Table 3.

Multilevel estimates for models predicting total diurnal cortisol output (AUC).

Variable	Null model		Model 1		Model 2		Model 3
	Coefficient	t	Coefficient	t	Coefficient	t	Coefficient	t
Intercept	17.134 (0.54)	31.75***	17.064 (0.52)	32.95***	17.064 (0.52)	32.95***	17.063 (0.52)	32.96***
Age			0.049 (0.06)	0.76	0.049 (0.06)	0.76	0.049 (0.06)	0.76
Gendera			−0.575 (1.06)	−0.54	−0.575 (1.06)	−0.54	−0.575 (1.06)	−0.54
Job tenure			0.041 (0.08)	0.50	0.041 (0.08)	0.50	0.041 (0.08)	0.50
Educationb			1.222 (0.69)	1.78	1.222 (0.69)	1.78	1.220 (0.69)	1.78
Smokingc			0.045 (1.25)	0.04	0.045 (1.25)	0.04	0.046 (1.25)	0.04
Exercisec			0.538 (1.25)	0.43	0.538 (1.25)	0.43	0.539 (1.25)	0.43
BMI			−0.016 (0.13)	−0.11	−0.016 (0.14)	−0.11	−0.016 (0.14)	−0.12
Negative work events					0.204 (0.53)	0.39	0.163 (0.52)	0.31
Time of waking							0.187 (0.15)	1.28
−2 log likelihood (FIML)	842.51		835.45		835.30		833.68
Δ−2 log likelihood			7.06		0.15		1.62
Number of estimated parameter	3		10		11		12

The number of daily observations was 148, nested within 62 individuals. Coefficients are unstandardized estimates of regression coefficients. Standard errors appear in parentheses. FIML, full information maximum likelihood estimation.

^aGender is coded as 0, male; 1, female.

^bEducation is coded as 1, no formal education; 2, professional training; 3, higher professional training; 4, university degree.

^cSmoking and exercise are coded as 0, no; 1, yes.

^***p < 0.001.