The Case for Early, Targeted Interventions to Prevent Academic Failure

Abstract

The persistent achievement gaps among children of different race/ethnicity and socioeconomic status in the United States represent an issue that has commanded public, policy, and research attention on and off for about 100 years now, and it is once again in the forefront of policy-making agendas. Debates nevertheless abound on the most promising and cost-effective strategies to address the problem. We examine critically the available evidence on the benefits and costs of early childhood education and conclude that early, vigorous interventions targeted at disadvantaged children offer the best chance to substantially reduce gaps in school readiness and increase the productivity of our educational systems. The available evidence fails to provide a complete road map for future investments, however. Hence, we propose a program of challenge grants to states and their subunits, coupled with waivers from regulation, to spur innovation and experimentation within this important research area. We provide examples of the types of experiments that could be funded and discuss important considerations in the development and implementation of such a research grants program.

Notes

^a Source: Federal Interagency Forum on Child and Family Statistics (2006) and additional detailed tables available at http://www.childstats.gov/americas-children/tables.asp.

^ba Source: Bloom and Dey (2006).

^c Source: Dye and Johnson (2006).

^d Source: U.S. Census Bureau, 2005 American Community Survey, Tables S1501 and S1701 (available at http://factfinder.census.gov).

^e Source: Brooks-Gunn and Duncan (1997).

¹We estimate gaps in the Item Response Theory scale score units provided in the Early Childhood Longitudinal Study–kindergarten Cohort data rather than the standardized units presented in Figure 1, because the standard deviations of test scores increase as children age. Use of standardized measures can thus mask changes in the achievement gaps over time.

²Children assigned to the school-age intervention received support through Grade 3 from a home/school teacher. The home/school teacher met with parents about every 2 weeks to provide supplemental educational activities to do at home and served as a general liaison between the child's regular school teacher and parents. The home/school teacher also provided support to families with employment, social services, transportation, health care, and other needs to help improve their ability to care for their children.

³The kindergarten experiences of program and comparison group children in the CPC study differed according to the structure of kindergarten programs in their neighborhoods. Some members of both groups attended part-day programs, while others attended full-day programs.

^a Source: Schweinhart et al. (2005).

^b Source: Ramey and Ramey (2004); Campbell et al. (2002); Masse and Barnett (2001); Ramey et al. (2000).

^c Source: Reynolds et al. (2001); Reynolds and Temple (1995).

⁴The Barnett, Lamy, and Jung (2005) study includes Oklahoma's pre-kindergarten program but evaluates it on a statewide basis. Gormley et al. (2005) examined the program's effects for a sample of children from Tulsa, Oklahoma, only.

⁵Cost and benefit estimates for the Perry Preschool program are in constant 2000 dollars, discounted at a rate of 3%.

⁶Cost and benefit estimates for the Abecedarian program are in constant 2002 dollars, discounted at a rate of 3%.

⁷Cost and benefit estimates for the Chicago CPC program are in constant 1998 dollars, discounted at a rate of 3%.

⁸Gorey's review includes, for example, the evaluations of the Perry Preschool, Abecedarian, and Chicago CPC programs. It does not include findings from the evaluations of state-funded pre-kindergarten programs or the Head Start impact study, because these were only recently published. His review includes 35 experimental and quasi-experimental studies.

⁹Gorey used Cohen's U ₃ statistic as the index of effect size and practical significance for his analysis. The statistic compares all the scores for members of the intervention group with the average score for the members of the control group. A U ₃ statistic of 85% is roughly equivalent to an effect size (d) of one full standard deviation.

¹⁰After accounting for program intensity, Gorey found no relationship between program duration and the persistence of program effects. However, it is unclear whether his duration variable counts only years of preschool services or also includes follow-through services provided in the elementary or later grades. Notably, evaluations of the Abecedarian program often estimate program effects according to whether participants received preschool services only (starting at 6 weeks until kindergarten entry), elementary services only (from kindergarten to second grade), or both types of services. Generally, elementary services seem important in maintaining the effects of preschool services but, on their own, yield weaker effects than preschool services (see, e.g., Ramey et al., 2000). Differences in outcomes between participants who received preschool services only versus preschool and elementary services are not always statistically significant.