The Future of Children, Princeton - Brookings: Providing research and analysis to promote effective policies and programs for children.

The evaluation results of the Salinas Valley and Teen PAT programs suggest that, when comparing PAT or PAT-plus-case-management groups with control groups, PAT had little effect on parenting knowledge, attitudes, or behaviors as measured in these demonstrations. Nor were significant effects noted on child health or health care. There is some evidence, however, that PAT improved children's cognitive development and perhaps their social and self-help development. The following section describes the major program evaluation results.³²

Despite the fact that increasing parents' understanding of parenting and child development was the main focus of PAT home visits, Table 2 shows that no benefit emerged in terms of parenting knowledge for the full sample in either program. Changes in parenting attitudes also were small and inconsistent in direction. In terms of parenting behavior, although the combined intervention group of the Teen PAT Demonstration had a positive and significant impact on the HOME subscale that measures acceptance of children's behavior, a negative effect of similar magnitude was noted on that same subscale in the Salinas Valley PAT group (in other words, the control group outperformed the PAT group on that subscale). The only other notable program effect, reflecting the availability of appropriate play materials in the PAT intervention group of the Teen PAT program, also was negative.³³

Given that changes in parenting knowledge, attitudes, and behaviors were the means by which improvements in child health and development were expected to occur, these small and inconsistent parenting outcomes would argue against seeing any systematic positive program benefits for children. However, Table 3 shows that small to moderate program benefits were found for some aspects of child development.

PAT children in the Salinas Valley program benefitted significantly in the area of self-help development, advancing by approximately two months over the control group. Although not statistically significant, small positive effects on social development were noted for the children in the PAT groups in both studies. In Teen PAT, the children in the combined PAT-and-case-management group experienced statistically significant gains of one or more months over the control group in cognitive development, but the children in the PAT group experienced no significant benefits in cognitive development.

In terms of child health and health care, no significant benefits were associated with PAT alone in either demonstration. In Teen PAT, the combined PAT-plus-case-management intervention was associated with significantly fewer opened cases of child abuse or neglect,³⁴ and the case management intervention was associated with a significantly higher rate of full immunization.

Multivariate analyses were conducted to further explore the impact of the PAT program. Such analyses permitted researchers to control statistically for some of the possible effects of individual, family, and household characteristics, such as age, language, and income, that might affect the outcomes, either by creating spurious differences between the PAT and control groups or by hiding real differences that might be present. Ideally, of course, random assignment would equalize the groups at the outset, and equivalence would be maintained over time, such that statistical controls would not be necessary. But, as mentioned above, the groups did differ at the outset on a few characteristics, and this form of analysis provides an additional exploration of the data.

Multivariate analyses were conducted for each parent and child outcome. For Salinas Valley PAT, characteristics held constant in the multivariate analysis were mother's education, age, ethnicity, primary language, and self-reported experience with infants at enrollment; household income; child's gender; divorce during the course of the demonstration project (which, as reported later, may have had a disturbing effect on children's development); and child care setting (center-based care compared with any other type of care). In the Teen PAT multivariate analysis, factors included mother's age, ethnicity, school attendance, pregnancy status, and self-reported experience with infants at enrollment; whether the father lived in the household; and whether the household received AFDC or welfare.

In general, the results of these multivariate analyses, summarized in Table 4, suggested that participation in PAT conferred greater benefits than were apparent in the analyses of the entire randomized group. Although the multivariate analyses did not demonstrate significant positive impacts on parenting outcomes, the small negative effects associated with PAT in the initial analyses were eliminated or reduced. Participation in PAT was associated with a statistically significant 1.3- to 1.5-month advantage in cognitive development over the control group for Salinas Valley PAT children and for Teen PAT case management and combined intervention children, as measured by the DPII (p<.05).

In the areas of social and self-help development, participation in PAT was associated with statistically significant 2.4- and 2.3-month advantages, respectively, for Salinas Valley PAT children on the DPII (p<.05). With the exception of physical development, all relationships were positive in both studies, although no others were statistically significant.

In sum, the analyses suggested that PAT alone produced modest benefits in the area of child development (especially cognitive, social, and self-help) but no measurable benefits for parents or the home environment. In the Teen PAT Demonstration, children in the combined PAT-plus-case-management group were less likely to be subjects of child abuse investigations (p<.05).