Factors reflecting current experience-for example, number of medications us
ed, poor function, negative affects, and positive affect-had stronger assoc
iations with self-assessments of health (SAH) than factors related to prior
events (medical history, prior smoking) in baseline data from 830 elderly
retirement-community residents (mean age = 73). Participants appear to have
implicit knowledge of the factors affecting their SAH: The rank order of t
he beta weights relating factors to SAH was correlated with the rank order
of participant ratings of the attention given to each factor when making SA
H judgments. Relationships of SAH and each of the factors to five year mort
ality showed that subjectively salient factors such as function and lack of
energy predicted five-year mortality, reduced the relationship of SAH to m
ortality, and accounted for most of the relationship of medical factors to
mortality. Affective variables, however, had no relationship to mortality d
espite their impact on SAH.