Skip to main content
. 2016 Sep 24;74(2):318–328. doi: 10.1093/geronb/gbw112

Table 2.

Multilevel Models Examining WC as a Function of Chronic Stress Exposure, NMQ, and Gender (N = 2,042)

Predictor variables Model 1: direct effects of stress and NMQ Model 2: moderating role of NMQ on stress–WC link Model 3: do the effects vary by gender?
b (SE) b (SE) b (SE)
Actor stress .00 (0.04) .00 (0.05) .01 (0.05)
Partner stress .11 (0.05)* .10 (0.05)* .12 (0.05)*
Actor × partner stress −.03 (0.03) −.03 (0.03) −.02 (0.03)
Actor NMQ .02 (0.10) .01 (0.11) .02 (0.11)
Partner NMQ −.05 (0.10) −.06 (0.10) −.03 (0.11)
Actor × partner NMQ .12 (0.16) .09 (0.18) .06 (0.18)
Actor stress × actor NMQ .02 (0.07) .04 (0.07)
Actor stress × partner NMQ −.02 (0.08) −.07 (0.08)
Partner stress × actor NMQ .02 (0.08) .02 (0.09)
Partner stress × partner NMQ .06 (0.07) .13 (0.08)
Actor stress × gender .01 (0.05)
Partner stress × gender −.02 (0.05)
Actor stress × partner stress × gender −.03 (0.03)
Actor NMQ × gender −.16 (0.12)
Partner NMQ × gender .26 (0.12)*
Actor NMQ × partner NMQ × gender .18 (0.13)
Actor stress × actor NMQ × gender .04 (0.07)
Actor stress × partner NMQ × gender −.09 (0.08)
Partner stress × actor NMQ × gender −.17 (0.08)*
Partner stress × partner NMQ × gender .24 (0.07)**
−2 Log likelihood 9,909.2 9,908.1 9,883.8
Change in likelihood 6.9 1.1 24.3**

Notes: NMQ = negative marital quality; WC = waist circumference. Change in likelihood in Model 1 is based on comparison with a covariate-only model (9,916.1). Change in likelihood for Models 2 is change from Model 1 and for Model 3 is change from Model 2. Models controlled for Wave 1 waist circumference (of both actor and partner), education, number of years married, marital order, number of children, race, gender, and age.

*p < .05, **p < .01.