. 2021 Jun 3;77(2):284–294. doi: 10.1093/geronb/gbab063

Table 2.

Conditional Growth Curve Model of Depressed Mood Over Time (N = 5,689)

	Intercept				Slope
	Estimate	95% CI, lower bound	95% CI, upper bound	Standard estimate	Estimate	95% CI, lower bound	95% CI, upper bound	Standard estimate
Fixed effects^a
	−2.279*	−2.720	−1.838	−1.643	−0.076	−0.165	0.012	−0.315
Age	−0.001	−0.006	0.005	−0.001	0.005*	0.003	0.007	0.189
Gender^b	−0.454*	−0.587	−0.321	−0.163	0.014	−0.023	0.052	0.030
Education^c	−0.111	−0.269	0.047	−0.028	−0.016	−0.065	0.032	−0.023
Ethnicity^d	0.866*	0.370	1.362	0.078	−0.081	−0.199	0.037	−0.042
Phys. limitations	0.179*	0.137	0.221	0.301	0.005	−0.005	0.014	0.017
VRF burden	0.081*	0.024	0.138	0.060	0.004	−0.014	0.022	0.017
VRF burden × Age	−0.007*	−0.013	−0.002	−0.045	0.001	−0.002	0.002	0.013
Random effects
Variance	1.618*	0.757	2.478	0.841	0.056*	0.024	0.088	0.954
Covariance	−0.085	−0.206	0.036	−0.282
Model fit
χ ² (df)					612.33 (567)
CFI					0.978
RMSEA					0.004
SRMR					0.051

Notes: ADL = activities of daily living; CFI = comparative fit index; CI = confidence interval; phys. = physical; RMSEA = root mean square error of approximation; SRMR = standardized root mean square residual; VRF = vascular risk factor.

^aThe fixed effect of the intercept in this model reflects the logit for answering the first items (“felt depressed”) them with “yes”; converting the estimates to probabilities (by $P (y_{1 i} = 1) = 1 / 1 + e x p (- i n t e r c e p t)$ , indicates that the probability of answering the first with “yes” is 1.10, accounting for all covariates. ^bReference category = female. ^cReference category = lower education. ^dReference category = White.

*p < .05.