The National Instititute of Education sponsored a conference on the future of research in adolescent reasoning. This report reflects this conference discussion, which focused on: (1) the theoretical basis for research programs in the area of adolescent reasoning; (2) research programs likely to be effective in gathering information relevant to education; and (3) methodical procedures that might foster these research programs. It was found that two major issues face decision makers considering future directions for research on adolescent complex problem solving. The first concerns appropriate theoretical formulations to guide research programs which, unlike Piagetian theory, would be able to offer predictions concerning current educational issues such as transfer, task specific variance and instruction. The second concerns appropriate methodical tools, such as task analysis and multivariate analysis, which gather information to develop the theories. These methodological tools were considered valuable for understanding why and how task information is encoded and why and how conceptual schemes become available to a subject and are activated. Also, it was proposed that by analyzing the likely causes of failure on tasks studied, researchers could begin to understand how to design instruction in adolescent reasoning. (Author/RF)