Simulation for Data Science with R

Design-based simulation

Design-based simulations are particularly important when the selection probabilities for statistical units of a finite sampling frame are not equal, that is, when samples are drawn with a complex sampling design. This primarily relates to any sampling from finite populations, for example, samples drawn from a population register.

The costs of a sample survey can be reduced if the sample is drawn with a certain complex sampling design. For example, for poverty measurement, a household with a single parent and children might be included with a higher probability than households with another composition of household members, because it's likely that the single parent household is poor (basically the target variable).

Tip

Basically, ...