Table 6. Regression-Adjusted Program Impacts on Child Care Setting by Child Age at Random Assignment (Logistic Regressions).

Program impacts by child age	Any nonmaternal care	Any center-baseda	Any home-basedb
Under age 1c	−0.013 (0.397)	−0.452 (0.465)	0.167 (0.385)
Age 1	0.209 (0.148)	0.278 (0.161)	−0.034 (0.135)
Age 2	0.347** (0.131)	0.361** (0.118)	0.143 (0.111)
Age 3	0.409** (0.136)	0.254* (0.125)	0.220 (0.119)
Pseudo R2	.203	.139	.115
χ2
Overall model	631.03**	578.73**	565.37**
Equality of coefficientsd	1.74	2.99	2.09
N	4,492	4,489	4,511

Note. Standard errors are in parentheses. Models controlled for study-site dummies and the following baseline characteristics: follow-up length, prior earnings, prior earnings², prior Aid to Families With Dependent Children receipt, prior years of employment, high school degree, teen parent, marital status, number of children, age of youngest child in family, and race.

^aCenter-based child care includes care provided in a child care center, an extended day program at a school or community organization, and summer programs.

^bHome-based child care includes care provided by a relative, the mother's spouse or partner, or a nonrelative babysitter.

^cThis category includes children 6–11 months in age at random assignment. There were insufficient numbers of children under 6 months to include in this analysis.

^dA significant chi-square statistic for the equality of coefficients indicates a rejection of the null hypothesis that all coefficients presented in the column are equal to one another.

^*p < .05.

^**p < .01.