TABLE 2.

Results for the binary outcome and risk difference (RD) estimand in the hospitalized population

n	Estimator type	Effect	P(reject H 0)	MSE	Bias	Variance	Rel. Eff.
100	Unadjusted	0.000	0.030	0.995	0.022	0.996	1.000
100	Adjusted	0.000	0.052	0.900	0.023	0.900	0.904
100	Unadjusted	−0.161	0.307	0.877	0.011	0.878	1.000
100	Adjusted	−0.161	0.420	0.791	0.009	0.792	0.902
100	Unadjusted	−0.201	0.463	0.829	0.025	0.829	1.000
100	Adjusted	−0.201	0.607	0.755	0.023	0.755	0.911
200	Unadjusted	0.000	0.038	1.006	−0.024	1.007	1.000
200	Adjusted	0.000	0.049	0.907	−0.030	0.906	0.901
200	Unadjusted	−0.147	0.527	0.917	0.002	0.918	1.000
200	Adjusted	−0.147	0.633	0.801	−0.009	0.802	0.873
200	Unadjusted	−0.201	0.821	0.864	0.010	0.865	1.000
200	Adjusted	−0.201	0.895	0.749	−0.001	0.750	0.867
500	Unadjusted	0.000	0.036	1.038	0.020	1.039	1.000
500	Adjusted	0.000	0.043	0.897	0.024	0.898	0.864
500	Unadjusted	−0.093	0.542	0.994	−0.017	0.995	1.000
500	Adjusted	−0.093	0.611	0.863	−0.012	0.863	0.868
500	Unadjusted	−0.126	0.798	0.979	−0.013	0.980	1.000
500	Adjusted	−0.126	0.862	0.850	−0.007	0.851	0.868
1000	Unadjusted	0.000	0.033	0.932	0.012	0.933	1.000
1000	Adjusted	0.000	0.038	0.829	0.019	0.829	0.889
1000	Unadjusted	−0.058	0.440	0.932	0.014	0.933	1.000
1000	Adjusted	−0.058	0.507	0.857	0.021	0.857	0.919
1000	Unadjusted	−0.091	0.837	0.898	0.012	0.899	1.000
1000	Adjusted	−0.091	0.892	0.817	0.020	0.818	0.910

BCa bootstrap is used for confidence intervals and hypothesis testing. “Effect” denotes the true estimand value; “MSE” denotes mean squared error; “Rel. Eff.” denotes relative efficiency, which we approximate as the ratio of the MSE of the estimator under consideration to the MSE of the unadjusted estimator. In each block of six rows, the first two rows involve no treatment effect and the last four rows involve a benefit from treatment. MSE and variance are scaled by n; bias is scaled by $n^{1/2}$.