Table 4.

Unstandardized Coefficients From Regressions Predicting Changes in Adolescents’ Closeness to Biological Mothers by Family Structure Trajectory

	Change in Closenessa		Change in Being Closeb		Decline to Not Close (vs. Remain Close)c
Predictor	B	SE B	B	SE B	B	SE B
Married stepfatherd	.01a	.08	.01(a)	.05	.20a	.45
Cohabiting stepfatherd	−.19a+	.10	−.13(a)*	.06	1.12a**	.36
Adolescent age	.02	.01	.01*	.01	−.02	.09
Male adolescent	.01	.04	−.01	.02	−.08	.28
Black (vs. White)	.07	.06	.04	.02	.46	.31
Hispanic (vs. White)	.13	.11	.03	.04	−.26	.48
Other (vs. White)	−.07	.15	−.05	.08	.96	.54
Immigrant	−.01	.09	−.02	.04	−.08	.44
Marital birth	−.01	.05	.01	.02	−.25	.25
Income	−.10**	.03	−.04*	.01	.58***	.15
Mother education	.01	.02	.01	.01	−.02	.09
Father education	.01	.02	.01	.01	−.28*	.12
Years since lived with father	−.01	.01	−.01	.01	.04	.03
Child support paid	.04	.05	−.01	.02	.14	.27
R2		.02		.03
Fe		1.88*		2.17*		2.99***

Note: All values are weighted. Coefficients within a column that share subscripts do not significantly differ from one another at p < .05. Coefficients within a column that share subscripts within parentheses differ at p < .10. N = 1753.

^aOrdinary least-squares regression; Wave 2 closeness (1 to 5) - wave 1 closeness (1 to 5), where 1 = not at all close and 5 = extremely close.

^bOrdinary least-squares regression; Wave 2 close (0, 1) - wave 1 close (0, 1), where 0 = not at all, not very, or somewhat close and 1 = quite or extremely close.

^cMultinomial logistic regression, where decline to not close = quite or extremely close at W1 but not at all, not very, or somewhat close at W2, and remain close = quite or extremely close at W1 and W2.

^dOmitted category is stable single mother.

^eDesign-based F.

⁺p = .06.

^*p < .05.

^**p < .01.

^***p < .001.