Bridging Disciplinary Boundaries (January 11 - 14, 2007)

SEED is a national policy, practice, and research initiative designed to test the efficacy of progressively funded asset-building accounts for children and youth. There are twelve community-based organizations, and one state, operating children's savings account programs across the United States and in Puerto Rico as part of SEED. Social work researchers at three universities are conducting multi-year, multi-method SEED studies.

One SEED study is a large pre-school demonstration using a quasi-experimental design. In fall 2004, parents of pre-school children in treatment centers (n=323) and comparison centers (n=409) serving low-income families were interviewed. The treatment center parents were then offered the opportunity to open progressively funded college savings accounts for their pre-school children. The SEED program also provides financial education and support services to treatment center children and their parents.

Of the 323 parents in the treatment centers, 179 (55%) opened SEED accounts for their children, and 144 (45%) did not. In early analyses of the baseline data, we found few demographic variables associated with account opening. This is consistent with the theoretical underpinnings of the SEED initiative in that we expect asset-building outcomes to be explained primarily by institutional, rather than individual, factors.

In this paper, we report findings on the measurement of parent well-being. One of the key questions of our larger study has to do with the impact of SEED on parents and families over time. In preparation for longitudinal analysis, we developed a parental well-being measurement model that includes three latent variables (parenting stress, personal mastery, and economic strain). We use structural equation modeling to address our question: Does a parental well-being measurement model fit the data equally well for both account openers and non-openers?

We find that the measurement model was invariant, fitting the data for both account openers and non-openers with no significant differences [c2 = 737.27, 448 df, p <.001, RMSEA=.0667, NNFI = .0934, CFI = .936]. Further, tests of homogeneity of the variances and covariances of the latent constructs did not yield evidence of significant differences between account openers and non-openers on the basis of parent well-being [c2 = 722.11, 435 df, p < .001, RMSEA = .0681, NNFI = .0932, CFI = .936]. These results strongly suggest that we can measure parental well-being similarly for account openers and non-openers

Paired with our prior findings of few demographic variables that are reliably associated with account opening, the results of our measurement model tests lend support to an institutional understanding of asset-building. Social work practitioners, researchers, and policy makers who want to understand what leads some parents, but not others, to open progressively funded savings accounts for their children may need to look beyond individual level factors. It may be important to consider characteristics of the organizations that administer asset building programs, as well as the nature and structure of the savings accounts themselves.