Forecasting in the Social and Natural Sciences

Social and natural scientists often are called upon to produce, or participate, in the pro duction of forecasts. This volume assembles essays that (a) describe the organizational and political context of applied forecasting, (b) review the state-of-the-art for many fore casting models and methods, and (c) discuss issues of predictability, the implications of forecaSt errors, and model construction, linkage and verification. The essays should be of particular interest to social and natural scientists concerned with forecasting large-scale systems. This project had its origins in discussions of social forecasts and forecasting method ologies initiated a few years ago by several social and natural science members of the Social Science Research Council's Committee on Social Indicators. It became appar ent in these discussions that certain similar problems were confronted in forecasting large-scale systems-be they social or natural. In response, the Committee hypothesized that much could be learned through more extended and systematic interchanges among social and natural scientists focusing on the formal methodologies applied in forecasting. To put this conjecture to the test, the Committee sponsored a conference at the National Center for Atmospheric Research in Boulder, Colorado, on June 10-13, 1984, on forecasting in the social and natural sciences. The conference was co-chaired by Committee members Kenneth C. Land and Stephen H. Schneider representing, respectively, the social and natural science mem bership of the Committee. Support for the conference was provided by a grant to the Council from the Division of Social and Economic Science of the National Science Foundation.