Diagnostics through Residual Plots in Binomial Regression Addressing Chemical Species Data

Table 5

Empirical power of different residuals based on Grubb’s test statistic for outlier’s detection in binomial regression.

		RES	PRES	APRES	CERES (L)	CERES (K)
10	2	0.018	0.534	0.822	0.342	0.600
	10	0.024	0.540	0.806	0.292	0.590
	20	0.014	0.550	0.788	0.290	0.602
	95	0.016	0.568	0.794	0.310	0.582
20	2	0.000	0.684	0.876	0.352	0.830
	10	0.001	0.698	0.894	0.386	0.878
	20	0.000	0.686	0.892	0.360	0.838
	95	0.002	0.670	0.902	0.420	0.832
30	2	0.003	0.738	0.932	0.510	0.950
	10	0.004	0.746	0.920	0.502	0.940
	20	0.001	0.728	0.900	0.476	0.946
	95	0.000	0.716	0.904	0.502	0.940
40	2	0.000	0.704	0.902	0.584	0.944
	10	0.000	0.848	0.906	0.516	0.972
	20	0.005	0.862	0.904	0.500	0.976
	95	0.001	0.976	0.974	0.500	0.978