Although laboratory stress research is a popular and vibrant area of resear
ch activity, there is surprisingly little evidence that laboratory stress m
odels are clinically useful (i.e. that they can explain and predict the dev
elopment of disease). This article summarizes evidence that the usefulness
of lab stress research can be improved with the use of social stressors. Tw
o lines of evidence are presented in support of this argument: (a) studies
comparing physiological reactivity to different lab stressors with ambulato
ry activity, and (b) a meta-analysis of investigations of cortisol response
s to laboratory stressors. Further issues of importance in understanding so
cial stressors are gender differences and the vulnerability (i.e. weak reli
ability) of social stressor impact to relatively small changes in the exper
imental protocol itself.