Asymptotic power calculations: Description, examples, computer code

This paper describes two asymptotic methods for sample size and power calcu lation for hypothesis testing. Both methods assume that the distribution of the likelihood ratio is approximately distributed as a central chi(2) dist ribution under the null hypothesis and as a non-central chi(2) under the al ternative hypothesis. The approximation to the non-centrality parameter dif fers between the methods. It is shown how these methods can be automaticall y extended from constraints setting parameters to constant values to constr aints positing equality of parameters. Two very simple examples are present ed; one demonstrates that the information method can produce arbitrarily in correct results. Four more comprehensive examples are then discussed. In ad dition to demonstrating the wide range of applicability of these methods, t he examples illustrate techniques that may be used in cases in which there is insufficient initial information available to perform a realistic calcul ation. The availability of a computer implementation of these methods in S- plus is announced, as are routines for computing the cumulative distributio n function of the non-central chi(2) and its inverse. Copyright (C) 1999 Jo hn Wiley & Sons, Ltd.