In the 1990s a wide array of non-governmental certification initiatives emerged as a way to promote the sustainable management of resources in sectors such as fisheries and forestry. In this paper, we examine two related questions about these initiatives: how does the institutional design of certification initiatives affect the way science is used in the development of certification standards and in whose interest is science employed? Using the empirical case of forest certification and the specific standards various initiatives have created to address the management of forest genetic resources, we show how structural aspects of decision-making processes affects the standards adopted and the rationalization for their appropriateness. Two basic models of decision-making-stakeholder participation and technical expertise-are discussed in relation to three certification initiatives active in North America-the Canadian Standards Association, the Sustainable Forestry Initiative and the Forest Stewardship Council. By examining the standards these initiatives set for the management of forest genetic resources, we illustrate how two dimensions of science-uncertainty and the logic of cause and effect-are used to rationalize cautious and rigid versus flexible and discretionary standards for the management of forest genetic resources. Our findings indicate that the design or structure of certification decision-making processes, and their embedded balance of authority, mediate the form of standards initiatives will justify on the basis of science.