The Long-Term Impacts of Child Development Accounts on Parental Educational Expectations and College Preparation

Background and Objective: Parental expectations and preparation for their children’s postsecondary education have positive effects on children’s academic performance, educational attainment, and healthy development. Studies have documented that Child Development Accounts (CDAs) have positive impacts on early educational outcomes. This study investigates longer-term impacts of CDAs on parental educational expectations and college preparation.

Methods: Data are drawn from the Wave 3 survey of the SEED for Oklahoma Kids experiment (SEED OK). Parental educational expectations are measured by a 5-level Likert question “How far in school do you think that your child will go?” Based on the descriptive distribution of this variable, we create two dummy variables: a) whether the parent expected their child to go to graduate school at Wave 3 (1= graduate school and 0=below graduate school), and b) whether the parent maintained their educational expectations at the graduate school level since wave 1 (1=maintain at the graduate level and 0=otherwise). College preparation is a latent variable measured by five ordinal indicators. We asked the five questions regarding planning for the child’s education or training after high school, with responses ranging from strongly disagree to strongly agree. The key independent variable is the treatment group status. Control variables include demographic and socioeconomic characteristics of the child, mother, and household. Binary logistic regressions were used to assess parental educational expectations, and Structural Equation Models (SEM) were used to evaluate college preparation. We performed two sets of analyses using the pre-COVID sample (n = 626) and the whole sample (n =1,556). With directional hypotheses, we apply one-tailed statistical tests and set significance at the .10 level.

Results: At Wave 3, in the pre-COVID sample, mothers in the treatment group were more likely to expect their children to attend graduate school than those in the control group (OR = 1.37, p = 0.07). Furthermore, mothers in the treatment group have greater odds of maintaining the same levels of educational expectations between Wave 1 and Wave 3 (Pre-COVID: OR = 1.41, p = 0.08; Whole sample: OR = 1.27, p = 0.07). Results from the SEM models have a reasonable fit to the data [CFI = 0.97, TLI = 0.96, RMSEA = 0.05 (90% CI: 0.04–0.05)]. Moreover, we find that the CDAs have a positive effect on college preparation (pre-COVID: b = 0.09, p = 0.02; Whole sample: b = 0.04, p = 0.08). Analyses stratified by racial groups suggest that CDAs have larger effects on educational expectations for Black mothers, and greater impacts on college preparation for White mothers.

Conclusion and Implications: The study finds that CDAs have positive long-term impacts on parental educational expectations and college preparation. Over a 13-year span, mothers in the treatment group were more likely to form and maintain educational expectations and prepare for children’s college education. Findings inform social work practitioners and policy makers to incorporate SEED OK’s universal, automatic, and progressive model in asset-building policies for education.