The influence of peer’s social networks on adolescent’s cannabis use: a systematic review of longitudinal studies

Aim: A systematic review was performed to summarize the key findings of the peer influence on cannabis use through Social Network Analysis (SNA) studies and identify limitations and gaps with the purpose of informing future research and practice. Longitudinal studies were included since they provide robust information about social relationships change over time.

Background: Adolescents’ cannabis use is a global problem, which has awakened an interest in its determinants such as social influences. Research has shown the importance of these influences on cannabis uptake and use. SNA is an useful relational approach to examine socialization mechanisms related to the onset of cannabis use in adolescents.

Method: A search was conducted in PyscINFO, PubMed, Scopus and Web of Science for longitudinal articles published until February 2023, to examine cannabis use and peer’s social networks. We focus on peers’ influence of peers on cannabis use. Additionally, information about effect of cannabis use for peer selection was collected.

Results: The results of the included studies (n = 8) showed that friends’ cannabis use was most often/strongly associated with cannabis use. There was also an increase of cannabis use when the adolescent did not feel close to the school’s peers, had a higher proportion of friendships relative to the total number of ties in the neighborhood, had a central position, did not belong to any group but had ties to members of two or more groups, had cannabis user friends (especially in early ages), and lived in a neighborhood where cannabis was used.

Conclusion: Cannabis use is mainly related to friends’ use. Yet, future studies are warranted to control for relevant selection effects to further knowledge on network effects on cannabis use, improving the design, and improving the modeling of the network. This systematic review may inform about the critical aspects of preventing cannabis use among adolescents, taking into consideration their complex social environment.

1 Introduction

Cannabis is the third most consumed drug among adolescents (1, 2). According to the World Drug Report issued in 2021 (2), 13.8 million of adolescents aged between 15 and 16 years old used cannabis worldwide. This figure is equivalent to 5.6% of the global population and exceeds the prevalence rate among the general population (3). Furthermore, the European report on drugs (2021) declared cannabis to be the most established drug used in Europe, finding that 15.4% of young people had consumed cannabis in the previous year (1). It is also important to note that cannabis has become a legal drug in many high-income countries in recent decades (4). This legalization of cannabis has led to a 3.8% increase in recreational use in those states (5).

The onset age of cannabis use is around 15 years old (1, 2). Adolescents are particularly vulnerable to cannabis use because their brain is still developing (6). Some consequences of cannabis use are short-term, such as cognition and coordination problems, toxicity or traffic injuries, but also long-term, such as mental disorders, addiction, suicide risk, and cardiovascular or pulmonary diseases (6–8). These consequences, furthermore, can affect the social and family sphere, leading to, for instance, school dropout and family/friend’s problems (e.g., breach of family rules, physical and psychological violence, deterioration, or loss of relationships) (7–9). Moreover, adolescents are sensitive to joining and remaining in peer groups. Peers are individuals who share similar ages and interests, and tend to belong to a similar social group. Adolescents are highly influenced by peers since they try to resolve disagreements by adopting peers’ behaviors (10).

The importance of peer influence on adolescent’s recreational cannabis use has been demonstrated in a substantial body of research (11–13). For instance, peers close to adolescents exert a stronger influence on adolescent cannabis use than peers less close to adolescents (12). Additionally, well-stablished socio-cognitive theories discuss the importance of peer influences on risk behaviors (11–13). A first type of peer influence concerns social modelling. This factor, proposed by Bandura in his Social Cognitive theory (14), implies that behaviors of others such as cannabis use can be adopted by merely observing their behavior (15, 16). When such a behavior is reinforced, this behavior is more likely to become adopted (14). This process may occur unconsciously. Furthermore, occupied social position within a peer group can also play a significant role in how adolescents are influenced toward cannabis use. Adolescents may see high-status peers or more popular peers (usually cannabis users) as role models, in order to improve their own social standing into the group (13). Another type of social influence concerns social norms, a construct originally proposed by Fishbein and Ajzen in the theory of Reasoned Action (17). Norms of other people have shown to influence other persons behavior, including cannabis use such as previous studies have shown regarding adolescent peer’s social norms favoring cannabis use (10, 11). Furthermore, during adolescence, it is expected that youth may not reject the cannabis use offers because they want to fit in with their peers (11). An even more explicit type of social influence concerns direct peer pressure from others (18, 19). These three types of social influence can operate and can have unique contributions (20–23).

The traditional theoretical concepts about social influences did not look to the complex constellation of peer’s influence processes that can modulate the social modelling, norms or pressure (e.g., the influence of social structure of the friendships or the interactions between the adolescents within a social network). Taking such a social network approach implies the use of Social Network Analysis (SNA). SNA is focusing on examining the social structure and interactions among social actors within a social network (24). A network is comprised of nodes (i.e., individuals/actors) and the ties/relationships between those nodes. SNA assumes that social actors and the network they are embedded in are interdependent (25). SNA also makes it possible to control correctly for possible selection and confounding effects while examining peer influence processes, for instance, regarding the cannabis use. Besides, adolescent’s cannabis use can also become similar to their peers because they select each other based on similar cannabis use behavior or due to peers jointly being influenced by an external source (i.e., both watching a movie in which actors use cannabis) (26, 27). This also adds value in regard to previous approaches in cannabis prevention field. Several studies have now applied SNA to examine peer network influences in adolescent cannabis use (28, 29). So it is now timely and relevant to review and summarize their findings.

Longitudinal studies may provide important insights into dynamic social relationships, since they address temporality (30). They consider the dynamics of social phenomena as being the result of a time-process, where observations are made at different time points. Reviewing longitudinal studies is needed in order to understand dynamic relationships between individuals and peer groups patterns regarding cannabis use, and broader patterns of social change over time (31). Thus, this review included only longitudinal studies in order to tracking the influence of peers on cannabis use over time (causal relationship), reducing the bias inherent in cross-sectional studies, and facilitating future comprehensive experimental research.

The goal of this study is to systematically synthesize the scientific literature on longitudinal applications of social network approach to study peer influences on adolescent recreational cannabis use. The research question was: do peer networks characteristics influence in recreational cannabis use in adolescents between 12 and 21 years? This review will summarize the key characteristics and findings of peer influences on recreational cannabis use and identify limitations and gaps in the literature, with the purpose of informing future research and practice. Additionally, measures for selection effects on cannabis use will be also included.

2 Methods

2.1 Search strategy and retrieval system

For conducting and communicating this systematic review, the recommendations of PRISMA 2020 statement were followed (32). The protocol was pre-registered in PROSPERO (33). Eligible studies were identified in October 2021 and the search was updated in February 2023 by conducting an in-depth literature search on electronic databases used (peer-reviewed): PsycINFO, Scopus, PubMed and Web of Science. The search strategy used a combination of Boolean connections and search terms relevant to key concepts used in this review: social network, cannabis use and adolescent. As example, the specific search strategy in the PubMed database was:

((Network*[tiab]AND friend*[tiab]) OR relation*[tiab] OR peer*[tiab] OR social*[tiab] OR media*[tiab] OR acquaintance*[tiab] OR team*[tiab] OR mate*[tiab] OR partner*[tiab] OR leisure[tiab] OR hobby[tiab] OR school[tiab] OR highschool[tiab] OR university[tiab] OR junior high* OR senior high*[tiab] OR colleague*[tiab]) AND (“Adolescent”[Mesh] OR “Young Adult”[Mesh] OR adolescen*[tiab] OR teen*[tiab] OR young*[tiab] OR youth*[tiab] OR puber*[tiab] OR minor*[tiab] OR juvenil*[tiab] OR student*[tiab] OR pupil*[tiab]) AND (“Marijuana Use”[Mesh] OR “Marijuana Smoking”[Mesh] OR “Cannabis”[Mesh] OR cannabi*[tiab] OR blunt*[tiab] OR marijuana*[tiab] OR marihuana*[tiab])

Additionally, we searched reference lists from each study (backward) and articles that cite back to a specific article (forward), for additional articles. No restrictions were placed on the language and the time window of search.

2.2 Study selection criteria

2.2.1 Inclusion criteria

• Study design: longitudinal empirical studies.

• Study subject, target population: adolescents (between 12 and 21 years old). Although organizations as the World Health Organization or American Academy of Pediatric usually define the adolescence as the period of 10–21 years, adolescents from 12 years were chosen because cannabis is a substance which use starts late (34, 35). In case there were studies where they include our age’s rank but also older ages, we will only take the data from our target population, if it is possible.

• Statistical method: Social Network Analysis (SNA) descriptive analysis and/or statistical models for social networks.

2.2.2 Exclusion criteria

• Medicinal cannabis use and patients.

• Intervention studies.

• No collection of complete social network data (i.e., no information on social actors and the ties among them, collection of individual data or only dyad and triad level data).

• Simulation studies that focused on varying intervention parameters.

• Exclusive use of conventional statistical method (e.g., regression) rather than SNA in data analysis.

2.3 Data extraction and preparation

The process was developed in the following phases: firstly, after deleting the duplicates, a screening of articles based on title and summary following the inclusion / exclusion criteria was done. The second screening was through the reading of full texts, to which methodological evaluation was also done. This process was done independently by two authors (MCTG and MABJ), who also independently checked results to evaluate inclusion and exclusion criteria. Discrepancies between these two authors were discussed and a final determination decided by a third author (LM). Data were extracted by a structured form. The following variables were considered: instrument, article ID, publication year, country, design, sample, type of network, social network measures, SNA performed, type of software used to conduct the SNA, and main findings.

2.4 Data synthesis

The studies were heterogeneous in the types of measures used, and it was therefore more appropriate to follow a narrative data synthesis strategy. The review was focused on variables related to peer social network influences and not on other types of network ties (such as family, social kin or broader social contacts).

All included peer network influence effects on cannabis use were extracted from each included manuscript. They were classified in three groups: (1) Endogenous network influence effects on cannabis use (the influence of the network structure and interactions itself without considering other variables such as the network’ behavior or personal characteristics), (2) Cannabis-related network influence effects on cannabis use (the influence of the network considering the cannabis use of actors in the network) and, (3) Other risk behavior related network effects on cannabis use (e.g., alcohol drinking, or personal characteristics). Interaction effects between peer network influence effects were also extracted (e.g., popularity x density).

A few examples of endogenous network influence effects on cannabis use are the effect of number of nominated friends (Outdegree); the effect of number of incoming friendship ties/relationships (Indegree/popularity); the effect of being in a central position within the network (centrality).

Examples of cannabis-related network influence effects on cannabis use are the effect of cannabis use of someone’s friends (friends’ cannabis use); the effect of number of steps it takes for someone to reach the nearest cannabis user in the network considered friends and friend’s friends (distance to nearest cannabis users considered friends and friend’s friends).

Examples of other risk behavior related network effects on cannabis use are the effect of smoking behavior of someone’s friends (friends’ smoking behavior); the effect of the age of someone’s friends (friends’ age).

Additionally, for those studies that controlled for selection effects, cannabis use related selection effects were extracted (being the outcome the selection of peers instead of the cannabis use). Examples of cannabis use related selection effects are the effect of own cannabis use on selecting an actor as a friend (cannabis use Ego); the effect of cannabis use of another actor on selecting that actor as a friend (cannabis use alter); the effect of having similar cannabis use to another actor on selecting that actor as a friend (cannabis use similarity).

2.5 Study quality assessment

Two reviewers (MCTG and MABJ) independently assessed risk of bias for all studies that met eligibility criteria, using the National Institutes of Health’s Quality Assessment Tool for Observational Cohort, Case–Control Studies, Systematic reviews and Meta-Analyses (36). The assessment tool rates each study based on 14 criteria. For each criterion, a score of one was assigned if ‘yes’ was the response, whereas a score of zero was assigned otherwise (i.e., an answer of ‘no,’ ‘not applicable,’ ‘not reported’ or ‘cannot determine’). The study-specific global score could range from 0 to 14. Discrepancies between MCTG and MABJ were discussed, and a final determination was decided by a third author (LM).

3 Results

3.1 Study selection

As depicted in the flow diagram (Figure 1), the initial search yielded 3,261 scientific articles. Adding the articles identified through backward and forward searching, a total of 4,196 articles were found.

FIGURE 1

Figure 1. PRISMA 2020 flow diagram for new systematic reviews of reviewed studies.

During the review process, many studies that had used the traditional regression-based approach to study the influence of peers on cannabis use were found, that is that they estimated peer effect without developing a complete network.

The main reasons for exclusion were that articles focused on medicinal cannabis use, patients, did not use statistical models for social networks such as SNA, did not include a whole or complete social network, were not longitudinal, or did not meet the age criterion (12–21 years). Finally, 8 articles were included for review (37–44).

3.2 Study quality assessment

Table 1 reports the specific scoring rubric used for assessing each study’s methodological quality. All included studies had a high-quality level according to the used tool. Each study clearly defined their research question, their target population, their design, and their analytic methods. However, there were three studies that did not control for the most important confounder in peer influence research, namely selection effects, implying that the estimations of these effects may be bias (40, 43, 44). Thus, their results were treated with caution.

TABLE 1

Table 1. Summary of the criteria used for assessing each study’s methodological quality³⁶.

3.3 Basic characteristics of the included studies

Table 2 summarizes the characteristics of the eight longitudinal studies that met the inclusion criteria. The reviewed studies had been published in six journals that mainly emphasized adolescence. Three of the eight studies had common authors (39, 42, 43). Every study had been conducted in the USA and used at least three data measurement points. The included studies were published between 2006 and 2018. Six of the eight studies used data from the Add Health study (The National Longitudinal Study of Adolescent to Adult Health) (37–42), one study used data from the PROSPER study (Promoting School-Community-University Partnerships to Enhance Resilience) (43), and one had collected its own school network data (44). The sample size ranged from 1,373 to 9,500 participants, and mostly examined adolescents aged 15–16 years old. To focus on friendship ties among students within the same school, each school was defined as a separate social network, and most of the included studies had created a specific model for each school group (37, 39, 41, 42, 44).

TABLE 2

Table 2. Characteristics of the longitudinal included studies.

Table 2 also reports the outcome measures, social network characteristics, and statistical models used. Cannabis use in the last 3 months was the outcome of one study (44), cannabis use in the last month was the outcome of the remaining seven studies (37–43).

In all studies, social network data were constructed based on nominations within school grades, provided by respondents via survey data. The studies differed in the type of nominations and the maximum allowed number of nominations. Regarding the number and the type of allowed nominations, six studies asked respondents to nominate ten of their best friends (five females and five males) (37–42), one study asked to nominate two best friends and five additional friends (42)^, and the last one asked to nominate a maximum of five best friends (44). The last studies did not make a distinction by gender. Regarding the social network methods used, five studies applied stochastic actor-based/orientated models (SABM/SAOM) via R Siena (37–39, 41, 42), one of them used network measures that was already provided on Add Health (39), one used their own network analysis routine (43), and the last one calculated network measures in UCINET that were later treated as variables in three-level hierarchical generalized linear models (44).

3.4 Key findings regarding peer network influence effects on cannabis use

The main objectives of the study and key findings about peer network influences on adolescent cannabis use are presented in Table 3. Furthermore, the definitions of the included effects (Supplementary Table S1) and a more concise overview of the significance of all included peer influence (Supplementary Table S2) are available in the supplementary materials.

TABLE 3

Table 3. Aim and major findings related to peer influence and selection effects on adolescent’s cannabis use.

3.4.1 Endogenous network influence effects on cannabis use

The most examined effect was own popularity (37, 40, 43, 44). Two of the studies did not find a significant effect (37, 43), while two others found a significant positive effect indicating that the more incoming ties one had, the more cannabis they used (40, 44). One study examined the effect of their own popularity on cannabis use in three different age groups (44), and found that when age increase, the effect of their own popularity decreases. Finally, one study examined whether the effect of own popularity on cannabis use could be moderated by density, school level’s alcohol and tobacco use, and school connectedness (40). Only the interaction with school connectedness was statistically significant, indicating that the effect of own popularity on cannabis use was less strong for schools where students felt happy and close to the people at school (40).

Outdegree, neighborhood density, and reciprocity were jointly examined in one study (44). Although adolescents nominating a higher number of friends (high outdegree) had a significantly higher risk of cannabis use, being nested in denser network neighborhoods protected against cannabis use. Reciprocity did not appear to significantly influence cannabis use (44). However, in this study, possible selection effects were not modelled.

Three studies examined the effect of being in a central position within the network on cannabis use (40, 43, 44). Although different centrality measures were modelled, most centrality effects were not significant (40, 44). Only reach centrality significantly influenced cannabis use in two studies (43, 44). This indicates that adolescents that are closer to all other actors in the network, have a higher tendency to use cannabis in comparison with those who are less close to others. Unfortunately, none of the studies controlling for selection included centrality effects in their models. The same applies for group position effects that were also only examined in two studies not controlling for selection effects (43, 44). Only being a liaison compared to being a core member significantly increased cannabis use (42). These two studies which included influence effects related to the adolescent’s position in the network created different dummy variables to make the comparison (43, 44). The first study distinguished between one cohesive group (core and peripheral members), multiple groups (liaisons), or no group (isolates and other non-members). They were represented through a set of dummy variables that contrast all other positions to core members (43). The second study measured adolescents’ position in the network as a group member, bridge, or isolate. They were measured by two dummy-coded variables with group member as the reference (44). See Supplementary material for more information.

3.4.2 Cannabis-related network influence effects on cannabis use

The cannabis use of friends was included in six studies (37–39, 41–43). Three of them showed a positive significant influence of friends’ cannabis use on adolescent cannabis use (37, 38, 43), while one study showed no significant results (41). The two remaining studies (39, 42) showed mixed results, and one of them had two different outcomes that gave different results (42).

In the first study, adolescent’ lifetime cannabis use was significant influenced by friend’s lifetime cannabis use only at one school (39). In the second study, adolescent’s lifetime cannabis use was not statistically significantly influenced by friend’s lifetime cannabis use. Yet, it was significantly influenced by friend’s last month cannabis use at one school but not the other (42). Additionally, no significant results of friend’s cannabis use were found for adolescent’s last month cannabis use (42).

Only one study modelled the specific effect of social proximity with a cannabis user, where not having a best friend cannabis user, having fewer neighborhood cannabis users, and being at a greater distance to the nearest user among friends and friends’ friends, led to less adolescent cannabis use (44). The influence of best friend’s use was greatest at early adolescence (44).

One study found that the higher mean of friends’ cannabis users, the more adolescent’s cannabis use (43). This study also examined whether the effect of friends’ cannabis use could be moderated by the number of friends. The effect of the average of friends’ users is stronger on own cannabis use in case of lower number of friends, possibly due to the fact that adding a single friend is more consequential for adolescents with fewer friends than for those with many friends (43). Another study examined whether friends’ cannabis use could be moderated by reciprocity, friends’ popularity and popularity difference in two schools (41). The interaction with reciprocity was significant only in one school. In this, the effect of friend’s cannabis use on adolescent’ cannabis use was stronger for adolescents who tended to adopt the cannabis use behaviors of their mutual friends. The interaction with friend popularity was significant also in one of both schools, where adolescents were likely to adopt the cannabis use behaviors of their more popular friends (41).

3.4.3 Other effects of network-related risk behavior related network effects On cannabis use

One study found that adolescents were more likely to use cannabis if their peers used alcohol or tobacco (40). Yet, another study found that the number of friends who smoke and drink at school was not statistically significant with cannabis use (37).

3.5 Findings regarding cannabis use related selection effects

There were five SIENA studies that incorporated a co-evolution approach, in which peer influence was controlled for selection effects. Furthermore, they incorporated alternative network and behavior change mechanisms to avoid misdiagnosis of selection or influence effects when another social process is operating. In all five studies, adolescents had the significant tendency to select friends based on similarities in cannabis use behavior, regardless of the outcome (lifetime cannabis use or last month use) and regardless of which operationalization was used for the selection effects in the estimation procedure (37–39, 41, 42). Additionally, cannabis users tended to make fewer friends (39, 42).

4 Discussion

This review is the first to longitudinally examine the influence of peer social networks on cannabis use, taking a social network approach. During the review process, the methodological quality of the included studies was critically evaluated, considering the high quality of those studies that controlled for selection effects (37–39, 41, 42). This review found multiple outcomes that sometimes result in different results, such as, for instance, some studies controlling for selection effects and others not.

The most often found effect was the positive influence of friend cannabis use on adolescent cannabis use, meaning that adolescents are more likely to use cannabis if their close friends use cannabis. Yet, different operationalizations were used for assessing the influence of friends’ cannabis use. Other studies not included in this review, have also shown that the influence of peers is an important factor in cannabis use (11, 21, 22, 45). Although these studies did not take a network approach, they showed that adolescents are influenced by the perceived cannabis use of friends (modeling) (21, 22, 39), norms regarding cannabis use among Friends (11, 21, 22), and that they may experience pressure to conform to the behavior of their cannabis’ consuming Friends (22). Based on these studies and our results in the present review, we can conclude that it is important to consider the influences of peers in cannabis prevention efforts.

Although the own popularity was explored by four studies (43, 46–48), results were inconclusive. Those studies, that showed significant positive effects, unfortunately did not control for selection effects (46, 48). The one study that controlled for selection did not find significant effects of popularity (37). Hence, the development of more longitudinal studies controlling for selection effects is needed to study the effects of popularity and possible interactions with popularity on cannabis use.

Regarding the effects of social position on cannabis use, only one study suggested that an adolescent who takes a liaison position in the network could be a potential diffusion agent of behaviors towards cannabis use (43). There is very little evidence to draw any conclusions about the effect of social position on cannabis use. In other research, no association was found between social network position and cannabis use (46). More research is needed to analyze how social positions are related to cannabis use. If different social positions do influence cannabis use, prevention efforts need to consider them as they can possibly interfere in intervention process by empowering (e.g., when multiple non-cannabis users are in a bridge position) or diminishing (e.g., when multiple cannabis users are in a bridge position) intervention effects.

It is also relevant to further explore the changes in peer influence effects over time. One included study showed that age seems to act as a moderator of the peer influence and selection effects on cannabis use (44). For instance, the influence of best friend’s use was greatest at early adolescence, when they are most vulnerable to peer influence. Previous research has shown that at early adolescence, peers become the reference and the main source of support, and adolescents are concerned about the need to feel accepted by those around them (14, 23, 46). This feeling faded when adolescents were older (44, 46, 49).

Our findings showed a large diversity in effects, indicating that more research is needed to clarify peer influence processes on cannabis use. The use of stochastic actor-orientated modelling (SAOM) implies the use of a complete network, as well as leads to a specific and better approach to study peer influence effects controlling for selection effects (47, 50). Data about peer influence on substance use applying SAOM have been published in the last decade, especially for alcohol and tobacco (47, 50). A similar approach is recommended for studying cannabis use.

Some limitations are evident in this review. First, the number of included studies is low, most of them belonged to the Add Health study, and the peer influence are measured differently. Although Add Health is a highly powerful study in terms of social networks, additional research, through other validated surveys, examining social networks in other communities is needed to be able to identify common patterns regarding social networks. Second, the instrument used to assess the methodological quality of this literature does not include questions to assess the control of selection effects, neither the number of adolescents on the network nor its changes over time (i.e., the number of joining and leaving participants in the network). It would be recommended to develop a standard tool to assess the methodological quality of SNA studies.

Despite these limitations, this systematic review supported earlier findings suggesting the role of peer influence on cannabis use among adolescents. An unique point of this review is the specificity and adequacy of the study selection criteria, for instance, including only longitudinal studies and completed social networks.

4.1 Implications for future research and practice

Identifying the effect of peer influence on cannabis use is crucial in order to improve the social approach of the traditional socio-cognitive models, as well as prevention and early intervention programs (11–13, 20–23). Once more research would include a broad range of effect in a correct longitudinal design, it could start to translate these peer network influence effects into tools to help intervention programs to become more effective, taking advantage of the composition of the network. Then, it will be of great value to direct prevention efforts towards peer groups that are at risk of cannabis use.

Future research should conduct more longitudinal studies, controlling for selection effects, improving the modelling of the network (i.e., using a complete network), and with more effects of peer influence effects. For instance, studying the role of own popularity and friend popularity, and the influence of the number of alcohol and tobacco friend’s users on cannabis use. It would be recommended that studies control for selection effects. Moreover, as indicated by previous literature (48, 51), it is important to examine more data about the frequency of cannabis use and patterns according to gender and ethnicity, since these could be different in these populations. Additionally, although a growing body of evidence indicates that school environments have a strong influence on adolescent cannabis use (51), future studies could aim to have a broader scope of the network around adolescents, including, for example, out-of-school networks.

It is also important to include direct observations of interactions between targets and peers in order to better understand how peers affect each other and how this influence works overtime. There is already an European intervention trial with social network data for smoking cigarettes (52, 53). Yet, SNA was not yet used within cannabis use prevention, probably due to the lack of adequate SNA studies to support the use of SNA in such programs. Moreover, the European Monitoring Centre for Drugs and Drug Addition has recognized in 2021 the importance of social responses to cannabis-related problems (54). It might be relevant to include the SNA in this type of standards. Finally, further studies may also focus on the impact of the legalization of recreational cannabis use on peer network influence, considering that most of the included articles were conducted before the legalization of marijuana in the USA. The legalization could potentially amplify the positive perception of cannabis use within peer groups.

5 Conclusion

To develop effective cannabis prevention programs, it is essential to comprehend the impact of social networks on cannabis use behavior. This systematic review underscores the significance of peer influences on adolescent cannabis use. This review provides a summary of longitudinal findings from scientific evidence regarding peer influence on cannabis use, utilizing the social network approach, and confirming the importance of peer influences on adolescent cannabis use. Different operationalization of the influence of friends’ cannabis use was the most significant peer influence effect on adolescent’s cannabis use. Regrettably, disparities in reporting other peer influence effects hinder optimal comparisons, making it advisable to identify the best methods for measurement in the future (e.g., through consensus meetings or Delphi studies). Therefore, more extensive, and improved research is warranted. Additional studies that control for relevant selection effects are necessary to advance our understanding of these network effects on cannabis use, with an emphasis on expanding the range of network influence effects considered.

Author contributions

MCTG: Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Writing – original draft. MABJ: Data curation, Formal analysis, Investigation, Methodology, Supervision, Writing – review & editing. MLS: Conceptualization, Funding acquisition, Resources, Supervision, Writing – review & editing. HV: Conceptualization, Supervision, Writing – review & editing. LM: Conceptualization, Formal analysis, Investigation, Methodology, Supervision, Writing – review & editing.

Funding

The author(s) declare financial support was received for the research, authorship, and/or publication of this article. MCTG acknowledges the support of the PIF research program of the University of Seville (VPPI-US). The authors wish to acknowledge the financial support of the grant PID2019-107229RA-I00 funded by MCIN/AEI/10.13039/501100011033.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fpsyt.2023.1306439/full#supplementary-material

References

* Included in the literature review

Google Scholar

1. World Drug Report 2021 (United Nations publication, Sales No. E.21.XI.8).

Google Scholar

2. European Monitoring Centre for Drugs and Drug Addiction. European drug report 2021: Trends and developments. Luxembourg: Publications Office of the European Union (2021).

Google Scholar

3. Page, N, Hallingberg, B, Brown, R, Lowthian, E, Hewitt, G, Murphy, S, et al. Change over time in adolescent smoking, cannabis use and their association: findings from the school Health Research network in Wales. J Public Health (Oxf). (2020) 43:e620–8. doi: 10.1093/pubmed/fdaa174

PubMed Abstract | CrossRef Full Text | Google Scholar

4. Hopfer, C. Implications of marijuana legalization for adolescent substance use. Subst Abus. (2014) 35:331–5. doi: 10.1080/08897077.2014.943386

PubMed Abstract | CrossRef Full Text | Google Scholar

5. Camps, Camila, Exploring the impact of legalization of recreational marijuana on adolescent drug use (2021). Thesis. Rochester Institute of Technology.

Google Scholar

6. Volkow, ND, Baler, RD, Compton, WM, and Weiss, SR. Adverse health effects of marijuana use. N Engl J Med. (2014) 370:2219–27. doi: 10.1056/NEJMra1402309

PubMed Abstract | CrossRef Full Text | Google Scholar

7. Verbanck, P. Effets à court et à long Terme de la consommation de cannabis [short-term and long-term effects of cannabis use]. Rev Med Brux. (2018) 39:246–9.

PubMed Abstract | Google Scholar

8. Hall, WD, Renström, M, and Poznyak, V: The health and social effects of nonmedical cannabis use. Alcohol, Drugs and Addictive Behaviors. World Health Organization (2016) 72.

Google Scholar

9. La fundacion de Ayuda contra la Drogradicción (FAD). Prevencion familiar de problemas asociados al consume de cannabis Consejeria de Salud y Familias de la Junta de Andalucía (2022).

Google Scholar

10. Tomé, G, Matos, M, Simões, C, Diniz, JA, and Camacho, I. How can peer group influence the behavior of adolescents: explanatory model. Glob J Health Sci. (2012) 4:26–35. doi: 10.5539/gjhs.v4n2p26

PubMed Abstract | CrossRef Full Text | Google Scholar

11. Stone, AL. Adolescent Cannabis use and perceived social norm trends pre- and post-implementation of Washington State’s liberalized recreational Cannabis policy: healthy youth survey, 2008–2018. Prev Sci. (2020) 21:772–83. doi: 10.1007/s11121-020-01136-0

PubMed Abstract | CrossRef Full Text | Google Scholar

12. Zaharakis, N, Mason, MJ, Mennis, J, Light, J, Rusby, JC, Westling, E, et al. School, friends, and substance use: gender differences on the influence of attitudes toward school and close friend networks on Cannabis involvement. Prev Sci. (2018) 19:138–46. doi: 10.1007/s11121-017-0816-y

PubMed Abstract | CrossRef Full Text | Google Scholar

13. Dijkstra, JK, Cillessen, AHN, Lindenberg, S, and Veenstra, R. Basking in reflected glory and its limits: why adolescents hang out with popular peers. J Res Adolesc. (2010) 20:942–58. doi: 10.1111/j.1532-7795.2010.00671.x

CrossRef Full Text | Google Scholar

14. Bandura, A. Social foundations of thought and action. Englewood Cliffs, NJ: Prentice Hall (1986).

Google Scholar

15. Burdzovic Andreas, J, Pape, H, and Bretteville-Jensen, AL. Who are the adolescents saying "no" to cannabis offers. Drug Alcohol Depend. (2016) 163:64–70. doi: 10.1016/j.drugalcdep.2016.03.025

PubMed Abstract | CrossRef Full Text | Google Scholar

16. Jalilian, F, Mirzaei-Alavijeh, M, Ahmadpanah, M, Mostafaei, S, Kargar, M, Pirouzeh, R, et al. Extension of the theory of planned behavior (TPB) to predict patterns of marijuana use among young Iranian adults. Int J Environ Res Public Health. (2020) 17:1981. doi: 10.3390/ijerph17061981

PubMed Abstract | CrossRef Full Text | Google Scholar

17. Ajzen, I, and Fishbein, M. Attitude-behavior relations: a theoretical analysis and review of empirical research. Psychol Bull. (1977) 84:888–918. doi: 10.1037/0033-2909.84.5.888

CrossRef Full Text | Google Scholar

18. Evans, RI, Rozelle, RM, Mittelmark, MB, Hansen, WB, Bane, AL, and Havis, J. Deterring the onset of smoking in children: knowledge of immediate physiological effects and coping with peer pressure, media pressure, and parent modeling 1. J Appl Soc Psychol. (1978) 8:126–35. doi: 10.1111/j.1559-1816.1978.tb00771.x

CrossRef Full Text | Google Scholar

19. de Vries, H, Mudde, A, Leijs, I, Charlton, A, Vartiainen, E, Buijs, G, et al. The European smoking prevention framework approach (EFSA): an example of integral prevention. Health Educ Res. (2003) 18:611–26. doi: 10.1093/her/cyg031

CrossRef Full Text | Google Scholar

20. De Vries, H, Backbie, E, Kok, G, and Dijkstra, M. The impact of social influences in the context of attitude, self-efficacy, intention and previous behavior as predictors of smoking onset. J Appl Soc Psychol. (1995) 25:237–57. doi: 10.1111/j.1559-1816.1995.tb01593.x

CrossRef Full Text | Google Scholar

21. Kam, JA, Matsunaga, M, Hecht, ML, and Ndiaye, K. Extending the theory of planned behavior to predict alcohol, tobacco, and marijuana use among youth of Mexican heritage. Prev Sci. (2009) 10:41–53. doi: 10.1007/s11121-008-0110-0

PubMed Abstract | CrossRef Full Text | Google Scholar

22. Torrejón-Guirado, MC, Lima-Serrano, M, Mercken, L, and de Vries, H. Which factors are associated with cannabis use among adolescents in Andalusia? An application of the I-change model. J Nurs Scholarsh. (2023) 55:739–50. doi: 10.1111/jnu.12806

PubMed Abstract | CrossRef Full Text | Google Scholar

23. De Vries, H. An integrated approach for understanding health behavior; the I-change model as an example. Psychol Behav Sci Int J. (2017) 2:555585. doi: 10.19080/PBSIJ.2017.02.555585

CrossRef Full Text | Google Scholar

24. Ioana-Alexandra, A. An overview of software applications for social network analysis. Int Rev Soc Res. (2013) 3:71–7. doi: 10.1515/irsr-2013-0023

CrossRef Full Text | Google Scholar

25. Wasserman, S, and Faus, K. Social network analysis in the social and behavioural sciences In: Social network analysis: methods and aplications. New York: Cambridge University Press (1994). 1-27.

Google Scholar

26. Borgatti, SP, and Martin, GE. Network analysis of 2-mode data. Soc Networks. (1997) 19:243–69. doi: 10.1016/S0378-8733(96)00301-2

CrossRef Full Text | Google Scholar

27. Peterson, S. (2017). Why Can’t we be friends? Exploring short-term peer selection and peer influence dynamics using longitudinal social network analysis. University of Cincinnati.

Google Scholar

28. Lorant, V, and Tranmer, M. Peer, school, and country variations in adolescents’ health behaviour: a multilevel analysis of binary response variables in six European cities. Soc Networks. (2019) 59:31–40. doi: 10.1016/j.socnet.2019.05.004

CrossRef Full Text | Google Scholar

29. Mathys, C, Burk, WJ, and Cillessen, AHN. Popularity as a moderator of peer selection and socialization of adolescent alcohol, marijuana, and tobacco use. J Res Adolesc. (2013) 23:513–23. doi: 10.1111/jora.12031

CrossRef Full Text | Google Scholar

30. Snijders, TA. Longitudinal methods of network analysis. Encyclopedia Complexity Syst Sci. (2009) 24:5998–6013. doi: 10.1007/978-0-387-30440-3_353

CrossRef Full Text | Google Scholar

31. Ryan, L, and D'Angelo, A. Changing times: migrants’ social network analysis and the challenges of longitudinal research. Soc Networks. (2018) 53:148–58. doi: 10.1016/j.socnet.2017.03.003

CrossRef Full Text | Google Scholar

32. Page, MJ, McKenzie, JE, Bossuyt, PM, Boutron, I, Hoffmann, TC, Mulrow, CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Syst Rev. (2021) 10:89. doi: 10.1186/s13643-021-01626-4

PubMed Abstract | CrossRef Full Text | Google Scholar

33. Torrejón-Guirado, MC, Lima-Serrano, M, Baena-Jiménez, MA, De Vries, H, Jolani, S, and Mercken, L. Peer's social influences on cannabis use in adolescents: a systematic review of longitudinal studies. PROSPERO 2020 CRD42020202873 (2020). Available at: https://www.crd.york.ac.uk/prospero/display_record.php?ID=CRD42020202873

Google Scholar

34. World Health Organization. Adolescence health (2022) Available at: https://www.who.int/health-topics/adolescent-health#tab=tab_1

Google Scholar

35. Hagan, JF, Shaw, JS, and Duncan, PM eds. Bright futures: Guidelines for health supervision of infants, children, and adolescents. 4th ed. Elk Grove Village, IL: American Academy of Pediatrics (2017).

Google Scholar

36. National Heart, Lung, and Blood Institute. Quality assessment tool for observational cohort and cross-sectional studies (on-line). (2017). Available at: https://www.nhlbi.nih.gov/health-pro/guidelines/in-develop/cardiovascular-risk-reduction/tools/cohort

Google Scholar

37. *Wang, C, Hipp, JR, Butts, CT, and Lakon, CM (2018) The interdependence of cigarette, alcohol, and marijuana use in the context of school-based social networks. PLoS One 13::e0200904. doi: 10.1371/journal.pone.0200904

PubMed Abstract | CrossRef Full Text | Google Scholar

38. *Schaefer, D.R. A network analysis of factors leading adolescents to befriend substance-using peers. J Quant Criminol 34, 275–312 (2018). doi: 10.1007/s10940-016-9335-4

CrossRef Full Text | Google Scholar

39. *de la Haye, K, Green, HD Jr, Pollard, MS, Kennedy, DP, and Tucker, JS. Befriending risky peers: factors driving Adolescents' selection of friends with similar marijuana use. J Youth Adolesc (2015) 44, 1914–1928. doi: 10.1007/s10964-014-0210-z

PubMed Abstract | CrossRef Full Text | Google Scholar

40. *Vogel, M, Rees, CE, McCuddy, T, and Carson, DC. The highs that bind: school context, social status and marijuana use. J Youth Adolesc (2015) 44, 1153–1164. doi: 10.1007/s10964-015-0254-8

CrossRef Full Text | Google Scholar

41. *Tucker, JS, de la Haye, K, Kennedy, DP, Green, HD Jr, and Pollard, MS. Peer influence on marijuana use in different types of friendships. J Adolesc Health (2014) 54, 67–73. doi: 10.1016/j.jadohealth.2013.07.025

PubMed Abstract | CrossRef Full Text | Google Scholar

42. *de la Haye, K, Green, HD Jr, Kennedy, DP, Pollard, MS, and Tucker, JS. Selection and influence mechanisms associated with marijuana initiation and use in adolescent friendship networks. J Res Adolesc (2013) 23. doi: 10.1111/jora.12018, 474–486

PubMed Abstract | CrossRef Full Text | Google Scholar

43. *Osgood, DW, Feinberg, ME, Wallace, LN, and Moody, J. Friendship group position and substance use. Addict Behav (2014) 39, 923–933. doi: 10.1016/j.addbeh.2013.12.009

PubMed Abstract | CrossRef Full Text | Google Scholar

44. *Ennett, ST, Bauman, KE, Hussong, A, Faris, R, Foshee, VA, Cai, L, et al. The peer context of adolescent substance use: findings from social network analysis. J Res Adolesc (2006) 16, 159–186, doi: 10.1111/j.1532-7795.2006.00127.x

CrossRef Full Text | Google Scholar

45. Stritzel, H. Peer and community influences on adolescent substance use in the context of adverse childhood experiences. Sociol Perspect. (2022) 65:413–32. doi: 10.1177/07311214211018718

PubMed Abstract | CrossRef Full Text | Google Scholar

46. Steinberg, L, and Monahan, KC. Age differences in resistance to peer influence. Dev Psychol. (2007) 43:1531–43. doi: 10.1037/0012-1649.43.6.1531

PubMed Abstract | CrossRef Full Text | Google Scholar

47. Henneberger, AK, Mushonga, DR, and Preston, AM. Peer influence and adolescent substance use: a systematic review of dynamic social network research. Adolescent Res Rev. (2020) 6:57–73. doi: 10.1007/s40894-019-00130-0

CrossRef Full Text | Google Scholar

48. Jacobs, W, Goodson, P, Barry, AE, and McLeroy, KR. The role of gender in Adolescents' social networks and alcohol, tobacco, and drug use: a systematic review. J Sch Health. (2016) 86:322–33. doi: 10.1111/josh.12381

PubMed Abstract | CrossRef Full Text | Google Scholar

49. Jeon, KC, and Goodson, P. US adolescents' friendship networks and health risk behaviors: a systematic review of studies using social network analysis and add health data. PeerJ. (2015) 3:e1052. doi: 10.7717/peerj.1052

PubMed Abstract | CrossRef Full Text | Google Scholar

50. Snijders, TAB, van de Bunt, GG, and Steglich, CEG. Introduction to stochastic actor-based models for network dynamics. Soc Networks. (2010) 32:44–60. doi: 10.1016/j.socnet.2009.02.004

CrossRef Full Text | Google Scholar

51. Dudovitz, RN, Perez-Aguilar, G, Kim, G, Wong, MD, and Chung, PJ. How urban youth perceive relationships among school environments, social networks, self-concept, and substance use. Acad Pediatr. (2017) 17:161–7. doi: 10.1016/j.acap.2016.10.007

PubMed Abstract | CrossRef Full Text | Google Scholar

52. Mercken, L, Snijders, TA, Steglich, C, and de Vries, H. Dynamics of adolescent friendship networks and smoking behavior: social network analyses in six European countries. Soc Sci Med. (2009) 69:1506–14. doi: 10.1016/j.socscimed.2009.08.003

CrossRef Full Text | Google Scholar

53. Mercken, L, Steglich, C, Sinclair, P, Holliday, J, and Moore, L. A longitudinal social network analysis of peer influence, peer selection, and smoking behavior among adolescents in British schools. Health Psychol. (2012) 31:450–9. doi: 10.1037/a0026876

PubMed Abstract | CrossRef Full Text | Google Scholar

54. European monitoring Centre for drugs and drug addiction. Cannabis: health and social responses (2021). Available at: https://www.emcdda.europa.eu/publications/mini-guides/cannabis-health-a

Google Scholar