ABSTRACT
As argued in Section 6.1.1, the goal of an epidemiological study is estimation of effects rather than mere assessment of the independence or nonindependence of an exposure and disease outcome. How large is the effect of prenatal care on the incidence of low birthweight babies? How effective is a vaccine in preventing a communicable disease? How strong is the association between elevated dietary fat and heart disease? We now focus on estimation of the measures of association introduced in Chapter 4 using data arising from any of the designs discussed in Chapter 5. A key component of our analysis is determination of the level of uncertainty associated with proposed estimators, expressed via calculation of confidence intervals. We again use the notation for entries in a 2 × 2 table given in Table 6.1.
