Adolescence is a critical life-course period for young people’s health and wellbeing (Patton et al., 2016). Many young people are exposed to psychoactive substances such as alcohol, tobacco and cannabis for the first time during this time (Donoghue et al., 2017; Hindocha et al., 2015). An estimated 1 in 5 adolescents have a diagnosable mental health problem (Green, McGinnity, Meltzer, Ford, & Goodman, 2005). Mental health problems developed during adolescence have been attributed with a tenfold greater long-term healthcare cost than those developed during adulthood (Lee et al., 2014). Mental health and subjective wellbeing have some conceptual overlap, but can be considered distinct constructs, with subjective wellbeing typically defined in terms of holistic aspects of young people’s happiness with their lives (Greenspoon & Saklofske, 1998, 2001; Keyes, 2006). While some studies indicate that subjective wellbeing is negatively correlated with mood disorders (Proctor, Linley, & Maltby, 2009) and externalizing behaviours (MacDonald, Piquero, Valois, & Zullig, 2005), recent data from the Millennium Cohort Study indicate only a weak correlation between mental health problems and subjective wellbeing in 11 year olds (Patalay & Fitzsimons, 2016). Co-morbidity of substance use with subjective wellbeing and mental health problems is however consistently demonstrated in epidemiological surveys (Lai, Cleary, Sitharthan, & Hunt, 2015), and hence, identification of common risk and protective factors is imperative in order to inform intervention approaches that simultaneously disrupt these trajectories.

According to Bronfenbrenner’s (1992) ecological systems theory, the behaviours and wellbeing of the developing person are influenced in part by biology and individual pre-dispositions, but in interaction with multiple layers of environmental influence. Within Bronfenbrenner’s model, “micro-systems” are conceptualised as those environments most proximal to the young person, which exert direct influence on health and wellbeing (e.g. schools, families and peer networks). “Meso-systems” represents the interactions between these microsystems (e.g. family-school interactions, or peer-family interactions). These are in turn nested within “exosystems” (e.g. broader education and health systems) and “macro-systems” (e.g. prevailing social and cultural norms). Finally, the “chrono-system” encompasses changes over time in young people’s interactions with, and responses to, their environments as they move through childhood and adolescence.

Influences of micro-systems such as family and friends on young people’s substance use have commonly been investigated using the Social Development Model (SDM) (Cleveland, Feinberg, Bontempo, & Greenberg, 2008). SDM hypothesises that social behaviours are learned through social interactions, which give rise to formation of attachments with lasting effect on behaviour, through supporting development of skills, norms and values (Catalano & Hawkins, 1996; Catalano et al., 2005). Attachment to others who offer opportunities for and reward prosocial behaviour protects against antisocial behaviour, while attachments to those who support and reward anti-social behaviours may increase risk behaviour (Catalano & Hawkins, 1996; Catalano et al., 2005; Schor, 1996). These relationships form much of the basis for mental health throughout adolescence and beyond (Elgar, Craig, & Trites, 2013); limited pro-social communication within families for example is linked with increased risk of mental health symptoms (Barrett, Duffy, Dadds, & Rapee, 2001; Greenberg, Domitrovich, & Bumbarger, 1999). While close relationships with family are almost universally discussed as protective for substance use, subjective wellbeing and mental health, peer relationships have been considered simultaneously as risk and protective factors. In recent analyses of data from Wales for example, young people who reported that they could count on friends reported better subjective wellbeing, though greater risk of binge drinking (Long et al., 2017).

A growing body of literature focuses on the influence of school systems on young people’s substance use and mental health, with a particular emphasis on school connectedness (Bonell et al., 2016; Markham & Aveyard, 2003). Definitions of school connectedness vary, but typically centre around student-staff cohesiveness, and relatedness to other pupils within the school setting (Markham & Aveyard, 2003; Stewart, McWhirter, Rowe, Stewart, & Patterson, 2007). Markham and Aveyard argue that structuring school social environments in a manner which minimises boundaries between teachers and pupils for example, may lead to increased engagement with norms and practices of schools, and in turn, better health and wellbeing (Markham & Aveyard, 2003; Stewart et al., 2007). The Gatehouse Project, drawing on attachment theory (Bowlby, 1998; Bowlby, 2005), focused on enhancing young people’s connectedness to school and peers, leading to improved behaviours, though not mental health outcomes (Bond et al., 2004). In a longitudinal analysis, Bond et al. (2007) found that school connectedness and connectedness to peers independently predicted mental health and substance use outcomes. More recent UK studies have replicated these associations of teacher-pupil relationships with substance use and subjective wellbeing (Long et al., 2017; Moore et al., 2017).

The impacts of these various micro-systems most likely cluster. According to attachment theory, interactions within the family environment can enable or thwart the development of capabilities to build positive relationships elsewhere (Bowlby, 2005). Young people with less supportive family environments might therefore engage less with schooling, or develop fewer good quality friendships, with cumulative effects on health and wellbeing (Bronfenbrenner, 1986). Considering independent and combined influences of these micro-systems on young people’s health and wellbeing is therefore vital in informing intervention. A recent longitudinal analysis using a global measure of social connectedness reported that high social connectedness was associated with better subjective wellbeing. While significant associations are observed across all measured domains, associations were strongest for family, followed by school, peers then community (Jose, Ryan, & Pryor, 2012).

While a growing body of studies demonstrate the importance of these micro-systems, few have moved toward understanding how interactions between them form “meso-systems” of influence on young people’s substance use and mental health. Some studies from the sociology of education have pointed to benefits of positive student-teacher interactions for educational attainment of children whose families place less emphasis on parent-child communication (Epstein, 1983). Likewise, Gorard (2010) argues that while schools cannot entirely compensate for the effects of external systems, they can provide young people with some insulation from social disadvantage, through providing opportunities for interaction with a mixed peer group and mutually respectful adult-child relationships. Few studies have tested interactions between family relationships and relationships within the school setting in predicting substance use and mental health symptoms, though it is plausible that supportive relationships with school staff may have a disproportionately beneficial influence for young people with less family support (Moore et al., 2017).

In their aforementioned longitudinal analysis, Bond et al. (2007) reported a significant interaction between school connectedness and connectedness to peers; substance use outcomes were highest among young people with low school connectedness but high connectedness to peers. Furthermore, connectedness to peers was not associated with better mental health in the absence of school connectedness. The idea that connectedness to peers may have differential impacts among subgroups of young people is supported by qualitative findings from the UK, which suggest that young people who feel disconnected from school commonly form sub-cultures whose identities are constructed around “deviant” behaviours (Fletcher & Bonell, 2013). Formalised school norms and practices may run parallel, and often in competition with, a student-led system that develops its own set of rules and regulations. Friendships may therefore serve risk or protective roles, varying according to levels of connectedness to school. Similarly, much theorisation of the interaction between family and peers in shaping young people’s behaviour has centred on the importance of risk taking in adolescent identity construction (Pound & Campbell, 2015). Where young people perceive limited closeness to family members, influence from peers may become greater than that of family members (Vitaro, Brendgen, & Tremblay, 2002). Young people who report that their parents are usually aware of their whereabouts for example are less likely to engage in substance misuse (Moore, Rothwell, & Segrott, 2010). Less positive parent-child relationships may therefore exacerbate potentially negative influences of adolescent social relationships (Martino, Ellickson, & McCaffrey, 2009).

This paper draws upon a cross-sectional school-based survey in Wales, to test the following hypotheses. First, in line with other recent studies (Jose et al., 2012), we anticipate an independent association between relationships across a range of micro-systems and young people’s health and wellbeing. More specifically, we test the hypothesis that positive family relationships, peer relationships and higher connectedness to school teachers and students will be independently associated with lowered risk of substance use and mental health symptoms, and with increased subjective wellbeing. Moving beyond a focus on individual microsystems, and toward understanding how these micro-systems combine to form meso-systems of influence (Bronfenbrenner 1992), we then test the hypotheses that i) support from teachers will play a stronger protective role for young people who report less supportive family relationships and less support from friends, ii) peer relationships will play a stronger protective role where young people also report high levels of family support.