Table 2.

Frequentist analysis of the diabetes data [21]. The entry in row i and column j gives one minus the one-sided p-value of rejecting the null hypothesis that the treatment in the row (i) is worse than the treatment in the column (j) in favor of superiority of i compared to j

	acar	benf	metf	migl	piog	plac	rosi	sita	sulf	vild
acar	–	0.62	0.13	0.37	0.18	1.00	0.07	0.74	0.95	0.63
benf	0.38	–	0.11	0.28	0.13	0.99	0.05	0.64	0.80	0.53
metf	0.87	0.89	–	0.74	0.50	1.00	0.26	0.93	1.00	0.87
migl	0.63	0.72	0.26	–	0.29	1.00	0.14	0.82	0.94	0.72
piog	0.82	0.87	0.50	0.71	–	1.00	0.32	0.91	0.99	0.85
plac	0.00	0.01	0.00	0.00	0.00	–	0.00	0.05	0.04	0.02
rosi	0.93	0.95	0.74	0.86	0.68	1.00	–	0.96	1.00	0.92
sita	0.26	0.36	0.07	0.18	0.09	0.95	0.04	–	0.64	0.40
sulf	0.05	0.20	0.00	0.06	0.01	0.96	0.00	0.36	–	0.25
vild	0.37	0.47	0.13	0.28	0.15	0.98	0.08	0.60	0.75	–

Abbreviations: acar acarbose, benf benfluorex, metf metformin, migl miglitol, piog pioglitazone, plac placebo, rosi rosiglitazone, sita sitagliptin, sulf sulfonylurea alone, vild vildagliptin