Validation Study Methods for Estimating Odds Ratio in 2×2×𝐽 Tables When Exposure is Misclassified

Table 5

Simulation results for a 2×2×2 table when binary exposure is misclassified.

Validation sample rate	Method	Strata 1	Se11 = 0.90; Sp11 = 0.90; Se10 = 0.90; Sp10 = 0.90	Se11 = 0.85; Sp11 = 0.90; Se10 = 0.85; Sp10 = 0.90	Se11 = 0.80; Sp11 = 0.85; Se10 = 0.80; Sp10 = 0.85	Se11 = 075; Sp11 = 080; Se10 = 075; Sp10 = 0.80	Se11 = 0.90; Sp11 = 0.90; Se10 = 0.90; Sp10 = 0.90	Se11 = 0.85; Sp11 = 0.90; Se10 = 0.85; Sp10 = 0.90	Se11 = 0.80; Sp11 = 0.85; Se10 = 0.80; Sp10 = 0.85	Se11 = 0.70; Sp11 = 0.80; Se10 = 0.70; Sp10 = 0.80
		Strata 2	Se21 = 0.90; Sp21 = 0.90; Se20 = 0.90; Sp20 = 0.90	Se21 = 0.85; Sp21 = 0.90; Se20 = 0.85; Sp20 = 0.90	Se21 = 0.80; Sp21 = 0.85; Se20 = 0.80; Sp20 = 0.85	Se21 = 075; Sp21 = 080; Se20 = 075; Sp20 = 0.80	Se21 = 0.85; Sp21 = 0.85; Se20 = 0.85; Sp20 = 0.85	Se21 = 0.80; Sp21 = 0.85; Se20 = 0.80; Sp20 = 0.85	Se21 = 0.70; Sp21 = 0.80; Se20 = 0.70; Sp20 = 0.80	Se21 = 0.90; Sp21 = 0.95; Se20 = 0.90; Sp20 = 0.95
𝑛𝑣=0.2	Matrix		1.717; (0.0298)	1.711; (0.0325)	1.671; (0.0400)	1.652; (0.0483)	1.720; (0.0301)	1.713; (0.0338)	1.682; (0.0418)	1.715; (0.0375)
	Inverse matrix		1.724; (0.0232)	1.723; (0.0234)	1.732; (0.0238)	1.686; (0.0240)	1.727; (0.0218)	1.731; (0.0232)	1.634; (0.0252)	1.762; (0.0233)
	Likelihood		1.723; (0.0146)	1.725; (0.0155)	1.730; (0.0176)	1.689; (0.0182)	1.724; (0.0165)	1.730; (0.0182)	1.682; (0.0203)	1.724; (0.0187)
	Weighted likelihood		1.737; (0.0560)	1.726; (0.0554)	1.730; (0.0561)	1.685; (0.0539)	1.738; (0.0534)	1.730; (0.0532)	1.691; (0.0560)	1. 721; (0.0544)
𝑛𝑣=0.3	Matrix		1.718; (0.0247)	1.720; (0.0263)	1.721; (0.0300)	1.659; (0.0358)	1.721; (0.0272)	1.711; (0.0330)	1.694; (0.0362)	1.718; (0.0335)
	Inverse matrix		1.720; (0.0188)	1.723; (0.0194)	1.723; (0.0237)	1.678; (0.0310)	1.723; (0.0223)	1.729; (0.0240)	1.673; (0.0322)	1.728; (0.0232)
	Likelihood		1.719; (0.0140)	1.719; (0.0144)	1.716; (0.0178)	1.671; ) 0.0174(	1.711; (0.0145)	1.722; (0.0181)	1.696; (0.0181)	1.723; (0.0173)
	Weighted likelihood		1.727; (0.0372)	1.740; (0.0383)	1.745; (0.0431)	1.682; ) 0.0364(	1.721; (0.0375)	1.728; (0.0388)	1.624; (0.0443)	1.733; (0.0362)
𝑛𝑣=0.4	Matrix		1.714; (0.0214)	1.719; (0.0228)	1.713; (0.0249)	1.662; (0.0293)	1.718; (0.0224)	1.712; (0.0235)	1.697; (0.0250)	1.711; (0.0233)
	Inverse matrix		1.712; (0.0162)	1.725; (0.0167)	1.723; (0.0140)	1.684; (0.0178)	1.715; (0.0166)	1.727; (0.0165)	1.692; (0.0162)	1.729; (0.0165)
	Likelihood		1.710; (0.0132)	1.722; (0.0140)	1.716; (0.0147)	1.673; (0.0163)	1.710; (0.0131)	1.716; (0.0140)	1.688; (0.0142)	1.711; (0.0162)
	Weighted likelihood		1.726; (0.0280)	1.724; (0.0278)	1.727; (0.0275)	1.678; (0.0277)	1.729; (0.0281)	1.728; (0.0282)	1.674; (0.0275)	1.729; (0.0273)

Numbers in each cell reflect mean of adjusted odds ratios (MSE) based on 1000 simulated data sets, with true OR=1.71.