The designs of studies to evaluate and compare imaging techniques are reviewed. Thirteen principles for the design of studies of diagnostic accuracy are given. Because of the "independence principle" these studies are not able directly to evaluate the contribution of a technique to clinical management. For the latter, the "clinical value" study design is recommended. A classification of study designs is proposed in parallel with the standard classification of clinical trials. Studies of diagnostic accuracy are analogous to Phase II, whereas studies evaluating the contribution to clinical management correspond to the Phase III category. Currently the majority of published studies employ the Phase II design. More emphasis on Phase III studies is required.