Encuentro Presencial 3
TEMA 3 Trabajo en equipo
The theoretical concepts described above assume that youngsters use substances because they are influenced by similar behaviour of peers. However, network analysts studying substance use often argue on the basis of longitudinal research that similarity cannot be equated with peer influence and that adolescents may just select peers based on similar patterns of delinquent behaviour (Valente et al., 2004). As commented in chapter 3, section 3.7.4, this issue of causality is also a matter of concern when evaluating the effectiveness of SN theory which assumes that personal behaviour is influenced by peers. In fact, much of the earlier substance use related network research has been focussed on disentangling influence from selection (Hussong et al., 2001).
In this regards, a number of researchers provide support for peer influence (Bauman and Ennett, 1994; Friedman et al., 1997; Sieving et al., 2000). For example, Friedman and others (1997) from a cross sectional study of 767 drug injectors in New York city report that being connected to a large group of people who used drugs was associated with own drug use. Sieving and others (2000) examine adolescent friendships longitudinally for three years and demonstrate that over time, higher levels of friends’ drug use led to increased alcohol use. Some analysts have looked into the alcohol consumption and social networks of couples (Leonard et al., 2000; Homish and Leonard, 2008). These studies provide evidence that a partner’s drinking has an influence on one’s own drinking during the transition to marriage and through the early years of marriage (Homish and Leonard, 2008).
In contrast, other network research provides evidence of peer selection. For example, Donohew and colleagues (1999) in a longitudinal study conducted over 2 years find that individuals who were high on sensation seeking tended to select friends who were also high on sensation seeking, and were more likely to experiment with alcohol, marijuana, and other substances. Consistent with this assertion, Pearson and West (2003) report findings from three data collection points of a longitudinal study of secondary school adolescents at a Scottish school and suggest that people who became substance users transitioned from belonging to non risk- taking groups to risk taking-groups. The research on married couples also suggests that individuals select social networks which are consistent with their drinking behaviours and that of their partners’ (Leonard et al., 2000). The study explains that these social networks which are moderately stable overtime then shape one’s drinking patterns.
The majority of network research argues that both the processes of selection and influence are responsible for similarity of substance use behaviour among peers (Fisher and Bauman, 1988; Ennett and Bauman, 1994; Engels et al., 1997; Kirke, 2004; Kirke, 2006; Reifman et al., 2006; Kirke, 2009; Rosenquist et al., 2010). Kandel (1985) provides one of the earliest evidences. In a longitudinal network investigation of high school students the study demonstrates that models which included both selection and influence explained initiation into marijuana use more fully then either factor alone. Ennett and Bauman (1994) use network techniques to study the contribution of influence and selection to cigarette smoking homogeneity in adolescent peer groups. Their work comprises two rounds of data collection and assessment of friendship links and smoking behaviour of 926 8th graders and that of their friends. The results show equal contribution of selection and influence towards similarity in smoking behaviour of the participants and their peer groups. Reifman and colleagues (2006) report similar results in
relation to college students’ heavy drinking based on a 3 wave longitudinal study at a large South Western University in the US. Having found the evidence for both phenomenon of interest, the study further addresses specialized issues related to each. It reports that greater presence of network members whom the focal respondent regarded as ‘drinking buddies’ was predictive of one’s own later drinking after having controlled for potential confounds. The study also reports that changes in the overall network drinking over time appeared to be driven predominantly by the dropping and adding of network members.
A recent network study (Rosenquist et al., 2010) provides the most thorough findings in relation to selection and influence. The findings are based on a longitudinal investigation examining person to person spread of alcohol consumption in 12,067 people spanning 32 years. The study uses data from the Framingham Heart Study which is a population based longitudinal, observational cohort study initiated in 1948 to prospectively investigate risk factors for cardiovascular disease. The researchers provide evidence for both selection and interpersonal influence up to three degrees of separation (e.g. a person’s friends’ friends’ friends). Further, the study also reports that changes in alcohol consumption behaviour of an individual’s social network had a statistically significant effect on one’s subsequent consumption of alcohol. In addition, the study also reports that female contacts were significantly more likely to influence the spread of heavy alcohol consumption behaviour than male contacts.
The most comprehensive work in the domain of social networks and substance use conducted in Ireland comes from a series of articles by Deirdre Kirke based on a dataset collected in 1987. Kirke’s work involves total coverage of all adolescents (298) aged 14-18 in one DED
(District Electoral Division) in Dublin County Borough (Kirke, 2004; Kirke, 2006; Kirke, 2009). Kirke (2004) demonstrates from three case studies that youngsters adjacent to each other in peer networks were likely to be similar in their substance use and to form chains of users of similar substances which she refers to as ‘chain reactions’. Her work shows that the similarity in substance use among adolescents occurred because of both selection and influence. Kirke (2006, 2009) further explain how chain reactions drew together those who were similar with those who were not and resulted in similarity of substance use among teenagers and their peer groups. The studies also elaborate that gender played an important role in all aspects of chain reactions which involved peer selection, the patterning of peer ties and peer influence. Specifically, the findings indicate that teenagers generally selected same sexed peers, males clustered into large more dense networks than females and male teenagers who were substance users were predominantly influenced by other males while females were predominantly influenced by both males and females.
Other network studies focus on HIV risk and prevention among drug users. Several studies focus primarily on sexual and needle sharing networks and identify risk factors associated with network characteristics (Curtis et al., 1995; Latkin et al., 1995b; Morris et al., 1995; Friedman et al., 1997). Network characteristics that have been found to shape injection risk and behaviour include network size and its cohesiveness (Latkin et al., 1995a; Latkin et al., 2003). Network research also suggests that Greek membership and frequency of discussing particular alcohol related issues are positively related to drinking in excess among college students (Dorsey et al., 1999b).
To date, most network research related to substance use has collected sociocentric data, a (near) complete enumeration of the population of interest as described in section 4.3. This research as is evident from the above discussion has been conducted in both academic and non academic settings.