Table 3.

Relative treatment effects (RTE) table for HAMA from the non-parametric longitudinal factorial analysis.

	Group	Time	Count	Rank Means	RTE	Bias	Variance	Lower	Upper
1	Placebo	BLa	20	231.4	0.7697	6.00E-04	0.0152	0.7367	0.799
2	Placebo	D15	20	222.125	0.7388	−9.00E-04	0.0242	0.6972	0.7758
3	Placebo	D30	20	221.6	0.737	−4.00E-04	0.0162	0.7033	0.7677
4	Placebo	D45	20	229.425	0.7631	4.00E-04	0.0169	0.7283	0.794
5	Placebo	D60	20	236.75	0.7875	4.00E-04	0.0256	0.7435	0.8245
6	AE120	BLa	20	216.25	0.7192	4.00E-04	0.076	0.6435	0.7824
7	AE120	D15	20	141.35	0.4695	−3.00E-04	0.0413	0.4186	0.5211
8	AE120	D30	20	92.375	0.3062	0.00E+00	0.0182	0.2734	0.3416
9	AE120	D45	20	68.875	0.2279	1.00E-04	0.0218	0.1932	0.2679
10	AE120	D60	20	22.05	0.0718	−2.00E-04	0.0035	0.0594	0.0899
11	AE60	BLa	20	234.75	0.7808	2.00E-04	0.0761	0.7013	0.841
12	AE60	D15	20	148.225	0.4924	−3.00E-04	0.0266	0.4513	0.5336
13	AE60	D30	20	91.425	0.3031	0.00E+00	0.0183	0.2702	0.3385
14	AE60	D45	20	76.1	0.252	2.00E-04	0.0141	0.2234	0.2835
15	AE60	D60	20	24.8	0.081	−2.00E-04	0.005	0.066	0.1023

^aBL: Baseline; RTE tending below 0.5 means that AE60 and AE120 were able to considerably reduce the HAMA scores whereas placebo can be seen with minimal change to the RTE. The systematic deviation of the estimated treatment effect from the true treatment effect is given as the bias. How much the estimated treatment effects vary across different samples or iterations of the analysis is given by the variance.