Infant mortality in a Mexican-American community: Laredo, Texas, 1950-1977.

Published infant mortality rates (IMR's) for Mexican-American populations frequently are lower than expected given the socioeconomic status (SES) of these populations. It has been speculated that this is due to bias or incompleteness in Mexican-American vital statistics. In this paper an extensive genealogical data base constructed from Catholic church records and civil records for the border city of Laredo, Texas is used to study this problem. The infant mortality probabilities (IMP's) since 1950 are compared to conventional IMR's, both based strictly on the population at risk defined by baptisms, in which the deaths are a proper subset of the denominator, and these are compared with IMR's calculated in the usual way from aggregate civil records of births and infant deaths for Laredo. We find that when these data are used, the IMR's for the most recent years are about twice the conventional rates computed from registered vital statistics.