. Author manuscript; available in PMC: 2014 Oct 2.

Published in final edited form as: Eur J Epidemiol. 2012 Mar 3;27(4):267–279. doi: 10.1007/s10654-012-9666-x

Table 3.

Comparison of different model selection approaches with all the SNPs in the NPHS-II study

Gene	SNPs	Mean log of fibrinogen	Rank of P value	Conditional regression analysis	Adjusted R²	Mallows’ Cp	RMS residuals	AIC	AICc	BIC	Cross- validation	Stepwise regression using AIC	Stepwise regression using P = 0.1^a
FGB	rs7673587	−0.0066	12
	rs4508864	0.0300^****	1	X						X
	rs1800791	−0.0020	16
	rs1800790	0.0297^****	2			X	X	X	X		X	−2768.462	2.01e – 08
	rs1800788	−0.0120^*	8
	rs4220	0.0307^****	3
	rs4463047	−0.0095	11
FGA	Taq1	−0.0093^*	9
	rs2070022	−0.0031	15
	rs2070018	−0.0077	13
	rs2070016	0.0304^****	4
	rs2070011	−0.0135^****	5	X	X	X	X	X	X		X	−2773.900	0.006
	rs7659613	−0.0124^**	6
	rs7659024	−0.0094^*	10
FGG	rs1118823	−0.0042	14
FGG	rs1800792	0.0114^**	7

The best-subset models cross (X) denotes inclusion in the best model

AIC Akaike’s information criterion, BIC Bayesian information criterion; RMS root mean square; AIC given;

P value are shown;

<0.05;

^**

<0.01;

^***

<0.001;

^****

<0.0001