This paper is concerned with measurement of the benefits of scientific and technical information. The first part discusses the theoretical problems of benefit measurement in this field, and the second part reviews critically a number of recent cost studies.