Reducing Hazardous Alcohol Use in Social Networks Using Targeted Intervention

Alcohol use is almost ubiquitous on college campuses and first-year students are at particularly high risk of alcohol-related harm when they first make the transition to college. Peers are important agents in socializing both healthy and unhealthy behaviors, but despite the clear role of peer behavior in the maintenance of college problem drinking, there have been no efforts to measure the effect of individual change on the reduction of alcohol-related risks in the broader student body. That is, despite the importance of social connections for inducing and maintaining alcohol use in youth, intervention approaches have not measured nor capitalized on the potential of social influences for changing this problem behavior. It is essential that we understand the indirect effects of individual interventions and the impact such interventions have on the social structure and social connections. The best way to evaluate such effects is to use a research design that experimentally manipulates drinking using the best available intervention and measures its effects on the social network and its members. The purpose of this research is to investigate whether using an established individual Brief Motivational Intervention (BMI) administered to a small number of influential network members embedded in a social network significantly reduces heavy drinking and alcohol consequences among close peers who do not receive any intervention. In addition, the investigators will investigate social influence mechanisms of this transmitted effect, investigate how specific types of network connections and relationships moderate the indirect intervention effect, and investigate the effects of the intervention on network position and structure. First-year students at Brown will be enrolled and assessed early in their fall 2016 academic semester. Heavy drinkers in each dormitory who are in the top quartile of betweenness centrality, a social network construct that reflects high connectivity and potential influence, will either receive BMI or serve as controls, according to their dormitory's intervention assignment. All participants will be assessed again 5 and 12 months after baseline to measure changes in behavior and in peer ties. The long-term objective of this research is to understand how peer influences function in social networks in order to leverage those mechanisms to reduce problematic alcohol use in heavy drinking populations.

The proposed design and network analytic methods will allow the investigators to investigate the extent to which an intervention conducted with careful attention to network connection is transmitted to others (Aim 1), how those effects occur (Aim 2), conditions under which those effects are more likely (Aim 3), and how the connections themselves change as a function of the intervention (Aim 4). AIM 1. To investigate the efficacy of targeted Brief Motivational Intervention for reducing heavy drinking and alcohol consequences in network members who received no intervention. The investigators expect that heavy drinking participants residing in dormitories assigned to BMI but who receive no intervention will show lower frequency of heavy drinking and alcohol consequences at follow-up than their comparison group in NHC dormitories (n = 480; 240 in each condition). AIM 2. To identify the social influence mechanisms through which the intervention effect is conveyed. The investigators expect that reductions in the following social influence mechanisms will mediate the indirect intervention effect: (1) behavioral modeling of heavy drinking; (2) social reinforcement for drinking or not drinking; (3) offers and provision of alcohol; (4) the perception of peer heavy drinking (descriptive norms); and (5) the perception of peer approval of heavy drinking (injunctive norms). AIM 3. To identify the network and relationship features that moderate intervention efficacy. 3A. Proximity to Intervention. The investigators expect that heavy drinking network members who: (1) have first-degree ties (i.e., a direct tie) with intervention recipients, (2) have a higher proportion of intervention recipients in their close network (i.e., have the highest exposure to transmitted intervention effects), and (3) have an intervention recipient as a roommate will show the strongest indirect intervention effect. 3B. Quality of Peer Relationships. The investigators expect that stronger relationships with intervention recipients as measured by: (1) best-friend status, (2) higher perceived relationship closeness, (3) higher perceived social support, and (4) reciprocated network nominations, will be related to greater indirect intervention effects. AIM 4. To investigate the intervention effect on personal-level network position and on drinking-based selection in the network. 4A. Network position. Differences between the intervention recipients in the BMI and NHC groups at follow-up will be investigated on: (1) betweenness centrality (how often the participant falls on the shortest path between two others) and (2) prestige (number of nominations of the participant by others). There is little literature on how network positions change following behavioral intervention, so directional hypotheses are not proposed. 4B. Drinking-based selection. At follow-up, the investigators expect less drinking-based selection in the BMI group compared to the NHC group.