The proliferation of hydrologic models suggests either a situation where-in hydrologists are homing onto the true model or a situation wherein hydrologists are uncertain about the appropriate model. The latter seems more likely in the face of constraints on data availability, budgets, time, technology and capable manpower to build models. Any formal analysis of the cost that arises must set forward goals of a model, goal specifications, and noncommensurate measures of effectiveness for judging model performance. Factors are empodied in the cost-effectiveness technique as outlined in the paper. This methodology is put forward as a logical basis for choosing models. Use of the technique is outlined in terms of forecasting water quality and quantity.