Table V.

Robustness of X_M to Misspecification of k_f and k_m

Assumed kf and km	Multiple of true $k_f^{\ast }$ and $k_m^{\ast }$	Estimated Power (s.e.)
.053 and .010	$\frac{1}{4}$	.67 (.006)
.071 and .014	$\frac{1}{3}$	.66 (.006)
.106 and .020	$\frac{1}{2}$	.69 (.006)
.212 and .041	1	.72 (.006)
.423 and .082	2	.75 (.006)
.635 and .122	3	.71 (.006)
.846 and .163	4	.43 (.006)

Note. — Simulation model is II-b, with sample configuration 1. Power is assessed at significance level .0001, on the basis of 5,000 simulated replicates. $k_f^{\ast }$ and $k_m^{\ast }$ are the female and male population prevalences of the trait, while k_f and k_m are the values used in the calculation of X_M. The ${\widehat{\sigma }}_2^2$ of equation (7) is used for the calculation of X_M.