Key points are not available for this paper at this time.
A nonparametric strategy for correcting for unobserved heterogeneity in hazard models is described and used examine covariates of child mortality and birth interval length in Republic of Korea. An attempt is also made to discover whether specific allowance for heterogeneity alters covariate effects estimates and conclusions that would have been reached in hazard models that ignore unobserved heterogeneity. In addition authors examine the sensitivity of covariate-effect estimates imposition of a specific parametric form on hazard and sensitivity different choices of parametric hazards...and offer suggestions about actually estimating such models investigator who is concerned about correcting for heterogeneity. This is a revised version of a paper originally presented at 1983 Annual Meeting of Population Association of America (see Population Index Vol. 49 No. 3 Fall 1983 p. 420). (EXCERPT)
Trussell et al. (Tue,) studied this question.