Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2026 Sep 5.
Published before final editing as: Subst Use Misuse. 2026 May 27:1–11. doi: 10.1080/10826084.2026.2672656

Longitudinal Associations Between Other-Sex Close Friends and Alcohol Use is a Between-, Not Within-, Person Process

Amy M Loviska a, Kristine Marceau a, Valerie S Knopik a, Michelle L Rogers b, Kristina M Jackson c
PMCID: PMC13543608  NIHMSID: NIHMS2181854  PMID: 42204435

Abstract

Background:

Adolescents are more likely to use alcohol if close friends are characterized as majority other-sex. The extent to which the association between proportion of other-sex close friends and adolescent alcohol use is directional and/or reciprocal is unclear. Although evidence is mixed for male adolescents, higher proportion of other-sex close friends is associated with increased risk of alcohol use for female adolescents.

Methods:

This study examined whether proportion of other-sex close friends and alcohol use predicted each other longitudinally and if associations differed by sex. Participants were 1021 adolescents (Mage = 12.22 years), 52.20% female, 71.36% White non-Hispanic, from a cohort-sequential study. Adolescent alcohol use was measured via progression on alcohol use milestones. Proportion of other-sex close friends was calculated via adolescent report of their three closest friends’ sex. A random-intercept cross-lagged panel model tested for reciprocal, longitudinal association between other-sex close friends and alcohol use across six timepoints. Sex differences were examined via a multigroup framework.

Results:

Alcohol use and proportion of other-sex close friends were relatively stable traits, with other-sex close friends’ autoregressive accumulation more pronounced at later timepoints. There were no observed concurrent nor cross-lagged associations, indicating no reciprocal association within individuals. However, there was a positive between-person association, wherein adolescents with a higher proportion of other-sex close friends had, on average, greater alcohol involvement.

Conclusions:

Results suggest that the influence of other-sex close friends on alcohol use is a between-, not within-, person process. Thus, theory purporting a within-person reciprocal mechanism may not be pertinent within the context of other-sex close friendships.

Keywords: Adolescence, alcohol, close friendships, other-sex close friends

Introduction

Adolescent alcohol use is associated with short- and long-term negative cognitive and physiological health outcomes (Boden & Fergusson, 2011; Hanson et al., 2011). Adolescent alcohol use can be influenced by social factors, such as close friends’ attitudes toward and use of alcohol (Leung et al., 2014; Mason et al., 2014). For instance, the perception of best friend alcohol use has been strongly positively associated with adolescents’ own alcohol use (Salvy et al., 2014). Increases in close friend drinking behaviors have been associated with increased likelihood of target adolescent drinking (vs. not drinking) in the past 12 months (Ali & Dwyer, 2010), whereas lower levels of best friends’ alcohol use have been associated with lower target adolescent alcohol use, measured as a factor score of adolescent alcohol quantity and frequency across lifetime, past 30 days, and past 7 days (LaFreniere et al., 2022). Yet, there is some evidence to suggest that the sex of adolescents’ close friends may play a role in the association between close friends’ alcohol attitudes and use behaviors and adolescents’ own use. For example, adolescents who reported other-sex close friends were found to be at three times greater risk for alcohol use (Dick et al., 2007). Adolescents are also more likely to report alcohol use in close friend groups characterized as (i) majority other-sex friends or (ii) balanced representation of same- and other-sex friends (Jacobs et al., 2017; Mrug et al., 2011). As interest in other-sex friends typically begins in fifth grade and peak levels of other-sex close friendships typically occur during adolescence (i.e., between eighth and eleventh grade; Arndorfer & Stormshak, 2008; Molloy et al., 2014; Mrug et al., 2011; Poulin et al., 2011; Poulin & Pedersen, 2007), it is important to understand if other-sex close friends create a unique social context with increased opportunity for alcohol initiation and use.

Theory suggests a reciprocal relationship between other-sex close friends and adolescents’ own alcohol use. According to the Peer Socialization hypothesis, friendships with older, other-sex friends serve as contexts for delinquency, such as alcohol use, and engagement in delinquent behaviors provide opportunities for friendships with older, other-sex peers (Magnusson et al., 1985; Stattin et al., 2011). There is some empirical evidence supporting reciprocal and cross-lagged relationships between close friends’ and adolescent’s own alcohol behaviors and attitudes across time (Becker & Curry, 2014; Cleveland et al., 2012), although few studies have examined these reciprocal effects of other-sex friends with alcohol use. Other studies utilizing multilevel or multiple regression models have found null associations between alcohol use and close friendship sex composition (Deutsch et al., 2014; Molloy et al., 2014), or associations in the opposite direction (Mrug et al., 2011). For example, one study in a sample of rural youth demonstrated associations between alcohol use and future other-sex friendship (but not the converse); however, associations at the level of the person (stable between-person phenomena) and individual-level associations (within-person processes that vary over time) were examined via separate models (Molloy et al., 2014). Thus, more work is needed to determine whether there are reciprocal longitudinal effects for other-sex close friends and adolescent alcohol use. The present study rigorously tests whether proportion of other-sex close friends reciprocally influences the development of alcohol use at the person and individual level across adolescence within a more geographically diverse sample.

The nature of the association between other-sex close friends and male adolescents’ alcohol use has been mixed (Arndorfer & Stormshak, 2008; Mrug et al., 2011). In contrast, having a higher proportion of, or closeness in, friendship with other-sex friends is associated with increased risk of alcohol use for female adolescents (Arndorfer & Stormshak, 2008; Gaughan, 2006; Poulin et al., 2011), who develop a larger proportion of other-sex friendships faster and initiate other-sex friendships at a younger age (Poulin et al., 2011; Poulin & Pedersen, 2007). Therefore, the observed sex difference of other-sex close friends influence on alcohol use may be due, in part, to female adolescents entering a majority other-sex friendship social context earlier than male adolescents. Accordingly, we also examine sex differences.

Current study

This study extends current literature by testing whether proportion of other-sex close friends has a reciprocal, within-person association with alcohol use over the course of three years using an analytically robust approach well-suited to this question. Specifically, we utilize a random-intercept cross-lagged panel model (RI-CLPM) to simultaneously examine contemporaneous and reciprocal effects of other-sex close friends and alcohol use within a sample of adolescents who are within the developmental range of increasing and peak interest in other-sex close friendships (i.e., between eighth and eleventh grade; Arndorfer & Stormshak, 2008; Molloy et al., 2014; Mrug et al., 2011; Poulin et al., 2011; Poulin & Pedersen, 2007). We predicted (H1) a reciprocal relation between reporting a larger proportion of other-sex close friends and alcohol use that would lead to increased use over time. Additionally, we expected sex differences, such that (H2) reciprocal associations would be salient for female adolescents only. We consider a number of confounders: (a) timing of other-sex close friendship uptake (i.e., developing other-sex friendships after initiating sixth grade; Poulin et al., 2011; Poulin & Pedersen, 2007), (b) demographic characteristics, such as pubertal timing, or (c) characteristics of friends within the friend group, such as friends’ age, age relative to the focal adolescent, friendship stability, relationship to the adolescents, and friend drinking behaviors as covariates in focal and sensitivity analyses.

Materials and methods

Participants and procedures

The study was approved by Brown University Institutional Review Board (IRB; 0808992539) and obtained a Certificate of Confidentiality from the National Institute on Alcohol Abuse and Alcoholism (AA-075-2009). Secondary data analysis was approved by Purdue University IRB (2021-1046). We report all data exclusions, manipulations, and measures included in the study. Data management, manipulation, and descriptive statistics were completed in Stata 16 (StataCorp, 2021) or SAS 9.4 and structural equation models were completed in Mplus 8 (Muthen & Muthen, 2017). Given the complexity of the data due to the cohort-sequential design, access to study data is available upon request from the study principal investigator (last author). Study code and outputs are available for purposes of reproducing results or replicating procedures upon request from the first author.

Participants were from a cohort-sequential study of 1,023 adolescents on the progression of alcohol use, Internet Surveys About You, aged 10–15 years (M = 12.22; SD = 0.98) at enrollment (Jackson et al., 2021). Participants and procedures have been previously described in detail (Jackson et al., 2014, 2015; Marceau et al., 2015). Briefly, adolescents were recruited from six geographically diverse middle schools in the Northeastern United States in five semiannual cohorts. Study information and consent forms were mailed to students’ homes and distributed in schools with ~38% of all consent forms returned. Of the returned forms, 65% indicated parental consent to participate, of which 88% of individuals were enrolled. Cohorts were approximately equally distributed by sex (52.2% female) and grade (33% sixth, 32% seventh, and 35% eighth) at baseline assessment. After completing a 45-minute baseline assessment during the two-hour in-school orientation, participants completed web-based surveys every six months for the first two years (T2 through T5) and a final follow-up survey one year later (T6) resulting in six total assessments. Six-to-twelve-month time interval between assessments is adequate for addressing research hypotheses because youth, on average, take about half a year to transition from one drinking milestone to the next, with the exception of sip to full drink (Jackson et al., 2021) and close friendships tend to be highly stable month-to-month (Chan & Poulin, 2007).

Adolescents self-reported as White, non-Hispanic (71.36%), Black, non-Hispanic (4.01%), Hispanic (12.12%), and Mixed or Other, non-Hispanic (11.53%). Approximately a third (32.84%) of participants qualified for free or reduced lunch, 29.38% had parent-reported household income of <$50,000, and about half (48.58%) had a parent with a college degree. Descriptive statistics shown in Table 1.

Table 1.

Descriptive statistics.

n Percent (%)
Demographic characteristics
 Female 1021 52.20
 White, non-Hispanic 1021 71.36
 Hispanic 1018 12.12
 Qualified/received free/reduced lunch 935 32.84
Time-invariant measures n Mean (SD) Min. Max.
 Average age across study 1021 13.43 (0.97) 10.00 16.60
 Average friendship stability 1021 1.20 (0.63) 0 3
 Average proportion of younger friends 1017 0.10 (0.14) 0 1
 Average proportion of older friends 1017 0.06 (0.12) 0 0.93
 Average close friend drinking 1021 0.34 (0.46) 0 3
Time-varying measures
 Alcohol uptake (Timepoint 1) 1015 0.70 (0.76) 0 2.45
 Alcohol uptake (Timepoint 2) 924 0.58 (0.73) 0 2.45
 Alcohol uptake (Timepoint 3) 882 0.66 (0.73) 0 2.45
 Alcohol uptake (Timepoint 4) 853 0.65 (0.76) 0 2.45
 Alcohol uptake (Timepoint 5) 829 0.73 (0.79) 0 2.45
 Alcohol uptake (Timepoint 6) 833 0.95 (0.86) 0 2.45
 Other-sex friends (Timepoint 1) 1017 0.18 (0.23) 0 1
 Other-sex friends (Timepoint 2) 923 0.21 (0.24) 0 1
 Other-sex friends (Timepoint 3) 884 0.21 (0.25) 0 1
 Other-sex friends (Timepoint 4) 857 0.23 (0.25) 0 1
 Other-sex friends (Timepoint 5) 824 0.25 (0.26) 0 1
 Other-sex friends (Timepoint 6) 830 0.29 (0.28) 0 1
Open in a new tab

Measures

See supplemental materials for additional details regarding measurement creation.

Alcohol uptake was assessed using self-endorsement of binary alcohol milestones (lifetime sip, full drink, heavy drinking [3+ drinks at one drinking occasion], and feeling drunk), and commitment not to drink (responding “definitely not” on two items: drink alcohol if offered by best friend and by classmate) at each assessment. Uptake continuum scores were coded into categories; 0 = committed to not using alcohol, 1 = susceptible to using alcohol, 2 = alcohol sipper, 3 = non-current experimenter, 4 = current experimenter, 5 = non-current established alcohol use, and 6 = current established alcohol use (see supplemental materials for information on categories). The variable was treated as continuous, as it indexes a theoretically continuous alcohol use progression. The majority of participants were categorized as committed to not using alcohol (ranging from 60 to 42%, declining across timepoints, see Figures S1a and S1b) yielding skewed data (see Table S1), with no obvious outliers. As Winsorization did not correct skew (Table S1), alcohol uptake scores were square root transformed (scores from 0 to 2.45). Square root transformations are frequently used to alleviate non-normally distributed data (Osborne, 2010). However, there is some concern regarding the validity of square root transformations for data with many zeros (Boulton & Williford, 2018). Therefore, sensitivity analyses were completed with binary coded no use (0) and any use, defined as at least a sip of alcohol (1), to assess potential differences in results via measurement of alcohol uptake (see sensitivity analysis).

Proportion of other-sex close friends was generated using participant report of up to three closest friends’ sex at each timepoint (# other-sex close friends/# friends with reported sex; see Table S2 for details). Reporting was limited to an adolescent’s three closest friends to align with perceived friend alcohol use question, which was derived from Add Health (W1 In-Home Interview, variable H1TO29; Add Health, The National Longitudinal Study of Adolescent to Adult Health, 2025). Reporting of three close friends has been previously found to be sufficient in capturing concurrent and lagged associations between other-sex close friends and delinquent behavior (Arndorfer & Stormshak, 2008). Scores ranged from 0 (all same-sex friends) to 1 (all other-sex close friends). Close friends could include romantic partners and relatives. As other-sex close friendship formation and effect on alcohol use could be influenced by secondary relationships to the adolescent, we conducted additional analyses including only non-relative, non-romantic friendships (see sensitivity analysis).

Demographic variables: sex (male= 1, female= 0), grade (sixth, seventh, or eighth at baseline), and self-reported age.

Average friendship stability was generated using participant report of first and last initials of their three closest friends at each timepoint (see Table S2 for details). Total number of stable friendships (same initials reported in the previous assessment) for each timepoint was averaged over timepoints (higher scores = more stable/longer-lived friendships over the study).

Average proportion of older and younger friends was generated using participants’ report of their three closest friends’ age at each timepoint (# younger or older friends/# friends reported). Same-age friendships were defined as plus or minus one-year age difference (Burk et al., 2012). Proportion of younger friends ranged from 0 (all older/same-age friends) to 1 (all younger friends) with proportion of older friends from 0 (all younger/same-age friends) to 1 (all older friends). Some individuals reported very young (e.g., <8 years old) and very old (e.g., >80 years old) friends. Theoretically, the developmental processes examined here should pertain to friends in a similar life period. Thus, we conducted a sensitivity analysis including only friends who were between 8 and 25 years old.

Average close friends drinking was generated using participants’ report on how many of their three closest friends drink alcohol at each timepoint. The total number of friends who drink was averaged over timepoints.

Pubertal timing was estimated using adolescent self-report on the Pubertal Development Scale (PDS; Petersen et al., 1988). The PDS is a five-item questionnaire wherein adolescents report the extent they experienced development on pubertal milestones in three shared and two sex-specific domains using a 4-point Likert scale from 1, not yet started puberty, to 4, puberty complete. Menarche was dichotomously coded as 1, pre-menarcheal, and 4, post-menarcheal. A logistic growth mixed-effects model was estimated separately for each sex using the PROC NLMIXED command in SAS 9.4 following procedures from Grimm and Ram (2009). Estimates of individual pubertal timing were obtained from λi, which represents the individual-specific slope of pubertal development, to be used in the subsequent moderated model.

Analysis plan

Hypothesis testing

We tested a RI-CLPM for proportion of other-sex close friends and alcohol use across six timepoints (see supplemental materials for model building, Table S3; Hamaker et al., 2015). For the current study data, timepoints are nested within individuals. The RI-CLPM decomposes stable between-person differences in other-sex friends and alcohol use from within-person changes in phenotypes over time. Reciprocal cross-lagged paths between proportion of other-sex close friends and adolescent alcohol use would support our primary hypothesis (H1). Adolescent sex, average participant age across the study, average friendship stability, and average proportion of older and younger friends were included as covariates on the RI. To test if lagged effects differed by sex (H2), final RI-CLPMs with covariates were run using a multigroup framework (Mulder & Hamaker, 2021). Unconstrained and constrained multigroup models were compared using the Satorra–Bentler scaled chi-square (Satorra & Bentler, 2010).

Sensitivity analyses

First, to confirm a RI-CLPM was sufficient, we estimated an autoregressive latent trajectory (ALT) model (Bollen & Curran, 2004) which additionally captures potential linear change in between-person differences (slope), rather than only a RI. As change in between-person differences in alcohol use and other-sex close friends may not be linear, an ALT with a quadratic slope was also estimated (Bauldry & Bollen, 2018). Additionally, we estimated a traditional CLPM to compare RI-CLPM model fit. Second, a series of sensitivity analyses were run on sub-samples: (a) including friends who were adolescents and emerging adults (i.e., excluding friends younger than 8 and older than 25, n = 465 friends [3%]; see Table S4), (b) excluding participants who completed <2 timepoints (n = 50 individuals), (c) excluding friends who were relatives (n = 1,178 friends; 8%) and romantic partners (n = 1,050 friends; 7%), (d) including dichotomously coded variables of grade at baseline as covariates on the RI, (e) including participants who reported no other-sex close friends at baseline with other-sex close friendships endorsed at any subsequent timepoint (n = 459 approximately evenly distributed across grade at T1 with 51% female adolescents), (f) including average close friend drinking as a covariate on the RI in both main and multigroup analyses, (g) including friend drinking as a time-varying predictor, (h) including an interaction term between the lagged component of other-sex close friends and pubertal timing, and (i) as a robustness check, alcohol use was binary-coded and included in an RI-CLPM in lieu of the continuous uptake measure.

A Bayesian estimator was utilized for sensitivity analyses (g), (h), and (i). A three-variable RI-CLPM was used to test for reciprocal associations between other-sex close friends, friends drinking, and adolescent alcohol use. A moderated RI-CLPM (M-RI-CLPM), following procedures from Speyer et al. (2023), was estimated including an interaction term between pubertal timing and the within-person lagged association of other-sex close friends on adolescent alcohol use. Convergence criterion was established as a potential scale reduction (PSR) of 1.05 or lower (Asparouhov & Muthen, 2023). As number of iterations were set using the FBITER command, convergence was confirmed by doubling the number of iterations to ensure stability of PSR (Speyer et al., 2023). Models specified the use of two Markov Chain Monte Carlo chains. The three-variable RI-CLPM (g) achieved convergence at 127,400 iterations, the M-RI-CLPM (h) achieved convergence at 6,900 iterations, and the binary-coded alcohol model (i) achieved convergence at 20,000 iterations. All Bayes estimation used non-informed priors, which produced parameter estimates analogous to Maximum Likelihood estimation (Muthén, 2010).

Results

Descriptive statistics

Based on inspection of the descriptive statistics presented in Table 1, mean levels of alcohol use were relatively stable across the study, whereas mean proportion of other-sex close friends increased from T1 to T6. The distributions of alcohol use and proportion of other-sex close friends changed across time, with reduced frequency of committed nonuse endorsement (Figures S1a and S1b) and same-sex friendship endorsement (Figure S2). Prevalence of any alcohol use (vs. no alcohol use) showed recanting from T1 to T2 (Figure S3) and increased with grade level at T1 (Figure S4). Proportion of other-sex friend endorsement was similar across grade levels at T1 (Figure S5).

Alcohol use and higher proportion of other-sex close friends were weakly correlated at some timepoints (Pearson’s r = 0.07–0.12, see Table 2). Adolescent age was consistently positively correlated with alcohol uptake. Average friendship stability was consistently negatively correlated with other-sex close friendships. Average age of adolescents’ friends was consistently positively correlated with alcohol use as well as proportion of other-sex close friends (Figure S6). Average close friend drinking was consistently positively correlated with alcohol uptake and proportion of other-sex close friends. Correlations support the inclusion of covariates as time-invariant on the RI for final models.

Table 2.

Correlation matrix for alcohol use and proportion of other-sex close friends with covariates.

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12.
1. Alcohol uptake (T1) – – – – – – – – – – – –
2. Alcohol uptake (T2) 0.65*** – – – – – – – – – – –
3. Alcohol uptake (T3) 0.59*** 0.66*** – – – – – – – – – –
4. Alcohol uptake (T4) 0.52*** 0.61*** 0.66*** – – – – – – – – –
5. Alcohol uptake (T5) 0.47*** 0.55*** 0.60*** 0.67*** – – – – – – – –
6. Alcohol uptake (T6) 0.36*** 0.37*** 0.47*** 0.49*** 0.61*** – – – – – – –
7. Other-sex close friends (T1) 0.11** 0.08* 0.08* 0.12** 0.08* 0.04 – – – – – –
8. Other-sex close friends (T2) 0.09* 0.05 0.04 0.07 0.09* 0.01 0.38*** – – – – –
9. Other-sex close friends (T3) 0.07* 0.05 0.09* 0.09* 0.09* 0.05 0.37*** 0.38*** – – – –
10. Other-sex close friends (T4) 0.09* 0.04 0.05 0.09* 0.11* 0.03 0.36*** 0.33*** 0.44*** – – –
11. Other-sex close friends (T5) 0.09* 0.05 0.09* 0.11* 0.12** 0.10* 0.26*** 0.28*** 0.40*** 0.44*** – –
12. Other-sex close friends (T6) 0.09* 0.03 0.06 0.07 0.08* 0.06 0.34*** 0.29*** 0.36*** 0.40*** 0.48*** –
Male 0.01 −0.03 −0.07 −0.11* −0.13** −0.09* 0.04 0.11** 0.05 0.07 −0.02 0.03
Avg. age across study 0.21*** 0.22*** 0.20*** 0.22*** 0.23*** 0.18*** 0.03 0.05 0.10* 0.08* 0.08* 0.03
Avg. friendship stability −0.04 −0.02 −0.03 −0.04 −0.03 −0.03 −0.20*** −0.16*** −0.23*** −0.25*** −0.28*** −0.24***
Avg. Prop. of younger friends 0.14*** 0.15*** 0.17*** 0.11** 0.11* 0.06 0.18*** 0.15*** 0.13*** 0.21*** 0.16*** 0.14***
Avg. Prop. of older friends −0.01 −0.06 −0.01 −0.02 −0.01 −0.06 0.09* 0.05 0.09* 0.09* 0.16*** 0.07*
Avg. close friend drinking 0.23*** 0.22*** 0.22*** 0.28*** 0.25*** 0.25*** 0.09* 0.06 0.14*** 0.11* 0.09* 0.07*
Open in a new tab

Gray boxes are correlations between alcohol use and proportion of other-sex friend variables at each timepoint.

Avg.: average; Prop.: proportion.

Estimates <0.009 are rounded to 0.01.

*

p < .05.

**

p < .001.

***

p < .0001.

Random intercept cross-lagged panel model

The final RI-CLPM model (Figure 1) including covariates fit the data adequately. Although there was some misfit based on the χ2 value, χ2(101) = 235.80, p < .0001, this metric is known to be overly conservative with large sample sizes. Other key fit indices, CFI = 0.960, SRMR = 0.039, and RMSEA = 0.036, were within established ranges for fit to the data (Hu & Bentler, 1999). The RI-CLPM also fit the data better than the traditional CLPM, χ2(114) = 3374.01, p < .0001, CFI = 0.800, SRMR = 0.109, RMSEA = 0.084. We report standardized coefficients obtained using the STDYX command in MPlus for all models except for the M-RI-CLPM as inclusion of an interaction term requires TYPE = random, which cannot be standardized (Muthen & Muthen, 2017).

Figure 1.

Figure 1.

Open in a new tab

RI-CLPM with standardized coefficients. Number of observations =1,021. Black lines indicate significant paths whereas light gray indicate non-significant paths. Only significant paths from covariates to random intercepts are shown and are represented by paths with the dash dot pattern. χ2(101) = 235.80, p < .0001, CFI = 0.96, RMSEA = 0.036, SRMR = 0.039. *p < .05, **p < .001, ***p < .0001.

There was evidence of autoregressive paths for alcohol use, with slightly decreasing magnitude across time (from 0.34 to 0.25). Alcohol uptake at all timepoints loaded similarly and strongly on the RI (from 0.62 to 0.73). Conversely, there was only evidence of autoregressive paths for proportion of other-sex close friends from T3 onward, with slightly increasing magnitudes (from 0.13 to 0.29). Factor loadings for proportion of other-sex close friends were relatively stable (standardized coefficients 0.51 to 0.60).

Our hypothesis (H1) that proportion of other-sex close friends would prospectively predict adolescent alcohol use and vice versa was not supported, given no evidence of cross-lagged paths. Additionally, there were no significant concurrent paths for proportion of other-sex close friends and alcohol use: timepoint-specific variation in proportion of other-sex close friends and alcohol use were not related. However, the RIs between proportion of other-sex close friends and alcohol use were correlated (r = 0.16) suggesting that individuals who tend to have a higher overall proportion of other-sex close friends also tend to have higher severity of alcohol use. Therefore, the association between proportion of other-sex close friends and alcohol use is a stable, between-person phenomenon, not a within-person process that varies over time, during adolescence.

Older youth (older average age across the study) had a higher overall (across the study) proportion of other-sex close friends (β = 0.10) and more alcohol use (β = 0.32). A higher average proportion of younger friends (relative to same-age friends) also predicted having a higher overall proportion of other-sex close friends (β = 0.25) and more alcohol use (β = 0.18). A higher average proportion of older friends (relative to same-age friends) was also associated with greater overall proportion of other-sex close friends (β = 0.09) and less alcohol use (β = −0.08). Male adolescents, compared to female adolescents, had a greater overall proportion of other-sex close friends (β = 0.12) but less alcohol use (β = −0.08). Less average friendship stability (i.e., more changes in friends over time) was associated with a higher overall proportion of other-sex close friends (β = −0.31; i.e., more stable friendships tended to be same-sex).

Multigroup analyses

Results for the multigroup RI-CLPM are shown in Figure 2 (female adolescents) and Figure 3 (male adolescents). Results for females were highly consistent with the full model (all participants) with few exceptions. The RIs between proportion of other-sex close friends and alcohol use trended toward correlated (r = 0.14, p = 0.06), but did not reach significance, whereas in the final model the RIs were significantly correlated.

Figure 2.

Figure 2.

Open in a new tab

RI-CLPM with standardized coefficients for female adolescents. Number of observations= 534. Black lines indicate significant paths whereas light gray indicate non-significant paths. Only significant paths from covariates to random intercepts are shown and are represented by paths with the dash dot pattern. χ2(166) = 293.685, p < .0001, CFI = 0.96, RMSEA = 0.039, SRMR = 0.048. *p < .05, **p < .001, ***p < .0001.

Figure 3.

Figure 3.

Open in a new tab

RI-CLPM with standardized coefficients for male adolescents. Number of observations= 487. Black lines indicate significant paths whereas light gray indicate non-significant paths. Only significant paths from covariates to random intercepts are shown and are represented by paths with the dash dot pattern. χ2(166) = 293.685, p < .0001, CFI = 0.96, RMSEA = 0.039, SRMR = 0.048. *p < .05, **p < .001, ***p < .0001.

Results for male adolescents were highly consistent with the final model with a few exceptions. There was a single cross-lagged path from T3 alcohol use to T4 proportion of other-sex close friends (β = −0.17, p < 0.05). Specifically, higher alcohol severity at T3 predicted a lower proportion of other-sex close friends at T4 (see supplemental materials for details).

Comparison of the unconstrained and constrained multigroup model revealed no sex difference in lagged effects of proportion of other-sex close friends and alcohol use, χ2(36) = 41.34, p = 0.25 (Table S5), counter to our hypothesis (H2). As the cross-lagged path for male adolescents between T3 alcohol use and T4 proportion of other-sex close friends did not significantly differ from the non-significant findings for females, we lack confidence that this lagged association for male adolescents is not spurious.

Sensitivity analyses

ALT model with a linear slope (Figure S7) as well as a quadratic slope (Figure S8) results show that the between-person change in proportion of other-sex close friends and alcohol use over time was not correlated, regardless of whether change was modeled linearly or quadratically. Yet, the stable, between-person differences in intercept levels were correlated in both the linear and quadratic ALT models. Concurrent and lagged associations in the ALT models did not differ from the final RI-CLPM model.

A RI-CLPM (a) including only friends identified as adolescents and emerging adults (i.e., 8–25 years; Figure S9a), was highly consistent with the final model except there was a single significant cross-lagged path (β = 0.09) such that a higher proportion in other-sex close friends at T5 predicted increased T6 alcohol use. To probe this path, age was restricted to adolescence (i.e., 8–19 years old, Figure S9b). When friends age was restricted to those more proximal in age, the cross-lagged effect was no longer significant. As the cross-lagged effect is small and was no longer present when cross-lagged path was probed, we conclude that the association is spurious.

Results for RI-CLPM model (b) excluding participants with fewer than two timepoints (Figure S10), (d) dichotomously coded enrollment grade variables as covariates on the RI (Figure S11), and (i) including alcohol use as binary any vs. no use (Figure S12) were consistent with final models. Results for RI-CLPMs (c) excluding relatives or romantic partners (i.e., model only including friends and best friends; Figure S13) and (e) including only participants who reported no other-sex close friends at T1 yet other-sex close friendships at any subsequent timepoint (Figure S14) were consistent with the final model except RIs for proportion of other-sex close friends and alcohol use were not correlated. Results for RI-CLPM (f) including close friend drinking as a covariate on the RI in the full model (Figure S15) and multigroup framework models were consistent with final models with a few exceptions. For female adolescents (Figure S16a), when close friend drinking was included proportion of older friends was negatively associated with RI for alcohol use and T3 alcohol use was no longer associated with T4 alcohol use (see supplemental materials for details). For male adolescents, when close friend drinking was included, the model was consistent with the final male adolescent model (Figure S16b). Comparison of the unconstrained and constrained multigroup models including close friend drinking as a covariate was consistent with final multigroup comparisons in revealing no sex difference (Table S6).

The three-variable RI-CLPM (g) was consistent with the final model (Figure S17), as there were not concurrent nor lagged associations between proportion of other-sex close friends and adolescent alcohol use as well as between proportion of other-sex close friends and close friend drinking. There were concurrent associations between close friend drinking and adolescent alcohol use at Timepoints 1, 4, and 6. Adolescent alcohol use at Timepoint 5 predicted close friend drinking at Timepoint 6. Consistent with the final model, individuals who tend to have a higher overall proportion of other-sex close friends also tend to have higher severity of alcohol use. Individuals who tend to have more friends who drink tend to have higher severity of alcohol use. The M-RI-CLPM (h) was consistent with the final model (Figure S18). Pubertal timing did not moderate the association between other-sex friends and alcohol use.

Discussion

Contrary to hypotheses, we found no cross-lagged associations between proportion of other-sex close friends and alcohol use, implying no reciprocal, within-person relationship. Moreover, there were no concurrent associations, suggesting no short-term relationships between friendship sex composition and drinking at any timepoints, nor were there correlated between-person linear changes over time. However, in line with previous studies (Arndorfer & Stormshak, 2008; Poulin et al., 2011), there was a positive between-person association of proportion of other-sex close friends and alcohol use, indicating a between-person phenomenon, not a within-person process. Trait-like proportion of other-sex close friends and alcohol use severity were positively associated, indicating that adolescents who had a higher proportion of other-sex close friends on average tended to have greater alcohol involvement, and this was true for both female and male adolescents.

Results are consistent with some reports that having other-sex close friends is not associated with adolescent alcohol use (Molloy et al., 2014), but not other reports (Poulin & Pedersen, 2007). This discrepancy may be explained by analytic differences. Previous literature which found a positive longitudinal association between other-sex close friendships and alcohol use did not disaggregate between-person and within-person processes (Arndorfer & Stormshak, 2008; Poulin et al., 2011), which can yield biased estimates, leading to an error in inference where between-person effects are generalized to the individual thereby inflating the causal relationship on the individual level (Curran & Bauer, 2011). Therefore, previous work in this area may reflect between-person differences rather than within-person processes. This has important implications for intervention as our findings suggest that encouraging same-sex friendships in youth would not likely reduce adolescent alcohol consumption. Findings revealed no sex differences in lagged associations between proportion of other-sex close friends and alcohol use, which is in contrast to previous work (Mrug et al., 2011; Poulin et al., 2011). Considered together, results suggest reciprocal processes of proportion of other-sex close friendships and alcohol use may not be strong over time, regardless of sex.

An alternative explanation of the divergence of our findings is conceptual differences stemming from measurement strategy. We limited respondents to three friends to align with the measure of friend alcohol use derived from Add Health (W1:H1TO29; Add Health, The National Longitudinal Study of Adolescent to Adult Health, 2025), likely capturing changes in close friends sex composition during adolescence, which may be socially different than generally increasing one’s proportion of other-sex close friends in one’s entire friend network. Thus, our findings should not be generalized to friendships in general or other network-based measures of friendship (e.g., reciprocity and popularity). Our findings are specific to adolescents’ self-reported perceived close friendships.

Although we found a between-person association between other-sex close friends and adolescent alcohol use, we cannot determine the mechanisms underlying the observed association. However, covariates suggest possible explanations. For instance, adolescents with fewer same-age friends (i.e., a higher proportion of younger and older friends in a friendship group) tend to report a higher proportion of other-sex close friends. This may suggest that other-sex close friendship groups may be more age heterogenous. As alcohol use increases with age (Duncan et al., 2006), age heterogeneity within friendship groups may increase opportunity for alcohol use. Prior work has shown that friendships with older, other-sex peers are associated with substance use and delinquency, particularly for female adolescents and in the context of early pubertal development (Poulin et al., 2011; Stattin & Magnusson, 1990). Moreover, female adolescents’ other-sex close friends tend to be older than their same-sex friends (Poulin & Pedersen, 2007), and older friends have been found to exert greater influence on younger friends’ alcohol use (Popp et al., 2008). Notably, in this study, average proportion of younger friends was positively associated with alcohol use severity, whereas average proportion of older friends was negatively associated. The direction of these covariate effects may reflect age-related social processes. For example, increased drinking in the presence of younger friends may function to establish or maintain status for the older friend (Gommans et al., 2017) whereas disapproval of alcohol use at that age by older friends may discourage drinking among younger friends (Mrug & McCay, 2013). Future work should examine whether age heterogeneity within other-sex friend groups is associated with adolescent alcohol use.

Sensitivity analyses excluding friends who were also relatives or romantic partners revealed no between-person association between proportion of other-sex close friends and alcohol use. This finding may suggest that relationships beyond friends may contribute to the observed association. Prior research on sibling gender composition and substance use has yielded mixed results, with some studies reporting greater similarity in substance use within same-sex dyads (Trim et al., 2006) and others finding no effects of sibling gender composition (Whiteman et al., 2016). Importantly, other-sex siblings may provide initial or increased exposure to other-sex peers and facilitate engagement with other-sex individuals, potentially influencing the timing and rate of other-sex friendship formation (Doughty et al., 2015). Romantic relationships may also play a role, as evidence suggests romantic partners’ alcohol attitudes exert stronger influences on adolescent alcohol use than platonic friendships (Wesche et al., 2019). Additionally, early initiation of romantic relationships has been associated with greater alcohol use during adolescence (Zimmer-Gembeck et al., 2004). Future research should more fully explore how relationship type differentially relates to adolescent alcohol use.

Parents may also influence the development of both adolescent alcohol use and other-sex friendships. Although parental behaviors are not explored in this study, behaviors, such as parental monitoring, has been associated with delayed substance initiation (Yap et al., 2017), lower rates of binge drinking (Mustanski et al., 2017), and reduced alcohol use (Mills et al., 2021; Yap et al., 2017). Some evidence suggests parental monitoring may influence alcohol use indirectly through other-sex friendships for girls, but not boys, such that lower parental knowledge is associated with more other-sex friendships, which in turn are associated with more alcohol use (Poulin & Denault, 2012). Future studies should explore a broader social network perspective to better understand how peers, parents, siblings, and romantic partners jointly shape adolescent alcohol use.

Several limitations should be noted. First, adolescents were only allowed to nominate and report on a maximum of three close friends. The developmental trend toward heterophily could be more challenging to observe because constraining to the three closest friendships could increase homophily. We found that the gender correlation between friendship pairs for our sample ranged from 0.63 (T1) to 0.43 (T6). Although gender homophily is high in our sample, particularly at T1, it is somewhat lower than correlations recently reported when nominating up to seven friends (r = 0.88 to 0.73 across middle school and high school; Osgood et al., 2022), and indeed shows the trend toward heterophily even constraining to three friendships. Further, previous work that limited friend nominations to a maximum of three friends has observed an association between other-sex close friends and adolescent delinquency both concurrently and longitudinally (Arndorfer & Stormshak, 2008). However, limiting the number of nominations to a small threshold could lead to low variability within the nominated friends thereby increasing the chance for type II error (Babcock et al., 2014). Thus, the three-friend limit may contribute to the null within-person association.

Second, measures of friendship characteristics and alcohol use are self-report, contributing to the potential for shared method variance and increasing the associations among constructs. Thus, the observed between-person association of other-sex close friends and alcohol use should be considered in this context. However, as our study found no evidence of within-person processes when shared method variance would increase an association, this strengthens our argument that there truly is no meaningfully large reciprocal relationship. Future research on other-sex close friends and alcohol use should include multiple reporters to offset this limitation.

Third, each timepoint included participants across a range of ages due to the cohort-sequential nature of the data. The accelerated longitudinal design is a strength for assessing the range of alcohol use across a wider span of adolescence (Jackson et al., 2021), yet it presents a challenge: if other-sex close friends influence on alcohol use substantively differ by age, then the range of ages at each timepoint may diminish or mask effects. This limitation is mitigated by sensitivity analyses showing that grade at baseline was not associated with overall proportion of other-sex close friends (Figure S10).

Fourth, alcohol prevalence at enrollment may have been lower than population rates. The study attempted to overcome this limitation by using the alcohol uptake continuum, which captures early stages of alcohol use progression, such as alcohol susceptibility and commitment not to drink, though this does not prevent selective enrollment. However, if the reciprocal processes between other-sex close friends and alcohol use is specific to later stages of alcohol progression, then low prevalence of use may contribute to the null within-person association. Future research should test for reciprocal associations between other-sex close friends and alcohol use within samples who have achieved later stages of alcohol use progression.

Conclusion

Our findings suggest there may be no causal relationship between other-sex close friends and alcohol use on an individual level. However, there was a between-person association in that adolescents with a higher proportion of other-sex close friends on average tended to have more severe alcohol use. This has implications for future research as theory purporting within-person reciprocal mechanisms for other-sex close friends and alcohol use may not be pertinent within the context of close friendships. However, studies exploring the influence of close friends on adolescent alcohol use may want to control for between-person differences in exposure to other-sex friends. Future research should explore the association between other-sex friends and alcohol use within broader friendship networks.

Supplementary Material

Supp 1

Supplemental data for this article can be accessed online at https://doi.org/10.1080/10826084.2026.2672656.

Funding

This study was supported by grants from the National Institute of Health; R01 AA016838 (PI: Jackson), K02 AA13938 (PI: Jackson), and K01 DA039288 (PI: Marceau).

Footnotes

Disclosure statement

The authors report there are no competing interests to declare.

References

  1. Add Health, The National Longitudinal Study of Adolescent to Adult Health. (2025, April 26). Add Health Codebook Explorer (ACE) W1:H1TO29. https://addhealth.cpc.unc.edu/documentation/codebook-explorer/#/display_variable/7602 [Google Scholar]
  2. Ali MM, & Dwyer DS (2010). Social network effects in alcohol consumption among adolescents. Addictive Behaviors, 35(4), 337–342. 10.1016/j.addbeh.2009.12.002 [DOI] [PubMed] [Google Scholar]
  3. Arndorfer CL, & Stormshak EA (2008). Same-sex versus other-sex best friendship in early adolescence: Longitudinal predictors of antisocial behavior throughout adolescence. Journal of Youth and Adolescence, 37(9), 1059–1070. 10.1007/s10964-008-9311-x [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Asparouhov T, & Muthen BO (2023). Bayesian analysis using Mplus: Technical implementation. Technical Report. Version 4. Los Angeles: Muth’en & Muth’en. https://www.statmodel.com [Google Scholar]
  5. Babcock B, Marks PEL, Cillessen AHN, & Crick NR (2014). Limited nomination reliability using single- and multiple-item measures. Social Development, 23(3), 518–536. [Google Scholar]
  6. Bauldry S, & Bollen KA (2018). Nonlinear autoregressive latent trajectory models. Sociological Methodology 48(1), 269–302. 10.1177/0081175018789441 [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Becker SJ, & Curry JF (2014). Testing the effects of peer socialization versus selection on alcohol and marijuana use among treated adolescents. Substance Use & Misuse, 49(3), 234–242. 10.3109/10826084.2013.824479 [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Boden JM, & Fergusson DM (2011). The short- and long-term consequences of adolescent alcohol use. In Saunders JB& Rey JM(Eds.), Young People and Alcohol: Impact, Policy, Prevention, Treatment (pp. 32–44). Blackwell Publishing Ltd. [Google Scholar]
  9. Bollen KA, & Curran PJ (2004). Autoregressive latent trajectory (ALT) models a synthesis of two traditions. Sociological Methods & Research, 32(3), 336–383. 10.1177/0049124103260222 [DOI] [Google Scholar]
  10. Boulton AJ, & Williford A (2018). Analyzing skewed continuous outcomes with many zeros: A tutorial for social work and youth prevention science researchers. Journal of the Society for Social Work and Research, 9(4), 721–740. 10.1086/701235 [DOI] [Google Scholar]
  11. Burk WJ, van der Vorst H, Kerr M, & Stattin H (2012). Alcohol use and friendship dynamics: Selection and socialization in early-, middle-, and late-adolescent peer networks. Journal of Studies on Alcohol and Drugs, 73(1), 89–98. 10.15288/jsad.2012.73.89 [DOI] [PubMed] [Google Scholar]
  12. Chan A, & Poulin F (2007). Monthly changes in the composition of friendship networks in early adolescence. Merrill-Palmer Quarterly, 53(4), 578–602. 10.1353/mpq.2008.0000 [DOI] [Google Scholar]
  13. Cleveland MJ, Feinberg ME, & Jones DE (2012). Predicting alcohol use across adolescence: Relative strength of individual, family, peer, and contextual risk and protective factors. Psychology of Addictive Behaviors: Journal of the Society of Psychologists in Addictive Behaviors, 26(4), 703–713. 10.1037/a0027583 [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Curran PJ, & Bauer DJ (2011). The disaggregation of within-person and between-person effects in longitudinal models of change. Annual Review of Psychology, 62, 583–619. 10.1146/annurev.psych.093008.100356 [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Deutsch AR, Steinley D, & Slutske WS (2014). The role of gender and friends’ gender on peer socialization of adolescent drinking: A prospective multilevel social network analysis. Journal of Youth and Adolescence, 43(9), 1421–1435. 10.1007/s10964-013-0048-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Dick DM, Pagan JL, Holliday C, Viken R, Pulkkinen L, Kaprio J, & Rose RJ (2007). Gender differences in friends’ influences on adolescent drinking: A genetic epidemiological study. Alcoholism, Clinical and Experimental Research, 31(12), 2012–2019. 10.1111/j.1530-0277.2007.00523.x [DOI] [PubMed] [Google Scholar]
  17. Doughty SE, Lam CB, Stanik CE, & McHale SM (2015). Links between sibling experiences and romantic competence from adolescence through young adulthood. Journal of Youth and Adolescence, 44(11), 2054–2066. 10.1007/s10964-014-0177-9 [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Duncan SC, Duncan TE, & Strycker LA (2006). Alcohol use from ages 9 to 16: A cohort-sequential latent growth model. Drug and Alcohol Dependence, 81(1), 71–81. 10.1016/j.drugalcdep.2005.06.001 [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Gaughan M (2006). The gender structure of adolescent peer influence on drinking. Journal of Health and Social Behavior, 47(1), 47–61. 10.1177/002214650604700104 [DOI] [PubMed] [Google Scholar]
  20. Gommans R, Müller CM, Stevens GWJM, Cillessen AHN, & Ter Bogt TFM (2017). Individual popularity, peer group popularity composition and adolescents’ alcohol consumption. Journal of Youth and Adolescence, 46(8), 1716–1726. 10.1007/s10964-016-0611-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Grimm KJ, & Ram N (2009). Nonlinear growth models in Mplus and SAS. Structural Equation Modeling: A Multidisciplinary Journal, 16(4), 676–701. 10.1080/10705510903206055 [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Hamaker EL, Kuiper RM, & Grasman RPPP (2015). A critique of the cross-lagged panel model. Psychological Methods, 20(1), 102–116. 10.1037/a0038889 [DOI] [PubMed] [Google Scholar]
  23. Hanson KL, Medina KL, Padula CB, Tapert SF, & Brown SA (2011). Impact of adolescent alcohol and drug use on neuropsychological functioning in young adulthood: 10-year outcomes. Journal of Child & Adolescent Substance Abuse, 20(2), 135–154. 10.1080/1067828X.2011.555272 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Hu L-T, & Bentler PM (1999). Cutoff criteria for fit indexes in covariance structure analysis: Conventional criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6(1), 1–55. 10.1080/10705519909540118 [DOI] [Google Scholar]
  25. 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: Journal of the Society of Psychologists in Addictive Behaviors, 29(3), 766–778. 10.1037/adb0000072 [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. 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. 10.1111/acer.14723 [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Jackson KM, Roberts ME, Colby SM, Barnett NP, Abar CC, & Merrill JE (2014). Willingness to drink as a function of peer offers and peer norms in early adolescence. Journal of Studies on Alcohol and Drugs, 75(3), 404–414. 10.15288/jsad.2014.75.404 [DOI] [PMC free article] [PubMed] [Google Scholar]
  28. Jacobs W, Goodson P, Barry AE, McLeroy KR, McKyer ELJ, & Valente TW (2017). Adolescent social networks and alcohol use: Variability by gender and type. Substance Use & Misuse, 52(4), 477–487. 10.1080/10826084.2016.1245333 [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. LaFreniere LS, Newman MG, & Graham JW (2022). Parental support and monitoring influences on adolescent alcohol use: A peer selection mediation model. Mental Health and Addiction Research, 6(2), 1–8. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Leung RK, Toumbourou JW, & Hemphill SA (2014). The effect of peer influence and selection processes on adolescent alcohol use: A systematic review of longitudinal studies. Health Psychology Review, 8(4), 426–457. 10.1080/17437199.2011.587961 [DOI] [PubMed] [Google Scholar]
  31. Magnusson D, Stattin H, & Allen VL (1985). Biological maturation and social development: A longitudinal study of some adjustment processes from mid-adolescence to adulthood. Journal of Youth and Adolescence, 14(4), 267–283. 10.1007/BF02089234 [DOI] [PubMed] [Google Scholar]
  32. Marceau K, Abar CC, & Jackson KM (2015). Parental knowledge is a contextual amplifier of associations of pubertal maturation and substance use. Journal of Youth and Adolescence, 44(9), 1720–1734. 10.1007/s10964-015-0335-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Mason MJ, Mennis J, Linker J, Bares C, & Zaharakis N (2014). Peer attitudes effects on adolescent substance use: The moderating role of race and gender. Prevention Science: The Official Journal of the Society for Prevention Research, 15(1), 56–64. 10.1007/s11121-012-0353-7 [DOI] [PubMed] [Google Scholar]
  34. Mills R, Mann MJ, Smith ML, & Kristjansson AL (2021). Parental support and monitoring as associated with adolescent alcohol and tobacco use by gender and age. BMC Public Health, 21(1), 2000. 10.1186/s12889-021-12119-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Molloy LE, Gest SD, Feinberg ME, & Osgood DW (2014). Emergence of mixed-sex friendship groups during adolescence: Developmental associations with substance use and delinquency. Developmental Psychology, 50(11), 2449–2461. 10.1037/a0037856 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Mrug S, & McCay R (2013). Parental and peer disapproval of alcohol use and its relationship to adolescent drinking: Age, gender, and racial differences. Psychology of Addictive Behaviors: Journal of the Society of Psychologists in Addictive Behaviors, 27(3), 604–614. 10.1037/a0031064 [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Mrug S, Borch C, & Cillessen AHN (2011). Other-sex friendships in late adolescence: Risky associations for substance use and sexual debut? Journal of Youth and Adolescence, 40(7), 875–888. 10.1007/s10964-010-9605-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Mulder JD, & Hamaker EL (2021). Three extensions of the random intercept cross-lagged panel model. Structural Equation Modeling: A Multidisciplinary Journal, 28(4), 638–648. 10.1080/10705511.2020.1784738 [DOI] [Google Scholar]
  39. Mustanski B, Swann G, Newcomb ME, & Prachand N (2017). Effects of parental monitoring and knowledge on substance use and HIV risk behaviors among young men who have sex with men: Results from three studies. AIDS and Behavior, 21(7), 2046–2058. 10.1007/s10461-016-1607-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Muthén B (2010). Bayesian analysis in Mplus: A brief introduction. Los Angeles: Muth’en & Muth’en. https://www.statmodel.com [Google Scholar]
  41. Muthen LK, & Muthen BO (2017). Mplus User’s Guide Eighth Edition. Muthen & Muthen. [Google Scholar]
  42. Osborne J (2010). Improving your data transformations: Applying the Box-Cox transformation. Pratical Assessment, Research, and Evaluation, 15(1), 12. [Google Scholar]
  43. Osgood DW, Ragan DT, Dole JL, & Kreager DA (2022). Similarity of friends versus nonfriends in adolescence: Developmental patterns and ecological influences. Developmental Psychology, 58(7), 1386–1401. 10.1037/dev0001359 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Petersen AC, Crockett L, Richards M, & Boxer A (1988). A self-report measure of pubertal status: Reliability, validity, and initial norms. Journal of Youth and Adolescence, 17(2), 117–133. 10.1007/BF01537962 [DOI] [PubMed] [Google Scholar]
  45. Popp D, Laursen B, Kerr M, Stattin H, & Burk WK (2008). Modeling homophily over time with an actor-partner interdependence model. Developmental Psychology, 44(4), 1028–1039. 10.1037/0012-1649.44.4.1028 [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Poulin F, & Denault A-S (2012). Other-sex friendships as a mediator between parental monitoring and substance use in girls and boys. Journal of Youth and Adolescence, 41(11), 1488–1501. 10.1007/s10964-012-9770-y [DOI] [PubMed] [Google Scholar]
  47. Poulin F, & Pedersen S (2007). Developmental changes in gender composition of friendship networks in adolescent girls and boys. Developmental Psychology, 43(6), 1484–1496. 10.1037/0012-1649.43.6.1484 [DOI] [PubMed] [Google Scholar]
  48. Poulin F, Denault A-S, & Pedersen S (2011). Longitudinal associations between other-sex friendships and substance use in adolescence. Journal of Research on Adolescence, 21(4), 776–788. 10.1111/j.1532-7795.2011.00736.x [DOI] [Google Scholar]
  49. Salvy S-J, Pedersen ER, Miles JNV, Tucker JS, & D’Amico EJ (2014). Proximal and distal social influence on alcohol consumption and marijuana use among middle school adolescents. Drug and Alcohol Dependence, 144, 93–101. 10.1016/j.drugalcdep.2014.08.012 [DOI] [PMC free article] [PubMed] [Google Scholar]
  50. Satorra A, & Bentler PM (2010). Ensuring positiveness of the scaled difference chi-square test statistic. Psychometrika, 75(2), 243–248. 10.1007/s11336-009-9135-y [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Speyer LG, Ushakova A, Blakemore S-J, Murray AL, & Kievit R (2023). Testing for within × within and between × within moderation using Random Intercept Cross-Lagged Panel Models. Structural Equation Modeling: A Multidisciplinary Journal, 30(2), 315–327. 10.1080/10705511.2022.2096613 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. StataCorp. (2021). Stata statistical software: Release 17. StataCorp LLC. [Google Scholar]
  53. Stattin H, & Magnusson D (1990). Pubertal maturation in female development. Routledge. 10.4324/9781315789019 [DOI] [Google Scholar]
  54. Stattin H, Kerr M, & Skoog T (2011). Early pubertal timing and girls’ problem behavior: Integrating two hypotheses. Journal of Youth and Adolescence, 40(10), 1271–1287. 10.1007/s10964-011-9696-9 [DOI] [PubMed] [Google Scholar]
  55. Trim RS, Leuthe E, & Chassin L (2006). Sibling influence on alcohol use in a young adult, high-risk sample. Journal of Studies on Alcohol, 67(3), 391–398. 10.15288/jsa.2006.67.391 [DOI] [PubMed] [Google Scholar]
  56. Wesche R, Kreager DA, & Lefkowitz ES (2019). Sources of social influence on adolescents’ alcohol use. Journal of Research on Adolescence: The Official Journal of the Society for Research on Adolescence, 29(4), 984–1000. 10.1111/jora.12439 [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Whiteman SD, Jensen AC, Mustillo SA, & Maggs JL (2016). Understanding sibling influence on adolescents’ alcohol use: Social and cognitive pathways. Addictive Behaviors, 53, 1–6. 10.1016/j.addbeh.2015.09.007 [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Yap MBH, Cheong TWK, Zaravinos-Tsakos F, Lubman DI, & Jorm AF (2017). Modifiable parenting factors associated with adolescent alcohol misuse: A systematic review and meta-analysis of longitudinal studies. Addiction, 112(7), 1142–1162. 10.1111/add.13785 [DOI] [PubMed] [Google Scholar]
  59. Zimmer-Gembeck MJ, Siebenbruner J, & Collins WA (2004). A prospective study of intraindividual and peer influences on adolescents’ heterosexual romantic and sexual behavior. Archives of Sexual Behavior, 33(4), 381–394. 10.1023/B:ASEB.0000028891.16654.2c [DOI] [PubMed] [Google Scholar]

Associated Data

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

Supplementary Materials

Supp 1
NIHMS2181854-supplement-Supp_1.docx (6.2MB, docx)

RESOURCES