1. Measuring norm-sensitivity and social-network position

In the MECHANISMS study protocol, we describe how we combined game-theoretic tasks (rule-following and coordination games) with friendship-network mapping in schools in the UK and Colombia to operationalise students’ sensitivity to social norms and their network positions (e.g., closeness, centrality). This mixed approach allows us to capture both the structural (who is connected to whom) and relational (how strongly they respond to norm signals) dimensions of social influence.

2. Norms, network structure, and behaviour change

Our papers explore how interventions (e.g., peer-led programmes in schools) affect change not purely by delivering content, but by interacting with peer networks and shifting perceived norms. For example, analyses examine how students with high centrality or high norm-sensitivity may serve as conduits of change, or conversely, how weaker network ties may hamper diffusion. While each paper addresses different contexts (smoking prevention, physical-activity norms, etc), a recurring finding is that the mechanism of change is often relational: network-embedded students adapt more when the intervention activates norm-signals through their peer ties.

3. Mediation and network-mechanism modelling

Several MECHANISMS papers use mediation analysis and/or network-analytics to test hypothesised chains: e.g., intervention → change in normative beliefs → friend-group norm shift → change in behaviour/intent. These underscore that shifts in perceived peer norms and peer behaviour modelling are often the intermediate steps between intervention and outcome. They also highlight that network structure (e.g., density of ties, clustering, centrality of change-agents) moderates the strength of those mediation pathways.

4. Context matters: cross-cultural, school-setting variability

A distinctive feature of the MECHANISMS work is the cross-setting design (UK and Bogotá, Colombia). This allows examination of how context (baseline prevalence of behaviour, school culture, norm strength, network structure) influences whether mechanisms get triggered. Some mechanisms that operate well in high-norm-strength contexts may fail in weaker-norm environments, and vice versa.

Implications for design and evaluation of network- and norm-based interventions:

Our findings suggest that designing effective adolescent-health interventions via social networks requires:

• Mapping the network ahead of time: identifying who has centrality, embeddedness, ties across cliques, and peers with high norm-sensitivity.
• Selecting or engaging change-agents not only on popularity but on connectivity and influence potential (both structural and relational).
• Measuring norm perceptions and network positions before and after intervention to detect mechanism activation.
• Tailoring intervention content to engage not just the individual, but peer norms and network diffusion (for instance, peer discussion, visible normative shifts).
• Accounting for contextual moderators: network structure (tight vs loose networks), baseline normative climate, cultural factors, and school dynamics.