Table 3:

Difference in number of positive houses in the top 30% of probabilities of infestation using our approach of map distortion compared to using the true map with no distortion (S = 1). Model fit with true values of S = 1, S = 2.5, and S = 4 when a randomly selected one-third of points were observed. Intercept fixed at β₀ = −5. 100 simulated datasets were run at each value. We report number of positive houses in the top 30% of probabilities which were treated as unobserved (ie. no gain for houses that were observed as positive).

κ	${\mathit{\sigma }}_{\mathit{u}}^{\mathbf{2}}$	S	$\widehat{\mathit{S}}$	+ houses under $\widehat{\mathit{S}}$	+ houses under S = 1	Difference
0.005	20	4	4.4	316	302	14
		2.5	3.1	324	319	5
		1	1.1	315.5	316.5	−1
	10	4	4.3	230	216	14
		2.5	3.4	245	241	4
		1	1.1	252	252	0