A Gaussian Approximation to the Distribution of the Sample Variance for Nonnormal Populations

A Gaussian approximation to the distribution of the sample variance is developed using the Wilson-Hilferty (1931) approach. This approximation is compared with the well-known approximations due to Box (1953), Roy and Tiku (1962), and Tan and Wong (1977) by taking the exponential, Laplace, uniform, product normal, Weibull, lognormal, Tine, logistic, and various mixtures of normal distributions as parent populations. The Gaussian approximation, which can be used for probabilities as well as percentiles, compares favorably with the other three approximations.