emissions may change considerably if centrally planned economies become extinct. To estimate the size of any such effect, it is necessary to have studies at the national level. In addition, policy responses, particularly mitigation responses, require understanding of the activities that drive global change at the level at which the responses will be made. Depending on the topic, it may be important to conduct studies at the level of the nation-state, the community, the industry, the firm, or the individual. For studies below the global level, priorities should be set on the basis of the potential to gain understanding of the global picture or to make significant responses to global change. Thus, a high-priority study might be one that focuses on a country or activity that by itself contributes significantly to global change; or one that is expected to generalize to a sufficient number of individuals, firms, or communities to matter on a global level; or one that illuminates variables that explain important differences between actors at the chosen level of analysis. At each level of analysis, projects that meet such criteria are worthy of support, independently of what is known at the global level.
5. Important questions should be studied at different time scales. The full effects of technological and social innovations—both on society and on the natural environment—are often unrecognized for decades or centuries. The CFC case shows how the effects of human activities can look very different depending on the time scale used for analysis: a technology developed to refrigerate food had much wider global implications several decades later, after it was applied to refrigerating buildings. Such cases need to be collected so they can be studied systematically and testable hypotheses derived about what kinds of innovations are likely to acquire the social momentum that produces long-lasting and increasing effects on the global environment, such as has resulted from CFC technology or from the Brazilian development strategy used in the Amazon Basin. Theory is particularly weak for this purpose. Historians can offer convincing accounts of the current effects of changes of the distant past, but social scientists have little ability to project the effects of current changes in human systems equally into the future.
6. Research should build understanding of the links between levels of analysis and between time scales. For example, social movements mediate between individual attitudes and national policies; the interactions of individuals and firms can result in the creation of national and global markets; and national policies can stand or fall depending on whether thousands of firms or millions of individuals willingly comply. Because of these linkages, hu-