In a 110-village Honduran trial, whether treated people cut or added ties depended on coverage
A randomized trial in 110 Honduran villages found that treated people dropped untreated advice contacts at low coverage and added ties at high coverage. The Nature analysis of 8,331 people says the rewiring generally lasted three years after the 22-month program ended.
The Scientist · Science desk
In low-dosage villages, 5 to 30% of households treated, treated people cut ties with untreated advice contacts. In high-dosage villages, 50 to 100% of households, treated people added ties. Rewiring generally lasted three years after the intervention ceased.
| Figure | Value | Note | Claim |
|---|---|---|---|
| Low-dosage band, at least: treated people cut ties | 5% of households | band runs 5–30% of households | 3 |
| High-dosage band, at least: treated people added ties | 50% of households | band runs 50–100% of households | 4 |
| Years rewiring generally lasted | 3 years | 7 |
What happened
- The trial varied the share of households treated from village to village and used villages where no one was treated as the main reference group.
- Treated-treated pairs persisted and formed new links while treated-untreated ties dissolved, and those pair dynamics explain most of the average change.
- The intervention delivered health information, yet it reshaped friendship and financial ties as well as health-advice connections.
Why it matters
- Who has to decide decision Anyone sizing an informational program now has coverage as a variable with social side effects, and a thin rollout is the setting where treated people cut ties.
- What it limits constraint Monitoring that tracks only how many connections each person has would miss this, because village-wide degree distributions stayed relatively stable while individual ties changed.
- Who's exposed exposure Untreated people who had been a neighbor's source of health advice are the ones cut loose when coverage is low.
The comparison group is what makes the sign flip credible. Because some villages had no treated households [2], a change in a treated person's ties can be read against villages where nothing was delivered. That rules out a drift common to every village over the program's run.
The dosage bands leave a gap. Low-dosage villages had 5-30% of households treated, and high-dosage villages had 50-100% [3][4]. That leaves 20 percentage points, from 30% to 50%, in between [10]. The abstract reports directions rather than sizes, and it does not describe villages in that band, so the point where the sign changes is unknown.
The tie-level result suggests a simple counting story. Treated-treated pairs persisted and formed new links, while treated-untreated ties dissolved [5]. At 5-30% coverage a treated person has few treated partners to build on; at 50% or more there are many. The counting is our inference. The authors' mediation analysis points to information itself: absorbing the health information raised the likelihood of tie formation and persistence [6]. In our view that finding carries less weight than the dosage comparison, because mediation is a statistical decomposition laid over the randomization.
The setting limits how far any of this travels. The authors describe small-scale traditional or rural societies in which people rely on friends and family for loans in place of banks, for security in place of police or courts, and for health care in place of doctors or clinics [12]. The villages were remote [1]. Whether a program in a town with a working clinic and a bank would rewire anything is a question this trial was not built to answer.
The authors write that many informational public-health and development initiatives "might be inadvertently rewiring social connection in locales around the world, for better or worse" [11]. This trial supplies a direction and a dose for that claim in one region. The thing it doesn't tell you is whether the new ties left anyone better off.
What to watch
- The full paper's effect sizes and any results for villages between 30% and 50% coverage, where the sign of the change is not described.
- Replication in settings with clinics and banks, where friends and family are less of a substitute for formal institutions.
- Follow-up past the three-year mark, to see whether the rewired ties persist or revert.
Clarity's read
What the record supports and how the coverage leans. The claims behind it follow.
Reality
- Evidence75
- Adoption
- Not scored
- Hype gap+10
- Incentives
- Not scored
- Confidence68
Claim ledger
Ranked by verification strength, evidence, and original report placement.
- [1]
A randomized controlled trial involving a total of 8,331 people in 110 remote Honduran villages evaluated the secondary effects of a 22-month public-health intervention on pre-existing social-network structure.
- [2]
The design varied the fraction of village households that received the intervention, and villages in which no one received the intervention were used as the primary reference group.
- [3]
In 'low-dosage' villages (in which 5-30% of households received the intervention), people who received the intervention tended to sever ties with untreated individuals to whom they previously went for health advice.
- [4]
In high-dosage villages (50-100% of households), treated individuals increased their ties.
- [5]
The average changes are explained mainly by tie dynamics: treated-treated pairs persist and form new links, whereas treated-untreated ties dissolve.
- [6]
A mediation analysis indicates that absorption of health information increases the likelihood of tie formation and persistence.
- [7]
The rewiring patterns generally lasted for three years after the intervention ceased.
- [8]
Although the intervention delivered health information, it reshaped friendship and financial ties, not just health-advice connections.
- [9]
The overall degree distribution of village-wide networks remained relatively stable despite the rewiring of individuals' connections.
- [10]
The low-dosage band ends at 30% of households and the high-dosage band starts at 50%, leaving a 20 percentage-point stretch that the abstract's two bands do not cover.
- [11]
The authors write that many public-health and development initiatives that are ostensibly of a purely informational nature 'might be inadvertently rewiring social connection in locales around the world, for better or worse.'
- [12]
The authors describe many small-scale traditional or rural societies as ones where people rely on friends and family for loans rather than banks, for security rather than police or courts, and for health care rather than doctors or clinics.
Get this ledger inside ChatGPT. Ask about any of our stories and the answer brings its figure, each claim’s verdict and the sources. Free, no account.
Sources
1 independent publisher whose own reporting we read for this story.
- nature.comEducational intervention can rewire village social networks
1 article · October 13, 2026
Topics and entities
Follow any of these and your For You feed starts watching them — no settings page required.
Topics
- Public health interventionsFollow
- Randomized controlled trialsFollow
- Social networksFollow