Negotiators and decision-makers are typically embedded within a web of existing social relations. Taken as a whole, these relations define an implicit social structure which can influence the decision outcome. One aspect of this structure is founded on interpersonal affinity between parties to the negotiation, and on the recognition of status characteristics, such as domain expertise. This paper presents a methodology aimed at extracting an explicit representation of such social structures using meeting transcripts as a data source. Use of this method is demonstrated on the transcripts of U.S. Food and Drug Administration (FDA) advisory panel meeting transcripts; nevertheless, the approach presented here is extensible to other domains and requires only a meeting transcript as input. Preliminary results demonstrate that the method presented here can identify groups of decision-makers with a contextual affinity (i.e., membership in a given medical specialty or voting clique), can extract meaningful status hierarchies, and can identify differing facilitation styles among committee chairs.