Each theoretical estimand is linked to an empirical (observational) estimand, which is by predicting the relationship through the data. These assumptions can be conveyed in a Directed Acyclic Graph (DAC). Eventually, one finds the empirical estimand (e.g. Regression model). Exemplifying both the theoretical estimand and the empirical estimand allows for the link between theory and evidence, and different modes of arguments. The distinction is subtle but important: the former may involve unobservable quantities while the latter involves only observable data.