This article discusses theoretical assumptions underlying physiological stress reactivity research. It examines early conceptualizations of activation and recovery and contrasts these with current practices in designing, analyzing, and reporting stress reactivity studies. Study protocols from four major journals covering the last 2 years of publication were examined for current practices. Of the 105 studies which tested physiological reactivity, 63% collected recovery data but only 23% reported the recovery data. We concluded that stress recovery issues are neglected and a renewed case is made for their conceptual and ecological importance. The case for studying recovery is further supported by a selective review of studies using recovery protocols that revealed positive findings not apparent in reactivity comparisons only. Finally, options for sound design of recovery protocols, statistical processing of resulting data, and interpretation of findings are presented.