Table 2.

Multilevel Models Predicting Negative Marital Quality in Wave 2 as a Function of Drinking Status in Wave 1

	Model 1		Model 2
	b	SE	b	SE
Intercept	1.928***	0.015	1.932***	0.015
Negative marital quality Wave 1	0.608***	0.012	0.609***	0.012
Gender (Wife)	0.022**	0.007	0.029***	0.008
Education	−0.001	0.003	−0.001	0.003
Length of marriage	−0.001	0.001	−0.001	0.001
Race (White)	−0.013	0.014	−0.013	0.014
Age	0.001	0.001	0.001	0.001
Actor drinking	−0.008	0.008	−0.003	0.009
Partner drinking	−0.011	0.008	−0.015	0.008
Actor drinking × Partner drinking	−0.025**	0.009	−0.029**	0.009
Gender × Actor drinking			0.011	0.009
Gender × Partner drinking			−0.004	0.009
Gender × Actor drinking × Partner drinking			−0.016*	0.008
−2 Log likelihood	8,023.6		8,018.6
Proportion of between variance accounted for	.70		.70
Proportion of within variance accounted for	.23		.23
Change in likelihood	13.6**		18.6**

Notes: All models were adjusted for the weighted data. Changes in likelihood are based on comparison with a covariates-only model (−2 log likelihood = 8,037.2). Estimates are unstandardized. Variance accounted for is determined in comparison with the unrestricted model.

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