To enhance the efficient utilization of health providers, medical care researchers are increasingly concerned with the measurement of task performance. One job analysis methodology that is widely used is the task inventory technique, characterized by retrospective self-reporting. The purpose of this study was to validate the task inventory method by using the observational technique. Two data collection instruments were developed, a Task Inventory and an Observer Check List; both contained the same set of task statements. Trained observers shadowed 13 primary care physicians in their offices for one week recording the frequency and duration of task performance on the Observer Check Lists. At the end of the week, the 13 physicians completed a Task Inventory indicating how often they had performed each task in their offices and how long it normally took. The observers' and physicians' respective sets of data were analyzed to determine the degree of agreement. The results indicate that the agreement is not sufficient between the task data obtained by the two different methods to conclude that one is substitutable for the other. The validity of the Task Inventory method is therefore not supported by the results.