The present study investigated whether bonding social capital in the form of connectedness and networks within neighbourhoods (neighbourhood social capital) and bridging social capital in the form of general social trust were associated with substance use in a large Swedish adolescent population. We found that both low neighbourhood social capital and low general social trust were associated with higher alcohol consumption and higher prevalence of smoking and illicit drug use. The results remained after adjusting for possible confounding from demographic and socioeconomic factors, thus demonstrating robust findings of the models.
Low neighbourhood social capital was associated with an approximately 60% increased odds of high alcohol consumption, more than three times increased odds of smoking and more than double the odds of having used illicit drugs. Low general social trust was associated with an approximately 50% increased odds of high alcohol consumption and double the odds of smoking and having used illicit drugs.
It seems plausible that low levels of bonding social capital within the neighbourhood might be associated with a lack of social control, lower community reinforcements of desired behaviours, and lower tendency to conform to the norms of society, including norms and laws prohibiting underage drinking, smoking, and drug use. Low bonding social capital of the neighbourhood may also be more prevalent in socioeconomically deprived areas, with higher prevalence of social problems, youth delinquency and availability of illegal drugs. Moreover, low levels of bridging social capital in the form of general social trust might result in a weaker identification with the society and the norms of society and less tendency to conform to these norms, including norms prohibiting underage drinking, smoking, and drug use. It may also involve feelings of insecurity and alienation which could lead to higher stress levels and antisocial coping behaviours including substance use.
The results are in accordance with previous findings in adult populations [22, 23, 25–27] as well as previous findings regarding alcohol consumption , smoking and illicit drug use [34, 39] among adolescents. However, our finding of an increased risk of high alcohol consumption among adolescents with lower levels of general social trust differed from the previous finding by Lundborg . This may be explained by differences in alcohol measures, the use of one community setting in the Lundborg study compared with the ten community/whole county setting of the present study, and the larger population sample of the present study (n > 7000). The weaker results found for alcohol consumption in relation to neighbourhood social capital in comparison with smoking and illicit drug use may also be explained by the fact that alcohol consumption was more frequently occurring in the adolescent population than smoking and illicit drug use. Moreover, it is possible that alcohol consumption may be more socially accepted than smoking and illicit drug use among Swedish adolescents considering the substantially higher proportion of alcohol users versus tobacco and illicit drug users . This could involve less controlling and regulating effects of neighbourhood social capital on alcohol use. This possible explanation would be interesting to investigate in future studies.
At the contextual level, weak relations were found. In the adjusted two-level model, housing area social trust in relation to alcohol consumption showed decreased odds of 1–3%, whereas an increased odds of 1–3% was found in relation to illicit drug use. Contextual level neighbourhood social capital was related to illicit drug use with increased odds of 1–3%. In previous contextual level analyses of social trust within schools Takakura found a 30% increased odds for smoking among students in schools with lower levels of trust although the association was not statistically significant . Our findings of contextual level estimates in relation to alcohol consumption, smoking, and illicit drug use were distinctly weaker. Our significant findings are probably a consequence of the larger population sample which elevated the power in the present analyses. However, an interpretation of these results might also be that contextual level effects and social processes within groups have smaller effects in Sweden than in Japan. Sweden is known for its highly egalitarian and individualistic society. Swedes have been suggested to have a strong need for social autonomy and to not be dependent on other individuals, such as neighbours, relatives, employers, etc. . This could be an explanation for the weak contextual level cluster effects of housing area in the present study. Moreover, it is interesting that even in a highly egalitarian country such as Sweden there is an association between individual level neighbourhood social capital, general social trust and substance use in an adolescent population. Social capital has been suggested as one important mediating factor of the relations between income inequality and ill health [9, 57–59] and substance use is strongly related to public health [11–13].
The results of the present study should be interpreted in the light of several limitations. Firstly, the debate regarding the concepts of social capital and social trust and how they are supposed to be measured is not settled. Social capital is operationalized in many different ways depending on which theoretical framework the researcher uses. The different schools of sociology often employ either a consensus or a conflict perspective, or a macro or micro perspective, which always attract criticism from the opposing side . For example, Putnam [16, 17] describes social capital by using a functionalist, consensual perspective, whereas Bourdieu  defines social capital as more based on conflict and exploitation. In the present study, we aimed to determine whether two forms of bonding and bridging social capital  were related to substance use among adolescents. Our measures are corresponding to Putnam’s description of social capital [16, 17]. However, this way of using the concept of it can always be criticised depending on whether a consensus, conflict, macro, or micro perspective is used .
Secondly, the analyses are based mainly on self-reports, which involves a risk of information bias due to false or inaccurate responses from the participants. Individuals are often inclined to underreport behaviours that are not approved of in society and/or against cultural norms. Thus, the underreporting of alcohol consumption and illicit drug use are well-known problems within the research field of substance use and misuse when using self-reports [61–63]. The prevalence of adolescent substance use corresponded to previous national studies in Sweden . However, the AUDIT-C has only been validated in the oldest age group of the present study . The measures of smoking and illicit drug use were not validated. To our knowledge, there is no validated gold standard for the measurement of social capital. Regarding the validity of the social capital measures, we have considered the construct validity, face validity, and content validity of the measurements as suggested by Harpham et al. . Moreover, this study design always involves a risk of confounding from unmeasured variables, for example parental education, peer attitudes to substance use, group pressure, etc. However, we have adjusted for several other well-known and previously suggested confounders in our models .
Thirdly, there is the problem of causality regarding neighbourhood social capital, general social trust and substance use, as the cross-sectional design of the study involved no possibility to distinguish the directions of cause and effect. Although the directions of causality implied in the models of the study, where low neighbourhood social capital and general social trust were associated with higher levels of substance use, seems plausible, substance use may also alter and influence levels of neighbourhood social capital and general social trust. The data moreover did not allow analyses of how an individual’s neighbourhood referred to in the neighbourhood social capital measurement was related to school and housing area.
Lastly, adolescents as a group may be particularly affected by social capital of the neighbourhood as a cause of limited mobility . Limited mobility can also be related to the outcome, e g underage drinking if alcohol outlets are not dense in the residential area. However, in Sweden there are no legal ways for adolescents to obtain alcohol, due to the alcohol monopoly of the Swedish government not permitting individuals under the age of 20 to purchase alcohol. Therefore, limited mobility in regard to alcohol outlets would not be of any great influence in this study.
The study also has several strengths, particularly regarding the large adolescent community sample from a county that is considered to be fairly representative of Sweden as a whole, because of its distribution of educational, income, and employment levels as well as urban and rural areas . The rather high participation rate reduces the risk of selection bias. The results may thus be possible to generalize to other adolescent populations.