This article begins by reviewing laboratory studies of stress and affiliation. This review is not exhaustive, but rather focuses on tests of the emotional similarity hypothesis, which predicts that people faced with novel, physically threatening situations will prefer to affiliate with someone of relatively similar emotional status. Purportedly, such selective affiliation is motivated by the desire to determine the appropriateness of one’s emotional response to the threat. Research involving affiliation responses to a physically threatening medical procedure (viz., surgery) is then discussed, which challenges the generality of the laboratory conclusions and at the same time extends stress and affiliation research into a more applied domain. This field-setting work is used primarily as a vehicle for illustrating some of the potential pitfalls for both theory development and clinical application of relying exclusively on laboratory demonstrations of health-related phenomena.