Role of personal network attributes in adoption of clean stoves among Congolese refugees in Rwanda

Praveen Kumar, Naira Kalra, and Anita Shankar

World Development, Volume 177 (2024), Article 106539


This article examines the effect of personal network attributes on the adoption of clean cookstove technology by Congolese women living in a Rwandan refugee camp. This study was conducted in the Kigeme refugee camp located in the Nyamagabe district of the southern province, about 150 km from Kigali. The camp is home to nearly 19,000 Congolese refugees.

The research was conducted as part of a randomized controlled trial (RCT) where women were offered the opportunity to receive a free clean tier 4+ cookstove along with a subscription for pellets that could be paid for by a cash for fuel program. Initially, between 2016 and 2018, the clean cookstoves were offered on a first-come first serve basis to “early adopters” who paid out of pocket. Subsequently, from 2018, clean cookstoves were financially subsidized and offered to households based on random allocation.

Network size, number of peers living inside the camp (camp homogeneity) and number of peers having already adopted a clean cookstove (clean cookstove homogeneity) were examined at baseline and at six-months follow-up. The baseline social network data were collected in September 2018 from women aged 18–45 and included socio-demographic data and information on adoption of clean cooking technology. The endline survey was conducted in March/April 2019. Out of 935 households that participated in both the baseline and endline surveys, 503 households were in the intervention group, and 432 households were in the waitlisted control group.

Main findings:

  • Early adopters of clean cookstoves in the refugee camp were more likely to have peers who also had a clean cookstove (clean cookstove homogeneity) and were less likely to have peers living inside the camp (camp homogeneity). Additionally, early adopters were more likely to have higher levels of education.
  • Clean cookstove homogeneity was significantly associated with clean cookstove adoption. Respondents who adopted tier 4+ stoves reported a higher increase in peers having tier 4+ stoves from baseline to follow-up compared to those who did not adopt tier 4+ stoves. A unit positive increase in the number of peers having tier 4+ stoves from baseline to follow-up was associated with 1.2 times increase in the odds of adopting tier 4+ stoves at follow-up, even after adjusting for demographic predictors and randomization status. There was no significant difference for network size and camp homogeneity across the two outcome groups of stove adoption.
  • There was a significant difference in the employment status of the respondents at follow-up between the two outcome groups. However, other demographic predictors such as age, education, employment status of the partner, and employment status of the respondents at the baseline did not show a significant difference.

The authors conclude that personal network attributes played a significant role in clean cookstove adoption in a humanitarian setting. They suggest that power of social networks is an untapped influence on technology adoption and should be more effectively integrated in future research and programs.