MONTE-CARLO GENERATION OF S-B AND S-U DISTRIBUTIONS

Generating random variables from specific bounded or unbounded distrib utions is a problem frequently encountered in Monte Carlo studies, Of particular interest is that of transforming variables from a standard normal distribution to a given Johnson's S-B Or S-U distribution with specified (gamma, delta) parameters. We use a composite standard norma l generator N (0,1) to generate Johnson's S-B or S-U distributions wit h (gamma, delta) parameters. Goodness-of-fit tests and graphical illus trations demonstrate the adequacy of the empirical distributions.