In recent years, large-scale assessments of the state of scientific knowledge have become an important part of the scientific and policy landscape. Structured assessments have played a major role in debates over, and policy decisions about, acid rain, ozone depletion, and global warming; they have also engaged large numbers of scientists and cost a great deal of money. Yet while scholars have mainly investigated how assessments influence (or fail to influence) public policy, there has been relatively little consideration of how these assessments actually work. However, if assessments are presumed to provide a robust basis for policy, then it is important to understand them. By focusing on the assessments of three major environmental problems of recent decades — acid rain, ozone depletion, and sea level rise — this symposium will analyze the internal dynamics of assessments: How do scientists assess their colleagues' research, evaluate its reliability, understand its limits and degrees of uncertainty, and come to consensus (or not)? How do scientists respond to the subtle or overt pressures that arise when they know their conclusions will be widely disseminated, beyond the community of peer experts? What factors in the process may lead to systematic bias, error, or distortion? Understanding the internal dynamics of the assessment process puts us in a better position to judge the quality of any particular assessment as well as to suggest potential means of improvement.