Monte Carlo simulation of the power of a test for the exponential distribution of survival times

An illustrative analysis is presented demonstrating the use of the Monte Carlo method in estimating the power of a test to detect deviations from distributional form. New estimates of power are obtained for the WE₀ test, a statistical test used to assess a set of observed survival times for deviation from the exponential distribution. An IBM PC/AT was used to process a series of simulations, in which samples of survival times were drawn from a series of populations whose survival was distributed according to the gamma distribution. Each sample was then tested for difference from the exponential to see how often the WE₀ test would correctly classify the sample as drawn from a non-exponential population.