Association for Public Policy Analysis & Management

In line with research identifying the family as the key context influencing children’s development (Kreppner & Lerner, 2013), school districts and policymakers have sought to engage parents in children’s learning, particularly low-income families (Booth & Dunn, 2013). Increasing parental engagement has been a focal point of No Child Left Behind and Race to the Top (Robinson & Harris, 2014). Such interest has prompted developers of some school-based interventions to integrate programming targeted at low-income parents. Yet, little is known about the factors that are predictive of low-income parents’ participation in services (Wilson, 2012). Similarly, little research has examined whether parent participation in program services translates into gains for children. With limited resources for parent services, it is critical to determine whether they do in fact benefit children.

This paper examines the parenting component of one early childhood program called INSIGHTS. INSIGHTS is a manualized curriculum for teachers, students, and parents that aims to improve children’s non-cognitive outcomes in order to improve their academic skills. Results from a cluster-randomized trial revealed that INSIGHTS improved low-income children’s achievement, attention, and behaviors. Leveraging data from this randomized trial, the current study first determines the key factors that predict parent take-up of program services. The paper then determines whether program impacts differed by levels of parent participation.

Participants included 435 children and parents from 22 low-income urban schools. Eighty-seven percent of children qualified for free lunch. Seventy-five percent of children were black; 25% were Hispanic or biracial. Most parents had low levels of education. Schools were randomly assigned to participate in INSIGHTS or the control condition. The intervention was implemented across two years of programming.

Given bimodal parent participation, analyses compared outcomes for high dosage parents (> 80% of the intervention) versus low dosage parents (< 80% of intervention). Logistic regressions revealed that high-dosage parents were more likely to be older, married, and working full-time. High-dosage parents had about a year more education than low-dosage parents. Children of high-dosage parents had higher levels of baseline math and reading skills.  

Inverse probability of treatment weighting was then used to determine intervention impacts for children of high vs. low dosage parents. This method allows one to examine a moderator measured post-treatment by comparing the treatment group members with high dosage to control groups members predicted to be high-dosage had they been given the opportunity to participate. This analysis was then repeated for low-dosage parents. Results revealed that children of high-dosage parents showed small impacts on math and reading achievement. Children of low-dosage parents, however, showed larger impacts on math and reading achievement, attention, and behaviors. Results suggest larger impacts across multiple outcomes for children whose parents participated at lower levels.

Due to demographic risks and lower baseline academic skills, it appears that children of low-dosage parents had more to gain from intervention. Implications regarding cost for engaging higher-need parents in services, and the intensity of services that may be necessary to promote behavior change will be discussed.