Perceiving is Believing:
Exploring Young People’s Perceptions of Interventions Strategies to Reduce Alcohol-related Social Media Posts
Tu Le 10353682
Master Thesis
Graduate School of Communication
Research Master’s programme Communication Science
Supervisor: dr. H. (Hanneke) Hendriks Date of Completion: 31 January 2020
Abstract
Alcohol use is associated with harmful outcomes, such as brain damage, vandalism, traffic accidents, and future alcohol-related problems. One factor that influences alcohol use is interpersonal communication that nowadays often takes place on social media. Especially young people use social media daily and often share posts with explicit references to alcohol use. Evidence shows that both sharing and seeing alcohol-related social media posts are linked to increased alcohol use, binge drinking, and alcohol-related problems. A mixed-methods study examined whether young people perceive alcohol-related social media posts as a problem (RQ1) and explored how they evaluated eight intervention strategies in terms of effectiveness to reduce the number of such posts and their negative effects on alcohol use (RQ2). Focus groups (Study I, N = 27) and surveys (Study II, N = 282) showed that (1) young people do not truly perceive alcohol-related posts as a real problem, except for posts
displaying drunk people, and (2) young people perceived an automated warning as the most effective intervention strategy. A warning that appears when one intends to share a post that is detected as an alcohol-related post could cause young people to post less. This strategy was perceived as direct, not forceful and non-judgmental, as it mainly focused on alcohol-related posts instead of young people themselves. Additionally, (3) young people perceived a regret-induced strategy as the least effective strategy. This strategy was felt as unnecessary and not memorable. The findings emphasise the relevance of increasing young people’s problem awareness. Future research should focus on strategies to increase problem awareness and testing interventions with automated warnings to reduce young people’s alcohol-related social media posts.
Keywords: alcohol-related posts, social media, mixed-methods design, young adults, adolescents, alcohol use
Exploring Young People’s Perceptions of Interventions Strategies to Reduce Alcohol-related Social Media Posts
Although alcohol use has decreased among young people (i.e., adolescents and young adults), many young people still engage in frequent and excessive use (Stevens et al., 2018). In the Netherlands, young people consume on average more alcohol compared with young people in other European countries (Van Laar & Van Ooyen-Houben, 2016), as 68.6% of the Dutch adolescents aged 16 to 20 consume alcohol (Centraal Bureau voor de Statistiek [CBS], 2019b). Young adults engage in even more frequent use of alcohol than adolescents, with reports showing that 85.4% of the Dutch young adults aged 20 to 30 consume alcohol (CBS, 2019b). These percentages are alarming, as alcohol consumption is associated with harmful outcomes, such as brain damage, liver diseases, vandalism, traffic accidents, and death (Hughes, Anderson, Morleo, & Bellis, 2008; Rehm, Samokhvalov, & Shield, 2013; Ventura-Cots, Ballester-Ferré, Ravi, & Bataller, 2019; National Institute on Alcohol Abuse and Alcoholism, 2017).
Furthermore, reports showed that among the Dutch young people who consume alcohol, 7.7% of the adolescents and 12.5% of the young adults are heavy drinkers who consume more than 14 (women) or 21 (men) drinks per week (CBS, 2019b). Also, a fifth are binge drinkers who consume five or more drinks per occasion (i.e., 19.4% of the adolescents and 19.3% of the young adults, CBS, 2019b). Although this seems a small group, binge drinking has increased in the last few decades. More young people reported engaging in this behaviour in the previous month, from 64% in 2003 compared with 70% in 2017 (Stevens et al., 2018). These numbers are concerning, as alcohol use at a young age may be an indicator for alcohol-related problems later in life (see for an overview Verdurmen et al., 2006; Boelema, Ter Bogt, Van den Eijnden, & Verdurmen, 2009).
One factor that influences alcohol use is interpersonal communication. That is, discussions about alcohol, which nowadays often takes place online on social media (Hendriks, Gebhardt, & Van den Putte, 2017), have been shown to affect attitudes and intentions towards alcohol use (Hendriks, Van den Putte, De Bruijn, & De Vreese, 2014). In the Netherlands, almost all young people (95.2%) report being active on social media platforms, spending an average of 92 minutes per day on these platforms (CBS, 2019a; Anderson & Jiang, 2018; Waterloo, Wennekers, & Wiegman, 2019). Especially social media platforms such as Facebook, Instagram and Snapchat are the most popular among young people (Van der Veer, Boekee, & Hoekstra, 2019; Anderson & Jiang, 2018), where they can share and create content about anything (Muntinga, Moorman, & Smit, 2011), including alcohol.
Studies demonstrated that young people post content often showing explicit alcohol use on social media (e.g., Moreno et al., 2009; Hendriks, Van den Putte, Gebhardt, & Moreno, 2018; Hendriks, Van den Putte, & Gebhardt, 2018; Moreno et al., 2010) and are also regularly exposed to alcohol-related content (e.g., see for an overview Jernigan, Landon, Thornton, & Lobstein, 2017). For example, content analyses of Facebook profiles highlighted a significant display of alcohol-related posts in 85% and 96% of the profiles (Egan & Moreno, 2011; Beullens & Schepers, 2013, respectively). These posts that depict explicit use of alcohol range from moderate to extreme posts. Moderate posts include photos in which alcoholic beverages are in the back- or foreground (e.g., at a dinner or a close-up of a beer), whereas extreme posts consist of photos with intoxicated people or drinking games (Hendriks et al., 2017; 2018).
A number of studies have established that alcohol-related posts are associated with increased risk of alcohol use (e.g., Boyle, LaBrie, Froidevaux, & Witkovic, 2016; Curtis et al., 2018; Critchlow, MacKintosh, Hooper, Thomas, & Vohra, 2019). On the one hand, studies revealed that posting is positively related to binge drinking (e.g., Geusens & Beullens, 2017).
When young people posted more alcohol-related posts on social media, their alcohol use increased six months later (Boyle et al., 2016). On the other hand, recent studies found that not only posting, but also seeing alcohol-related posts can lead to increased alcohol use (Geusens, Bigman-Galimore, & Beullens, 2020; Hendriks, Gebhardt, Van den Putte, & De Nooy, 2019). These alcohol-related posts on social media may be conceived as accurate behaviour by young people, which could lead to incorrect perceptions and may foster that alcohol use is normal (Moreno et al., 2009; 2010).
Considering the negative influence of alcohol-related posts on alcohol use, reducing the number of such posts is critical. Despite extensive research on the effects of alcohol-related posts, until now, no intervention studies have been conducted to reduce these posts (see Moreno et al., 2009 for an exception). Investigating potential intervention strategies can help researchers and healthcare practitioners in the development of effective interventions, and ultimately, reducing the undesirable effects of alcohol-related posts on alcohol use. Therefore, this study extends past research by investigating (1) the extent to which young people perceive alcohol-related posts as a problem, (2) how young people evaluated eight intervention strategies in terms of effectiveness to reduce the number of such posts and their negative effects on alcohol. By using a mixed-methods design, in which data are collected through focus groups (Study I) and surveys (Study II), more in-depth insights and nuances of such posts and potential interventions for young people can be revealed.
Theoretical framework Alcohol-related posts on social media
The rise of social media has enabled people to connect by sharing and creating information, opinions and experiences about anything (Muntinga, Moorman, & Smit, 2011), including alcohol. Research has shown that alcohol is a recurring theme on social media among young people (e.g., Jernigan et al., 2016). Young people often display posts about
alcohol on their social media profiles (varying from 85% to 96% of profiles, Egan & Moreno, 2011; Beullens & Schepers, 2013; Moreno, D’Angelo, & Whitehill, 2016).However, young people seem to be unaware of sharing alcohol-related posts as such posts are often shared without the explicit intention to show alcohol (Hendriks et al., 2017).
Hendriks et al. (2018) identified four types of alcohol-related posts shared on young people’s social media profiles, namely (1) an alcohol-in-background post, (2) an alcohol focus post, (3) a drunken post, and (4) a drinking-game post. Moderate alcohol-related posts refer to posts that capture alcohol coincidentally in the image, such as a photo of a dinner with alcoholic beverages on the table (alcohol-in-background post), or intentionally in the
foreground as the focal point, such as a close-up photo of a beer on a terrace (alcohol focus post, Hendriks et al., 2018). Extreme alcohol-related posts include someone who is drunk or intoxicated (drunken post) or display drinking games, such as a post showing a game of beer pong (drinking-game post, Hendriks et al., 2018).
The influence of alcohol-related posts on alcohol use
Several studies have established the link between exposure to alcohol-related social media posts and alcohol use (e.g., Boyle et al., 2016, see Curtis et al., 2018). On the one hand, posting alcohol-related posts affects future drinking behaviour, including heavy drinking (e.g., Moreno et al., 2016; Critchlow et al., 2019). Posting about alcohol use was both linked to increased drinking behaviour (Moreno, Cox, Young, & Haaland, 2015) and alcohol abuse (Geusens & Beullens, 2017).
On the other hand, seeing alcohol-related posts also contributes to alcohol use (e.g. Jernigan et al., 2016; Geusens et al., 2020). Boyle et al. (2016) showed that seeing alcohol-related posts of peers was alcohol-related to increased alcohol consumption six months later. Furthermore, Hendriks et al. (2019) found that seeing an alcohol-related post increased the frequency and quantity of alcohol use. It can thus be argued that not only posting but also
seeing alcohol-related posts is associated with increased alcohol use and binge drinking, more frequent alcohol use and alcohol-related problems (Hoffman, Pinkleton, Weintraub Austin, & Reyes-Velázquez, 2014; Curtis et al., 2018; Davis et al., 2019).
The influence of alcohol-related posts on alcohol use of young people can be explained by theories of social norms and social learning. Social Norms Theory (Perkins & Berkowitz, 1986) posits that behaviour is based on both perceived (1) descriptive norms that are perceptions of actual behaviours of other people, and (2) injunctive norms that are perceptions of behaviours other people approve or disapprove (Chung & Rimal, 2016;
Cialdini, Reno, & Kallgren, 1991). Similarly, Social Learning Theory (Bandura, 1977) posits that behaviour is learned from observing others and modelling their behaviour.
Supporting these theories, young people may perceive alcohol-related posts as credible sources of information (Moreno et al., 2009). Studies showed that young people view posts about alcohol as an indicator of actual alcohol use (Moreno et al., 2010; Moreno, Grant, Kacvinsky, Egan, & Fleming, 2012), which might lead to incorrect perceptions of how much others drink (i.e., pluralistic ignorance). Moreover, alcohol-related posts are usually displayed in a positive and social context and posted for entertainment reasons (Hendriks et al., 2018; Beullens & Schepers, 2013), which can be misperceived as approval of alcohol use. By displaying alcohol frequently, and in a positive context, these positive posts may influence norms, such that alcohol use is perceived as normal and positive, and thus, encourage frequent and excessive drinking behaviour (Moreno et al., 2009; Hendriks et al., 2018).
Problem awareness of alcohol-related posts
Effective interventions are needed that reduce the number of alcohol-related posts, and ultimately, decrease excessive alcohol use among young people. Although researchers point out that alcohol-related posts are problematic, it remains unclear whether young people similarly perceive these posts. As mentioned earlier, young adults indicate that they do not
intentionally share alcohol-related posts (Hendriks et al., 2017). Hence, it might be
imaginable that they are also not aware of the negative influence. This awareness might be of importance when aiming to develop effective interventions.
Without problem awareness, interventions for behaviour change could be less effective. That is, several theoretical perspectives (e.g., Integrative Model of Behaviour, Protection Motivation Theory, Health Belief Model) suggest that not only attitudes, but also the applicability of the problem to oneself is essential to induce behaviour change. People may be likely to form intentions or engage in behaviours linked to the extent of the likelihood, and similarly, the importance of the problem. The more problematic one thinks
alcohol-related posts are, the more likely one is to change their behaviour to avoid the problem (i.e., perceived risk, Fishbein, 2000; perceived threat, Hochbaum, Rosenstock, & Kegels, 1952; Rogers, 1975; susceptibility and severity, Witte & Allen, 2000). For example, Yan and Brocksen (2013) found that among adolescents who are more aware of the problematic behaviour of alcohol use, the likelihood of that behaviour was lowered. Investigating whether young people have concerns about alcohol-related posts is, therefore, an important first step to the effectiveness of interventions, leading to the following first research question:
RQ1: To what extent is problem awareness of alcohol-related posts present among young people?
Interventions strategies to reduce alcohol-related posts
In addition to examining young people’s problem awareness, understanding how they perceive intervention strategies regarding alcohol-related posts is also essential. Dillard, Shen and Vail (2007) have studied that perceived effectiveness influences actual effectiveness. Interventions that are seen as effective can induce a greater change in intentions and
behaviour than interventions that are perceived as less effective (Fishbein & Cappella, 2006). On this basis, the effectiveness of interventions depends on the extent to which interventions
will be favourably evaluated by young people (Dillard & Ye, 2008).The question then becomes what intervention approach is likely to be perceived as effective, and, therefore, an important second step is to explore how young people evaluate potential intervention strategies regarding perceived effectiveness.
To date, almost no studies have focused on developing interventions for alcohol-related posts. As the sole exception, Moreno et al. (2009) investigated how a single email from a physician targeted at profiles with alcohol-related posts could reduce the number of these posts. However, they concluded that this intervention was ineffective because there was no difference in the number of alcohol-related posts on the profiles that received the email compared with the profiles that did not receive the email (Moreno et al., 2009). As a possible explanation, Moreno et al. (2009) propose that adolescents might not have appreciated the instruction that was communicated, as it could be perceived as unsolicited advice about the risks of disclosing alcohol-related posts. From the perspective of the Theory of Psychological Reactance (Brehm, 1966), when the email presents the recommended advice as something that should be agreed with and would despise anyone who disagrees with it, people might experience a negative emotion (i.e. reactance) and perceive an increased threat to their freedom. Perceiving reactance can then serve as a motivation, for example, to avoid the message or to engage even more in the current, undesired behaviour (Brehm, 1966; Jenkins & Dragojevic, 2011). This implies that interventions could backfire if these are perceived as not pleasant or forceful. Collaborating with young people is thus needed in the design of
interventions, to understand why and when strategies are perceived as effective in addressing young people (Moreno et al., 2016).
Therefore, this study proposes and tests eight intervention strategies based on five potential intervention approaches, which will be explained later, namely: (1) problem
gain a better understanding of young people’s evaluations of these strategies, the following research questions are proposed:
RQ2a: What do young people perceive to be the most and the least effective intervention strategy?
RQ2b: Why do young people prefer some intervention strategies? And under what conditions? (Only Study I)
In the remainder of this section, a theoretical rationale is provided for each of the five approaches from which eight intervention ideas will be proposed.
Approach 1: Problem awareness. Problem awareness can be considered not only as an important factor for intervention effectiveness, but also as a potential approach to
interventions. According to Rogers (1975; 1983), people are motivated to protect themselves against a problem or threat when they perceive it as severe and real. His Protection
Motivation Theory posits that motivations (i.e., threat and coping appraisals) influence behaviour. It could explain which factors may persuade young people to engage in less frequent posting behaviour to ensure their health. Alternatively, the Theory of Reasoned Action (Fishbein & Ajzen, 1972) posits that behaviour is determined by personal beliefs and the information one has about the behaviour. If young people are knowledgeable about the harmful effects of alcohol-related posts, this may decrease positive beliefs and intentions towards sharing such posts (Fishbein & Ajzen, 2011). Thus, increasing problem awareness could be a potential intervention strategy, for example, by demonstrating that alcohol-related posts can be a problem (idea: alcohol post problem).
Approach 2: Regret. Recent research into the motivations of not posting alcohol-related posts has shown that regret can be of importance in preventing this posting behaviour (Hendriks et al., 2017). Participants reported that they might regret sharing alcohol-related posts, and thus, do not engage in this behaviour (Hendriks et al., 2017). Regret is experienced
when one realises that the present situation could have been better had one acted differently (Sandberg & Conner, 2008). Consequently, it can be assumed that regret is an emotion that one prefers to avoid (Sandberg & Conner, 2008), which could affect the intention to engage in certain behaviour.
According to the Theory of Planned Behaviour (Ajzen, 1991) influences on intentions and behaviours are mediated through factors such as attitude, subjective norms, and perceived behavioural control. However, Sandberg and Conner (2012) proposed regret as an additional factor, as it has a direct impact on behavioural intentions. A study into smoking initiation demonstrated that more regret about smoking was correlated to stronger intentions not to smoke (Conner, Sandberg, McMillan, & Higgins, 2010). Following this study, regret should negatively influence the posting behaviour of alcohol-related posts. Therefore, a strategy that emphasises regret after sharing alcohol-related posts (idea: regret alcohol posts) could lower the number of posts.
Approach 3: Socialisation and modelling. As argued, social norms and social learning theories (Perkins & Berkowitz, 1986; Bandura, 1977) may explain why alcohol-related posts lead to undesirable effects (i.e., frequent and excessive alcohol use). It is possible, however, to use these theories to trigger desirable behaviour by showing the
opposite, but recommended behaviour. Hence, following both theories, a reasonable strategy is to depict alcohol use negatively. When others express disapproval of alcohol use or alcohol-related posts on social media, young people may become more likely to model these
behaviours. Friends or social media influencers (i.e., opinion leaders who can influence large audiences following them on social media, Hendriks, Wilmsen, Van Dalen, & Gebhardt, 2019; Boerman, 2020) could share messages, in which they express negatively about alcohol or alcohol-related posts.
Another potential strategy is adjusting perceived norms, which was suggested by Davis et al. (2019), as young people often overestimate their peers drinking behaviour. Davis et al. (2019) highlighted the development of social media interventions to correct
misperceptions by providing accurate normative information. By providing the actual
drinking behaviour of peers, young people could learn how others actually act. Thus, it can be expected from both ideas that perceptions may be adjusted through modelling and
socialisation. This results in a strategy concerning messages from influencers or peers (i.e., idea: popular young people) and a strategy based on accurate norms (i.e., idea: correcting misperceived norms).
Approach 4: Identity. Another approach concerns identity on social media.
According to the Social Identity Theory (Tajfel & Turner, 1979), people have multiple social identities that are linked to a broader social group and demonstrate their social identity by aligning attitudes and behaviours with the perceived social group norms. Regarding alcohol-related posts, Geusens and Beullens (2017) found that what young people shared was strongly related to what their friends shared. Pegg, O’Donnell, Lala and Barber (2018) found that the association between exposure to alcohol-related posts and alcohol use is stronger for young people who strongly identify with their online social network (i.e., strong online social identity). When they strongly identify themselves with online social networks that share alcohol-related posts, they are likely to conform to the groups’ displayed identity by also sharing such posts.
Building upon this, young people’s online social networks consists of multiple audiences (e.g., friends, family members, strangers), which all may have different
expectations about one’s displayed identity (Binder, Howes, & Sutcliffe, 2009; i.e., multiple audience problem, Leary, 1995). Thus, posting alcohol-related posts may not be perceived as acceptable to everyone in young people’s networks. Hendriks et al. (2017) found two
identity-related motives for not posting such posts, namely the fear that parents and potential future employers could see their alcohol-related posts. These motives could imply that as one’s displayed, and thus, perceived identity is not in line with the expectations of the different audiences, this may cause negative impressions (Binder et al., 2009). By incorporating these motives in an intervention strategy, young people may be aware of being selective about what they share about themselves and controlling how they are perceived by others (Leary, 1995). As a result, young people may stop posting as a precaution.
Another possibility is that young people reflect on their online identity, for instance, by asking questions about their profile content (Moreno et al., 2012). LaBrie, Pedersen, Neighbors, and Hummer (2008) found that reflection predicted less alcohol use, as inward thoughts and feelings about one’s own desires, needs and ideas that become more salient may undermine to conform to social group norms (i.e., private self-consciousness). For example, by asking to count the number of alcohol-related posts, this may induce reflection, and in turn, result in the realisation that they share too many posts, as they may not desire to appear as someone who engages in excessive drinking. Therefore, a strategy concerning parents and future employers (i.e., idea: perceived identity of alcohol posts) and the number of shared posts (i.e., idea: counting alcohol posts) will be explored.
Approach 5: Media literacy. One more approach is media literacy (Thompson & Romo, 2016; Hoffman et al., 2014). In the context of this study, media literacy refers to critically thinking about media that includes having “the knowledge and skills to analyse, evaluate, produce, and participate in social media” (Thoman & Jolls, 2003 in Tamplin, McLean, & Paxton, 2018, p. 30). It may serve as a moderator between exposure to alcohol-related posts and alcohol use, as young adults with higher media literacy were found to consume less alcohol (Thompson & Romo, 2016; Chang et al., 2016). One potential strategy is a juxtaposition by showing alcohol-related posts (e.g., a dinner with alcoholic beverages)
next to what actually happened in certain situations (e.g., fights or hangovers afterwards). This could make young people conscious about the influences of such posts.
Another strategy could be a real-time or automated intervention that automatically detects alcohol-related posts and provides a warning when one intends to upload such a post. When a warning informs young people about the alcoholic content in the post, this may trigger critical evaluations about the post to be shared, for example, to use social media differently by not sharing the post. As it aims to make people mindful of their behaviour, warnings enhance people to critically think about their intention to post. This differs from the idea of counting alcohol-related posts that is more focused on self-reflection of what already has been posted, which relates to identity rather than media literacy. Thus, a strategy with a juxtaposition (i.e., idea: alcohol posts unrealistic) and a warning will be tested (i.e., idea: warning alcohol posts).
Method Design
The current study used a mixed-methods design with focus groups and surveys to gain a fuller understanding of how alcohol-related posts and intervention strategies are viewed by young people. While focus groups provide participants time for thoughts, stimulate in-depth communication, and might reveal patterns related to group dynamics as people talked to each other instead of to a researcher, surveys can gather data from a larger representative sample and complement the qualitative data from a smaller subgroup (Braun & Clarke, 2013). The university's ethics committee approved this study (ERP 2019-PC-11419).
Study I: Focus groups Sample
Purposeful sampling was used, a qualitative sampling method, to select participants based on two criteria suitable to the research topic (Braun & Clarke, 2013): two age
categories (adolescents and young adults) and three educational levels (havo, vwo1,
university). First, young people aged 16 years and older in high school and university were invited, because these groups may differ in alcohol use and motivations to post alcohol-related posts (CBS, 2019b; Stevens et al., 2018; Hendriks et al., 2017). Second, alcohol use is also related to educational level, as young people enrolled in pre-university education
consume less than those enrolled in general secondary education (Rijksinstituut voor
Volksgezondheid en Milieu, 2017; Stevens et al., 2018). Hence, adolescents and young adults participated in focus groups related to their educational level and age.
In total, seven focus groups sessions were conducted (N = 29, Appendix C). For the sample regarding adolescents in high school, three focus groups with six participants each were conducted (N = 18). However, two participants were excluded from the analysis, as they were younger than 16 years, resulting in a sample of sixteen participants (10 women, 6 men, Mage = 16.75 years, rangeage = 16-19 years). 62.5% of the participants followed general
secondary education and 37.5% pre-university education. For the sample regarding university students, four focus groups with two to three participants each (N = 11) were conducted (5 women, 6 men, Mage = 22.60 years, rangeage = 19-28). Six gift vouchers were raffled among
participants (€10).
Data collection and Procedure
The author conducted the focus groups in November and December 2019. High school participants were recruited with permission from a high school teacher in the North-eastern region of the Netherlands. University students were recruited online on Facebook, via the moderator’s personal network and face-to-face at university canteens. The sessions took place either at high school (Hoogeveen) or university (Amsterdam) in a classroom and lasted
1 Havo and vwo are the Dutch equivalent of general secondary education and pre-university education,
between 40 to 60 minutes. The focus groups were held in Dutch, audio-recorded and transcribed (Appendix D).
The moderator used an interview guide for conducting focus groups (Appendix A). This guide started with an opening question by which participants introduced themselves and reported their most used social medium. Participants were then introduced to alcohol-related posts, asking whether they posted or saw such posts from friends on social media. After these questions, problem awareness and related evaluations of alcohol-related posts were discussed with the questions “What do you think of seeing alcohol posts?” and “To what extent do you think these posts are a problem?” Finally, two open-ended questions, a ranking exercise and a rounding up question had to gauge participants’ ideas for and evaluations of intervention strategies.
In the ranking exercise, participants were first asked to rank eight different intervention ideas from best to worse strategy in dyads to elicit evaluations for a joint
discussion. After a joint discussion about what participants liked and disliked about each idea, the whole group was asked to jointly rank the three most effective ideas. Finally, participants were asked whether they had additional ideas that were not mentioned yet. External stimuli were created for the ranking exercise (Table 1) and for illustrating alcohol-related posts (adopted from Hendriks et al., 2018, Appendix B) to elicit ideas and perspectives of the participants (Greenbaum, 2000) and to facilitate comparisons across groups, as participants were on the same page and reminded of the topic (Bloor et al., 2001).
Table 1
Example intervention idea2 Description Idea 1: Alcohol posts problem
(Approach: problem awareness)
Create awareness among young people that alcohol posts can be problematic. For example, by providing facts and figures (“research shows that seeing one alcohol post increases the odds of drinking alcohol by 15%”).
Idea 2: Alcohol posts unrealistic (Approach: media literacy)
Show young people that alcohol posts are not an accurate representation of reality. These posts are often too positive and do not show the negative sides of alcohol. For example, a juxtaposition in which the first photo depicts “what you think happened” (laughing individuals with beers) and the second photo “what happened after” (drunk individuals/vomiting).
Idea 3: (Too) many alcohol posts (Approach: identity)
Let young people come to an understanding that they post (too) many alcohol posts. For instance, by letting them scroll through their timelines and count the number of alcohol posts.
Idea 4: Warning alcohol posts (Approach: media literacy)
Provide an automatic warning when young people intend to upload an alcohol post. This could be implemented with machine
learning (to recognise alcohol on images in an automated way). When you want to upload an alcohol post, you will get a message “You are about to upload a post that includes alcohol. Are you sure that you want to upload this?”
Idea 5: Perceived identity of alcohol posts (Approach: identity)
Point out to young people that alcohol posts are not well perceived by e.g., family
2 Two images were designed for each idea which were identical in content. For the examples depicting
individuals, the second example portrayed individuals of the opposite sex to rule out potential gender differences.
members, parents and future employers. An example could be showing a resume that includes an alcohol post as a profile picture with the statement: “Is this how your future employer should perceive you?”
Idea 6: Regret alcohol posts (Approach: regret)
Emphasise that young people could regret sharing alcohol posts the day after uploading the posts.
Idea 7: Popular young people
(Approach: socialisation and modelling)
Let popular young people (e.g., good friends or influencers) share messages on social media that are negative about alcohol or emphasise not to post alcohol posts.
Idea 8: Correcting misperceived norms (Approach: socialisation and modelling)
Young people unfairly hold the idea that many young people drink much, which could be reinforced by alcohol posts. This idea could be corrected by disseminating messages on social media such as: “Most young people consume 0 to 4 drinks when they go out”.
Analysis
For the analyses, a grounded theory approach (Glaser & Strauss, 1967) was used, which aims to inductively construct theory in an open and iterative process from data (Braun & Clarke, 2013). In the first step of the analysis (open coding), transcripts were read line-by-line and divided into smaller fragments with data analysis software ATLAS.ti. These
fragments were labelled with codes that briefly describe participants’ formulations or
expressed meanings. These fragments were coded without pre-defined codes but linked to the explored topics (i.e., problem awareness or intervention ideas), for example, ‘Idea_5’ would be linked to fragments concerning this idea.
During the second stage (axial coding), possible categorisations between or among the coded fragments were explored (Braun & Clarke, 2013). Codes were also reviewed, and overlapping codes were merged. As it was noted for each code to which topic it was related, patterns and categorisations in the codes could be identified (Appendix F).
The final step was to define the links between topics, categories or codes. For the five intervention approaches, a taxonomy of categories was created that reported clusters of common and related participants’ evaluations that emerged from the data.
Three techniques were used to enhance the validity of the results (Appendix E). First, the immediate transcription of the focus groups and memos written during the research process avoid biases and assure transparency, which enhances the trustworthiness and credibility (Flick, 2014). Second, a thick description was created, which described
participants’ behaviour (i.e., gestures and facial expressions), and a reflection of the focus groups, which enhance transferability and credibility as the steps in the research process can be followed (Van der Goot, 2017). Finally, triangulation enhances credibility and
transferability as data were collected using two methods (Braun & Clarke, 2013, see Study II). Study II: Online surveys
Participants and Procedure
The online questionnaire was held in Dutch and presented in Qualtrics. The data collection took place in November 2019. Respondents were first informed they had to be the age of 16 or 17 (adolescents) or 16 or older (young adults) to participate. After agreeing with the informed consent, respondents were shown a filter question asking for age and a filter
question asking whether they had a Facebook or Instagram account. Underage respondents and respondents without Facebook and Instagram were redirected to the end of the
questionnaire. Hereafter, questions related to alcohol-related posts and the eight intervention strategies were presented, including two attention checks. Upon completion of the
questionnaire, respondents received credits for their participation.
Adolescents. Data were gathered from a survey that was part of a more extensive survey regarding Instagram among 16- and 17-year-olds by market research panel PanelClix. The sample consisted of 117 respondents. However, five respondents were omitted from the analyses, because they failed both attention checks (n = 3) or gave invalid answers (e.g., ‘d’ for every open-ended question, n = 2). This exclusion left a total of 112 respondents with 64 women (57.1%) and 48 men (Mage = 16.63 years, SDage = 0.49). Regarding educational levels,
24.1% attended pre-university education (vwo), 23.2% vocational education (mbo), 22.3% general secondary education (havo), 19.6% professional higher education (hbo), 3.6% university, 2.7% pre-vocational secondary education (vmbo), 0.9% practical training (praktijkonderwijs) and 3.6% did not attend any education at the time.
Young adults. The questionnaire was presented on research platform LAB of the University of Amsterdam. The sample consisted of 188 respondents. However, 18
respondents were excluded as they (a) completed the questionnaire more than once (n = 12), (b) failed both attention check questions (n = 3), or (c) were unable to complete the
questionnaire on the first attempt but succeeded on the second attempt (n = 3). In the case of the latter, six responses were merged into three responses by substituting missing answers of respondents’ first attempt with the answers given on their second attempt. This exclusion left
a total of 170 respondents, 139 women (81.8%), 29 men, and 2 non-binaries (Mage = 20.28 years, SDage = 2.10, rangeage = 17-28). All respondents studied Communication Science3.
Measures
Frequency and occurrence of alcohol-related posts. Descriptions of four types of alcohol-related posts were presented (adopted from Hendriks et al., 2018) to which
respondents indicated whether, and if so, how often they posted such posts on social media. Additionally, they indicated whether and how often they saw such posts from friends (0 = Yes; 1 = No, 1 = Once a year or less to 12 = 10 times a day or more).
Problem awareness. Respondents were asked to which extent they find alcohol posts a problem using a 5-point scale ranging from (1) not a problem at all to (5) a serious problem.
Ranking intervention ideas. Respondents were asked to rank eight intervention ideas from (1) best to (8) worst idea. As some answers were missing, because some participants filled out only the first half of the questionnaire, the scores of 105 (adolescents) and 168 respondents (young adults) were calculated.
Perceived effectiveness. For each intervention idea, the Perceived Effectiveness Scale from Dillard, Shen, and Vail (2007) measured respondents’ perceived effectiveness with six 7-point scale items anchored by not persuasive/persuasive, ineffective/effective, not
convincing/convincing, not compelling/compelling, misleading/straightforward, and
forgettable/memorable (α range = .82-.93). An overall measure for each idea was computed of the mean scores of the six items (M range = 3.71-4.97, SD range = 1.22-1.45). Because of missing answers, the scores were calculated for 99 adolescents and 164 young adults (165 for the first three ideas).
3 When comparing samples, adolescents and young adults did differ in age, t(195.687) = -21.848, p < .001,
95%CI [-3.99;-3.33], education, χ² (1, N = 282) = 265.68, p < .001, and gender, χ² (1, N = 282) = 23.46, p < .001. The sample with young adults is older, higher educated and consisted of more women than men. However, this study focuses on young people in general.
Demographics. Respondents’ current level of education (praktijkonderwijs, vmbo, havo, vwo, mbo, hbo, universiteit, I am not studying right now), gender (0 = male, 1 = female, 2 = other, please specify) and age in years were asked.
Attention checks. Respondents’ attentiveness was measured with two 7-point Likert scale items anchored by strongly disagree/strongly agree. Respondents had to answer the first item with slightly disagree and the second with slightly agree.
Results Frequency and occurrence of alcohol-related posts
Before answering the research questions, the frequency and occurrence of alcohol-related posts of the focus groups participants (Study I) and survey respondents (Study II) are reported first.
Study I. Regarding the focus groups, participants reported that alcohol-related social media posts were common. It was revealed that most young people shared alcohol-related posts: “Oh, I do post photos. Of alcohol. (…) [alcohol post example] 1, 2, and 4” (Gerda). Also, most of them indicated seeing friends’ posts often: “Yeah, I would see [example] 1, 2 and 4, I basically see them weekly. And I see 3 only in private groups” (Francis).
Study II. Regarding the surveys, more young adults (70.6%) than adolescents (29.5%) reported posting alcohol-related posts, χ² (1, N = 282) = 46.01, p < .001. On average, both young adults and adolescents seemed to post once every three months (M = 2.97, SD = 1.41), t(151) = -1.853, p = .066, 95%CI [-1.05;0.03].
Also, a strong presence of alcohol-related posts from friends is reported. Most of the adolescents (79.5%) and all young adults (100%) reported seeing posts from friends, χ² (1, N = 282) = 38.01, p < .001. Young adults reported seeing friends’ posts more often than adolescents, t(257) = -3.129, p = .002, 95%CI [-0.97;-0.22]. On average, adolescents see
friends’ alcohol-related posts 2-4 times per month (M = 4.99, SD = 1.55), young adults 2-4 times per month to 2-6 times per week (M = 5.58, SD = 1.40).
RQ1: Problem awareness of alcohol-related posts
To answer RQ1, the extent to which problem awareness of alcohol-related posts is present among young people was investigated.
Study I. When asked about participants’ problem awareness and related evaluations regarding alcohol-related posts, they were unanimous in the view that moderate posts are “socially accepted”. While these posts were judged positively, the extreme posts, especially the drunken post, were evaluated negatively, as “problematic” and “unacceptable”. This was illustrated by the comments below:
Humberto: The first two [examples] are still good.
Tanya: The other two are mainly showing the problems. Fernand: Dra-ma-tic.
Moderator: What did you say, Jerry?
Jerry: I think the third one is the worst. 4 is quite fun, having a drinking game, in my opinion, could be fun.
(…)
Fernand: Number 1 and 2 are just fine.
Rebekah: And yeah, I do find it more difficult, as soon as there really is someone really drunk who just can’t walk anymore, to make videos of it, yes, that…
Fernand: 4 is also fine but with 3 I am like: So you’re taking a picture instead of helping, that is what I feel about 3.
Moreover, as one participant put it (Atiyah): “Ehm, I, depends on, what you define as a problem of course, because, look [example] 1 and 2, you cannot be fired for that but 3, I
would find it problematic, I just would not put it anywhere”. Thus, most participants indicated that mainly the drunken post could be perceived as a problem.
Study II.
Problem awareness. Young people reported perceiving such posts as not actually as a problem (M = 2.33, SD = 1.07, range = 1-5). Only 2.5% of the young people indicated it as a serious problem. More specifically, young adults seems to have a slightly lower level of problem awareness than adolescents, t(207.497) = 2.071, p = .040, 95%CI [0.01;0.54]. Young adults reported alcohol-related posts as ‘not actually a problem’ (M = 2.22, SD = 0.99), while adolescents reported such posts as ‘not actually a problem’ to being neutral about it (M = 2.50, SD = 1.17).
From a triangulation perspective, the results show a divergence in the data. The findings of the survey demonstrated low levels of problem awareness among young people. However, these results seemed more nuanced when interpreting the focus groups, as moderate posts (e.g., alcohol-in-background and alcohol focus posts) are perceived as acceptable and normal and extreme posts (e.g., drunken and drinking-game posts) as problematic.
RQ2a: Effectiveness of intervention strategies
To answer RQ2a, it was explored which of the eight proposed intervention ideas would be perceived as the most effective intervention strategy to reduce alcohol-related posts.
Study I.
Group ranking. Idea 4 (warning alcohol post) was considered as an effective strategy as it was the only strategy that was listed in all seven group rankings.
Study II.
Ranking intervention ideas. The results of ranking the eight intervention strategy examples show that idea 4 (warning alcohol post), was ranked as best according to young
people (M = 3.44, mode = 1), while idea 6 (regret alcohol posts) was ranked as worst (M = 5.51, mode = 8).
Perceived effectiveness. When looked at the eight intervention strategies separately, young people perceived idea 5 (M = 4.97, range = 1-7) as the most effective strategy, whereas idea 6 (M = 3.71) was perceived as the least effective strategy4.
Taken together, the data showed inconsistent findings as to what strategy is perceived as most effective. Data from focus groups and surveys corroborate the finding that idea 4 (warning alcohol post) was ranked as the best strategy. However, according to the surveys, when perceived effectiveness of the strategies was interpreted, idea 5 (perceived identity) appeared as most effective strategy. Furthermore, the survey findings demonstrated that idea 6 (regret alcohol posts) was the worst strategy.
RQ2b: Evaluations of intervention strategies
To answer RQ2b, in-depth evaluations of the intervention strategies during the joint discussion of the ranking exercise in dyads in the focus groups were reported. Common views (participants’ evaluations) of the strategies that emerged from the data are presented
according to the five intervention approaches. Problem awareness.
Lack of awareness. Most participants reported being unaware of the effects of
alcohol-related posts (idea 1). As Atiyah noted: “(...) young people probably don’t know that they’re influenced by these posts. Like, I was even unaware about it, you just told me that, that research found… Like, I wasn’t thinking about that”. Ciara commented that “problem awareness should be raised first, before conducting any of the other strategies”.
4 Adolescents indicated to perceive both idea 4 and idea 6 more effectively (M
idea4 = 4.98, Midea6 = 4.07) than
Problem not perceived as threat. Most participants seem not to appeal to research findings. Results and statistics were not perceived as inducing threat according to Jerry: “I myself think that 15 percent increased chance is not captivating for most people” and Hugh: “15 percent might not say that much to people”. Instead, it was perceived as forceful “That’s why I said [idea] 1 would not work”. It was suggested that it would be more appealing by “using visuals”, “humour” or converting it into “a fun fact to recall it better”.
Regret.
Unaffected by regret. Participants were mainly negative about idea 6. Some participants argued that regret would be only experienced for a “short time”, while others report that they would regret the negative consequences of drinking rather the shared alcohol-related posts. Furthermore, comments revealed that this idea was perceived as “not
memorable”, “unnecessary” and “ineffective”:
Karen: I was just about to say that it is not necessary to set up a campaign for that. Dorian: If you check what’s posted at 3 in the night, and at 10 in the morning… Moderator: And if you wake up the next morning and check it?
Dorian: Almost nothing. Socialisation and modelling.
Credibility and endorser fit. Idea 7 was evaluated positively. However, it was suggested that it would be effective with a credible endorser. Participants reported that Lil’ Kleine (artist, example male influencer) “seems to drink a lot” and is “not credible”. His messages about reducing alcohol would feel “insincere”. Popular young people with a healthy lifestyle were suggested, such as “Doutzen Kroes” and athletes to “show the negative effects of alcohol on their body”.
Ineffective communicated norm. Idea 8 was perceived as weak as participants argued it would backfire: “And so it seems that most young people drink. Only 4 [drinks], but they do
drink” (Andrea). This idea was also perceived as ineffective when one already complies to the norm, as Atiyah commented that she is “a responsible drinker but still could place alcohol posts”. She further explains: “I think, yes: I usually keep it under four. Or I can’t even drink more than four. Look. What’s the problem, then?”
Identity.
Thinking of the future. Young adults emphasised the importance of future employers, as Hugh explained: “I hear a lot of people saying: Yeah, I want to build up my resumé. I’d like to come across as a bit of a fun person”. For those who are ignorant about the future, on the other hand, it would be ineffective. Iris noted: “(…) I think that someone at uni is likely to think: Oh, right, I need to focus on that, but someone enrolled in vocational education would not think about it”. Idea 5 (i.e., perceived identity) would be useful when someone is actively working on their career or highly educated.
(In)attentive about parents. Opinions differed as to whether stressing parents could be a useful strategy. Some participants felt positive about stressing young people’s identity towards parents, as Ciara revealed: “I teach people. (…), imagine that I have my Facebook publicly visible (…), and I post pictures every weekend of me being totally drunk and
intoxicated. And then, on Tuesday I’m standing in front of a group of eight-year-olds. I don’t think parents would find that enjoyable”. Others reported to be less worried as close others (e.g., friends and family) “know them well” and do not care about alcohol-related posts. Gerda pointed out about her parents: “(…) once if I go to a party, they also know that I go to a party and what I do there. So I don’t care if they see that or not”.
Unmotivated to count. Regarding idea 3 (i.e., counting alcohol posts), participants indicated young people would only scroll through their profile to look at “funny” content instead of counting. This idea was perceived as “requires too much effort” and “most people are too lazy to count”. Only counting is not sufficient, as Jan noted: “There you only
determine how much they post in terms of alcohol, but you don’t specify or so what kind of influence it has”. Interestingly, posts would be removed if it had “not many likes”.
Media literacy.
Susceptibility to negative sides of alcohol. Positive views show that participants experienced idea 2 (i.e., alcohol posts unrealistic) as a reminder that the negative sides of alcohol use exist. It would be perceived as more effective if it were coming from someone they know. Negative views revealed that idea 2 was perceived as a “meme” or “funny”. The depicted negative sides of alcohol were “unrealistic” and “unrepresentative” as most
participants mentioned this would not occur to them.
Disclosures for alcohol posts. Most participants were positive about idea 4 (i.e., warning), as it felt “non-judgmental”, “in your face” and would create awareness “without some effort”. Besides that, participants also named additional strategies. Disclosures such as “18+ content”, “this is an alcohol-related post” or blurring alcohol-related posts for underage people were also perceived to increase awareness in a similar way.
Discussion and Conclusion
The present study was designed to explore whether young people perceive alcohol-related social media posts as a problem (RQ1) and how young people evaluated eight
intervention strategies in terms of effectiveness to reduce the number of alcohol-related social media posts and their negative effects on alcohol use (RQ2). Three main findings can be distinguished based on the data from focus groups and surveys. First, young people do not truly perceive alcohol-related posts as a real problem, except for posts showing drunk people. Second, an intervention strategy that shows a warning before uploading alcohol-related posts on social media was perceived as most effective, because it was seen as a direct, not forceful and non-judgmental strategy. Third, an intervention strategy regarding regretting
alcohol-related posts was perceived as least effective, as it was observed as an unnecessary and non-memorable strategy.
Findings and Implications
First, the results show that young people do not perceive alcohol-related posts as an actual problem, except for the posts showing drunk people, which were reported as
problematic. These results are in line with previous research of Hendriks et al. (2018) that demonstrated that moderate posts were acceptable (i.e., alcohol-in-background and alcohol focus posts), while extreme posts were not (i.e., drunken and drinking-game posts). Despite the fact that alcohol-related posts are common on social media and negatively affect alcohol use, young people do not seem to worry about such posts. Hence, if young people seem to be largely unaware that alcohol-related posts are a real problem, interventions aiming to reduce these posts would be ineffective. This confirms the idea in line with behavioural theories (e.g., Integrative Model of Behaviour, Fishbein, 2000; Protection Motivation Theory, Rogers, 1975; 1983) that not only attitudes, but also the applicability of the problem to oneself is needed to motivate behaviour change. The more problematic young people think alcohol-related posts are to themselves, the more likely they are motivated to change their behaviour to avoid the problem (i.e., threat appraisal, Rogers, 1975).
The question then becomes how problem awareness could be increased among young people. On the one hand, one approach could be highlighting the problems of alcohol-related posts by showing the negative effects of such posts. Such an approach was tested in the current study by showing findings from research that demonstrate the effects of alcohol-related posts (idea 1), but young people evaluated this strategy as forceful and not threat-inducing. On the other hand, as highlighting research findings seem ineffective, another, more indirect approach could be portraying young people who mimic alcohol-related posts and experience the negative effects of such posts. A portrayal could be showing young people
who buy and consume the identical excessive quantity of alcohol that is depicted in alcohol-related posts of other young people. Future studies are needed to investigate this approach or other potential approaches for increasing young people’s problem awareness.
Second, the results revealed that an intervention strategy concerning a warning (idea 4) was collectively perceived as most effective in both studies. That is, by providing an
automatic warning when young people intend to upload an alcohol-related social media post that shows “you are about to share a post in which alcohol is present, are you sure you want to upload that?”. This strategy could be implemented with machine learning, by which alcohol in images is detected in an automated way. With this result, new insights are offered, in
particular, that a media literacy approach can be relevant for reducing social media posts. Media literacy posits that people think critically about media, such that they have the ability and competences to deal with social media (Thoman & Jolls, 2003 in Tamplin et al., 2018). When young people see such a warning, they may be triggered to think about their post critically and re-evaluate their intention to upload that post, as one is made conscious of the alcoholic content in the post. Thus, this result offers a starting point for future work that focuses on testing such a warning strategy and assessing its actual effectiveness. This strategy could be tested with a longitudinal study by which young people interact in an app for a fictitious social media platform and receive notifications when they upload alcohol-related posts.
Additional in-depth insights showed that young people preferred this strategy as it was perceived as a direct and not a forceful approach. Moreover, a warning was experienced as a non-judgmental approach, focusing on the post and not the person who intended uploading the post. These insights are in line with both the study by Moreno et al. (2012), implying that young people are drawn towards intervention strategies that are direct and respectful. These findings suggest that media literacy can be an important factor in inducing behaviour change.
This is in contrast with behavioural theories that posit that behaviour change is established through attitudes, subjective norms, and perceived behavioural control (e.g., Theory of Planned Behaviour, Ajzen, 1991). Thus, an interesting question that asks for follow-up research is whether problem awareness plays a role in behaviour change. As problem awareness was argued to be needed for decreasing positive attitudes and intentions towards posting alcohol-related posts, young people could also get a sense of awareness through a warning without influencing factors (e.g., attitudes, intentions) of the Theory of Planned Behaviour first. Further research into intervention strategies could combine both problem awareness and a warning strategy, examining whether problem awareness could be an
underlying mechanism for behaviour change. For instance, by comparing the effectiveness of warnings focused on either highlighting the negative consequences of uploading alcohol-related posts or, similar to the example provided in this study, making young people aware of the alcohol in the content.
Third, the results demonstrated that an intervention strategy aimed at regretting alcohol-related posts (idea 6) was perceived as the least effective. This strategy emphasises that young people could regret uploading alcohol-related posts the day after uploading the posts. Additional in-depth evaluations revealed that this strategy was observed as unnecessary and non-memorable. Thus, these findings are not in line with the notion that regret should be taken into account as an additional factor that could induce behaviour change (Sandberg & Conner, 2012). As Sandberg and Conner (2012) argued that regret is a negative emotion that one prefers to avoid, it may be possible that this strategy did not induce high levels of regret such that young people would feel motivated to change their behaviour. While this strategy emphasises regret after uploading alcohol-related posts, future research could also examine strategies focused on anticipated regret combined with possible negative outcomes. A broader
strategy could be pointing out that young people might regret as they miss out on things, for example, during a party, while they are busy with uploading alcohol-related posts.
Limitations
Although this study provides important new insights into intervention strategies for alcohol-related posts, it does have some limitations. First, the evaluations of the intervention strategies may be influenced by the external stimuli that were provided to the focus groups. Focus group participants viewed both the descriptions and visual examples for each
intervention strategy, while survey respondents only viewed the descriptions. This limitation was also stressed by one of the focus group participants. He mentioned that although he knew he should not focus on the layout and colours, he liked the idea but not the imagery of it in the external stimuli. Although it was also included in the interview guide as the moderator
emphasised multiple times that participants should focus on the idea and not the way it was presented, it could have influenced the evaluation of the intervention strategies. Future studies could consider whether to provide external stimuli to both focus groups and surveys to rule out potential differences in evaluations.
Second, this study was conducted with young people primarily enrolled in upper secondary education and university. This suggests that the results may not be generalisable to all young people. Although adolescents enrolled in vocational education were invited to participate in focus groups, unfortunately, they did not sign up. Additionally, it was unfortunate that the focus groups with university students did not have the ideal size of a minimum of four to six participants, as there were two to three participants in each focus group. It was challenging to recruit university students as the focus group sessions took place close to the exam period. Due to many cancellations and no-shows, it was decided to proceed with the small size of the groups. Further research is needed to understand whether these findings can be generalised to young people from a broader range of educational levels.
Conclusion
Altogether, this study shows that young people do not truly perceive alcohol-related social media posts as a real problem. Additionally, a strategy with automated warnings was perceived as the most effective strategy to reduce alcohol-related social media posts and their effects on alcohol use. Given that an automated warning appears when someone intends to upload alcohol-related posts on social media, it could increase awareness about the alcoholic content and reduce the intention to post such posts. Future research should focus on strategies to enhance problem awareness and testing interventions with automated warnings.
References
Ajzen, I. (1991). The theory of planned behavior. Organizational Behavior and Human Decision Processes, 50(2), 179-211. https://doi.org/10.1016/0749-5978(91)90020-T Anderson, M., & Jiang, J. (2018). Teens, social media, & technology 2018. Washington, DC:
Pew Research Center.
Bandura, A., & Walters, R. H. (1977). Social learning theory (Vol. 1). Englewood Cliffs, NJ: Prentice-hall.
Beullens, K., & Schepers, A. (2013). Display of alcohol use on Facebook: A content analysis. Cyberpsychology, Behavior, and Social Networking, 16(7), 497-503.
https://doi.org/10.1089/cyber.2013.0044
Binder, J., Howes, A., & Sutcliffe, A. (2009). The problem of conflicting social spheres: Effects of network structure on experienced tension in social network sites. Conference on Human Factors in Computing Systems - Proceedings, 965–974. https://doi.org/10.1145/1518701.1518849
Bloor, M., Frankland, J., Thomas, M., & Robson, K. (2001). Focus groups in social research. London: Sage Publications.
Boelema, S., Ter Bogt, T., van den Eijnden, R., & Verdurmen, J. (2009). Fysieke, functionele en gedragsmatige effecten van alcoholgebruik op de ontwikkeling van 16-18 jarigen. Utrecht: Trimbos Instituut/Universiteit van Utrecht.
Boerman, S. C. (2020). The effects of the standardized instagram disclosure for micro- and meso-influencers. Computers in Human Behavior, 103, 199–207.
https://doi.org/10.1016/j.chb.2019.09.015
Boyle, S. C., LaBrie, J. W., Froidevaux, N. M., & Witkovic, Y. D. (2016). Different digital paths to the keg? How exposure to peers’ alcohol-related social media content
influences drinking among male and female first-year college students. Addictive Behaviors, 57, 21-29. https://doi.org/10.1016/j.addbeh.2016.01.011
Braun, V., & Clarke, V. (2013). Successful qualitative research: A practical guide for beginners. London: Sage Publications.
Brehm, J. (1966). A theory of psychological reactance. New York: Academic Press.
Centraal Bureau voor de Statistiek. (2019a). Internet jongeren; toegang, gebruik, faciliteiten en online aankopen. Retrieved January 11, 2020, from
https://jmopendata.cbs.nl/#/JM/nl/dataset/20255NED/table?dl=1FB93
Centraal Bureau voor de Statistiek. (2019b). Leefstijl en (preventief) gezondheidsonderzoek; persoonskenmerken. Retrieved January 11, 2020, from
https://opendata.cbs.nl/statline/#/CBS/nl/dataset/83021NED/table?fromstatweb Chang, F., Miao, N., Lee, C., Chen, P., Chiu, C., & Lee, S. (2016). The association of media
exposure and media literacy with adolescent alcohol and tobacco use. Journal of Health Psychology, 21(4), 513–525. https://doi.org/10.1177/1359105314530451 Chung, A., & Rimal, R. N. (2016). Social norms: A review. Review of Communication
Research, 4, 1-28. https://doi.org/10.12840/issn.2255-4165.2016.04.01.008 Cialdini, R. B., Kallgren, C. A., & Reno, R. R. (1991). A Focus Theory of Normative
Conduct: A Theoretical Refinement and Reevaluation of the Role of Norms in Human Behavior. Advances in Experimental Social Psychology, 24(C), 201–234.
https://doi.org/10.1016/S0065-2601(08)60330-5
Conner, M., Conner, M., Sandberg, T., McMillan, B., & Higgins, A. (2006). Role of anticipated regret, intentions and intention stability in adolescent smoking initiation. British Journal of Health Psychology, 11(1), 85-101.
Critchlow, N., MacKintosh, A. M., Hooper, L., Thomas, C., & Vohra, J. (2019). Participation with alcohol marketing and user-created promotion on social media, and the
association with higher-risk alcohol consumption and brand identification among adolescents in the UK. Addiction Research & Theory, 1-12.
Curtis, B. L., Lookatch, S. J., Ramo, D. E., McKay, J. R., Feinn, R. S., & Kranzler, H. R. (2018). Meta-Analysis of the Association of Alcohol-Related Social Media Use with Alcohol Consumption and Alcohol-Related Problems in Adolescents and Young Adults. Alcoholism: Clinical and Experimental Research, 42(6), 978–986. https://doi.org/10.1111/acer.13642
Davis, J. P., Pedersen, E. R., Tucker, J. S., Dunbar, M. S., Seelam, R., Shih, R., & D’Amico, E. J. (2019). Long-term Associations Between Substance Use-Related Media
Exposure, Descriptive Norms, and Alcohol Use from Adolescence to Young Adulthood. Journal of Youth and Adolescence, 48(7), 1311–1326.
https://doi.org/10.1007/s10964-019-01024-z
Dillard, J. P., Shen, L., & Vail, R. G. (2007). Does Perceived Message Effectiveness Cause Persuasion or Vice Versa? 17 Consistent Answers. Human Communication Research, 33(4), 467–488. https://doi.org/10.1111/j.1468-2958.2007.00308.x
Dillard, J. P., & Ye, S. (2008). The Perceived Effectiveness of Persuasive Messages:
Questions of Structure, Referent, and Bias. Journal of Health Communication, 13(2), 149–168. https://doi.org/10.1080/10810730701854060
Egan, K. G., & Moreno, M. A. (2011). Alcohol references on undergraduate males’ Facebook profiles. American Journal of Men’s Health, 5(5), 413–420.
https://doi.org/10.1177/1557988310394341
Flick, U. (2014). An Introduction To Qualitative Research (5th ed.). Los Angeles, London, New Delhi, Singapore, Washington DC: SAGE.
Fishbein, M. (2000). The role of theory in HIV prevention. AIDS Care - Psychological and Socio-Medical Aspects of AIDS/HIV, 12(3), 273–278.
https://doi.org/10.1080/09540120050042918
Fishbein, M., & Ajzen, I. (1972). Attitudes and opinions. Annual review of psychology, 23(1), 487-544.
Fishbein, M., & Ajzen, I. (2011). Predicting and changing behavior: The reasoned action approach. Psychology press.
Fishbein, M., & Cappella, J. N. (2006). The Role of Theory in Developing Effective Health Communications. Journal of Communication, 56(suppl_1), S1–S17.
https://doi.org/10.1111/j.1460-2466.2006.00280.x
Geusens, F., & Beullens, K. (2017). The reciprocal associations between sharing alcohol references on social networking sites and binge drinking: A longitudinal study among late adolescents. Computers in Human Behavior, 73, 499–506.
https://doi.org/10.1016/j.chb.2017.03.062
Geusens, F., Bigman-Galimore, C. A., & Beullens, K. (2020). A cross-cultural comparison of the processes underlying the associations between sharing of and exposure to alcohol references and drinking intentions. New Media & Society, 22(1), 49–69.
https://doi.org/10.1177/1461444819860057
Glaser, B. G., & Strauss, A. L. (1967). The Discovery of Grounded theory: Strategies for Qualitative Research. Chicago: Aldine Publishing Company.
Greenbaum, T. L. (2000). Moderating focus groups: A practical guide for group facilitation. London: Sage Publications.
Hendriks, H., Gebhardt, W. A., & Van den Putte, B. (2017). Alcohol-Related Posts from Young People on Social Networking Sites: Content and Motivations.
Cyberpsychology, Behavior, and Social Networking, 20(7), 428–435. https://doi.org/10.1089/cyber.2016.0640
Hendriks, H., Gebhardt, W. A., Van den Putte, B., & De Nooy, W. (2019). What’s there to like? Testing the causal effects of alcoholposts on Facebook on drinking behavior. Paper presented at the European Conference for Health Communication (ECHC), Zurich, 2019
Hendriks, H., van den Putte, B., de Bruijn, G.-J., & de Vreese, C. H. (2014). Predicting Health: The Interplay Between Interpersonal Communication and Health Campaigns. Journal of Health Communication, 19(5), 625–636.
https://doi.org/10.1080/10810730.2013.837552
Hendriks, H., Van den Putte, B., & Gebhardt, W. A. (2018). Alcoholposts on Social
Networking Sites: The Alcoholpost-Typology. Cyberpsychology, Behavior, and Social Networking, 21(7), 463–467. https://doi.org/10.1089/cyber.2017.0729
Hendriks, H., Van den Putte, B., Gebhardt, W. A., & Moreno, M. A. (2018). Social Drinking on Social Media: Content Analysis of the Social Aspects of Alcohol-Related Posts on Facebook and Instagram. Journal of Medical Internet Research, 20(6), e226.
https://doi.org/10.2196/jmir.9355
Hendriks, H., Wilmsen, D., Van Dalen, W., & Gebhardt, W. A. (2019). Picture me drinking: Alcohol-related posts by Instagram influencers popular among adolescents and young adults. Frontiers in Psychology, 10, 2991.
Hochbaum, G., Rosenstock, I., & Kegels, S. (1952). Health belief model. United states public health service, W432W8784.
Hoffman, E. W., Pinkleton, B. E., Weintraub Austin, E., & Reyes-Velázquez, W. (2014). Exploring College Students’ Use of General and Alcohol-Related Social Media and
