Table 4:

Simulation results from 10,000 MC runs when the true model is an upper hinge logistic model. Mean estimate, %bias, coverage, and MC SD of the parameter estimate are given for both the segmented and upper hinge models. The ratio of the MC standard deviation of the parameter estimates from the two models fit is given. The ratio of the MC interquartile range of the parameter estimates from the two model fits is also given.

		Segmented Model			Upper Hinge Model
	n	Est(%Bias)	Coverage	MC SD	Est(%Bias)	Coverage	MC SD	SD Ratio	IQR Ratio
${\mathit{\beta }}_{\mathit{z}}$
	50	1.39 (39)	0.989	0.68	1.29 (29)	0.985	0.57	1.19	1.12
	100	1.14 (14)	0.966	0.30	1.11 (11)	0.968	0.28	1.07	1.06
	500	1.02 (2)	0.952	0.09	1.02 (2)	0.953	0.09	1.01	1.01
	2000	1.00 (0)	0.951	0.05	1.00 (0)	0.951	0.05	1.00	1.00
${\beta }_{(x-e)-}$
	50	1.58 (58)	0.963	2.19	1.97 (97)	0.976	1.87	1.17	1.06
	100	1.51 (51)	0.939	1.69	1.45 (45)	0.949	1.23	1.38	1.21
	500	1.14 (14)	0.915	0.54	1.04 (4)	0.920	0.22	2.48	1.14
	2000	1.02 (2)	0.928	0.10	1.01 (1)	0.933	0.10	1.05	1.01
$e$
	50	3.05	0.923	1.44	2.99	0.919	1.28	1.12	1.23
	100	3.01	0.894	1.45	3.02	0.911	1.06	1.37	1.54
	500	3.01	0.908	0.74	3.02	0.931	0.43	1.72	1.52
	2000	3.00	0.934	0.28	3.00	0.938	0.20	1.43	1.37