Table 3.

Comparison of nested model statistics

Covariance Models		−2LL	D.F.	AIC	BIC	Difference χ2	p-value
BMI × IL-6	ACE	39.22	405	−770.77	−922.81
	CE	41.52	406	−770.48	−924.00	2.29	0.13
	AE	40.69	406	−771.31	−924.41	1.46	0.23
BMI × sIL-6r	ACE	−216.23	405	−1026.23	−1050.54
	CE	−215.63	406	−1027.63	−1052.57	0.60	0.44
	AE	−215.99	406	−1027.99	−1052.75	0.25	0.62
BMI × CRP	ACE	1396.80	401	594.80	−232.80
	CE	1398.10	402	594.10	−234.47	1.31	0.25
	AE	1396.81	402	592.81	−235.12	0.01	0.91
CRP × IL-6	ACE	1415.15	401	613.15	−223.63
	CE	1415.46	402	611.46	−225.79	0.31	0.57
	AE	1415.48	402	611.47	−225.78	0.33	0.57

−2LL (−2 * Log-Likelihood), maximum likelihood estimate; AIC, Akaike’s Information Criterion; BIC, Bayesian Information Criterion; Lower -2LL, AIC and BIC indicate models that more parsimoniously fit the data; Non-significance in chi-square (χ²) values between saturated and nested models indicate equivalent fit of data; Bold highlighted font indicates the model that most parsimoniously fit the data, based on AIC and BIC.