Table 3.

The results from 1000 simulated case-control studies from a population where HWE and the independence between haplotypes (H) and environmental covariate (X) holds within strata defined by S. Each replicate contains 1000 cases and 1000 controls and is analyzed using (a) the proposed unconditional retrospective maximum-likelihood (RML) method that assumes HWE and H – X independence hold in the entire population; (b) the proposed conditional RML method assuming HWE and H-X independence hold conditional on S and (c) the proposed modified prospective score-equation (PSE) method

		βX	βH	βHX	βS	βHS
(a) Unconditional RML
	Bias	0.009	−0.547	−0.033	−0.102	0.755
	Empirical SE	0.059	0.290	0.064	0.305	0.264
	Estimated SE	0.062	0.303	0.063	0.318	0.281
	Covergage Prob	0.956	0.551	0.897	0.941	0.223
(b) Conditional RML
	Bias	−0.004	−0.012	0.010	0.008	0.002
	Empirical SE	0.061	0.313	0.065	0.331	0.307
	Estimated SE	0.063	0.327	0.067	0.339	0.317
	Covergage Prob	0.955	0.963	0.954	0.957	0.965
(c) Modified PSE
	Bias	0.001	−0.022	−0.003	0.010	0.029
	Empirical SE	0.070	0.326	0.075	0.337	0.302
	Estimated SE	0.069	0.350	0.076	0.349	0.330
	Covergage Prob	0.942	0.966	0.947	0.960	0.972