Social network determinants of alcohol and tobacco use: A qualitative study among out of school youth in South Africa.

An important determinant of alcohol and tobacco use is the adolescent's social network, which has not been explored among out of school youth (OSY). OSY are adolescents not currently enrolled in school and have not completed their schooling. This study aims to qualitatively understand how OSY's social networks support or constrain alcohol and tobacco use. Respondent-driven sampling was used to select 41 OSY (aged 13-20 years) for individual in-depth interviews in a South African urban area. The data were analysed using content analysis. Smoking and drinking friends, family close in age to OSY that drank and smoked, and lack of parental support were associated with alcohol and tobacco use among OSY. Household norms, romantic partners and non-smoking or non-drinking friends were suggested to mitigate alcohol and tobacco use. Understanding how the social network of OSY plays a role in alcohol and tobacco use is useful for gaining an insight into the profile of OSY at risk for alcohol and tobacco use. Registration of OSY youth and community-based peer led programmes that include influential OSY family and friends could be beneficial.

Introduction

Tobacco and heavy use of alcohol result in millions of deaths annually, with the majority of tobacco and alcohol related deaths occurring in low- and middle-income countries (LMICs) [1]. Alcohol and tobacco use, like in many other LMIC’s is prevalent among adolescents [2, 3]. In South Africa, categorised as an LMIC, national studies found that past month alcohol (54.7%) and tobacco use (50.4%) were significantly higher among those who dropped out of school [4, 5] compared to school-going learners who reported 35% and 21% alcohol and tobacco use, respectively [3]. These risk behaviours are associated with poor educational outcomes, [6] diseases, morbidity, and mortality [7] which is a major public health concern.

Previous South African studies have mainly focused on school-going learners and their alcohol and tobacco use [8, 9]; however, those who drop out of school in South Africa have received less attention. According to the latest UNESCO report globally, 258 million children and adolescents (ages 7–19 years) had either never started or dropped out of school, and more than half (58%) of these youth were living in sub-Saharan Africa [10]. In South Africa, only 52% of the age-appropriate population remained in school until the last grade of high school in 2016 [11, 12]. South African studies show that reasons for leaving school include poverty [13-15], high use of substances [5, 13, 14, 16, 17], bullying [18], boredom [17, 19], family needs (helping support the families, being pregnant, traditional family role expectations) [13, 14, 20], illness [14, 21], disability [14], community violence [14], and school related factors (academic performance, disliking school, not getting along with teachers, being too old for school and disciplinary consequences) [13, 14, 21, 22]. Negative social, health and economic consequences such as unemployment, substance use, delinquency, and poor mental and physical health are usually associated with early school leaving [5, 13, 16]. Given the high prevalence of alcohol and tobacco use among OSY in South Africa, it would be useful to understand the determinants of alcohol and tobacco use in this population.

To our knowledge, no previous studies have examined social network determinants among OSY regarding tobacco and alcohol use. One South African study found that leisure motivation and leisure boredom may be associated with substance use among school dropouts [17], which may extend to the characteristics of the social network of OSY. Previous studies including school-going adolescents showed that drinking and smoking behaviour tend to be modelled after friends’ drinking [23-26] and smoking behaviour [27-31]. Partners [32-35] and immediate family (siblings and parents) [36-40] were also significantly associated with adolescent tobacco and alcohol use. However, these studies only considered adolescents attending school who spent a substantial amount of time with school-going peers and teachers. OSY do not have the protective factor of schools, such as the supervision and positive mentoring of teachers and peers, and are more vulnerable to the experimentation and uptake of alcohol and tobacco use [16, 41–44]. Our study will be the first to explore the composition of OSY’s social network, and how these social networks support or constrain OSY’s alcohol and tobacco use.

Numerous theoretical frameworks have been used to explain the processes by which interpersonal relationships may influence an individual’s health-risk behaviour. Adolescence is a transitional period during which peers gain more importance as a means of developing a sense of belonging, self-concept, and support [45]. In the context of alcohol and tobacco use among OSY, the social learning theory [46], the social identity theory [47] and the social network theory [48] provide frameworks for understanding the individual’s interpersonal relationships and their cognitions concerning the larger social system [49]. Social networks refer to the connections and nature of interactions between individuals in a social system, which may facilitate the uptake and spread of resources and behaviours [26, 40, 50]. The most distinguishing feature of the social network theory is its two-fold focus on both the individual actors and the social relationships connecting them [50], which can facilitate or inhibit behaviour. The social learning theory considers the acquisition and continuation of behaviour [51]. Adolescents are likely to imitate those with whom they have the greatest amount of contact and continue behaviour based on the rewards and punishments [49]. The social identity theory proposes that a portion of the self-concept of individuals is dependent on the normative values and behaviours of groups that they belong to [47]. Similarity, smoking and drinking behaviour may be explained by the social identity theory whereby individuals tend to act following group norms, adopting them as their own [49]. These theories will be considered in this study to understand the role that interpersonal relationships play in alcohol and tobacco use among OSY.

Understanding the adolescent’s social network may be useful to identify the profile of OSY that are most vulnerable to alcohol and tobacco use. Although the WHO defines adolescents as individuals between ages 10–19 years [52], this study focuses on adolescents between 13–20 years as grade repetition in high school is high [53]. Also, a slightly older sample of adolescents was targeted because dropout tends to increase from the age of 15 (grade 9) [5]. National studies show that alcohol and tobacco are the most prevalent among adolescents compared to other addictive behaviours such as illegal and other drug use [2, 3]. Moreover, alcohol and tobacco use is usually initiated between 12–14 years and is highly prevalent in the slightly older age group [3]. To our knowledge, no qualitative study has been conducted, specifically focusing on 13-20-year-old dropouts in South Africa. Qualitative studies may assist in acquiring a deeper understanding of the composition of OSY social networks and interactions as determinants of alcohol and tobacco use among OSY. Therefore, this study aims to qualitatively understand the composition of OSY social networks and explore how these social network relationships and interactions facilitate or constrain alcohol and tobacco use.

Materials and methods

Study setting

This study was conducted in the Western Cape, the fourth largest province in South Africa within an urban district characterised by high rates of school dropout [54]. Schooling is compulsory for all South African children from the age of 6 years (grade 1) to the age of 15 years (grade 9). Primary education consists of two phases: the foundation phase, and the intermediate phase. All school governing boards of public schools must supplement government funding by charging school fees and doing other reasonable forms of fund-raising. Caregivers who cannot afford to pay school fees may apply to the school governing board for conditional, partial, or full exemption from paying school fees. Eligibility for full and partial school fee exemptions is calculated based on parental income in relation to the fees [55]. The right not to charge school fees is limited to the schools that have been declared ‘no fee schools,’ which is based on the economic level of the community around the school. From grade 9, however, children do not benefit from the no-fee policy. Although we did not ask respondents the school they previously attended, they were recruited from urban areas that comprised of both no paying and fee-paying schools.

Respondents and sampling

Adolescents between the ages of 13–20 years who were not currently enrolled in secondary or high school for that academic year and have not completed their secondary or high schooling were eligible to participate in the study. Those who were enrolled in college or vocational training were excluded. Eligibility was further confirmed through contacts that the potential participant provided. Because there was no available register of OSY, respondents were recruited using respondent-driven sampling (RDS) [56]. Using RDS, the initial sample or “seeds” of OSY were purposefully obtained by data collectors through a community youth group and approaching young people who appeared to meet the pre-determined criteria in selected communities. Initially, eight seeds (4 smokers and 4 non-smokers) were obtained, and they were required to identify up to two other OSY. These respondents recruited by the seeds formed the “first wave” of sampling and were themselves asked to identify and refer a further two more school dropouts (Fig 1). Up to two waves of recruitment were conducted. The seeds consisted of smokers and non-smokers because the initial focus of the paper was on tobacco use among OSY. However, the paper evolved to focus on both alcohol and tobacco use. Despite targeting smokers and non-smokers as the initial seeds, the researchers still acquired a sample that contains almost an even number of alcohol and tobacco users while minimising potential bias being introduced, due to the nature of the respondent driven sampling strategy.

Fig 1

Respondent-driven sampling for out of school youth–a schematic representation of four seeds.

W 1: wave 1, W 2: wave 2.

Data collection

The interviews were conducted in-person as well as using text messaging on the cellular phone application WhatsApp. Studies have shown that adolescents prefer sharing information and communicating via instant messaging (IM) [57, 58]. Online interviewing allows access to the voices and experiences of “hidden” populations [59-62], that are difficult for researchers to access due to the lack of a sampling frame and geographical location [63]. Compared to the traditional face-to-face interviewing method, qualitative online interviewing saves cost and time, and is convenient for both the researcher and the participant [57, 64]. In the case of text based online interviewing, the use of emoticons and internet slang abbreviations such as “LOL” are commonly used to convey emotion, and has been used previously by interviewers to build rapport [64, 65]. One study systematically compared face-to-face interviewing techniques to instant messaging interviewing, mediated by a computer [57]. Despite taking longer and producing fewer words in the online condition, data quality was unaffected by the mode of data collection (online versus face-to-face) with no differences in the number, depth and type of themes discussed [57]. This suggests that online data collection using cell phone mediated IM, namely WhatsApp, may be a novel, appropriate and feasible method for obtaining sensitive information from school dropouts. Testing of this latter assumption was not part of the present study.

The data collectors recruited and conducted the interviews in the respondents’ language of preference (English, Afrikaans or IsiXhosa). In-person interviews were conducted at private locations agreed to by the respondent and interviewer, lasting approximately 30 minutes. Interviews conducted using WhatsApp took approximately 3–5 days on average with one hour of chat time a day. Respondents who completed the interview received monetary compensation (50 ZAR) for their time. Also, monetary incentives (20 ZAR) were provided for every respondent who successfully recruited more respondents who met the eligibility criteria and participated in the study. Of the respondents that were approached, one respondent was unavailable to do the interview and another did not meet the criteria of being between the ages of 13–20 years. Although the target was 56 respondents, data saturation was achieved after 41 respondents, as by that number no new views were emanating from the interviews since the last three respondents. Of those, 12 respondents conducted the interview over a cellular phone application, WhatsApp, and 29 respondents conducted the interview face-to-face. The transcripts were translated by the data collectors to English from Afrikaans or isiXhosa and then back translated to check for consistency and correct translation.

Data collection tools

Open-ended questions with probes were used to guide the individual in-depth interviews. All authors of the study designed the questions through a reflective, iterative, and dialogic process. Respondents were asked to describe their relationships with their family, friends, and other important people in their lives to gain a sense of their social network. They were further asked to discuss their own alcohol and tobacco use, and alcohol and tobacco use of their social network, as well as to elaborate on how important social network members played a role in their alcohol and tobacco use. The semi-structured discussion guide was designed in English and translated into isiXhosa and Afrikaans. The English version is available as a S1 File.

Analysis

Atlas.ti version 8 was used to code and analyse the data using a content analysis approach [66]. The six-phase approach by Braun and Clarke (2018) was followed to code and analyse the interview data. First, the first author read the transcripts several times and noted initial ideas. Secondly, two independent researchers (RD and AM) blindly coded four transcripts, and preliminary codes were developed and defined based on the objectives of the study. The two researchers compared and discussed the coding until consensus was reached. A final codebook was developed, and the remaining transcripts were coded. The final codebook was reviewed by all authors. Thirdly, the coded transcripts were analysed by running query reports and primary document tables of emerging themes. The fourth phase included review and refining of the themes and sub-themes by all authors. In phase five, the themes and sub-themes were named, and clear working definitions were provided for each theme. Lastly, extracts were selected, and the results of the analyses were organised and presented in this manuscript.

Ethics and consent

Ethics approval was obtained from the Human Sciences Research Council (Protocol number: REC 2/23/08/17). Youth who are out of school, maybe out of school without their caregiver’s knowledge or have unstable family support structures. The nature of the research included minors who might have been unwilling to participate if they had to divulge the nature of the research to their parents or caregivers to obtain parental permission. Given the low risk of the study on respondents, obtaining independent consent from the youth respondents themselves was feasible and increased participation. Therefore, in line with the South African National Department of Health Ethics guidelines (2015) section 3.2.2.4., permission for independent consent for minors was obtained [67]. Once eligibility for participation was established, full written informed consent was obtained from each respondent. All names in the transcripts and extracts presented below were removed or replaced with pseudonyms.

Results

Characteristics of the sample

A total of 41 OSY participated in the study. The study had slightly more females (n = 22) than males (n = 19). The average age of respondents was 18 years. At the time of the interview, 31 respondents reported using tobacco, 23 respondents reported drinking alcohol, 10 did not use tobacco, and 18 reported not drinking (Table 1). Using either tobacco or alcohol was similar across gender. However, using both alcohol and tobacco was more common among males than females. Those who drank reported drinking during special occasions, the weekend and payday. Those who smoked reported smoking daily, and the frequency of smoking increased on weekends.

Table 1

Behavioural characteristics of the OSY sample.

			Male		Female
Characteristics	%	n	%	n	%	n
Total	100	41	46	19	54	22
Tobacco users	65	31	55	17	45	14
Non-tobacco users	34	10	20	2	80	8
Alcohol users	56	23	48	11	52	12
Non-alcohol users	43	18	44	8	56	10
Alcohol and tobacco users	39	16	63	10	38	6
Non- alcohol and non-tobacco users	12	5	20	1	80	4

The social network of OSY

Respondents reported that their social network usually consisted of their friends (in-school and out of school friends and romantic partners) and family (parents, siblings, and extended family). As seen in the extracts of conversations between the interviewer (I) and respondent (R), each of these network members seemed to play a role in a respondent’s inclination and frequency to smoke and/or drink alcohol among OSY users and non-users of alcohol and tobacco. The following paragraphs demonstrate the role that these different individuals play on OSY alcohol and tobacco use.

Friends

OSY friends played a major role in their alcohol and tobacco use. OSY claimed to have met their friends from when they were attending school or from residing in the same neighbourhood. OSY claimed to have friends who were both in and out of school.

Initiation of alcohol and tobacco

Most respondents who were smoking and drinking at the time of the interview initiated and continued to use alcohol and tobacco in the company of friends who smoked and drank. To strengthen bonds between friends, respondents would increase their alcohol and tobacco frequency by adjusting their behaviour to their friends. Moreover, friends would place direct pressure on respondents to smoke and drink by offering alcohol or cigarettes, show respondents how to smoke, and coerce respondents to smoke or drink alcohol.

“I: What made you start drinking and smoking tobacco?

R: I wanted to know the feeling they are getting when they are drunk and smoking tobacco. Sometimes you want to please your friends and be in the same vibe. When your friends are smoking and drinking then you cannot say no, that’s what happened to me.” (Female respondent ID 31, 20 years old)

Often respondent’s initiation with alcohol and tobacco was driven by curiosity and experimentation. Respondents began experimenting with alcohol and tobacco while they were in school with their school friends and continued to drink and/or use tobacco once they left school. Once leaving school, OSY claimed to drink and/or smoke more frequently compared to when they were in school due to having more free time, being unsupervised and being addicted. Despite national policy prohibiting the sale of cigarettes to minors under 18 years, OSY were able to purchase cigarettes from the shop on their transport routes.

“The first time that I smoked was in grade 10 in Johannesburg, it was me and my friend. I had to take her to the taxi rank every afternoon. We just tried to feel what it felt like and every afternoon we bought a cigarette and just breathed out the smoke. We did not inhale at that time and then I came back to Cape Town and became use to smoking and started smoking that I am still smoking today.” (Female respondent ID 13, 19 years old)

“I: How old were you when you first started to use tobacco and drink alcohol?

R: When I was 13 years and I was doing Grade 8

I: With whom?

R: With my primary school friends, we were together at high school even.” (Male respondent ID 39, 20 years)

OSY who were non-users claimed to have experimented with alcohol or tobacco in the presence of their friends. The experience of the unpleasant physiological and emotional feelings associated with using tobacco and alcohol played a role in their decision not to continue. Respondents would associate alcohol and tobacco use to other negative risky behaviours such as violence, sex, and bad company. Although OSY would initially adjust their behaviour to that of their friends, eventually OSY would develop strong opinions on the negative consequences of using tobacco and alcohol from their own experiences.

R: Yes, I tried smoking, consuming alcohol, smoking marijuana, and hookah pipe.

I: Okay, so you have tried everything? Can you tell me about the first time that you smoked and consumed alcohol or tried to use it?

R: I wasn’t the same person that I am now. That was when I tried to do everything friends do. Sitting at a shebeen, walking around in the night and became friends with older men.

I: And you feel that it was not for you?

R: No, it was definitely not for me.” (Female respondent ID12, 19 years old)

R: “…I don’t consume alcohol. So, I don’t know what is the purpose of consuming alcohol because it makes people violent and most people don’t know how to conduct themselves when they have consumed alcohol because they abuse it. So, in my opinion alcohol has a bad influence on people and that is why I don’t consume alcohol.” (Female respondent ID 13, 19 years old)

Financial resources

Friendship groups provided a support structure and social opportunities which would facilitate alcohol and/or tobacco use among OSY. Cited commonly among the males of this sample, the amount of alcohol and tobacco consumed was dependent on the number of friends and the amount of finances available. Each member of the friend group would contribute financially towards purchasing cigarettes and alcohol, and this would be shared amongst them. Purchasing alcohol and cigarettes in bulk made it more affordable. Thus, those who did not have the financial resources had their alcohol or cigarettes purchases partially or fully subsidised by their friends. Given national regulations prohibiting the sale of alcohol and tobacco to minors, older OSY may have also purchased the alcohol and tobacco on behalf of the group.

I: “How many drinks do you have in one week?

R: It depends what we have on the table and how many people are drinking at that moment because we are four in my group. We usually contribute with a R150 each person.

I: In that amount of money what do you buy?

R: It’s a bottle of brandy 750ml, wine and few of 6 dozen of cans of beer/ciders.” (Male respondent ID 37, 19 years old).

R: We are coming from different backgrounds and our homes are not the same. For example, we plan to contribute with R150, and it happens that you do not have the whole amount, maybe you have R90 instead of R150. It doesn’t mean I cannot go with them. That’s the amount you have, at least you came with something” (Male respondent ID 39, 20 years old)

Leisure boredom

Given that respondents were not attending school and mostly stayed at home, male participants (n = 6) reported that they were bored in their free time. The participants spent much of their free time during the week at their friends’ homes, who were also possibly OSY. Boredom with friends may have inevitably led to tobacco and/or alcohol use as well as contributed to their addiction. The quotations below suggest that OSY youth and their OSY friends would smoke and/drink due to boredom, they had more unstructured time and unsupervised opportunities to use tobacco and or alcohol.

“I: And what do you do every day to keep yourself busy?

R: Every day I am with my friends.

I: And there, what do you do?

R: We smoke the hookah pipe, maybe walk with the dogs but we are just at home." (Male respondent ID 26, 19 years old)

“…all of us smoke together. We walk together every day then we smoke marijuana, cigarettes and all of that.” (Male respondent ID 8, age unknown)

Reluctance to alcohol and tobacco

In three instances, there was a reduction in smoking and drinking among female respondents using alcohol and/or tobacco. This occurred when friends would collectively decide to reduce their alcohol and tobacco use, or when respondents had non-drinking or non-smoking friends, who were usually in school.

“R…I think that’s also one of the reasons I like to be friends with my old school friends because they are not drinking. I buy [a soft drink] and that's it. When I go with other friends, I buy a carry pack [6 bottles of alcohol], but when I go with other ones I don’t drink but I feel happy…” (Female respondent 13, 17 years old)

Some of the respondents (n = 8) who reported not using alcohol and/or tobacco still had smoking and/or drinking friends. In the company of their friends who would smoke and/or drink, they would not feel coerced by their friends to change their behaviour, their friends were more accepting of their non-using behaviour and claimed that their friends would not drink or smoke excessively in front of them. Participants would engage in other leisure activities with friends such as going to the mall, attending church, playing sports, or making music. These respondents tend to not change their behaviour to that of their smoking or drinking friends.

“I: maybe smoke more or less or consumes more alcohol or less?

R: Neither one of them.

I: Is it? After you left school?

R: After I left school, neither one of the two. I only chilled with my friends. That is all and even if they smoked, I would not smoke. I will only sit by them and the music is playing that is all.” (Male respondent ID 38, 17 years old)

Three female respondents reported that their romantic partners contributed towards their cigarette and alcohol purchases. Of those responses who reported using alcohol and/or tobacco, most of their partners discouraged this behaviour by associating smoking and drinking with being unsuitable and unattractive. Some respondents tended to smoke and drink less in the presence of their partners compared to when smoking and drinking with friends.

“I: Where do you get money for drinking alcohol?

R: I get pocket money from my boyfriend and my father. I don't stay with my boyfriend, but he stays just around the corner. He usually spends his weekends with his friend, and I enjoy myself with my friends too.

I: He doesn’t say anything when you are drunk?

R: I don’t stay in the same place as him when I am going to drink with my friends because he doesn't want me to drink. He told me when I am drunk, I don’t want to listen, I am too noisy and using vulgar language.” (Female respondent ID 31, 20 years old)

I: “Ok. What information have you been told about drinking by him?

R: I will end up being something else and being ugly” (Female respondent ID 30, 20 years old)

Family

The majority of the respondents resided in a household consisting of biological parents, siblings and at least one non-parental member such as an aunt, uncle or grandparents. Family tended to play a facilitating and inhibiting role in OSY alcohol and tobacco use.

Family members of similar ageIn some cases, siblings played a role in the respondents’ alcohol and tobacco use. Siblings of a similar age to the respondent were more likely to drink together (n = 6). Some respondents would also consume alcohol with their siblings’ friends. Siblings who drank and smoked also contributed financially towards the respondents’ alcohol and tobacco use.

“I: Are there are other family members who encourage you to stop drinking?

R: No, but sometimes I do drink with my older brother when he drinks with his friends

I: Can you tell me why you think your brother doesn't mind when you drink with him?

R: There is not much difference in age between us." (Male respondent ID 39, 20 years old)

Extended family also seemed to play a role in facilitating respondents’ alcohol and tobacco use. Extended family members of a similar age to the respondents who smoked or drank alcohol, were less likely to discourage cigarette and alcohol use. Some respondents reported that they initiated and used cigarettes with their cousins.

“I: Okay. What information were you given by friends and family regarding smoking and the consumption of alcohol? Is there anyone that gave you information?

R: Yes, my cousin. It is almost like he taught me to smoke. He told me I must just do it once and I coughed. So I told him that I want to do it continuously because he also smokes and that was how he taught me to smoke.

I: And he didn’t tell you not to smoke because it is bad for you or nothing like that?

R: No.” (Female respondent ID 24, 17 years old)

Elder family members

The majority of older family members discouraged alcohol and tobacco use by highlighting its adverse health effects and social consequences, even though they smoked or consumed alcohol. Most respondents continued to use tobacco or alcohol, even though their family did not support it. Often respondents hid their alcohol and tobacco use from their family or use little to no tobacco or alcohol in the presence of their older family members out of respect or fear.

“…My grandmother always told me that my grandfather use to smoke and it was not easy for him to stop. The day when he stopped smoking was the day he died. So, you better stop now before it is too late. My grandmother also said that I think that it is easy to stop but it is not. She also told me about the dangers of smoking that my lungs will collapse. But then I just tell myself, arg to hell old women. I am at that stage that when I want to smoke, I want to smoke.” (Female respondent ID 13, 19 years old)

In one case, a non-user of alcohol and tobacco reported residing with a sibling who was older and took on the role of the caregiver. In this case, the sibling played a role in the respondent's decision not to consume alcohol after being reprimanded for initiating alcohol use.

R: “My sister is happy that I am no longer drinking. I remember on the day after my birthday she was mad with me. She didn't hit me because she respects my special day, but she warned me. All I can say she was not happy with my decision” (Female respondent ID 28, 17 years old).

Lack of parental support

Respondents highlighted that their relationship with parents affected their alcohol and tobacco use. Some respondents attributed their alcohol and tobacco use to not getting along with parents, feelings of neglect, grief and coping with household issues. Those neglected by their parents attributed their alcohol and tobacco use to the stress of contributing towards the household income.

“R…Cause the stress levels went higher from like my house problems and my mother is also sick, she like has a heart problem, she has a hole in the heart. And another thing, uhm, we like, my whole family, my mother and my sisters, my two sisters and my dad, we’re not into talking a lot. We like don’t talk a lot. We talk, but we like, how can I say that, family time is not there, how can I say that, you see family time is not there, to put it in that way. Sometimes, with me it’s like, my family is not really away from me but sometimes I feel so, then I go smoke…” (Male respondent ID 25, 18 years old)

"I: What made you start drinking?

“R: It’s a frustration of losing parents. My mother passed away and I don’t know my father” (Female respondent ID 27, 17 years old)

“R: I find it difficult to stop [drinking] because I'm staying alone and I am always thinking about food and clothes to wear, nothing else.” (Female respondent ID 34, age unknown)

Household norms

Respondents who are non-users of tobacco and alcohol use reported lower instances of parental smoking and drinking. Although these OSY had some family members who smoke and or drink, these behaviours were seen to be controlled, away from the home, out of sight or only done on special occasions.

“R: Only my father smokes but he never smokes in the house. He always smokes outside. And they don’t drink. It will only happen occasionally maybe just wine or so, but they won’t drink in front of us.” (Male respondent ID 38, 17 years old).

Moreover, it was found that among non-users of alcohol and tobacco (n = 3), families with religious and traditional affiliations were found not to engage in smoking or drinking and tended to condone these behaviours, which also extended to elders in the extended family

I: “What information have you been told about smoking and drinking by your family or friends?

R: I must not drink again because we are not the family who is drinking. We are the Christians and I was not raised to be the drinker. Even my uncle is not a person who’s drinking but only on special occasions, like on ceremonies. He is not supporting negatives things and he is very traditional” (Female respondent ID 28, 17 years old)

Discussion

This study aimed to understand how social network determinants play a role in alcohol and tobacco use behaviours in a group of OSY in South Africa. Overall, the social environment, which included friends and family, played a facilitating or inhibiting role in OSY’s alcohol and tobacco use. The findings of this study have provided valuable insight into the profile of OSY at risk for alcohol and tobacco use.

The findings from this study suggest the value of examining the friendship group of OSY as a determinant of OSY alcohol and tobacco use. Similar to studies conducted among in-school learners, OSY would imitate their friend’s behaviour concerning alcohol and tobacco use, and continue to use it with their friends [31, 49, 50, 68]. In addition, OSY who were current tobacco and alcohol users tended to spend more time with other OSY friends due to boredom, offering more opportunities and unsupervised time to engage in risk behaviours, rather than spending that time in school. A previous study also found that OSY tend to have more OSY friends [69] and that the smoking behaviour of OSY is associated with that of their OSY friends [70]. In a few cases, being in the company of non-smoking and non-drinking friends or romantic partners would reduce OSY’s alcohol and tobacco use. These findings suggest that OSY tobacco and alcohol use is a learned behaviour with reinforcement in the peer context [71].

In this study, OSY and their friends often drank and smoked in groups, conveying a sense of collective identity and group membership [47]. Furthermore, each group member contributed to the purchase of cigarettes and alcohol. Those who could not contribute financially were assisted by their friends or partners and were thus able to continue using cigarettes and alcohol. The national policy states that tobacco and alcohol may not be purchased by minors under 18 years [72, 73]. Older OSY may have also purchased the alcohol and tobacco for underage OSY. The availability of these resources may explain how female dropouts experiencing financial difficulties had access to cigarettes, as seen in a previous study by Desai and colleagues [5]. Enforcement of national policy prohibiting the sale of alcohol and tobacco to minors or raising the minimum age of tobacco and alcohol purchases may be beneficial. Those OSY who were not users of alcohol or tobacco reported having friends who engaged in these risk behaviours but would not play a role in their decision not to smoke and/or drink, possibly due to individual characteristics or the characteristics of their friends.

The family social network comprising of parents, siblings, and extended family, was also a determinant of alcohol and tobacco use among OSY in this study. OSY mostly resided with parents and at least one non-parental adult, which included grandparents, aunts, uncles, cousins, and other relatives. Interestingly, respondents described their home as consisting of backyard dwellings where the extended family or the children resided. Backyard dwellings are informal shacks, typically erected in the yards of other properties, and is unique to South Africa [74]. The extended family plays a crucial social safety net in sub-Saharan Africa [75]. In cases where a child loses their parents or when parents do not have the resources to support the child, it is common for the child to reside with their relatives [75, 76]. Our study also found that older family members such as parents, aunts, uncles and grandparents and siblings would discourage alcohol and tobacco use, even if they used alcohol or tobacco. However, those OSY who were smoking, or drinking were not affected by these discouragements. This finding may be attributed to members in the household smoking and drinking, the lack of in-home smoking rules, lack of religious norms, poor quality and frequency of caregiver-child communication, as well as the lack of value OSY place on their caregivers’ opinions about smoking and drinking [36, 77, 78]. This calls for further exploration on how to incorporate broader caregiving strategies into alcohol and tobacco prevention interventions among OSY.

Some OSY in this study attributed their alcohol and tobacco use to coping with household issues such as not being able to get along with parents, the stress of contributing towards the household income, feelings of neglect by parents or the loss of a parent. Moreover, previous studies have found that living in single-parent families, step-parent families or no-parent families are all associated with higher odds of ever/lifetime smoking and drinking [79-82]. Many children in South Africa grow up in fractured families usually without one or both biological parents. These families are also faced with poverty and unemployment [15, 76]. Studies have shown that socioeconomic disadvantaged adolescents are more likely to take up tobacco and alcohol use [83, 84]. Similarly, leaving school due to financial difficulties and lack of family support was associated with alcohol and tobacco use among OSY in South Africa, but the strength and direction of these associations were dependent on gender and geographical area [4, 5]. Currently, South African policies exempting students who do not have the affordability from paying tuition fees should be enacted [85]. Coping mechanisms and strategies can also be incorporated into alcohol and tobacco prevention interventions among OSY.

To our knowledge, this was the first study that explored the social network determinants of tobacco and alcohol use among OSY in South Africa. A limitation of the study was that the views expressed in the interviews are more representative of an older sample of OSY (average age 18 years) who are legally permitted to purchase tobacco and alcohol in South Africa. More research is needed among younger OSY between ages 13–15. The different interviewer methods may have influenced participant responses. However similar a previous study [57], the findings of this study suggest that data quality is unaffected by the mode of data collection. The findings of this study may not be unique to OSY and therefore the study may need to be replicated among in-school learners to draw further comparisons and conclusions with older and younger OSY. We acknowledge that sections of data may be included in multiple themes with some overlap between themes. However, the researchers considered how each theme fit into the overall story about the entire data set. This study does however provide some insight into the unique social network determinants of alcohol and tobacco use among OSY, which are confirmed by aspects of previous studies among OSY and in-school learners. Furthermore, due to OSY being a hard-to-reach population, transcripts were not returned to respondents for comment and/or correction due to difficulty in contacting them again.

Conclusions and implications

While similarities in social network determinants exist with in-school learners, this study additionally found that facilitators of alcohol and tobacco use include having more access to financial resources from peers and OSY having more OSY friends who may use alcohol and tobacco. OSY’s experience of the lack of parent support, coping with the loss of a parent and the larger role that extended family may play were also facilitators of alcohol and tobacco among OSY.

The findings of this study have important implications for alcohol and tobacco prevention and cessation programs among OSY. Attempts should be made to register all those who drop out of school to allow for tracking of school dropouts for intervention. Given that OSY do not have the protective factor of school-based interventions, interventions appropriate and accessible through the community is warranted. Furthermore, peer led programmes which have been shown to be successful among in-school learners [86, 87] could include a peer led OSY component in that encourage the diffusion of non-smoking and non-drinking norms in the community.

(DOCX)

Click here for additional data file.

(XLSX)

Click here for additional data file.

2 Jul 2020

PONE-D-20-17565

Social network determinants of alcohol and tobacco use: a qualitative study among out of school youth in South Africa

PLOS ONE

Dear Dr. Desai,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

In particular, please pay attention to 1) the rationale (why the focus on tobacco and alcohol?), 2) presentational issues (further details are required in your methods section), 3) depth of findings (in your results section, more quotes or longer quotes are needed to support your conclusions) and 4) interpretation (additional limitations need to be considered; what does this study add?).

Please submit your revised manuscript by Aug 16 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Lion Shahab, MA MSc MSc PhD CPsychol

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

2. Please include additional information regarding the interview guide or script used in the study and ensure that you have provided sufficient details that others could replicate the analyses.

For instance, if you developed a guide as part of this study and it is not under a copyright more restrictive than CC-BY, please include a copy, in both the original language and English, as Supporting Information.

3. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

4. Thank you for stating the following in the Acknowledgments Section of your manuscript:

'We would like to acknowledge the scholarship of the Foundation Study Fund for South African students in the Netherlands.'

We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:

'The authors received no specific funding for this work.'

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Partly

Reviewer #2: Partly

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: N/A

Reviewer #2: N/A

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: No

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: Abstract

1.Lines 28-30: "Interventions preventing the use of alcohol and tobacco may benefit from considering the family and friend social networks of OSY.” There is little mention of the development of interventions based on the findings of this study elsewhere in the manuscript. The authors may consider expanding upon this if it is to be included in the abstract. What kind of interventions? How would considering these factors help prevent alcohol and tobacco use? Why is it important to target OSY?

Introduction

1.Please provide some more detailed background about the schooling system in south Africa (public schools are fee paying/subsidised, free for low-income etc), and the system that most/all participants were recruited from. This will help frame the sample population for the reader.

2.Lines 37-39: Are there any estimates of prevalence other than past-month alcohol and tobacco use? For instance, weekly or daily use? The past month indicator will capture experimentation as well as more frequent usage, so it would be helpful to have another indicator of more frequent usage (if available!).

3.Lines 42-53: What is the socio-economic patterning of OSY? Presume it is largely driven by those who suffer greater disadvantage?

5.Lines 49-51: Might delinquency, poor mental health, and poor physical health also influence the likelihood of dropping out of school (i.e. some reverse causality?)

Materials and methods

1.The inclusion/exclusion criteria for participants in the study could be made clearer.

2.Lines 108-109: Please provide further details about the 4 smoker and 4 non-smoker sampling strategy, what is the rationale for this and how does this apply to alcohol use?

3.Lines 118-120: Why was WhatsApp used as a data collection tool in addition to in person interviews? Has this been done before and is it a valid method to collect this information? Why was telephone interviewing not considered? It may be a more flexible approach but it does it allow for skill of the interviewer to elicit deeper and more revealing responses?

4.Lines 122-124: How verify that individual was not in school? Monetary incentive might encourage recruitment of ineligible participants?

5.Why is qualitative methodology useful here? If the aim was to identify the composition of the social network, and how this influenced tobacco/alcohol use (e.g asking whether friends/family make the individual less likely/more likely to use product) might a simple survey have been more appropriate and less resource intensive?

Data collection

1.How many interviews were in person and how many were conducted via text message?

2.As above, is using text messages a valid methodology? It may be flexible but do the responses warrant inclusion with more in-depth interviews? Might people using WhatsApp to respond be qualitatively different to those who agreed to have in-person interviews?

Results

1.A summary table including information on the characteristics of the sample is necessary. While this is not a quantitative analysis, it would still be helpful for the reader to get an understanding of the participants. This could include the age range, gender, the number that were smokers/drinkers, how long they've been out of school etc)

2.Please give an indication of how many respondents are being referred to, rather than ‘some’. While quantifying the results may not be the goal of this qualitative research, it is helpful to know how many participants are represented in each result, especially in a large (n=41) qualitative sample.

3.Was any attempt made to explore the different ways in which tobacco or alcohol is used by OSY? Is there more frequent usage, or more experimentation?

4.Were there any differences between male and female OSY? Or between younger vs older participants?

5.The data appear to be lacking in depth. Most of the example quotes appear limited/short, and were not explored further. Might this reflect the WhatsApp interview technique? The data presented are do not currently seem to be sufficient to represent/support the findings.

6.Lines 149-152 in the analysis section mention the generation of themes and sub-themes. It is not clear what these themes are as they are not specifically mentioned in the results or discussion. The results are grouped under different members of the social network – are these the themes? Some clearer structure to the results, and corresponding discussion is needed.

7.The quotes could do with an anonymous identifier. Otherwise it is unclear whether the quotes used came from a wide range of participants or from a select few who gave more coherent responses. E.g. participant 1, female, age 19 or participant 2, male, age 20 etc…

A coding matrix in the supplementary materials would also be helpful. This could show what the collection of quotes were that generated the themes in this study. The method clearly outlines the process to reach each theme, but without some more examples of what content actually created them, it is difficult for the reader to get an idea about how valid the themes are. The form submitted with the manuscript mention that data is readily available, but it does not seem to have been submitted with the manuscript for review?

8.As above, it would be helpful to see the interview discussion/topic guide that was used.

9.Is the sample representative of younger OSY? What is the range of ages?

10.Line 184-186: Friends either played a facilitating or inhibiting role in the tobacco and alcohol use of current OSY users, but did not play a role among non-users of alcohol and/or tobacco.” This discrepant finding is interesting but requires more elaboration. What was the explanation as to why smoking/drinking friends did they not play a role initiation or use among these non-users?

11.Lines 213-214: The word ‘affordability’ does not make sense in this context. Affordability refers to the cost of something or its inexpensiveness. Perhaps ‘financial resources’ can be used instead?

Discussion

1.Lines 364-365: What about the participants who drank/smoke despite their partner disapproving? There is no discussion of this apparently discrepant finding.

2.Lines 390-392 This is the first time that religiosity is mentioned in the manuscript. If it is a key theme then it should be referred to/exemplified in the results section, before being discussed later on.

3.Lines: 408-412: Additional limitations need to be considered.

-It is possible that the findings in this study are not unique to OSY, but also reflect youth who attend school.

-Representativeness is mentioned as a limitation due to the geographic region, but does this not also apply to younger OSY? If there were no/few younger members in the sample then these results may reflect an older demographic (average age is 18, which in terms of ability to purchase substances is an important difference.)

-There is no mention of reflexivity of the researchers. How might this have influenced the conduct and analysis of this study? Has the relationship between researcher and participants been adequately considered? It would be helpful to see the researchers critically examine their own role, potential bias and influence during (a) formulation of the research questions (b) data collection, including sample recruitment and choice of location.

-Are there any limitations in using different methods of data collection? Using WhatsApp is convenient but it may limit the depth of responses due to it being a burden to write out a long response.

-Some more time spent outlining the importance of these findings is warranted. What new knowledge has been created? The introduction lines 58-61 outline some findings from research among school attending youth – these appear to be very similar to this study’s current findings/conclusions. What are the differences? This could be clearer in the discussion section.

4. As above - despite there being no comparison with school attendees in this study, how might these results differ compared to previous research?

5.This study focuses on the social networks of OSY, but the wider determinants of smoking/alcohol need to be considered in the discussion. E.g. The availability of cigarettes/alcohol, any wider socio-economically patterns in usage, government policies influencing price/availability.

6.More discussion is warranted on contribution the study makes to existing knowledge or understanding. Use in intervention development is mentioned but what exactly would this look like? What targets have been created for intervention development? Why would an OSY tailored intervention be more beneficial than a more generic intervention targeting all youth (given that the social-network determinants appear to be similar between the two)?

Reviewer #2: In this manuscript, the authors explored through qualitative research social network determinants of alcohol and tobacco use among out of school youth (OSY) in South Africa. They conducted 41 interviews with OSY in a South African urban area and their findings suggest that determinants of alcohol and tobacco use include drinking friends, family members close in age to OSY that drank and smoked and household issues. On the other hand, they found that religiosity, parental control over alcohol and tobacco, romantic partners, and non-smoking or non-drinking friends alleviated alcohol and tobacco use.

The topic is very interesting, but the manuscript needs more attention to the use of English language. Additionally, it could be improved in the following respects:

1)In the abstract the authors mention that ‘an important determinant of alcohol, and tobacco use is the adolescent's social network, which has globally not been explored among out of school youth (OSY). However, the paper only includes participants from South Africa.

2) It is also good practice for qualitative research to state the analytic method in the abstract.

3)The last sentence of the abstract is very general. Authors only look at OSY, so their suggestions for interventions should be relevant to this group of people.

4)Authors should also give a better definition of adolescent and out of school youth regarding the age range. They provide information for different age groups in the Introduction and then they state that their sample included participants age between 13-20 years old. But there is no justification why they chose this age range.

5) Authors present % of tobacco and alcohol use among school going learners and OSY. It would be clearer if they present these figures in the same paragraph.

6) There is no clear explanation why authors are focusing on alcohol and tobacco use and not other addictive behaviours?

7) It is not clear why authors needed 52 participants for their study and if they recruited more than 52. Additionally, why did authors initial recruit 4 smokers and 4 non-smokers and not 2 smokers, 2 non-smokers, 2 drinkers and 2 non-drinkers?

8) How many interviews conducted in person and how many through whatsapp? The authors could perhaps also reflect on whether any differences in quality were observed across those interviewed via the different modalities.

9) Data collection tools would be clearer if they include examples of questions and probs (e.g. including the topic guide as a supplementary file). Did the authors assess participants’ alcohol and tobacco use and how?

10) The authors need to specify who gave them ethical approval.

11) Results section: How many participants use both alcohol and tobacco?

12) It is not clear if the friends of participants are also OSY or not.

13) Results section could benefit with more examples, interview quotations to support the results. Especially results about parents the example provided did not support the findings. The authors also mention that ‘Often respondents hid their alcohol and tobacco use from their family or use little to no tobacco or alcohol in the presence of their older family members out of respect or fear’, but the example provided did not support such findings. It also appears that the analysis is a bit shallow. For example, the theme ‘Friends’ is very broad, and the presented quotation suggests that participants were bored and drinking/smoking because there isn’t much else to do. They also mentioned having resources/money to buy alcohol/tobacco, which is more of an ‘opportunity’ rather than ‘social’ variable that influences use. Again, the theme ‘Parent’ doesn’t quite seem to capture what’s expressed by participants (e.g. drinking due to lack of parental support/grief.

14) Some discussion points are not supported by the findings i.e. ‘Among OSY who were current tobacco and alcohol users, OSY tended to spend more time with other OSY friends’; ‘ These findings may explain why a different sample of females, who dropped out of school due to financial difficulties in South Africa were smoking cigarettes, as seen in a previous study by Desai and colleagues (13)’. In the second example it is not also clear where different refers to?

15) Authors should avoid use language such as substance users when they are only referring to alcohol and tobacco use

16) Discussion about romantic partners is not clear. For example, the point that ‘compared to friends, OSY in this sample would smoke and drink less with their romantic partners, possibly due to spending less time with romantic partners in the context of smoking and drinking’ is not clear how is supported by the results.

17) It would be useful to discuss implications for policy and avenues for future research.

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: No

Reviewer #2: Yes: Dimitra Kale

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

16 Aug 2020

We thank the reviewers for these helpful comments, which we have used to strengthen our paper. We respond to each comment below and indicated the changes in track changes in the paper. Reference to lines in the document pertain to the version with no track changes

1. Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

https://journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and

https://journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

We have formatted the paper to meet the journals style requirements

2. Please include additional information regarding the interview guide or script used in the study and ensure that you have provided sufficient details that others could replicate the analyses.

For instance, if you developed a guide as part of this study and it is not under a copyright more restrictive than CC-BY, please include a copy, in both the original language and English, as Supporting Information.

The English version of the interview guide has been provided as Supporting information

3. We note that you have stated that you will provide repository information for your data at acceptance. Should your manuscript be accepted for publication, we will hold it until you provide the relevant accession numbers or DOIs necessary to access your data. If you wish to make changes to your Data Availability statement, please describe these changes in your cover letter and we will update your Data Availability statement to reflect the information you provide.

The data has been provided as Supporting information titled “coding matrix”. We have requested for the data availability statement to be changed in the revised cover letter.

4. Thank you for stating the following in the Acknowledgments Section of your manuscript:

'We would like to acknowledge the scholarship of the Foundation Study Fund for South African students in the Netherlands.'

We note that you have provided funding information that is not currently declared in your Funding Statement. However, funding information should not appear in the Acknowledgments section or other areas of your manuscript. We will only publish funding information present in the Funding Statement section of the online submission form.

Please remove any funding-related text from the manuscript and let us know how you would like to update your Funding Statement. Currently, your Funding Statement reads as follows:

'The authors received no specific funding for this work.'

Please include your amended statements within your cover letter; we will change the online submission form on your behalf.

The acknowledgments section has been removed. The authors did not receive specific funding for this work. We have included these changes in the revised cover letter.

Review Comments to the Author

Reviewer #1: Abstract

1.Lines 28-30: "Interventions preventing the use of alcohol and tobacco may benefit from considering the family and friend social networks of OSY.” There is little mention of the development of interventions based on the findings of this study elsewhere in the manuscript. The authors may consider expanding upon this if it is to be included in the abstract. What kind of interventions? How would considering these factors help prevent alcohol and tobacco use? Why is it important to target OSY?

We thank the reviewer for the comment. After revising the manuscript and taking all the comments into consideration, it would appropriate to state the following conclusion in the abstract: “Understanding how the social network of OSY plays a role in alcohol and tobacco use is useful for gaining an insight into profile of OSY at risk for alcohol and tobacco use. Registration of OSY youth and community-based peer led programmes that include influential OSY family and friends could be beneficial. We feel that this conclusion sits closer to the study findings. We feel that this conclusion sits closer to the study findings.

Introduction

1.Please provide some more detailed background about the schooling system in south Africa (public schools are fee paying/subsidised, free for low-income etc), and the system that most/all participants were recruited from. This will help frame the sample population for the reader.

Thank you for your comment.

This study was conducted in the Western Cape, the fourth largest province in South Africa within an urban district characterised by high rates of school dropout. Schooling is compulsory for all South African children from the age of 6 (grade 1) to the age of 15 (grade 9). Primary education consists of two phases: the foundation phase; and the intermediate phase. All school governing boards of public schools must supplement government funding, by charging school fees and doing other reasonable forms of fund-raising. Caregivers who cannot afford to pay school fees may apply to the school governing board for conditional, partial, or full exemption from paying school fees. The right not to charge school fees is limited to the schools that have been declared ‘no fee schools,’ which is based on the economic level of the community around the school. From grade 9 however, children do not benefit from the no-fee policy. Eligibility for full and partial school fee exemptions is worked out on the basis of parental income in relation to the fees.

Although we did not ask respondents the school they previously attended, they were recruited from urban areas that comprised of both no paying and fee-paying schools.

We added a sub-section “study setting” in the methods section to indicate the details of the south African schooling system and the context in which the study took place. We trust that these edits will allow for the paper to read better and provide more context.

2.Lines 37-39: Are there any estimates of prevalence other than past-month alcohol and tobacco use? For instance, weekly or daily use? The past month indicator will capture experimentation as well as more frequent usage, so it would be helpful to have another indicator of more frequent usage (if available!).

Thank you for the comment. Past month and lifetime alcohol and tobacco use prevalence rates are available for in-school and out of school youth between 13-20 years, which is the age range considered in this study. Although national prevalence of weekly use of alcohol and daily use of tobacco among adolescents aged 15-19 years is available from the South African Demographic Health Survey, it is not clear whether these adolescents are in-school or out of school. We therefore used past month tobacco and alcohol use prevalence rates. Moreover, we will follow the second reviewer’s suggestion of presenting the prevalence of past month tobacco and alcohol use among school going learners and OSY in the same paragraph to allow for clear comparison.

3.Lines 42-53: What is the socio-economic patterning of OSY? Presume it is largely driven by those who suffer greater disadvantage?

We have added more detail on the profile of OSY. The South African literature shows that reasons for leaving school include poverty, high use of substances, bullying, boredom, family needs (helping support the families, being pregnant, traditional family role expectations), illness, disability, community violence, and school related factors (academic performance, disliking school, not getting along with teachers, being too old for school and disciplinary consequences). We trust that his information provides a better understanding of the profile of OSY in South Africa.

5.Lines 49-51: Might delinquency, poor mental health, and poor physical health also influence the likelihood of dropping out of school (i.e. some reverse causality?)

The reverse is certainly true and therefore we changed the phrasing of this sentence to not imply causality. The sentence now reads as “Negative social, health and economic consequences such as unemployment, substance use, delinquency, and poor mental and physical health is associated with early school leaving.”

Materials and methods

1.The inclusion/exclusion criteria for participants in the study could be made clearer.

Adolescents between the ages of 13-20 years who is not currently enrolled in secondary or high school for that academic year and has not completed their secondary or high schooling were eligible. Those who were enrolled in college or vocational training were excluded. We trust that this inclusion and exclusion criteria provides more clarity.

2.Lines 108-109: Please provide further details about the 4 smoker and 4 non-smoker sampling strategy, what is the rationale for this and how does this apply to alcohol use?

The initial focus of the paper was tobacco use among OSY and their social network. As the paper evolved, we focused on both alcohol and tobacco use. Despite targeting smokers and non-smokers as the initial seeds, we still acquired a sample that contains almost an even number of alcohol and tobacco users while minimising potential bias introduced, due to the nature of the respondent driven sampling strategy.

3.Lines 118-120: Why was WhatsApp used as a data collection tool in addition to in person interviews? Has this been done before and is it a valid method to collect this information? Why was telephone interviewing not considered? It may be a more flexible approach but it does it allow for skill of the interviewer to elicit deeper and more revealing responses?

Given the wide use of cell phones among adolescents, cell phone mediated online communication technologies and instant messaging (IM) services have become popular ways of communication. Studies have shown that compared to other online modalities, adolescents prefer sharing information and communicating via instant messaging (IM) (Lee, 2007; Shapka, Domene, Khan, & Yang, 2016).

Online interviewing allows access to the voices and experiences of “hidden” populations (Adler & Zarchin, 2002; Ayling & Mewse, 2009; Mathy et al., 2002; Turney & Pocknee, 2005), characterised as socially disadvantaged groups that are difficult for researchers to access due to the lack of a sampling frame and geographical location (Heckathorn, 1997). Compared to the traditional face-to-face interviewing method, qualitative online interviewing saves cost and time, and is convenient for both the researcher and the participant (Jowett et al., 2011; Shapka et al., 2016).

In the case of text based online interviewing, the use of emoticons and internet slang abbreviations such as “LOL” are commonly used to convey emotion, and has been used previously by interviewers to build rapport (Jowett et al., 2011; Kazmer & Xie, 2008).

One study systematically compared face-to-face interviewing techniques to instant messaging interviewing, mediated by a computer (Shapka et al., 2016). Shapka et al. (2016) found that despite taking longer and producing fewer words in the online condition, data quality was unaffected by the mode of data collection (online versus face-to-face) with no differences in the number, depth and type of themes discussed. This suggests that online data collection using cell phone mediated IM may be a novel, appropriate and feasible method for obtaining sensitive information from school dropouts.

Given the novelty of interviewing over WhatsApp, we will be writing a separate paper detailing the equivalence of interview data collected online using text over WhatsApp, with that obtained face-to-face to obtain qualitative data from a hidden population of school dropouts. Given the focus and length of the current paper, the above information was omitted. However, we are open to suggestions on how we should include this information in the manuscript.

4.Lines 122-124: How verify that individual was not in school? Monetary incentive might encourage recruitment of ineligible participants?

Eligibility was further confirmed through contacts that the potential participant provided. This detail has been added to the manuscript.

5.Why is qualitative methodology useful here? If the aim was to identify the composition of the social network, and how this influenced tobacco/alcohol use (e.g asking whether friends/family make the individual less likely/more likely to use product) might a simple survey have been more appropriate and less resource intensive?

Limited research is known about the composition of OSY social network and the similarities and differences of their social network to in-school learners. Qualitative methods provide a voice to this hidden population and ensures that study findings are grounded in participants' context and experiences. Qualitatively understanding how OSY’s social networks support or constrain alcohol and tobacco use is useful for providing additional validity and contextualization when developing survey instruments.

Data collection

1.How many interviews were in person and how many were conducted via text message?

12 respondents conducted the interview over a cellular phone application WhatsApp, and 29 respondents conducted the interview face-to-face. We added this detail to the manuscript.

2.As above, is using text messages a valid methodology? It may be flexible but do the responses warrant inclusion with more in-depth interviews? Might people using WhatsApp to respond be qualitatively different to those who agreed to have in-person interviews?

Text-based WhatsApp interviews are a valid methodology for the reasons stated above. In the context of the coronavirus, this way of conducting qualitative interviews may offer a good alternative to the traditional face-to-face interviews.

To minimise bias introduced in the interview conditions, simple randomisation was used to assign the seeds into either the in-person interview condition or WhatsApp interview condition. The seeds further recruited into the interview conditions they were assigned into. If a participant could not conduct the interview in the condition they were assigned into, the alternative interview condition was offered, and they were no longer required to further recruit more participants. Once again, we have omitted this information due to the length and focus of the paper but we are open to suggestions on how to include this information if necessary.

*Results

1.A summary table including information on the characteristics of the sample is necessary. While this is not a quantitative analysis, it would still be helpful for the reader to get an understanding of the participants. This could include the age range, gender, the number that were smokers/drinkers, how long they've been out of school etc)

We thank the reviewers for the suggestion and included a table of characteristics available to us. Alcohol and tobacco behavioural characteristics were presented across gender. Age was described in the text following the table.

2.Please give an indication of how many respondents are being referred to, rather than ‘some’. While quantifying the results may not be the goal of this qualitative research, it is helpful to know how many participants are represented in each result, especially in a large (n=41) qualitative sample.

We have indicated the number of respondents referred to where possible

3.Was any attempt made to explore the different ways in which tobacco or alcohol is used by OSY? Is there more frequent usage, or more experimentation?

Respondents were asked to indicate if they used tobacco and or alcohol at the time of the interview, their age of initiation, frequency of smoking and/or drinking per week. Those who did not drink, or smoke were asked about their lifetime drinking or smoking and reasons for quitting. A brief summary description of the frequency of smoking and drinking has been added under the characteristics of the sample. The discussion guide and coding matrix has been included as a supplementary file.

4.Were there any differences between male and female OSY? Or between younger vs older participants?

We unfortunately only had older OSY sample in this study. Using either tobacco or alcohol was similar across gender. However, using both alcohol and tobacco was common among males compared to females. The gender breakdown of alcohol and tobacco use has been added to the sample characteristics table.

Examining the gender differences was not the main focus of this study however, findings somewhat indicate that being assisted financially by friends and using alcohol and tobacco due to leisure boredom was commonly cited among males. OSY resistance to alcohol and tobacco use and discouragement from romantic partners was mostly cited among the females in this study. We felt that given the small sample size differences found between genders, strong conclusion could not be made between male and female in the study.

5.The data appear to be lacking in depth. Most of the example quotes appear limited/short, and were not explored further. Might this reflect the WhatsApp interview technique? The data presented are do not currently seem to be sufficient to represent/support the findings.

We have taken the reviewers comments into consideration and added more example quotations to support the findings. We also considered re-ordering the presentation of the quotations, added subthemes where appropriate and explored those quotations in more depth. We trust that the revised presentation of results strengthens the paper.

6.Lines 149-152 in the analysis section mention the generation of themes and sub-themes. It is not clear what these themes are as they are not specifically mentioned in the results or discussion. The results are grouped under different members of the social network – are these the themes? Some clearer structure to the results, and corresponding discussion is needed.

We thank the reviewer for the comment. We have clarified the main themes to be friends and family. The subthemes for friends were initiation of alcohol and tobacco, financial resources, leisure boredom and resistance to tobacco and alcohol. Family subthemes were those of similar age, those elderly, lack of parental support and household norms. The results have been restructured to reflect these themes and sub-themes.

7.The quotes could do with an anonymous identifier. Otherwise it is unclear whether the quotes used came from a wide range of participants or from a select few who gave more coherent responses. E.g. participant 1, female, age 19 or participant 2, male, age 20 etc…

A coding matrix in the supplementary materials would also be helpful. This could show what the collection of quotes were that generated the themes in this study. The method clearly outlines the process to reach each theme, but without some more examples of what content created them, it is difficult for the reader to get an idea about how valid the themes are. The form submitted with the manuscript mention that data is readily available, but it does not seem to have been submitted with the manuscript for review?

We thank the reviewer for the comment. We included an anonymous identifier after each quotation. We have also included a coding matrix with code groups and corresponding quotations, which has been included as a supplementary file. All coded data has been included in this file and will be readily available. We trust that this will reflect the process of how we reached each theme.

8.As above, it would be helpful to see the interview discussion/topic guide that was used.

This has been included as a supplementary file

9.Is the sample representative of younger OSY? What is the range of ages?

Given the average age range of participants was 18 years (SD=1.2), the sample in this study represents older OSY, possibly due to the inclusion of older OSY as initial seeds in the RDS recruitment strategy. The study could have benefited from including seeds from younger OSY groups to get a more representative sample. We have included this in the limitations.

10.Line 184-186: Friends either played a facilitating or inhibiting role in the tobacco and alcohol use of current OSY users, but did not play a role among non-users of alcohol and/or tobacco.” This discrepant finding is interesting but requires more elaboration. What was the explanation as to why smoking/drinking friends did they not play a role initiation or use among these non-users?

We realise that it would be better to move lines 184-186 to the discussion section as it is more appropriate as a summary of findings about friends. We provide additional quotes and context to smoking/drinking friends not playing a role in the initiation or use among the non-users. We further analysed non-users and it was found that they would engage in other leisure activities such as going to the mall, sports and attending church. They also claim that their friends did not excessively use alcohol and tobacco.

An in-depth explanation as to why smoking/drinking friends did not play a role initiation or use among these non-users is offered in the discussion section and we trust that this is also sufficient.

11.Lines 213-214: The word ‘affordability’ does not make sense in this context. Affordability refers to the cost of something or its inexpensiveness. Perhaps ‘financial resources’ can be used instead?

We agree with this comment and made the changes in the manuscript

*Discussion

1.Lines 364-365: What about the participants who drank/smoke despite their partner disapproving? There is no discussion of this apparently discrepant finding.

Our results mainly reflected that respondents would drink and smoke less and were discouraged by their romantic partners. We did not find quotations that support OSY who drank and smoked despite their partner disapproving. Given the restructuring of the results section, we now decided to combine the finding pertaining to romantic partners in the previous paragraph with friends. Romantic partners are not the main finding of this study and therefore a full discussion is not warranted. We trust that this presentation of results allows for the paper to read better.

2.Lines 390-392 This is the first time that religiosity is mentioned in the manuscript. If it is a key theme then it should be referred to/exemplified in the results section, before being discussed later on.

We acknowledge that the way religiosity is presented in the discussion section makes this concept seem like a main finding. We have therefore in the results section considered re-ordering the presentation of the quotations, added subthemes where appropriate and explored those quotations in more depth. Religiosity is not necessarily a key theme but it is worth mentioning in the discussion section that lack of religious norms along with lack of in-home smoking rules, poor quality and frequency of caregiver-child communication, as well as how much adolescents value their caregivers’ opinions about smoking and drinking play a role on OSY decision to smoke or drink alcohol (lines 468-471). We trust that this presentation of results and discussion provides more clarity

3.Lines: 408-412: Additional limitations need to be considered.

-It is possible that the findings in this study are not unique to OSY, but also reflect youth who attend school.

We agree that the findings of this study may not be unique to OSY and therefore the study may need to be replicated among younger and older in-school learners to draw further comparisons and conclusions. This study does however provide some insight into the unique social network determinants of alcohol and tobacco use among OSY, which are confirmed by aspects of previous studies among OSY and in-school learners. We have included this statement in the limitations section.

-Representativeness is mentioned as a limitation due to the geographic region, but does this not also apply to younger OSY? If there were no/few younger members in the sample then these results may reflect an older demographic (average age is 18, which in terms of ability to purchase substances is an important difference.)

We thank the reviewer for this comment and mentioned that the OSY in this sample were representative of an older OSY who are legally permitted to purchase tobacco and alcohol in South Africa.

-There is no mention of reflexivity of the researchers. How might this have influenced the conduct and analysis of this study? Has the relationship between researcher and participants been adequately considered? It would be helpful to see the researchers critically examine their own role, potential bias and influence during (a) formulation of the research questions (b) data collection, including sample recruitment and choice of location.

All authors of the study formulated the questions through a reflective, iterative and dialogic process. Respondents were given the lead in ‘setting the pace’ of the interview either over WhatsApp or face-to-face. Data collectors deliberately adopted a ‘back seat’ approach for participants to feel that they were exercising a measure of control over the interview process. During the data collection and recruitment, data collectors kept memos for each interview containing reflections, feelings towards the participant and potential biases introduced. A focus group with data collectors and the authors was conducted to understand how the data collectors’ interactions with participants might be influenced by their own professional background, experiences, and prior assumptions. Given the length of the current paper, we did not include these details in the manuscript, but would certainly do so if the editor will allow it.

-Are there any limitations in using different methods of data collection? Using WhatsApp is convenient but it may limit the depth of responses due to it being a burden to write out a long response.

There were no limitations in using the different data methods of data collection. Studies have shown that compared to other online modalities, adolescents prefer sharing information and communicating via instant messaging (IM) (Lee, 2007; Shapka, Domene, Khan, & Yang, 2016). Moreover, the results of this study concur with the results found in the study by Shapka et al (2016). Interviews conducted online produced fewer words and took longer to complete, however, there were no mean differences in the number and kind of themes that emerged or in the depth to which the themes were discussed. The findings suggest that despite taking longer and producing fewer words, data quality is unaffected by the mode of data collection

-Some more time spent outlining the importance of these findings is warranted. What new knowledge has been created? The introduction lines 58-61 outline some findings from research among school attending youth – these appear to be very similar to this study’s current findings/conclusions. What are the differences? This could be clearer in the discussion section.

Thank you for the comment. we have improved the discussion section by clearly stating from the outset the unique contributions of this study pertaining to OSY. The first paragraph of the discussion reads as:

The findings of this study have provided valuable insight into the profile of OSY at risk for alcohol and tobacco use. Overall, the social environment, which included friends and family, played a facilitating or inhibiting role in OSY’s alcohol and tobacco use. While similarities in social network determinants exist with in-school learners, this study additionally found that facilitators of alcohol and tobacco use include having more access to financial resources from peers and OSY having more OSY friends who may use alcohol and tobacco. OSY’s experience of the lack of parent support, coping with the loss of a parent and the larger role that extended family may play were also facilitators of alcohol and tobacco among OSY.

4. As above - despite there being no comparison with school attendees in this study, how might these results differ compared to previous research?

The findings of this research provide a better understanding of the composition of the OSY social network as well as the unique social network facilitators and inhibitors of alcohol and tobacco use among OSY. We have revised the discussion section to highlight how the findings of this study corroborate previous studies among in-school and OSY, as well as offer additional insights into the social network determinants of alcohol and tobacco use among OSY.

5.This study focuses on the social networks of OSY, but the wider determinants of smoking/alcohol need to be considered in the discussion. E.g. The availability of cigarettes/alcohol, any wider socio-economically patterns in usage, government policies influencing price/availability.

Thank you for the comment. We have revised the results and discussion section to include wider determinants of alcohol and tobacco use

The following additions were made:

Line 234: Despite national policy prohibiting the sale of cigarettes to minors, OSY were able to purchase cigarettes from the shop on their rout to the transport routes.

Line 278-280: Given national regulations prohibiting the sale of alcohol and tobacco to minors, older OSY may have also purchased the alcohol and tobacco on behalf of the group.

Line 453-455: Enforcement of national policy prohibiting the sale of alcohol and tobacco to minors or raising the minimum age of tobacco and alcohol purchases may be beneficial.

Line 483: Studies have shown that socioeconomic disadvantaged adolescents are more likely to take up tobacco and alcohol use

Line 486: South African policies exempting students who do not have the affordability from paying tuition fees should be enacted.

6.More discussion is warranted on contribution the study makes to existing knowledge or understanding. Use in intervention development is mentioned but what exactly would this look like? What targets have been created for intervention development? Why would an OSY tailored intervention be more beneficial than a more generic intervention targeting all youth (given that the social-network determinants appear to be similar between the two)

We thank the reviewer for the comments and used this suggestion to strengthen the paper. We included a “conclusions and implications” section highlighting the unique contributions of this study as well as possible targets for intervention. Attempts should be made to register all those who drop out of school to allow for tracking of school dropouts for intervention. Given that OSY do not have the protective factor of school-based interventions, interventions appropriate and accessible through the community is warranted. Furthermore, peer led programmes which have been shown to be successful among in-school learners could include a peer led OSY component in that encourage the diffusion of non-smoking and non-drinking norms in the community

Reviewer #2: In this manuscript, the authors explored through qualitative research social network determinants of alcohol and tobacco use among out of school youth (OSY) in South Africa. They conducted 41 interviews with OSY in a South African urban area and their findings suggest that determinants of alcohol and tobacco use include drinking friends, family members close in age to OSY that drank and smoked and household issues. On the other hand, they found that religiosity, parental control over alcohol and tobacco, romantic partners, and non-smoking or non-drinking friends alleviated alcohol and tobacco use.

The topic is very interesting, but the manuscript needs more attention to the use of English language. Additionally, it could be improved in the following respects:

We thank the reviewer for the compliment and have used the comments to improve the manuscript

1)In the abstract the authors mention that ‘an important determinant of alcohol, and tobacco use is the adolescent's social network, which has globally not been explored among out of school youth (OSY). However, the paper only includes participants from South Africa.

We recognise that the sentence creates confusion and we therefore removed the word “globally”.

2) It is also good practice for qualitative research to state the analytic method in the abstract.

The data was analysed using thematic content analysis, which has been added to the abstract

3)The last sentence of the abstract is very general. Authors only look at OSY, so their suggestions for interventions should be relevant to this group of people.

Thank you for the comment. We have improved the abstract by stating the following: understanding how the social network of OSY plays a role in alcohol and tobacco use is useful for gaining an insight into the profile of OSY at risk for alcohol and tobacco use. Registration of OSY youth and community-based peer led programmes that include influential OSY family and friends could be beneficial.

4) Authors should also give a better definition of adolescent and out of school youth regarding the age range. They provide information for different age groups in the Introduction and then they state that their sample included participants age between 13-20 years old. But there is no justification why they chose this age range.

Although the WHO defines adolescents as those between ages 10-19 years, this study has focused on adolescents between 13-20 years. A slightly older sample of adolescents was targeted because studies have shown that dropout tends to increase from the age of 15 (grade 9). Moreover, alcohol and tobacco use is usually initiated between 12-14 years and is highly prevalent in this age group. To our knowledge, no qualitative study has been conducted, specifically focusing on 13-20-year-old dropouts in South Africa. We added this detail to the introduction.

5) Authors present % of tobacco and alcohol use among school going learners and OSY. It would be clearer if they present these figures in the same paragraph.

We agree with your suggestion and moved the prevalence rates to reflect in the same paragraph.

6) There is no clear explanation why authors are focusing on alcohol and tobacco use and not other addictive behaviours?

National studies show that alcohol and tobacco are the most prevalent among adolescents compared to other addictive behaviours such as illegal and other drug use. We added this statement in the introduction and hope this provides clarity. While we acknowledge that other addictive behaviours warrant investigation, a separate study focusing would be beneficial as it may be burdensome on participants to cover all addictive behaviours in a single study. Moreover, eliciting specific information on alcohol and tobacco would be more meaningful for the development of specific interventions.

7) It is not clear why authors needed 52 participants for their study and if they recruited more than 52. Additionally, why did authors initial recruit 4 smokers and 4 non-smokers and not 2 smokers, 2 non-smokers, 2 drinkers and 2 non-drinkers?

In order to complete the full respondent driven sampling process of two waves of recruitment with 8 seeds, it was proposed that up to 52 participants can be recruited. This has been made clearer in the manuscript. As indicated under the data collection section, 41 participants were sufficient as no new views were emanating from the interviews since the last three respondents.

The initial focus of the paper was tobacco use among OSY and their social network. As the paper evolved, we focused on both alcohol and tobacco use. Despite targeting smokers and non-smokers as the initial seeds, we still acquired a sample that contains almost an even number of alcohol and tobacco users while minimising potential bias introduced, due to the nature of the respondent driven sampling strategy.

8) How many interviews conducted in person and how many through whatsapp? The authors could perhaps also reflect on whether any differences in quality were observed across those interviewed via the different modalities.

12 respondents conducted the interview over a cellular phone application WhatsApp, and 29 respondents conducted the interview face-to-face. From our preliminary unpublished analysis, the content from both Whatsapp and face-to-face interviews were relevant. Studies have shown that compared to other online modalities, adolescents prefer sharing information and communicating via instant messaging (IM) (Lee, 2007; Shapka, Domene, Khan, & Yang, 2016). Moreover, the results of this study concur with the results found in the study by Shapka et al (2016). Interviews conducted online produced fewer words and took longer to complete, however, there were no mean differences in the number and kind of themes that emerged or in the depth to which the themes were discussed. The findings suggest that despite taking longer and producing fewer words, data quality is unaffected by the mode of data collection

9) Data collection tools would be clearer if they include examples of questions and probs (e.g. including the topic guide as a supplementary file). Did the authors assess participants’ alcohol and tobacco use and how?

Participants were asked to discuss their own alcohol and tobacco use and alcohol and tobacco use of their social network. These details have been added to the manuscript and we have included the full interview guide as a supplementary file.

10) The authors need to specify who gave them ethical approval.

Ethical approval was obtained from the Human Sciences Research Council. This detail has been added to the manuscript.

11) Results section: How many participants use both alcohol and tobacco?

We have included this in the table of sample characteristics

12) It is not clear if the friends of participants are also OSY or not.

We found that in general the participants had friends who were both in-school, and out of school

13) Results section could benefit with more examples, interview quotations to support the results. Especially results about parents the example provided did not support the findings. The authors also mention that ‘Often respondents hid their alcohol and tobacco use from their family or use little to no tobacco or alcohol in the presence of their older family members out of respect or fear’, but the example provided did not support such findings. It also appears that the analysis is a bit shallow. For example, the theme ‘Friends’ is very broad, and the presented quotation suggests that participants were bored and drinking/smoking because there isn’t much else to do. They also mentioned having resources/money to buy alcohol/tobacco, which is more of an ‘opportunity’ rather than ‘social’ variable that influences use. Again, the theme ‘Parent’ doesn’t quite seem to capture what’s expressed by participants (e.g. drinking due to lack of parental support/grief.

We thank the reviewers for their comment and have revised the results section. We have included subthemes under the broad themes “friends” and “family” to closely reflect and organise the data better.

The subthemes for friends were initiation of alcohol and tobacco, financial resources, leisure boredom and resistance to tobacco and alcohol. Family subthemes were those of similar age, those elderly, lack of parental support and household norms. The results have been restructured to reflect these themes and sub-themes. More example quotations under each theme were added followed by a revised and in-depth analysis of the theme.

We provide additional quotes and context to smoking/drinking friends not playing a role in the initiation or use among the non-users. We further analysed non-users and it was found that they would engage in other leisure activities such as going to the mall, sports and attending church. They also claim that their friends did not excessively use alcohol and tobacco

We trust that the results section reads better.

14) Some discussion points are not supported by the findings i.e. ‘Among OSY who were current tobacco and alcohol users, OSY tended to spend more time with other OSY friends’; ‘ These findings may explain why a different sample of females, who dropped out of school due to financial difficulties in South Africa were smoking cigarettes, as seen in a previous study by Desai and colleagues (13)’. In the second example it is not also clear where different refers to?

Now that we have introduced the subtheme leisure boredom and included more quotation examples, this discussion point is supported. In the second example, the word “different” refers to the previous study’s sample not being the same sample as the current study. We rephrased this sentence to make it clear. We trust that the paragraph reads better with these changes.

15) Authors should avoid use language such as substance users when they are only referring to alcohol and tobacco use

We have removed instances where we use this language in the methods, results and discussion section

16) Discussion about romantic partners is not clear. For example, the point that ‘compared to friends, OSY in this sample would smoke and drink less with their romantic partners, possibly due to spending less time with romantic partners in the context of smoking and drinking’ is not clear how is supported by the results.

Thank you for the comment. The way the results section has been reorganised, we have combined the discussion on romantic relationships with friends. Romantic partners are not the main finding of this study and therefore a full discussion on this point is not warranted. Further exploration of romantic relationships is needed in future studies. We trust that this presentation provides more clarity.

17) It would be useful to discuss implications for policy and avenues for future research.

We have included a conclusions and implications paragraph highlighting the recommendations for policy and programmes.

Submitted filename: Response to Reviewers.docx

Click here for additional data file.

26 Aug 2020

PONE-D-20-17565R1

Social network determinants of alcohol and tobacco use: a qualitative study among out of school youth in South Africa

PLOS ONE

Dear Dr. Desai,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

As you can see, Reviewer 2 has a number of additional suggestions for how to improve presentation and interpretation of your results (see below). In particular, please pay attention to 1) reporting transparently the change in focus of the study 2) ensure that presented themes and subthemes are supported by appropriate quotes and 3) discuss study limitations in greater detail.

Please submit your revised manuscript by Oct 10 2020 11:59PM. If you will need more time than this to complete your revisions, please reply to this message or contact the journal office at plosone@plos.org. When you're ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). You should upload this letter as a separate file labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. You should upload this as a separate file labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. You should upload this as a separate file labeled 'Manuscript'.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter. Guidelines for resubmitting your figure files are available below the reviewer comments at the end of this letter.

If applicable, we recommend that you deposit your laboratory protocols in protocols.io to enhance the reproducibility of your results. Protocols.io assigns your protocol its own identifier (DOI) so that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

We look forward to receiving your revised manuscript.

Kind regards,

Lion Shahab, MA MSc MSc PhD CPsychol

Academic Editor

PLOS ONE

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. If the authors have adequately addressed your comments raised in a previous round of review and you feel that this manuscript is now acceptable for publication, you may indicate that here to bypass the “Comments to the Author” section, enter your conflict of interest statement in the “Confidential to Editor” section, and submit your "Accept" recommendation.

Reviewer #1: All comments have been addressed

Reviewer #2: (No Response)

**********

2. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Partly

**********

3. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: N/A

Reviewer #2: N/A

**********

4. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

5. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: Yes

Reviewer #2: Yes

**********

6. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: The authors have responded well to the reviewer comments and made appropriate changes. If the editors will allow it, some more space to allow the authors to provide their justification for the whatsapp interview technique will strengthen the methods section.

Reviewer #2: The authors have been mostly responsive to the issues raised during the previous review. The manuscript has been improved, but a few comments remain:

1) Introduction, lines (44-45) ‘Alcohol and tobacco use, like in many other countries is prevalent among adolescents’. Not clear if authors are referring to LMICs or South Africa.

2) Authors provide a definition of adolescents, which defines adolescents as those aged 10-19 years old. Not clear why then they include participants 20 years old?

3) Authors state that they added this statement in the introduction ‘National studies show that alcohol and tobacco are the most prevalent among adolescents compared to other addictive behaviours such as illegal and other drug use’. But I can’t find it in the Introduction. Could they please indicate the lines?

4) In the methods section authors state (lines 148-151) ‘Initially, eight seeds were obtained, and they were required to identify up to two other OSY. These respondents recruited by the seeds formed the “first wave” of sampling and were themselves asked to identify and refer a further two more school dropouts (Fig 1). Up to two waves of recruitment were conducted. participants describe the method of recruitment’. Based on this description, the target should have been 56 participants and not 52 as stated in line 170.

5) Authors should also report the initial focus of the paper as mentioned in their response, in order to justify why they targeted only smokers and non-smokers.

6) Regarding my comment on ‘How many interviews conducted in person and how many through whatsapp? The authors could perhaps also reflect on whether any differences in quality were observed across those interviewed via the different modalities’. The authors provided the requested numbers, but they did not include any reflection in the manuscript.

7) Table 1 does not read well. % and number of whole sample should be closer together. Provide total numbers of males and females in the first line.

8) Authors have changed the Results section and they provided more examples to support their findings. They have provided new themes and sub-themes, but a few issues remain.

9) For example, the sub-theme ‘Initiation of alcohol and tobacco’ is only included in the friend section, however examples provided in sub-theme ‘family members of similar age’ of family section also supports the sub-theme ‘initiation of alcohol and tobacco’.

10) In line 352, authors state ‘Those respondents (n=8) who reported not using alcohol and/or tobacco still had smoking and/or drinking friends’. Not clear which group of participants are they referring to based on Table 1. Maybe the authors should state ‘Some of the respondents (n=8)…’

11) In the sub-theme ‘lack of parental support’ authors mention ‘Some respondents attributed their alcohol and tobacco use to not getting along with parents, feelings of neglect, grief and coping with household issues. Those neglected by their parents attributed their alcohol and tobacco use to the stress of contributing towards the household income’. However, the examples provided (lines 492-497) does not support such findings.

12) Sub-theme ‘resistance to alcohol and tobacco’. The quotations provided to support this sub-theme does not suggest resistance. Authors should consider a different title for this sub-theme.

13) Discussion about attributing alcohol and tobacco use to coping with household issues such as not being able to get along with parents, the stress of contributing towards the household income, feelings of neglect by parents and lack of parental support are not supported by results section.

14) Limitations of the study should discuss the different methods of data collection.

**********

7. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Dr Loren Kock

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email PLOS at figures@plos.org. Please note that Supporting Information files do not need this step.

29 Sep 2020

Reviewer #1: The authors have responded well to the reviewer comments and made appropriate changes. If the editors will allow it, some more space to allow the authors to provide their justification for the WhatsApp interview technique will strengthen the methods section.

We thank the reviewer for the constructive feedback. We included a justification for the WhatsApp interview technique in the methods section under the sub-heading “data collection” on page 8.

Reviewer #2: The authors have been mostly responsive to the issues raised during the previous review. The manuscript has been improved, but a few comments remain:

We thank the reviewer for the comments and used them to further improve the quality of the paper.

1) Introduction, lines (44-45) ‘Alcohol and tobacco use, like in many other countries is prevalent among adolescents’. Not clear if authors are referring to LMICs or South Africa.

We are refereeing to LMIC’s. This clarification has been made in line 44

2) Authors provide a definition of adolescents, which defines adolescents as those aged 10-19 years old. Not clear why then they include participants 20 years old?

In the South African education system, repetition rates are known to be high from Grade 9 up to Grade 11 resulting in students’ older students attending high school. We therefore feel it is appropriate to include participants 20 years old. We have included this detail in line 105.

3) Authors state that they added this statement in the introduction ‘National studies show that alcohol and tobacco are the most prevalent among adolescents compared to other addictive behaviours such as illegal and other drug use’. But I can’t find it in the Introduction. Could they please indicate the lines?

We thank the reviewer for pointing this out and added this statement to lines 107-108.

4) In the methods section authors state (lines 148-151) ‘Initially, eight seeds were obtained, and they were required to identify up to two other OSY. These respondents recruited by the seeds formed the “first wave” of sampling and were themselves asked to identify and refer a further two more school dropouts (Fig 1). Up to two waves of recruitment were conducted. participants describe the method of recruitment’. Based on this description, the target should have been 56 participants and not 52 as stated in line 170.

We thank the reviewer for pointing this out and corrected this to 56

5) Authors should also report the initial focus of the paper as mentioned in their response, in order to justify why they targeted only smokers and non-smokers.

We have included this statement in lines 146-151.

6) Regarding my comment on ‘How many interviews conducted in person and how many through whatsapp? The authors could perhaps also reflect on whether any differences in quality were observed across those interviewed via the different modalities’. The authors provided the requested numbers, but they did not include any reflection in the manuscript.

We thank the reviewer for the constructive feedback. Despite taking longer and producing fewer words in the online condition, data quality was unaffected by the mode of data collection (online versus face-to-face) with no differences in the number, depth and type of themes discussed

In lines 158-173, we included a justification for the WhatsApp interview technique as well as a reflection on the quality of the different interviewing methods.

7) Table 1 does not read well. % and number of whole sample should be closer together. Provide total numbers of males and females in the first line.

We have made the recommended adjustments and trust that the table reads better

8) Authors have changed the Results section and they provided more examples to support their findings. They have provided new themes and sub-themes, but a few issues remain.

9) For example, the sub-theme ‘Initiation of alcohol and tobacco’ is only included in the friend section, however examples provided in sub-theme ‘family members of similar age’ of family section also supports the sub-theme ‘initiation of alcohol and tobacco’.

We acknowledge that sections of data may be included in multiple themes with some overlap between themes. However, the researchers considered how each theme fit into the overall story about the entire data set. Although there were some instances on initiation with family, we feel that this quote also supports the theme of “family members of similar age”. We have included this statement in lines 543. We trust that this decision still allows for the paper to read well.

10) In line 352, authors state ‘Those respondents (n=8) who reported not using alcohol and/or tobacco still had smoking and/or drinking friends’. Not clear which group of participants are they referring to based on Table 1. Maybe the authors should state ‘Some of the respondents (n=8)…’

We have made the adjustment in the manuscript.

11) In the sub-theme ‘lack of parental support’ authors mention ‘Some respondents attributed their alcohol and tobacco use to not getting along with parents, feelings of neglect, grief and coping with household issues. Those neglected by their parents attributed their alcohol and tobacco use to the stress of contributing towards the household income’. However, the examples provided (lines 492-497) does not support such findings.

We have included an additional quote where respondents attributed their alcohol and tobacco use to not getting along with parents and coping with household issues. We feel that the other quotations capture parental neglect, grief and the stress of contributing towards the household income. We trust that the presentation of quotes support the findings.

12) Sub-theme ‘resistance to alcohol and tobacco’. The quotations provided to support this sub-theme does not suggest resistance. Authors should consider a different title for this sub-theme.

We have changed this subtheme to reluctance to alcohol and tobacco

13) Discussion about attributing alcohol and tobacco use to coping with household issues such as not being able to get along with parents, the stress of contributing towards the household income, feelings of neglect by parents and lack of parental support are not supported by results section.

We hope that the revised presentation of quotes for this theme on page 20 now support the discussion points.

14) Limitations of the study should discuss the different methods of data collection.

We thank the reviewer for this suggestion. The different interviewer methods may have influenced participant responses. However similar to the study by Shapka et al (2016), the findings of this study suggest that data quality is unaffected by the mode of data collection. We have added this statement in lines 541.

1 Oct 2020

Social network determinants of alcohol and tobacco use: a qualitative study among out of school youth in South Africa

PONE-D-20-17565R2

Dear Dr. Desai,

We’re pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it meets all outstanding technical requirements.

Within one week, you’ll receive an e-mail detailing the required amendments. When these have been addressed, you’ll receive a formal acceptance letter and your manuscript will be scheduled for publication.

An invoice for payment will follow shortly after the formal acceptance. To ensure an efficient process, please log into Editorial Manager at http://www.editorialmanager.com/pone/, click the 'Update My Information' link at the top of the page, and double check that your user information is up-to-date. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to help maximize its impact. If they’ll be preparing press materials, please inform our press team as soon as possible -- no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

Kind regards,

Lion Shahab, MA MSc MSc PhD CPsychol

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

9 Oct 2020

PONE-D-20-17565R2

Social network determinants of alcohol and tobacco use: a qualitative study among out of school youth in South Africa

Dear Dr. Desai:

I'm pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please let them know about your upcoming paper now to help maximize its impact. If they'll be preparing press materials, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

If we can help with anything else, please email us at plosone@plos.org.

Thank you for submitting your work to PLOS ONE and supporting open access.

Kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Dr. Lion Shahab

Academic Editor

PLOS ONE