Purpose: Asset theory suggests that assets may have influence on behaviors (Sherraden, 1991). The mediating effect of mother's expectation on the relationship between assets and child's educational outcomes was supported by Zhan and Sherraden (2003). However, there could be an alternative theory, where mother's expectation could moderate the relationship between assets and child's outcome. One missing link is how assets affect parents' behavior related to child's education. In this regard, this study examines whether the relationship between assets and mother's involvement in child's education is mediated or moderated by mother's expectation of child's education.

Methods: The data of this study is from an experimental survey of Individual Development Accounts (IDA) program, which is a matched savings program for the poor (N = 532). Multivariate ordinary least square (OLS) regression was mainly used to test the relationship. The mother's expectation of child's education (an ordinal variable) was regressed on four assets variables (homeownership, education savings, amount in savings account, and amount in checking account) and control variables using ordinal logit regression. The hierarchical regression analysis was used to assess the mediating or moderating effects of mother's expectation of child's education on the relationship between mother's assets and mother's involvement in child's education.

Findings: The results of the ordinal logit regression (c²=65.72, df=14, p<.001) and the hierarchical regression (R2=.088, F = 3.31, p<.001) indicate that mother's expectation do not mediate the relationship between mother's assets and mother's involvement in child's education. On the contrary, the interaction model (R2=.095, F = 4.45, p<.001) indicates that the relationship between mother's homeownership and mother's involvement is moderated by mother's expectation (b=.46, p<.01). In the interaction model, homeownership (b=.497, p<.05) and savings for child's education (b=.423, p<.05) were significantly related to mother's involvement in child's education. However, the other assets variables were not significantly associated with mother's involvement in child's education.

Implications: This study shows that the effects of mother's assets on their involvement are moderated by mother's expectation of child's education. However, the moderating effects are not consistent through four assets. Further study will be required to examine why and how the mechanisms differ. Moreover, asset accumulation program should work together with education programs for mothers to increase their expectation related to child's education.