later academic achievement and reduce the subsequent need for special education services. After more than 30 years of empirical study, the research literature on this issue is uneven but promising.
Beginning with the data syntheses of the Consortium for Longitudinal Studies (Lazar et al., 1982), early childhood researchers in growing numbers have looked beyond the disappointing fade-out of early IQ effects after the intervention is completed, focusing increasingly on intervention-control group differences in school performance during middle childhood and adolescence (i.e., differences in later performance between children who received the intervention during the preschool years and those who did not). This approach began with aggregated findings from 11 program evaluations reported by Lazar and his colleagues (1982), which revealed significant impacts of early intervention on both grade retention (i.e., repeating a grade) and the need for special education services, with greater differences found for those studies that had more nearly randomized research designs. More recently, the Abecedarian Project demonstrated a statistically nonsignificant trend toward less grade retention and special education at age 12, which reached significance at age 15 (Campbell and Ramey, 1994, 1995). Notwithstanding their statistical significance, however, the small magnitude of the intervention-control differences in many of these studies have led some critics to question their value (e.g., Locurto, 1991). However, since the one-time costs of repeating a grade are roughly $6,000 per year and the continuing costs of special education are approximately $8,000 per year, relatively small impacts on grade retention and especially the use of special education services can produce substantial financial benefits (Currie, 2000).
The frequently replicated finding of positive impacts of early intervention services on school performance, however, has not been universal. For example, no differences in either special education or grade retention were found in follow-up investigations of the Houston Parent-Child Development Center to age 11 (Johnson and Walker, 1991) or for the Syracuse Family Development Research Program up to age 15 (Lally et al., 1988). Researchers in the Infant Health and Development Program also found no differences in either retention or special education at age 8 (McCarton et al., 1997). However, the sample children were only in first and second grade at the time of the follow-up assessments, and the intervention-control group differences in special education placement for the Perry Preschool sample did not appear until the third grade (Weikart et al., 1978).
The interpretation of these discrepant findings is not entirely clear. Beyond obvious differences in the nature of the preschool intervention and the program participants, it is difficult to determine how much these findings are related to differences in criteria for repeating a grade or for special