The purpose of this paper is to stimulate more systematic research into the link between stress and illness, by proposing a model summarizing the variables previously related to illness onset and suggesting an improved methodology for measurement of life events. Despite the long-recognized role of factors other than the presence of a disease agent in the etiology of illness, only a weak, though statistically significant, link between stressful life events and illness onset has been supported thus far. The somewhat disappointing results of previous stress research may be due to problems in measurement of both life stress and illness and/or inadequate conceptualization of a model incorporating the multitude of variables significant to illness onset.