In this work, we investigate whether the effects of a particular intervention vary depending on network size. The intervention we consider is a randomised trial in Nepal which focuses on perinatal care. The programme, by encouraging communal engagement in maternal and infant issues amongst women, and providing basic information and resources, was found to have important effects on newborn mortality. Our research adds to the evidence by investigating whether the effects vary depending on (pre-programme) networks in the community. For this purpose, we collected data from the field on networks in around May through July 2008.