Abstract

The objective in environmental epidemiology is to quantify the association between an environmental exposure and a health outcome measure, in a population under study, with the intention of generalizing to the population at large or to sub-groups of interest. The standard approach is to fit a regression model with health outcomes y_i and corresponding exposures x_i for subjects i = 1,…,n with additional health model covariates z _i = (z_{i 1},…,z_ip ) ∈ ℝ ^p , including an intercept. For example with a continuous outcome we might fit the linear model, 25.1 y i   =   x i β + z i β z +   ∈ i ,