These deal with how to best handle data to not lose the detail. G. van Belle sees the public health implications of a set of risk factors would favor additive models since rates and changes in rates can be directly calculated in an additive model. However, these days, sophisticated computer programs can do the more complicated calculations in a multplicative model.

Walter (2000) found that there is little numerical difference among the common methods of comparing risks: risk ratio, odds ratio, and risk difference. He suggests that risk difference is easier to understand than odds ratio.