Table 3.

This shows the results of search model analyses. The μ parameter refers to the mean of the Gaussian from which the signal-site decision variables are sampled (the standard deviation is assumed to be unity) and λ and ν parameters are the x and y coordinates of the true end-point. In addition to LLF_α, which depends on the choice of α, the search model provides a second figure-of-merit ${\theta }_1=\theta (\mu ,\lambda ,\nu ,\stackrel{\perp }{1})$ which is independent of α. NA: confidence interval cannot be calculated due to floating point overflow.

Data set	μ	λ	ν	LLFα	θ1	p-value
CAD_A	2.00 (1.82,2.45)	2.92 (2.79, 5.79)	0.936 (0.900,0.970)	0.792 (0.761,0.807)	0.747 (0.722,0.778)	7.33E-3
CAD_B	2.06 (1.68,2.57)	4.02 (2.47, 10.7)	0.941 (0.895,0.989)	0.814 (0.781,0.840)	0.736 (0.706,0.768)	0.440
CAD_C	1.49 (0.924,2.66)	5.33 (2.57, 37.5)	0.925 (0.782,1.00)	0.560 (0.528,0.594)	0.624 (0.585,0.706)	0.226
CAD_D	2.16 (1.60,2.80)	5.96 (2.60, 17.4)	0.992 (0.937,1.00)	0.763 (0.734,0.790)	0.737 (0.696,0.783)	0.573
CAD_E	2.19 (2.02,2.55)	6.13 (6.13,9.06)	0.927 (0.886,0.958)	0.822 (0.799,0.835)	0.821 (0.796,0.847)	0.757
CAD_F	2.84 (NA)	95.6 (4.64,1096)	1.00 (NA)	0.789 (NA)	0.771 (NA)	0.210
CAD_G	2.42 (2.17,3.04)	6.13 (6.13, 14.8)	1.00 (1.00,1.00)	0.925 (0.894,0.932)	0.833 (0.793,0.877)	0.303