Generating Random Numbers

The need for random numbers has long existed within the realm of empirical educational research, especially with regard to the random assignment of students to experimental conditions. Because of the growing popularity of simulation studies designed to investigate complex educational models, the quantity and quality of random numbers required by researchers have also increased dramatically. This article discusses the nature of random numbers, approaches to generating them with various distributional properties, tests of randomness, and other features that should be considered. Researchers are encouraged to thoroughly investigate and report on the random number generators they employ, and to avoid the substandard ones that are still widely available.