Virtual sampling from Johnson distributions

From the point of internal and external validity, in designing the robustness studies of statistical techniques, the levels of distribution dimension should span a wide diversity of non-normal distributions which may possibly be come across with in the real applications. In meeting this requirement, the most appropriate distribution family is the non-normal distributions obtained by the Johnson distributions. However, it is observed that Johnson distributions are specified as the distribution dimension in only a very few of robustness studies probably due to the fact that the random sampling functions do not readily exist in the widely used statistical programming languages. In this work, the fundamental information on Johnson distributions has been briefed, the virtual sampling algorithms from these distributions have been developed, and the reliability of these algorithms have been presented by a Monte Carlo experiment. The given algorithms can be easily turned into readily available functions of virtual sampling from the Johnson distributions in any of the programming languages by the prospective researchers who are to perform such robustness studies.