Perinatal mortality in a population of 403,057 births was studied with respect to age, parity, and race in an attempt to see whether an additive model could satisfactorily account for the effects of these variables. It was found that additive models, as applied to both the mortality rates and their legit transforms gave only a very rough description of the data. First-order interactions, and in particular the interaction between age and parity, were found to be important. Secondorder interactions were negligible.

The text of this article is only available as a PDF.
You do not currently have access to this content.