Rapid advances in digital imaging and computer-supported decision making are finding their way into orthodontists' offices. This is requiring reconsideration of the traditional role of professional clinical judgment. The emerging digital algorithms are complex, but precisely operationalized. Clinical judgment is also complex, but remains vaguely defined. This case explores the use of a common statistical technique, multiple regression analysis, for identifying what orthodontists are focusing on when making global assessments of the severity of a presenting patient's malocclusion. Two types of data were collected from 120 patients. Severity of malocclusion was expressed as the average judgment of 69 teachers of orthodontics. Potential predictor data consisted of the operational measures of physical dimensions of the same cases performed by three orthodontic residents. The physical measures included 11 components of the Peer Assessment Rating system and 12 components of the Diagnostic Index. These are widely accepted evaluation systems developed by panels of experts to identify the critical characteristics of malocclusions. Component scores in the indices were used to predict professional judgment. Multiple regression analysis determined differential weights for the physical measures. Importance was given by experts to a small number of components, whereas other components are not considered at all. The profile of important dimension of malocclusion is known to vary with the role of professionals (generalists versus specialists, for example) and with the ethnicity of the patient pool. It is apparent that "objectivity" and completeness are not sufficient criteria for characterizing clinical judgment.