POPULATION AND DEVELOPMENT REVIEW 26(4):795–819 (DECEMBER 2000) 795 NOTES AND COMMENTARY Perceiving Mortality Decline M ARK R. M ONTGOMERY THE TOPIC of mortality risk perceptions has somehow failed to engage the interest of demographers. Perhaps this is because in demography, as in so- cial science more generally, the passage from socioeconomic change to in- dividual perceptions of change has been assumed to be immediate, or to involve lags of little consequence. If individual perceptions adjust quickly to changing empirical realities, such perceptions can justifiably be ignored. In the case of mortality decline in developing countries, however, there is good reason to believe that perceptions are likely to be diffuse, are quite possibly biased upward in relation to the empirical risks, and are probably rather slow to adjust to declines in those risks. I base these propositions not on a large empirical literature in demog- raphy—there is almost no literature of this kind to cite—but on findings from cognitive and social psychology, disciplines in which probability and risk assessment have been intensively studied. A sketch of the findings from their literatures is as follows. When laypersons attempt to understand mor- tality decline, they are apt to bring to the task a great variety of rules of thumb and heuristic standards of judgment (see, among others, McKenzie 1994; Mellers, Schwartz, and Coale 1998; Nisbett and Ross 1980; Tversky and Kahneman 1974). The layperson’s understanding of probability and risk is imperfect at best, and is particularly so when the probabilities are changing or when a correct assessment requires fine discrimination among levels of risk. Individual perceptions and individual experience alone may not then suffice. Furthermore, social learning—that is, learning from the experiences of family, peers, social network partners, and mass media— may suffer from biases and uncertainties that are similar to those affecting individual perception. The usual perceptual difficulties are greatly amplified in the case of mortality by the fact that mortality decline is not fully exogenous. Rather,