Table V.

Estimated sensitivity and positive predicted value of the multinomial, ordinal, and Bernoulli models.

Ha	Sensitivity							PPV
	Multi	Ord	Ord*	Br1	Br2	Br3	Br4	Multi	Ord	Ord*	Br1	Br2	Br3	Br4
80 cases in cluster
A: p= (0.05,0.15,0.35,0.45)	0.896	0.898	0.563	0.876	0.250	0.273	0.596	0.887	0.890	0.611	0.838	0.222	0.325	0.683
B: p= (0.05,0.25,0.25,0.45)	0.854	0.877	0.518	0.876	0.069	0.057	0.625	0.838	0.863	0.579	0.835	0.101	0.128	0.668
C: p= (0.10,0.10,0.40,0.40)	0.853	0.873	0.535	0.624	0.569	0.429	0.370	0.840	0.859	0.559	0.553	0.500	0.485	0.480
D: p= (0.15,0.15,0.15,0.55)	0.837	0.861	0.487	0.313	0.304	0.314	0.878	0.864	0.883	0.494	0.293	0.243	0.239	0.898
60 cases in cluster
A: p= (0.05,0.15,0.35,0.45)	0.844	0.866	0.425	0.851	0.238	0.204	0.507	0.829	0.852	0.491	0.745	0.157	0.211	0.587
B: p= (0.05,0.25,0.25,0.45)	0.791	0.832	0.402	0.845	0.077	0.099	0.521	0.706	0.760	0.394	0.732	0.063	0.086	0.568
C: p= (0.10,0.10,0.40,0.40)	0.772	0.832	0.501	0.547	0.516	0.319	0.299	0.754	0.818	0.449	0.396	0.348	0.413	0.350
D: p= (0.15,0.15,0.15,0.55)	0.742	0.814	0.410	0.249	0.202	0.255	0.845	0.775	0.844	0.392	0.184	0.141	0.193	0.870

Multi=multinomial model, Ord=ordinal model, Ord*=ordinal model tested under unordered alternatives A′ :p=(0.45,0.05,0.35,0.25), B′ :p= (0.45,0.05,0.25,0.25), C′ :p= (0.40,0.10,0.40,0.10), D′ :p= (0.15,0.55,0.15,0.15), Br1= Bernoulli model for category 1 (versus the others), Br2= Bernoulli model for category 2 (versus the others), Br3= Bernoulli model for category 3 (versus the others), Br4= Bernoulli model for category 4 (versus the others).