Although it has been assumed that increasing maternal education or family income will improve children's well-being, considering the impact on child care arrangements and home environment raises the possibility of negative effects. This paper reviews experimental evaluations of seven programs designed to enhance welfare families' self-sufficiency, and develops a model describing the mechanisms through which these programs affect children. Variables included in the model are maternal education, family economic status, maternal subjective well-being, child care arrangements, and home environment. For each pathway variable, the review identifies specific markers that have been examined, whether program impacts have been detected, and whether differences emerge in the short- or long-term. The review finds that changes were most universally examined in type of child care used and less consistently examined in maternal well-being, movement out of poverty, and quality of child care used. Conclusions differed depending on how variables were measured and how long data were collected. Program impact was reported on earnings and AFDC receipt. In evaluations measuring both educational attainment and achievement, none showed impact on achievement. There were significant program impacts on participation in mental health services or counseling but no reported effects on depression, locus of control, or stress. Programs clearly affected children's participation in formal nonmaternal care, but one evaluation found that program participation was tentatively related to reduced quality of care. Four of the studies found that program mothers were more warm and less harsh with their children. (Contains 10 references.) (KB)