Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Dec 1.
Published in final edited form as: J Res Adolesc. 2018 Jul 28;29(4):984–1000. doi: 10.1111/jora.12439

Sources of social influence on adolescents’ alcohol use

Rose Wesche a, Derek A Kreager b, Eva S Lefkowitz c
PMCID: PMC6349521  NIHMSID: NIHMS978114  PMID: 30054964

Rates of lifetime drunkenness increase by a factor of four during the teenage years (Johnston, O’Malley, Miech, Bachman, & Schulenberg, 2016), a concerning statistic given alcohol’s potentially deleterious effects on health, safety, and development (National Institute on Alcohol Abuse and Alcoholism [NIAAA], 2017). Although there are legal consequences of any alcohol use in adolescence, frequent drunkenness is particularly problematic because it raises risks of physical injury, substance dependence, and detrimental neurodevelopmental outcomes (Centers for Disease Control [CDC], 2015; NIAAA, 2017; Squeglia, Jacobus, & Tapert, 2009; Substance Abuse and Mental Health Services Administration, 2014). Peer relationships, such as friendships and romantic relationships, are major factors in determining adolescents’ alcohol use (Fischer & Wiersma, 2012; Kreager, Haynie, & Hopfer, 2013). In this paper, we examine the associations of peers’ drunkenness, peers’ alcohol-related attitudes, and unstructured socializing with adolescents’ self-reported frequency of drunkenness. We distinguish between the contributions of friends, romantic partners, and romantic partners’ friends to determine which peer relationships uniquely contribute to changes in frequency of drunkenness during adolescence. In addition, we explore multiple social processes (peers’ frequency of drunkenness, peers’ alcohol-related attitudes, and unstructured socializing) to determine whether the mechanisms of social influence on frequency of drunkenness apply to a range of peer relationships.

Transmission of Alcohol Use in Friendship Groups

Throughout adolescence, friends and romantic partners may influence risk behavior through mechanisms corresponding to multiple theories of behavioral influence, including differential association, social learning, and social ecological theories. Differential association theory (Sutherland, 1947) proposes that individuals learn favorable attitudes toward criminal and delinquent deviant behaviors from significant others; in turn, these attitudes become principal mechanisms for individuals’ own future deviant behavior (Ragan, 2014; Bruinsma, 2014). Although there is overlap between differential association theory and social learning theory, they differ in that differential association theory necessitates attitude transference to explain behavior, whereas social learning theory emphasizes that behavior change can occur without attitude transference due to modeling and operant conditioning processes (Akers, Krohn, Lanza-Kaduce, & Radosevich, 1979; Akers, 2017).

In contrast to differential association and social learning theories, social ecological theories emphasize that the spatial and temporal organization of daily life creates opportunities to engage in specific behaviors (Hawley, 1950). Unstructured socializing with peers increases adolescents’ opportunities for alcohol use and other deviant behaviors because (1) the presence of peers makes engaging in these behaviors easier and more rewarding, (2) the absence of authority figures reduces the potential for punishment for these behaviors, and (3) the lack of structured activities means that individuals have time to engage in these behaviors (Osgood, Wilson, O’Malley, Bachman, & Johnston, 1996).

Consistent with the above theories, studies separately examining peer alcohol use, alcohol-related attitudes, and unstructured socializing have identified each of these variables to be associated with alcohol use. Although the present paper focuses on frequency of drunkenness, past research has utilized a variety of measures of alcohol use, such as whether/how often adolescents drink any alcohol or engage in heavy episodic drinking. In support of differential association theory, friends’ alcohol-related attitudes predict changes in individuals’ own alcohol use (Borsari & Carey, 2003; Ragan, 2014). In support of social learning theory, adolescents whose friends and romantic partners use alcohol increase their own alcohol use over time, relative to adolescents whose peers do not use alcohol (Cheadle, Walsemann, & Goosby, 2015; Leung, Toumbourou, & Hemphill, 2014; Osgood et al., 2013). In support of social ecological explanations of alcohol use, unstructured socializing with friends predicts alcohol use (Hoeben, Meldrum, & Young, 2016; Sun & Longazel, 2008).

Romantic Partners and Alcohol Use

The bulk of research and interventions has focused on friendships as primary social influences on adolescents’ alcohol use (Brechwald & Prinstein, 2011). Despite their importance, however, friends are not adolescents’ only close and potentially influential peer relationships. Romantic relationships become increasingly common across adolescence—by age 18, the majority of adolescents have had a romantic relationship (Connolly & McIsaac, 2011)–and these relationships are distinct from friendships in key ways that may alter their associations with alcohol use. Both friendships and romantic relationships are close and voluntary relationships that provide emotional support and affiliation, but romantic relationships are marked by distinctive emotional intensity and physical intimacy (Collins, Welsh, & Furman, 2009) and become increasingly committed and emotionally intimate across adolescence (Connolly & McIsaac, 2011; Meier & Allen, 2009). Because of the intensity of romantic relationships, adolescents may be particularly motivated to decrease discrepancies between their own and their romantic partners’ alcohol-related behavior and attitudes. Therefore, romantic partners’ alcohol-related behavior and attitudes may predict changes in adolescents’ own alcohol use beyond that of other peer relationships. Alternatively, it is also possible that romantic partners have little influence on changes in adolescents’ own alcohol use. Specifically, alcohol use and attitudes may be characteristics that adolescents rely on to select romantic partners, creating behavioral homophily that weakens the potential for behavioral influence (Kreager et al., 2013).

In addition to romantic partners’ behavior and attitudes, unstructured socializing with romantic partners may also relate to adolescents’ alcohol use. On the one hand, unstructured socializing creates opportunities for risk behavior, and thus “hanging out” with romantic partners may be positively associated with adolescents’ alcohol use. On the other hand, unstructured socializing with romantic partners may involve other behaviors, such as intimate and/or sexual behavior (Barnes, Hoffman, Welte, Farrell, & Dintcheff, 2007; Cohen, Farley, Taylor, Martin, & Schuster, 2002), which replace alcohol use and weaken the association between unstructured socializing with romantic partners and adolescents’ own alcohol use.

Romantic Partners’ Friends and Alcohol Use

Romantic partners may also serve as bridges or liaisons that connect different groups of friends (Kreager & Haynie, 2011). Although adolescents’ romantic relationships typically do not begin as friendships (Kreager, Molloy, Moody, & Feinberg, 2016), the friends of one’s dating partners may become one’s own friends and acquaintances. Upon exposure to these new peers, individuals may alter their own behavior to match their partners’ friends to (1) enhance status with new peers, (2) please their romantic partners, or (3) forge friendships with their partners’ friends. Therefore, the alcohol-related behavior and attitudes of romantic partners’ friends may influence individuals’ own alcohol use.

When their romantic partners’ friends drink more, adolescents increase their own alcohol use over time (Kreager & Haynie, 2011), indicating that friends of romantic partners contribute to social learning of alcohol use. However, to our knowledge, it is unknown whether romantic partners’ friends’ alcohol-related attitudes predict adolescents’ own alcohol use. Consistent with differential association theory, becoming integrated into romantic partners’ social networks may motivate individuals to adopt behavior consistent with their partners’ friends’ attitudes, and therefore change their alcohol use to be more congruent with these attitudes.

In addition to normative social influence on behavior, the friends of an adolescent’s romantic partner may influence opportunities to use alcohol through unstructured socializing. The time that romantic partners spend socializing with their friends may organize adolescents’ own social opportunities in ways that promote alcohol use. If romantic partners include each other in unstructured socializing with their friends, adolescents may find themselves in situations where they are able and encouraged to drink alcohol. To our knowledge, it is unknown whether unstructured socializing involving romantic partners’ friends predicts adolescents’ alcohol use.

The Present Research

In this paper we examine diverse social influences on adolescents’ self-reported frequency of drunkenness. We utilize a sample of rural U.S. adolescents, a group characterized by higher rates of alcohol use and trajectories of increasing heavy alcohol use across adolescence, relative to urban adolescents (Chan et al., 2016; Martino, Ellickson, & McCaffrey, 2008). We build on past research by addressing multiple peer relationships: friends, romantic partners, and friends of romantic partners. We also explore multiple mechanisms of social influence: peers’ frequency of drunkenness, alcohol-related attitudes, and unstructured socializing. We use a longitudinal design, which strengthens conclusions about the possibility of a causal association between variables by assessing within-person change. We also use social network measures, collected from participants and individuals they nominate as friends or romantic partners. By relying on first-hand reports rather than adolescents’ perceptions of their peers’ behaviors and attitudes, social network measures reduce perception biases, providing more accurate estimates of friends’ and partners’ behaviors and attitudes.

In addition to these predictors, we control for variables that are associated with more frequent/heavier alcohol use in adolescence: age (wave of data collection), school grades, SES (receiving free/reduced price school lunch), and living with both parents (Brown & Rinelli, 2010; Johnston et al., 2016; Poonawalla, Kendzor, Owen, & Caughy, 2014). Because norms for alcohol use and susceptibility to social influence change across adolescence (Brechwald & Prinstein, 2011; CDC, 2015; Johnston et al., 2016; Steinberg & Monahan, 2007), we explore whether age moderates each association we examine. We propose the following research questions:

  1. Are peers’ frequency of drunkenness, alcohol-related attitudes, and unstructured socializing uniquely associated with within-person changes in adolescents’ own frequency of drunkenness?

  2. Do these associations apply to friends, romantic partners, and romantic partners’ friends?

Method

Participants and Procedure

The longitudinal PROSPER study (Spoth, Greenberg, Bierman, & Redmond, 2004; Spoth et al., 2007) follows two successive cohorts of students from 28 rural communities in Iowa (n = 14) and Pennsylvania (n = 14), with 1,300 to 5,200 enrolled public school students per community. The communities were predominantly White (61% to 97%) and had a median household income of $37,000. Students were surveyed in their classrooms in the fall and spring of grade 6 (W1 and W2) and then every spring thereafter until grade 12 (W8). All study procedures were approved by the supervising institutions’ Institutional Review Boards. Participation rates ranged from 86–90% across waves for all eligible students, with an average of 87.2% participation and about 11,000 students responding at each wave. Enrollment in the study was open at each wave, drawing the sample from the entire student body at each occasion (Osgood et al., 2013). The present research focuses on waves 4–8 (8th–12th grade).

Criteria for inclusion in the present analyses were thus that students participated in at least two waves of data collection between waves 4 and 8, provided data on friendships and romantic relationships, reported an other-gender romantic partner during at least two waves of data collection, and whose romantic partner(s) was identified as a study participant. Thus, we excluded approximately 9,200 of the 14,000 PROSPER participants who provided data between waves 4–8 because they did not report a matched romantic partner at any wave (although some reported a romantic partner in a different grade or school), approximately 2,900 for reporting a romantic partner at only one wave, and 74 because they did not report any friends at any wave of data collection. We excluded same-sex romantic partnerships for two key reasons. First, it was necessary to conduct analyses separately by gender to account for interdependence in the data. Second, the effects of friends and romantic partners on heavy alcohol use may differ for people in heterosexual versus same-sex romantic relationships, but the small number of same-sex romantic relationships in the sample (n = 10) prevented us from testing these differences. Finally, we excluded approximately 300 participants because they were missing at least one variable included in the regression at every wave of data collection. The final sample included 1,439 adolescents and 2,943 measurement occasions, with 2–5 measurement occasions per participant.

The sample was 46% female. Regarding race/ethnicity, 90% of participants were White, 4% Hispanic, 2% Black, 1% Asian, less than 1% American Indian, and 3% reported another race or more than one race/ethnicity. On average, the sample was 14.28 years old (SD = 0.38) in 8th grade (W4). Twenty-nine percent received free/reduced price school lunch on at least one data collection occasion (a proxy for socioeconomic disadvantage). We performed four χ2 tests and three t-tests to compare the analytic sample to the rest of the PROSPER sample on demographic and friendship variables. All of these tests were significant. Participants in the analytic sample were more likely to be male (p < .05), living with both parents at W4 (p < .001), and were less likely to have received free/reduced lunch on at least one occasion (p < .001). Participants in the analytic sample were less likely to be Hispanic (p < .001), Black (p < .001), or to report other/multiple race/ethnicities (p < .001), but did not differ in their likelihood of being Asian (p > .05). Participants in the analytic sample were slightly younger at W4 than those not in the analytic sample (p < .01) and tended to nominate more friends at each wave (p < .001 at each wave).

Measures

Friends, romantic partners, and romantic partners’ friends.

At each wave, participants nominated up to seven same-grade friends (mean friendship nominations = 4.03). Participants responded to the question, “Who are your best and closest friends in your grade? Spell out the names the best you can.” Project staff used school rosters to match the names that participants reported to the names of other participants in the PROSPER study, creating ties between participants. For the present analyses, outdegree (number of outgoing friendship nominations) indicated friendships. If a participant nominated someone as a friend, their relationship was recorded as a friendship, even if this friendship nomination was not reciprocated. Thus, in this paper, an individual’s network of (same grade) friends is composed of every student s/he nominated as a friend. Descriptive statistics for outdegree by wave and gender are reported in Table 1.

Table 1.

Descriptive Statistics by Wave and Gender

Boys Girls
Mean or proportion SD Mean or proportion SD
Outdegree (number of friendship nominations made in school)
8th grade (W4) 4.73 1.72 5.33 1.39
9th grade (W5) 4.45 1.79 5.05 1.66
10th grade (W6) 4.05 1.74 4.44 1.75
11th grade (W7) 3.82 1.87 4.33 1.73
12th grade (W8) 3.76 1.82 3.89 1.76
Frequency of drunkenness
8th grade 1.34 0.77 1.20 0.58
9th grade 1.70 1.22 1.48 0.92
10th grade 1.76 1.23 1.48 0.92
11th grade 2.04 1.31 1.80 1.04
12th grade 2.27 1.44 2.01 1.21
Friends’ frequency of drunkenness
8th grade 1.38 0.51 1.29 0.43
9th grade 1.69 0.82 1.56 0.64
10th grade 1.88 0.87 1.80 0.86
11th grade 2.15 0.96 2.06 0.93
12th grade 2.38 1.03 2.25 0.99
Partners’ frequency of drunkenness
8th grade 1.33 0.79 1.32 0.78
9th grade 1.52 0.98 1.75 1.21
10th grade 1.59 1.02 1.61 1.10
11th grade 1.85 1.13 2.00 1.29
12th grade 2.05 1.23 2.26 1.46
Partners’ friends’ frequency of drunkenness
8th grade 1.37 0.51 1.38 0.52
9th grade 1.65 0.77 1.72 0.87
10th grade 1.81 0.86 1.90 0.94
11th grade 2.06 0.97 2.11 0.99
12th grade 2.28 1.06 2.42 1.07
Friends’ alcohol-related attitudes
8th grade 1.79 0.52 1.67 0.45
9th grade 2.10 0.59 1.97 0.53
10th grade 2.18 0.61 2.09 0.62
11th grade 2.32 0.69 2.14 0.59
12th grade 2.40 0.79 2.24 0.64
Partners’ alcohol-related attitudes
8th grade 1.68 0.73 1.78 0.77
9th grade 1.92 0.84 2.10 0.99
10th grade 1.90 0.81 2.04 0.93
11th grade 1.94 0.80 2.29 1.00
12th grade 2.15 0.85 2.35 1.00
Partners’ friends’ alcohol-related attitudes
8th grade 1.72 0.47 1.81 0.52
9th grade 2.06 0.59 2.15 0.66
10th grade 2.10 0.64 2.18 0.71
11th grade 2.15 0.57 2.29 0.70
12th grade 2.28 0.66 2.47 0.77
Unstructured socializing with friends
8th grade 3.05 1.00 2.87 1.00
9th grade 3.15 1.03 2.80 0.88
10th grade 3.12 0.98 2.85 0.89
11th grade 3.28 0.88 2.91 0.85
12th grade 3.51 0.90 2.94 0.90
Unstructured socializing with partners
8th grade 3.12 1.30 2.88 1.40
9th grade 3.14 1.36 3.12 1.36
10th grade 3.39 1.41 3.58 1.33
11th grade 3.70 1.38 3.90 1.14
12th grade 4.12 1.22 4.19 1.09
Partners’ unstructured socializing with friends
8th grade 2.94 0.96 3.16 1.06
9th grade 2.93 0.94 3.25 1.01
10th grade 2.95 0.89 3.24 0.95
11th grade 2.93 0.92 3.36 0.89
12th grade 2.98 0.88 3.44 0.87
School grades
8th grade 4.12 0.84 4.39 0.72
9th grade 4.01 0.84 4.32 0.78
10th grade 4.08 0.89 4.33 0.73
11th grade 4.13 0.83 4.41 0.65
12th grade 4.20 0.77 4.61 0.57
Free/reduced price school lunch at any wave .27 .25
Lives with both parents W4 .71 .67
Open in a new tab

Participants also reported the name of their current or most recent same-grade boyfriend or girlfriend, if they had any within the past year. Project staff matched these romantic nominations to the names of other participants in the PROSPER study, creating ties between participants. Like friendship ties, romantic ties were based on the single outgoing nomination. Romantic partners’ friends refer to the individuals that a participant’s romantic partner nominated as friends.

Some individuals reported out-of-grade or out-of-school friends and romantic partners; these nominations are not included in analyses because there are no behavioral data from these friends and partners. Of the entire PROSPER sample, 96% of friendship nominations made were matched to an in-grade friend. On 78% of occasions when individuals indicated that they had a romantic partner, the romantic partner was matched to an in-grade romantic partner.

We chose to focus on outgoing nominations for both theoretical and practical reasons. From a theoretical standpoint, an individual’s perceived friends/partners should be the people s/he considers as sources of behavioral and attitudinal norms, whether or not those relationships are perceived mutually. In support of this statement, past research on friends and alcohol use has found socialization effects on alcohol use regardless of whether the friendship was reciprocated (Bot, Engels, Knibbe, & Meeus, 2005; Giletta et al., 2012).

From a measurement standpoint, reciprocated nominations may not be a valid measure of reciprocated romantic relationships for our sample. In our question about romantic involvement, we asked about current or most recent partner. If a couple broke up and one person had a subsequent relationship, person A’s most recent partner is not person B’s most recent partner, even if both people agreed that they had previously been in a romantic relationship with each other. Therefore, we may not be able to appropriately assess all past reciprocated romantic relationships.

Frequency of drunkenness.

At each wave, participants responded to the question, “During the past year, how many times have you been drunk from drinking beer, wine, wine coolers, or other liquor?” Response options ranged from 1 (Not at all) to 5 (More than 12 times). For each participant, friends’ frequency of drunkenness was the average response across all of the participant’s nominated friends in a particular wave of data collection. Romantic partner’s frequency of drunkenness was the response of the participant’s nominated romantic partner. Romantic partner’s friends’ frequency of drunkenness was the average response across all the nominated friends of a participant’s romantic partner. We report descriptive statistics by wave and gender in Table 1. We use a self-reported measure of drunkenness rather than other measures of heavy alcohol use (such as number of drinks, having more than 4/5 drinks, or estimates of blood alcohol content) because it emphasizes the subjective state of intoxication, and therefore may be more strongly associated with negative outcomes of alcohol use than measures that rely on quantity of alcohol consumed (Midanik, 1999). However, we recognize that a drawback of this measure is that it does not account for differing levels of drunkenness, and that one person’s subjective state of drunkenness may differ from others’ perceptions of what it means to be drunk.

Alcohol-related attitudes.

Participants answered a 6-item measure of alcohol-related attitudes at each wave (Shin, 2011). The scale included one question about how wrong participants think it is for someone their age to drink beer, wine, or liquor, on a scale of 1 (Not at all wrong) to 4 (Very wrong); one question about likelihood of saying no when someone tried to get them to drink on a scale of 1 (Definitely would say no) to 5 (Definitely would not say no); and four questions about their beliefs about alcohol consequences (e.g., “Drinking alcohol lets you have more fun,”) on a scale of 1 (Strongly disagree) to 5 (Strongly agree). The average of these six items assessed positive alcohol-related attitudes. We calculated friends’, romantic partners’, and partners’ friends’ attitudes in the same manner as frequency of drunkenness. Table 1 includes descriptive statistics. Reliability was acceptable for both boys (α ranged from .70 to .80 across waves) and girls (α .70 to .77).

Unstructured socializing.

At each wave, for each nominated friend and romantic partner, participants responded to the question, “How often do you spend time just hanging out with this person outside of school (without adults around)?” (Siennick & Osgood, 2012). Response options ranged from 1 (Never) to 5 (Almost every day). For unstructured socializing with friends, we averaged responses across each nominated friend to create a measure of how frequently participants engaged in unstructured socializing with all of their friends. Unstructured socializing with romantic partners refers to how frequently participants reported unstructured socializing with their romantic partner. For romantic partners’ unstructured socializing with friends, we calculated the average frequency that an individual’s reported romantic partner spent socializing with each of his/her nominated friends. The measure for partners’ friends’ unstructured socializing differs conceptually from participants’ unstructured socializing with friends and romantic partners. The participant may or may not be present for romantic partners’ unstructured socializing with friends. However, the data do not allow us to create a measure of participants’ unstructured socializing with partners’ friends. We report descriptive statistics in Table 1.

School grades.

At each wave, participants reported what grades they usually get in school, on a scale of 1 [Mostly A’s (90–100)] to 5 [Mostly lower than D’s (below 60–69)]. We reverse coded these responses so that higher scores indicate higher grades (see Table 1).

Free/reduced price lunch.

At each wave, participants responded to the question, “What do you usually do for lunch on school days?” We coded participants as 1 (Receives free/reduced price lunch) or 0 (Does not receive free/reduced price lunch). We combined these responses across waves to indicate whether a participant received free/reduced price lunch on at least one wave (see Table 1).

Two-parent household.

At each wave, participants responded to the question, “Who do you live with most of the year?” We coded students who responded that they lived with their mother and father as 1 (Living with both parents). We coded students who responded that they lived with a parent and a step-parent, only their mother, only their father, or in another type of household at W4 as 0 (Not living with both parents). Because this measure was highly stable over time, we used participants’ response at W4 (see Table 1).

Analytic Plan

The data are interdependent in several ways. First, measurement occasions are nested within individuals, and participants’ responses are not independent across measurement occasions—for example, one person’s response at W4 is likely associated with his/her response at W5. Second, individuals are nested within schools. It is likely that there is interdependence within schools—that is, students’ behaviors and attitudes are more similar to other students in their school than to students at other schools. Third, the relationships between students introduce interdependence in responses. For example, if A and B are romantic partners, A’s value for partners’ friends’ attitudes is equal to B’s value for friends’ attitudes. In addition, if C and D are friends, C’s friends’ attitudes are likely similar to D’s friends’ attitudes because of overlap in friendship networks.

It is necessary to account for these multiple layers of interdependence to produce accurate coefficient estimates. Therefore, we assessed change in alcohol use using multilevel models in STATA (the xtreg procedure) with bootstrapped standard errors (50 iterations) clustered by community and study cohort. The models measure within-person changes in frequency of drunkenness, accounting for the nesting of measurement occasions within individuals. The models include predictors at two levels. At Level 1 (wave of data collection), we assessed whether within-person changes in peer characteristics are associated with within-person changes in frequency of drunkenness. At Level 2 (person), we controlled for between-person effects of peers’ frequency of drunkenness, alcohol-related attitudes, and unstructured socializing. In order to create Level 2 variables for peers’ frequency of drunkenness, alcohol-related attitudes, and unstructured socializing, we averaged each variable for the participant across measurement occasions. This strategy for analyzing longitudinal data measures within-person variability in frequency of drunkenness, essentially treating each participant as his/her own control. With this strategy, we are able to rule out time-stable confounding factors, bringing us closer to causal associations between peer characteristics and adolescents’ frequency of drunkenness. Bootstrapping accounts for the nesting of students within schools within measurement occasions and mitigates the problems posed by the interdependence of friendship networks (Snijders & Borgatti, 1999). We conducted all analyses separately for boys and girls to avoid the interdependence of romantic partners’ responses. In addition to these steps to account for interdependence, we controlled for wave of data collection, school grades, outdegree (number of friends nominated), romantic partner’s outdegree, whether the participant received free lunch at any wave, and whether the participant lived in a two-parent household at W4.

The data pose an additional challenge in that measures of romantic partners’ and partners’ friends’ attitudes, frequency of drunkenness, and unstructured socializing only exist when an individual reports a romantic partner. Therefore, the analyses focus on waves in which participants report a romantic partner. The analyses address the question, “When adolescents have a romantic partner, are the characteristics of friends, romantic partners, and romantic partners’ friends associated with changes in frequency of drunkenness, compared to the last time they reported a (same or a different) romantic partner?”

In Model 1, we predicted within-person change in adolescents’ frequency of drunkenness from their friends’, romantic partners’, and romantic partners’ friends’ frequency of drunkenness. In Model 2, we predicted within-person change in adolescents’ frequency of drunkenness from their friends’, romantic partners’, and romantic partners’ friends’ alcohol-related attitudes. In Model 3, we predicted within-person change in adolescents’ frequency of drunkenness from their unstructured socializing with friends, unstructured socializing with romantic partners, and romantic partners’ unstructured socializing with friends. Models 1–3 allowed us to test whether different peer relationships are uniquely associated with change in adolescents’ frequency of drunkenness. In Model 4, we predicted within-person change in adolescents’ frequency of drunkenness from all others’ frequency of drunkenness, alcohol-related attitudes, and unstructured socializing. This model allowed us to test whether peers’ frequency of drunkenness, attitudes, and unstructured socializing all contribute uniquely to changes in adolescents’ frequency of drunkenness. In unreported models, we added interactions of each predictor (others’ frequency of drunkenness, attitudes, and unstructured socializing) by time to Model 4 by multiplying the value of each predictor by wave of data collection (which is coded so that W4 = 0, W5 = 1, etc.). We tested each interaction separately, for a total of 18 models (9 each for boys and girls). Model 5 includes all of the time interactions that were statistically significant in these models.

Results

We first estimated multilevel models with no predictors (i.e., null models) to assess the intraclass correlation for frequency of drunkenness. The intraclass correlation coefficient for our outcome was .41 for boys and .46 for girls, indicating, for instance, that 41% of the variance in boys’ frequency of drunkenness was at the within-person, versus the between-person, level. Results for all multilevel models are reported in Tables 2 (for boys) and 3 (for girls).

Table 2.

Multilevel Model Results Predicting Boys’ Frequency of Drunkenness (N = 780 participants, 1603 observations)

Model 1 Model 2 Model 3 Model 4 Model 5
Peers’ drunkenness Peers’ attitudes Unstructured socializing Full model Full model plus time interactions
Coefficient bootstrapped SE Coefficient bootstrapped SE Coefficient bootstrapped SE Coefficient bootstrapped SE coefficient bootstrapped SE
Within-person effects
 Friends’ drunkenness 0.29*** 0.07 0.28*** 0.05 0.27*** 0.07
 Partners’ drunkenness 0.10* 0.05 0.08 0.05 0.07 0.06
 Partners’ friends’ drunkenness 0.20** 0.06 0.22*** 0.06 0.21** 0.07
 Friends’ attitudes 0.22** 0.07 0.02 0.06 0.01 0.07
 Partners’ attitudes 0.14*** 0.04 0.09* 0.04 0.08 0.04
 Partners’ friends’ attitudes 0.15* 0.07 −0.05 0.07 −0.04 0.06
 Unstructured socializing with friends 0.02 0.03 <−0.01 0.03 −0.08 0.04
 Unstructured socializing with partners −0.01 0.02 <−0.01 0.02 <0.01 0.02
 Romantic partners’ unstructured socializing with friends 0.02 0.04 <−0.01 0.04 <0.01 0.04
 Time* unstructured socializing with friends 0.05* 0.02
 Time 0.09*** 0.02 0.16*** 0.02 0.25*** 0.02 0.09*** 0.02 −0.08 0.05
 School grades −0.10 0.07 −0.12 0.07 −0.15** 0.06 −0.10 0.06 −0.10 0.08
 Outdegree −0.01 0.02 −0.02 0.03 −0.02 0.03 −0.01 0.02 −0.01 0.02
 Partner’s outdegree −0.01 0.02 −0.01 0.02 −0.02 0.01 −0.01 0.02 −0.01 0.02
Between-person effects
 Free/reduced lunch −0.05 0.07 −0.04 0.08 <−0.01 0.07 −0.02 0.08 −0.02 0.08
 Lives with both parents 0.03 0.07 0.05 0.06 0.02 0.09 0.02 0.06 0.03 0.07
 Grades averaged across waves −0.05 0.08 −0.03 0.08 −0.02 0.10
 Friends’ drunkenness averaged across waves 0.15 0.10 0.12 0.10 0.13 0.12
 Partners’ drunkenness averaged across waves <0.01 0.08 0.01 0.07 0.01 0.09
 Partners’ friends’ drunkenness averaged across waves 0.13* 0.06 0.08 0.11 0.07 0.09
 Friends’ attitudes averaged across waves 0.22* 0.09 <0.01 0.10 0.01 0.10
 Partners’ attitudes averaged across waves 0.01 0.08 −0.08 0.10 −0.08 0.08
 Partners’ friends’ attitudes averaged across waves 0.22* 0.09 0.08 0.14 0.08 0.09
 Unstructured socializing with friends averaged across waves .28*** 0.05 0.16*** 0.06 0.17** 0.05
 Unstructured socializing with partners averaged across waves −.03 0.04 −0.02 0.05 −0.02 0.04
 Romantic partners’ unstructured socializing with friends averaged across waves .12* 0.06 0.08 0.06 0.08 0.05
 Outdegree averaged across waves −0.04 0.02 −0.02 0.04 −0.01 0.03 −0.03 0.03 −0.03 0.03
 Partner’s outdegree averaged across waves 0.05 0.03 0.05 0.03 0.05 0.03 0.04 0.03 0.04 0.03
Variance components
 Within-person .65 .70 .75 .64 .64
 Between-person .79 .81 .83 .79 .79
 ICC .40 .43 .45 .40 .40
Open in a new tab

Notes:

*

p < .05,

**

p < .01,

***

p < .001.

Table 3.

Multilevel Model Results Predicting Girl’ Frequency of Drunkenness (N = 659 participants, 1,340 observations)

Model 1 Model 2 Model 3 Model 4 Model 5
Peers’ drunkenness Peers’ attitudes Unstructured socializing Full model Full model plus time interactions
coefficient
 bootstrapped SE coefficient bootstrapped SE coefficient bootstrapped SE coefficient bootstrapped SE coefficient bootstrapped SE
Within-person effects
 Friends’ drunkenness 0.31*** 0.05 0.33*** 0.06 0.29*** 0.06
 Partners’ drunkenness 0.08** 0.03 0.05 0.03 0.05 0.03
 Partners’ friends’ drunkenness 0.13*** 0.03 0.13*** 0.04 0.12** 0.04
 Friends’ attitudes 0.25*** 0.06 −0.02 0.06 −0.01 0.06
 Partners’ attitudes 0.12* 0.05 0.08* 0.04 0.07 0.04
 Partners’ friends’ attitudes 0.11* 0.05 −0.04 0.05 −0.03 0.05
 Unstructured socializing with friends −0.01 0.04 <−0.01 0.03 −0.07 0.04
 Unstructured socializing with partners 0.02 0.02 0.01 0.02 0.02 0.02
 Romantic partners’ unstructured socializing with friends 0.03 0.03 0.02 0.02 −0.04 0.03
 Time* unstructured socializing with friends 0.05** 0.02
 Time*romantic partners’ unstructured socializing with friends 0.05** 0.01
 Time 0.07*** 0.02 0.13*** 0.02 0.20*** 0.02 0.06** 0.02 −0.22** 0.06
 School grades −0.14* 0.06 −0.17* 0.07 −0.22*** 0.06 −0.15* 0.07 −0.13* 0.06
 Outdegree −0.01 0.02 −0.01 0.03 <−0.01 0.02 −0.02 0.02 −0.01 0.02
 Partner’s outdegree 0.01 0.01 0.01 0.01 <−0.01 0.01 0.01 0.01 0.01 0.01
Between-person effects
 Free/reduced lunch −0.06 0.07 −0.05 0.09 −0.07 0.08 −0.06 0.06 −0.06 0.07
 Lives with both parents −0.13 0.07 −0.13 0.07 −0.16 0.08 −0.13 0.07 −0.13 0.07
 Grades averaged across waves −0.04 0.07 <−0.01 0.07 −0.03 0.08 −0.03 0.08 −0.04 0.06
 Friends’ drunkenness averaged across waves 0.13* 0.07 0.08 0.08 0.11 0.08
 Partners’ drunkenness averaged across waves 0.08 0.05 0.10 0.06 0.09 0.06
 Partners’ friends’ drunkenness averaged across waves 0.03 0.05 0.04 0.08 0.05 0.08
 Friends’ attitudes averaged across waves 0.26* 0.12 0.06 0.13 0.04 0.11
 Partners’ attitudes averaged across waves 0.01 0.06 −0.10 0.06 −0.08 0.07
 Partners’ friends’ attitudes averaged across waves 0.08 0.08 −0.01 0.08 −0.02 0.10
 Unstructured socializing with friends averaged across waves 0.12 0.07 0.05 0.05 0.04 0.05
 Unstructured socializing with partners averaged across waves 0.04 0.04 −0.01 0.03 <0.01 0.03
 Romantic partners’ unstructured socializing with friends averaged across waves 0.15** 0.05 0.03 0.04 0.03 0.04
 Outdegree averaged across waves 0.03 0.03 0.03 0.03 0.04 0.03 0.03 0.03 0.03 0.02
 Partner’s outdegree averaged across waves 0.02 0.02 0.02 0.03 0.03 0.02 0.02 0.02 0.02 0.02
Variance components
 Within-person .50 .57 .63 .49 .50
 Between-person .60 .62 .64 .60 .59
 ICC .41 .46 .49 .40 .42
Open in a new tab

Notes:

*

p < .05,

**

p < .01,

***

p < .001

Peers’ Frequency of Drunkenness

In Model 1, which tested the associations between changes in peers’ frequency of drunkenness and adolescents’ own frequency of drunkenness, boys and girls were drunk more frequently at waves when their friends, romantic partners, and romantic partners’ friends were drunk more frequently.

Peers’ Alcohol-related Attitudes

In Model 2, which tested the associations between peers’ alcohol-related attitudes and adolescents’ frequency of drunkenness, both boys and girls were drunk more frequently at waves when their friends, romantic partners, and partners’ friends held more positive alcohol-related attitudes.

Unstructured Socializing with Peers

In Model 3, which tested the associations between unstructured socializing and adolescents’ frequency of drunkenness, there were no significant coefficients for unstructured socializing predicting changes in boys’ or girls’ frequency of drunkenness.

Combined Models

In Model 4 (the fully specified model), the coefficients for adolescents’ friends’ and partners’ friends’ frequency of drunkenness remained significant; boys and girls were drunk more frequently at waves when their friends and romantic partners’ friends were drunk more frequently. However, partners’ frequency of drunkenness no longer significantly predicted adolescents’ own frequency of drunkenness. There were no significant time interactions for peers’ frequency of drunkenness for either girls or boys; thus, these interactions are not included in Model 5.

For peers’ attitudes, in Model 4, only the estimate of romantic partners’ attitudes remained statistically significant. Both boys and girls were drunk more frequently at waves when their romantic partners held more positive alcohol-related attitudes. There were no significant time interactions for peers’ alcohol-related attitudes for either girls or boys; thus, these interactions are not included in Model 5. However, in Model 5, which includes time interactions for other predictors, romantic partners’ attitudes no longer significantly predicted boys’ or girls’ frequency of drunkenness.

There were no significant coefficients for unstructured socializing predicting changes in boys’ or girls’ frequency of drunkenness in Model 4, consistent with the results from Model 3. Examining time interactions revealed that, for boys and girls, there was a significant time*unstructured socializing with friends interaction. This interaction is included in Model 5. The addition of an interaction term changes the interpretation of the main effect of unstructured socializing with friends. In Model 5, this coefficient should be interpreted as the effect of unstructured socializing with friends at the first occasion of data collection used in the analyses, W4. The time interaction represents the change in the main effect in each subsequent wave of data collection. Together, the main effect and interaction coefficients indicated that, although the association between unstructured socializing with friends and adolescents’ frequency of drunkenness was not significant at W4, this association increased over time. Thus, older adolescents increased their drunkenness frequency when they engaged in more frequent unstructured socializing with friends. In addition, for girls, there was a significant interaction for time*romantic partners’ unstructured socializing with friends. This interaction indicated that, although romantic partners’ unstructured socializing with friends and girls’ frequency of drunkenness were not significantly associated at W4, this association increased over time. Thus, older girls increased their drunkenness frequency when their romantic partners engaged in more frequent unstructured socializing with friends.

Gender Comparisons

We compared the coefficients of each predictor for boys versus girls in Models 1–4 using a series of z-tests. None of the coefficients for peers’ drunkenness, attitudes, or unstructured socializing differed significantly for boys versus girls (z ranging from −0.99 to 1.08, ps > .05), indicating that there were no gender differences in the estimates.

Discussion

In the present paper, we examined how multiple peer relationships and mechanisms of social influence are associated with changes in adolescents’ frequency of drunkenness. When examined separately, changes in friends’, romantic partners’ and romantic partners’ friends’ frequency of drunkenness and alcohol-related attitudes each predicted changes in adolescents’ own frequency of drunkenness. However, in a combined model, friends’ and partners’ friends’ frequency of drunkenness predicted changes in adolescents’ frequency of drunkenness, whereas romantic partners’ attitudes predicted frequency of drunkenness. Changes in unstructured socializing with peers increasingly predicted changes in frequency of drunkenness as adolescents aged. The results expand understanding of the social transmission of alcohol use in adolescence and inform future intervention efforts.

Diverse Peer Relationships and Drunkenness in Adolescence

Past research on social influences on adolescent alcohol use has typically examined one relationship type (e.g., Cheadle et al., 2015; Osgood et al., 2013) and/or one mechanism of influence (Kreager et al., 2013; Kreager & Haynie, 2011). Examining multiple peer relationships and mechanisms of social influence in the same analyses indicates both who is important for alcohol use and how they are important.

Regarding who is important, the findings of the present research indicate that multiple peer relationships are important in determining adolescents’ frequency of drunkenness. Each type of peer relationship (friends, romantic partners, and romantic partners’ friends) was significantly correlated with changes in adolescents’ frequency of drunkenness in at least one model. This result extends prior research that has separately demonstrated that characteristics of friends, romantic partners, and partners’ friends predict adolescents’ alcohol use (e.g., Ragan, 2014; Kreager et al., 2013; Osgood et al., 2013) by demonstrating that each association is independent of the estimates of other peer relationships.

The findings of the present research highlight that the mechanisms of social influence on alcohol use may differ for friends, romantic partners, and partners’ friends. Although past research indicates that peers’ alcohol use and alcohol-related attitudes both predict adolescents’ alcohol use (Burk, Van Der Vorst, Kerr, & Stattin, 2012; Cheadle et al., 2015; Borsari & Carey, 2003; Ragan, 2014), we found that within-person changes in friends’ and partners’ friends’ frequency of drunkenness accounted for the association between these peers’ attitudes and changes in adolescent drunkenness frequency. In contrast, changes in romantic partners’ attitudes were uniquely associated with changes in adolescents’ frequency of drunkenness, accounting for partners’ frequency of drunkenness. Stated simply, what friends and partners’ friends do matters more for individuals’ drunkenness than what friends and partners’ friends think. This finding is consistent with research finding that friends’ deviant behaviors, but not friends’ attitudes toward deviant behaviors, are independently associated with adolescents’ own deviant behaviors (Warr & Stafford, 1991). However, what romantic partners think matters more than what they do for adolescents’ drunkenness.

Together, these findings advance both social learning and differential association explanations of social influences on alcohol use, demonstrating that each theory may explain different types of social influence. Whereas social learning explanations are key to understanding friends’ and partners’ friends’ influence on alcohol use, differential association theory may be more applicable to romantic partners than to other peers. Characteristics of these relationships may explain their differential mechanisms of social influence. Friends and partners’ friends may serve as role models for drinking more than romantic partners do (Yancey, Grant, Kurosky, Kravitz-Wirtz, & Mistry, 2011), which may explain the increased importance of behavior modeling in explaining friends’ and partners’ friends’ associations with adolescents’ own alcohol use.

In contrast, heterosexual adolescents may be less likely to view their romantic partners as models of alcohol use behavior because they are a different gender, and therefore different behavioral norms for alcohol use apply to them (de Visser & McDonnell, 2012). However, adolescents may find it important to adhere to partners’ alcohol-related attitudes in order to maintain their partners’ approval. For instance, college women’s beliefs about whether men want them to use alcohol are associated with women’s own alcohol use (Hummer, LaBrie, Lac, Sessoms, & Cail, 2012). In addition, college women report adapting their alcohol use to what they perceive will make a favorable impression on men (Young, Morales, McCabe, Boyd, & D’Arcy, 2005). Together, these findings suggest that maintaining partners’ approval is important, and adjusting drinking to align with partners’ attitudes is one way to maintain this approval.

Developmental Changes in the Mechanisms of Social Influence

In addition to findings for peers’ frequency of drunkenness and alcohol-related attitudes, we also found that changes in unstructured socializing may increase drunkenness frequency later in adolescence. Although changes in unstructured socializing with friends were not associated with changes in adolescents’ frequency of drunkenness in 8th grade, unstructured socializing with friends predicted increased frequency of drunkenness later in adolescence. In addition, changes in girls’ romantic partners’ unstructured socializing with friends increasingly predicted changes in girls’ frequency of drunkenness over time. As adolescents age, obtaining alcohol becomes easier because they gain increasing access to social sources such as older friends and parties (Harrison, Fulkerson, & Park, 2000). Because alcohol becomes more accessible with age, unstructured socializing may provide more opportunities for alcohol use later in adolescence. In addition, because rates of alcohol use increase with age (Johnston et al., 2016), the purpose of unstructured socializing may change to be increasingly centered on alcohol use. As unstructured socializing increasingly becomes a context for drinking alcohol, more time spent in unstructured socializing is linked more strongly to alcohol use.

Prevention Implications

The results of this research can inform the targets and content of alcohol use interventions. Regarding targets, the results suggest that in addition to focusing on friends, interventions aimed at reducing social influence on alcohol use may address the roles of romantic partners’ and partners’ friends in determining alcohol use. Regarding content, the present findings suggest that intervention content should be tailored according to relationship type and adolescents’ ages.

For addressing friends’ and partners’ friends’ influence on drunkenness, our results suggest that focusing on these peers’ alcohol-related behavior (including frequency of drunkenness and unstructured socializing), rather than alcohol-related attitudes, may be beneficial. For example, interventions that alter the structure of friendship networks to discourage ties with individuals who drink alcohol (e.g., Valente, Gallaher, & Mouttapa, 2004; Valente et al., 2007) may be sufficient to produce change in frequency of drunkenness; altering the attitudes of central people to be less favorable toward alcohol use may not produce additional change in drunkenness. However, for romantic partners’ influence on drunkenness, addressing how attitudes may influence alcohol use may be key to intervention success. For example, relationship education programs often teach decision-making skills to help individuals choose healthy relationships (e.g., Fincham, Stanley, & Rhoades, 2011; Holt et al., 2016). These programs may advise adolescents to choose romantic partners whose attitudes toward alcohol use align with individuals’ own values, and avoid partners who have more favorable alcohol-related attitudes than their own.

Clinical interventions, which are predominantly aimed at older adolescents, have used normative feedback as a tool for decreasing individuals’ motivation to drink heavily (Lewis & Neighbors, 2006). Clinicians who work with adolescents could use similar strategies, informing clients that their peers drink less, or have less favorable attitudes toward drinking, than clients perceive they do. For both clinical practice and larger scale interventions, the findings of the present research support tailoring messages about friends and partner’s friends versus romantic partners. For individuals with other-gender romantic partners, educating clients/participants about partners’ attitudes may improve intervention outcomes. This strategy may be particularly useful for individuals whose partners have stronger anti-alcohol attitudes than clients/participants do themselves.

Intervention strategies should also be tailored according to participants’ ages. In particular, interventions to reduce opportunities for unstructured socializing (e.g., Smith, 2007; Tebes et al., 2007) may be more effective in reducing drunkenness for older adolescents. Leisure educations programs conducted with younger adolescents have demonstrated effectiveness in improving adolescents’ ability to restructure boredom, take initiative, and participate in activities (Caldwell, Baldwin, Walls, & Smith, 2004). These effects may translate to less alcohol use later in adolescence. To ensure that early intervention effects lead to decreased alcohol use, practitioners may consider adding booster sessions later in adolescence, when unstructured socializing is more strongly associated with alcohol use.

Limitations

The results of this research must be interpreted in light of its limitations. First, it is impossible to determine the extent to which covariation of peers’ drunkenness, attitudes, and unstructured socializing with adolescents’ own drunkenness is due to selection (individuals choose peers who are similar to them in alcohol use) versus influence (individuals become more similar in alcohol use to their peers over time). Selection into friendships and romantic partnerships may partially account for any associations between peers’ behaviors/ attitudes/ unstructured socializing and individuals’ alcohol use (e.g., Brechwald & Prinstein, 2011; Kreager & Haynie, 2011; Rhule-Louie & McMahon, 2007). The fact that individuals may have had different romantic partners and friends at each wave of data collection increases the possibility of selection bias. Although some statistical tools can help distinguish selection and influence processes with longitudinal social network data (e.g., SIENA; Steglich, Snijders, & West, 2006), we were unable to use these tools because of the sparseness of romantic connections within the networks. Without accounting for selection effects, estimates of friends’, partners’, and partners’ friends’ influence may be inflated. Thus, although theoretical discussions about peers’ behavior, attitudes, and unstructured socializing propose that these are mechanisms of social influence, the present research does not offer concrete evidence that these mechanisms are responsible for changes in adolescents’ alcohol use.

Second, the sample and data collection method limit our ability to generalize our findings. The participants in this study are adolescents who reported a romantic partner in their same school and grade. Adolescents who have more frequent romantic relationships may differ in their susceptibility to peer or partner influence, compared to adolescents who have less frequent romantic relationships. For example, susceptibility to peer influence on alcohol use is associated with reduced stability in adolescent friendships (Allen, Porter, & McFarland, 2006). If this process is also true for romantic relationships, individuals who are more susceptible to partners’ influence on alcohol use may be less likely to be in a romantic relationship at any given point, and therefore may be underrepresented in the present analyses. In addition, the romantic relationships captured in this study may not be representative of adolescent romantic relationships because boyfriends tend to be older than girlfriends (Carver, Joyner, & Udry, 2003). Furthermore, due to the data collection strategy, we do not know about the influence of out-of-school or out-of-grade friends and romantic partners on adolescents’ frequency of drunkenness, which may differ from the influence of same-grade peers (Halpern, Kaestle, & Hallfors, 2007).

Third, sample characteristics limit the generalizability of results. The PROSPER study sampled predominantly White, heterosexual adolescents in rural communities. In addition, the data analysis approach used to account for interdependence precluded us from including same-gender romantic relationships. Future research should examine how peer relationships are associated with alcohol use in diverse samples.

Finally, our inclusion of unreciprocated nominations has important implications for demonstrated associations between peer relationships and alcohol use. Although some research has found that friendships do not need to be reciprocated in order to exert influence on adolescent substance use (Bot et al., 2005; Giletta et al., 2012), other research suggests that reciprocated relationships may exert a stronger influence on substance use (Fujimoto & Valente, 2012; Lin & Weinberg, 2014). Similarly, social influence processes may differ according to other relationship features such as closeness and length (Fujimoto & Valente, 2012). Future research may consider how different characteristics of peer relationships moderate associations of peers’ heavy alcohol use, attitudes, and unstructured socializing with adolescents’ own heavy alcohol use.

Another alternate measurement strategy for researchers to consider is to collect measures of adolescents’ unstructured socializing with their romantic partners’ friends, instead of measuring romantic partners’ unstructured socializing with friends. In addition, researchers might examine additional dimensions of alcohol use (such as frequency of any alcohol use, binge drinking, and alcohol-related consequences) and other risk behaviors (such as risky sexual behavior and other substance use) in order to determine whether the same social influence processes account for multiple outcomes.

Conclusion

Despite its limitations, the present research offers valuable contributions to understanding how diverse peer relationships are associated with alcohol use during adolescence. The results highlight that, although multiple peer relationships are associated with alcohol use, the mechanisms of influence may differ across relationship types. Friends’ and partners’ friends’ frequency of drunkenness, and romantic partners’ attitudes, were important in predicting adolescents’ frequency of drunkenness. In addition, the mechanisms of social influence differ across adolescence; unstructured socializing with peers increasingly predicted frequency of drunkenness as adolescents aged. These findings suggest that intervention efforts to reduce adolescent drunkenness should be tailored to different types of peer relationships. In addition, intervention strategies may differ for younger versus older adolescents.

Acknowledgments

This research was supported by grants R01 DA013709 and R01 DA018225 from the National Institute on Drug Abuse, and grants P30 MH0522776 and T32 MH019985 from the National Institute on Mental Health. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institute on Drug Abuse, the National Institute on Mental Health, or the National Institutes of Health.

Footnotes

All authors have declared that they have no conflicts of interest.

References

  1. Akers R (2017). Social learning and social structure: A general theory of crime and deviance. New York, NY: Routledge. [Google Scholar]
  2. Akers RL, Krohn MD, Lanza-Kaduce L, & Radosevich M (1979). Social learning and deviant behavior: A specific test of a general theory. American Sociological Review, 44, 636–655. doi: 10.2307/2094592 [DOI] [PubMed] [Google Scholar]
  3. Allen JP, Porter MR, & McFarland FC (2006). Leaders and followers in adolescent close friendships: Susceptibility to peer influence as a predictor of risky behavior, friendship instability, and depression. Development and Psychopathology, 18, 155–172. doi: 10.1017/S0954579406060093 [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Barnes GM, Hoffman JH, Welte JW, Farrell MP, & Dintcheff BA (2007). Adolescents’ time use: Effects on substance use, delinquency and sexual activity. Journal of Youth and Adolescence, 36, 697–710. doi: 10.1007/s10964-006-9075-0 [DOI] [Google Scholar]
  5. Borsari B, & Carey KB (2003). Descriptive and injunctive norms in college drinking: A meta-analytic integration. Journal of Studies on Alcohol, 64, 331–341. doi: 10.15288/jsa.2003.64.331 [DOI] [PMC free article] [PubMed] [Google Scholar]
  6. Bot SM, Engels RC, Knibbe RA, & Meeus WH (2005). Friend’s drinking behaviour and adolescent alcohol consumption: The moderating role of friendship characteristics. Addictive Behaviors, 30, 929–947. doi: 10.1016/j.addbeh.2004.09.012 [DOI] [PubMed] [Google Scholar]
  7. Brechwald WA, & Prinstein MJ (2011). Beyond homophily: A decade of advances in understanding peer influence processes. Journal of Research on Adolescence, 21, 166–179. doi: 10.1111/j.1532-7795.2010.00721.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Brown SL, & Rinelli LN (2010). Family structure, family processes, and adolescent smoking and drinking. Journal of Research on Adolescence, 20, 259–273. doi: 10.1111/j.1532-7795.2010.00636.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Bruinsma G (2014). Differential association theory In Encyclopedia of criminology and criminal justice (pp. 1065–1075). New York, NY: Springer. [Google Scholar]
  10. 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, 89–98. doi: 10.15288/jsad.2012.73.89 [DOI] [PubMed] [Google Scholar]
  11. Caldwell LL, Baldwin CK, Walls T, & Smith E (2004). Preliminary effects of a leisure education program to promote healthy use of free time among middle school adolescents. Journal of Leisure Research, 36, 310–335. [Google Scholar]
  12. Carver K, Joyner K, & Udry JR (2003). National estimates of adolescent romantic relationships In Florsheim P (Ed.), Adolescent romantic relations and sexual behavior: Theory, research, and practical implications (pp. 23–56). Mahwah, NJ: Erlbaum. [Google Scholar]
  13. Centers for Disease Control. (2015). Youth risk behavior surveillance - United States, 2015. Surveillance Summaries, 65, 1–174. [DOI] [PubMed] [Google Scholar]
  14. Chan GC, Leung J, Quinn C, Kelly AB, Connor JP, Weier M, & Hall WD (2016). Rural and urban differences in adolescent alcohol use, alcohol supply, and parental drinking. The Journal of Rural Health, 32, 280–286. doi: 10.1111/jrh.12151 [DOI] [PubMed] [Google Scholar]
  15. Cheadle JE, Walsemann KM, & Goosby BJ (2015). Teen alcohol use and social networks: The contributions of friend influence and friendship selection. Journal of Alcoholism and Drug Dependence, 3, 224–233. doi: 10.4172/2329-6488.1000224 [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Cohen DA, Farley TA, Taylor SN, Martin DH, & Schuster MA (2002). When and where do youths have sex? The potential role of adult supervision. Pediatrics, 110, e66–e72. doi: 10.1542/peds.110.6.e66 [DOI] [PubMed] [Google Scholar]
  17. Collins WA, Welsh DP, & Furman W (2009). Adolescent romantic relationships. Annual Review of Psychology, 60, 631–652. doi: 10.1146/annurev.psych.60.110707.163459 [DOI] [PubMed] [Google Scholar]
  18. Connolly J, & McIsaac C (2011). Romantic relationships in adolescence In Underwood MK & Rosen LH (Eds.), Social development: Relationships in infancy, childhood, and adolescence (pp. 180–206). New York, NY: Guilford Press. [Google Scholar]
  19. de Visser RO, & McDonnell EJ (2012). ‘That’s OK. He’s a guy’: A mixed-methods study of gender double-standards for alcohol use. Psychology & Health, 27, 618–639. doi: 10.1080/08870446.2011.617444 [DOI] [PubMed] [Google Scholar]
  20. Fincham FD, Stanley SM, & Rhoades GK (2011). Relationship education in emerging adulthood: Problems and prospects In Fincham FD & Cui M (Eds.), Advances in personal relationships. Romantic relationships in emerging adulthood (pp. 293–316). New York, NY: Cambridge University Press. [Google Scholar]
  21. Fischer JL, & Wiersma JD (2012). Romantic relationships and alcohol use. Current Drug Abuse Reviews, 5, 98–116. doi: 10.2174/1874473711205020098 [DOI] [PubMed] [Google Scholar]
  22. Fujimoto K, & Valente TW (2012). Decomposing the components of friendship and friends’ influence on adolescent drinking and smoking. Journal of Adolescent Health, 51, 136–143. doi: 10.1016/j.jadohealth.2011.11.013 [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Giletta M, Scholte RH, Prinstein MJ, Engels RC, Rabaglietti E, & Burk WJ (2012). Friendship context matters: Examining the domain specificity of alcohol and depression socialization among adolescents. Journal of Abnormal Child Psychology, 40, 1027–1043. doi: 10.1007/s10802-012-9625-8 [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Halpern CT, Kaestle CE, & Hallfors DD (2007). Perceived physical maturity, age of romantic partner, and adolescent risk behavior. Prevention Science, 8, 1–10. doi: 10.1007/s11121-006-0046-1 [DOI] [PubMed] [Google Scholar]
  25. Harrison PA, Fulkerson JA, & Park E (2000). The relative importance of social versus commercial sources in youth access to tobacco, alcohol, and other drugs. Preventive Medicine, 31, 39–48. doi: 10.1006/pmed.2000.0691 [DOI] [PubMed] [Google Scholar]
  26. Hawley A (1950). Human ecology: A theory of community structure. New York, NY: Ronald. [Google Scholar]
  27. Hoeben EM, Meldrum RC, & Young JT (2016). The role of peer delinquency and unstructured socializing in explaining delinquency and substance use: A state-of-the-art review. Journal of Criminal Justice, 47, 108–122. doi: 10.1016/j.jcrimjus.2016.08.001 [DOI] [Google Scholar]
  28. Holt LJ, Mattanah JF, Schmidt CK, Daks JS, Brophy EN, Minnaar PY, & Rorer K (2016). Effects of relationship education on emerging adults’ relationship beliefs and behaviors. Personal Relationships, 23, 723–741. doi: 10.1111/pere.12147 [DOI] [Google Scholar]
  29. Hummer JF, LaBrie JW, Lac A, Sessoms A, & Cail J (2012). Estimates and influences of reflective opposite-sex norms on alcohol use among a high-risk sample of college students: Exploring Greek-affiliation and gender effects. Addictive Behaviors, 37, 596–604. doi: 10.1016/j.addbeh.2011.11.027 [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Johnston LD, O’Malley PM, Miech RA, Bachman JG, & Schulenberg JE (2016). Monitoring the Future national survey results on drug use, 1975–2015: Overview, key findings on adolescent drug use. Ann Arbor, MI: Institute for Social Research, The University of Michigan. [Google Scholar]
  31. Kreager DA, & Haynie DL (2011). Dangerous liaisons? Dating and drinking diffusion in adolescent peer networks. American Sociological Review, 76, 737–763. doi: 10.1177/0003122411416934 [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Kreager DA, Haynie DL, & Hopfer S (2013). Dating and substance use in adolescent peer networks: A replication and extension. Addiction, 108, 638–647. doi: 10.1111/j.1360-0443.2012.04095.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Kreager DA, Molloy LE, Moody J, & Feinberg ME (2016). Friends first? The peer network origins of adolescent dating. Journal of Research on Adolescence, 26, 257–269. doi: 10.1111/jora.12189 [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. 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, 426–457. doi: 10.1080/17437199.2011.587961 [DOI] [PubMed] [Google Scholar]
  35. Lewis MA, & Neighbors C (2006). Social norms approaches using descriptive drinking norms education: A review of the research on personalized normative feedback. Journal of American College Health, 54, 213–218. doi: 10.3200/JACH.54.4.213-218 [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Lin X, & Weinberg BA (2014). Unrequited friendship? How reciprocity mediates adolescent peer effects. Regional Science and Urban Economics, 48, 144–153. doi: 10.1016/j.regsciurbeco.2014.06.001 [DOI] [Google Scholar]
  37. Martino SC, Ellickson PL, & McCaffrey DF (2008). Developmental trajectories of substance use from early to late adolescence: A comparison of rural and urban youth. Journal of Studies on Alcohol and Drugs, 69, 430–440. doi: 10.15288/jsad.2008.69.430 [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Meier A, & Allen G (2009). Romantic relationships from adolescence to young adulthood: Evidence from the National Longitudinal Study of Adolescent Health. The Sociological Quarterly, 50, 308–335. doi: 10.1111/j.1533-8525.2009.01142.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Midanik LT (1999). Drunkenness, feeling the effects and 5+ measures. Addiction, 94, 887–897. doi: 10.1046/j.1360-0443.1999.94688711.x [DOI] [PubMed] [Google Scholar]
  40. National Institute on Alcohol Abuse and Alcoholism. (2017, February). Underage drinking. Retrieved from http://pubs.niaaa.nih.gov/publications/UnderageDrinking/Underage_Fact.pdf
  41. Osgood DW, Ragan DT, Wallace L, Gest SD, Feinberg ME, & Moody J (2013). Peers and the emergence of alcohol use: Influence and selection processes in adolescent friendship networks. Journal of Research on Adolescence, 23, 500–512. doi: 10.1111/jora.12059 [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Osgood DW, Wilson JK, O’Malley PM, Bachman JG, & Johnston LD (1996). Routine activities and individual deviant behavior. American Sociological Review, 61, 635–655. doi: 10.2307/2096397 [DOI] [Google Scholar]
  43. Poonawalla IB, Kendzor DE, Owen MT, & Caughy MO (2014). Family income trajectory during childhood is associated with adolescent cigarette smoking and alcohol use. Addictive Behaviors, 39, 1383–1388. doi: 10.1016/j.addbeh.2014.05.005 [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Ragan DT (2014). Revisiting “what they think”: Adolescent drinking and the importance of peer beliefs. Criminology, 52, 488–513. doi: 10.1111/1745-9125.12044 [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Rhule-Louie DM, & McMahon RJ (2007). Problem behavior and romantic relationships: Assortative mating, behavior contagion, and desistance. Clinical Child and Family Psychology Review, 10, 53–100. doi: 10.1007/s10567-006-0016-y [DOI] [PubMed] [Google Scholar]
  46. Shin C (2011). PROSPER technical report. Unpublished manuscript.
  47. Siennick SE, & Osgood DW (2012). Hanging out with which friends? Friendship‐level predictors of unstructured and unsupervised socializing in adolescence. Journal of Research on Adolescence, 22, 646–661. doi: 10.1111/j.1532-7795.2012.00812.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  48. Smith EP (2007). The role of afterschool settings in positive youth development. The Journal of Adolescent Health, 41, 219–220. doi: 10.1016/j.jadohealth.2007.06.010 [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Snijders TA, & Borgatti SP (1999). Non-parametric standard errors and tests for network statistics. Connections, 22, 161–170. [Google Scholar]
  50. Spoth R, Greenberg M, Bierman K, & Redmond C (2004). PROSPER community–university partnership model for public education systems: Capacity-building for evidence-based, competence-building prevention. Prevention Science, 5, 31–39. 10.1023/B:PREV.0000013979.52796.8b [DOI] [PubMed] [Google Scholar]
  51. Spoth R, Redmond C, Shin C, Greenberg M, Clair S, & Feinberg M (2007). Substance-use outcomes at 18 months past baseline: The PROSPER community–university partnership trial. American Journal of Preventive Medicine, 32, 395–402. 10.1016/j.amepre.2007.01.014 [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Squeglia LM, Jacobus J, & Tapert SF (2009). The influence of substance use on adolescent brain development. Clinical EEG and Neuroscience, 40, 31–38. doi: 10.1177/155005940904000110 [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Steglich C, Snijders TA, & West P (2006). Applying SIENA. Methodology, 2, 48–56. doi: 10.1027/1614-2241.2.1.48 [DOI] [Google Scholar]
  54. Steinberg L, & Monahan KC (2007). Age differences in resistance to peer influence. Developmental Psychology, 43, 1531–1543. 10.1037/0012-1649.43.6.1531 [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Substance Abuse and Mental Health Services Administration. (2014). Results from the 2013 National Survey on Drug Use and Health: Summary of national findings (NSDUH Series H-48, HHS Publication No. (SMA) 14–4863). Rockville, MD: DIANE Publishing. [Google Scholar]
  56. Sun IY, & Longazel JG (2008). College students’ alcohol-related problems: A test of competing theories. Journal of Criminal Justice, 36, 554–562. doi: 10.1016/j.jcrimjus.2008.09.007 [DOI] [Google Scholar]
  57. Sutherland EH (1947). Principles of criminology, 4th ed Philadelphia, PA: Lippincott. [Google Scholar]
  58. Tebes JK, Feinn R, Vanderploeg JJ, Chinman MJ, Shepard J, Brabham T, … & Connell C. (2007). Impact of a positive youth development program in urban after-school settings on the prevention of adolescent substance use. Journal of Adolescent Health, 41, 239–247. doi: 10.1016/j.jadohealth.2007.02.016 [DOI] [PubMed] [Google Scholar]
  59. Valente TW, Gallaher P, & Mouttapa M (2004). Using social networks to understand and prevent substance use: A transdisciplinary perspective. Substance Use & Misuse, 39, 1685–1712. doi: 10.1081/JA-200033210 [DOI] [PubMed] [Google Scholar]
  60. Valente TW, Ritt‐Olson A, Stacy A, Unger JB, Okamoto J, & Sussman S (2007). Peer acceleration: Effects of a social network tailored substance abuse prevention program among high‐risk adolescents. Addiction, 102, 1804–1815. doi: 10.1111/j.1360-0443.2007.01992.x [DOI] [PubMed] [Google Scholar]
  61. Warr M, & Stafford M (1991). The influence of delinquent peers: What they think or what they do? Criminology, 29, 851–866. doi: 10.1111/j.1745-9125.1991.tb01090.x [DOI] [Google Scholar]
  62. Yancey AK, Grant D, Kurosky S, Kravitz-Wirtz N, & Mistry R (2011). Role modeling, risk, and resilience in California adolescents. Journal of Adolescent Health, 48, 36–43. doi: 10.1016/j.jadohealth.2010.05.001 [DOI] [PubMed] [Google Scholar]
  63. Young AM, Morales M, McCabe SE, Boyd CJ, & D’arcy H (2005). Drinking like a guy: Frequent binge drinking among undergraduate women. Substance Use & Misuse, 40, 241–267. [DOI] [PubMed] [Google Scholar]

RESOURCES