99 0.99 0.05
Descriptive Norms 16.88 11 .111 1.57 0.02 0.99 0.99 0.03
Bivariate
Structured Activity Availability and Reasons against drinking 249.45 138 <.0001 1.35 0.03 0.95 0.93 0.06
Structured Activity Availability and Descriptive Norms 232.91 139 <.0001 1.31 0.03 0.96 0.95 0.05
Structured Activity Engagement and Reasons against drinking 225.39 120 <.0001 1.16 0.03 0.97 0.95 0.05
Structured Activity Engagement and Descriptive Norms 211.35 123 <.0001 1.17 0.03 0.98 0.96 0.05
Bivariate Multiple Group Analysis
Structured Activity Availability and Reasons against drinking 657.13 488 <.0001 1.14 0.03 0.92 0.92 0.09
Structured Activity Availability and Descriptive Norms 633.48 489 <.0001 1.14 0.03 0.94 0.94 0.08
Structured Activity Engagement and Reasons against drinking 574.93 463 .0003 1.10 0.03 0.96 0.96 0.08
Structured Activity Engagement and Descriptive Norms 586.76 473 .0003 1.11 0.03 0.96 0.96 0.08
Open in a new tab

Note. SCF=scaling correction factor, RMSEA=root mean square error of approximation, CFI=comparative fit index, TLI= Tucker-Lewis index, SRMR= standardized root mean square residual. The bivariate LCM-SR fit includes covariates whereas the univariate models do not include covariates.

Figure 2.

Figure 2.

Open in a new tab

Unconditional latent growth curve trajectories across SES levels. Dashed lines depict the 95% confidence for the latent growth trajectories.

Bivariate LCM-SR

Bivariate LCM-SR fit information can be found in Table 1. Given the large number of covariates in the bivariate LCM-SR, we summarize SES associations below and the remaining covariate associations in the Supplemental Materials (also see Table 2).

Table 2.

Covariate associations with for bivariate LCM-SR

Structured Activity Availability and Reasons Against Drinking
Availability Intercept Availability Linear Slope RAD Intercept RAD Linear Slope RAD Quadratic Slope
β p β p β p β p β p β p
Age −0.15 <.001 0.1 .126 - - 0.04 0.269 −0.14 .012 0.12 .023
Male Sex (0=Female) 0.05 .255 0.06 .369 - - −0.14 <.001 −0.05 .345 0.08 .132
Lifetime Alcohol Sip −0.04 .386 0.01 .906 - - 0.09 .040 −0.14 .014 0.10 .079
Lifetime Full Drink −0.16 .001 0.13 .104 - - −0.34 <.001 0.00 .941 −0.06 .288
Neighborhood Social Control −0.03 .620 −0.01 .890 - - −0.05 .330 0.06 .35 −0.06 .350
Neighborhood Reciprocated Exchange 0.05 .408 −0.05 .538 - - −0.01 .834 0.06 .353 −0.02 .785
Suburban School (0= Rural) 0.02 .647 −0.01 .841 - - −0.02 .756 −0.12 .072 0.12 .064
Urban School (0= Rural) −0.04 .531 −0.06 .545 - - −0.13 .017 −0.03 .708 0.06 .428
SES 0.19 .001 −0.06 .540 - - 0.19 <.001 0.17 .016 −0.20 .002
Factor Mean M=3.52, p<.001 M=−0.07, p=.110 - M=3.12, p<.001 M=0.38, p=.028 M=−0.07, p=.018
R 2 0.12 0.04 - 0.18 0.09 0.08
Structured Activity Availability and Descriptive Norms
Availability Intercept Availability Linear Slope Norms Intercept Norms Linear Slope Norms Quadratic Slope
β p β p β p β p β p β p
Age −0.15 <.001 0.10 .129 - - 0.22 <.001 0.49 <.001 −0.36 <.001
Male Sex (0=Female) 0.05 .244 0.06 .371 - - 0.12 .013 −0.3 .001 0.25 .011
Lifetime Alcohol Sip −0.04 .378 0.01 .915 - - 0.10 .026 0.08 .381 −0.04 .689
Lifetime Full Drink −0.16 .001 0.13 .103 - - 0.24 <.001 0.03 .771 0.13 .261
Neighborhood Social Control −0.02 .669 −0.02 .856 - - −0.04 .517 0.02 .891 0.01 .913
Neighborhood Reciprocated Exchange 0.05 .405 −0.05 .547 - - 0.05 .408 −0.12 .306 0.15 .254
Suburban School (0= Rural) 0.02 .659 −0.01 .836 - - −0.01 .774 −0.06 .529 0.02 .854
Urban School (0= Rural) −0.04 .497 −0.05 .590 - - 0.20 .002 0.28 .019 −0.33 .013
SES 0.18 .001 −0.05 .610 - - −0.26 <.001 0.18 .093 −0.17 .163
Factor Mean M=3.51, p<.001 M=−0.07, p=.113 - M=−0.76, p=.029 M=−1.25, p<.001 M=0.16, p<.001
R 2 0.12 0.04 - 0.34 0.41 0.26
Structured Activity Engagement and Reasons Against Drinking
Engagement Intercept Engagement Linear Slope W1-W4 Engagement Linear Slope W4-W6 RAD Intercept RAD Linear Slope RAD Quadratic Slope
β p β p β p β p β p β p
Age −0.14 <.001 0.18 .002 0.00 .995 0.04 .263 −0.14 .009 0.13 .016
Male Sex (0=Female) 0.10 .006 −0.05 .303 −0.13 .032 −0.14 <.001 −0.05 .397 0.07 .160
Lifetime Alcohol Sip −0.11 .006 −0.02 .742 0.01 .853 0.09 .04 −0.14 .014 0.10 .079
Lifetime Full Drink −0.13 .002 0.03 .681 −0.07 .335 −0.34 <.001 −0.01 .909 −0.06 .309
Neighborhood Social Control −0.08 .111 −0.03 .690 0.06 .398 −0.05 .307 0.06 .367 −0.06 .375
Neighborhood Reciprocated Exchange 0.11 .035 −0.08 .291 0.06 .484 −0.01 .835 0.07 .321 −0.02 .741
Suburban School (0= Rural) −0.02 .680 0.07 .265 −0.03 .594 −0.02 .750 −0.11 .089 0.11 .082
Urban School (0= Rural) −0.09 .110 −0.07 .408 0.20 .023 −0.13 .017 −0.02 .762 0.05 .463
SES 0.22 <.001 0.03 .640 0.19 .015 0.19 <.001 0.16 .019 −0.19 .002
Factor Mean M=6.56, p<.001 M=−0.58, p=.005 M=−0.03, p=.895 M=3.12, p<.001 M=0.40, p=.024 M=−0.07, p=.013
R 2 0.17 0.04 0.07 0.18 0.09 0.08
Structured Activity Engagement and Descriptive Norms
Engagement Intercept Engagement Linear Slope W1-W4 Engagement Linear Slope W4-W6 Norms Intercept Norms Linear Slope Norms Quadratic Slope
β p β p β p β p β p β p
Age −0.14 <.001 0.19 .001 0.00 .997 0.21 <.001 0.49 <.001 −0.37 <.001
Male Sex (0=Female) 0.11 .006 −0.05 .367 −0.13 .029 0.12 .012 −0.30 <.001 0.25 .010
Lifetime Alcohol Sip −0.11 .005 −0.01 .811 0.01 .921 0.10 .027 0.08 .376 −0.04 .682
Lifetime Full Drink −0.13 .002 0.03 .673 −0.06 .359 0.24 <.001 0.03 .790 0.13 .249
Neighborhood Social Control −0.08 .124 −0.02 .741 0.05 .426 −0.04 .522 0.02 .871 0.01 .953
Neighborhood Reciprocated Exchange 0.11 .034 −0.08 .285 0.06 .446 0.05 .400 −0.12 .296 0.16 .237
Suburban School (0= Rural) −0.02 .727 0.07 .269 −0.04 .556 −0.01 .769 −0.06 .501 0.03 .808
Urban School (0= Rural) −0.09 .118 −0.07 .428 0.19 .026 0.2 .002 0.28 .020 −0.33 .014
SES 0.22 <.001 0.03 .663 0.19 .015 −0.26 <.001 0.18 .084 −0.17 .154
Factor Mean M=6.56, p<.001 M=−−0.60, p=.003 M=−0.02, p=.908 M=−0.76, p=.028 M=−1.26, p<.001 M=0.16, p<.001
R 2 0.17 0.07 0.07 0.34 0.41 0.27
Open in a new tab

Note. RAD=reasons against drinking

Structured Activity Availability and Reasons Against Drinking.

At the between- and within-person levels, there were no significant associations between structured activity availability and reasons against drinking (see Supplemental Materials for detailed information regarding nested model test results). Greater SES levels were associated with significantly greater initial levels of structured activity availability and reasons against drinking. Adolescents with greater SES had significantly greater initial increases in reasons against drinking and then more rapid declines in reasons against drinking.

Structured Activity Availability and Descriptive Norms.

At the between- and within-person levels, there were no significant associations between structured activity availability and descriptive norms. SES levels were associated with significantly greater initial levels of structured activity availability and lower levels of descriptive norms.

Structured Activity Engagement and Reasons Against Drinking.

At the between-person level, greater initial levels of relative engagement in structured activities were significantly associated with greater initial levels of reasons against drinking (see Figure 3). The linear slope of structured activity engagement from W1-W4 was positively associated with the linear slope of reasons against drinking, suggesting that individuals with slower declines in structured activity engagement demonstrated greater increases in reasons against drinking. The linear slope of structured activity engagement from W1-W4 was negatively associated with the quadratic slope of reasons against drinking, such that adolescents with slower declines in structured activity involvement demonstrated less deceleration (i.e., a flatter curvature) in reasons against drinking across waves. The linear slope of structured activity engagement from W4-W6 was positively associated with the quadratic slope of reasons against drinking, indicating that adolescents who had slower declines in structured activity engagement during W4-W6 had slower downward accelerations (i.e., declines) in their reasons against drinking. At the within-person level, from W1 to W2, W2 to W3, and W4 to W5, and W5 to W6, when adolescents had greater relative engagement in structured activities than their typical levels, they had significantly lower reasons against drinking at the next wave. There was an opposite association from W3 to W4, such that when adolescents had greater relative engagement in structured activities than their typical levels, they had significantly greater reasons against drinking at the next wave. SES was positively associated with initial levels of relative engagement in structured activities and faster growth in relative engagement in structured activities from W4-W6. Adolescents with greater SES had significantly greater initial levels of reasons against drinking, initial increases in reasons against drinking, and then more rapid declines in reasons against drinking.

Figure 3.

Figure 3.

Open in a new tab

Relative Engagement in Structured Activities and Reasons Against Drinking LCM-SR

Note. RAD= reasons against drinking, RE = relative engagement in structured to total activities, W=wave. Only significant associations are depicted. Unstandardized coefficients are presented first and the standardized coefficient is presented second, after the slash. Model covariates, age, biological sex, lifetime alcohol sipping, lifetime full drink, neighborhood social control, neighborhood reciprocal exchange, and socioeconomic status are not depicted to reduce figure complexity.

Structured Activity Engagement and Descriptive Norms.

At the between- and within-person levels, there were no significant associations between structured activity engagement and descriptive norms. SES was negatively associated with initial levels of descriptive norms and positively associated with initial levels of relative engagement in structured activities and faster growth in relative engagement in structured activities from W4-W6.

Moderation by SES

Table 1 provides a summary of model fit information for the multiple group analyses comparing between and within-person associations across SES levels. Across all bivariate LCM-SR between the relative availability of and engagement in structured activities with reasons against drinking and descriptive norms, nested model tests indicated that constraining between- and within-person associations to be equal across SES did not result in a significant decrement in fit. Said otherwise, multiple group LCM-SR did not find evidence of moderation by SES at the between- or within-person level of analysis.

Discussion

The current study extended prior work by examining reciprocal between- and within-person associations between the relative availability and engagement in structured activities and reasons against drinking and descriptive drinking norms across six waves spanning early to middle adolescence. The results provide a nuanced picture of extracurricular activity-alcohol-related cognition and norms associations.

Growth Trajectories

Potentially due to increased competition for opportunities to participate or a decline in age-appropriate activities (Mahoney et al., 2005), the relative availability of structured activities significantly decreased across the study period. There was consistency (i.e., no change) in adolescent activity engagement across time. SES-specific analyses replicated prior work (Leventhal et al., 2015; Morris, 2015) by demonstrating that adolescents with high SES consistently had the highest levels of relative availability and engagement, whereas low SES adolescents consistently had the lowest levels of relative availability and engagement. For relative availability, these differences were driven by increased availability of structured and unstructured activities as SES level increased. Interestingly, whereas low SES adolescents had the least and high SES adolescents had the highest access to unstructured activities, low SES youth engaged the most, and high SES youth engaged the least in these activities. This trend suggests that increased access to, and as a result increased opportunity for engagement in structured activities can offset unstructured activity engagement, even when these activities are widely available.

The findings that reasons against drinking significantly decreased across early to middle adolescence and descriptive norms increased from early to middle adolescence replicated prior findings (Meisel & Colder, 2020; Merrill et al., 2016a). SES-specific analyses were the first to our knowledge to examine how these alcohol-related cognitions and norms differ as a function of SES level. SES-specific growth results indicating the lowest reasons against drinking and highest descriptive drinking norms by middle adolescence among adolescents with low and high SES, compared to adolescents with moderate SES, is consistent with prior research suggesting that both groups of youth may be at elevated risk for alcohol use through unique mechanisms (Luthar & Latendresse, 2005a, 2005b).

Extracurricular Activities and Reasons Against Drinking

Whereas there was no support for Hypotheses 1 and 2 (i.e., no between- or within-person associations) for relative availability of structured activities and reasons against drinking, there was mixed support for Hypotheses 1 and 2 for the relative engagement in structured activities and reasons against drinking. At the between-person level, individuals who continued to engage in activities such as organized sports or after-school activities (e.g., clubs) demonstrated more reasons against drinking alcohol and less deceleration (i.e., a flatter curvature) in reasons against drinking across waves. These findings align with prior work demonstrating protective effects of structured activities on adolescent substance use behaviors (Bone et al., 2025; Spillane et al., 2020; Spillane et al., 2021). Further, the current findings extend prior work by highlighting that structured activity engagement may facilitate stronger reasons against drinking during a developmental period marked by more positive alcohol attitudes (Merrill et al., 2016a; Treloar Padovano et al., 2020).

Whereas findings examining people who overall engaged in more structured activities were largely consistent with the prior literature and study hypotheses, findings examining a person’s variations over time in their structured, compared to total, activity engagement were predominantly in the opposite direction of our hypotheses. At W1, W2, W3, and W4, when adolescents had greater relative engagement in structured activities than their typical levels, they had significantly lower reasons against drinking at the next time point. Even in structured activities, conversations may occur between adolescents reflecting positive sentiments about alcohol (e.g., discussions about upcoming parties), which may reduce reasons against drinking. Some prior work suggests that the specific peers partaking in an extracurricular activity may influence whether structured activities promote or reduce alcohol use risk (Eisman et al., 2018; Mahoney et al., 2005). From W3 to W4, when adolescents had greater relative engagement in structured activities than their typical levels, they had significantly greater reasons against drinking at the next time point. It is unclear why there was a significant positive association at this one lag. Overall, the results between relative engagement in structured activities and reasons against drinking reflected Simpson’s Paradox (Kievit et al., 2013), where relative engagement in structured activities demonstrated protective between-person effects but risky within-person effects on reasons against drinking.

Extracurricular Activities and Descriptive Norms

There was no support for hypothesized between- and within-person associations between extracurricular activity availability and engagement and descriptive norms (Hypotheses 3 and 4). Zero-order within-time (r range= −.11 to −.05) and prospective correlations (r range=−.15 to −.03) between relative availability and descriptive norms and zero-order within-time (r range=−.14 to −.06) and prospective correlations (r range=−.14 to −.06) between relative engagement and descriptive norms (see Supplemental Materials for full correlation matrix) were significant across most time points. Albeit small in magnitude, the current findings demonstrate the importance of parsing within- and between-person effects, as correlations may not generalize to either level of analysis. These findings suggest that there may be no reciprocal associations between the relative availability and engagement in structured activities and descriptive norms across between- and within-person levels of analyses. One consideration to provide a more stringent test of these null findings will be to align reference groups in future research. The selection of reference group for descriptive norms measures has been associated with the magnitude of drinking misperceptions (i.e., over-perceiving the drinking behaviors of others) as well as the strength of the association between descriptive norms and drinking behaviors (Larimer et al., 2009; Neighbors et al., 2008). Using a more proximal reference group (e.g., peers participating on your sports team, peers who attend your community center, and peers who you play computer or video games with), rather than same-age and biological sex students in their grade, may better inform whether the relative availability and engagement in descriptive norms predict and are predicted by descriptive norms (Larimer et al., 2009).

Moderation by SES

Contrary to our expectations, SES did not moderate between- or within-person associations between the relative availability of or engagement in structured activities with descriptive drinking norms or reasons against drinking. Through disaggregating between- and within-person effects and examining reciprocal associations across six waves, the current study provides the most robust test to date of the resource compensation theory. Failure to find evidence of moderation builds on existing studies that have not found SES to moderate structured activity engagement and behavioral outcomes (Fredricks & Eccles, 2006; Khoddam et al., 2018; Larson et al., 2006). Importantly, and consistent with prior work (Baldwin & O’Flaherty, 2018; Snellman et al., 2014), adolescents from lower SES backgrounds had less access to and engagement in structured activities compared to high SES adolescents, as well as greater engagement in unstructured activities. These findings suggest that despite adolescents across SES levels experiencing different levels of structured and unstructured activity availability and engagement, the associations of these extracurriculars on alcohol-related cognitions are the same across SES.

Future Directions

Deepening our understanding of the interpersonal processes that occur during extracurricular structured and unstructured activities may facilitate more informative tests of the hypotheses analyzed in the current study (Acuff et al., 2019; Farb & Matjasko, 2012; Mahoney et al., 2005). Existing measures of extracurricular structured and unstructured activities, like the ARSS, provide limited information about what is occurring (or not occurring) within each of these activities. For example, an adolescent in band might engage in conversations during informal breaks about parties from the prior weekend or plans to drink the upcoming weekend. Another adolescent could affiliate with different friends in the same band club who do not promote alcohol use and discuss plans to see a concert over the weekend. Future measures of extracurricular activities and engagement that capture the network members and relational processes (e.g., positive peer influences, deviancy training) occurring during these activities may provide a more accurate understanding of how these activities relate to alcohol-related cognitions and alcohol behaviors.

Further, sum and mean scores on these measures place greater weight on the aggregate of extracurricular activities (i.e., total sum available or average total engagement) rather than the potential importance of a single extracurricular activity. Two teens could have the same average score on engagement in structured activities, with one teen having high engagement in a single activity and another teen having low engagement in several activities. Two teens could have the same score on engagement in unstructured activities, with one teen’s score driven by their time spent on their phone and social media, whereas another teen with the same score could be a function of engagement in multiple offline activities. To capture the potential influence of specific activities as well as being able to model extracurricular activity availability and engagement simultaneously, future work may benefit from adopting person-centered analytic approaches such as mixture modeling (Masyn, 2013). Mixture modeling would be able to identify subgroups of individuals based on their levels of availability and engagement in specific activities, which may provide additional nuance regarding the developmental significance of specific activities (e.g., structured sport participation, social media) and combinations of availability and engagement on adjustment. In addition to these improvements in the measurement and modeling of extracurricular activities, future work should also explicitly link reciprocal associations between extracurricular activities, norms, and reasons against drinking in early to middle adolescence to alcohol use occurring in middle to late adolescence, a time during which adolescents typically initiate and escalate alcohol use (Johnston et al., 2023).

Limitations

This study contained several limitations. First, extracurricular activity was modeled by accounting for the relative availability of structured activities and engagement in structured activities to the total availability and engagement of structured and unstructured activities. This approach was taken to provide a more parsimonious strategy for modeling structured and unstructured activities, however, unique associations for structured or unstructured activities with descriptive norms or reasons against drinking or for specific activities with these alcohol-related cognitions could not be evaluated. Accordingly, our findings help inform the risk and protective effects of extracurricular activity availability and engagement at the aggregate level, and cannot speak to the protective or risky nature of any specific structured or unstructured extracurricular activity. Second, and relatedly, although mixed (Barnes et al., 2006; Bohnert & Garber, 2007), there is some evidence suggesting that specific structured activities, like sports participation, are associated with increased substance use (Fauth et al., 2007; Murray et al., 2021). Third, there was age heterogeneity within study waves, and depending on the cohort, each wave varied on time of year (e.g., spring of fall depending on cohort). Associations between extracurricular activities and our outcomes may differ as a function of age or time of year. Fourth, our sample included 12–15-year-olds, on average, who were predominantly White and were recruited from the Northeast. Findings should not be generalized to other age periods and findings should be replicated in other geographic regions with a more racially diverse sample. Lastly, we had to create a categorical variable for SES in order to conduct our multiple group LCM-SR. SES exists along a continuum and forming a categorical variable from a continuous measure negatively impacts statistical power.

Conclusion

The present study examined how having access to and taking part in structured and unstructured activities across early through middle adolescence is associated with reasons for not drinking and teen perceptions of how common drinking is among their peers. reasons against drinking These associations were examined by comparing across people and comparing fluctuations within the same person over time as well as whether these associations differed as a function of SES. Overtime, adolescents’ alcohol cognitions and norms became riskier such that reasons against drinking reasons for not drinking decreased and descriptive drinking norms increased. Youth’s access to structured activities decreased yet their participation in these structured activities did not change. Adolescents who had more subtle decreases in their relative engagement in structured activities had greater initial increases and slower subsequent declines in reasons for not drinking. For all but one lagged association, times when an adolescent had greater than relative engagement in structured activities, compared to their overall average, was associated with lower reasons for not drinking. No other between- or within-person associations were found across models and there was no support for moderation by SES, despite different mean levels of structured and unstructured activities across SES levels. Overall, the findings provided limited support for bidirectional associations between extracurricular activities and alcohol-related cognitions and norms. The current study extended prior work through the disaggregation of between- and within-person effects and provided a more robust test of the effects of extracurricular activities on substance-related behaviors.

Supplementary Material

Supplemental text and graphics

Public Health Significance:

Engagement in structured extracurricular activities (e.g., organized sports) have been associated with reduced adolescent alcohol use, yet we know little about how these activities alter and are altered by alcohol-related cognitions and perceptions of peer alcohol use that often precede drinking. This study found that teens who engage in more structured activities hold stronger reasons against drinking that decrease less rapidly across adolescence, but when teens engage in more structured activities than their usual level, they have lower reasons against drinking roughly 6 months later in time. The findings suggest that in addition to focusing on access and engagement in structured activities, future research should consider the conversations and behaviors of teens in these settings.

Acknowledgments

This work was supported by grants from the National Institute on Alcohol Abuse and Alcoholism (PI Jackson: R01AA016838; PI Meisel: R00AA030030). The content is solely the responsibility of the authors and does not necessarily represent the official views or policy of the National Institutes of Health.

References

  1. Acuff SF, Dennhardt AA, Correia CJ, & Murphy JG (2019). Measurement of substance-free reinforcement in addiction: A systematic review. Clin Psychol Rev, 70, 79–90. 10.1016/j.cpr.2019.04.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Acuff SF, MacKillop J, & Murphy JG (2023). A contextualized reinforcer pathology approach to addiction. Nature Reviews Psychology. 10.1038/s44159-023-00167-y [DOI] [Google Scholar]
  3. Adachi-Mejia AM, Gibson Chambers JJ, Li Z, & Sargent JD (2014). The relative roles of types of extracurricular activity on smoking and drinking initiation among tweens. Acad Pediatr, 14(3), 271–278. 10.1016/j.acap.2014.02.002 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Ajzen I (1991). The theory of planned behavior. Organizational behavior and human decision processes, 50(2), 179–211. [Google Scholar]
  5. Anderson KG, Briggs KE, & White HR (2013). Motives to drink or not to drink: longitudinal relations among personality, motives, and alcohol use across adolescence and early adulthood. Alcohol Clin Exp Res, 37(5), 860–867. 10.1111/acer.12030 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Anderson KG, Grunwald I, Bekman N, Brown SA, & Grant A (2011). To drink or not to drink: motives and expectancies for use and nonuse in adolescence. Addict Behav, 36(10), 972–979. 10.1016/j.addbeh.2011.05.009 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Baldwin E, & O’Flaherty M (2018). Enriching the rich? A review of extracurricular activities, socioeconomic status and adolescent achievement.. In T. U. o. Q. Institute for Social Science Research (Ed.), Life Course Centre Working Paper Series, 2018-17. [Google Scholar]
  8. Bandura A (1978). The self system in reciprocal determinism. American psychologist, 33(4), 344. [Google Scholar]
  9. Barnes GM, Hoffman JH, Welte JW, Farrell MP, & Dintcheff BA (2006). Adolescents’ time use: Effects on substance use, delinquency and sexual activity. Journal of Youth and Adolescence, 36(5), 697–710. 10.1007/s10964-006-9075-0 [DOI] [Google Scholar]
  10. Benson PL (2002). Adolescent development in social and community context: A program of research. New directions for youth development, 2002(95), 123–148. [Google Scholar]
  11. Bohnert AM, & Garber J (2007). Prospective relations between organized activity participation and psychopathology during adolescence. J Abnorm Child Psychol, 35(6), 1021–1033. 10.1007/s10802-007-9152-1 [DOI] [PubMed] [Google Scholar]
  12. Bone JK, Fancourt D, Sonke JK, & Bu F (2025). The Changing Relationship Between Hobby Engagement and Substance Use in Young People: Latent Growth Modelling of the Add Health Cohort. J Youth Adolesc, 54(1), 133–145. 10.1007/s10964-024-02047-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Brooks-Russell A, Simons-Morton B, Haynie D, Farhat T, & Wang J (2014). Longitudinal relationship between drinking with peers, descriptive norms, and adolescent alcohol use. Prev Sci, 15(4), 497–505. 10.1007/s11121-013-0391-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Catalano RF, & Hawkins JD (1996). The social development model: a theory of antisocial behavior. [Google Scholar]
  15. Chassin L, & Barrera M (1993). Substance use escalation and substance use restraint among adolescent children of alcoholics. Psychology of Addictive Behaviors, 7(1), 3. 10.1037/0893-164X.7.1.3 [DOI] [Google Scholar]
  16. Curran PJ, & Hancock GR (2021). The Challenge of Modeling Co-Developmental Processes over Time. Child Development Perspectives, 15(2), 67–75. 10.1111/cdep.12401 [DOI] [Google Scholar]
  17. Curran PJ, Howard AL, Bainter SA, Lane ST, & McGinley JS (2014). The separation of between-person and within-person components of individual change over time: a latent curve model with structured residuals. J Consult Clin Psychol, 82(5), 879–894. 10.1037/a0035297 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Curtis BL, Lookatch SJ, Ramo DE, McKay JR, Feinn RS, & Kranzler HR (2018). Meta-Analysis of the Association of Alcohol-Related Social Media Use with Alcohol Consumption and Alcohol-Related Problems in Adolescents and Young Adults. Alcohol Clin Exp Res, 42(6), 978–986. 10.1111/acer.13642 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Denault A-S, & Poulin F (2019). Trajectories of participation in organized activities and outcomes in young adulthood. Applied developmental science, 23(1), 74–89. [Google Scholar]
  20. DesRoches A, & Willoughby T (2014). Bidirectional associations between valued activities and adolescent positive adjustment in a longitudinal study: positive mood as a mediator. Journal of Youth and Adolescence, 43(2), 208–220. 10.1007/s10964-013-9957-x [DOI] [PubMed] [Google Scholar]
  21. Eisman AB, Stoddard SA, Bauermeister JA, Caldwell CH, & Zimmerman MA (2018). Organized Activity Participation and Relational Aggression: The Role of Positive Youth Development. Violence Vict, 33(1), 91–108. 10.1891/0886-6708.VV-D-16-00208 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Farb AF, & Matjasko JL (2012). Recent advances in research on school-based extracurricular activities and adolescent development. Developmental Review, 32(1), 1–48. 10.1016/j.dr.2011.10.001 [DOI] [Google Scholar]
  23. Fauth RC, Roth JL, & Brooks-Gunn J (2007). Does the neighborhood context alter the link between youth’s after-school time activities and developmental outcomes? A multilevel analysis. Dev Psychol, 43(3), 760–777. 10.1037/0012-1649.43.3.760 [DOI] [PubMed] [Google Scholar]
  24. Fredricks JA, & Eccles JS (2006). Is extracurricular participation associated with beneficial outcomes? Concurrent and longitudinal relations. Dev Psychol, 42(4), 698–713. 10.1037/0012-1649.42.4.698 [DOI] [PubMed] [Google Scholar]
  25. Georgeson AR, Waddell JT, Paxton L, & Chassin L (2024). Disaggregating within- and between-person associations to test the aversive transmission of alcohol use in late adolescence through adulthood. Psychol Addict Behav, 38(5), 578–590. 10.1037/adb0001013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Greenfield TK, Guydish J, & Temple MT (1989). Reasons students give for limiting drinking: a factor analysis with implications for research and practice. Journal of studies on alcohol, 50(2), 108–115. 10.15288/jsa.1989.50.108 [DOI] [PubMed] [Google Scholar]
  27. Herrnstein RJ (1974). Formal properties of the matching law 1. Journal of the experimental analysis of behavior, 21(1), 159–164. [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Hoeben EM, Osgood DW, Siennick SE, & Weerman FM (2020). Hanging Out with the Wrong Crowd? The Role of Unstructured Socializing in Adolescents’ Specialization in Delinquency and Substance Use. Journal of Quantitative Criminology, 37(1), 141–177. 10.1007/s10940-019-09447-4 [DOI] [Google Scholar]
  29. Holmes GR, Heckel RV, Chestnut E, Harris N, & Cautela J (1987). Factor analysis of the Adolescent Reinforcement Survey Schedule (ARSS) with college freshmen. Journal of clinical psychology, 43(4), 386–390. [DOI] [PubMed] [Google Scholar]
  30. Jackson KM, Colby SM, Barnett NP, & Abar CC (2015). Prevalence and correlates of sipping and full drink of alcohol. Psychology of Addictive Behaviors, 29(3), 766–778. 10.1037/adb0000072 [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Jackson KM, Marceau K, Colby SM, Barnett NP, Rogers ML, & Hayes KL (2021). Trajectories of early alcohol use milestones: Interrelations among initiation and progression. Alcoholism: Clinical and Experimental Research, 45(11), 2294–2308. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Johnston LD, Miech RA, Patrick ME, O’Malley PM, Schulenberg JE, & Bachman JG (2023). Monitoring the Future National Survey Results on Drug Use, 1975–2022: Overview, Key Findings on Adolescent Drug Use. Institute for Social Research. [Google Scholar]
  33. Jones SC, Andrews K, & Francis K (2017). Combining Social Norms and Social Marketing to Address Underage Drinking: Development and Process Evaluation of a Whole-of-Community Intervention. PLoS One, 12(1), e0169872. 10.1371/journal.pone.0169872 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Khoddam R, Cho J, Jackson NJ, & Leventhal AM (2018). Diminished alternative reinforcement as a mechanism linking conduct problems and substance use in adolescence: a longitudinal examination. Addiction, 113(6), 1139–1148. 10.1111/add.14160 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Kievit RA, Frankenhuis WE, Waldorp LJ, & Borsboom D (2013). Simpson’s paradox in psychological science: a practical guide. Front Psychol, 4, 513. 10.3389/fpsyg.2013.00513 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Kraus MW, Piff PK, & Keltner D (2011). Social Class as Culture. Current Directions in Psychological Science, 20(4), 246–250. 10.1177/0963721411414654 [DOI] [Google Scholar]
  37. Larimer ME, Kaysen DL, Lee CM, Kilmer JR, Lewis MA, Dillworth T, Montoya HD, & Neighbors C (2009). Evaluating level of specificity of normative referents in relation to personal drinking behavior. Journal of Studies on Alcohol and Drugs, Supplement(16), 115–121. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Larson RW, Hansen DM, & Moneta G (2006). Differing profiles of developmental experiences across types of organized youth activities. Dev Psychol, 42(5), 849–863. 10.1037/0012-1649.42.5.849 [DOI] [PubMed] [Google Scholar]
  39. Leventhal AM, Bello MS, Unger JB, Strong DR, Kirkpatrick MG, & Audrain-McGovern J (2015). Diminished Alternative Reinforcement as a Mechanism Underlying Socioeconomic Disparities in Adolescent Substance Use. Prev Med, 80, 75–81. 10.1016/j.ypmed.2015.05.021 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Luthar SS, & Latendresse SJ (2005a). Children of the Affluent: Challenges to Well-Being. Curr Dir Psychol Sci, 14(1), 49–53. 10.1111/j.0963-7214.2005.00333.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Luthar SS, & Latendresse SJ (2005b). Comparable “risks” at the socioeconomic status extremes: preadolescents’ perceptions of parenting. Dev Psychopathol, 17(1), 207–230. 10.1017/s095457940505011x [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Magid V, & Moreland AD (2014). The Role of Substance Use Initiation in Adolescent Development of Subsequent Substance-Related Problems. Journal of Child & Adolescent Substance Abuse, 23(2), 78–86. 10.1080/1067828x.2012.748595 [DOI] [Google Scholar]
  43. Mahoney JL, Larson RW, & Eccles JS (2005). Organized activities as contexts of development: Extracurricular activities, after school and community programs. Psychology Press. [Google Scholar]
  44. Mahoney JL, & Stattin H (2000). Leisure activities and adolescent antisocial behavior: The role of structure and social context. Journal of Adolescence, 23(2), 113–127. [DOI] [PubMed] [Google Scholar]
  45. Marsh H, & Kleitman S (2002). Extracurricular School Activities: The Good, the Bad, and the Nonlinear. Harvard Educational Review, 72(4), 464–515. 10.17763/haer.72.4.051388703v7v7736 [DOI] [Google Scholar]
  46. Marsh HW, Hau K-T, & Wen Z (2004). In search of golden rules: Comment on hypothesis-testing approaches to setting cutoff values for fit indexes and dangers in overgeneralizing Hu and Bentler’s (1999) findings. Structural Equation Modeling, 11(3), 320–341. 10.1207/s15328007sem1103_2 [DOI] [Google Scholar]
  47. Masyn KE (2013). Latent class analysis and finite mixture modeling. [Google Scholar]
  48. McCabe KO, Modecki KL, & Barber BL (2016). Participation in Organized Activities Protects Against Adolescents’ Risky Substance Use, Even Beyond Development in Conscientiousness. J Youth Adolesc, 45(11), 2292–2306. 10.1007/s10964-016-0454-x [DOI] [PubMed] [Google Scholar]
  49. Meier A, Hartmann BS, & Larson R (2018). A Quarter Century of Participation in School-Based Extracurricular Activities: Inequalities by Race, Class, Gender and Age? J Youth Adolesc, 47(6), 1299–1316. 10.1007/s10964-018-0838-1 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Meisel SN, & Colder CR (2020). Adolescent Social Norms and Alcohol Use: Separating Between- and Within-Person Associations to Test Reciprocal Determinism. J Res Adolesc, 30 Suppl 2, 499–515. 10.1111/jora.12494 [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Meisel SN, Diab E, Acuff SF, & Jackson KM (2024, April 16). Reciprocal Associations Between Structured Activities and Reasons for Abstaining or Limiting Drinking and Descriptive Drinking Norms During Adolescence. 10.17605/OSF.IO/G64MJ [DOI] [Google Scholar]
  52. Meisel SN, Diab E, Acuff SF, & Jackson KM (2025, June 16). Analytic Code and Model Output ( 10.17605/OSF.IO/8WGZ4 [DOI] [Google Scholar]
  53. Meldrum RC, & Leimberg A (2018). Unstructured Socializing With Peers and Risk of Substance Use: Where Does the Risk Begin? Journal of Drug Issues, 48(3), 452–471. 10.1177/0022042618774263 [DOI] [Google Scholar]
  54. Merrill JE, Martin S, Abar CC, & Jackson KM (2016a). Trajectories and correlates of reasons for abstaining or limiting drinking during adolescence. Addict Behav, 52, 1–7. 10.1016/j.addbeh.2015.08.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Merrill JE, Martin S, Abar CC, & Jackson KM (2016b). Trajectories and correlates of reasons for abstaining or limiting drinking during adolescence. Addictive Behaviors, 52, 1–7. 10.1016/j.addbeh.2015.08.003 [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Moilanen KL, Markstrom CA, & Jones E (2014). Extracurricular activity availability and participation and substance use among American Indian adolescents. J Youth Adolesc, 43(3), 454–469. 10.1007/s10964-013-0088-1 [DOI] [PubMed] [Google Scholar]
  57. Morris DS (2015). Actively closing the gap? Social class, organized activities, and academic achievement in high school. Youth & Society, 47(2), 267–290. [Google Scholar]
  58. Murray RM, Sabiston CM, Dore I, Belanger M, & O’Loughlin JL (2021). Longitudinal associations between team sport participation and substance use in adolescents and young adults. Addict Behav, 116, 106798. 10.1016/j.addbeh.2020.106798 [DOI] [PubMed] [Google Scholar]
  59. Neighbors C, O’Connor RM, Lewis MA, Chawla N, Lee CM, & Fossos N (2008). The relative impact of injunctive norms on college student drinking: the role of reference group. Psychol Addict Behav, 22(4), 576–581. 10.1037/a0013043 [DOI] [PMC free article] [PubMed] [Google Scholar]
  60. Randall ET, & Bohnert AM (2009). Organized activity involvement, depressive symptoms, and social adjustment in adolescents: ethnicity and socioeconomic status as moderators. J Youth Adolesc, 38(9), 1187–1198. 10.1007/s10964-009-9417-9 [DOI] [PubMed] [Google Scholar]
  61. Roberts ME, Spillane NS, Colby SM, & Jackson KM (2017). Forecasting disparities with early substance-use milestones. Journal of Child & Adolescent Substance Abuse, 26(1), 56–59. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Roth JL, & Brooks-Gunn J (2016). Evaluating Youth Development Programs: Progress and Promise. Appl Dev Sci, 20(3), 188–202. 10.1080/10888691.2015.1113879 [DOI] [PMC free article] [PubMed] [Google Scholar]
  63. SAMHSA. (2024). Key substance use and mental health indicators in the United States: Results from the 2023 National Survey on Drug Use and Health (HHS Publication No. PEP24-07-021, NSDUH Series H-59). Center for Behavioral Health Statistics and Quality, Substance Abuse and Mental Health Services Administration.. [Google Scholar]
  64. Sampson RJ (1997). Collective regulation of adolescent misbehavior: Validation results from eighty Chicago neighborhoods. Journal of Adolescent Research, 12(2), 227–244. [Google Scholar]
  65. SAS Institute Inc. (2013). SAS 9.4 Statements: Reference. In
  66. Schuz B, Brick C, Wilding S, & Conner M (2020). Socioeconomic Status Moderates the Effects of Health Cognitions on Health Behaviors within Participants: Two Multibehavior Studies. Ann Behav Med, 54(1), 36–48. 10.1093/abm/kaz023 [DOI] [PubMed] [Google Scholar]
  67. Sherman DK, Updegraff JA, Handy MS, Eom K, & Kim HS (2022). Beliefs and Social Norms as Precursors of Environmental Support: The Joint Influence of Collectivism and Socioeconomic Status. Pers Soc Psychol Bull, 48(3), 463–477. 10.1177/01461672211007252 [DOI] [PubMed] [Google Scholar]
  68. Snellman K, Silva JM, Frederick CB, & Putnam RD (2014). The Engagement Gap. The ANNALS of the American Academy of Political and Social Science, 657(1), 194–207. 10.1177/0002716214548398 [DOI] [Google Scholar]
  69. Spillane NS, Schick MR, Kirk-Provencher KT, Hill DC, Wyatt J, & Jackson KM (2020). Structured and Unstructured Activities and Alcohol and Marijuana Use in Middle School: The Role of Availability and Engagement. Subst Use Misuse, 55(11), 1765–1773. 10.1080/10826084.2020.1762652 [DOI] [PMC free article] [PubMed] [Google Scholar]
  70. Spillane NS, Schick MR, Nalven T, Goldstein SC, Kirk-Provencher KT, Hill D, & Kahler CW (2021). Testing the competing life reinforcers model for substance use in reserve-dwelling First Nation youth. Am J Orthopsychiatry, 91(4), 477–486. 10.1037/ort0000543 [DOI] [PMC free article] [PubMed] [Google Scholar]
  71. Steinberg D, & Simon V (2019). A Comparison of Hobbies and Organized Activities Among Low Income Urban Adolescents. J Child Fam Stud, 28(5), 1182–1195. 10.1007/s10826-019-01365-0 [DOI] [PMC free article] [PubMed] [Google Scholar]
  72. Tandon PS, Kroshus E, Olsen K, Garrett K, Qu P, & McCleery J (2021). Socioeconomic Inequities in Youth Participation in Physical Activity and Sports. Int J Environ Res Public Health, 18(13). 10.3390/ijerph18136946 [DOI] [Google Scholar]
  73. Timonen J, Niemela M, Hakko H, Alakokkare A, & Rasanen S (2021). Associations between Adolescents’ Social Leisure Activities and the Onset of Mental Disorders in Young Adulthood. J Youth Adolesc, 50(9), 1757–1765. 10.1007/s10964-021-01462-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  74. Treloar Padovano H, Janssen T, Sokolovsky A, & Jackson KM (2020). The Altered Course of Learning: How Alcohol Outcome Expectancies Are Shaped by First Drinking Experiences. Psychol Sci, 31(12), 1573–1584. 10.1177/0956797620959006 [DOI] [PMC free article] [PubMed] [Google Scholar]
  75. Trucco EM, & Hartmann SA (2021). Understanding the Etiology of Adolescent Substance Use through Developmental Perspectives. Child Dev Perspect, 15(4), 257–264. 10.1111/cdep.12426 [DOI] [PMC free article] [PubMed] [Google Scholar]
  76. Voogt CV, Larsen H, Poelen EAP, Kleinjan M, & Engels RCME (2012). Longitudinal associations between descriptive and injunctive norms of youngsters and heavy drinking and problem drinking in late adolescence. Journal of Substance Use, 18(4), 275–287. 10.3109/14659891.2012.674623 [DOI] [Google Scholar]
  77. Walczak B, Walczak A, Tricas-Sauras S, & Kolodziejczyk J (2023). Does Sport Participation Protect Adolescents from Alcohol Consumption? A Scoping Review. Int J Environ Res Public Health, 20(7). 10.3390/ijerph20075417 [DOI] [Google Scholar]
  78. Wood MD, Read JP, Mitchell RE, & Brand NH (2004). Do Parents Still Matter? Parent and Peer Influences on Alcohol Involvement Among Recent High School Graduates. Psychology of Addictive Behaviors, 18(1), 19–30. 10.1037/0893-164X.18.1.19 [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supplemental text and graphics
NIHMS2142569-supplement-Supplemental_text_and_graphics.docx (504.2KB, docx)

RESOURCES