Table 2.

Power of tests for detecting departures from multiplicative and additive models at the 5% nominal significance level

Model	Power
Multiplicative	HL test	0.09	0.07	0.12	0.08	0.28	0.12	0.57	0.27
	Windmeijer test	0.05	0.06	0.07	0.06	0.05	0.05	0.06	0.04
	Proposed method	0.13	0.08	0.22	0.12	0.47	0.20	0.86	0.44
	Proposed method	0.13	0.08	0.19	0.10	0.42	0.17	0.83	0.39
Additive	HL test	0.06	0.21	0.09	0.41	0.13	0.76	0.23	0.99
	Windmeijer test	0.13	0.96	0.20	0.99	0.35	1	0.62	1
	Proposed method	0.10	0.94	0.16	0.99	0.25	1	0.50	1
	Proposed method	0.10	0.94	0.14	0.99	0.23	1	0.45	1

The proposed method is evaluated using or with the corresponding risk regions being defined by the combinations of evenly placed grid points at various upper and lower quintiles of the risk distribution. Results are based on simulated datasets each involving a total of subjects with equal number of cases and controls. In each simulation, data are generated with models involving SNPs or SNPs .