This thesis reconceptualizes prediction markets as sociotechnical assemblages rather than simple information aggregation mechanisms. Drawing on empirical analysis of 489 election contracts on Polymarket, I demonstrate how these markets operate through structured trader ecologies and hierarchical information flows that challenge conventional economic understandings. Contrary to efficiency-centered frameworks that treat prediction markets as neutral aggregators of distributed knowledge, I show how they function as distinctive social arrangements that distribute agency across human and non-human elements while creating multiple forms of value. The assemblage framework reveals how prediction markets achieve forecasting accuracy despite systematic inefficiencies, through specialized competition rather than democratic wisdom aggregation. This theoretical reframing contributes to economic sociology by illuminating new forms of economic coordination and digital sociality that emerge through blockchain-based financial platforms.