It is well established that discussion networks have meaningful consequences for a variety of sociopolitical attitudes and behavior. In this project, we explore how social structure shapes reactions to disaster; in particular, the 2010 BP oil spill. We address the questions of how networks are relied upon following community‐wide disaster, and to what extent these networks mirror social structures in other domains. To examine these questions, we analyze data that experimentally vary the commonly employed discussion “name‐generator” questions to see if oil spill discussants are fundamentally different from important matters discussants. Relative to “important matters” discussants, we find strong support for a specialist model in response to disaster; oil spill discussants tend to be less intimate, more knowledgeable, more active, and more talkative about the oil spill. Ultimately, this suggests a contextual basis for the formation of and reliance on discussion networks.