ABSTRACT
Counter-intuitively, the use of random numbers is an important component of many simulations. Using random numbers ensures that any structure that authors find in the simulation is either due to the mechanism they built for the simulation or is purely accidental. The workhorse of simulation is the generation of random numbers in the range from zero to one, with each possibility being equally likely. Perhaps the widest use of simulation in data analysis involves the randomness introduced by sampling, resampling, and shuffling. Other important functions for building simulations are those that generate random numbers with certain important properties. Such a simulation is easy to write: generate uniform random numbers between 0 and 60 minutes after 7:00 pm. It turns out that many such people will have a better understanding of a simulation than of theoretical concepts.
