The use of logit models to investigate social and biological factors in infant mortality. I: Methodology.

Infant mortality data for England and Wales, cross classified by mother's age, parity and social class, have been published on two occasions, the first giving the relevant data for 1949/50, the second for 1975, some 25 years later. Published analyses of these separate data sets have been based on graphical and tabular methods. This paper describes the statistical methodology appropriate to the use of logit models to investigate these data sets and shows how such models may be used to supplement the more informal approach. The stillbirth data for 1975 are used for illustration. Paper II, which follows, explores the stillbirth data in greater detail, with emphasis on the changes between 1949/50 and 1975. Subsequent papers describe neonatal and post-neonatal deaths using the same framework. A final paper summarizes the overall changes in infant mortality by means of a multilogit model.