Skip to main content
. 2015 Apr 7;71(5):775–785. doi: 10.1093/geronb/gbv023

Table 1.

Multilevel Models Examining Blood Pressure as a Function of Chronic Stress, Negative Relationship Quality, and Gender

Model 1: Direct effects of stress and NRQ Model 2: Do direct effects vary by gender? Model 3: Moderating role of negative quality on stress–BP link Model 4: Does the moderating role vary by gender?
Systolic BP, b Systolic BP, b Systolic BP, b Systolic BP, b
Actor stress −.62* −.62* −.56* −.55*
Partner stress .44 .34 .25 .20
Actor × partner stress −.00 −.06 .01 .06
Actor NRQ −.07 .22 −.16 −.22
Partner NRQ −.22 −.64 −.33 −.48
Actor × partner NRQ 3.23*** 3.40*** 3.10** 3.15**
Actor stress × actor NRQ .05 .39
Actor stress × partner NRQ −.73 −.93
Partner stress × actor NRQ .25 .30
Partner stress × partner NRQ 1.12* .99*
Actor stress × gender −.16 −.08
Partner stress × gender .75** .80**
Actor stress × partner stress × gender .33*
Actor NRQ × gender −.25 −.30
Partner NRQ × gender .55 .26
Actor NRQ × partner NRQ × gender −.13
Actor stress × actor NRQ × gender −.22
Actor stress × partner NRQ × gender .28
Partner stress × actor NRQ × gender 1.10*
Partner stress × partner NRQ × gender −.39
−2 Log likelihood 17,680.0 17,662.7 17,669.8 17,652.5
Change in likelihood 20.4** 17.3** 10.2* 17.3*

Notes: BP = blood pressure; NRQ = negative relationship quality. Change in likelihood in Model 1 is based on comparison with a covariates-only model. Change in likelihood for Models 2 and 3 is change from Model 1. Change in likelihood of Model 4 is based on comparison with Model 3. To compare model fit across models, we restricted all four models to individuals with complete data on all predictors and covariates (N = 2,221).

*p < .05. **p < .01. ***p < .001.