Error distribution for a proportion
When a sample proportion, p, is used to estimate the corresponding population probability, π, we know that the error distribution is approximately normal with mean zero and standard deviation
standard error =
However for other types of parameter estimate, the error distribution's properties are unknown. A simulation can be used to find the error distribution's properties numerically.
We will illustrate with a simulation to find the error distribution for a proportion. Of course, the formulae above should always be used so this would never be done in practice but it provides a simple illustration of the methodology .
Rice survey