From a theoretical perspective what is the difference between sampling from a statistical distribution to generate a synthetic data set versus using a Monte Carlo simulation to generate synthetic data? If I am understanding them correctly They seem like the same thing to me.