Table 2.

Maximum likelihood estimates of dynamic random effects model for Weight, HDL and LDL of the subjects in the Framingham offspring study in the 4 exams explained by demographic, behavioral and anthropometric variables^1,2

	Dependent variable
Independent variable	Weight n = 2481		HDL n = 2481		LDL3 n = 2503
	Coefficient	SE	Coefficient	SE	Coefficient	SE
Constant	1.303*	0.136	3.6750*	0.032	2.106*	0.451
Sex	-0.050*	0.008	0.100*	0.012	-0.023	0.023
Age2	0.240*	0.018	0.035*	0.004	-0.203*	0.060
Age-squared2	-0.033*	0.002	-0.005*	0.001	0.053*	0.004
Physical activity score	-0.001*	0.0006	0.002*	0.001	0.001	0.002
Alcohol index	0.0007*	0.0002	0.007*	0.001	-0.001	0.001
Cigarettes smoked	-0.0007*	0.0001	-0.002*	0.0002	0.001*	0.0003
Height2	0.796*	0.097	0.747*	0.087	-1.291*	0.174
Weight2	-	-	-0.469*	0.030	0.461*	0.055
Lagged dependent variable2	0.495*	0.053	0.453*	0.017	0.305*	0.114
Indicator variable for Exam 3	0.012*	0.002	0.077*	0.005	0.005	0.007
Indicator variable for Exam 4	0.014*	0.003	0.032*	0.004	-	-
Chi-square statistic4, df = 4	-		20.2*		26.7*
Chi-square statistic5, df = 1	-		31.3*		57.3*

¹ Values are slope coefficients and standard errors; see equations (1) and (3) in the text for the explanation of the models. ² The dependent and the marked independent variables were in natural logarithms. ³ Only 3 time observations were available on LDL. ⁴ Chi-square test for the exogeneity of the mean over time of body weight. ⁵ Chi-square test for combining height and weight as the BMI. * P < 0.05.