Disentangling the Physical, Social, and Situational Contexts of Young Adolescents’ Initiation to Alcohol Use and Intoxication: A Mixed Methods Study

Abstract

Background

The onset of alcohol use during adolescence is associated with concurrent and subsequent related problems. Research on drinking contexts that underly these key first-time experiences and how they differ by initiation type is needed. The current study examined the physical, social, and situational characteristics of three types of initiation: first drink, first heavy episodic drinking (HED), and first intoxication and considered variations between early and later initiating adolescents.

Methods

A mixed-methods approach was used to survey and interview adolescents who reported lifetime drinking. Survey responses from 471 participants were analyzed using multilevel multinomial and logistic regressions accounting for nesting of drinking events (i.e., type of initiation) within respondents. A subsample of 50 participants recruited at baseline took part in in-depth interviews that were thematically coded.

Results

After controlling for demographics, initiation of HED and intoxication, compared to initiation of a whole drink, were more likely to occur when more close friends are present and when those close friends are also drinking. The likelihood of early initiation of a whole drink and intoxication was also positively associated with being in an outdoor setting. Narratives identified distinct and shared patterns of context characteristics across the types of alcohol initiation.

Conclusion

The quantitative and qualitative findings revealed several parallels as well as aspects that differ, enriching our understanding of early drinking contexts. Results highlight the importance of considering contextual characteristics by initiation type for prevention efforts.

1. Introduction

Adolescent initiation of drinking poses substantial risks, affecting concurrent and subsequent developmental, health, and social problems (Hingson et al., 2009, 2006; Kim et al., 2017; Stueve and O’Donnell, 2005; Warner et al., 2007; Warner and White, 2003; Windle and Zucker, 2010). Notably, adolescents who begin to use alcohol before the age of 15 years are more likely to develop symptoms of dependency than their peers who delay initiation (Dawson et al., 2008; Grant and Dawson, 1998, 1997; Warner et al., 2007). Given that such associations remain after accounting for family history of problems with alcohol and personality disorders, a socio-ecological framework of drinking contexts and alcohol-related problems (Freisthler et al., 2014) can help to broaden our understanding by considering the environments where adolescents initiate drinking. Prior research shows that drinking contexts are key predictors of alcohol use and problems (e.g., Bersamin et al., 2016; Lipperman-Kreda et al., 2017a, 2017b; Mair et al., 2015); however, contexts of first use, which represent formative experiences, have been understudied. The current investigation examines the contextual factors of drinking environments where initiation occurs to help identify characteristics that can be targeted by preventive interventions for adolescents.

Several studies have drawn on socio-ecological approaches to illuminate the critical role of contexts of risky drinking behavior. Conceptualized as domains of influence, drinking contexts comprise: location (e.g., own home, outdoor places), the social (e.g., number of people, age composition), and situational (e.g., time of the day, alcohol availability) characteristics of those locations that enable or discourage alcohol use (Freisthler et al., 2014). These domains interact to structure alcohol use contexts that influence drinking behavior through different mechanisms of risk (Freisthler et al., 2014). Social contexts have received relatively more attention showing for example that larger, mixed-gender gatherings are associated with elevated risks of alcohol use and moreover, non-party occasions are related to increased drinking (Lipperman-Kreda et al., 2018). Studies investigating the social contexts of initial drinking experiences further suggest that the presence of peers enables early onset of heavier drinking, whereas presence of parents and friends are more likely to contribute to the early onset of lighter drinking (Jackson et al., 2016). The importance of location for adolescents’ drinking has been established, such that youth are more likely to adopt frequent drinking behavior if they are in an entertainment venue or at the home of a friend for their first drink, compared to adolescents who were in other places (Chen et al., 2008). It has also been suggested that the provision of alcohol within first-time contexts can differentially shape future drinking depending on the source (i.e., parents, friends), highlighting the significance of initiation environments (Kelly et al., 2012). Considering this limited research, there is a need to collectively study the physical, social, and situational domains to identify the unique contexts of initiation that facilitate greater risk.

The current study builds on and advances the extant research in important ways. First, this work is grounded in a socio-ecological model of drinking that recognizes the mechanistic role of contexts in adolescent alcohol use and potential problems (Freisthler et al., 2014). Hence, we focus on initiation contexts inclusive of physical, social, and situational domains. Second, we address the complexity of initiation and investigate three drinking experiences: first full drink of alcohol, first heavy episodic drinking (HED; consuming ≥5 alcoholic drinks in a single occasion), and first intoxication. Third, since specific contexts may be selected by adolescents to facilitate their drinking (Mair et al., 2015), we acknowledge that these aspects may not be fully captured with quantitative data alone. Thus, we use a mixed-method approach with quantitative-based survey data and qualitative-based interview data from a sample of adolescents in California to address the following goals:

1. Identify and describe the physical, social, and situational context characteristics associated with first whole drink of alcohol, first HED, and first intoxication; and

2. Differentiate the physical, social, and situational context characteristics associated with first whole drink of alcohol, first HED, and first intoxication between adolescents of early (≤15 years) and later age (≥16 years) of initiation.

2. Materials and methods

2.1. Overview

The study data are from a longitudinal, statewide study of 1,350 adolescents in California. Our quantitative analyses use a cross-sectional approach to examine event-level data about initiation drinking events from adolescents who reported lifetime drinking at baseline or one of two consecutive 6-month follow-up surveys. Baseline data were collected between January 2018 to June 2019. Data for the first and second follow-ups were collected between May 2019 to February 2020 and from October 2019 to June 2020. Interviews were conducted between December 2018 and November 2019. All data were collected after institutional review board approval was obtained.

2.2. Survey procedures

The sample was recruited from commercially available listings of households with valid mailing addresses associated with a landline or cellphone as well as social media. Invitations to participate in the study were mailed to those recruited through the listings with follow-up telephone calls to determine that an adolescent aged 12–16 years resided within the household. A random selection process was used for households with more than one age-eligible adolescent. Those recruited through social media were routed to an online screener and then contacted by phone to participate in the study after confirming eligibility. Consent and assent were also obtained. Baseline surveys were mainly conducted using computer-assisted telephone interviews (N=1,325); few participants opted for a web-based survey (n=25). Each follow-up was web-based, and participants received $25 per completed survey. Baseline response rates were 49.9% with completion rates of 89% and 93% for each of the two follow-ups from the baseline cohort (AAPOR, 2008).

2.2.1. Survey sample

517 adolescents reported initiating drinking at baseline, follow-up 1, or follow-up 2. The results are based on 822 initiation events from 471 adolescents who provided complete data for all analytic variables. This sample had a mean age of 15.0 years (SD=1.14) at baseline with 52% identifying as female and 47% identifying as non-Latino white. Table 1 displays additional sample demographics.

Table 1.

Demographics of survey sample (N = 471 adolescents)

	%	Mean	SD	Range
Age at baseline		15.02	1.14	12–16
Age at first whole drink		14.24	1.83	3–17
Age at first HED		14.81	1.32	8–17
Age at first intoxication		14.69	1.39	7–17
Annual household income1		7.31	2.14	1–9
Education of the primary household wage earner2		6.83	1.87	1–9
Subjective socioeconomic status3		3.76	1.23	1–7
Gender (female)	52.14
Race/ethnicity
Non-Latino, white	47.44
Non-Latino, American Indian or Alaska Native	0.64
Non-Latino, Native Hawaiian	0.21
Non-Latino, Asian	3.21
Non-Latino, African American or Black	4.70
Latino, white	7.91
Latino, non-white	24.79
More than one race and “other”	11.10

¹Assessed on a 9-point scale from 1 (0-$4,999 yearly) to 9 ($75,000 or more yearly) in response to “In the last year, what was your household income from all sources before taxes?”

²Assessed on a 9-point scale from 1 (less than 8th grade) to 9 (graduate of professional school after college)

³Assessed on a 7-point scale from 1 (well above average) to 7 (well below average) in response to “Compared with other people in America, how rich or poor do you consider yourself?”

2.2.2. Outcome measures

Participants were asked if they had ever had a whole drink of an alcoholic beverage (i.e., lifetime drinkers). Lifetime drinkers were then asked two dichotomous (yes/no) items: (1) if they ever had ≥5 whole drinks of an alcoholic beverage within a couple of hours and (2) if they ever had enough to drink to feel drunk. Given the study focus on characterizing contexts of three types of initiations, we used these measures to construct a multilevel dataset with each type of initiation event nested within participants. Our first outcome measure was a multinomial variable indicating the type of alcohol initiation event (i.e., first drink, HED, and intoxication). Participants who reported any initiation were also asked how old they were at the time. Three dichotomous variables indicating first whole drink/HED/intoxication before age 16 (early versus later initiation) were created from these responses.

2.2.3. Drinking context measures

Using an event-level data approach, respondents were asked to think about the first time they consumed a full drink of alcohol, engaged in HED, and felt intoxicated and answer a series of questions for each initiation event. First, they were asked where they were (coded as home settings, outdoor settings, or other places) and whether it was a party (yes/no). Next, adolescents were asked to estimate the number of people present and as well as the number of close friends and how many of their close friends were drinking alcohol at the event. Additional (yes/no) questions asked whether a responsible adult was present, and when applicable, if that adult was their own parent/guardian and whether their parents/guardian knew that the respondent was initiating alcohol use. Lastly, adolescents rated how easy it was to get alcohol this first time (1=very easy to 4=somewhat difficult). Contextual characteristics identified as important to understanding underage drinking in previous research were selected (Freisthler et al., 2014; Jackson et al., 2016; Lipperman-Kreda et al., 2015; Lipperman-Kreda et al., 2017b; Mair et al., 2015).

2.2.4. Demographic measures

Demographic characteristics, including gender, race/ethnicity, age, and socioeconomic status (SES), were obtained from a brief household survey conducted with a parent/guardian at recruitment and from the baseline survey completed by adolescent participants. SES was reported through perceived SES, household income, and education level of the primary wage-earner (Table 1).

2.3. Interview procedures

Baseline survey respondents reporting any lifetime drinking were eligible to take part in a 60-minute online interview. A convenient sampling approach was used to select equal subgroups of males and females, and throughout the recruitment process, participants were also stratified by age of first whole alcoholic drink as either early (≤15 years old) or later initiators (≥16 years old) (Dawson et al., 2008; Kuntsche et al., 2013). Initial contact was made via telephone to parents/guardians (n=71) following the baseline survey; 94% (n=67) provided permission to contact their child. Among those adolescents, 17 were not interviewed due to unsuccessful scheduling, lack of interest, or unwillingness to be recorded. The remaining 50 participants were interviewed by the research team and compensated with $50 gift cards. Audio-recordings of the interviews were professionally transcribed.

2.3.1. Interview sample

Of the 50 interviewees, 27 were early initiators. Overall mean age was 15.6 years with roughly half of participants identifying as female (n=27) and non-Latino white (n=26). Nearly one-third identified as Hispanic/Latino (n=16) regardless of race. Demographically, this subsample was similar to the survey sample.

2.3.2. Interview guide

Interviews were semi-structured and informed by the research questions and theoretically relevant concepts (e.g., physical, social, situational factors). Each interview began with a short introduction where participants shared details about their personal background. Questions then asked about peer-specific normative beliefs about initiation before shifting focus to their initiation experiences and related context characteristics. The open-ended nature of the interview guide allowed for capturing multidimensional aspects of initiation contexts and incorporating probes with language presented by the participant. All interviews closed with an opportunity for participants to pose questions and feedback.

2.4. Data analysis

2.4.1. Quantitative analyses

First, to assess context characteristics associated with type of initiation event (first drink, HED, and intoxication), we used a multi-level multinomial logistic regression with cluster robust standard errors to account for the clustering of 822 drinking events within 471 participants. Then, we conducted three logistic regression analyses to assess context characteristics associated with early age of initiation (≤15 years old) of any drinking, HED, and intoxication. All analyses controlled for parental education, household income, perceived SES, gender, age, and race/ethnicity; all variables were entered simultaneously. Analyses were conducted using Stata 15 (StataCorp., 2017).

2.4.2. Qualitative analysis

To identify emerging themes, ten randomly selected interviews were coded by three analysts (SI, JPW, EK). Preliminary and a priori themes were discussed, and a draft codebook was developed. Another group of 10 randomly selected interviews were then coded and further refinements were made to complete the codebook. Each interview was then randomly assigned to two analysts; weekly team meetings were held throughout the analyses to discuss disagreements and reach consensus. Subsequent iterations of this process allowed for a dynamic, hybrid approach that allowed for unanticipated and anticipated themes. In total, 60 theory- and data-derived codes relating to the context of initiation were created. All interviews, including those analyzed for the development of the codebook, were coded using the final version of the codebook in Atlas.ti 7.

3. Results

3.1. Quantitative Results

3.1.1. Context characteristics and alcohol initiation behaviors

Descriptive statistics for context characteristics are presented by outcome category (i.e., initiation type) (Table 2). Multinomial logistic regression analyses assessed contexts of initiation of HED and intoxication compared to initiation of first whole drink (Table 3). The initiation of HED and intoxication were more likely to occur at a party and within contexts with fewer close friends present, but a greater number of close friends drinking. Supervision by own parent was negatively associated with initiation of HED and intoxication. No other event-level predictors were significant.

Table 2.

Contexts by type of initiation (N = 822 observations)

Context Characteristics	Proportion or Mean (95% CI)
	First Whole Drink	First HED	First Intoxication
Drinking place
Home settings	0.78 (0.74, 0.82)	0.80 (0.71, 0.86)	0.80 (0.75, 0.85)
Outdoor settings	0.08 (0.06, 0.11)	0.11 (0.06, 0.18)	0.10 (0.07, 0.15)
Other settings	0.12 (0.10, 0.16)	0.08 (0.04, 0.15)	0.08 (0.05, 0.13)
Party	0.40 (0.35, 0.44)	0.56 (0.47, 0.65)	0.48 (0.42, 0.55)
Number of people	14.82 (12.08, 17.56)	17.03 (12.59, 21.47)	16.86 (12.66, 21.05)
Number close friends	3.22 (2.67, 3.78)	4.54 (3.66, 5.42)	3.89 (3.30, 4.47)
Number close friends drinking	2.35 (1.94, 2.76)	4.16 (3.30, 5.01)	3.35 (2.79, 3.90)
Adult supervision	0.42 (0.38, 0.47)	0.31 (0.23, 0.40)	0.27 (0.22, 0.33)
Adult supervision, parent/guardian	0.25 (0.21, 0.29)	0.07 (0.04, 0.14)	0.07 (0.05, 0.12)
Parent/guardian knew you were drinking	0.31 (0.27, 0.35)	0.18 (0.12, 0.27)	0.17 (0.12, 0.22)
Alcohol perceived availability1	3.30 (3.23, 3.38)	3.50 (3.37, 3.62)	3.39 (3.28, 3.49)

¹Assessed on a 4-point scale from 1 (very difficult) to 4 (very easy)

Table 3.

Multilevel multinomial regressions to account for nesting of drinking events within respondents (822 observations from 471 adolescents) to examine context characteristics associated with type of initiation

Context Characteristics	First HED1		First intoxication1
	RRR (95% CI)	p	RRR (95% CI)	p
Drinking place2
Outdoor settings	1.45 (0.77, 2.71)	0.24	1.16 (0.76, 1.77)	0.71
Other settings	0.82 (0.41, 1.65)	0.59	0.83 (0.48, 1.42)	0.50
Party	1.66 (1.05, 2.63)	0.03	1.20 (0.86, 1.67)	0.27
Number of people	1.00 (0.99, 1.00)	0.90	1.00 (0.99, 1.00)	0.39
Number close friends	0.73 (0.60, 0.88)	0.002	0.84 (0.76, 0.93)	0.002
Number close friends drinking	1.49 (1.20, 1.84)	< 0.001	1.26 (1.09, 1.46)	0.002
Adult supervision	0.95 (0.59, 1.53)	0.84	0.81 (0.58, 1.13)	0.22
Adult supervision, parent/guardian	0.26 (0.10, 0.66)	0.005	0.32 (0.17, 0.62)	0.001
Parent/guardian knew you were drinking	1.10 (0.59, 2.04)	0.74	0.96 (0.61, 1.50)	0.86
Alcohol perceived availability3	1.25 (0.95, 1.63)	0.10	1.09 (0.92, 1.30)	0.27
Age	1.20 (1.02, 1.41)	0.27	1.06 (0.95, 1.18)	0.28
Gender (female)	0.84 (0.61, 1.17)	0.31	1.01 (0.84, 1.23)	0.84
Race/ethnicity (non-Latino white)	1.11 (0.79, 1.57)	0.51	0.93 (0.76, 1.14)	0.54
Household income	0.98 (0.89, 1.07)	0.70	0.97 (0.92, 1.03)	0.46
Education of the primary household wage earner	1.01 (0.91, 1.12)	0.74	1.03 (0.98, 1.09)	0.19
Subjective socioeconomic status	1.05 (0.90, 1.23)	0.50	0.96 (0.87, 1.05)	0.35

¹First whole drink is the reference category

²Home setting is the reference category

³Assessed on 4-point scale from 1 (very difficult) to 4 (very easy)

3.1.2. Context characteristics and early initiation

Three logistic regressions assessed context characteristics associated with early age of first drinking, heavy drinking, and intoxication (Table 4). Controlling for demographics, results showed likelihood of early (≤15 years) initiation of a whole drink was positively associated with being at an outdoor setting. Although adult supervision reduced the likelihood of early initiation of a whole drink, supervision by own parent was positively associated with early initiation of a whole drink. Greater perceived availability of alcohol was associated with increased risks of early initiation of a whole drink. Perceived availability of alcohol was associated with increased risks of early HED initiation. Finally, early age of intoxication was positively associated with being in an outdoor setting.

Table 4.

Results of three logistic regressions to assess context characteristics associated with early age (< 16 years) of initiations

Context Characteristics	Early age of first whole drink N = 466		Early age of first HED N = 112		Early age of first intoxication N = 225
	OR (95% CI)	p	OR (95% CI)	p	OR (95% CI)	p
Drinking place2
Outdoor settings	3.60 (1.07, 12.15)	0.03	10.30 (0.82, 128.24)	0.07	7.40 (1.45, 37.62)	0.01
Other settings	0.77 (0.31, 1.86)	0.56	0.12 (0.01, 1.70)	0.11	0.71 (0.16, 3.00)	0.64
Party	1.12 (0.62, 2.02)	0.69	0.39 (0.07, 1.93)	0.25	0.58 (0.24, 1.41)	0.23
Number of people	0.99 (0.98, 1.00)	0.41	1.02 (0.98, 1.07)	0.26	0.99 (0.98, 1.00)	0.51
Number close friends	1.07 (0.95, 1.20)	0.25	1.41 (0.62, 3.22)	0.40	1.08 (0.78, 1.48)	0.63
Number close friends drinking	0.89 (0.76, 1.04)	0.15	0.75 (0.32, 1.73)	0.50	0.95 (0.69, 1.31)	0.77
Adult supervision	0.42 (0.21, 0.85)	0.01	2.13 (0.45, 9.92)	0.33	1.74 (0.70, 4.31)	0.22
Adult supervision, parent/guardian	2.94 (1.17, 7.36)	0.02	0.16 (0.01, 1.93)	0.15	0.79 (0.09, 6.58)	0.83
Parent/guardian knew you were drinking	0.76 (0.34, 1.69)	0.50	1.30 (0.25, 6.75)	0.75	1.50 (0.46, 4.84)	0.49
Alcohol perceived availability3	1.46 (1.04, 2.04)	0.02	3.44 (1.09, 10.87)	0.03	0.94 (0.59, 1.51)	0.82
Age	0.13 (0.07, 0.22)	< 0.001	0.01 (0.00, 0.12)	< 0.000	0.09 (0.04, 0.22)	< 0.000
Gender (female)	0.67 (0.39, 1.12)	0.13	0.71 (0.19, 2.69)	0.62	0.47 (0.22, 1.01)	0.05
Race/ethnicity (non-Latinx white)	1.16 (0.67, 2.00)	0.59	0.41 (0.11, 1.43)	0.16	1.28 (0.58, 2.82)	0.52
Household income	0.84 (0.72, 0.99)	0.04	1.09 (0.72, 1.63)	0.67	0.90 (0.71, 1.13)	0.37
Education of the primary household wage earner	0.98 (0.83, 1.15)	0.81	0.72 (0.48, 1.07)	0.10	0.87 (0.68, 1.11)	0.28
Subjective socioeconomic status	0.80 (0.61, 1.05)	0.12	0.88 (0.49, 1.58)	0.67	0.94 (0.65, 1.36)	0.75

¹First whole drink is the reference category

²Home setting is the reference category

³Assessed on a 4-point scale from 1 (very difficult) to 4 (very easy)

3.2. Qualitative results

3.2.1. Physical context

3.2.1.1. Selection of initiation sites was determined by perceptions of privacy

Participants differentiated between private and public locations for their initiation experiences. Private locations commonly consisted of residential settings, hotel rooms, and other venues that could be reserved for special events like weddings. P41 [male (M), later initiator (L)], for instance, described having his first whole drink while attending a sleepover birthday party in the home of a friend. He explains that being able to spend the night was important to his decision to drink:

“I was kind of relieved… People not being irresponsible outside the house, that they [adults] could monitor us and make sure we were safe.”

Few participants reported drinking in public spaces; however, the element of privacy in those locations was significant. In detailing her first drink, which also included HED initiation, P16 [female (F), early initiator (E)] reflects:

“We couldn’t be at our houses. We didn’t want to be in a very public space, and I had been in that place before because my friends and I used to hang out there when we weren’t drinking. It was secluded so you had to go over a fence and [with] trees and stuff, no one would see who and no one came back there…”

For participants who were constrained by a lack of options within home settings, public areas served as sites of their initiation experiences, including P35 [F, L], who had both her first drink and intoxication experience in a neighborhood park. Centrally located to her close friends, this spot was frequented for their hanging out and deemed intimate and appropriate for their drinking together. P35 also disclosed being well acquainted with patrons of the park, including adults that could be approached to purchase alcohol on behalf of her and her friends.

3.2.2. Social context

3.2.2.1. Initiation was informed by interpersonal closeness

Participants reported that their initiation experiences were strongly shaped by their relationships with those surrounding them. Specifically, the presence of close friends at first drinking events was common across a range of scenarios. For example, P13 [M, E] shared that his first drink was a social experience with well-known people:

“One of them, he’s been my best friend since I was in the second grade. I’ve known him for eight years, and the other one, I had known for just a year but still pretty good friends… And I know one of my best friend’s brothers… I’ve grown up with him. And the parents, I’ve known them just as long as I’ve known my friend. About eight years I’ve known all of them. I was comfortable.”

Later describing his first intoxication and HED, P13 said,

“I was at my friend J’s house the first time I got drunk. There were probably eight boys and eight girls. And all the guys, I’m really close with…”

Several participants noted that drinking together was a special shared experience as participants and their friends were initiating together. Others linked their fears of consuming large amounts of alcohol and feeling drunk, leading to a loss of self-control and embarrassment around non-close peers, as important to initiating with close friends who provided a sense of security.

3.2.3. Situational context

3.2.3.1. Availability of alcohol contributes to unplanned initiation

Participants frequently reported contextual factors central to their drinking experiences that were incidental, such as having access to alcohol rather than planning to secure a location and obtain alcohol. Indeed, some participants did not engage in any planning for their initiation, wherein drinking on that occasion was determined largely by proximity to alcohol.

“The fact that the hard liquor was present was probably prompting all of us to just try it, so it was not planned before, but it was sort of just sitting there.”

[P27, F, E]

3.2.3.2. Initiation with parents varied by types of initiation

The presence or absence of parental and adult supervision at first drinking events was another major characteristic of situational contexts. Even if parents and adults were nearby, there often was a general lack of awareness as illustrated with this narrative,

“We snagged a bottle of Patron from his parents, and we were like, ‘okay, we’re gonna see what getting drunk feels like.’ We drank until we felt something… His parents were home. They have a really large house, though. We were on the other end of the house, blasting music. His parents didn’t do anything.”

[P11, M, E]

Some initiation experiences were, therefore, not limited to a complete absence of guardians, but involved drinking in locked bedrooms or areas of a house where adults were not present. Counternarratives, though minimal, emphasized that parental presence was not a hindrance to drinking. While parents did not encourage consumption, they did not prevent adolescents from drinking alcohol. In their responses, participants reported light drinking in such circumstances and that parents tended to limit the amount of drinking.

“As long as there’s not too much, they allow us to have maybe a sip or one shot of alcohol but that’s pretty much it.”

[P29, F, L]

4. DISCUSSION

Although it is well documented that initiation of alcohol use at a younger age increases the risk of short- and long-term adverse impacts on health, (DeWit et al., 2000; Gruber et al., 1996; Guttmannova et al., 2011; LaBrie et al., 2008; Pilatti et al., 2014) to our knowledge there are no mixed-methods studies investigating first-time drinking contexts. Also unique, our research differentiated three types of initiation: first drink, intoxication, and heavy episodic drinking. Using survey and interview data, our analyses provide a nuanced examination of the physical, social, and situational characteristics of drinking environments and how adolescents describe these aspects as collectively compelling, such that specific contexts interact to shape initiation experiences. The results are congruent with prior research demonstrating that specific contextual factors work together to influence when and how individuals use substances like alcohol (Egan et al., 2019; Lipperman-Kreda et al., 2017b, 2015; Mair et al., 2015; Trucco et al., 2014). Our findings further highlight the importance of making distinctions between initiation types for future research and prevention.

Our mixed-method findings revealed several patterns that are shared across early drinking contexts as well as aspects that differ. It is apparent that adolescents are cognizant of their drinking contexts, make mindful choices regarding their drinking behaviors based upon favorable and unfavorable attributes of those contexts, and attempt to assess risk throughout the drinking event. Generally, adolescents perceived the physical context as a location that does or does not accommodate for privacy from unwanted people and consequences. This characteristic was distinctly important in HED and intoxication initiation, which called for greater discreetness in maintaining a favorable context for their experimentation of drinking limits. Similarly, compared to first drink experiences, participants tended to drink more heavily when they were able to spend an extended amount of time in a single location and were with close friends also engaged in drinking. Narratives often mentioned the collaborative process enabling initiation, wherein adolescents planned mutual first-time experiences of getting drunk with trusted peers. Such elements of social bonding represented a form of risk mitigation that were not as notable to alcohol initiation unless the first drink co-occurred with first HED and intoxication.

Relatedly, presence of parents/guardians were predominantly linked to situations where adolescents had their first full drink of alcohol, a pattern largely associated with early-initiating participants. While some participants were monitored while first engaging in light drinking, when the drink was not offered by an adult, adolescents concealed it because their parents were in the vicinity. Parental reactions of disappointment were commonly mentioned in these events. This pattern may suggest that parental supervision imparts a sense of safety and security that does not require pre-situational coordinating, such as would be the case with friends and other peers and provides a barrier to HED and intoxication. In contrast, later initiators and participants exploring HED and intoxication for the first time, seem to plan more with close friends and negotiate risk and protection more carefully to socially drink in ways that are likely to be disapproved by their parents (Jackson et al., 2016). The ease of obtaining alcohol is consistent with the literature that accessibility makes drinking more possible (Bersamin et al., 2016; Friese et al., 2013; Komro et al., 2007). Among our interview sample, several adolescents acknowledged that availability of alcohol in the home, whether their own or that of their friends, was an influence on initiation whether or not the participant was interested in drinking.

Furthermore, the examination of early and later initiators demonstrated a key difference: Early initiators were more likely to have their first drink in an outdoor setting in the quantitative analysis. This difference did not emerge as strongly in the qualitative analyses. Still, accounts of initiation were more often within outdoor locations among early initiators. This difference, however, is a bit more nuanced, wherein some early initiators had their first drink in outdoor settings with their parents or on occasions where adult family members and friends were, at least, aware of their drinking or had supplied the alcohol. Outdoor locations were often public spaces that nonetheless could provide some degree of privacy. Example contexts from interviews included a variety of semi-public environments like secluded neighborhood parks, a family trip to a lake, and campsites.

These results align with previous research that shows adolescents’ drinking contexts matter for alcohol use and have implications for problems (Friese et al., 2013; Lipperman-Kreda et al., 2018, 2015); moreover, the current study builds upon this body of work by showing that physical, social, and situational contexts are instrumental in varying ways for adolescent initiators and between types of initiation. The interdependence of context characteristics and efforts taken by adolescents to reduce risk also contribute to the literature, suggesting that prevention programs may be more impactful if they can attend to the choices that adolescents make in specific contexts. The prevalence of planning found in the qualitative data suggests that adolescents are aware of optimal context characteristics for initiation of riskier drinking behaviors, often weighing positive and negative characteristics of contexts when initiating drinking, especially HED and intoxication. Some adolescents also contrasted this behavior by explaining the unfavorable aspects of large groups, including past experiences in such settings and refusing pressures to drink due to concerns that their non-close peers would not look after them if they felt bad or sick from the alcohol.

4.1. Study limitations, strengths, and conclusion

Study limitations should be noted. First, although there is a possibility that we reduced self-report and social desirability biases by only including adolescents in the study whose parents felt comfortable with their child taking part and ensuring adolescents’ privacy during the survey and interviews, these issues must be acknowledged. In addition, our multi-coder method of analyzing interviews helped to ensure that saturation was reached, and that subjective bias was, at most, minimized in our interpretations. The HED measure assessed as ≥5 drinks and did not account for gender-specific thresholds. Demographically, our overall sample is limited to California and tended to come from households with higher socioeconomic statuses. Finally, both survey and interview data are subject to potential recall biases from retrospective reporting and limited by the cross-sectional nature of the data.

Notwithstanding any limitations, our mixed methods study provides rich data allowing us to examine how contexts are related to different types of alcohol initiation among adolescents. The one-on-one interviews allowed adolescents to articulate their responses in their own words, thus augmenting our understanding of drinking contexts. As the quantitative component of our study is ongoing, longitudinal analyses will help to identify how drinking contexts are associated with early initiation of drinking and the progression of problems over time. This focus on drinking contexts in shaping developmental trajectories for alcohol-related problems has the potential to guide tailored interventions for adolescents. The results of our mixed-method investigation raise the question of whether traditional interventions are equipped to respond to early drinking contexts that are most closely related to initiation of drinking, risky drinking, and later problems, given that they are not contextually informed. A shift in focus towards moderating key contextual risks is needed to change adolescent choices related to context. Adaptive interventions using emerging technologies that account for the dynamics of drinking contexts that confer differential risks may be a promising approach.