Field studies of the sick building syndrome have generally used questionnaires developed to define symptoms present over weeks and months, such as those asking ". . . do you usually . . .?" Although such studies have implicated ventilation systems as a major contributor to indoor air quality complaints, no specific exposures have been identified. An alternative approach is based on the short-term quantification of symptoms and characterization of the indoor environment with direct-reading or short-term sampling techniques. This paper summarizes the development of such a method in two studies in problem buildings and its subsequent application in a study in nonproblem buildings. The method correctly identified the postulated cause in a first building and generated a hypothesis for intervention in the second. In the nonproblem buildings, the levels of complaints appear related to the levels of VOCs, lighting, hours spent at desks, and crowding. This approach has suggested that dose-response relationships do exist in the sick building syndrome. Weaknesses of this approach include an inability to adequately address microbial characterization of indoor environments and the social and organizational predictors of complaints that are recognized to be of importance.