Data are of value in the design and operation of water resource projects because they provide information about uncertain states of nature. This information can in turn be used to reduce the risks inherent in the project. In this paper, risk calculations from the preposterior analysis of statistical decision theory are extended to cover nonoptimal decisions. Risk is measured in terms of the project objectives by embedding the data network into an information-response (IR) system. The value of addition or improvement to the data base is measured by the change in the risk. Measures of the performance of the IR system are developed. From these measures, the efficiency of the information subsystem, the response subsystem, and the overall IR system are defined. These concepts are illustrated by a simple example in which the additional data sample for estimating the 6¼-year flow is shown to have different values (negative in one case), depending on the estimation method or decision rule used. Five anomalies and paradoxes in the literature concerning the worth of data are explained.