and, in particular, time in center-based arrangements, they had no effect on parenting behavior, the high rates of depressive symptoms that characterized the mothers in both programs, or the home environment (as measured by the HOME scale). To the extent that positive child outcomes emerge from these experiments, they are restricted largely to school-age children.
A fourth set of studies were based on work and education strategies that did not include economic incentives, such as those included in the three experimental programs described above (Hamilton, 2000; McGroder et al., 2000). These were conducted in 11 sites of the National Evaluation of Welfare-to-Work Strategies (NEWWS). There were no consistent impacts on young children's development.
Nonexperimental research using longitudinal data has shifted from studying poverty as an unchanging status (poor versus not poor) to understanding how particular characteristics of poverty affect development for different age groups. This has focused attention on the depth, duration, and timing of poverty in childhood (Brooks-Gunn and Duncan, 1997; McLeod and Shanahan, 1993). What have we learned? Of particular importance is emerging evidence suggesting that family income may exert its most powerful influence on children during the earliest years of life (Duncan et al., 1998b). Using data from a nationally representative sample of children and families, Duncan and his colleagues related children's completed schooling to family income averaged over three age spans: 0-5, 6-10, and 11-15 years. Family income during children's preschool years, which are most distant from their decisions about leaving school, appeared far more important than income during middle childhood. Income during adolescence mattered, but primarily for entry into college. Moreover, early childhood income effects were particularly strong in the lower income ranges. Controlling for income later in childhood as well as for demographic characteristics of households, a $10,000 increment to income averaged over the first five years of life for children in low-income families was associated with a 2.8-fold increase in the odds of finishing high school. This analysis suggests that for children in families experiencing economic hardship, income in the preschool years matters more for children 's education attainment than does income later in childhood.
We have also learned that a household's long-term economic status has a much greater association with achievement and behavior problems than do single-year income measures (Blau, 1999). There appear to be larger impacts of income increments on low-income than higher-income families (Duncan et al., 1998b; Mayer, 1997; Smith et al., 1997), although this is not found consistently (Blau, 1999). Finally, although we've seen that