This paper addresses shallow-water ship model testing, and the determination of water depth effects on ship resistance. It is shown how effects of water depth, towing tank width, and the scaling of viscous resistance interact, and are of large importance for ship performance prediction. A new method to correct model tests for tank width effects is briefly described. For a set of shallow-water model-test data, it is shown how tank width effects apparently double the water depth influence. Also the model-to-ship extrapolation strongly affects the predicted water depth effect for the ship, easily leading to an overestimation. The procedure to be used, with a depth-dependent form factor, is proposed. A new procedure to correct ship speed trials for water-depth effects is introduced, and validations are shown with full-scale trial data, with the same set of model-test data and with the full-scale predictions based on these.