Articles suggesting the zoonotic potential of certain human diseases (eg, multiple sclerosis, rheumatoid arthritis, and leukemia) periodically appear in the literature and frequently receive considerable attention in the popular press. Although various epidemiologic study designs have been utilized to test these hypotheses, defining and accurately measuring animal exposure has been a problem common to most of these studies and in some instances has limited their usefulness. The relative strengths and limitations of the definitions and measurements used most commonly by investigators evaluating potential zoonoses are discussed. In addition, several recommendations for assessing animal exposure in future studies are made.